Insider Ownership, Shareholder Structures and Corporate Governance [1 ed.] 9783896444295, 9783896734297

Citation preview

Entrepreneurial and Financial Studies Hrsg.: Prof. Dr. Dr. Ann-Kristin Achleitner Prof. Dr. Christoph Kaserer

Benjamin Moldenhauer

Insider Ownership, Shareholder Structures and Corporate Governance

Verlag Wissenschaft & Praxis

Insider Ownership, Shareholder Structures and Corporate Governance

Entrepreneurial and Financial Studies

Herausgegeben von Prof. Dr. Dr. Ann-Kristin Achleitner Prof. Dr. Christoph Kaserer

Band 9

Entrepreneurial and Financial Studies

Benjamin Moldenhauer

Insider Ownership, Shareholder Structures and Corporate Governance

Verlag Wissenschaft & Praxis

Bibliografische Information der Deutschen Bibliothek Die Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über http://dnb.ddb.de abrufbar.

ISBN 978-3-89673-429-7 © Verlag Wissenschaft & Praxis Dr. Brauner GmbH 2007 D-75447 Sternenfels, Nußbaumweg 6 Tel. 07045/930093 Fax 07045/930094

Alle Rechte vorbehalten Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlages unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Printed in Germany

Preface

The question whether a relationship between insider ownership — i.e. direct shareholdings of managers in their firms — and corporate performance exists increasingly attracts attention in academia as well as in investment practice. The introduction of the German Entrepreneurial Index (GEX) through Germany’s leading stock exchange operator Deutsche B¨orse AG along with its remarkable performance may have contributed to this development as well. The concept of this new stock index is based on both theoretical and empirical research. Departing from the owner-management conflict, which has already been extensively discussed in literature, the quintessence of these works is that management’s incentives can be considerably improved, if managers also benefit from the success of their firms through direct shareholdings. The work on hand is an important contribution to this research strand especially as most of the relevant empirical studies refer to Anglo-Saxon capital markets. This fact may not be neglected in the comparison of corresponding empirical studies as insider ownership in these countries is mainly driven by share-based compensation schemes. Furthermore, these countries do not have a notable tradition of midsize family owned companies, as they are common in Germany and other Continental European countries. Therefore, this study investigates the relationship between insider ownership and corporate performance on the basis of a new and unique data set. From a methodical point of view, this data set has the advantage that econometrical biases may be less severe than in comparable Anglo-Saxon studies as the determinants of insider ownership are rather exogenous for the case of Germany. The results of this study are very interesting from both an academic and practical perspective. They provide evidence that the development of investment strategies should not disregard companies’ ownership structures. Finally, this work constitutes the first comprehensive overview on the development of shareholder structures of German listed stock corporations. M¨unchen, 21 January 2007 Prof. Dr. Christoph Kaserer

Acknowledgements

My debts of gratitude to people and institutions who helped me with the completion of this book, which results from my work at the Center for Entrepreneurial and Financial Studies (CEFS) at the Technische Universit¨at M¨unchen (TUM), are enormous. First of all, I am deeply indebted to my doctoral thesis supervisor Professor Christoph Kaserer. He provided me with both trustful academic freedom and thoughtful guidance which enabled me to complete this work. Furthermore, working with him for three years was intellectually and personally enriching as well as exciting which made my working experience at the CEFS exceptional. Similarly, I owe tremendous thanks to Professor Ann-Kristin Achleitner for both acting as referee for my dissertation thesis and making my working experience at the CEFS even more valuable. With her intellectual ingenuity which she willingly shared, as well as her outstanding personal care, she helped me to get great enjoyment out of my work. I am also grateful to Professor Rainer Kolisch for assuming the chair of my doctoral examination committee and enabling the speedy completion of the process. I owe a special debt to several colleagues and friends which supported me over the last three years. Firstly, Martin Brixner was my CEFS colleague from the first hour and made the experience of building up and working at the Center an exciting and enjoyable effort. His helpfulness remains unmatched and without his enduring support this thesis would never have been finished on LaTeX as it is. Special thanks for their probing questions and tedious reviewing are to Markus Ampenberger, Benedikt Kormaier, Felix Moldenhauer, Oliver Kl¨ockner, Peter Heister, Christopher Angelo and Bes Achi. I have also benefited from the discussions with and experience of all my colleagues at the CEFS which are not limited to Dr. Christian Diller, Dr. Christian Fingerle, Annabel Geidner, Dr. Thorsten Groth, Kay M¨uller, Dr. Eva Nathusius, PD Dr. Niklas Wagner and Dr. Simon Wahl.

I was fortunate to enjoy the help of several excellent research assistants: Marcus Hampl, Marcus Duttler and Viktoria Hajas performed meticulous efforts in the process of collecting and structuring shareholder data. Their work was a great help and they contributed to the pleasant working environment at the CEFS. And my thanks as well to Tina Kaltenecker and Monika Paul for helping in so many aspects of the daily work. I must acknowledge explicitly the Deutsche B¨orse AG for the co-operation in the German Entrepreneurial Index (GEX) project which initially triggered my interest to examine insider ownership on the German capital market. Insights from this practical project contributed to this academic work. Furthermore, the help in providing stock exchange data was of enormous help. Without the emotional support of family and friends this project would never have been finished. When I acknowledge that Karin supported me in all thinkable aspects throughout the times of writing this thesis I can only hint at her contribution. I also want to thank my brothers, family and friends for moral support in completing the dissertation. Finally, I would like to thank my parents very much who supported me throughout my education in a unique and outstanding manner. This work is dedicated to them. Even with all this help and support I am certain that flaws remain. They are the author’s responsibility alone. Frankfurt-on-Main, 29 January, 2007 Dr. Benjamin Moldenhauer

IX

Table of Contents

Table of Contents

Table of Contents

IX

List of Figures

XV

List of Tables

XVII

List of Abbreviations

XXI

1 Corporate Governance and the Role of Insider Ownership 1.1 Why Do Corporate Governance and Insider Ownership Matter? . 1.2 Aims of the Study . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Structure of the Analysis . . . . . . . . . . . . . . . . . . . . . 2 Fundamentals of Corporate Governance 2.1 Definition and Understanding of Corporate Governance . 2.2 Corporate Governance Mechanisms . . . . . . . . . . . 2.2.1 Business Objective and Shareholder Value . . . . 2.2.2 Conflicts between Shareholders and Management 2.2.3 Mechanisms to Control Management . . . . . . . 2.3 Basic Types of Corporate Governance Systems . . . . . 2.3.1 The Capital Market Oriented System . . . . . . . 2.3.2 The Corporate Law Oriented System . . . . . . . 2.3.3 Which Corporate Governance System is Best? . . 2.4 The German System of Corporate Governance . . . . . . 2.4.1 Historical Development . . . . . . . . . . . . . . 2.4.2 Status Quo . . . . . . . . . . . . . . . . . . . . 2.4.3 Characteristics and Idiosyncrasies . . . . . . . . 2.4.4 Trends and Outlook . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

1 1 6 8 11 11 14 14 16 17 22 23 25 26 28 28 33 36 38

X

Table of Contents

3 Causes and Consequences of Shareholder Structures in Theory and Practice of Corporate Governance 3.1 The Separation of Ownership and Control . . . . . . . . . . . 3.2 Constitution of the German Stock Corporation . . . . . . . . . 3.2.1 Management Board . . . . . . . . . . . . . . . . . . . 3.2.2 Supervisory Board . . . . . . . . . . . . . . . . . . . 3.2.3 Shareholders’ Meeting . . . . . . . . . . . . . . . . . 3.3 Shareholders’ Objectives and Control Incentives . . . . . . . . 3.3.1 Systematization of Shareholders’ Interests . . . . . . . 3.3.2 Shareholder Identity as Differentiating Factor . . . . . 3.3.2.1 Insider Individuals . . . . . . . . . . . . . . 3.3.2.1.1 Members of the Management Board 3.3.2.1.2 Members of the Supervisory Board 3.3.2.1.3 Former Members of the Boards . . 3.3.2.2 Non-financial Companies . . . . . . . . . . 3.3.2.2.1 Business Group . . . . . . . . . . 3.3.2.2.2 Corporations . . . . . . . . . . . . 3.3.2.3 Financial Institutions . . . . . . . . . . . . . 3.3.2.3.1 Banks . . . . . . . . . . . . . . . . 3.3.2.3.2 Insurance Companies . . . . . . . 3.3.2.3.3 Investment Companies . . . . . . . 3.3.2.3.4 Domestic Institutional Investors . . 3.3.2.3.5 Foreign Institutional Investors . . . 3.3.2.4 Federal, State and Municipal Government . . 3.3.2.5 Miscellaneous . . . . . . . . . . . . . . . . 3.3.2.5.1 Outsider Individuals . . . . . . . . 3.3.2.5.2 Employees . . . . . . . . . . . . . 3.3.2.5.3 Others and Treasury Shares . . . . 3.3.3 Shareholder Concentration as Differentiating Factor . . 3.3.3.1 Minority Shareholders . . . . . . . . . . . . 3.3.3.1.1 Smallholders . . . . . . . . . . . . 3.3.3.1.2 Simple Blockholders . . . . . . . . 3.3.3.1.3 Dominating Blockholders . . . . . 3.3.3.2 Majority Shareholders . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43 43 47 48 49 51 53 53 55 56 56 58 60 61 61 62 63 63 65 66 66 69 70 70 70 71 72 73 74 74 76 78 79

XI

Table of Contents

3.4

3.3.3.2.1

Simple Majority Blockholders . . . . 79

3.3.3.2.2

Super Majority Blockholders . . . . 80

Shareholders, Management and Conflicts of Interest . . . . . . . 81 3.4.1

Introductory Remarks . . . . . . . . . . . . . . . . . . . 81

3.4.2

Systematical Analysis of Potential Conflicts of Interest . 82

3.4.3

3.4.2.1

Between Shareholders and Management . . . 82

3.4.2.2

Among Shareholders . . . . . . . . . . . . . . 86

Ways and Means for Overcoming Conflicts of Interest . 87

4 Insider Ownership and its Effects on Corporate Performance 4.1

4.2

91

Insider Ownership According to Different Theoretical Views . . 91 4.1.1

The Neoclassical View . . . . . . . . . . . . . . . . . . 91

4.1.2

The Institutional View . . . . . . . . . . . . . . . . . . 93 4.1.2.1

Introductory Remarks . . . . . . . . . . . . . 93

4.1.2.2

Transaction Cost Theory . . . . . . . . . . . . 95

4.1.2.3

Property Rights Theory . . . . . . . . . . . . 96

4.1.2.4

Principal Agent Theory . . . . . . . . . . . . 98

Empirical Evidence on Insider Ownership and Performance . . . 102 4.2.1

Determinants and Effects of Insider Ownership . . . . . 102

4.2.2

Overview of Empirical Studies on Insider Ownership . . 104

4.2.3

4.2.4

4.2.2.1

Descriptive Studies on the German Market . . 106

4.2.2.2

Studies on Insider Ownership and Performance 110

Selected Studies on Insider Ownership and Performance 120 4.2.3.1

Results from Capital Market Oriented Systems 120

4.2.3.2

Results from Corporate Law Oriented Systems 123 4.2.3.2.1

Germany . . . . . . . . . . . . . . . 123

4.2.3.2.2

Other Nations . . . . . . . . . . . . 125

Shortcomings of German Research . . . . . . . . . . . . 127

5 Hypotheses on Causes and Effects of Insider Ownership

129

5.1

Determinants of Insider Ownership . . . . . . . . . . . . . . . . 129

5.2

Insider Ownership and Corporate Control . . . . . . . . . . . . 131

5.3

Insider Ownership and Corporate Performance . . . . . . . . . . 137

XII

Table of Contents

6 Empirical Analysis of Insider Ownership in German Public Stock Corporations 6.1 Study Design . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Sample and Data . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1 Sample Selection . . . . . . . . . . . . . . . . . . . . 6.2.2 Data and Sources . . . . . . . . . . . . . . . . . . . . 6.3 Definition of Variables . . . . . . . . . . . . . . . . . . . . . 6.3.1 Shareholder Structure Variables . . . . . . . . . . . . 6.3.1.1 Introductory Remarks . . . . . . . . . . . . 6.3.1.2 Shareholder Identity . . . . . . . . . . . . . 6.3.1.3 Shareholder Concentration . . . . . . . . . . 6.3.2 Performance Variables . . . . . . . . . . . . . . . . . 6.3.3 Control and Other Variables . . . . . . . . . . . . . . 6.4 Descriptive Results . . . . . . . . . . . . . . . . . . . . . . . 6.4.1 Shareholder Structure Variables . . . . . . . . . . . . 6.4.1.1 Shareholder Identity . . . . . . . . . . . . . 6.4.1.1.1 Control Rights . . . . . . . . . . . 6.4.1.1.2 Cash Flow Rights . . . . . . . . . 6.4.1.1.3 Control and Cash Flow Rights . . . 6.4.1.1.4 Ultimate and Direct Ownership . . 6.4.1.2 Shareholder Concentration . . . . . . . . . . 6.4.1.2.1 Largest Shareholder . . . . . . . . 6.4.1.2.2 Blockholders . . . . . . . . . . . . 6.4.1.2.3 Number of Blockholders . . . . . . 6.4.2 Performance Variables . . . . . . . . . . . . . . . . . 6.4.3 Control and other Variables . . . . . . . . . . . . . . . 6.5 Overview of Hypotheses . . . . . . . . . . . . . . . . . . . . 6.6 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 Results I: Determinants of Insider Ownership . . . . . . . . . 6.7.1 Simple Regression Analysis . . . . . . . . . . . . . . 6.7.2 Difference in Means . . . . . . . . . . . . . . . . . . 6.7.3 Multiple Regression Analysis . . . . . . . . . . . . . . 6.7.4 Discussion of Results . . . . . . . . . . . . . . . . . . 6.8 Results II: Insider Ownership and Corporate Control . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

143 143 146 146 150 156 156 156 157 161 162 166 170 170 170 170 181 182 185 187 188 191 195 198 200 204 206 208 208 209 212 213 215

XIII

Table of Contents 6.8.1 Simple Regression Analysis . . . . . . . . . . . . . 6.8.2 Correlation Analysis . . . . . . . . . . . . . . . . . 6.8.3 Difference in Means . . . . . . . . . . . . . . . . . 6.8.4 Discussion of Results . . . . . . . . . . . . . . . . . 6.9 Results III: Insider Ownership and Corporate Performance . 6.9.1 Correlation Analysis . . . . . . . . . . . . . . . . . 6.9.2 Difference in Means . . . . . . . . . . . . . . . . . 6.9.3 Multiple Regression Analysis . . . . . . . . . . . . . 6.9.3.1 Base Case (Cross Section 2003) . . . . . . 6.9.3.2 Base Case Extended (Cross Section 2003) . 6.9.3.3 Base Case (Cross Section 1998) . . . . . . 6.9.3.4 Base Case (Cross Section 1993) . . . . . . 6.9.3.5 Base Case (Pooled Sample) . . . . . . . . 6.9.3.6 Models of Dynamic Ownership . . . . . . 6.9.4 Robustness Tests . . . . . . . . . . . . . . . . . . . 6.9.4.1 Alternative Performance Measures . . . . 6.9.4.2 Endogeneity . . . . . . . . . . . . . . . . 6.9.4.3 Multi-Collinearity . . . . . . . . . . . . . 6.9.5 Discussion of Results . . . . . . . . . . . . . . . . . 6.10 Limitations of the Empirical Study . . . . . . . . . . . . . . 7 Conclusion and Implications 7.1 Conclusion . . . . . . . . . . . 7.2 Implications . . . . . . . . . . . 7.2.1 For Research and Theory 7.2.2 For Practice . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . . . . . . . . . . . .

. . . .

. . . . . . . . . . . . . . . . . . . .

215 217 219 222 224 224 227 229 229 234 238 240 242 244 246 246 251 255 256 261

. . . .

265 265 266 266 270

Appendices A Exemplary Shareholder Structure Classification

277

B Results I: Determinants of Insider Ownership

279

C Results II: Insider Ownership and Corporate Control

283

D Results III: Insider Ownership and Corporate Performance

289

XIV

Table of Contents

E Results III: Insider Ownership and Corporate Performance Robustness

309

F Exemplary Ownership Filings in the U.S. and Germany

315

References

317

XV

List of Figures

List of Figures

Figure 1-1: Chart of the German Entrepreneurial Index (GEX) . . . Figure 1-2: Structure of the Analysis . . . . . . . . . . . . . . . . .

7 9

Figure 2-1: Overview of Control Mechanisms . . . . . . . . . . . . 18 Figure 3-1: Shareholders’ Set of Objectives . . . . . . . . . . . . . 54 Figure 3-2: Shareholders’ Concentration Scheme . . . . . . . . . . 74 Figure 6-1: Figure 6-2: Figure 6-3: Figure 6-4: Figure 6-5: Figure 6-6: Figure 6-7: Figure 6-8: Figure 6-9: Figure 6-10: Figure 6-11: Figure 6-12: Figure 6-13: Figure 6-14: Figure 6-15: Figure 6-16: Figure 6-17:

Sample Composition 1993-2003 . . . . . . . . . . . Control Rights 1993-2003 . . . . . . . . . . . . . . Control Rights 1993-2003 (MGMT) . . . . . . . . . Control Rights in Industries 2003 (MGMT) . . . . . Control Rights 1993-2003 (CORP) . . . . . . . . . . Control Rights 1993-2003 (FINC) . . . . . . . . . . Control Rights 1993-2003 (MISC) . . . . . . . . . . Control Rights Weighted by Market Value 1993-2003 Control Rights in Size Groups 1993-2003 (MGMT) . Cash Flow Rights 1993-2003 . . . . . . . . . . . . . Excess Control Rights 1993-2003 . . . . . . . . . . Excess Control Rights 1993-2003 (MGMT) . . . . . Direct Ownership Stakes 1993-2003 . . . . . . . . . Ultimate and Direct Control Rights 1993-2003 . . . Largest Shareholder’s Identity 1993-2003 . . . . . . Overview of Methodology . . . . . . . . . . . . . . Changes of Insider Ownership 1993-2003 . . . . . .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

150 172 174 175 176 177 179 180 181 182 183 184 186 187 189 207 252

Figure F-1: SEC Filing of Beneficial Ownership (U.S.) . . . . . . . 315 Figure F-2: BaFin Database of Directors’ Dealings (Germany) . . . 316

XVII

List of Tables

List of Tables

Table 2-1: Corporate Governance Initiatives 1998-2002 . . . . . . . 32 Table 2-2: Corporate Governance Initiatives 2003-2006 . . . . . . . 34 Table 3-1: Determinants of Control Incentives . . . . . . . . . . . . 57 Table 3-2: Systematization of Equity Agency Conflicts . . . . . . . 83 Table 4-1: Empirical Studies on German Shareholder Structures . . 106 Table 4-2: Empirical Studies on Insider Ownership and Performance 110 Table 6-1: Table 6-2: Table 6-3: Table 6-4: Table 6-5: Table 6-6: Table 6-7: Table 6-8: Table 6-9: Table 6-10: Table 6-11: Table 6-12: Table 6-13: Table 6-14: Table 6-15: Table 6-16: Table 6-17: Table 6-18: Table 6-19: Table 6-20:

Listings at the Frankfurt Stock Exchange 2005 . . . . Sample Selection 1993-2003 . . . . . . . . . . . . . Hoppenstedt Shareholder Structure Data 1993-2003 . Board Characteristics 1993-2003 . . . . . . . . . . . Shareholder Identity Classification Scheme . . . . . . Shareholder Concentration Classification Scheme . . Overview of Control and Other Variables . . . . . . . Industry Classification Scheme . . . . . . . . . . . . Incidences of Control Enhancing Devices 1993-2003 . Overview of Largest Shareholder Identity 1993-2003 . Overview of Blockholder Identity 1993-2003 . . . . . Blockholder Identity and Types 1993 . . . . . . . . . Blockholder Identity and Types 1998 . . . . . . . . . Blockholder Identity and Types 2003 . . . . . . . . . Descriptive Statistics of Performance Variables . . . . Descriptive Statistics of Control Variables I . . . . . . Descriptive Statistics of Control Variables II . . . . . Descriptive Statistics of Industry Split . . . . . . . . . Overview of Hypotheses . . . . . . . . . . . . . . . . Results I: Simple Regression . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

147 148 154 155 159 162 168 169 185 190 192 194 196 197 200 202 203 204 205 209

XVIII

List of Tables

Table 6-21: Table 6-22: Table 6-23: Table 6-24: Table 6-25: Table 6-26: Table 6-27: Table 6-28: Table 6-29: Table 6-30: Table 6-31: Table 6-32: Table 6-33: Table 6-34: Table 6-35: Table 6-36: Table 6-37: Table 6-38: Table 6-39: Table 6-40: Table 6-41: Table 6-42:

Results I: Difference in Means I . . . . . . . . . . . . . Results I: Difference in Means II . . . . . . . . . . . . Results I: Multiple Regression . . . . . . . . . . . . . . Results I: Review of Hypotheses . . . . . . . . . . . . Results II: Simple Regression . . . . . . . . . . . . . . Results II: Correlation Analysis . . . . . . . . . . . . . Results II: Difference in Means . . . . . . . . . . . . . Results II: Review of Hypotheses . . . . . . . . . . . . Results III: Correlation Analysis I . . . . . . . . . . . . Results III: Correlation Analysis II . . . . . . . . . . . Results III: Difference in Means . . . . . . . . . . . . . Results III: Multiple Regression - Base Case (2003) . . Results III: Multiple Regression - Base Case Extended . Results III: Multiple Regression - Base Case (1998) . . Results III: Multiple Regression - Base Case (1993) . . Results III: Multiple Regression - Base Case (Pooled) . Results III: Multiple Regression - Dynamic Ownership . Results III: Descriptive Statistics of APM . . . . . . . . Results III: Multiple Regression - APM . . . . . . . . . Results III: Multiple Regression - Endogeneity . . . . . Results III: Multiple Regression - Multi-Collinearity . . Results III: Review of Hypotheses . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . .

210 211 212 214 216 218 220 222 225 226 228 232 236 239 241 243 245 249 250 253 257 259

Table B-1: Table B-2: Table B-3: Table B-4:

R I: Simple Regression . . R I: Difference in Means I . R I: Difference in Means II R I: Multiple Regression . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

279 280 280 281

Table C-1: Table C-2: Table C-3: Table C-4: Table C-5: Table C-6:

R II: Simple Regression I . . . . . R II: Simple Regression II . . . . . R II: Simple Regression III . . . . R II: Correlation Analysis . . . . . R II: Difference in Means (T-Test) . R II: Difference in Means (U-Test)

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

283 284 285 286 287 288

. . . .

. . . .

. . . .

Table D-1: R III: Correlation Analysis I . . . . . . . . . . . . . . . . 290

XIX

List of Tables

Table D-2: R III: Correlation Analysis II . . . . . . . . . . . . . . . 291 Table D-3: R III: Difference in Means (T-Test) . . . . . . . . . . . . 292 Table D-4: R III: Difference in Means (U-Test) . . . . . . . . . . . . 293 Table D-5: R III: Multiple Regression - Base Case (2003) I . . . . . 294 Table D-6: R III: Multiple Regression - Base Case (2003) II . . . . . 295 Table D-7: R III: Multiple Regression - Base Case Extended (2003) I 296 Table D-8: R III: Multiple Regression - Base Case Extended (2003) II 297 Table D-9: R III: Multiple Regression - Base Case (1998) I . . . . . 298 Table D-10: R III: Multiple Regression - Base Case (1998) II . . . . . 299 Table D-11: R III: Multiple Regression - Base Case (1993) I . . . . . 300 Table D-12: R III: Multiple Regression - Base Case (1993) II . . . . . 301 Table D-13: R III: Multiple Regression - Base Case (Pooled) I . . . . 302 Table D-14: R III: Multiple Regression - Base Case (Pooled) II . . . . 303 Table D-15: R III: Multiple Regression - Base Case (Beta) I . . . . . . 304 Table D-16: R III: Multiple Regression - Base Case (Beta) II . . . . . 305 Table D-17: R III: Multiple Regression - Dynamic Ownership (2003) . 306 Table D-18: R III: Multiple Regression - Dynamic Ownership (1998) . 307 Table D-19: R III: Multiple Regression - Dynamic Ownership (Pooled) 308 Table E-1: Table E-2: Table E-3: Table E-4: Table E-5:

R III: Multiple Regression - APM (2003) . . . . . R III: Multiple Regression - APM (1998) . . . . . R III: Multiple Regression - APM (1993) . . . . . R III: Multiple Regression - APM (Pooled) . . . . R III: Multiple Regression - Endogeneity (Pooled)

. . . . .

. . . . .

. . . . .

. . . . .

310 311 312 313 314

List of Abbreviations

XXI

List of Abbreviations 2SLS . . . . . . . Two-Stage Least Square 3SLS . . . . . . . Three-Stage Least Square AG . . . . . . . . . Aktiengesellschaft (“Limited Liability Stock Corporation”) AktG . . . . . . . Aktiengesetz (“Stock Corporation Act”) AMEX . . . . . American Stock Exchange ANOVA . . . . Analysis of Variance AnSVG . . . . . Anlegerschutzverbesserungsgesetz (“Investor Protection Improvement Act”) APM . . . . . . . Alternative Performance Measures BaFin . . . . . . . Bundesamt f¨ur Finanzdienstleistungsaufsicht (formerly: BaWe) (“Federal Financial Supervisory Authority”) BaWe . . . . . . . Bundesaufsichtsamt f¨ur den Wertpapierhandel (now: BaFin) (“Federal Financial Supervisory Authority”) BetrVG . . . . . Betriebsverfassungsgesetz (“Works Constitution Act”) BilKoG . . . . . Bilanzkontrollgesetz (“Balance Sheet Auditing Law”) BilReG . . . . . Bilanzrechtsreformgesetz (“Accounting Law Reform”) bn . . . . . . . . . . billion CAR . . . . . . . Cumulative Abnormal Return CEO . . . . . . . . Chief Executive Officer cf. . . . . . . . . . . confer CG . . . . . . . . . Corporate Governance DCGK . . . . . . Deutscher Corporate Governance Kodex (“German Corporate Governance Code”) e.g. . . . . . . . . . exempli gratia (“for example”) ECGN . . . . . . European Corporate Governance Network EDGAR . . . . Electronic Data Gathering, Analysis, and Retrieval System et al. . . . . . . . . et alii (“and the following”)

XXII

List of Abbreviations

EU . . . . . . . . . European Union EUR . . . . . . . . Euro excl. . . . . . . . . excluding FCF . . . . . . . . Free Cash Flow FMFG . . . . . . Finanzmarktf¨orderungsgesetz (“Financial Market Promotion Law”) GDP . . . . . . . . Gross Domestic Product GEX . . . . . . . German Entrepreneurial Index GmbH . . . . . . Gesellschaft mit beschr¨ankter Haftung (“Limited Liability Company”) HGB . . . . . . . Handelsgesetzbuch (“Commercial Code”) i.e. . . . . . . . . . id est (“that is”) IFRS . . . . . . . International Financial Reporting Standards inc. . . . . . . . . . incorporated incl. . . . . . . . . including InsO . . . . . . . . Insolvenzordnung (“Insolvency Statute”) InvG . . . . . . . . Investmentgesetz (“Investment Act”) IPO . . . . . . . . Initial Public Offering IV . . . . . . . . . . Instrumental Variable KAGG . . . . . . Gesetz u¨ ber Kapitalanlagegesellschaften (“Investment Companies Act”) KapAEG . . . . Kapitalaufnahmeerleichterungsgesetz (“Law to Ease Capital Procurement”) KapInHaG . . Kapitalmarktinformationshaftungsgesetz (“Law Governing Liability for Capital Market Information”) KapMuG . . . . Kapitalanleger-Musterverfahrensgesetz (“Capital Markets Model Case Act”) KonTraG . . . . Gesetz zur Kontrolle und Transparenz im Unternehmensbereich (“Law on Control and Transparency in the Corporate Sector”)

List of Abbreviations

XXIII

KStG . . . . . . . K¨orperschaftsteuergesetz (“Corporation Tax Law”) KuMaKV . . . Verordnung zur Konkretisierung des Verbotes der Kurs- und Marktpreismanipulation (“Regulation for the Concretion of the Prohibition of Stock Price and Market Price Manipulation”) LBO . . . . . . . . Leveraged Buy-Out LSX . . . . . . . . London Stock Exchange m . . . . . . . . . . million M&A . . . . . . . Mergers & Acquisitions MaKonV . . . . Marktmanipulations-Konkretisierungsverordnung (“Regulations for the Concretion of the Prohibition of Market Manipulation”) N.N. . . . . . . . . nomen nescio (“name unknown”) Nasdaq . . . . . National Association of Securities Dealers Automated Quotation System NPV . . . . . . . . Net Present Value NYSE . . . . . . New York Stock Exchange OLS . . . . . . . . Ordinary Least Square OSE . . . . . . . . Oslo Stock Exchange p./ pp. . . . . . . page/ pages p.a. . . . . . . . . . per annum (“yearly”) R&D . . . . . . . Research and Development S.p.A. . . . . . . Societ`a per azioni S&P 500 . . . . Standard & Poor’s 500 Stock Index SEC . . . . . . . . Security Exchange Commission SOX . . . . . . . . Sarbanes-Oxley Act SSRN . . . . . . Social Sciences Research Network St¨uckAG . . . . St¨uckaktiengesetz (“Law on Admission of Nonpar Shares”) TransPuG . . . Transparenz- und Publizit¨atsgesetz (“Transparency and Disclosure Law”)

XXIV

List of Abbreviations

TUG . . . . . . . Transparenzrichtlinie-Umsetzungsgesetz (“Transparency Directive Ratification Act”) U.K. . . . . . . . . United Kingdom U.S. . . . . . . . . United States of America UBGG . . . . . . Gesetz u¨ ber Unternehmensbeteiligungsgesellschaften (“Law on Private Equity Companies”) UMAG . . . . . Gesetz zur Unternehmensintegrit¨at und Modernisierung des Anfechtungsrechts (“Act on Corporate Integrity and Modernization of Rescission Law”) UmwG . . . . . Umwandlungsgesetz (“Restructuring Act”) USD . . . . . . . . United States Dollar VAG . . . . . . . . Versicherungsaufsichtsgesetz (“German Insurance Supervision Act”) VIF . . . . . . . . Variance Inflation Factor VorstOG . . . . Vorstandsverg¨utungs-Offenlegungsgesetz (“Management Compensation Disclosure Act”) vs. . . . . . . . . . versus ¨ . . . . . . Wertpapiererwerbs- und Ubernahmegesetz ¨ WpUG (“Acquisition and Takeover Act”) WpHG . . . . . . Wertpapierhandelsgesetz (“Securities Trading Act”)

Corporate Governance and Insider Ownership

1 1.1

1

Corporate Governance and the Role of Insider Ownership Why Do Corporate Governance and Insider Ownership Matter?

In the aftermath of large corporate scandals as Enron, Inc. (U.S.), MCI Worldcom, Inc. (U.S.), Parmalat S.p.A. (Italy) and Bankgesellschaft Berlin AG (Germany), public discussion about corporate governance has increased dramatically. The consequences of individual managers’ criminal and selfish behavior in such cases are dire and diverse: The wealth of shareholders and creditors of the affected companies was diminished, investors’ trust in corporate managers has been ruined on a much broader scale and policy makers reacted by initiating and passing numerous regulations. For example, the introduction of the Sarbanes-Oxley Act (SOX) in the U.S. in 2002 aimed at preventing the recurrence of such cases but also imposes significant administrative costs.1 Often, the occurrence of such scandals is reasoned by a failure of corporate governance. Though the diversity of definitions of corporate governance will be dealt with later on in section 2.1, it should be noted that corporate governance is commonly referred to in at least three different contexts: First, as already mentioned, in public discussion, the criminal and deceptive activities of corporate managers and the resulting corporate scandals are often attributed to a lack of adequate corporate governance. Second, in academic discussion, corporate governance primarily deals with the question of how to prevent corporate managers from misusing corporate resources and how to ensure that managers act in the best interest of those who provide funds, i.e. the shareholders. The second meaning, even though similar at the fist glance, must be clearly differentiated from illegal practices and corporate scandals2 : It usually assumes that managers act in accordance with law but managers are also seen as individuals which pursue their own — and not necessarily shareholders’ — interests. Third, from an economics and law perspective, corporate governance often refers to the whole system of corporate governance, i.e. the institutional characteristics and regulations, which may vary significantly across countries. The study at hand3 will (almost) exclusively focus on the second aspect, i.e. the question of 1 2 3

Cf. Ziener, Riecke, Hussla and Fockenbrock (2006), p. 2. See Helmis (2002), p. 2. The terms this study or this analysis will always refer to the works at hand in the following if not stated otherwise.

2

Corporate Governance and Insider Ownership

how conflicts of interest between managers and shareholder can be prevented or resolved. Thereby, the role of ownership structures — and especially insider ownership — will be at the very focus. To begin with, the importance of corporate governance and ownership structures in general and for the German capital market will be highlighted. Besides public discussion, also academic research extensively deals with corporate governance. A search at the Social Sciences Research Network (SSRN)4 for scientific papers submitted between August 2, 2005 and August 2, 2006 containing “Corporate Governance” in the title yields a total of 227 hits.5 Thereby, also the question of how corporate governance affects corporate performance has attracted broad attention: For example, G OMPERS , I SHII AND M ETRICK (2003) report that a trading strategy according to which firms with bad corporate governance, i.e. low shareholder rights (“dictatorship firms”), are sold and those with good corporate governance, i.e. strong shareholder rights (“democracy firms”), are bought yields abnormal annual returns of 8.5%.6 Similarly, in a recent survey among German listed companies 62.2% state that high standards in corporate governance have a “high” or “very high” importance for their stock market performance.7 But what constitutes good corporate governance or how can it be measured? The number of possible approaches is large: Some studies examine the compliance with (newly introduced) corporate governance codes while others analyze the effectiveness of single corporate governance mechanisms, e.g. supervisory board composition, compensation or control through the takeover market.8 A less obvious aspect refers to the ownership structures of public stock corporations, which have far-reaching implications for potential corporate governance issues. A prominent, current example illustrating the importance of ownership structures in the case of public German stock corporations is the case of Deutsche B¨orse AG. In 2005, foreign institutional investors opposed the plans of the Chief Executive Officer (CEO) to acquire the London Stock Exchange (LSE). The 4 5

6 7 8

For more information refer to the website http://papers.ssrn.com. Good surveys of corporate governance and overviews of the fast growing theoretical and empirical literature are provided amongst others by Shleifer and Vishny (1997); Farinha (2003); Gillan (2006). See Gompers, Ishii and Metrick (2003), p. 144. Cf. Daouk, Lee and Ng (2006). See ergo Kommunikation (2005), p. 51. See section 2.2.

Corporate Governance and Insider Ownership

3

CEO initially dismissed institutional investors concerns, last but not least because of their historical comparatively minor weight. However, the institutional investors revolted effectively and finally achieved the denial of the takeover attempt as well as the resignation of the CEO and the supervisory board chairman.9 Besides outside blockholders, as in the case of Deutsche B¨orse AG, also the shareholdings of insiders, i.e. the members of the management and supervisory boards, might play an important role. In the case of significant insider ownership10 levels, corporate managers also become shareholders themselves and the typical conflicts of interest between both groups may be alleviated or the interests might even become aligned. Similar arguments also can be applied to the advantages of family firms, where management and ownership usually are still (partially) united in the hands of the (founding) family. Therefore, the discussion about family firms should be seen closely related to the discussion about the effects of insider ownership. For example, family firms are often perceived to be less short-term oriented and more focused on value than on mere growth.11 Nevertheless, family firms sometimes are also characterized by a bad reputation as outside (minority) shareholder are supposed to become expropriated by potent family shareholders.12 As empirical research about the relationship between ownership — and especially insider ownership — and performance is rare and ambiguous for the case of Germany, these arguments should be regarded as speculations in the first place. The question of how insider dominated firms perform on the stock market becomes even more important if one argues that the ambiguity regarding their performance might play a role in the underdevelopment of the German capital market. This underdevelopment is a topic which has attracted widespread discussion in the past.13 Even though the importance of the German capital market has clearly increased in the last decade, a closer look at the actual figures still reveals a significant backlog demand in capital market development. 9 10

11 12

13

See Kamp and Krieger (2005), pp. 54-56. Insider ownership refers to the absolute level of shareholdings by corporate insiders as members of the management and supervisory boards. It has to be differentiated from insider trading which refers to the fact that corporate managers buy or sell shares in their own company. The latter is problematic (and illegal) if the managers act on information which is not available to other (outsider) market participants. See footnote 880; Weishaupt (2005). Cf. Hajek (2004), pp. 114-117; G¨ocmen and Meyer (2004), p. 25; N.N. (2004), p. 35; Dostert (2004), p. 21; Gr¨oneweg (2004), p. 22. See Monopolkommission (1998), pp. 18-63.

4

Corporate Governance and Insider Ownership

In 2004, the market capitalization of all domestic listed companies, measured as a percentage of the German gross domestic product (GDP), was only 43.6% in comparison to 134.9% in the U.K., 139.4% in the U.S. and 231.1% in Switzerland, which thereby takes the front-runner position. Even in a comparison with the (mainly) Continental European EU countries — which are largely characterized by civil law based corporate governance systems in contrast to the common law based corporate governance systems in the U.K. and the U.S.14 — only Austria (30.0%), Poland (29.6%) and Hungary (18.8%) show lower ratios than Germany.15 This underdevelopment is reflected on both sides of the market for public equity, i.e. the supply and the demand side. On the supply side, in 2005 only 16.6% of the German population above the age of 14 provided equity capital to listed companies by either owning directly shares of these companies or indirectly via investment funds. The ratio has almost doubled from only 8.9% in 1997 but remains comparatively low.16 This development was probably induced by legal initiatives to promote funded retirement provisions. It can be expected that this trend further nurtures the proliferation of stock ownership in the future. On the demand side, the number of domestic listed companies in the regulated and official markets (“Geregelter Markt” and “Amtlicher Markt”) experienced an absolute decline since 1997: It decreased by 7.2% from a total of 698 to 648 in 2005. In the same period, the number of listed companies in the remaining eight EU countries where data was available for both dates increased on average by 16.4%.17 This directly translates into a lack of initial public offerings (IPOs) on the German capital market. Primary offerings as percentage of total stock market capitalization reach an average of a mere 0.5% for the period ranging from 1997 to 2005 and even falls to 0.3% if the heyday of the new economy (i.e. the year 2000 with a ratio of 2.0%) is excluded.18

14 15 16 17 18

See section 2.3; La Porta, Lopez-de Silanes, Shleifer and Vishny (1998), pp. 1117-1119. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 05-3. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 08.3-Zahl-D. See Deutsches Aktieninstitut e.V. (DAI) (2006), pp. 02-1, 02-3. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 03-2. The problems stemming from a lack of German IPOs are even worsened by an increasing number of delistings (going privates). See Franke (2004), pp. 136139. Cf. Raettig (2004), p. 23.

Corporate Governance and Insider Ownership

5

As IPOs as means of financing for the corporate sector are rare, other sources of financing, especially bank-lending, dominate the German financing market.19 But what determines the actual resource pool of potential IPO candidates? Two obvious sources of IPO candidates come into mind first: The portfolios of private equity/venture capital companies on the one hand and small- and medium sized companies (often family firms) on the other hand. Concerning the former, German private equity activity only started to take off recently and therefore still a manageable pool of potential IPO candidates is held by private equity/venture capital firms.20 With regard to the latter, among the 153 German IPOs in the period from 1990 to 2000 (excluding financial, non-domestic, and preference share capital offerings), a total of 70% is estimated to have been family firms. Family firms could be defined — for example — as companies where at least 25% of the shares are owned by one family or several families which is/are represented in the management and/or supervisory board.21 In a European comparison, among the 50 (10) largest private companies, 26 (5) are German which thereby represent the by far largest group.22 But why do so many large German companies refrain from going public? Do they fear a loss of control? Do they believe that founder families have to give off majority control of their companies in the context of an IPO? Do they have sufficient capital from internal funds and/or bank financing? Or do they lack large investment projects or capital intensive expansion opportunities?23 Another frequently heard argument is that investors generally would dislike listed companies which are family controlled or show high levels of insider ownership as interest between insider and outsider shareholders may diverge. Investors’ lower interest in such firms would translate into less trading volume, lower firm valuations and finally higher equity costs for family firms.24 This in turn would make IPOs rather unattractive to

19

20

21

22 23 24

Therefore, the German system of corporate governance is also often labeled as a bank based system. Cf. Wenger and Kaserer (1998a); Gorton and Schmid (2000); Edwards and Nibler (2000); Hommel and Schneider (2003); Commission of the European Communities (2003a). See Bundesverband Deutscher Kapitalbeteiligungsgesellschaften (BVK) (2003), p. 6. Cf. Bottazzi and Da Rin (2003), pp. 3-9. See Schiereck and Jaskiewicz (2004), p. 13. Cf. Ehrhardt and Nowak (2003a). Of course, several differing definitions of family firms exists. Nevertheless, the definitions mostly include a criterion relating to the ownership stake of the (founding) family. See Chrisman, Chua and Sharma (1996), pp. 4-7. See N.N. (2005b), p. 11. Cf. Commission of the European Communities (2002), pp. 11-13. For evidence on privately-held firms’ decisions to go public see B¨ohmer and Ljungqvist (2004). Cf. Amihud (2002).

6

Corporate Governance and Insider Ownership

these types of firms. As indicated at the very beginning of this section, exactly this proposition is questioned and examined in this study. While this study aims at providing an academic contribution to the discussion of how insider dominated firms perform on the German capital market, practical evidence can be found in the German Entrepreneurial Index (GEX). The GEX is a new stock market index, which was introduced by Deutsche B¨orse — Germany’s largest stock exchange operator — on January 3, 2005. As its main novelty, the GEX applies shareholder structures as its primary selection criterion: It aggregates the performance of all German insider dominated firms where active and former members of the management and supervisory boards own at least 25% of the voting stock.25 As insider or owner dominated firms are often also referred to as entrepreneurial firms, the GEX bears the name German Entrepreneurial Index.26 Since its introduction, the GEX increased by 58.3% (as of July 7, 2006) and thereby outperformed the other well-known German stock indices (see figure 1-1). Of course, the 18 month period from January 2005 to July 2006 is not sufficient to prove a general outperformance of insider dominated firms. Nevertheless, the GEX documented for the first time that insider dominated firms performed at least as well as other German listed companies. Furthermore, first preliminary evidence indicates that the GEX, and the introduction of derivatives based on the GEX, increased the liquidity of index companies’ shares on the stock exchange. This in turn would finally lower their cost of equity as insider dominated firms are often comparably small and stock prices are assumed to incorporate a so-called illiquidity discount.27 1.2

Aims of the Study

Having highlighted the relevance of corporate governance, shareholder structures, insider ownership and their effects on corporate performance, four concise aims pursued by this study are presented now. 25

26

27

For more details on the construction of the GEX and its background see Deutsche B¨orse (2005a); Achleitner, Kaserer and Moldenhauer (2005). Cf. Kaserer, Achleitner and Moldenhauer (2005). See Gravelle and Rees (1984), p. 337; Achleitner, Kaserer and Moldenhauer (2005), p. 118. Note that especially earlier literature often differentiated between owner and management controlled firms where owner controlled firms are characterized by a concentrated ownership structure independent of the shareholders’ identity. See Holl (1975), p. 258. See Kaserer, Achleitner, Moldenhauer and Ampenberger (2006); Achleitner and Moldenhauer (2006).

7

Corporate Governance and Insider Ownership 190 180 170

03.01.2005 = 100

160 150 140 130 120 110 100

GEX

DAX

MDAX

03.07.2006

03.06.2006

03.05.2006

03.04.2006

03.03.2006

03.02.2006

03.01.2006

03.12.2005

03.11.2005

03.10.2005

03.09.2005

03.08.2005

03.07.2005

03.06.2005

03.05.2005

03.04.2005

03.03.2005

03.02.2005

03.01.2005

90

TecDAX

Source: Datastream; author’s illustration.

Figure 1-1: Chart of the German Entrepreneurial Index (GEX) To begin with, the first aim of this study is to provide a comprehensive analysis of the objectives and interests of diverse shareholder types against the background of the German system of corporate governance. As most existing related analyses are rather generic, the main novelty of this study is the detailed classification of shareholder types along two dimensions (i.e. shareholders’ identity and concentration) from a corporate governance perspective and the explicit consideration of the idiosyncrasies of the German corporate governance system. The second aim of this study is to enable new insights into the composition of shareholder structures of German listed stock corporations. Therefore, extensive descriptive statistics and figures on shareholders’ identity and concentration during the decade ranging from 1993 to 2003 are presented. The level of detail of this study is expected to exceed that of previous analyses, as the transparency of shareholder structures has improved only recently due to changes in the regulation of ownership disclosure. Furthermore, it will be interesting to observe how ownership structures develop over a period of time in which the German capital market experienced significant changes.

8

Corporate Governance and Insider Ownership

The third aim of this study is to organize and summarize existing empirical research on insider ownership. Therefore, the different methodological approaches are contrasted and — even more important — the ambiguity of the findings is explained. A selection of empirical studies from different corporate governance systems will be sufficient to highlight the main characteristics and findings of those works while an exhausting coverage of all related research would be beyond the scope of this study. Finally, the fourth aim of this study is to empirically analyze the causes and consequences of insider ownership for the German market. The main question to be answered is whether insider ownership has an impact on corporate performance or not. Nevertheless, the analysis of the determinants of insider ownership and its effects on the corporate control setting is necessary to fully understand the concept on insider ownership. While empirical evidence on this question is rather rich for the U.S., such an analysis does not exist for public German companies, yet. 1.3

Structure of the Analysis

According to the outlined research question and the stated aims, this study applies both deductive-theoretical as well as empirical analyses. The main part of this study consists of six chapters (see figure 1-2). Chapter 2 revisits the fundamentals of corporate governance in a condensed way. Departing from the variety of definitions of corporate governance (section 2.1), the conflicts of interests between shareholders management and the corresponding corporate governance mechanisms to resolve those conflicts are outlined in section 2.2. The two polar prototypes of corporate governance systems are contrasted in section 2.3 before historical and current developments of the German system of corporate governance as well as its main characteristics are summarized (section 2.4). Chapter 3 analyzes the causes and consequences of shareholder structures in theory and practice of corporate governance. Therefore, the issues related to the separation of ownership and control in public companies are clarified (section 3.1) and the legal distribution of rights and duties among owners and managers is drafted (section 3.2). The main part of this chapter, section 3.3, consists of a

9

Corporate Governance and Insider Ownership

Chapter 5

Chapter 6

• Hypotheses on Causes and Effects of Insider Ownership • Determinants of Insider Ownership • Effects of Insider Ownership on Corporate Control • Effects of Insider Ownership on Corporate Performance

• Empirical Analysis of Insider Ownership in German Public Stock Corporations • Study Design • Sample and Data • Definition of Variables • Descriptive Results

Chapter 3

Chapter 4

• Causes and Consequences of Shareholder Structures in Theory and Practice of Corporate Governance • Separation of Ownership and Control

• Insider Ownership and its Effects on Corporate Performance

• Constitution of German Stock Corporation • Shareholders’ Objectives and Interests • Conflicts of Interests Bottom-up Analysis

• Methodology • Empirical Results • Limitations

• Theoretical Views • Neoclassical • Institutional • Empirical Evidence • Capital Market Systems • Corporate Law Systems • Research Gap Top-down Analysis

Chapter 7 • Conclusion and Implications • Conclusion • Implications for Research and Theory • Implications for Practice

Chapter 2 • Fundamentals of Corporate Governance • Definitions and Understanding • Corporate Governance Mechanisms • Basic Types of Corporate Governance Systems • German System of Corporate Governance

Source: Author’s illustration.

Figure 1-2: Structure of the Analysis bottom-up analysis of the interests and objectives of different shareholder types. Thereby, shareholders will be differentiated along two dimensions: identity and concentration. The concluding section 3.4 organizes potential conflicts of interests between shareholders and managers and lists ways of overcoming these conflicts. Next, chapter 4 takes a top-down perspective on insider ownership and its effects on corporate performance. First, the role of insider ownership according to different theoretical views is discussed in section 4.1. Special importance is attached to the institutional views, i.e. the transaction cost, property rights and principal agent theories while the neoclassical theory is dealt with as a means of contrast. Second, the empirical evidence of insider ownership research is summarized in section 4.2. As for the German market, primarily descriptive results are available. They are presented separately from the primarily U.S. findings

10

Corporate Governance and Insider Ownership

regarding the relationship between insider ownership and performance. Finally, based on the ambiguous results with respect to the relationship in different corporate governance systems, the lack of research on the German market is highlighted. Condensing the findings of the two preceding chapters, chapter 5 proposes several hypotheses on the causes and effects of insider ownership. Based on theoretical analyses and empirical evidence, section 5.1 develops hypotheses on the determinants of insider ownership (cause-analysis). Then, hypotheses on the effects of insider ownership on the corporate control setting (section 5.2) and corporate performance (section 5.3) are elaborated (effect-analysis). In the very center of this study chapter 6 — the empirical analysis on the German public equity market — tests the previously developed hypotheses. Section 6.1 lays down the study design and section 6.2 describes the sample and data sources used. The shareholder structure, performance and control variables are defined in section 6.3 before comprehensive descriptive results are presented (section 6.4). Thereby, a special focus will be on shareholder structures and their development over the sample period (1993-2003). In section 6.5, the hypotheses developed in chapter 5 are presented once more in an overview table to lead over to the empirical tests. Section 6.6 will elaborate on the methodology of the empirical tests. The empirical results are divided into three parts: First, the findings concerning the determinants of insider ownership (“Results I”; section 6.7) are presented. Second, section 6.8 summarizes the results with respect to the effects of insider ownership on corporate performance (“Results II”). Finally, and most comprehensive, the results on the relationship between insider ownership and performance (“Results III”) are presented in section 6.9. The analyses are more extensive and subjected to a series of robustness tests as they provide the answer to the main research question. As the problems involved in any empirical corporate governance studies are non-trivial, section 6.10 concludes with a survey of limitations of the empirical study. Last, chapter 7 concludes this study. After providing a short summary of the main findings in the conclusion (section 7.1), the implications for research, theory and practice are outlined in section 7.2.

Fundamentals of Corporate Governance

2 2.1

11

Fundamentals of Corporate Governance Definition and Understanding of Corporate Governance

The term corporate governance does not refer to a clear-cut concept but is subject to a broad range of different definitions and understandings in literature which sometimes gives rise to confusion.28 In the context of this study a rather broad definition from an economic point of view is used, which was introduced by the comprehensive survey of corporate governance by S HLEIFER AND V ISHNY (1997): “Corporate Governance deals with the ways in which suppliers of finance to corporations assure themselves of getting a return on their investments.”29 This definition of corporate governance is based on principal agent theory and therefore bases on the fundamental problem arising from the separation of ownership and control, which was first brought into scientific discussion by the pathbreaking work of B ERLE AND M EANS (1932).30 According to the principal agent theory, managers (i.e. the “agents”) are ultimately employed by investors (i.e. the “principals”) to run firms on their behalf.31 Since the interests and objectives of both groups are likely to differ, corporate governance would be concerned with monitoring the management or aligning the interests of both parties and to ensure that firms are run in the best interest of the investors.32 Further, corporate governance in the context of this study only relates to publicly traded stock corporations (“Aktiengesellschaften” (“AGs”)), in which conflicts of interests are expected to be most severe.33 28 29 30

31

32 33

Cf. Kaserer (2001), p. 176-178. Shleifer and Vishny (1997), p. 737. See Berle and Means (1932); Shleifer and Vishny (1997), p. 738. The theoretic foundations of the principal agent theory and the main features of the separation of ownership and control will be discussed in more detail in sections 3.1 and 4.1.2.4 respectively. An alternative approach to corporate governance is the stewardship theory which assumes that management’s and shareholders’ interests are generally aligned. Therefore governance mechanisms which empower the management are preferred to those that control and monitor management. See Davis, Schoorman and Donaldson (1997). For common antecedents of both theories cf. Klein, Pieper and Jaskiewicz (2005b). See Mayer (1996), p. 10. The other, relatively rare corporate form qualifying for public listing is the “Kommanditgesellschaft auf Aktien” (“KGaA”), which is a partly limited partnership combining characteristics of a partnership with those of a stock corporation. Agency conflicts also may appear in non-public stock corporations and private limited liability companies (”Gesellschaften mit beschr¨ankter Haftung“ (“GmbHs”)). However, since the ownership of this type of companies is usually considerably less dispersed than in public companies, agency

12

Fundamentals of Corporate Governance

In Germany, corporate governance is often equated with the term corporate constitution (“Unternehmensverfassung”), which has been shaped mainly by juristic discussions about the legal and effective distribution of tasks and duties among the management and supervisory boards and owners.34 Similar, M ATTHES (2000) defines corporate governance as “[. . . ] institutional mechanisms which serve to control the management of a limited liability corporation and hence also arrange relationships among shareholders, managers and stakeholders.”35 In this definition, the control aspect is emphasized and that is why corporate governance is sometimes also referred to as corporate control (“Unternehmenskontrolle”). In this study, corporate governance and corporate control will not be used synonymously since the concept of corporate control is subordinate to the broader framework of corporate governance.36 In contrast to the two previous definitions, the perspective of S HLEIFER AND V ISHNY (1997) treats the corporate constitution and corporate control as only two components of a broader system of corporate governance. The system character of corporate governance is emphasized by the definition of the OECD (2004): “Corporate governance is one key element in improving economic efficiency and growth as well as enhancing investor confidence. Corporate governance involves a set of relationships between a company’s management, its board, its shareholders and other stakeholders. Corporate governance also provides the structure through which the objectives of the company are set, and the means of attaining those objectives and monitoring performance are determined.”37

34

35

36 37

conflicts play a comparable minor role. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 20. See Mann (2003), p. 30: For example, the commission “Corporate Governance” of the German government explicitly aims at laying the fundamentals for a modernization of the German corporate law. See Bundesregierung (2004). Cf. section 3.2. Matthes (2000), p. 17. Translated from German: “Gemeint sind [. . . ] die institutionellen Mechanismen, mit denen das Management einer Kapitalgesellschaft kontrolliert wird und die damit auch die Beziehungen der Shareholder, Manager und Stakeholder untereinander regeln.” Cf. Gompers, Ishii and Metrick (2003), p. 144, who identify as much as 24 distinct corporate governance mechanisms. See Mann (2003), p. 31; Fama and Jensen (1983b), pp. 303-304. OECD (2004), p. 11. Note that this definition explicitly incorporates the claim that the corporate governance structure is also responsible for goal setting and performance monitoring in corporations.

Fundamentals of Corporate Governance

13

Support for the system perspective of corporate governance is given by a more recent definition of D EUTSCHE BANK (2002) which states that “[. . . ] corporate governance is constituted through code of conducts, institutions, markets (for corporate financing) and organized industrial relationships, which should altogether be regarded from a systemic perspective. Besides the basic legal and regulative framework, the corporate practice of control plays an important role.”38 Without questioning the legitimation of this systemic view, this study will take a rather narrow view on one selected corporate governance mechanism, i.e. the shareholder structure and especially insider ownership, because it has attracted only limited research until now, at least in the case of Germany. A systemic view of the various corporate governance mechanisms, which will be outlined in section 2.2, offers promising future research opportunities. In contrast to the quite narrow German understanding, recently very broad definitions of corporate governance evolved which explicitly incorporate the role of stakeholders other than shareholders (i.e. employees, customers, suppliers and neighbors) whose welfare should be kept in mind. Such a corporate governance would then refer to “[. . . ] the design of institutions that induce or force management to internalize the welfare of stakeholders.”39 Another very broad definition describes corporate governance as the system of financing, management and control of corporations.40 The arguments for a very broad definition of corporate governance appear intuitively plausible, but applying such a broad definition was to inflate the scope of this study and would blur the focus on the basic principal agent conflict between shareholders and management, which will be shortly summarized in section 2.2.2 and later on more detailed in section 4.1.2.4. To sum up, the definition given by S HLEIFER AND V ISHNY (1997) focuses as clear on the very cause of the whole corporate governance discussion from a finance point of view as none of the other presented definitions. Thus, in the 38

39 40

Deutsche Bank (2002), p. 3. Translated from German: “Die CG wird gepr¨agt durch Verhaltensregeln, Institutionen, M¨arkte (f¨ur die Finanzierung von Unternehmen) und organisierte industrielle Beziehungen, die insgesamt in systemischer Perspektive betrachtet werden sollten. Neben den gesetzlichen und regulatorischen Rahmenbedingungen spielt die unternehmerische Praxis der Kontrolle eine wichtige Rolle.” Cf. Nassauer (2000), p. 44 and Schmidt (2003), pp. 3-4. Tirole (2001), p. 4. Cf. Mayer (1996), p. 10. See Deutsche Bank (2002), p. 1.

14

Fundamentals of Corporate Governance

context of this study corporate governance is about how investors assure to get their money back. 2.2

Corporate Governance Mechanisms

The need for corporate governance mechanisms is revealed in the initial definition of corporate governance which points to the fact that suppliers of finance might face legitimate doubts that managers will always act in their best interest. This might be especially true for those who provide equity capital, i.e. the shareholders. Before the origin of these potential conflicts of interests can be analyzed (section 2.2.2), the question of what corporations’ business objectives are or, even more, corporations’ right to exist is must be answered first (section 2.2.1). Finally, the various means for overcoming shareholders’ and management’s conflictive interests will be presented and analyzed (section 2.2.3). 2.2.1

Business Objective and Shareholder Value

The rationale for firms’ existence is explained by C OASE (1937) in his seminal work The Nature of the Firm as follows:41 The boundaries of the firm are determined by the range of exchanges where the market system is suppressed and resource allocation would take place by means of authority and direction. Activities would be performed inside the firm as long as the costs of using the market are greater than those of using direct authority.42 A LCHIAN AND D EM SETZ (1972) later on specify the meaning of authority by stating that firms are characterized by a complex nexus of contracts as an instrument for voluntary exchange.43 Hence, the business objective of a firm is harder to observe than the interests of individual contracting parties, or as J ENSEN (2000) put it, “The firm is not an individual. It is a legal fiction that serves as a focus for a complex process in which the conflicting objectives of individuals (some of whom may ‘represent’ other organizations) are brought into equilibrium within a framework of contractual relations.”44 41 42

43

44

Cf. section 4.1 for an overview of the underlying economic theories. See Coase (1937), pp. 390-395; Jensen (2000), pp. 87-88. Coase’s view represented an explicit denial of the then prevailing neoclassical theory of the firm, which explained the existence of firms mainly by three advantages of firms over individuals: 1) benefits of specialization, 2) extended use of capital goods and 3) economies of scale. See Ulen (1993), pp. 304-305. See Alchian and Demsetz (1972), pp. 781-783, 794-795. Cf. Fama (1980), p. 290; Jensen (2000), p. 88; Valc´arcel (2002), pp. 141-188. Jensen (2000), p. 89. Cf. Grossman and Hart (1986), pp. 691-719.

Fundamentals of Corporate Governance

15

The business objective of a stock corporation is plausibly derived from the interests of those who provide capital to the firm, i.e. the shareholders. They hold so-called residual claims on the assets and cash flows of the firm. These cash flows can be reinvested into business or distributed to claimholders. While other parties (e.g. employees or creditors) usually receive fixed claims, the owners — or more specifically in the case of a stock corporation the shareholders — receive what is left after all fixed claims have been served. As reward for bearing this risk, the shareholders are entitled to all remaining profits. Since the owners would not invest into a company without appropriate return prospects, the maximization of shareholders’ claims has to be seen as the primary business objective. Acting on this business objective has become popular under the term shareholder value management. Therefore, a shareholder value orientation should aim at maximizing the value of shareholders’ wealth through dividend income and increasing stock prices (i.e. capital gains).45 Without doubting the importance of shareholder value as a major business objective, a considerable number of scientists and practitioners calls for a broader view.46 Firms are obliged to the interests of a number of diverse groups with different — sometimes conflicting — interests, so-called stakeholders. Groups of stakeholders include amongst others employees, consumers, creditors, suppliers, government or society as a whole. According to this stakeholder approach, in the long run firms can only achieve their goal of shareholders’ wealth maximization by paying appropriate attention to the expectations of various stakeholders.47 It can be expected that the heterogeneity of stakeholders and their interests respectively increases along with firm size making the adequate treatment of all stakeholder groups particularly difficult in the case of large corporations.

45

46 47

See Rappaport (1999), pp. 1-14; Bischoff (1994); Achleitner and Bassen (2002), pp. 611-635; OECD (1998a); J¨urgens, Rupp and Vitols (2000). For a critical survey of the micro-foundations of shareholder value see Reb´erioux (2003). See Tirole (2001), pp. 3-4. Cf. Cornell and Shapiro (1987); Berrar (2001), pp. 27-28. Even CEOs of leading German companies are among those who highlight the importance of other stakeholders besides shareholders. For example, Wendelin Wiedeking, CEO of successful German luxury car manufacturer Porsche AG, recently ranked the importance of stakeholder groups as follows: 1) customers 2) employees 3) business partners, suppliers, and dealers 4) shareholders. See Wiedeking (2005), p. 19.

16

Fundamentals of Corporate Governance

This study assumes that the maximization of shareholder value is a company’s business objective. Thus, the empirical part of this study uses shareholder returns as a proxy of meeting business objective.48 2.2.2

Conflicts between Shareholders and Management

Whilst an extensive review of the conflicts of interest between management and shareholders will follow in section 3.4, this section summarizes the basic problem. Just as defined above, under the paradigm of shareholder value maximization, the business objective is absolutely in line with the shareholders’ interests; even more it is one and the same thing. In contrast, the objectives of management — or better managers — are less clear-cut.49 In addition to their salary as employees of the company managers could be self interested to extract private benefits (“consumption on the job”). For example, they might buy expensive corporate jets because of the associated prestige and comfort or invest in charity projects to increase their local reputation. While not all of these actions must necessarily be opposed to shareholders’ interests, the important fact is that the benefits of such actions disproportionately accrue to the managers themselves.50 The existence of such conflicts of interest is not a recent finding. It dates back as far as to S MITH ([1994] 1776), who stated in his pathbreaking work An Inquiry into the Nature and Causes of the Wealth of Nations: “The directors of such companies [author’s note: i.e. joint-stock] companies, however, being the managers rather of other people’s money than of their own, it cannot well be expected, that they should watch over it with the same anxious vigilance with which the partners in a private copartnery frequently watch over their own. Like the stewards of a rich man, they are apt to consider attention to small matters as not for their master’s honour, and very easily give themselves a dispensation from having it. Negligence and profusion, therefore, must always prevail, more or less, in the management of the affairs of such a company.”51 48 49

50 51

See section 6.3.2. In the beginning, the functional characteristics of shareholders and management will be analyzed separately. Later on, it will be analyzed what happens if the functions of management and shareholders overlap, i.e. in the case of insider ownership. Different forms of private benefits of control will be outlined later on in section 3.3.2.1. Smith ([1994] 1776), p. 800.

Fundamentals of Corporate Governance

17

Similarly, W ESTON , S IU AND J OHNSON (2001) put the possible consequences of unresolved conflicts of interest in explicit words: “The books are cooked to meet or exceed analysts’ forecasts; workers are underpaid and executives are overpaid; board members are hand picked by management to help plunder the firm at expense of shareholders and creditors; etc., etc.”52 The experiences from recent scandals in large American companies, e.g. Enron, Inc. and MCI Worldcom, Inc., show that the worries of the authors were not without cause and the factual collapse of the Russian economy in 1998 pointed out what consequences the lack of effective governance mechanisms can have for the whole economy.53 Following the view of the firm as a nexus of contracts, those governance problems arise from incomplete contracts, which link a firm and its various stakeholders and thereby create a network of relationships and interdependencies between the different stakeholder groups.54 However, when conflicts of interest are as extreme as indicated by the previous citations, one might ask why investors should entrust money to corporations and their selfish managers? The next section will describe and structure different control mechanisms, which generally are thought to prevent managers from value destroying behavior. 2.2.3

Mechanisms to Control Management

Sometimes it is argued that worrying too much about corporate governance mechanisms is redundant and counterproductive, since product market competition induces firms to practice cost minimization, which includes the use of rules and mechanisms that provide the company with access to outside capital at the lowest cost.55 Even though product market competition undoubtedly constitutes one important corporate governance mechanism56 , it should be seen as only one element of a whole set of governance mechanisms. These are usually

52 53

54 55

56

Weston, Siu and Johnson (2001), p. 595. See Cromme (2002a), pp. 3-4. For corporate governance issues involved in the collapse of Enron, Inc., see Munzig (2003), pp. 19-47; Harris and Kramer (2002), pp. 14-19; Permanent Subcommittee on Investigations (2002), pp. 3-60. Regarding corporate governance in Russia see Sprenger (2002); Shleifer and Vishny (1997), pp. 768-769. See Schmidt (2003), p. 5. Cf. Shleifer and Vishny (1997), p. 738. This view, the so-called evolutionary theory of economic change, was originally reasoned by Alchian (1950) and Stigler (1958). Cf. Vives (2000), pp. 10-12; Allen and Gale (2000), pp. 23-84; Januszewski, K¨oke and Winter (1999). Cf. Hart (1980).

18

Fundamentals of Corporate Governance

Legal regulations Internal Control Mechanisms • Board of directors / Supervisory board • Composition of the board • Compensation of board members • Management / Management board • Insider ownership / Managerial ownership • Management compensation • Internal labor market for managers • Ownership concentration • Financial policy • Creditors • Leveraged buy-outs (LBOs)

External Control Mechanisms • Capital market • Primary capital market and Secondary capital market • Equity capital market and Debt capital market • Proxy contests • Market for corporate control (Mergers & Acquisitions (M&A)) • Product market competition • Factor market competition • External labor market for managers • Private sources of external oversight

• Other contracting parties • Corporate Charter

Sometimes ‘internal’ and ’external’ mechanisms are referred to as ‘direct’ and ‘indirect’ mechanisms respectively. See Nassauer (2000), pp. 32-55. The board of directors does not accurately equate to the German form of a supervisory board. Since most corporate governance studies apply to one-tier board CG-systems — and not to the German two-tier system of management board (“Vorstand”) and supervisory board (“Aufsichtsrat”) — both types of boards are listed separately in the table. Source: Author’s illustration, following Jensen (1993), p. 850; Weston, Siu and Johnson (2001), pp. 598-615; Bott (2002), pp. 3-4; Gillan (2006), pp. 382-384.

Figure 2-1: Overview of Control Mechanisms categorized into mechanisms of internal and external control which are embedded into the framework of legal regulations (see figure 2-1).57 Legal regulations constitute the framework for both internal and external control mechanisms. They determine the power of the respective mechanisms as well as their interdependencies. For example, corporate law — primarily the Stock Corporation Act (“Aktiengesetz” (“AktG”)) — prescribes the distribution of rights and duties among management board, supervisory board and shareholders and therefore outlines a general framework of rules that could be called the corporate constitution. This will be described later on in section 3.2. Therefore, regulation has a direct impact on internal mechanisms, e.g. the impact of the AktG on the supervisory board’s rights and duties. Similarly, the Acquisition 57

Similarly, Jensen (1993), p. 850, names four control forces: capital markets, legal/political/regulatory system, product and factor markets and internal control system. Cf. Kaserer (2001), pp. 178–182; Gillan (2006), p. 384.

Fundamentals of Corporate Governance

19

¨ ¨ and Takeover Act (“Wertpapiererwerbs- und Ubernahmegesetz” (“WpUG”)) regulates investors’ scope (e.g. the market for corporate control) by prescribing mandatory offers for all outstanding shares if more than 30% of the voting rights of a publicly listed company are acquired.58 Among the internal control mechanisms the board of directors, or for the German case the supervisory board, plays an important role.59 Elected by shareholders at annual shareholders’ meeting, the supervisory board’s legal obligation is to preserve shareholders’ interests.60 The most examined features of boards of directors are their compositions as well as the compensation of their members. For the former, the question of the role of outside directors in monitoring management is often addressed for the case of one-tier board systems (e.g. in the U.S. or U.K.). It does not refer to the case of the German two-tier system, where only outsiders are eligible for membership in the supervisory board, i.e. insiders (members of the management board) are not electable.61 For the latter, the effectiveness of a well-structured compensation system, e.g. containing stock options to align board members’ interests with those of the shareholders, has been in the center of interest.62 While in theory monitoring by the board of directors should be capable of dealing with at least the major corporate governance problems, the effectiveness of such control must at least be doubted in practice.63 With regard to the management board, control mechanisms rather aim to prevent potential agency conflicts ex ante rather than to discipline agents once malpractice has occurred ex post. One important aspect is to — at least partly — align management’s incentives with those of the owners: This is obviously the 58

59 60

61

62

63

¨ Studies investigating the relation between corporate governance and investor protection Cf. §§ 29, 35 WpUG. or security laws are e.g. La Porta, Lopez-de Silanes, Shleifer and Vishny (2000); Daouk, Lee and Ng (2006). Cf. Gillan (2006) pp. 389-390. See Goergen, Manjon and Renneboog (2004), pp. 13-16. The constitution of the German stock corporation, as dictated by German corporate law, will be described in more detail in section 3.2. For studies which suggest or test the hypothesis of a superior monitoring quality by outside directors see amongst others Fama (1980); Weisbach (1988); Rosenstein and Wyatt (1990); Barnhart, Marr and Rosenstein (1994); Borokhovich, Parrino and Trapani (1996); Faccio and Lasfer (1999); Holderness, Kroszner and Sheehan (1999); Peasnell, Pope and Young (2003). For studies addressing executive compensation see amongst others Mehran (1995); Chung and Pruitt (1996); Weber and Dudney (2003); Kaserer and Wagner (2004). See Weston, Siu and Johnson (2001), pp. 598-602; Baums (1993), pp. 61-63; Jensen (1993). Wenger and Kaserer (1998b), pp. 43-47, describe the cases of Metallgesellschaft AG (MG) and Daimler-Benz AG as drastic cases of supervisory board’s failure. Cf. Morck, Shleifer and Vishny (1989).

20

Fundamentals of Corporate Governance

case if management board members also own stakes in the firm they are employed with. 64 Another possibility is to align management’s compensation with shareholders’ interests by tying it to the development of the company’s stock price.65 In addition, companies’ internal labor markets for managers may prevent especially young managers from value destroying activities, since such activities — once detected — are likely to hamper their future career chances within the company. Whereas in presence of a totally dispersed shareholder structure no single owner might be able to control management, a certain degree of ownership concentration could act as a counterbalance to a selfish management.66 Similarly, the financing policy could act as a disciplinary mechanism, if providers of debt capital, i.e. the creditors, exert control on management in order to guarantee the proper (re-) payment of interest and principal. In this case the borders between internal and external (or outside) control cancel each other out. For example, a bank providing significant funds will minimize its loss risk by the use of numerous covenants in the credit contract, which provide extensive control privileges and therefore certainly has to be seen as an internal control mechanism;67 whereas the market pressure from issuing corporate bonds would rather represent a form of external control.68 All other contracting parties, bound to the company closely by contracts, could also serve as internal control mechanisms. For example, employees could easily be discouraged or demoralized if they see their efforts to create value foiled by the actions of a self-serving management. As a consequence, they might whistle-blow such actions to the public or threat to leave the company. Finally, the corporate charter itself represents an important internal corporate 64

65

66

67 68

The fact that stock ownership is an important element of executive compensation has been observed by Lewellen (1969). He analyzes top executive earnings from fixed dollar rewards and from stock-based rewards for U.S. Fortune 500 companies and concludes that the stock component rose from 1-3% during 1940-1948 to 43-49% during 1960-1963. See Lewellen (1969), p. 312. However, Sanders (2001) argues that stock ownership and stock option pay have asymmetrical risk properties and hence managers might respond in different ways. See also von Heusinger (2004), p. 36, who argues that this aligns management’s incentives with those of short-term speculators rather than with long-term investors. Cf. Brick, Palmon and Wald (2006), pp. 403-423; Denis, Hanouna and Sarin (2006), pp. 467-488; Kaserer (2005), p. 29. The causes and effects of the total separation of ownership and control, which is the case for the typical widely held corporation, will be covered in detail in section 3.1. See Harvey, Lins and Roper (2004), pp. 5-7. Cf. Elston (2004), pp. 65-70. Cf. Jensen (1986) who discusses the role of debt in motivating organizational efficiency.

Fundamentals of Corporate Governance

21

governance feature: Especially in the U.S., e.g. poisson pills or shareholder right plans often raise barriers against potential takeover threats; thus, charter amendments should be seen as an element of the internal control system.69 While internal control takes place within the contractual nexus of the company, external control refers to the disciplining effect which stems from the competition to which companies are exposed in several markets.70 Next to the capital market, which will be addressed in more detail, this also includes the product, factor and external labor market for managers.71 The argument is that by the general rules of the market unsuccessful participants will be driven out of the market by fierce competition.72 In the capital market firms compete for outside funding in the primary capital market, where investors also evaluate management’s past efforts and companies’ future prospects before making their investment decisions. The secondary market does not provide funding to the company but indirectly influences its chances to raise money on the primary market in the future as it indirectly influences companies’ cost of equity. Therefore, management is disciplined indirectly by investors’ scrutiny and more directly by the market for corporate control, which has to be seen as a special type of secondary market control. The market for corporate control should force managers to act in the best interest of the shareholders. Otherwise current owners are likely to sell their stakes to those who see the company currently undervalued because of poor management and are willing to provoke a change in management.73 Hence, investors or corporations bidding for control over companies, which are perceived to suffer under the management in place, also may serve as a disciplinary device.74 As a consequence, defensive measures on behalf of the

69 70 71 72

73

74

See Gillan (2006), pp. 388-389. See Bott (2002), p. 3. Cf. Goergen, Manjon and Renneboog (2004), pp. 16-21. See Fama (1980); Easterbrook and Fischel (1982), p. 701; Hart (1983); Jensen and Ruback (1983). The main problem with regard to the control through product and factor markets — though inevitable effective in the long run — is their relative slowness and their inability to preserve investors’ capital from expropriating actions by the management once the capital is sunk. See Jensen (1993), p. 850; Shleifer and Vishny (1997), p. 738. Jensen and Ruback (1983), p. 6, view “[. . . ] the market for corporate control [. . . ] as a market in which alternative managerial teams compete for the rights to manage corporate resources” and, hence, as an “[. . . ] important component of the managerial labor market”. See Ulen (1993), p. 313; Jensen (1993), pp. 850-852; Goergen, Manjon and Renneboog (2004), pp. 16-21. For a detailed description of different corporate control events (i.e. proxy contests and takeover attempts) see Weston, Siu and Johnson (2001), pp. 611-617; Harris and Raviv (1988a), p. 55-75.

22

Fundamentals of Corporate Governance

target’s management are often perceived as attempts to evade market discipline and hence are associated with wealth losses for target’s shareholders.75 Another increasingly important corporate governance force are private sources of external oversight as the media or lawsuits. Particularly, the recent U.S. corporate scandals of Enron, Inc. or Worldcom, Inc. attracted broad media attention, which in turn put pressure on officials to indict assumedly guilty management board members.76 Furthermore, single journalists — e.g. Bethany Mclean of Fortune Magazine — played a key role in disclosing problems in the case of Enron, Inc. Similarly, contingent suits (including shareholder class actions, breach of contract or patent infringement suits) might deter management and supervisory board members from malpractice.77 At last, the special role of the shareholder structure as a control mechanism should be noticed. Beyond doubt, managerial or insider ownership should be classified as an element of internal ‘control’. On the other side, a takeover attempt by a (non-shareholding) company probably has to be categorized as an instrument of outside control. However, sometimes it might be difficult to clearly categorize ownership concentration as an internal or external control mechanism: Those owners could express their desire for control by aiming at seeds in the supervisory board (internal control) or by participating in or initiating proxy contests (outside control). Thus, the shareholder structure takes a hybrid position within the proposed control classification.78 2.3

Basic Types of Corporate Governance Systems

Corporate governance systems could be characterized as a combination of internal and outside control mechanisms, whereas the individual mechanisms stand in interdependent relationships and carry different weights across countries.79 While the boundaries between individual corporate governance systems have 75 76

77 78 79

See Dann and DeAngelo (1988), pp. 87-127. Besides, these scandals also provoked changes in legal regulations. For example, the introduction of the Sarbanes-Oxley Act (SOX) in 2002 aimed at increasing the personal liability of corporate managers. See Gillan (2006), pp. 394-395. See Bott (2002), p. 4. Goergen, Manjon and Renneboog (2004), p. 2, define corporate governance regimes (or systems) as “[. . . ] the amalgam of mechanisms which ensure that the agent (the management of a corporation) runs the firm for the benefit of one or multiple principals (shareholders, creditors, suppliers, clients, employees and other parties with whom the firm conducts its business).” Cf. Agrawal and Knoeber (1996); Berry, Fields and Wilkins (2006).

Fundamentals of Corporate Governance

23

been blurred, it is widely accepted to differentiate between two polar forms: the capital market oriented system and the corporate law oriented system. Even though numerous different labels are used for these two types of corporate governance systems, most differentiations consistently refer to the U.S. (or U.K.) and Germany (or Japan) respectively as polar prototypes.80 2.3.1

The Capital Market Oriented System

Capital market oriented systems rely on a high degree of investors’ legal protection, thereby encouraging individuals to invest in the capital market. Furthermore, these countries often are characterized by capital-based pension systems which further promotes the demand for securities in those countries. As a result, numerous and even relatively small firms are able to raise capital in the public capital market, e.g. by issuing shares on the equity capital market. Usually manifold safeguards for investor protection are in place: Minority rights are explicitly protected, transaction costs are kept low, accounting rules are oriented towards shareholder protection and shareholders have the possibility to remove directors without cause and sue them through class-actions for violations of fiduciary duty.81 As a consequence, capital market oriented systems are marked by higher market capitalizations in relation to the gross domestic product (GDP) as well as higher stock exchange turnover than corporate law oriented systems.82 Ownership concentration traditionally plays a subordinate role, except that ownership is concentrated only temporarily during takeover events. Nevertheless, the market for corporate control plays an important role in disciplining man80

81

82

Other notations include amongst others: shareholder-value vs. stakeholder system, outsider vs. insider system, capital market based vs. bank- and corporation based system, Anglo-American vs. Continental European system, common law vs. civil law systems, willingness to accept conflicts vs. consensus system, flexibility vs. stability of business relations system, dispersed ownership vs. concentrated ownership system, market-based vs. control-based systems and market-oriented vs. network-oriented system. See Matthes (2000), p. 21; Carlin and Mayer (2000), pp. 137-144; Andreani (2003), pp. 2-7; Berrar (2001), pp. 36-41; Coffee (2001), p. 3; La Porta, Lopez-de Silanes, Shleifer and Vishny (1998), pp. 1117-1126; Thomsen, Pedersen and Kvist (2006), p. 249; Van der Elst (2000), p. 4. See Shleifer and Vishny (1997), pp. 769–770; Dietl (1998), pp. 147-155; Scott (1999), p. 6. Note that U.S. corporate law falls under state legislature and hence differences in regulations between states do exist. In 2004, the market capitalization in relation to the GDP for the U.S. (139%) and for the U.K. (135%) was significantly higher than for Japan (76%) or Germany (44%). In fact, within the EU only Hungary (19%), Poland (30%) and Austria (30%) show lower ratios than Germany. Front-runner is Switzerland with 231%. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 05-3. Similarly, stock exchange turnover in percent of GDP in 2000 is significantly higher for the U.S. and U.K. (both 318%) than for Germany (112%). See Helmis (2002), p. 46. Cf. Dietl (1998), p. 121; Van der Elst (2000), pp. 8-9; OECD (1998a), p. 18.

24

Fundamentals of Corporate Governance

agers as the numerous (hostile) takeovers in the U.S. demonstrate.83 Likewise the position of banks — or more general creditors — is comparatively weak as extensive bankruptcy protection regulations limit their rights in cases of corporate crises.84 Moreover, banks were prohibited from owning directly large shares in industrial corporations by the Glass-Stegall Act for a long time, which limited their opportunities to gain weight within the internal control mechanisms, e.g. by sending representatives to the board of directors.85 Furthermore, interests of stakeholders other than shareholders are less protected and codetermination or a comparable form of representation of employees’ interests is not institutionalized in the one-tier board systems of the U.S. or the U.K.86 Notably, corporate governance systems appear to be rather dynamic than stable, thereby reacting on changes in the environment. For the example of the U.S., the traditional high influence of insurance companies, mutual funds, and banks was temporarily restricted by several laws in the aftermath of the great depression at the beginning of the 20th century. However, during the last two decades — and supported by recent decisions of the Securities and Exchange Commission (SEC) in 1992 and 1997, which promoted communication among shareholders — these institutional investors regained a more active role in corporate governance.87 Another trend can be seen in the increasing regulation: For example, the introduction of the Sarbanes-Oxley Act (SOX) in 2002 aimed amongst others at increasing corporate managers’ personal liability and improving the effectiveness of annual audits. Regarding the one-tier board system, there are signs that investors increasingly force companies to separate the roles of CEO and board chairman; a split of duties and responsibilities that is institutionalized in the German two-tier board system.88

83

84 85

86 87 88

For a comprehensive review of the value increase in (hostile) acquisition targets see Weston, Siu and Johnson (2001), pp. 199-222, 615. See Shleifer and Vishny (1997), pp. 769-770. See Roe (1990). In 1999, the U.S. Congress repealed the Glass-Stegall Act, thereby lifting restrictions which prevented U.S. banks from directly owning large U.S. equity stakes. See Matthes (2000), pp. 20-21. See Gillan and Starks (2003), pp. 9-10. See Felton (2002), pp. 28-41. Cf. section 2.4.4.

Fundamentals of Corporate Governance 2.3.2

25

The Corporate Law Oriented System

In corporate law oriented systems internal control mechanisms prevail. In theory, the protection of shareholders’ interests is primarily enforced by their legal representatives, i.e. for the case of Germany by the supervisory board, which is elected by shareholders at the annual shareholders’ meeting. While in Germany the two-tier board system (consisting of management and supervisory board) is still compulsory, other Continental European countries, i.e. France in 1966 and Italy in 2003, provided companies with a choice to implement one-tier AngloSaxon or two-tier German type board models.89 In Germany the explicit provision for employees’ interests — as natural stakeholders of the company — is guaranteed by the system of codetermination, a peculiarity of the German corporate governance system, which forces management to balance the coalition of interests between shareholders and employees.90 Banks usually play a predominant role in corporate law oriented systems as it is the case in Germany or Japan:91 First, they provide extensive long-term outside financing to companies in exchange for powerful control rights (so-called “house bank principle”). Second, they traditionally hold large equity stakes of their clients enabling them to actively participate in the internal control system by sending bank representatives to the supervisory board (or the board of directors). Third, proxy voting empowers banks to exercise additional control in excess of their voting rights stemming from holding direct equity stakes. The dominance of banks in corporate law oriented systems comes along with financial disclosure and auditing provisions which are primarily orientated at creditors’ needs. External control through the capital market is traditionally weak and hostile takeovers are rare. This could be partly explained by a high degree of ownership concentration and permanent large shareholders.92 These large shareholders have the possibility to participate actively in the control of 89

90

91 92

See Hopt and Leyens (2004), pp. 15-18. However, the recent establishment of a new type of company, i.e. the European Stock Corporation (“Societas Europaea” (“SE”)), also provides German companies with the theoretical option for a one-tier board. Cf. section 2.4.4. See Hopt and Leyens (2004), pp. 7-9. Similar, but less stringent forms of codetermination exist in Hungary, Slovenia, Slovakia, Poland, Austria, Luxembourg and the Netherlands. See Claassen (2003), p. 177; Wirtschaftsrat der CDU e.V. (2005), p. 7. For the development of codetermination in Germany cf. section 2.4.1. See Scott (1999), pp. 6-7. K¨oke (2002), pp. 47-74, argues that the dealing with large blockholdings might represent an effective substitute for the rather inactive German takeover market.

26

Fundamentals of Corporate Governance

companies (i.e. exercise “voice”). Contrarily, small investors lack opportunities to sell their stakes in underperforming companies to outside bidders trying to provoke changes in management (exercise “exit”) because the market for corporate control is rather inactive.93 The alternative sale of their shares on the stock exchange might only be possible at depressed prices as other investors can be expected to be aware of the incapable management as well. Relatively weak regulations concerning insider trading and market manipulation discourage small investors from directly investing into the capital market and promotes financial intermediation.94 2.3.3

Which Corporate Governance System is Best?

The diversity of corporate governance systems is remarkable: The U.S. and the U.K. systems are primarily capital market oriented and focus on a high degree of protection of investors. Germany and Japan rely on the surveillance through large investors and banks and in almost the rest of the world ownership concentration in family conglomerates prevails supported sporadically by some large investors and banks.95 Given the variety of corporate governance systems, one might be induced to ask which system is best or most effective?96 Even though one can hardly come up with a clear-cut answer, it appears worthwhile to investigate the outcome of the two polar prototypes. But what is the outcome of a corporate governance system and how could it be measured?97 At the first glance, the control efficiency appears to be the most obvious measure. Theoretically, it should be higher in corporate law oriented systems (i.e. insider systems) because the controlling parties are expected to have an informational advantage compared to outsiders. However, the risk of collusion between management and supervisory board or banks is high. For example, banks might prefer to get along with management because they fear negative consequences for their reputation if their refusal to prolong credits for underperforming compa93

94 95

96 97

Hirschman (1970) formulated the modes of “voice” and “exit” as two forms of behavior by which individuals can participate in a coalition of individuals. For the transfer of this concept to corporate governance see Schmidt (2003), p. 13; Mann (2003), pp. 45-47. See Matthes (2000), pp. 18-19; Dietl (1998), pp. 111-126; Shleifer and Vishny (1997), pp. 771-773. See Shleifer and Vishny (1997), p. 769; La Porta, Lopez-de Silanes, Shleifer and Vishny (1998); La Porta, Lopez-de Silanes and Shleifer (1999). Cf. Milbourn (1998), p. 174. See Scott (1999), p. 8.

Fundamentals of Corporate Governance

27

nies, as a means of exercising control, leads to insolvency and finally puts jobs at risk.98 Thinking one step further, control efficiency should be reflected in companies’ economic performance. In theory, capital market oriented systems should result in more efficient use of resources and hence better economic performance. Notwithstanding the plausible theoretical argument, empirical proof is hardly possible, since many other factors (e.g. labor costs and regulation) influence the economic performance as well.99 Finally, the adaptability of the corporate governance system might be an important feature. In the long run, the stability of the corporate governance system in Germany100 leads to fewer changes and lower transaction costs. However, in a fast changing economic environment delays in the adoption to new circumstances might cause significant costs. The relatively undeveloped capital markets in corporate law oriented systems make resource allocation less efficient, widespread grandfathering hinders structural change, and the consensus orientation slows decisions.101 To sum up, both types of corporate governance systems appear to have their strengths and both systems have shown flaws in the past: The excessive dispersion of ownership in U.S. in the early 20th century as well as the very close alliances between management board and supervisory board in Germany appeared to represent suboptimal developments.102 S HLEIFER AND V ISHNY (1997) draw the conclusion that “[. . . ] corporate governance systems of the United States, Germany, and Japan have more in common than is typically thought, namely a combination of large investors and a legal system that protects investor rights. Corporate governance systems elsewhere are less effective because they lack the necessary legal protections.”103 Besides the notion that a minimum of legal protection and ownership concentration is necessary for an 98

99

100 101

102 103

See Matthes (2000), pp. 22-25. German cases where tight relationships between managing and supervising bodies delayed changes (and sometimes even triggered crises) include amongst others Deutsche B¨orse AG (2005), Balsam AG (1994), and Metallgesellschaft AG (MG) (1993). See Albers (2002), pp. 225-250. For the U.S., Jensen (1993) observes the failure of insider systems in practice. Cf. Felton and Watson (2002), pp. 31-45. See Matthes (2000), pp. 25-26. Nevertheless, several hints suggest that e.g. the comparatively low German capital productivity hurts economic performance. See Januszewski, K¨oke and Winter (1999); B¨orsch-Supan (1998). See section 2.4.1. See Matthes (2000), pp. 27-28. Cf. Scott (1999), pp. 8-12; Deutsche Bank (2002), pp. 15-20; Wirtschaftsrat der CDU e.V. (2005), pp. 6-10. See Matthes (2000), p. 56. Shleifer and Vishny (1997), p. 770. Cf. Schmidt (2003), p. 7, and Mann (2003), pp. 126-129, who underline the importance of consistency in corporate governance systems because of their systemic character.

28

Fundamentals of Corporate Governance

efficient control system, their judgement gave rise to the hypothesis that both systems are converging. The reasons for this development are twofold: On the one side, experiences with the unilateral reliance on inside or outside control mechanisms were rather disappointing. On the other side, cross country initiatives, as the OECD Principles of Corporate Governance or certain EU directives104 , actively promote the unification of corporate governance systems.105 However, it is commonly perceived that in the course of this conversion process corporate law systems are approaching capital market systems rather than the other way around.106 2.4

The German System of Corporate Governance

In accordance with the applied understanding of corporate governance, which restricts it to the case of publicly traded companies, the following section will draft the historical development of corporate law up to now, examine the current formal and informal features of the German system of corporate governance and provide a compact perspective on the future development. 2.4.1

Historical Development

The first codification of German company law, which also included the public limited company (“Aktiengesellschaft” (“AG”)), dates back to the German Commercial Code (“Handelsgesetzbuch” (“HGB”)) of 1861.107 Even before the Commercial Code has been implemented by all German states, the amendments of 1870 abolished the requirements of government permits and government control of public limited companies.108 At the same time, a mandatory supervisory board was established giving birth to the two-tier board system, which until now is often seen as the distinguishing mark of German corporate governance par excellence. With the introduction of the supervisory board, the basic structure for the distribution of rights and duties among shareholders, supervisory and management board was laid out. Because of the easements to found public 104 105

106 107

108

See sections 2.4.1 and 2.4.2. See Berndt (2002); Nestor (2001), pp. 1-10; Nassauer (2000), pp. 266-279; OECD (2004), pp. 11-66; Cromme (2002b), pp. 119-121; Coffee (2001), pp. 12-24. For a discussion of the convergence of European corporate governance systems see Hopt and Leyens (2004), pp. 19-23; Berrar (2001), pp. 64-78. See Helmis (2002), p. 27. The roots of company law go back to medieval times or even further back to pre-medieval and Roman law. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 34. Cf. Schmidt (1986). See Ernst (2001), p. 4.

Fundamentals of Corporate Governance

29

limited companies grievances of defrauded investors initially soared. As a consequence, the amendments of 1884 specified details with respect to the founding procedure and the valuation of assets and strengthened the control rights and duties of supervisory board members, who also became liable for certain malpractices. Since smaller businesses felt overstrained with the new and more restrictive requirements, in 1892 the form of a private limited liability company (“Gesellschaft mit beschr¨ankter Haftung” (“GmbH”)) was established, which enabled direct shareholder control on management and was suitable for companies with a limited number of shareholders.109 At the turn of the century, the public limited company has proofed survivable and a couple of basic safeguards for shareholder and creditor protection were in place. The next milestone, which until now brands the German corporate governance system, represented the evolution of codetermination (“Mitbestimmung”). The development of codetermination can be traced back to as early as 1891, when modifications to existing law enabled larger companies to install works councils (“Betriebsr¨ate”). After initially becoming mandatory only for larger firms in 1916, two years later also smaller firms had to establish them. Through the Works Council Act (“Betriebsr¨ategesetz”) of 1920, these institutions were to appoint one or two employees to serve as members of the supervisory board. During the national socialist government from 1933 to 1945, employees’ rights were temporarily cut back and in 1937 the regulations concerning public limited companies were taken out of the Commercial Code, revised, and pooled in a separate act, i.e. the Stock Corporation Act (“Aktiengesetz” (“AktG”)). In the 1965 reform of the Stock Corporation Act, the law was again adjusted to market oriented principles.110 After the end of World War II, labor representatives regained influence and finally enforced that labor and shareholder representatives shared seats in the supervisory board. Later, in 1951 the Montan-Codetermination Act (“Montan-Mitbestimmungsgesetz”) dictated that in all companies employing more than 1,000 persons in the steel and mining industry both parties share seats in the supervisory board in equal parts. In 1952, the Works Constitution Act (“Betriebsverfassungsgesetz” (“Be109

110

See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 33-35. Cf. Schubert and Hommelhoff (1985). See Schneider and Heidenhain (2000), pp. 9-10; Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 35-37, 194-230. Cf. Schubert and Hommelhoff (1985).

30

Fundamentals of Corporate Governance

trVG”)) allocated one third of supervisory board seats to labor representatives in all companies with more than 500 employees. Finally, the Codetermination Act (“Mitbestimmungsgesetz”) of 1976 transferred the allocation of supervisory board seats in equal parts to shareholder and labor representatives to all public and private companies with more than 2,000 employees independently of their legal form or industry.111 The historical strong influence of labor representatives in Germany, amongst others, led to the “[. . . ] limited acceptance of the view that the large company is an instrument of the shareholders. It is frequently seen as a coalition of its stakeholders, and a higher priority is attributed to the survival of this entity than to any claim or interest of any stakeholder or stakeholder group.”112 As a consequence, the business objective of the public limited company in Germany became complex and vague and executives had to be “[. . . ] socially responsible skillful balancers of stakeholder interests”113 and it became “[. . . ] at best difficult to judge the performance of the executive board.”114 Even though shareholders formally appointed the majority of supervisory board members, proxy voting by banks or lacking coordination among individual shareholders led to corporate control problems and an inferior perception of corporate governance as a means of preserving shareholders’ interests.115 During the 22 year period from 1976 to 1998, the German corporate governance system was marked by an extraordinary high degree of stability. Hence, hardly any changes to the legal structure of the public stock corporation or the balance of power between the different stakeholder groups took place.116 The ongoing debate about corporate governance started with criticism on legal regulations

111

112

113 114 115 116

As the supervisory board chairman has two voting rights in the case of stand-offs, it is guaranteed that no decisions against the shareholders can be taken because the chairman is elected by a two third votes cast or — in case of a stand-off — by the shareholders’ representatives in a second ballot. See Gorton and Schmid (2002), p. 7; Dietl (1998), p. 113. Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 38. Cf. section 2.2.1 for a discussion of the concepts of shareholder value and stakeholder orientation. Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 38. Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 38. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 38-39. See Schmidt (2003), p. 19. Exemplary changes are the introduction of the Securities Trading Act (“Wertpapierhandelsgesetz” (WpHG)) and changes to the Restructuring Act (“Umwandlungsgesetz (UmwG)”) in 1995. See Goergen, Manjon and Renneboog (2004), pp. 22-25.

Fundamentals of Corporate Governance

31

with respect to the supervisory board.117 Whilst a comprehensive overview over the most recent changes in regulations concerning corporate governance is given in table 2-1, only selected aspects can be commented in more detail. An important impetus was brought by the introduction of the Law on Control and Transparency in the Corporate Sector (“Gesetz zur Kontrolle und Transparenz im Unternehmensbereich” (“KonTraG”)) in 1998. With this law, the legislator extended the obligations for executive care and management board’s obligations to report as well as strengthened the quality of supervisory board supervision and — with a two year transition period — abolished plural voting rights.118 In 2000, efforts to establish the so-called German Corporate Governance Code (“Deutscher Corporate Governance Kodex” (“DCGK”)) for publicly listed companies started. This code constitutes a compendium of German corporate governance standards by summarizing existing regulations and giving continuative recommendations.119 After the appointment of two government commissions, the first version of the DCGK became effective in 2002 through the adoption of the Transparency and Disclosure Law (“Transparenz- und Publizit¨atsgesetz” (“TransPuG”)). The TransPuG forced public limited liability stock corporations to declare whether they comply with the recommendations of the code and to explain possible deviations (“comply or explain principle”). The code is subject to regular reviews and was lastly modified in June 2005.120 Finally, the 4th Financial Market Promotion Law (“4. Finanzmarktf¨orderungsgesetz” (“4. FMFG”)) aimed to enhance investor protection by increasing transparency on security markets and prosecuting market manipulation more effectively.121

117

118

119

120

121

See Matthes (2000), p. 24. Another source of the current discussions are regulations by the European Union (“EU”) which required the national legislators to accept foreign legal forms of business organization on an equitable basis. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 39-41. See Matthes (2000), pp. 24, 55; OECD (1998b), pp. 19-25. For a comprehensive analysis of the development and effects of the KonTraG see Berrar (2001), pp. 52-63; Albers (2002). For an overview about the DCGK see Brink and Romeike (2005), pp.14-34. Similar corporate governance codes were also introduced in the U.K. (2003) and the Netherlands (2004). See Brink and Romeike (2005), pp. 156-169. A current version of the Kodex can be obtained at the official website of the government committee at http://www.corporate-governance-code.de/. For studies about the development and effects of the German Corporate Governance Kodex see Nowak, Rott and Mahr (2004); Bassen, Kleinschmidt, Prigge and Z¨ollner (2005); ergo Kommunikation (2005). Drobetz, Schillhofer and Zimmermann (2004) analyze the relationship between a self constructed corporate governance index, addressing issues very similar to those of the DCGK, and performance. See Bundesministerium der Finanzen (2006).

32

Fundamentals of Corporate Governance

Year

CG-Initiative/ Lawa

1998

3. FMFG

1998

KonTraG

1999

OECD Principles of Corporate Governance

2000

Code of Best Practice German Code of Corporate Governance 1. Government • “Corporate Governance — Corporate Management – Committee “CorpoCorporate Control – Modernization of Stock Corporarate Governance” tion Law” • Chairmanship: Prof. Theodor Baums • Numerous recommendations for the modernization of Report of the 1. Government existing regulations and a proposal to introduce a CorCommittee porate Governance Code for public stock corporations 2. Government • Chairmanship: Dr. Gerhard Cromme (Supervisory Committee “Corboard chairman of ThyssenKrupp AG) porate Governance Kodex” ¨ WpUG • Acquisition and Takeover Act (“Wertpapiererwerbs¨ und Ubernahmegesetz”) • Sets 30% threshold for mandatory offers DCGK • German Corporate Governance Code (“Deutscher Corporate Governance Kodex”) • Code for public stock corporations; yearly declaration of conformity according to the “comply-or-explain principle”; annual review TransPuG • Transparency and Disclosure Law (“Transparenz- und Publizit¨atsgesetz”) • Legal anchorage of the CG-Kodex in § 161 AktG 4. FMFG • 4. Financial Market Promotion Law (“4. Finanzmarktf¨orderungsgesetz ”)

2000

2000

2001

2001

2002

2002

2002

2002 a

Summary Description • 3rd Financial Market Promotion Law (“3. Finanzmarktf¨orderungsgesetz”) • Over 100 separate provisions to promote liberalisation and deregulation by e.g. revising stock exchange admission and prospectus requirements or simplifying the listing procedure • Law on Control and Transparency in the Corporate Sector (“Gesetz zur Kontrolle und Transparenz im Unternehmensbereich”) • The principles are thought as guidelines to assist member and non-member governments in their efforts to improve the legal, institutional and regulatory framework for corporate governance • Initiator: Frankfurter Grundsatzkommission Corporate Governance • Initiator: Berliner Initiativkreis

For the sake of clarity legal initiatives of comparatively minor relevance as the Law on Admission of Nonpar Shares (“St¨uckaktiengesetz” (“St¨uckAG”)), the Law to Ease Capital Procurement (“Kapitalaufnahmeerleicherungsgesetz” (“KapAEG”)) or the amendments to the Securities Trading Law (“Wertpapierhandelsgesetz” (“WpHG”)) are not explicitly listed. Source: Author’s illustration. See OECD (1998b), pp. 17-25; Theisen (2003), pp. 442-451.

Table 2-1: Corporate Governance Initiatives 1998-2002

Fundamentals of Corporate Governance 2.4.2

33

Status Quo

Once the evident underdevelopment of the German capital market was realized122 , numerous efforts to improve the efficiency of the capital market and corporate governance system have taken place during the past few years. Based on the ten-point program of the German Federal Government presented in February 2003, several old regulations have been modified and new ones have been introduced. While table 2-2 provides a comprehensive overview, only selected initiatives will be outlined here. In October 2004, the Investor Protection Improvement Act (“Anlegerschutzverbesserungsgesetz” (“AnSVG”)) was introduced which ratified the EUdirective regarding insider regulations, ad-hoc publicity and market manipulation rules. In January/July 2005, the Accounting Law Reform Act (“Bilanzrechtsreformgesetz” (“BilReG”)) and the Accounting Enforcement Act (“Bilanzkontrollgesetz” (BilKoG”)) were installed with the purpose of unifying accounting rules and safeguarding their proper use. Thereby a two-stage enforcement process was established.123 After a long-standing public discussion, the Management Compensation Disclosure Act (“Vorstandsverg¨utungsOffenlegungsgesetz” (“VorstOG”)) became finally effective in August 2005 and forced the disclosure of individual management board members’ fixed and variable compensation constituents.124 By then, only the aggregated compensation for all management board members had to be published in the annual reports. At the end of 2005, the introduction of the Capital Markets Model Case Act (“Kapitalanleger-Musterverfahrensgesetz” (“KapMuG”)) provided aggrieved investors with a more efficient mechanism to sue for damages due to false or delusive capital market information (e.g. regarding ad-hoc news or security prospectus). Finally, in May 2006 a draft of the Transparency Directive Ratification Act (“Transparenzrichtlinie-Umsetzungsgesetz” (“TUG”)) was presented, which would establish additional thresholds for the mandatory disclosure of shareholdings. As a result, supplementary to the 5%, 10%, 25%, 122 123 124

See Monopolkommission (1998), pp. 18-63. See von Rosen (2004), pp. 330-331; Blum (2005), pp. 131-142. In April 2005, 21 of the 30 largest German companies (DAX) disclosed the remuneration of its management board or its CEO. However, critics claim that even if the absolute level of remuneration is disclosed more transparency about the structure of the salary system is needed. See Fockenbrock (2005b), p. 2. Cf. Gillmann (2004), p. 3; Commission of the European Communities (2004), pp. 6-7; Koenen (2004), p. 2.

34

Fundamentals of Corporate Governance

Year

CG-Initiative/ Lawa

02/2003

Catalog of Countermeasures for the Fortification of Corporate Integrity and Investor Protection AnSVG

10/2004

01/2005

01/2005

07/2005

07/2005

08/2005

11/2005

11/2005

02/2006

05/2006

a

Summary Description • Ten-point program aiming at increasing stock market transparency and investors’ protection • Topics: liability of management/supervisory board members, DCGK, accounting and auditing

• Investor Protection Improvement Act (“Anlegerschutzverbesserungsgesetz”) • Changes to insider trading, ad-hoc publicity and market manipulation regulations Investment Modern- • Investment Modernization Act (“Investmentmoderniization Act sierungsgesetz”) • First-time accreditation of hedge funds BilReG • Accounting Law Reform Act (“Bilanzrechtsreformgesetz”) • Internationalization of accounting rules and fortification of the auditor’s independence BilKoG • Accounting Enforcement Act (“Bilanzkontrollgesetz”) • Establishes two-stage enforcement procedure for accounting rules; revision of annual reports Law Implementing • Law Implementing the Prospectus Directive (“Prothe Prospectus Dispektrichtlinie-Umsetzungsgesetz”) • Ratification of EU-directive for prospectus’ rective VorstOG • Management Compensation Disclosure Act (“Vorstandsverg¨utungs-Offenlegungsgesetz”) • Disclosure of the compensation (fixed and variable) of individual management board members of publicly listed companies UMAG • Act on Corporate Integrity and Modernization of Rescission Law (“Gesetz zur Unternehmensintegrit¨at und Modernisierung des Anfechtungsrechts”) • Changes to management/supervisory board’s liability KapMuG • Capital Markets Model Case Act (“KapitalanlegerMusterverfahrensgesetz”) • Introduction of a mechanism for group actions ¨ Takeover Guidelines • Takeover Directive Ratification Act (“Ubernahmerichtnd linie-Umsetzungsgesetz”) Ratification Act (2 reading) • Ratification of EU-directive for takeover practices TUG (Draft) • Transparency Directive Ratification Act (“Transparenzrichtlinie-Umsetzungsgesetz”) • Mandatory disclosure of shareholdings above additional thresholds of 3%, 15%, 20% and 30% (hitherto 5%, 10%, 25%, 50% and 75%)

For the sake of clarity legal initiatives of comparatively minor relevance as the Regulation for the Implementation of the Prohibition on Market and Price Manipulation (“Verordnung zur Konkretisierung des Verbotes der Kurs- und Marktpreismanipulation” (“KuMaKV”)) and the Market Manipulation Definition Regulation (“Marktmanipulations-Konkretisierungsverordnung” (“MaKonV”)) are not explicitly listed. Source: Author’s illustration. See Bundesministerium der Finanzen (2006).

Table 2-2: Corporate Governance Initiatives 2003-2006

Fundamentals of Corporate Governance

35

50% and 75% thresholds, shareholdings above 3%, 15%, 20% and 30% would have to be disclosed on part of the shareholders.125 While commenting on the content of these legal initiatives is beyond the scope of this synopsis, the general focus of these regulations is to increase corporate transparency, improve investors’ rights, improve supervision by the supervisory board126 and to implement sanctions for cases of security fraud, which was historically hardly pursuable under German legislation.127 Outside control through the capital market has slightly improved but is still comparatively underdeveloped. Public takeover bids are still an exception, which is last but not least due to the lack of volume in number and substance of publicly quoted companies.128 Other reasons include the extensive capital- and personnel linkages — and especially cross holdings129 — among public stock corporations, the role of banks in corporate financing, and management’s profit distribution capacity.130 Since these structural characteristics of the German capital market cannot radically be altered in the short-term, the focus of current corporate governance discussion lies on the improvement of internal control mechanisms, and especially on the role of the supervisory board and the compensation of management and supervisory board members. For the former, the once rather passive monitoring role is in the process of changing towards a more active advisory role. This development increases the required qualifications of supervisory board members, asks for the reduction of simultaneously exercised supervisory board mandates, and postulates reforms towards smaller board sizes, which until now are bound to minimum sizes prescribed by the 125 126

127

128 129

130

For a detailed description of disclosure requirements regarding shareholdings see section 6.2.2. The awareness of reform needs for the two previously named aspects is anything but new. Already in 1884, the legislator aimed for the empowerment of the shareholders’ meeting and the specification of supervisory board’s tasks, when the Stock Corporation Act was reformed. See Schubert and Hommelhoff (1985), p. 85. In 2003, the responsible Bundesamt f¨ur Finanzdienstleistungsaufsicht (BaFin) (“Federal Financial Supervisory Authority”) reported that a total of 83 actions related to insider trading were finalized. Of those only five actions led to a conviction. In seven cases the proceedings were stopped against the payment of a fine. One action ended with an acquittal. The remaining 70 proceedings were stopped by the public prosecutor’s office. See Bundesanstalt f¨ur Finanzdienstleistungsaufsicht (BaFin) (2004), p. 183. See footnotes 82 and 143. Since the abolishment of capital gain taxes for corporations on January 1, 2002, the network of cross holdings in Germany thinned out. See H¨opner and Krempel (2005), pp. 10-11; Fockenbrock and Stratmann (2006), p. 14. See Monopolkommission (1998), pp. 79-80; J¨urgens, Rupp and Vitols (2000), p. 5; Adams (1999), pp. 80109. For the extent of capital- and personell linkages among German public companies see Monopolkommission (2005), pp. 246-287. For the development of corporates’ and banks’ shareholdings from 1993 to 2003 cf. section 6.4.

36

Fundamentals of Corporate Governance

German codetermination law. For the latter, individual compensation should become even more transparent and disclosed and should be related to shortterm as well as long-term performance.131 Finally, current developments are increasingly provoked by European initiatives and EU directives or regulations: For example, one of the 24 propositions of the “Plan to Move Forward — Modernising Company Law and Enhancing Corporate Governance in the European Union” of the European Commission suggests that outgoing CEOs normally should not directly take over supervisory board chairmanship, which obviously would represent a rigorous renunciation from common practice in Germany.132 2.4.3

Characteristics and Idiosyncrasies

As already mentioned in the previous sections the German corporate governance system can hardly be described without pinpointing certain characteristics.133 Not shareholders in general but three influential stakeholder types exhibit great influence: blockholders, employees and banks.134 Ownership concentration is widespread among listed German stock corporations and most companies have at least one major outside shareholder, i.e. blockholder135 . Later on in the context of the empirical study (section 6.4), it will be shown that at the end of 2003, 62.1% of all companies in the examined sample had at least one blockholder and 36.2% had at least one blockholder with a stake exceeding 25% (cf. tables 6-11 and 6-14).136 Besides the strong influence of blockholders in general, cross holdings, i.e. bilateral holdings among large public corporations, stigmatized the German capital market as the “Deutsch-

131 132

133

134

135 136

See Cromme (2002b), pp. 123-129; Cromme (2003), pp. 141-151; Jensen (1993), p. 865. Cf. § 87 AktG. See Wirtschaftsrat der CDU e.V. (2005), pp. 20-24; Commission of the European Communities (2003b). For recent developments in German corporate Governance cf. Goergen, Manjon and Renneboog (2004). Wenger and Kaserer (1998b), p. 41, describe the German system of corporate governance as “[. . . ] a system of noncontrol that protects managers against the pressure that results from active capital-market oversight.” See Schmidt (2003), p. 11; Matthes (2000), p. 57; Goergen, Manjon and Renneboog (2004), p. 2. J¨urgens, Rupp and Vitols (2000), p. 5, add the predominant technology and production focus of management in industrial companies as an additional feature. In section 6.3 blockholders will be defined as those outside shareholders which own a stake of at least 5%. For evidence about the prevalence of blockholders see amongst others Becht and B¨ohmer (2003); K¨oke (1999); Edwards and Weichenrieder (2004); La Porta, Lopez-de Silanes and Shleifer (1999).

Fundamentals of Corporate Governance

37

land AG” (“Germany, Inc.”).137 Since (unrelated) corporations build the largest blockholder group, a significant part of monitoring power remains in the hands of an owner group, which itself is subject to agency-conflicts. Hence, long-term business relationships between blockholders and companies prevail over clean cut outside monitoring and underline the stakeholder orientation of the German corporate governance system.138 However, the possible effects of blockholders on the efficiency of control are twofold. On the one hand, the presence of large shareholders might be required up to a certain degree to avoid the free-rider problem of corporate control.139 On the other hand, the interests of blockholders are likely to differ from those of other small shareholders, giving rise to additional conflicts of interest.140 As shown in the historical abstract of German corporate governance, employees traditionally represent a very powerful stakeholder group. In international comparison German codetermination, i.e. the institutionalized presence of employee representatives in the supervisory board, constitutes a special case. Opinions about the consequences of German codetermination diverge: While some emphasize that codetermination fosters a more consensual relationship others highlight that firms would voluntarily adopt codetermination without legal pressure if it was beneficial for the company.141 However, the importance of codetermination seems to have declined in the past 20 years: While 49.4% of private sector companies were codetermined in 1984 on either the shop-floor level through works councils or through equal representation on the supervisory board — or both — this number declined to 39.5% in 1994/1996.142 The German takeover market was perceived to be rather inactive for a long time and only few successful hostile tender offers had taken place until 2002.143 However, critics argue that the importance of the takeover market is underesti137

138 139 140 141

142 143

The extent of German cross holdings seems to have declined during the last years. For an extensive analysis see Adams (1999); Faccio and Lang (2002); Monopolkommission (2005), pp. 246-287, 618; H¨opner and Krempel (2005), pp. 11-13; Fockenbrock and Stratmann (2006), p. 14. See Matthes (2000), p. 46; Schmidt (2003), p. 9. Cf. section 3.3.3.1.1. See Shleifer and Vishny (1997), pp. 754-757; Mayer (1998), pp. 151-153; B¨ohmer (2001), p. 111. See Jensen and Meckling (1979), p. 474; Matthes (2000), p. 57; Gorton and Schmid (2002), p. 16. For an overview of studies evaluating the effects of codetermination see J¨urgens, Rupp and Vitols (2000), pp. 9-10. See Gorton and Schmid (2002), p. 38. According to Helmis (2002), p. 52, six hostile takeover attempts had been successful until 2002 (Year: Acquirer/Target): 1989: Stora Enso/Feldm¨uhle Nobel AG, 1991: Krupp AG/ Hoesch AG, 2000: Vodafone plc/ Mannesmann AG, 2001: WCM/Kl¨ocker-Werke AG , 2001: INA GmbH/FAG Kugelfischer AG, 2002:

38

Fundamentals of Corporate Governance

mated as a more discrete form, namely the building of hostile stakes, is more widespread. If one relates the cases of hostile stakebuilding to the number of firms facing a realistic risk of hostile acquisitions – due to their shareholder structure — German hostile “takeover” activity is not less frequent or important than in the U.K.144 Finally, banks play an important role in the German corporate governance system because of several reasons: First, they take a predominant role in providing long-term external funding to corporations.145 Second, their presence in supervisory boards resulting from lending relationships, direct shareholdings146 , and depositary voting rights secures their influence in the internal control system of numerous companies. Third, banks traditionally exercise depositary voting rights on behalf of their clients, giving them considerable voting power far in excess of eventual direct shareholdings.147 In an investigation of the distribution of voting rights at the shareholders’ meetings of the 24 largest German corporations, 13% of banks’ voting rights resulted from direct shareholdings, 10% from bank related investment funds and 61% from depositary voting rights summing up to 84% of total voting rights.148 Finally, the historical prominence of banks as main providers of outside finance comes along with German financial disclosure and auditing rules, which are primarily orientated at creditors’ needs.149 However, several signs indicate that the power of German banks has dwindled. Time will tell whether they can regain their former status or need to refocus on their original role as financial intermediaries.150 2.4.4

Trends and Outlook

From a legal perspective European initiatives and regulations will continue to exert significant influence on German legislation. To promote the integration

144 145

146 147

148

149 150

Barilla Gruppe/Kamps AG. For example, H¨opner and Jackson (2001), pp. 23-43, analyze comprehensively the takeover of Mannesmann AG by Vodafone in 2000. Cf. Franks and Mayer (1990), p. 215. See Jenkinson and Ljungqvist (2001), p. 430. Cf. K¨oke (2002), pp. 47-74. According to Deutsche Bundesbank (2004b), p. 45, for Germany bank credits (corporate bonds) accounted for 38.2% (2.3%) of the GDP in 2003 in contrast to 7.9% (26.5%) in the U.S. See Monopolkommission (2005), pp. 249-255. See J¨urgens, Rupp and Vitols (2000), p. 5; Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 10; Wittkowski (2006), p. 4. Cf. Benders and Drost (2006), p. 23. See Baums and Fraune (1995), p. 103. Cf. Baums (1993); Wenger and Kaserer (1998a); Mann (2003), pp. 189-193; Emmons and Schmid (1998), p. 25. See Dietl (1998), p. 144. See H¨opner and Krempel (2005), pp. 11-13. Cf. footnote 129.

Fundamentals of Corporate Governance

39

of capital markets, in 2001 the member states of the European Union (“EU”) agreed on a uniform legal body, the so-called European Stock Corporation (“Societas Europaea” (“SE”)). While preserving the core elements of codetermination — at least for the case if interests of German employees are involved — it increases the flexibility for open negotiations about the degree of codetermination between employees, management and owners.151 Further, the SE provides alternatives of one-tier or two-tier board structures and the boards become more flexible with regards to their size. There are some signs that this development will also emanate on German corporate law in the long run.152 Furthermore, the adoption of the Financial Services Action Plan (FSAP) in 1999, aiming at the integration of the national markets for financial services, is continuing to exert significant influence on the regulation of German capital market laws.153 Other forces driving change in the German corporate governance system are the rising international competition, liberalization and the introduction of the euro in 1999. The transferability of goods, and especially funds, across borders has increased dramatically and transaction costs have diminished. This development automatically strengthens the mechanisms of outside control, since larger markets with more participants tend to promote intense competition. If corporate control is not effectively exercised in the German capital market, nowadays investors find it easy to allocate their funds to other markets. Certain signs promise a high growth period for the capital market in general: The boom of new financial instruments, especially derivatives, will augment transaction volume and liquidity. An increased pressure for funded — in contrast to the presently common pay-as-you-go — pension schemes and further privatizations will increase primary and secondary capital market demand and supply.154 However, retarding effects might result from current efforts to consolidate public finances and structural weaknesses of the German economic and political system.155 151

152

153 154 155

The insurance company Allianz AG was the first large German corporation to adopt the new type of company. However, it preserved the two-tier board structure with a supervisory board subject to codetermination. See Allianz (2006). See Clement (2004); Wirtschaftsrat der CDU e.V. (2005), pp. 20-24. For example, Cromme (2002b), p. 124, argues that supervisory boards with up to 30 people cannot be efficient and confidential discussions are not longer possible. See Deutsche Bundesbank (2004b), pp. 33-49. See Schr¨oder and Schrader (1998), pp. 19-20; Emmons and Schmid (1998), pp. 32-33. See Matthes (2000), pp. 28-30.

40

Fundamentals of Corporate Governance

Apart from the impact of changes in the broader economic environment and legislative pressure resulting from the European unification process, the balance of power among corporate shareholder (and stakeholder) groups is changing. Power seems to shift towards outside control while internal control simultaneously looses weight.156 This tendency can also be observed if e.g. the future power of employees, blockholders, banks and outside small shareholders is analyzed. For the case of employees, convergence tendencies in the EU put pressure on the traditional form of German codetermination.157 Otherwise, a trend to employee share participation programmes might increase their weight as a shareholder group and hence promote their weight as “outside” monitors.158 The future importance of blockholders appears ambiguous. On the one side, the slowly proceeding disintegration of the “Deutschland AG” reduces their weight. Moreover, trends to focus on core competencies as indicator of an increasing shareholder value orientation, e.g. by outsourcing and spin-offs of business units, also leads to a declining role of large corporate shareholders.159 On the other side, institutional investors evolve as a relatively new power. This investor type, often with Anglo-Saxon origin, forced German corporations to change to a more open information behavior and to professionalize investor relations.160 While the former blockholder type typically owns larger stakes — often above 25% of voting rights — and often is represented in the companies’ supervisory board, the latter usually owns smaller stakes. However, several institutional investors having similar interest combined are increasingly in the position to exercise effective outside control. As a result, blockholders’ control is likely to shift from internal involvement towards outside — and maybe tougher — monitoring.161 156

157 158 159

160 161

Several studies observed a diminishing control premium for the case of Germany, meaning that the relative benefits of active control are decreasing. See Schmidt (2003), p. 32; Dyck and Zingales (2004). Cf. Daske and Ehrhardt (2000), pp. 24-33. See Baums (2002), p. 137. Cf. section 2.4.3. See Matthes (2000), pp. 40-41; Monopolkommission (2005), pp. 257-265; Monopolkommission (2003), pp. 219-225. Cf. tables 6-11 and 6-14 and section 6.4.1. See Black (1992); Gillan and Starks (2003); K¨oke (1999). Cf. section 6.4. A current example is the case of Deutsche B¨orse AG: In 2005, institutional investors opposed the plans of the CEO to acquire the London Stock Exchange. The CEO initially dismissed institutional investors concerns last but not least because of their historical minor weight. However, the institutional investors effectively revolted and finally achieved the resignation of CEO and supervisory board chairman. See Kamp and Krieger (2005), pp. 54-56.

Fundamentals of Corporate Governance

41

From the four analyzed parties, banks are expected to suffer the highest absolute loss in control power. New capital requirements (Basel II) call for a risk adequate pricing of credits and an increasingly intense competition in the European bank market exerts pressure on the traditional German so-called housebank principle (“Hausbankprinzip”), which is often referred to as “relationship banking”.162 In addition, German corporations increasingly access public and private debt markets beyond the plain vanilla bank credit.163 The rising importance of private retirement provisions will presumably promote indirect shareholdings through investment funds and increase their relative weight compared to banks that nowadays often exercise depositary voting rights for their clients’ direct shareholdings. The gradually diminishing power of banks can also be observed in the declining number of banks’ representatives in German supervisory boards since the 1970’s.164 Given German banks’ once disproportionate control, this development can be seen as a reversion process towards their pro rata position of power within the companies’ control system.165 Finally, the position of outside small shareholders, whose interests often are opposed to both managements’ and large shareholders’ ones, will be strengthened.166 Besides an expansion of civil liability of management and supervisory board members of public stock corporations, corporate transparency with regard to both information quantity and quality will be enhanced. While blockholders and banks used to have privileged access to information, small outside shareholders had to rely on spare publicly disclosed information. Therefore, the German Federal Government plans to obligate companies to publish quarterly reports reviewed by the auditors and to further promote the auditors’ independence. Furthermore, the competencies of the shareholders’ meeting probably will be enlarged and the participation of individual shareholders, e.g. via internet voting, will be facilitated.167

162 163 164 165 166

167

See Ordemann, M¨uller and Brackschulze (2005), pp. 19-24. See Deutsche Bundesbank (2004a), pp. 15-26. Cf. Wahl (2004). See Matthes (2000), pp. 50-53; Sherman and Kaen (1997). Cf. Schr¨oder and Schrader (1998). The percentage of the German population owning directly or indirectly shares (through funds) in listed companies already increased from 8.9% in 1997 to 16.7% in 2005. See Deutsches Aktieninstitut e.V. (DAI) (2005c), p. 1. Cf. Helmis (2002), p. 39. See Bundesregierung (2004); Matthes (2000), p. 43; Deutsche Bank (2002), p. 4; Baums (2002), p. 134.

42

Fundamentals of Corporate Governance

To conclude, the traditional system of corporate governance has not been drastically unhinged by recent changes.168 However, it is safe to say that the corporate governance discussion has just started.169 On the one hand, the recent numerous legal initiatives adopted and proposed by government are aimed at providing higher corporate governance standards but also bear the threat of overregulation.170 On the other hand, the implementation of the German Corporate Governance Code is a clear sign of a rising awareness of further reform needs concerning the German corporate governance system.171

168 169

170 171

See Matthes (2000), p. 54. For example, the role and duties of the supervisory board are widely discussed in Germany. See amongst others Cromme (2003), pp. 144-146. See von Rosen (2003), pp. 1-2. However, even some of the largest German companies still publicly question some of the provisions stated in the DCGK. See Fockenbrock (2006), p. 1.

Shareholder Structures and Corporate Governance

3

43

Causes and Consequences of Shareholder Structures in Theory and Practice of Corporate Governance

This chapter will take a bottom-up perspective on the causes and consequences of shareholder structures for the corporate governance setting in publicly traded German corporations. After elaborating on the primary source of corporate governance problems — the separation of ownership and control (section 3.1) — the legal framework for the relation between shareholders and management, i.e. the German Stock Corporation Act (“Aktiengesetz” (AktG)), will be outlined (section 3.2). A detailed analysis of management’s and various shareholder groups’ objectives will be performed (section 3.3) and resulting conflicts of interests will be summarized (section 3.4). 3.1

The Separation of Ownership and Control

The bottom line of the pathbreaking work of B ERLE AND M EANS (1932) titled The Modern Corporation and Private Property is that ownership and control have been largely separated in large U.S. corporations at the beginning of the 20th century.172 The results of their analysis of the ownership structures of the 200 largest industrial corporations in the U.S. around 1929 can be summarized in the following three most crucial propositions: First, they observe that the majority of these firms is owned by a very large number of stockholders and typically not one or even a certain single group of them owns a significant portion of the outstanding shares.173 Second, also the corporate management owns at most a very small portion of its company’s shares as a consequence of the first observation.174 Third, they argue that shareholders’ and management’s interests 172

173

174

Shleifer and Vishny (1997), p. 740, argue that the term ‘separation of management and finance’ would be more appropriate than the standard terminology ‘separation of ownership and control’. Demsetz and Lehn (1985), p. 1173, report that the work of Berle and Means (1932) was already largely anticipated eight years earlier by Veblen (1924). They find that in at least 44% of the 200 largest non-financial corporations there is no effective control by shareholders. Furthermore, in the largest railroad, public utility and industrial company (namely, Pennsylvania Railroad, U.S. Steel, American Telephone and Telegraph) the largest reported holdings are only 0.34%, 0.70% and 0.90% respectively and even the twenty largest shareholders own at most 5.1%. See Berle and Means (1932), pp. 47-48, 98-100. Similar findings have already been reported earlier by Means (1931), pp. 68-100. Cf. Taussig and Barker (1925), pp. 11-18. Even though this fact was widely accepted at that time data limitations prohibited a separate documentation. See Stigler and Friedland (1983), p. 238. In contrast, an earlier survey by Taussig and Barker (1925), pp. 1130, documents the extent of executive ownership and reports that it is about 18% in larger and about 45-50% in smaller U.S. companies.

44

Shareholder Structures and Corporate Governance

most probably are not aligned or may even diverge largely. However, they neither elaborate on the particular interests of both groups (see section 3.4.2.1) nor empirically test their propositions.175 Instead, they refer to the changing statute and case law which tended to constrict shareholders’ voting and participation rights and hence “[. . . ] eliminated most legal restraints on managerial prerogatives.”176 The suggestion that management’s rising power might be detrimental to shareholders’ wealth is put into perspective by the following two statements later in the book: First, it is claimed that no provision in the statutes can deny or defeat the equitable treatment of shareholders. However, the main problem is that dispersed shareholders do not manage to file and finance suits against management. Consequently, the courts — though actually being able to deal with the problems evolving from the separation of ownership and control in an effective manner — have no opportunity to employ their powers. Second, future financing needs of corporations will deter management from abusing its power on a considerable scale.177 These slightly conflicting arguments lead S TIGLER AND F RIEDLAND (1983) to the following conclusion: “Thus the actual effects of the separation of ownership and control are left undetermined, and even unstudied. Yet the book closes on the fear of ‘corporate blundering’ by the non-owning controllers.”178 Departing from the analysis of team work, A LCHIAN AND D EMSETZ (1972) see shirking179 as a common characteristic of human behavior. They transfer their rather universal argument that “[t]he specialist who receives the residual rewards will be the monitor of the members of the team [. . . ] and earns his residual through the reduction in shirking that he brings about [. . . ]”180 to the case of corporations where the pooling of resources of many small investors makes it possible to carry out larger investments. Based on their comparatively well structured analysis, they conclude that some adaptations to the organizational form of an owner-managed firm are necessary since the residual claimants 175

176 177 178 179 180

Testing hypotheses was not very common in scientific economic studies of those days which makes criticism on that ‘a bit unfair’ as North (1983), p. 270, states. Stigler and Friedland (1983), p. 239. See Berle and Means (1932), pp. 219-243. Stigler and Friedland (1983), p. 239. For a discussion of the concept of shirking see Alchian and Demsetz (1972), pp. 779-781. Alchian and Demsetz (1972), p. 782. Accentuation taken over from the original.

Shareholder Structures and Corporate Governance

45

in the case of a public stock corporation are common shareholders. First, limited liability is necessary to prevent shareholders from large losses no matter how they are caused. Second, the usually large number of stock owners makes it necessary to delegate decision authority to a smaller group, i.e. the management, while corporate stockholders only retain the authority to supervise the management.181 Acknowledging the problems involved in the separation of ownership and control — referring solely to the phenomenon of shirking — they come to the following conclusion: “As a result a new modification of partnerships is induced — the right to sale of corporate shares without approval of any other stockholder. Any shareholder can remove his wealth from control by those with whom he has differences in opinion. Rather than try to control the decisions of the management, which is harder to do with many stockholders than with only a few, unrestricted salability provides a more acceptable escape to each stockholder from continued policies with which he disagrees.”182 With the recognition of this exit option and its effects on stock prices as an alternative action for disappointed shareholders, A LCHIAN AND D EMSETZ (1972) point to the pressure of the financial market and the possibility of a temporarily concentration of ownership (e.g. during a proxy battle) to discipline management. Hence, they implicitly confirm the observations and fears of B ERLE AND M EANS (1932) but contradict in their conclusions. While the latter two authors see a main problem in the insufficient enforcement of shareholders’ claims at the courts, the former authors believed that the pressure from financial markets is sufficient to resolve disputes between shareholders and management in the long run .183 The next seminal analysis tackling the problems arising from the separation of ownership and control was performed by J ENSEN AND M ECKLING (1976), who develop a theory of the ownership structure of the firm. Being the most formal analysis in this area up to that date, they depart from the case of an entirely owner-managed firm (i.e. a firm where the manager owns 100% of the residual 181 182 183

Alchian and Demsetz (1972), pp. 787-788. Cf. Easterbrook and Fischel (1991), pp. 8-11. Alchian and Demsetz (1972), p. 788. See Alchian and Demsetz (1972). Cf. Ulen (1993), pp. 310-312.

46

Shareholder Structures and Corporate Governance

claims) and show how so-called agency costs of outside equity arise when a portion of the residual claims is sold off to outsiders. The owner-manager is guided in his decision making by the goal to maximize his own utility. Thereby, he will not only consider pecuniary returns but also non-pecuniary benefits184 arising from his entrepreneurial activity. But since he has to bear all costs associated with non-pecuniary returns as they lower the residual claims he is entitled to, he will consume the optimal mix of pecuniary and non-pecuniary benefits. This optimal mix is achieved when the marginal utility from a dollar of additional expenditures is equal for all non-pecuniary items and matches the marginal utility associated with an additional dollar of (after tax) purchasing power.185 After selling off a portion of identically equipped equity claims, agency costs evolve because the manager “[. . . ] will then bear only a fraction of the costs of any non-pecuniary benefits he takes out in maximizing his own utility”186 and hence increase such activities beyond the level he would have chosen as a sole owner-manager. Therefore, outside stockholders have incentives to limit such value destroying activities and are motivated to spend resources on monitoring the management. This monitoring becomes even more desirable for outside shareholders as the owner-manager’s portion of the residual claims decreases because this seduces him to consume even larger amounts of corporate resources. Thus, J ENSEN AND M ECKLING (1976) explicitly postulate a positive, monotonic relationship between the management’s equity fraction (i.e. the level of insider ownership) and corporate value for the first time.187 Their analysis attracted broad attention among researchers but their hypothesis did not remain unquestioned for a long time: For example, FAMA (1980) claims that market competition will be sufficient to discipline management and D EMSETZ (1983) views the ownership structure as an endogenous outcome of a value maximizing process.188 184

185 186 187 188

Jensen and Meckling (1976), p. 312, list “[. . . ] the physical appointments of the office, the attractiveness of secretarial staff , the level of employee discipline, the kind and amount of charitable contributions, personal relations (‘love’, ‘respect’, etc.) with employees, a larger computer to play with, purchase production from friends, etc.” as examples for non-pecuniary benefits. See Jensen and Meckling (1976), p. 312. Jensen and Meckling (1976), p. 312. See Jensen and Meckling (1976), pp. 312-220. Empirical studies investigating the value impact of insider ownership will be presented later on in section 4.2. For different hypotheses on the effects of ownership on performance see section 5.3. Besides the presented so-called contractarian approaches to corporate governance, other approaches — based on the

Shareholder Structures and Corporate Governance

47

The separation of ownership and control, first observed by B ERLE AND M EANS (1932) as a typical characteristic of the modern corporation at the beginning of the 20th century in the U.S., remains highly relevant, even though shareholder structures have changed and thereby typically became less dispersed.189 Their observations partially refer to the case of German public companies as well but two distinctive restraints must be noted: First, large, non-quoted corporations are far more common in Germany than in the U.S.190 Second, shareholder structures of German listed companies have always been less dispersed as will be shown in the empirical study of the German market in chapter 6.191 However, since the basic problem resulting from the separation of ownership and control persists, the next section will show how the rights and duties of the owners (i.e. shareholders) and those who ‘control’ the company (i.e. management) are formally distributed.192 3.2

Constitution of the German Stock Corporation

This section provides a short overview about those articles of the German Stock Corporation Act (“Aktiengesetz” (AktG)) which regulate the interaction among the three corporate bodies (i.e. management board (“Vorstand”), supervisory board (“Aufsichtsrat”) and shareholders’ meeting (“Hauptversammlung”). The Stock Corporation Act can be seen as the constitution of German stock corporations.193 While a comprehensive presentation of the legal regulations is far beyond the scope of this section, its purpose is to illustrate the distribution of

189

190

191

192

193

human-capital investment, path-dependency, or competence based theory of the firm — evolved recently. For a synopsis on the main arguments see Lenoble (2003), pp. 23-52. Nevertheless, La Porta, Lopez-de Silanes and Shleifer (1999), pp. 492-495, find for the case of large and medium-sized U.S. public companies that at the end of 1995 still an average of 80-90% of the shares are widely held. In 2004, out of the 50 (10) largest European companies, 26 (5) are German. See N.N. (2005b), p. 11. In 2003, the 7,165 stock corporations represent only 0.2% of all German companies. Nevertheless, they account for 20.2% of all taxable deliveries and other performances which underlines that especially the largest companies are organized as stock corporations. This tendency becomes even more apparent for publicly quoted stock corporations — the subject of this study — which unfortunately are not registered as a separate category in the statistic. 736 companies have taxable deliveries and other performances over EUR 50m p.a., which approximately equals the number of companies listed in the official and regulated market at the Frankfurt Stock Exchange. Hence, even though only representing 0.03% of all companies they account for 19.6% of all taxable deliveries and other performances. See Statistisches Bundesamt (2005). La Porta, Lopez-de Silanes, Shleifer and Vishny (1998), pp. 1146-1148, find that the three largest shareholders of the ten largest non-financial companies own 48% in Germany vs. only 20% in the U.S. For a theoretical analysis of how residual claims should be distributed between risk takers (i.e. management) and risk bearers (i.e. owners) see Fama and Jensen (1983b). Nearly all publicly traded companies in Germany are stock corporations. The sporadic KGaAs are subject to the same provisions as the AGs to a large extent. Some supplementary and substitute provisions are

48

Shareholder Structures and Corporate Governance

rights, duties and responsibilities to provide a better understanding of the succeeding analysis of shareholders’ and management’s — partially diverging — interests.194 3.2.1

Management Board

The management board (“Vorstand”) bears the sole responsibility for managing the corporation. The members of the management board — at least two persons if the share capital exceeds EUR 3m (§ 76 (2) AktG) — are appointed by the supervisory board (see section 3.2.2) for terms of up to five years (§ 84 (1) AktG). Members of the management board may not serve as supervisory board members and vice versa.195 All members of the management board must act jointly in managing the corporation or representing it vis-´a-vis third parties (§ 77 (1) AktG) but usually the articles entitle any two members of the management board to represent the corporation.196 Management board members may have the right to participate in the corporations’ profits, but the supervisory board must assure that total compensation (i.e. salary, stock option plans, benefits and deferred compensation) bears a reasonable relationship to management board members’ duties and the condition of the company (§ 87 AktG). Management board members must not compete with their company in its line of business (§ 88 AktG). Unless the corporation is integrated with its parent company or subject to a control agreement, the management board has the non-delegable sole duty to manage the corporation (§ 76 (1) AktG).197 The management board always has to act independently for the mere well-being of the corporation. If no control agreement is signed, controlling parent companies are liable to the corporation and its (minority) share-

194

195

196 197

contained in book II of the Stock Corporation Act. Due to its rarity, the KGaA will not be dealt with in this study. Cf. footnote 33. Note that all non obligatory provisions of the Stock Corporation Act might be altered in the corporate articles (“Satzung”) if the Act explicitly permits (§ 23 (5) AktG). As an exception, in the case of vacancies members of the supervisory board may serve as management board members for a period of up to one year. However, they are not allowed to serve as supervisory board members during that period (§ 105 AktG). See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 75. Even though this formal right remains untouched so far, one large German Bank (Deutsche Bank AG) managed to assign extensive decision rights to a newly created institution, the so-called Executive Committee, which consists of management board members and divisional heads. The management board formally retains the right to appoint executive committee members and, consequently, bears the responsibility for decisions which effectively have been reached in the Executive Committee. See von Heusinger (2002), p. 26.

Shareholder Structures and Corporate Governance

49

holders if they induce the management (and consequently the AG) to carry out detrimental transactions (§ 311 AktG).198 The management board has to report to the supervisory board on the intended business policy and other fundamental matters regarding the future strategy of the company (§ 90 AktG). The management board has responsibility for calling the shareholders’ meeting on a regular basis (§ 121 AktG) and promptly if losses are expected to reduce the company’s share capital to less than a half (§ 92 (1) AktG). In case of insolvency, i.e. in cases of illiquidity as well as actual and impending over-indebtedness, the management board must file for insolvency at the courts within at most three weeks (§ 92 (2, 3) AktG).199 During the term of their appointment, management board members can only be dismissed for fundamental reasons, e.g. gross violations of duty, inability to manage the corporation or a vote of no confidence by the shareholders’ meeting (see section 3.2.3). Each member of the management board is obliged to use the care of a diligent and conscientious manager (§ 93 (1) AktG). In the case of violated duties, management board members shall be jointly and severally liable to the company for any resulting damage (§ 93 (2) AktG).200 Members of the management board have a statutory duty of confidentiality (§ 404 AktG) and certain breaches might even be seen as criminal offences (§§ 399-404 AktG).201 3.2.2

Supervisory Board

The supervisory board (“Aufsichtsrat”) appoints, supervises, controls and if necessary dismisses the management board but is not involved in the company’s operative day-to-day management (§§ 84 (1), 111 AktG). Its size varies between three and 21 members and depends on the stated share capital, the articles and — if applicable — codetermination statutes (§ 95 AktG). The supervisory board members are appointed by the shareholders’ meeting and — if any of the code198

199

200

201

For a more comprehensive review of the regulations concerning groups of companies see Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 77-93. See Schneider and Heidenhain (2000), pp. 8-9. Cf. §§ 17-19 Insolvency Statute (“Insolvenzordnung” (“InsO”)). Liability cannot be precluded by supervisory board approval but is precluded if the management board acts pursuant to a lawful resolution of the shareholders’ meeting (§ 93 (4) AktG). In 2004, attempts were made to enlarge the liability of management board members by proposing the Law Governing Liability for Capital Market Information (“Kapitalmarktinformationshaftungsgesetz” (“KapInHaG”)). However, it was finally withdrawn. See Drost (2005), p. 38. Cf. Kr¨amer (2002), pp. 7-26. See §§ 76-94 AktG; Schneider and Heidenhain (2000), pp. 8-9.

50

Shareholder Structures and Corporate Governance

termination statutes apply (§ 101 (1) AktG)202 — by the company’s employees. The terms of service are limited to a maximum of five years (§ 102 (1) AktG). Eligible for the supervisory board are natural persons (§ 100 (1) AktG) who are not already member of the supervisory board in ten corporations (chairmanship counts double)203 , which are required by law to form a supervisory board, or are legal representatives of an enterprise controlled by the company (§ 100 (2) AktG).204 Members elected by the shareholders can be removed prior to the expiration of their term of office pursuant to a resolution of the shareholders’ meeting with a majority of not less than three fourths of the votes cast (§ 103 (1) AktG).205 The supervisory board elects a chairman and one or more deputy chairmen, who may act in place of the chairman only if the latter is incapacitated (§ 107 (1) AktG). Members are not allowed to delegate a deputy to act in their absence (§ 111 (5) AktG) but may submit their votes in writing (§ 108 (3) AktG). The board must convene at least once every calendar half-year (at least twice for publicly traded corporations) but is expected to meet on a quarterly basis (§ 110 (3) AktG). Each member of the supervisory and management board may, upon stating the grounds therefore, request the supervisory board chairman to promptly call a supervisory board meeting. This meeting shall be held within a two week period (§ 110 (1) AktG). The board may form committees for certain purposes excluding some explicitly specified tasks (§ 107 (3) AktG).206 It is the supervisory board’s task to appoint and dismiss members of the management board (§ 84 (1) AktG) and it may appoint one as chairman of the management board (i.e. CEO) if the management board consists of more than one 202

203

204

205 206

For an overview of the codetermination statutes concerning the composition of the supervisory board (Works Constitution Act, Codetermination Act, Montan-Codetermination Act) see section 2.4.1. Cf. Dietl (1998), pp. 113-114. The concentration of supervisory board mandates in Germany is high: Only 18 managers hold 158 (9.1%) of all supervisory board mandates in the 80 largest German companies (DAX, MDAX) in 2005. See Fockenbrock (2005c), p. 16. With regard to age, supervisory board members of the DAX-companies are on average 57 years old, while the chairman is on average 66 years old. See Sommer (2006), p. 16. Former management board members are eligible to serve the supervisory board. As a German particularity, 35.9% of all supervisory board members were former management board members, and in 19.7% the former CEO serves as chairman to the supervisory board in 2004. See Fockenbrock (2005a), p. 14. Some critics argue that former management board members should not be eligible for the (chairmanship of the) supervisory board, at least within a certain standby period. See Steinbeis (2004), p. 6. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 75. See Schneider and Heidenhain (2000), pp. 10-11.

Shareholder Structures and Corporate Governance

51

person (§84(2) AktG).207 It represents the corporation — and therefore finally the shareholders and the employees if the company is subject to codetermination regulation — in all judicial and extrajudicial dealings with the management board (§ 112 AktG).208 The management board reports to the supervisory board (§ 90 AktG) whose members may inspect and examine the books and records of the company independently and which appoints the auditor of annual accounts (§ 111 (2) AktG). The articles of the corporation may demand that certain transactions require supervisory board’s approval but a refused approval can be annulled by a resolution of the shareholders’ meeting which requires at least three fourths of the votes cast (§ 111 (4) AktG).209 The due diligence requirements for management board members (§ 93 AktG) also apply to supervisory board members analogously (§ 116 AktG). Supervisory board members are jointly and severally liable unless they can prove that they have acted with the care of a diligent and conscientious manager (§ 117 (2) AktG). Persons exerting undue influence on the company and wilfully inducing members of the management or supervisory board to act to the disadvantage of the company or its shareholders are liable to the company for any resulting damages (§ 117 (1) AktG).210 3.2.3

Shareholders’ Meeting

The shareholders are supposed to exercise their rights with respect to the company at the shareholders’ meeting (“Hauptversammlung”), which management and supervisory board members are expected to attend (§ 118 AktG).211 According to the concept of separation of ownership and control, the shareholders’ meeting is only involved in decisions if required by law, the company’s articles or demanded by the management board (§ 119 AktG). Ordinary shareholders’ 207

208

209 210

211

As a consequence, the supervisory board is also in charge of laying down management board members’ remuneration, which ideally should be geared to corporate profits. See N.N. (2005a). Shareholders representing at least one percent of the company’s share capital or a pro-rata value of EUR 100,000 may sue the corporation for damages on their own behalf (but for payment to the corporation) only if the corporation (i.e. the supervisory board) did not do so within an appropriate period (§ 148 AktG). See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 75. See §§ 95-117 AktG. A more detailed description of the liability of supervisory board members is provided by Kr¨amer (2002), pp. 22-26. For an extensive analysis of the supervisory board’s role in the German corporate governance system see Albers (2002), pp. 29-38; Hopt and Leyens (2004), pp. 4-10. Cf. Mann (2003), pp. 170-180. Iber (1987), p. 40, lists seven material and formal shareholders’ rights. Of the latter, e.g. the right to call a shareholders’ meeting if the demanding group cumulatively holds at least 5% (§ 122 AktG) must be exercised before the shareholders’ meeting.

52

Shareholder Structures and Corporate Governance

meetings must take place once every year during the first eight months of the fiscal year to pass resolutions on the appropriation of distributable profits and to ascertain the approved annual financial statements and the annual report (§ 175 AktG). Shareholders whose aggregated holdings equal or exceed one-twentieth of the share capital may demand a shareholders’ meeting, stating the purpose and the reasons of such a meeting (§ 122 (1) AktG). Notice of the shareholders’ meeting shall be given no later than one month prior to the date of the meeting (§ 123 (1) AktG), and articles may provide that shareholders give notice of their intention to attend no later than the third day prior to the meeting (§ 123 (4) AktG).212 The decision rights of the shareholders’ meeting are limited to the election of members of the supervisory board (only to the extent that they are not appointed or elected as employee representatives according to codetermination regulations), the appropriation of distributable profits, the ratification of the actions of the management and supervisory board (“Entlastung”), amendments of the articles, measures to increase or reduce the share capital, the appointment of external auditors for the annual accounts and for the examination of matters in connection with the formation or the management of the company, the dissolution of the company (§ 119 (1) AktG) and to matters concerning the management of the company if required by the management board (§ 119 (2) AktG).213 Decisions at the shareholders’ meeting are determined by a (simple) majority of the votes cast unless law or the articles demand a larger majority or additional requirements (§ 133 AktG). Voting rights are determined by the nominal capital of the present shares or their number in case of no par value shares (§ 134 AktG). Proxy voting is permitted (§ 134 (3) AktG) but banks and shareholder associations exercising their customers’ voting rights within their discretion must assure that shareholders’ interests are preserved (§ 128 (2) AktG).214 The agenda of the shareholders’ meeting should be published together with notice of the meeting (§ 124 AktG), and the management board is supposed to 212

213

214

See Schneider and Heidenhain (2000), p. 12; Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 75. Because of the limited formal scope for decision making and the usually limited power in practice critics refer to the shareholders’ meeting also as a ‘staging of shareholders’ democracy’. See Maisch and Teschler (2005), p. 12. See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 75; Schneider and Heidenhain (2000), p. 13.

Shareholder Structures and Corporate Governance

53

publish countermotions and nominations that are consistent with statutory requirements and timely received (§§ 125, 126 AktG). Shareholders whose aggregated shares amount to at least one-twentieth of the share capital or the pro rata value of EUR 500,000 may demand that items for decision by the shareholders’ meeting are published on the agenda (§ 122 (2) AktG). At the shareholders’ meeting, each shareholder has the right to request information from the management board regarding the company’s affairs. However, this right is limited to the extent that such information is necessary to permit a proper evaluation of an agenda item. Such requests may only be refused if providing such information is likely to cause material damage to the corporation or in certain specified matters (§ 131 AktG). Detailed rules regulate the recording of minutes (§ 130 AktG) and the list of participants (§ 129 AktG).215 3.3

Shareholders’ Objectives and Control Incentives

Having presented shareholders’ formal rights according to the corporate constitution, i.e. the Stock Corporation Act, this section aims at analyzing shareholders’ objectives and interests from an economic perspective.216 After an introductory systematization of shareholders’ interests (section 3.3.1), two different dimensions of shareholder structures will be distinguished: shareholders’ identity (section 3.3.2) and shareholders’ concentration (section 3.3.3).217 While the focus will be on shareholders’ identity — especially with regard to insider ownership — the concentration aspect cannot be fully omitted in the analysis of shareholder structures as will be shown later on in section 6.218 3.3.1

Systematization of Shareholders’ Interests

In section 2.2.1 it was argued that shareholders hold residual claims on the assets and cash flows of the firm. They are willing to accept the risk involved (e.g. bankruptcy of the company) in exchange for unlimited participation on 215 216

217

218

See §§ 118-147 AktG; Schneider and Heidenhain (2000), pp. 12-14. Cf. Mann (2003), pp. 166-170. Due to the varying interests of different groups or types of shareholders, these groups may have different incentives to actively get involved into corporate governance, e.g. by monitoring management. Consequently, the different incentives to exercise corporate control determine their role within the corporate governance setting of the individual firm. Rennneboog (2000), p. 1965, amongst others claims that control incentives differ not only along with ownership concentration but also with the identity (or nature) of the shareholder. Similarly, Kretschmann (1976), pp. 79-84, differentiates between quantitative and qualitative elements of shareholder structures. Cf. Gleisberg (2003), pp. 33-35; Bott (2002), pp. 19-70.

54

Shareholder Structures and Corporate Governance

Basic interest • Attractive risk/return ratio • Return maximization • Distribution of profits (dividends) • Capital gains through stock price accretion • Risk minimization • Systematic (market) risk • Unsystematic (company) risk Auxiliary interests depending on shareholders’ concentration • Smallholders (free float) • Rational apathy (free-riding) • Simple blockholders • Inside control (supervisory board) • Dominating blockholders • Strategic influence (control stock) • Simple majority blockholders • Risk aversion (lack of diversification) • Super majority blockholders • Expropriation of minority shareholders

Auxiliary interests depending on shareholders’ identity • Insider individuals • Approval of own actions (discharge of the board) • Non-financial companies • Strategic motives (window-ontechnology, acquisitions) • Financial institutions • Operating motives (business relations, credit agreements) • Government • Political motives (job protection) • Miscellaneous

Source: Author’s illustration; cf. Bott (2002), p. 19.

Figure 3-1: Shareholders’ Set of Objectives the remaining profits once all fix claims have been served. Following classical finance theory and literature, investors assess the attractiveness of an investment (or in the case of investments in the public equity market of a stock) by its expected risk/return ratio.219 Risk averse investors, as assumed by classical finance theory, are only willing to accept higher risks in exchange for higher expected returns.220 However, as reality is by far more complex “[o]ne should be careful in specifying what is meant by shareholders and their interests.”221 In figure 3-1, this objective is labeled as basic interest meaning that all market participants are at least partially governed by this principle in their investment decisions. But, the basic interest must not necessarily represent the dominating interest: its weight relative to other auxiliary interests may vary for each individual shareholder. The prefix basic is just used to clarify that it is the one interest which all different types of investors have in common. 219

220

221

More precisely, the diversification potential could be seen as investors’ third decision criterion. For the purpose of simplification, this analysis will be limited to the risk and return aspects and neglect the diversification aspect, which becomes important for investors from a portfolio perspective. For a review of the portfolio theory by Markowitz and the capital asset pricing model (CAPM) see Brealey and Myers (2000), pp. 187–220; Bodie, Kane and Marcus (1999), pp. 147-397. OECD (1998a), p. 21.

Shareholder Structures and Corporate Governance

55

Exemplary auxiliary interests of different shareholder types — classified along the two dimensions, i.e. shareholders’ concentration and shareholders’ identity — are listed in the boxes on the left hand and right hand side of figure 3-1. The two upper arrows illustrate that the auxiliary interests might influence the relative weight of the basic interest.222 As the nature of each single shareholding is influenced by both dimensions, i.e. its size (shareholders’ concentration) and the owner’s identity (shareholders’ identity), the lower arrow shows that both dimensions may constitute an interdependent relationship.223 Following this framework, the next two sections will examine in detail the auxiliary interests of various shareholder types, which are classified along the two mentioned dimensions. 3.3.2

Shareholder Identity as Differentiating Factor

The differentiation of shareholders’ identities will be carried out on two levels. First, the shareholder universe is split into five main categories of shareholder types (inside individuals, non-financial companies, financial institutions, government and miscellaneous). In a second step, the breakdown is refined with a varying number of sub-categories. While the finer differentiation increases the homogeneity within the shareholder groups, the aggregation on the main category level will be necessary for the empirical analysis to assure an adequate number of occurrences for each group.224 As an extensive presentation of all shareholder types’ interests is beyond the scope of this study, the focus will be on insiders’ interests while those of other owner types will only be outlined. The analysis is largely based on a proprietary control incentive scheme shown in table 3-1, which evaluates the extent of: i) contractual relations, ii) internal agency conflicts and iii) auxiliary interest for the various shareholder groups.

222

223

224

For example, a non-financial company might acquire a rising competitor predominantly to eliminate an unpopular rival rather than because of a promising expected risk/return ratio. Similarly, government is likely to pass up privatization opportunities, e.g. to protect jobs which might be at risk if the government’s shareholding is reduced or completely sold. On the one hand, the degree of shareholders’ concentration determines the enforceability of interests depending on the identity of the shareholder. On the other hand, the identity will influence e.g. the way a powerful blockholder might be willing to expropriate minority shareholders. Therefore, Bott (2002), p. 41, presumes that some authors differentiate finer by the description of their samples than in the following analysis of a shareholder structure/ performance relationship. For an illustration of the shareholder identity classification scheme see table 6-5 and section 6.3.1.

56

Shareholder Structures and Corporate Governance

3.3.2.1 Insider Individuals (MGMT)225 3.3.2.1.1

Members of the Management Board (MB)

Members of the management board who own shares in the company they are contractually employed with may obviously have extensive auxiliary interests. As owners of the company, they attend the shareholders’ meeting and are entitled to elect the supervisory board, which in turn elects the management board members and negotiates their employment contracts. Provided the management board members accumulate sufficient control rights, they will (s)elect those supervisory board members who are most likely to support the management in place instead of those who might try to force management to adopt a value maximizing strategy and in the worst case may replace the incumbent management.226 However, as owners of the company they also have to (partially) bear the costs of any value decreasing action in contrast to managers without equity participation. With regard to a possible IPO, they probably choose the offering time by looking at a high company valuation, which will yield the highest return on their investment, rather than by evaluating the actual capital needs of the corporation.227 In their function as members of the management board, they have extensive auxiliary interests and are likely to be subject to certain selfish behavior which usually is not necessarily compliant with (outside) shareholders’ interests. Examples for such behavior will be given later on in the succeeding analysis of conflicts of interests in section 3.4. The numerous advantages management board members can draw from their current employment (income, prestige, power etc.) suggest that the continuity of their employment itself might represent an important objective.228 Since the mentioned observations let appear 225

226

227

228

The fact that shareholders are at the same time actively involved in the top management of the company is often referred to as insider ownership or managerial ownership. Both terms will be used synonymously in the following. As codes will be used for the ownership variables in the empirical study (see section 6.3.1), these codes are stated in capital letters in parentheses. MGMT stands for management or alternatively insider individuals. Holderness, Kroszner and Sheehan (1999), p. 103, describe this situation as follows: “[. . . ] allowing management to choose their own overseers might seem on a par with letting the fox guard the chicken coop [. . . ]”. Bhagat and Jefferis (2002), p. 74, report that insiders with sufficient strong ownership positions are unlikely to experience turnover. See Maug (2001), p. 27. However, this should not be necessarily seen as selfish action on the part of the management since other (former) shareholders also profit from a high issuing price. Shleifer and Vishny (1989), p. 126, put it slightly different: “A manager collecting rents at his current job will do what he can do to keep it. That usually involves making corporate control mechanisms more expensive to use, but also making the manager more valuable to shareholders than any alternative.”

57

Shareholder Structures and Corporate Governance Aspect/a Identity

Extent of Contractual Relations

Extent of Internal Agency Conflicts

Extent of Auxiliary Interests

MGMT MB SB FBM

++ ++ +

−− −−/(++)b −−

++ + +

CORP BGR COR

+ 

++ ++

++ +

FINC BANK INSR INVC INST INST-F

+  −− −− −−

++ ++ −  −

++ + −− − −−

−

++

++

−− ++ n/a

−− −− n/a

 ++ n/a

GOV MISC INVD EMP OTH/TRE a b

Expected characteristic are based on theoretical considerations and empirical observations. Internal agency conflicts are to be expected for representatives of corporations or banks serving the supervisory board. However, such shareholdings would be attributed to the owner types of corporations or banks (and not to supervisory board members) under the concept of ultimate ownership. See section 6.3.1.1. Legend: ++ = very high, + = rather high,  = moderate, − = rather low, −− = very low. MGMT = Management (with the sub-categories MB = Management board member, SB = Supervisory board member and FBM = Former board member), CORP = Non-financial corporation (with the sub-categories BGR = Business group (stakes above 50%) and COR = corporation (stakes below 50%)), FINC = Financial institutions (with the sub-categories BANK = Banks, INSR = Insurance companies, INVC = Investment companies, INST = Domestic institutional investors and INST-F = Foreign institutional investors), GOV = Government and MISC = Miscellaneous (with the sub-categories INVD = Outside individuals, EMP = employees, OTH = Others and TRE = Treasury shares). For a detailed definition of the shareholder identity codes see table 6-5. Source: Author’s analysis.

Table 3-1: Determinants of Control Incentives a positive or negative effect of management board ownership conclusive, the corresponding theories will be presented in chapter 4. The impact of management board members on the company’s affairs is extensive and its scope is formally determined by the Stock Corporation Act and the company’s statutes.229 In their function as management board members, they bear the sole responsibility for managing the corporation. If they simultane229

See section 3.2.1.

58

Shareholder Structures and Corporate Governance

ously represent an important shareholder group, which is not a rare case as the descriptive analysis of shareholder structures in section 6.4.1 will show, they are able to suspend control mechanisms to a certain degree and thus are able to gain virtually unlimited power over their company. Finally, it should be recognized that the interests among members of the management board and thus the interests of individual management board members and the board as a whole may diverge (personal vs. institutional view). This might be especially true if ownership is unequally distributed among management board members. As the empirical analysis in section 6 reveals, this is more the rule than the exception. With respect to tenure, a young manager will probably have more incentives to build up his reputation, e.g. to qualify for the management of a larger or more prestigious company, than a board member on the verge of retirement.230 However, if not mentioned otherwise the management board will be treated as a group with homogenous interests to not complicate matters even further. 3.3.2.1.2

Members of the Supervisory Board (SB)

Supervisory board members’ interests are even less homogenous than those of management board members because of the dual obligation they are confronted with: On the one hand, they have the contractual and legal duty to supervise the management board on behalf of all shareholders. On the other hand, they usually represent diverse lobbies, i.e. shareholder groups, banks or employees in the case of codetermination. Hence, the supervisory board unifies numerous, potentially conflicting interests in one single body. Obviously, this might limit supervisory board’s effectiveness in being the central institution of control.231 Thus, the board’s interests are likely to be a mixture which depends on the composition of the board. Besides the shareholder groups, which are described in the following sections, supervisory boards of companies subject to codetermination232 allocate employees’ interests a weight usually far above their importance as a shareholder group.233 230 231 232 233

Cf. Monsen and Downs (1965), p. 227. See Schmidt (2003), p. 11. See section 2.4.1. In the empirical analysis employees on average hold only between 0.1% (1993) and 0.5% (1998) of their company’s shares. See figure 6-7.

Shareholder Structures and Corporate Governance

59

If supervisory board members are representatives of shareholder groups which themselves face potential internal agency problems (see table 3-1), a somewhat paradoxical situation is created in which agents are supposed to monitor other agents.234 The resulting multi stage agency conflicts are assumed to further dilute the supervisory board’s control mandate. Furthermore, some critics argue that the supervisory board’s control mandate has not been defined properly.235 Supervisory board members representing the shareholdings of an individual or a family usually do not face internal agency problems but might be confronted with other conflicts of interests, e.g. with respect to the election of management board members: On the one hand, they might preferably elect family members into the management board and thereby preserve the company’s status of a family firm. On the other hand, and especially if the family resource pool is rather poor, the election of outside management board members might be the best solution for the well being of the company and consequently for the (family) shareholders’ wealth.236 As a particular point of the German corporate governance system, CEOs (i.e. the chairman of the management board) of large public corporations often become supervisory board chairmen after their period of service. This setting imposes two restrictions on the effectiveness of control through the supervisory board: First, it is unlikely that former CEOs will criticize their successors in questions of long-term strategies or acquisitions which have been initiated under their own leadership.237 Second, usually having served on the management board for years along with their fellow board members, the probability of close relationships among former and active management board members is high, which makes objective control even less likely.238 To sum up, the consequences of supervisory board ownership remain ambiguous. Being not involved in the corporation’s day-to-day management, they face 234 235 236

237

238

Cf. Woidtke (2002). See Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 116; Potthoff (1995), pp. 163-164. See Cf. Kaserer, Achleitner and Moldenhauer (2005), p. 8. For example, in the large non-listed German family firm Franz Haniel & Cie. GmbH as a general policy family members must not work within the whole group. See Kewes and Herz (2006), p. 15; Weishaupt (2006), p. 35. The differences between outsider and insider board members are also discussed by Demsetz and Villalonga (2001), p. 214. Mann (2003), p. 81, refers to the aspect that supervisory board members disciplining management board members implicitly attest themselves a bad choice in selecting the respective manager, who has been once appointed by the supervisory board. Cf. Sommer (2006), p. 16; Roth and W¨orle (2004), pp. 584-593.

60

Shareholder Structures and Corporate Governance

less direct opportunities to consume non-pecuniary private benefits compared to management board members. Consequently, the relative weight of their auxiliary interests might be rather low. Nevertheless, family ties between supervisory board members, management board members and family shareholders might dilute the control incentives associated with supervisory board ownership leaving the question of its impact on corporate performance an empirical issue. 3.3.2.1.3

Former Members of the Boards (FBM)

Former members of the boards are defined as those persons who have served on one (or both) of the company’s boards in the past but do not serve on any of the boards at the examined point in time. The inclusion of former board members into the group of insider is not self-explanatory: One might argue that these persons would have to be assigned to the group of outside individuals since they stopped working for the company in a managing or supervising position.239 However, at least four reasons suggest to include these persons into the group of insiders: First, ex-officers and ex-directors sometimes exercise significant influence on their former companies even without being formally in charge of it. This can be achieved through the selection of the successor (the so-called “crown prince”), who is likely to follow advices from former managers or through the placement of relatives or confidants in key positions during their time of service.240 Second, being an important shareholder group they can enforce their interests after their time of service in the shareholders’ meeting. Hence, their presence might deter incumbent management to deviate from the path pursued by its successors.241 Third, ownership of former board members builds the bridge from insider ownership to family ownership, which is examined more frequently for the case of Germany.242 Fourth, even after their terms 239

240

241

242

As figures 6-3 and 6-7 reveal, in the examined sample of the empirical study between 91% and 95% of all outside individuals — i.e. all individuals not being active members of the boards — are in fact former board members (including their families). Within the insider group they account for approximately 22-30% of the shareholdings. A prominent example is the Henkel KGaA, which was founded by Fritz Henkel in 1876. In 2003, the Henkel family still holds 52% of the ordinary shares but is not directly represented in the supervisory board. Instead, it continues to exert influence via a so-called shareholder committee. See Henkel (2006), pp. 3-4. See section 3.2.3 for the competencies of the shareholders’ meeting under the German Stock Corporation Act. The diversity of definitions of family firms is enormous. An overview of definitions of family businesses is provided by Chrisman, Chua and Sharma (1996), pp. 4-7. Cf. Astrachan, Klein and Smyrnios (2002), pp. 4558; Villalonga and Amit (2006), pp. 412-414; B¨uhler, Fueglistaller and Zellweger (2005), p. 10; Jaskiewicz, Gonzal´ez, Men´endez and Schiereck (2005), p. 11; Klein, Astrachan and Smyrnios (2005a), pp. 321-339.

Shareholder Structures and Corporate Governance

61

of service former board members can be expected to have inside knowledge of the company that others do not have. In contrast to active board members, who by definition stand in a contractual relationship with the company, former board members might negotiate advisory contracts after the end of their terms. But even if former executives may enter advisory contracts with their former company, it can reasonably be assumed that this income usually does not represent their main source of income as it is usually the case for active management board members. Therefore, one can conclude that the relative weight of auxiliary interests is somewhat lower for former board members compared to active members of the management board. Similar to management board members, former board members usually are not subject to internal agency conflicts (see table 3-1). 3.3.2.2 Non-financial Companies (CORP) Regarding non-financial companies (CORP), two sub-groups are differentiated: Ownership stakes up to 50% are labeled as corporations (COR) while ownership stakes above this threshold are called business groups (BGR) because the nature of such investments can be expected to be different.243 3.3.2.2.1

Business Groups (BGR)

There are basically two different ways in which a company can gain (or preserve) majority control over another listed corporation. First, the listed company can be the result of a split-off or equity carve-out where a fraction of the common shares of a solely owned subsidiary is sold to the public via an IPO.244 Second, the parent company might have acquired the stake in the listed company on the open-market, via a tender offer or a negotiated block trade.245 Such acquisitions (as well as split-offs) can be classified into vertical, horizontal and unrelated acquisitions. In the case of vertical acquisitions both companies operate in the same industry but in different stages of the value chain while in horizontal acquisitions both firms compete in the same kind of business activity. In 243 244 245

See section 6.3.1.2. See Weston, Siu and Johnson (2001), pp. 364-365. ¨ anybody who gains control, i.e. whose voting rights exceed 30%, Note that according to §§ 29, 35 WpUG must file and publish a mandatory takeover offer for the remaining shares.

62

Shareholder Structures and Corporate Governance

unrelated acquisitions both firms are engaged in unrelated markets.246 Considering these fundamentally different relationships between parent and subsidiary firm, the corresponding interests are likely to vary largely. If both companies operate in unrelated businesses, the extent of auxiliary interests is expected to be the smallest.247 Contrarily, if the companies are horizontally or vertically related, the parent company might be induced to use its dominant position to derive private benefits, e.g. by negotiating below- or above-market transfer prices.248 Besides contractual ties in the factor or product markets, the subsidiary might be subject to a control or profit transfer agreement (§ 291 AktG) giving the parent company extensive control.249 As a consequence, conflicts of interest between the majority shareholder and minority shareholders (see section 3.4.2.2) become more likely. 3.3.2.2.2

Corporations (COR)

While the use of the 50% threshold to distinguish between business groups (BGR) and corporations (COR) is motivated by formal shareholders’ rights, ownership stakes below 50% may allow substantial control as well.250 Horizontal relationships may be used to form strategic alliances aimed at gaining a competitive edge or expanding internationally. Vertical relationships can be used to strengthen existing customer or supplier relationships. While the optimal scope of the firm has been extensively discussed in literature251 , the question remains why firms invest into other firms if the shareholder could do it himself? Obviously, parent companies expect to capitalize on such investments — or the decision to retain significant stakes in a split-off company — beyond an attractive risk/return ratio. Whether these interdependencies profit or harm minority shareholders of one or both companies is an empirical question beyond the scope of this study.

246 247 248 249

250 251

See Weston, Siu and Johnson (2001), pp. 6-8. See Bott (2002), p. 56. See Barclay and Holderness (1989), p. 374. § 293 AktG requires a majority of three-fourth of the votes cast at the shareholders’ meeting for a control or profit transfer agreement to become effective. ¨ is gained if a party owns at least 30% of the voting rights. Note that control as defined by § 29 WpUG See amongst others Coase (1937); Williamson (1975); Grossman and Hart (1986).

Shareholder Structures and Corporate Governance

63

From an agency perspective, corporations appear to be less appropriate monitors as they are subject to internal agency problems themselves.252 While companies can manage their business portfolio by actively buying and selling corporate units, this process has been hindered for a long time by German tax legislation. As the share price of a subsidiary rises, so-called hidden reserves at the parent company increase according to HGB accounting rules since corporations generally have to value their long-term investments at the lowest of either the current price at the stock exchange or the price originally paid.253 Until capital gain taxes were abolished for corporations on January 1, 2002, corporations often would have faced significant liquidity straining tax-payments as a consequence of selling stakes above their book value.254 3.3.2.3 Financial Institutions (FINC) Financial institutions refer to the group of banks, insurance companies, investment companies and other institutional investors (domestic and foreign) whose primary purpose is to invest assets held in trust by them for others and their own assets. They are often analyzed as an important shareholder group.255 However, as theoretical considerations as well as empirical evidence suggest that the behavior of different types of institutional investors are likely to differ, they are analyzed separately in the following.256 3.3.2.3.1

Banks (BANK)

The power of banks in the German corporate governance system has been widely acknowledged257 even though their importance decreased during the last decade.258 Banks derive their power from at least four sources: direct equity 252

253

254

255 256

257 258

See Woidtke (2002) who discusses the problem of agents watching other agents considering pension funds as an example. These rules are also referred to as principle of the lower of cost and market. See Drukarczyk, Honold and Sch¨uler (1995), p. 119. Cf. §§ 253, 280 HGB. See W´ojcik (2003), p. 1455. Since then, the network of interlocking shareholdings has decreased significantly. See Fockenbrock and Stratmann (2006), p. 14; H¨opner and Krempel (2005), pp. 10-11; footnote 129. See amongst others Steiger (1998); Matthes (2000), pp. 34-38. For example, Brickley, Lease and Smith (1988), p. 284, find that mutual funds, endowments, foundations and public pension funds are more likely to oppose management than banks, insurance companies and trusts. See section 2.4.3. Cf. Baums (1993). Banks’ historically grown shareholdings in listed corporations decreased after the abolishment of capital gain taxes in 2002. See Fockenbrock and Stratmann (2006), p. 14; Busse (2005), p. 2; H¨opner and Krempel (2005), pp. 10-11. Cf. section 2.4.4.

64

Shareholder Structures and Corporate Governance

stakes, proxy voting rights, supervisory board representatives and their predominant role in providing debt capital.259 The latter three factor cause that banks enjoy disproportionate control on companies in which they own — often relatively small — equity stakes. Especially the fact that banks usually provide debt capital to those companies on a contractual basis might lead to certain interesting consequences. On the positive side, banks might have superior access to internal company information as debt contracts normally assign extensive information rights to the creditor. This internal capital market reduces information asymmetries between borrowers and lenders and, as a consequence, lowers banks’ costs of monitoring.260 Furthermore, the credit relationships are rather long-term compared to equity investments of allegedly myopic institutional investors261 which further increases banks’ informational advantage.262 On the negative side, banks also face extensive conflicts of interest which might limit their role as effective monitors.263 Banks have few incentives to act on behalf of other shareholders for at least three reasons: First, their control rights are usually significantly larger than their cash flow rights due to proxy votes which makes the consumption of private benefits of control less costly.264 Second, the amount of debt provided to firms they hold equity stakes in usually exceeds the equity stake by the factor of ten.265 And third, additional fees (e.g. from investment banking activities) are likely to come into reach if the incumbent management is supported.266 Regarding the predominance of the credit relationship, banks with comparatively small equity stakes can use their disproportionate voting power to divert earnings away from equity, e.g. by charging higher interest rates, or force management to adopt a sub-optimal level of risk 259

260

261 262 263 264

265 266

See Berrar (2001), p. 170; Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), pp. 65-67; Bott (2002), p. 50. Cf. footnote 145. See Emmons and Schmid (1998), p. 36; Helmis (2002), p. 65; Aglietta and Breton (2003), p. 123. Cf. Wenger and Kaserer (1998b), p. 66. See section 3.3.2.3.4. See Becht and B¨ohmer (2003), p. 2. See Wenger and Kaserer (1998a); Matthes (2000), pp. 48-53; Schmidt (2003), p. 10. Furthermore, the fiduciary responsibilities to their stock beneficiaries are likely to diverge from the objectives of the banks’ owners themselves. See Brickley, Lease and Smith (1988), p. 284. See B¨ohmer (2001), p. 110. See B¨ohmer (2001), p. 110.

Shareholder Structures and Corporate Governance

65

to lower the probability that the company defaults on the bank loans.267 However, since banks must also fear that companies change their bank in case of excessive scrutiny, they also have strong incentives to support management uncritically.268 Besides the fact that banks are frequently entangled in conflicts of interest, they are often subject to internal agency problems. At the same time they often do not face strong control.269 Hence, the effects of bank ownership in listed corporations remain ambiguous and empirical studies deliver rather inconsistent results.270 3.3.2.3.2

Insurance Companies (INSR)

Insurance companies are also perceived to be important shareholders even though their direct equity stakes are significantly smaller than those of banks.271 Insurance companies’ discretion regarding investment decisions is partially limited by the German Insurance Supervision Act (“Versicherungsaufsichtsgesetz” (“VAG”)). The act requires that continuing liquidity must be guaranteed and a maximum of safety and profitability should be achieved by an adequate degree of diversification. Therefore, not more than ten percent of the equity of a single company should be acquired. Shares and similar investment may account for a maximum of 30% of the actuarial reserve fund.272 Despite the narrow scope, these limits are traditionally not exhausted by German insurance companies.273 Insurance companies might have contractual relationships (i.e. primarily insurance contracts) with the companies they own stakes in but the importance of such contracts is significantly lower compared to the example of bank credits. Hence, the role of auxiliary interests may be somewhat lower compared to banks. As insurance companies themselves are typically large public corpora267 268 269 270

271

272 273

See Emmons and Schmid (1998), p. 37. See Wenger and Kaserer (1998b), p. 63. See Drukarczyk, Honold and Sch¨uler (1995), p. 118. A positive effect of bank ownership on firm performance is found e.g. by Gorton and Schmid (2000). A negative one is found by Wenger and Kaserer (1998a) , pp. 531-532. Ross (1977b), pp. 41-42, surveying eight early studies reports that five studies indicate a positive relationship, two no significant relationship at all and one a negative relationship. See Baums (1993). In this empirical study, banks own on average between 2.6% (1998) and 3.8% (1993) while insurance companies only own about 1.2%. See figure 6-6. See §§ 54, 54a VAG. See Monopolkommission (1998), pp. 25-26.

66

Shareholder Structures and Corporate Governance

tions, they also face substantial internal agency conflicts which in turn might dilute their control potential.274 3.3.2.3.3

Investment Companies (INVC)

In the context of this study, investment companies refer to private equity companies — both venture capital and buyout companies — that sometimes remain shareholders of listed companies after having sold them to the public.275 Their relationship towards portfolio companies is less ambiguous and less complex compared to banks or insurance companies. Investment companies are hardly restricted in their investment decisions as funds are traditionally raised from sophisticated investors and are not sold to the public. The time horizon of their engagement into a company is limited by the duration of each single fund to a period ranging from three to seven years. The remuneration of the investment company’s management significantly depends on the value it creates in the fund’s investments. Hence, the fund’s management has a crucial interest that the stock price increases — or at least does not fall — after a portfolio company has been floated. As a consequence, interests of investment companies can be expected to be well aligned with those of (other) minority shareholders, i.e. auxiliary interests presumably only play a subordinate role. Being active investors, they usually send representatives to the supervisory board. This first hand access to company information together with their typical industry expertise provides investment companies usually with comparative cost advantages in monitoring.276 3.3.2.3.4

Domestic Institutional Investors (INST)

The type of organization which primarily invests assets held in trust for others, e.g. asset management companies, pension funds, university endowment funds, are labeled as (domestic) institutional investors in the context of this study. They usually hold larger blocks of shares277 , which increases their incentives to acquire information for monitoring purposes because the per-unit costs of acquired information decrease as total costs are spread over a large 274 275 276 277

Cf. footnote 252. For the business model of venture capital firms see Fingerle (2005), pp. 49-101; Nathusius (2005), pp. 9-33. See Gleisberg (2003), p. 106. The empirical analysis shows that the mean size of their shareholdings in 2003 is 8.1%. Prior to 2003, stakes were too rare or not disclosed to prohibit a meaningful interpretation. See table 6-11.

Shareholder Structures and Corporate Governance

67

investment (efficient monitoring hypothesis).278 However, critics warn to treat institutional investors as a homogeneous group as their incentives to perform monitoring may diverge significantly.279 The degree of monitoring activities by institutional investors varies largely across countries.280 Their activism seems to depend on a variety of factors: size of their shareholdings, legal environment, second market liquidity, type of portfolio management (active vs. passive) and the degree of integration between institutional investors.281 German institutional investors are rather minority than majority stakeholders because a certain degree of diversification is necessary to attract investors and required by law for mutual funds. For example, the German Investment Act (“Investmentgesetz” (“InvG”)) — formerly Investment Companies Act (“Gesetz u¨ ber Kapitalanlagegesellschaften” (“KAGG”)) — prohibits investing more than five percent of the separate estate (“Sonderverm¨ogen”) into securities of the same issues and furthermore states that the cumulative voting rights from an investment must not exceed ten percent of the total voting rights of any issuer.282 Ideally, no strategic but solely financial motives should guide investment decisions of institutional investors283 but that may be questioned for the case of Germany, where the largest institutional investors are owned by banks and therefore are limited in their independence (conflicts of interest hypothesis).284 Consequently, institutional investors may align with management and vote against their own fiduciary interests if the management threats to withdraw other business (e.g. the opportunity to manage the company’s pension plans or credit relationships with the parent bank) from (non-cooperating) institutional investors (strategic alignment hypothesis).285 Moreover, the institutional investors’ money managers are also agents and hence their interests may diverge from those of the principals, i.e. the fund investors.286 However, these kind 278

279 280 281 282 283 284 285 286

See Chiang and Venkatesh (1988), p. 1041. Furthermore, synergies from monitoring a large number of companies may arise. See Bott (2002), p. 45. See Woidtke (2002), pp. 100-101. See Gillan and Starks (2003), p. 12. See Bassen (2002), pp. 56-79. Cf. Qiu (2003). See §§ 60 (1), 64 (2) InvG (formerly §§ 8a (1, 3) KAGG). See Helmis (2002), p. 10 See footnote 626. See Pound (1988), pp. 242-244. See Black (1992), p. 170.

68

Shareholder Structures and Corporate Governance

of agency problems can be expected to play a minor role for money managers as less opportunities to divert income exist compared to the case of corporate managers. To sum up, at least three different hypotheses for the monitoring potential of institutional investors can be proposed: efficient monitoring, conflict of interest and strategic alignment.287 Basically, institutional investors (and other investors as well) face two not mutually exclusive — but interdependent — alternatives to act: First, they may judge on the performance of their investment and consequently adjust their portfolio by selling shares (“exit”).288 Second, they may exert their influence by directly communicating with the management (“voice”).289 While the former stock picking (i.e. “exit”) approach was traditionally more widespread in Germany, especially institutional investors’ activism became much more important in recent times and they increasingly voted against incumbent management.290 If institutional investors choose the exit option, this also indirectly exerts pressure on incumbent management. The sale of large blocks usually puts stock prices under pressure and simultaneously can be interpreted as bad news by other investors. Finally, new and more active investors willing to provoke changes in management may gain weight within the shareholder structure.291 The voice option becomes more likely if bad performance is publicly perceived, the industry is accessible to informed outsiders (i.e. especially mature, low-tech industries) or agency problems within the firm are severe.292 One important way of institutional investors’ activism is to provoke management turnover.293 To summarize, despite the lack of contractual relations between institutional 287

288

289

290

291

292 293

See Pound (1988). Cf. Bøhren and Ødegaard (2001), pp. 6-7; McConnell and Servaes (1990), pp. 598-599. Cf. Bott (2002), pp. 44-50. The alternative modes of action “exit” and “voice” go back to Hirschman (1970). Here, “exit” also includes the opposite, i.e. accumulating more shares if the investment appears especially promising. See Helmis (2002), pp. 10-11; Gillan and Starks (2003), p. 4. Kahn and Winton (1998), p. 99, refer to the two approaches as speculation and intervention. Related are also the concepts by Tirole (2001), p. 9, of active and passive monitoring. See Hussla (2005b), p. 31. Agrawal and Mandelker (1990), pp. 143-145, conclude that institutions are rather active than passive monitors. Another indication of institutional investors’ activism is that they are less willing to acquire non-voting preference shares than private investors. See Deutsches Aktieninstitut e.V. (DAI) (2005b), p. 38. Cf. Kahn and Winton (1998); Brickley, Lease and Smith (1988), pp. 271–283. See Gillan and Starks (2003), p. 8. Similarly, passive monitoring is primarily performed by stock market analysts, underwriters, IPO investors and partly banks. See Kahn and Winton (1998), p. 119. See Kahn and Winton (1998), pp. 119-122. For example, institutional investors’ activism forced the CEO of Deutsche B¨orse AG in 2005 to resign. See footnote 161. Cf. Bhagat and Jefferis (2002), p. 74, who report that the likelihood of management turnover increases with strong institutional ownership.

Shareholder Structures and Corporate Governance

69

investors and their investments (i.e. portfolio companies), the existence of auxiliary interests cannot be ruled out as the independence of these investor type must be doubted for the case of Germany. 3.3.2.3.5

Foreign Institutional Investors (INST-F)

Foreign institutional investors are recorded as a separate group because of four reasons: First, foreign institutional investors are relatively new participants in the German capital market. While they had no registered shareholdings in 1993, they held an average fraction of 0.8% in 1998 and 2003.294 Second, they may be useful in importing shareholder activism to Germany as institutional investors occupy a more important role in other countries’ corporate control settings. For example, aggregated equity ownership by institutional investors in the U.S. grew from 6.1% in 1950 to more than 50% in 2002.295 Third, foreign institutional investors are usually independent from banks — and especially independent from the large German universal banks — which narrows the weight of potential auxiliary interests compared to domestic institutional investors. Finally, foreign investors may have an informational disadvantage because they lack knowledge about the country’s legal and institutional framework, the competitive environment, the strategy of the firm and the interest of other larger shareholders. That might lead to the so-called “home-bias” phenomenon, i.e. investors allocate a disproportionate amount of their wealth to the home market. While the second and third argument indicate that foreign institutional investors might represent a rather active monitoring shareholder group, the fourth suggests the opposite: Foreign institutional investors might be rather passive monitors and invest primarily for the purpose of diversification.296 Foreign institutional investors’ activism is further constrained by their difficulties to physically attend shareholders’ meetings as alternative ways to participate (e.g. voting through the internet) are just evolving in Germany.297 294 295 296 297

See figure 6-6. See Gillan and Starks (2003), p. 5. See Bøhren and Ødegaard (2001), p. 7. The OECD (2004), p. 20, Principles of Corporate Governance explicitly state that any impediments to crossborder voting should be abolished. Cf. Scheerer (2006), p. 23.

70

Shareholder Structures and Corporate Governance

3.3.2.4 Federal, State and Municipal Government (GOV) The state may own stakes in listed corporations on the federal, state and municipal level.298 As IPOs are a common form of privatization, which is often carried out in several steps and is especially adequate for very large state-owned companies, the German government takes the place of a significant shareholder group.299 While the extent of contractual relations between companies and the government is rather low the extent of internal agency conflicts and auxiliary interests may be enormous: With respect to the former, the incentives for government representatives to exert monitoring activities are rather limited as potential benefits finally accrue to the dispersed tax payers.300 Regarding the latter, auxiliary interests related to sociopolitical considerations, the economic cycle, the competitive environment and structural development may exist which are likely to be in conflict with the goal of value maximization. An obvious example would be the government’s interest to protect jobs as unemployed voters are likely to be dissatisfied with the incumbent government.301 3.3.2.5 Miscellaneous (MISC) 3.3.2.5.1

Outsider Individuals (INVD)

In the U.S., individual investors started to become involved in monitoring activities early on in the 1960s even before these activities were pursued by institutional investors, which nowadays are perceived to be more powerful due to their larger holdings.302 Private investors tend to see their shareholdings as rather long-term investments303 and are traditionally more concerned with social

298

299

300 301

302

303

Formally, direct governmental participations are administered by the Federal Ministry of Finance (Bundesministerium der Finanzen). Indirect participations are primarily held via the state-owned Kreditanstalt f¨ur Wiederaufbau (KfW). Cf. Bundesministerium der Finanzen (2005). Dyck and Zingales (2004), p. 578, even argue that the state is the one (minority) shareholder (or better stakeholder) common to all companies: The government has a general interest that firms are successful on the market as it expects to receive the corresponding tax payments. Cf. Moldenhauer (2003), pp. 29-34. See Bott (2002), pp. 64-65. See Bott (2002), p. 63. State interests — Bavaria in the case of Lufthansa AG and Lower Saxony in the case of Volkswagen AG — besides unions’ protests significantly slowed down privatization activities in the 1980s. See Portfolio Management (1988), p. 30. Cf. Shleifer and Vishny (1997), pp. 767-769. See Mahoney (1993), pp. 6-9. Known individual investors, which sometimes are also referred to as raiders, are e.g. Carl Icahn, Carl Lindner, David Murdock, Victor Posner and Charles Bluhdorn. See Holderness and Sheehan (1985). See Deutsches Aktieninstitut e.V. (DAI) (2001), p. 2.

Shareholder Structures and Corporate Governance

71

or environmental issues, while institutions are more focused on issues directly linked to share price reactions.304 Outsider individuals might actually be seen as better monitors compared to institutional or corporate shareholders because of at least two reasons: On the one side, they usually have no contractual relationship with the company. On the other side, they are not subject to internal agency conflicts which usually lowers the control incentives of those representatives which actually are in charge of executing the monitoring work.305 However, auxiliary interests may arise if the private investor also holds shareholdings in suppliers, customers or competitors of the company.306 Finally, outside individuals usually face wealth constraints and consequently become underdiversified by holding large stakes of listed companies which might make them more risk averse than desirable from a minority shareholder’s point of view.307 3.3.2.5.2

Employees (EMP)

At the beginning of 2001, of the estimated 6.0 million Germans who were holding equity stakes in listed companies, 1.0 million (16.7%) exclusively held employee shares while a total of 1.6m (26.7%) held both employee and other shares.308 Hence, employees holding stakes in their companies seem to represent an important shareholder group. Furthermore, the number of companies which issued employee shares increased sharply from 36 in 1970 to 140 in 1987.309 Employees by definition are connected with their companies on a contractual basis. As most employees hold their human capital rather undiversified in the company they are employed with, auxiliary interests beyond financial motives can be expected to be extensive if significant parts of their financial capital are invested in the same company as well. Employees with sufficient control rights might force management to pay themselves large quasi-dividends, e.g. in the form of perks, which are less visible and verifiable than ordinary dividends 304

305 306 307 308 309

See Mahoney (1993), p. 16. For more differences between private and institutional investors based on a survey see Deutsches Aktieninstitut e.V. (DAI) (2005b), pp. 37-38. See K¨oke (1999), p. 5. See Bott (2002), pp. 59-60. See K¨oke (1999), p. 5. See Deutsches Aktieninstitut e.V. (DAI) (2001), p. 2. See Portfolio Management (1988), p. 16.

72

Shareholder Structures and Corporate Governance

paid to normal shareholders.310 Further negative consequences of significant employee ownership can arise if compensation in the form of shares represents a significant portion of total compensation: If the company’s share price decreases over a prolonged period of time (or sharply declines suddenly) employees might easily get demotivated as they see their total compensation shrinking.311 However, employees (being natural persons) do not suffer from internal agency conflicts in contrast to (nearly) any kind of institution.312 Moreover, they enjoy direct access to insider information (e.g. about customers, suppliers or competition). Lastly, employee shareholders might limit potential efforts by large blockholders to extract private benefits of control without turning to the courts e.g. by showing lower motivation or even by withholding their services.313 3.3.2.5.3

Others (OTH) and Treasury Shares (TRE)

The owner group of others, i.e. other shareholders, refers to those registered shareholdings which are not attributable to any other category (e.g. retail traders or independent foundations), are not properly specified in the Hoppenstedt Aktienf¨uhrer (e.g. “other investors”) or cannot be definitely identified. Hence, this category is a melting pot of potentially very diverse shareholder types and, consequently, any further analysis would not make sense or is simply not possible respectively. As in the empirical study (see section 4.2), only 29 out of all shareholder entries (i.e. 1.02%) fall into the category OTH on the ultimate owner level. Therefore, its importance is relatively small and this shareholder group can be neglected in the further analysis. Under certain circumstances companies may acquire shares of their own company up to a maximum limit of 10% of the total share capital according to the Stock Corporation Act.314 These shares are labeled as treasury shares and the corresponding transactions are usually referred to as share repurchases.315 As a 310 311 312

313 314 315

See Tirole (2001), p. 28. See Kemper (2001), p. 10. For example, unions as institutions representing employees have both the power and incentive to monitor but are supposed to suffer under agency conflicts themselves . Cf. Stiglitz (1985). See Dyck and Zingales (2004), p. 578. See § 71 AktG. For a comparative analysis of the accounting treatment of treasury shares under HGB and International Financial Reporting Standards (IFRS) see G¨obel and Kormaier (2006).

Shareholder Structures and Corporate Governance

73

consequence, the company may be formally registered as its own shareholder. However, as the Act specifies that treasury shares do not entitle to any rights (neither dividend nor voting rights)316 , they cannot be treated as a normal owner group which has its own interests and means to enforce them.317 3.3.3

Shareholder Concentration as Differentiating Factor

As pointed out in the discussion of institutional investors’ interests (section 3.3.2.3.4), shareholders generally have two options to react if they are dissatisfied with a company’s performance: exit or voice.318 While the former is open to shareholders regardless of the size of their holding — even though the consequences may vary largely319 — the latter requires a certain degree of concentrated ownership since otherwise the possibility to intervene is de facto non-existent. While early literature often differentiated between firms with a dominant shareholder (owner controlled firms) and those with a dispersed ownership (manager controlled firms)320 , this approach nowadays seems oversimplified.321 In the context of this study, five (partially) proprietary categories of outside block ownership are distinguished whereas the borders are derived from voting thresholds defined by the Stock Corporation Act or the company’s charter (see figure 3-2)322 : 0.0-4.9% (smallholder/ free float), 5.0-24.9% (simple blockholder/ type 1), 25.0-49.9% (dominating blockholder/ type 2), 50.074.9% (simple majority blockholder/ type 3) and 75.0-100% (super majority blockholder/ type 4).323 316 317

318 319

320

321 322

323

See § 71b AktG. For the rationale, types and case studies of share repurchases see Weston, Siu and Johnson (2001), pp. 535559. See footnote 288. Small shareholders may sell their shares at any time without market impact, while larger stakes — depending on the absolute size of the company and the liquidity of the shares — can usually not be sold without depressing the market price. Therefore, especially larger stakes are often traded through negotiated block trades. Cf. Barclay and Holderness (1991). See Short (1994), pp. 207-215. Cf. Holl (1975). A slightly more differentiated classification is provided by Monsen and Downs (1965), p. 223, who divide firms into four groups: owner-managed firms, managerial firms (diffused ownership, concentrated ownership), non-ownership firms and fiduciarily owned firms. See Thonet and Poensgen (1979), p. 23; Kang and Sorensen (1999), p. 123. Cf. Cubbin and Leech (1983). Similarly, Jenkinson and Ljungqvist (2001), p. 402, differentiate five categories of block ownership: For example, they refer to blockholders which own between 25% and 50% as “blocking minority control”. The labels blockholder type 1-4 are used in the empirical study. See table 6-6. Even though in reality lower levels of ownership would yield control privileges associated with certain thresholds as usually not even close to 100.0% of voting rights are present at the shareholders’ meeting, these levels vary from company to company, from shareholders’ meeting to shareholders’ meeting and are not predictable which would let appear any other thresholds arbitrary. For example, Baums and Fraune

74

Shareholder Structures and Corporate Governance

Smallholder Simple Blockholder (Blockholder Type 1) Dominating Blockholder (Blockholder Type 2) Minority Shareholder Simple Majority Blockholder (Blockholder Type 3) Super Majority Blockholder (Blockholder Type 4) Majority Shareholder 0

5

25

50

75 100 Voting Rights in %

Source: Author’s illustration. Cf. table 6-6.

Figure 3-2: Shareholders’ Concentration Scheme 3.3.3.1 Minority Shareholders 3.3.3.1.1

Smallholders

Smallholders are shareholders whose individual stakes do not exceed 5% and hence these stakes usually are not disclosed to the public.324 As a result, no detailed information about this shareholder group is available and the aggregated stakes of all smallholders are usually referred to as “free float”. It is commonly assumed that auxiliary interests play a subordinate role as smallholders see their shareholdings primarily as financial investments.325 Moreover, their power is

324 325

(1995) report that in 1992 for the largest German companies with a free float over 50% on average about 58% of outstanding voting rights are present at the shareholders’ meeting. In 2005 and 2006, an average presence of 45.1% and 49.6% is reported for 19 DAX-companies by Hussla (2006), p. 25. Bott (2002), pp. 20-21, argues that it does not make much sense to differentiate shareholder groups according to certain voting thresholds and one should rather distinguish between blockholders with and without auxiliary interest. Cf. Deutsches Aktieninstitut e.V. (DAI) (2005a), p. 9; Jenkinson and Ljungqvist (2001), pp. 402-404. See footnote 602. See Iber (1987), p. 60.

Shareholder Structures and Corporate Governance

75

rather limited as underlined by the following statement: “At the first glance, minority shareholders appear to be the most vulnerable of financial agents. Unlike the debt contract, there is no contractual obligation to distribute dividends [. . . ] and unlike the controlling group, they have no means of direct control over the managers.”326 Along with an increasing free float portion, the total number of shareholders increases and, as a consequence, the wealth of each shareholder will typically depend less on the success of the company. Thus, it is more difficult to organize very diffuse ownership interests into an effective instrument adapted for monitoring management.327 Doubts of effective control through dispersed shareholders have been raised frequently, amongst others by B ERLE (1965): “Stockholders act like an unorganized, usually inert, political constituency. They are a ‘field of responsibility’ — far, indeed, from an entrepreneurial controlling force.”328 Similarly, G ROSSMAN AND H ART (1980) point to the problem of collective action regarding the production of the public good of monitoring: “In all but the smallest groups social choice takes place via the delegation of power from many to few. A fundamental problem with this delegation is that no individual has a large enough incentive to devote resources to ensuring that the representatives are acting in the interest of the represented. Since the representatives serve the Public Good, the social benefit to monitoring their activities is far larger than the private benefit to any individual. That is, the Public Good is a public good and each person attempts to be a free rider in its production.”329 The basic problem is that monitoring only becomes attractive to shareholders if the expected benefits — i.e. capital gains from increasing share prices or higher dividend payments — exceed the associated costs of monitoring; i.e. control is not for free. Since the benefits accrue to all shareholders, which participate on a pro-rata basis according to the size of their shareholdings, the monitoring 326 327 328 329

Aglietta and Breton (2003), p. 134. See Demsetz (1983), p. 386. Berle (1965), p. 31. Cf. Peterson (1965), pp. 18-24. Grossman and Hart (1980), p. 42.

76

Shareholder Structures and Corporate Governance

costs have to be exclusively borne by the monitoring party itself. As the ratio of costs to benefits is likely to be unfavorable for relatively small shareholdings, it is rational for smallholders not to engage in monitoring activities (“rationale apathy”).330 Efforts to bundle interests in associations dedicated to the protection of smallholders are only partially successful in Germany.331 And the (rare) cases where smallholders attend shareholders’ meetings, due to their limited experience as investors, they may probably vote rather in favor of the known, incumbent management than in favor of an unknown outside investor trying to provoke changes in management.332 Even if smallholders would be willing to participate more actively in the corporate control process, their opportunities to do so are rather limited and usually constrained to exercising their voting rights. Class action suits do not exist in Germany and as much as ten percent of the voting rights are necessary to directly sue management for damages.333 To sum up, companies with a large free float portion usually lack adequate control. Nevertheless, this also might imply some advantages: A dispersed ownership structure assures little interference between shareholders and management which might in turn motivate management to increase its efforts. Thus, there might also exist a trade-off between control and initiative.334 Finally, the free rider problem can be theoretically overcome by market forces and especially (hostile) takeovers. However, the market for corporate control is rather underdeveloped in Germany, yet, and the basic problem of externalities produced by monitoring is not overcome.335 3.3.3.1.2

Simple Blockholders

In the course of evaluating the control potential of blockholders not only the power of blockholders should be kept in mind but also their incentives to actually engage in monitoring.336 As discussed, smallholders hold their investments 330 331

332 333

334 335

336

See Bott (2002), p. 22. See Matthes (2000), p. 37. The largest and best known German association in this context is the “Deutsche Schutzvereinigung f¨ur Wertpapierbesitz” (“DSW”). However, its influence is perceived to be rather limited. See Easterbrook and Fischel (1983), p. 177. See footnote 208. However, the KapMuG introduced in November 2005 provides investors with a new instrument in capital market suits. See section 2.4.2. See Burkart, Gromb and Panunzi (1997), p. 694. Cf. Grossman and Hart (1980), pp. 42-43. A good overview about the advantages and disadvantages of dispersed/concentrated ownership/voting power is provided by Becht (1997), p. 25. See Scott (1999), p. 12.

Shareholder Structures and Corporate Governance

77

passively as they neither have the power nor the incentives to discipline management. In contrast, simple blockholders (i.e. the blockholder owns between 5.0% and 24.9% of the voting rights) can hold shares both passively or for control purposes. Descriptive statistics suggest that the control aspect actually is important as the stakes of outside blockholders are clustered around relevant voting thresholds.337 If blockholders normally use their voting power to secure benefits of control unavailable to other shareholders, such blocks would trade at a premium to exchange prices. Mean premiums of about 20% in the U.S. (1978/1982) or roughly 10% in Germany (1990-2000) suggest that benefits of control actually exist and vary significantly across countries.338 The Stock Corporation Act or the corporate charter specifies certain control rights for those shareholders who hold at least 5% or 10% respectively. The former are entitled i) to demand that a shareholders’ meeting be called, ii) to require the inclusion of certain items on the agenda, iii) to enforce these rights by going to the courts if necessary and iv) to file a petition to the courts for the replacement of the corporation’s liquidators if important reasons do exist.339 Blockholders owning at least 10% may i) block the settlement of certain claims of the corporation against the management or supervisory board, ii) petition to the courts under certain circumstances for the appointment of a special auditor and iii) force separate resolutions on individual or all members of the management and supervisory board regarding their discharge from responsibility.340 One of the positive aspects of the presence of blockholders is that “[. . . ] large shareholders can provide the public good associated with monitoring activity [. . . ]”.341 Generally, large minority shareholders have incentives to take control if management shirks and may effectively threat with takeovers, e.g. by raising their minority stakes or lining up with other blockholders.342 Furthermore, evi337 338

339 340

341 342

See Becht and B¨ohmer (2003), p. 10; Schmidt (2003), p. 13; section 6.4.1.2.2. See Barclay and Holderness (1989), p. 372; Dyck and Zingales (2004), p. 551. Dyck and Zingales (2004) report block premiums of only 1% for the U.S. (1990-2000) which is considerably lower than the 20% reported by Barclay and Holderness (1989). Barclay and Holderness (1989), pp. 378-383, also formulate alternative hypotheses for the existence of block premiums, i.e. the superior information hypothesis and the overpayment hypothesis, but the data are most supportive of the private benefits hypothesis. See §§ 122, 265 (3) AktG. See §§ 50, 53, 93 (4), 116, 117 (4), 120 (1), 142 AktG. Cf. Jenkinson and Ljungqvist (2001), p. 402; Schneider and Heidenhain (2000), pp. 13-14. Some of these rights also apply for shareholders owning shares of the pro-rata share capital amount or EUR 500,000 or EUR 1m. Mello and Parsons (1998), p. 103. See Buz (1994), p. 291.

78

Shareholder Structures and Corporate Governance

dence exists that outside block ownership significantly decreases management’s ability to limit voluntary disclosure of financial information.343 However, these potential benefits do not come without a cost. Blockholders usually face a tradeoff between control and liquidity as larger blocks usually cannot be sold on the stock exchange without depressing the share price.344 Finally, the absolute risk and opportunity cost of owning a given stake increases with its value (opportunity cost hypothesis) which might make blockholders risk averse beyond the optimal level desired by smallholders.345 3.3.3.1.3

Dominating Blockholders

Shareholders holding at least 25.0% of the company’s shares are referred to as dominating blockholders. If they hold 25% plus one share, they may block resolutions at the shareholders’ meeting which require three-quarters of the votes cast.346 Hence, they enjoy the power of veto over corporate charter amendments, supervisory board changes and profit transfer and control agreements.347 Especially the latter provide potential super majority blockholders348 with extensive possibilities to expropriate minority shareholders.349 Dominating blockholders’ ability to prevent changes in the supervisory board may serve as a defensive measure against hostile takeovers.350 The increase in the size of the blockholdings compared to simple blockholders is usually associated with growing involvement in value-increasing activities, i.e. monitoring. However, larger blockholdings play a dual role as they not only reduce managerial agency problems but also raise new agency problems (agency costs of expropriation) with regard to minority shareholders.351 It is argued that dominating blockholders are interested in securing their standing within the company and are often emotionally bounded to their shareholding 343 344

345 346 347

348 349 350 351

See Luo, Courtenay and Hossain (2006), p. 15. See Bolton and Thadden (1998). Cf. Maug (1998) who argues that a liquid stock market increases monitoring incentives because the associated costs can be covered through informed trading. See Thomsen, Pedersen and Kvist (2006), p. 248. Cf. Barclay and Holderness (1991). See Schneider and Heidenhain (2000), p. 14. See Jenkinson and Ljungqvist (2001), pp. 402-404, who provide an overview of control rights associated with certain block sizes. See section 3.3.3.2.2. See section 3.4.2.1. See Jenkinson and Ljungqvist (2001), p. 421. See Chirinko, van Ees, Garretsen and Sterlen (2004), p. 129.

Shareholder Structures and Corporate Governance

79

or the company respectively and might have significant auxiliary interests.352 For example, a dominating blockholder who is an important customer of the company might use its stake as a means of exercising strategic influence. As dominating shareholders represent the largest shareholder in 67.4% of the cases in the empirical study, they have to be seen as a very important power within the corporate control system.353 With regard to control transfers, block trades presumably are the preferred instrument as they help to maintain low ownership concentration (in contrast to tender offers) and large benefits of control. However, it should be noted that firms acquiring or attempting to acquire stakes above 30% are obliged to submit a mandatory takeover offer according to German takeover legislation.354 3.3.3.2 Majority Shareholders 3.3.3.2.1

Simple Majority Blockholders

Both types of majority blockholders by definition simultaneously take the position of the largest shareholder which is another frequently used measure of ownership concentration. Simple majority blockholders hold between 50.0% and 74.9% of a company’s voting rights.355 The power of simple majority shareholders crucially depends on the existence of other blockholders and particularly on the presence of a blocking minority.356 While a blocking minority may significantly curtail majority shareholder’s expropriation attempts, the non-existence of the same usually provides the majority shareholder with far reaching power. As a consequence, controlling ownership stakes may sell at a substantial premium.357 In the U.S., majority shareholders are typically either officers or directors in their firms. Hence, it can be argued that majority shareholders rather lead 352 353

354 355

356

357

See Iber (1987), p. 62. Out of the 129 reported blockholder type 2 stakes (for 1993, 1998 and 2003), 87 companies only had one type 2 (type 3 or type 4) blockholder, 22 had two type 2 blockholders and 3 companies had as much as three type 2 blockholders. ¨ table 2-1. See §§ 29-39 WpUG; Some empirical studies question whether small shareholders owning five to ten percent are sufficient to exercise effective control and hence use majority control, i.e. 50% of voting shares, as relevant threshold for blockholders. See amongst others Murali and Welch (1989), p. 386. See Jenkinson and Ljungqvist (2001), p. 402. In the empirical study, in 17 cases (20.5%) out of the 83 type 3 blockholders (1993, 1998 and 2003) a blocking minority was present. See Easterbrook and Fischel (1982), p. 699.

80

Shareholder Structures and Corporate Governance

than only monitor management teams.358 Especially in countries with relatively weak investor protection, e.g. Germany, high levels of concentrated ownership are often assumed to weaken minority shareholders’ wealth position because blockholders might try to extract private control rents and dilute minority shareholders cash flow rights.359 This might be even more true for majority shareholders as they do not have to fear hostile control activities, e.g. takeover attempts. Nevertheless, majority blockholders may of course deliberately engage in friendly transfers of control.360 Controlling shareholders might also try to profit from their relative informational advantage by insider trading.361 With regard to the effects on the investment behavior, two possible outcomes appear plausible: As fewer internal funds are left after expropriation through the majority shareholder, companies with majority shareholders might be induced to reduce investment expenditures to hide their mismanagement and thereby increase short-term cash flows. Alternatively, they might try to increase certain investments in the short run as a means of expropriating funds, e.g. “[. . . ] by making ‘sweetheart’ deals with other firms they control [. . . ]”.362 3.3.3.2.2

Super Majority Blockholders

Blockholders holding at least 75.0% of a company’s shares, so-called super majority blockholders, do not even have to worry about a blocking minority. Super majority control significantly increases majority shareholders’ effective control level: They are enabled to enforce amendments to the corporate charter, to change the composition of the supervisory board and to close profit transfer and control agreements.363 Considering the lack of control, one can argue that minority investors probably would not be willing to contribute additional capital to the firm or to maintain the stakes they already own.364 However, expropriation of minority shareholders may not be the predominant motive for super majority shareholders. Otherwise, it would remain unclear why they of358 359 360

361 362 363 364

See Holderness and Sheehan (1988), pp. 319, 324. See Ehrhardt and Nowak (2003b), p. 365. See Holderness and Sheehan (1988), p. 340. They also report that stock prices and management turnover increases when majority blocks are traded. See Demsetz (1986), pp. 313-316. Holderness and Sheehan (1988), p. 338. See Jenkinson and Ljungqvist (2001), p. 402. See Holderness and Sheehan (1988), p. 321.

Shareholder Structures and Corporate Governance

81

ten hold stakes above the 75% threshold since these additional shares do not ease expropriation.365 An explanation for the observed pattern might be that super majority shareholders could be tempted to force minority shareholders to accept a cash compensation in exchange for their shareholdings (“squeeze out”). This procedure requires an even higher super majority of at least 95% of the company’s share capital.366 A squeeze out facilitates the majority blockholder to save costs by e.g. delisting the subsidiary company, abandoning shareholders’ meetings and reducing financial reporting requirements. The majority shareholder determines the level of the cash compensation in accordance with the status of the company at the point of time when the squeeze out decision was made, i.e. at the relevant shareholders’ meeting.367 As the determination of an appropriate cash compensation is largely unregulated368 , minority shareholders may only legally challenge the appropriateness of the cash compensation but have no means of preventing the transmission of their shares.369 3.4

Shareholders, Management and Conflicts of Interest

3.4.1

Introductory Remarks

For the case that companies are exclusively financed through the issuance of homogenous securities and assuming a perfect capital market, it can be shown that the maximization of the company’s market value is the common objective of all shareholders. Assuming further that the management acts in the best interest of the shareholders, no conflicts of interest would emerge and therefore the value maximizing investment program can be carried out independently of the preferences of the providers of capital.370 However, the preceding analysis illustrates that management’s and shareholders’ (auxiliary) interests can differ 365

366 367 368 369

370

See Holderness and Sheehan (1988), p. 325. In the empirical analysis, the average size of the 152 super majority blockholdings varies between 90.2% (1993) and 91.6% (1998) which is considerably larger than the super majority threshold of 75%. See tables 6-12 and 6-13. See § 327a AktG. See § 327b AktG. See Jenkinson and Ljungqvist (2001), p. 402. The lack of a German squeeze-out regulation was criticized frequently up to its introduction in January 2002. An empirical analysis of squeeze outs in Germany is provided amongst others by Helmis (2003). See Bott (2002), p. 7. For a more detailed description of the separation theorem see amongst others Fisher (1930); Drukarczyk (1993), pp. 34-38, 46-50.

82

Shareholder Structures and Corporate Governance

significantly which gives rise to various conflicts of interest. These conflicts will be analyzed and classified in the following sections. Generally speaking, conflicts of interest can exist between the providers of capital and the management as well as within both groups.371 In this section, only conflicts of interest between the shareholders as providers of equity capital and the management, i.e. agency conflicts of outside equity, as well as within the shareholder group will be discussed as these aspects are closely related to insider ownership or more generally ownership structures. Besides, so-called agency conflicts of debt exist, which refer to inefficiencies emerging from conflicting interests of shareholders (or their representatives in the management board) and creditors, i.e. banks or bond owners. Since the latter constitute an independent branch of agency research, it is beyond the scope of this study to address these conflicts explicitly.372 3.4.2

Systematical Analysis of Potential Conflicts of Interest

3.4.2.1 Between Shareholders and Management Having referred to the potential of conflicts of interest between shareholders and management in sections 2.2.2 and 3.3.1, in this section a closer look at the nature of these conflicts will be taken.373 All these conflicts have in common, that “[. . . ] the delegation of control creates a conflict of interest between those who make decisions and those who bear the consequences.”374 It will be differentiated between those conflicts related to the manager’s job and position, i.e. his inputs into and withdrawals from the company, and those related to the company’s investment program as illustrated in table 3-2.375 The effort provision problem, the consumption of perquisites and entrenchment will be attributed to the former while the risk preference, time preference, overinvestment and sub-optimal investment problem will be related to the latter one. 371 372

373

374 375

Cf. Kaserer, Achleitner and Moldenhauer (2005), p. 4 Specifically, the underinvestment, asset substitution, value dilution, dividend payout, overinvestment and asset conversion problems are common types of agency problems between providers of debt capital and shareholders. For a short review of these problems see Bott (2002), pp. 12-13. Cf. Kim and Sorensen (1986); Anderson, Mansi and Reeb (2003); Winkens (2000); McConnell and Servaes (1995), pp. 132-133. These conflicts have been frequently discussed in earlier literature under the keyword managerialism. See Mann (2003), p. 58. Grossman and Hart (1988), p. 176. While the effects of the former impose direct costs on the company, the latter are of a more indirect nature as a sub-optimal investment policy will prevent revenues from reaching their value maximizing level. See Wolf (1999), p. 128.

Shareholder Structures and Corporate Governance

83

In a ceteris paribus analysis, a negative relationship between management’s effort provision and its utility is generally assumed.376 Consequently, management would be induced to raise its utility by reducing its efforts on the job, so-called shirking, which might finally lead to bad decision making. This in turn would finally be detrimental to firm value.377 From a more differentiated perspective, it also appears plausible that efforts (i.e. the job) increase the manager’s personal satisfaction and in the long run also his reputation leaving the net effect undetermined.378 Agency conflicts related to … … the manager’s job/position

… the company’s optimal investment program

• Effort provision problem • Shirking • Bad decisions

• Risk preference • Risk aversion • Overdiversification

• Consumption of perquisites • Private benefits of control • Abuse of corporate resources

• Time preference • Short term orientation

• Entrenchment • Increase management’s replacement costs

• Overinvestment • FCF-hypothesis • Empire building • Sub-optimal investment • Management’s private costs of investment

Source: Author’s systematization.

Table 3-2: Systematization of Equity Agency Conflicts Somewhat clearer appears the case for management’s consumption of perquisites.379 Perquisites can be described as non-cash benefits of monetary value exceeding management’s contractual compensation package, which management board members can consume due to their position. Admittedly, these benefits do not have a direct or indirect positive impact on the company and debit the company’s accounts.380 If the management does not own a fraction of 376 377 378

379

380

See Barnea, Haugen and Senbet (1985), p. 28. For a broader discussion of the concept of shirking see Alchian and Demsetz (1972), pp. 779-781. See Bott (2002), p. 11. Since the reputational argument could be brought up in favor of the management for each agency problem it will not be repeated for each conflict. Synonymously used are also the terms non-pecuniary benefits, fringe benefits, and private benefits of control (which includes other aspects as well). See e.g. Jensen and Meckling (1976), pp. 312-319; Dyck and Zingales (2004), pp. 540-543; Ehrhardt and Nowak (2003a), p. 225. See Bott (2002), p. 10.

84

Shareholder Structures and Corporate Governance

the company’s equity, it will only consume the associated benefits while bearing no costs at all. However, the distinction between use and abuse of corporate resources may become difficult in practice: While the use of a corporate jet can enable management to react fast on business opportunities and help to use management’s work time efficiently, the prominent example of the U.S. corporation RJR which almost maintained a whole airline — 10 aircrafts, 36 company pilots and a luxury corporate hangar for use by CEO Ross Johnson and numerous relatives and friends — represents a rather clear case of the consumption of perquisites.381 Actions of the management that primarily make it valuable to the company and costly to replace are summarized under the term entrenchment. Typical examples are manager-specific investments, where the incumbent management has much private information or where it has a competitive edge over potential rivals inside or outside the company. As a consequence, managers might be able to extract higher wages, prolonged employment contracts, larger perquisites and finally gain more discretion about the corporate strategy.382 With regard to the company’s investment program, differences in the risk preference of shareholders and management are likely to occur. While (welldiversified) shareholders are assumed to be risk neutral, managers can generally be assumed to be (company specific) risk averse because of the following three reasons: First, and especially for the case of Germany, managers’ claims against the company usually contain a significant fixed component. Hence, they profit disproportionately from a favorable development of the company while an unfavorable development might put their job at risk.383 Second, management’s reputation will probably suffer more from a corporate crisis (e.g. insolvency or bankruptcy) than it will benefit from an increase in earnings, which in turn will raise shareholders’ expectations regarding future profits. Third, managers’ cap381

382

383

See Burrough and Helgar (1990); Yermack (2005). Ang, Cole and Lin (2000) find that agency costs, measured as total standardized expenses, are higher for outsider managed firms and inversely related to managerial ownership. See Shleifer and Vishny (1989). This problem can also be interpreted as a hold-up conflict regarding company specific investments in real assets and human capital. See Mann (2003), pp. 59-60. Cf. Rajan and Zingales (2000). It can be argued that a higher sensitivity of management’s or CEO’s wealth to stock volatility would implement riskier policy choices, including relatively more investment in research and development (R&D), less investment in property, plant and equipment, more focus and a higher leverage. See Coles, Daniel and Naveen (2006).

Shareholder Structures and Corporate Governance

85

ital is undiversified because it is invested in their companies for the most part (e.g. salary, bonuses, non-transferable pension rights or health benefits).384 As a consequence, management has incentives to reduce firm specific risk and the variability of profits, e.g. by overdiversifying the firm’s activities or foregoing high-risk positive net present value (NPV) projects. Conflicts of interest do not only exist with respect to the risk of investment projects but also with regard to their time horizons, the so-called time preference problem. While shareholders prefer that all positive NPV projects are carried out independently of their planning interval, it can be assumed that managers have a strong preference for short-term projects. For example, managers are often promoted to other positions within the company or even leave the company. Hence, they do not profit from an investment’s long-term success. Furthermore, the pressure from the capital market to meet analysts’ quarterly earnings forecasts motivates management to invest in predictable short-term projects.385 This short-termism of managerial decision making translates into a failure to exploit long-term growth opportunities and hence does not maximize shareholders’ value.386 The term overinvestment refers to the observation that corporate managers might also undertake negative NPV projects as they derive benefits from the sheer size of the assets they are controlling. Their preference to invest beyond the optimal level is also labeled as “empire building”.387 Similarly, management may withhold free cash flows (FCF) instead of paying them out to owners in the form of dividends or share buybacks. By doing so, management reduces future needs for external financing which in turn would allow additional shareholders in the case of equity financing or creditors in the case of debt financing to get insights into management’s dealings.388 In contrast to the overinvestment problem, the exact opposite, i.e. a sub-optimal investment problem, may arise as well.389 It can be argued that managers in384 385

386 387 388 389

See Mann (2003), p. 58; Short (1994), p. 204. Cf. Benston (1985). As a consequence, some corporate executives, most prominent Wendelin Wiedeking, CEO of Porsche AG, forbear from publishing quarterly results. Cf. Wiedeking (2005). Cf. Roberts and Van den Steen (2000), p. 20. See Jensen (1986); Jensen (1993). See Wolf (1999), p. 128. The term “sub-optimal investment” is used instead of “underinvestment” as the latter usually is related to agency costs of debt. See Wolf (1999), p. 129.

86

Shareholder Structures and Corporate Governance

cur private costs of investment arising from additional oversight requirements. Firms which expand existing facilities or launch new product lines simply require managers to spend extra efforts. If managers generally prefer to work less, positive NPV projects may be foregone simply to reduce management’s work load.390 Another reason for underinvestment might be low free cash flows. Management could raise external financing to carry out positive NPV projects but might fear scrutiny from additional monitoring through the new investors.391 3.4.2.2 Among Shareholders It can be reasonably assumed that the interests of smallholders and blockholders diverge significantly as discussed in section 3.3.3. As a comparatively minor conflict, majority shareholders are often expected to prefer lower payouts and higher earnings’ retention because of the associated tax savings whereas smallholders — or more generally, minority shareholders — often are assumed to have a preference for dividend payments.392 However, the more serious problem is that large shareholders can extract private benefits of control which in turn harm minority shareholders’ wealth as they have to (partially) bear the costs of such behavior without receiving any benefits.393 This problem might be particularly important for Continental European countries as highlighted by the following statement of ROE (1994): “If the US/UK corporate governance problem is one of ‘strong managers, weak owners’ [. . . ] in continental Europe, and many other countries, the corporate governance problem is rather one of ‘strong block owners, weak minorities’.”394 In particular super majority blockholders have the power to expropriate minority shareholders, e.g. by closing profit transfer or control agreements. Concrete examples of such behavior would be transfers of hidden reserves to the controlling shareholders or group losses (profits) which can be transferred into (out of) the 390 391 392 393 394

See Aggarwal and Samwick (2006), pp. 489-491. See Stulz (1990), pp. 3-5. See Peterson (1965), pp. 19-20. Cf. Karpoff (1999), p. 245. See section 3.3.3.2.1. Roe (1994) cited according to Jenkinson and Ljungqvist (2001), p. 429. Cf. Kaserer, Achleitner and Moldenhauer (2005), pp. 4-5. With respect to expropriation rents of controlling shareholders in small Hong Kong firms cf. Cheung, Stouraitis and Wong (2005).

Shareholder Structures and Corporate Governance

87

subsidiary company at the expense of minority shareholders.395 Nevertheless, large blockholders in general have more incentives to actively monitor management due to the size of their holdings for at least two reasons: On the one hand, the benefits of such efforts (e.g. increases in the share price) are more likely to surpass the associated monitoring costs. On the other hand, large shareholdings are rather illiquid and, hence, less salable.396 As blockholders have cash flow incentives to make informed decisions, other investors are willing to delegate decisions to the blockholders. However, the blockholders might abuse this discretion and extract private benefits for themselves. Thus, large shareholders’ “[. . . ] abilities to create and to destroy value are closely related.”397 Generally, these conflicts might be alleviated if more than one blockholder is present as this significantly decreases large shareholders’ possibilities to extract private benefits.398 Similarly, voting caps may improve the protection of small shareholders against expropriation which in turn certainly increases the scope for entrenchment on behalf of the management.399 In summary, the variety, complexity and interdependency of conflicts of interest in the modern stock corporation seem to be enormous leading to the question how these conflicts can be overcome? 3.4.3

Ways and Means for Overcoming Conflicts of Interest

Even though the discussed conflicts among shareholders also represent a serious issue of corporate control, this study will predominantly focus on the conflicts of interest between shareholders and management. Evaluating the diverging interests of shareholders and management in the public stock corporation, S HLEIFER AND V ISHNY (1997) pinpoint to the apparent weak position of investors with the following statement: “[. . . ] why do investors part with their money, and give it to managers, when both the theory and the evidence suggests that managers have enormous discretion about what is done with that money, 395

396 397 398 399

See Jenkinson and Ljungqvist (2001), p. 403; Shleifer and Vishny (1997), p. 758. Cf. Barclay and Holderness (1989), p. 374. See Mann (2003), p. 66. Burkart, Gromb and Panunzi (2000), p. 648. Cf. Shleifer and Vishny (1997), pp. 758-761. See Guti´errez and Trib´o (2004), p. 22. See Goergen, Manjon and Renneboog (2004), p. 4. Cf. Edwards and Weichenrieder (2004).

88

Shareholder Structures and Corporate Governance often to the point of being able to expropriate much of it? [. . . ] Their investment is sunk and nobody — especially the managers — need them.”400

A rather early answer to this problem was given by FAMA AND J ENSEN (1983 B ) who presumed that the separation of decision and risk-bearing function in stock corporations exists because “[. . . ] these organizations separated the ratification and monitoring of decisions from initiation and implementation of the decisions.”401 However, this view seems a bit too myopic: It neither makes sense nor is it ex-ante possible to write complete contracts for management’s employment which would effectively prevent it from undertaking any actions harmful to shareholders.402 As a consequence, it is likely that other means play an important role as well. As outlined in section 2.2, several corporate governance mechanisms exist, which are expected to alleviate the presented problems. These instruments are most commonly differentiated into mechanisms of internal and external control. Another plausible and very appropriate distinction labels corporate governance mechanisms as monitoring and control mechanisms on the one side, and incentive and bonding mechanisms on the other side.403 Monitoring and control mechanisms reduce information asymmetries between shareholders and management but do not affect the utility functions of the involved parties. Management is simply prevented from maximizing its own utility by means of monitoring and control. “The preferences of principals and agents are only aligned in the sense that punishment up to the loss of the job is threatened in the case of non-obedience to the principals’ wishes.”404 The majority of the presented corporate governance mechanisms belongs into this group with compensation of management and supervisory board members and insider ownership being the only exceptions.405 Contrary to monitoring and control mechanisms, in the case of incentive mechanisms, the degree of information asymmetries between management and share400 401 402 403 404 405

Shleifer and Vishny (1997), p. 748. See Fama and Jensen (1983b), p. 302. Cf. Fama and Jensen (1983a). See Klein (1983), p. 368. See Wolf (1999), pp. 26-44. Wolf (1999), p. 39. Cf. figure 2-1.

Shareholder Structures and Corporate Governance

89

holders does not change while principals’ and agents’ utility functions are equal. While the hidden information problem persists406 , management will no longer exploit its informational advantage as its gains from maximizing shareholders’ utility are larger than its gains from exploitation.407 The remainder of this study will focus on insider ownership as the most important corporate governance incentive mechanism. Certain other corporate governance mechanisms will also be discussed from time to time.

406 407

See section 4.1.2.4. See Wolf (1999), p. 39.

Insider Ownership and Performance

4

91

Insider Ownership and its Effects on Corporate Performance

While the previous chapter applied a bottom-up analysis of shareholders’ interests and the resulting conflicts, this chapter takes a top-down perspective on the phenomenon of insider ownership and its effects on corporate performance. Therefore, the role of insider ownership according to different economic theories will be discussed first (section 4.1) before empirical evidence about the insider ownership-performance relationship in Germany and other countries will be presented (section 4.2). 4.1

Insider Ownership According to Different Theoretical Views

4.1.1

The Neoclassical View

Conventional economic theory — usually labeled as the neoclassical view — primarily analyses markets, i.e. the determination of prices and outputs, from an equilibrium perspective while firms are often “[. . . ] described as a ‘black box’.”408 The production process, i.e. the whole process from purchasing production factors to selling the produced goods, is largely disregarded and reproduced by a (simple) production function or as D EMSETZ (1983) put it: “The chief mission of neoclassical economics is to understand how the price system coordinates the use of resources, not to understand the inner workings of real firms.”409 The basic premisses of the neoclassical theory are based on a legal system with three basic components: i) the distribution of property rights is stemming from the concept of private property, ii) the transmittal of these rights is based on the concept of freedom of contract and iii) individuals’ liability for breaches of contract and illegal actions.410 Additionally, markets are characterized by perfect competition, i.e. atomicity of supply and demand, homogeneity of goods, perfect and complete information, equal access to all resources and no market entry barriers.411 Market participants act rationally, immediately and only ad408 409 410 411

Coase (1992), p. 714. Demsetz (1983), p. 377. See Richter and Furubotn (1999), p. 13. Cf. Gravelle and Rees (1984), pp. 1-14.

92

Insider Ownership and Performance

just the quantity of the goods bought or sold, as the price is set exogenously by the market.412 Under the neoclassical paradigm, the existence of firms — alternatively all transactions could be carried out on markets — is explained by three closely related economies: First, economies of specialization exist if the average output per employee in a team rises compared to the situation when each employee produces the same good individually. Therefore, the production process is broken down into various steps and each employee specializes in one of those tasks.413 Second, firms in contrast to individual production allow a more extensive use of capital goods, e.g. machinery in production, as individuals usually dispose of less savings and are not able to raise (limited liability) capital as companies can do. Third, an outcome of the first two arguments, economies of scale legitimate the existence of firms. Economies of scale refer to the more general fact that the average cost per unit falls along with larger volumes of output. These basic economic explanations for the existence of firms are elaborated mathematically in neoclassical theory.414 Thereby, firms’ decisions are orientated at marginal conditions with respect to inputs and outputs but no special attention is paid to the behavior of those employed with the firm. It is assumed that the owner (i.e. “risk-bearer”) and the manager (i.e. “risk-taker”) have the same utility functions which only appears plausible if both are indeed the same person, i.e. a single owner-manager.415 412 413

414 415

See Wolff (2000), p. 11. The most famous example is the pin making factory described in the first book of Smith ([1994] 1776)’s An Inquiry into the Nature and Causes of the Wealth of Nations (I. Division of Labour): “To take an example, therefore, from a very trifling manufacture; but one in which the division of labour has been very often taken notice of, the trade of the pin-maker; a workman not educated to this business (which the division of labour has rendered a distinct trade), nor acquainted with the use of the machinery employed in it (to the invention of which the same division of labour has probably given occasion), could scarce, perhaps, with his utmost industry, make one pin in a day, and certainly could not make twenty. But in the way in which this business is now carried on, not only the whole work is a peculiar trade, but it is divided into a number of branches, of which the greater part are likewise peculiar trades. One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on, is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct operations. [. . . ] [T]en persons, therefore, could make among them upwards of forty-eight thousand pins in a day. Each person, therefore, making a tenth part of forty-eight thousand pins, might be considered as making four thousand eight hundred pins in a day. But if they had all wrought separately and independently, and without any of them having been educated to this peculiar business, they certainly could not each of them have made twenty, perhaps not one pin in a day [. . . ]”. See Ulen (1993), p. 305. See Short (1994), p. 203.

Insider Ownership and Performance

93

However, the “[. . . ] curious disregard for what precisely it is that goes on within the firm [. . . ]”416 is in strict conflict with the science of business administration which by nature deals with problems on the business level. In this context, neoclassical models do not appear to be suitable for analysis as they immunize economic analysis against business problems.417 For example, shareholdings — the subject under investigation in this study — under the neoclassical view, simply represent a row of cash flows while institutional characteristics, e.g. voting rights, are largely neglected. The investor simply receives a proportional share of the surplus of cash inflows over cash outflows after all third party claims have been satisfied.418 As discussed in chapter 3, this pure cash flow perspective does not seem to reflect reality if the variety of auxiliary interests of the diverse shareholder groups is taken into account. As insider shareholders are not different from any other type of shareholder under the neoclassical view, a so-called institutional view is applied in the context of this study. 4.1.2

The Institutional View

4.1.2.1 Introductory Remarks The main proposition of the institutional view, also often referred to as new institutional economics419 , is that institutions matter in the economic process. This is a clear deviation from the neoclassical view which regards institutions as largely neutral elements.420 Generally, institutions could be defined as a system of formal and informal rules including their corresponding enforcement provisions.421 Examples of such institutions are the state, markets or corporations. The institutional view does not completely turn away from the equilibrium orientated, conceptional microeconomic analysis but rather tries to enlarge its scope of application by relinquishing (partially) its strict premises.422 The resulting institutional perspective can be characterized by the following five terms and hypotheses respectively: methodological individualism, utility maximizing 416 417 418 419

420 421 422

Ulen (1993), pp. 306-307. See Wolff (2000), p. 12. See Wolff (2000), p. 11. The origin of the new institutional economics goes back to the article The Nature of the Firm by Coase (1937). Cf. Coase (1998). See Richter and Furubotn (1999), p. 1. See Schmoller (1900), p. 61. See Sch¨uller (1983), p. 148.

94

Insider Ownership and Performance

individuals, individual rationality, imperfect markets and opportunistic behavior.423 The first aspect, methodological individualism, stresses the diversity of individuals and their utility preferences as ultimately people are the ones who make institutional decisions. As a consequence, corporations should not be seen as collectives and therefore have to be explained by theories focusing on the preferences of individuals and the interaction amongst them. Therefore, “[. . . ] the behavior of managers becomes the key for understanding the allocation and use of resources by corporations.”424 Unlike in neoclassical theory, utility maximizing individuals are the basic driving force and not corporations’ paradigm of profit maximization.425 Both objectives do not necessarily have to be identical as individuals will incorporate their own interests by making decisions within firms’ organizational structures. Individual rationality refers to the assumption that all decision makers have consistent and stable preferences. While earlier workings often assumed perfect individual rationality, which is similar to the neoclassical assumption, the view of bounded individual rationality becomes important. In doing so, the preferences of individuals are supposed to be incomplete and changing over time. Furthermore, the assumption of perfect markets is abolished in favor for imperfect markets, where the access to and the processing of information is limited and costly. As a result, information will not be distributed symmetrically. Finally, individuals are characterized by a certain degree of opportunistic behavior, i.e. they are pursuing their self-interest. For example, individuals might conceal their preferences or corrupt data to intentionally cause confusion.426 As corporations, or the existence of shareholdings, can also be seen as an institutional characteristic, shareholders’ preferences and interests are emphasized under the institutional view. Taking into account market participants’ different information levels, contractual relationships between shareholders, man-

423

424 425 426

See Richter and Furubotn (1999), pp. 3-5. They also list economic society, supervisory and enforcement system, institutions and organizations as further aspects. Cf. Richter and Furubotn (1999), pp. 5-9; Williamson (1975), p. 7; Coase (1998). Furubotn and Pejovich (1972), p. 1147. See Furubotn and Pejovich (1972), p. 1138. See Richter and Furubotn (1999), pp. 3-5.

Insider Ownership and Performance

95

agement and the corporation are examined.427 Information level differences and contractual rights might give rise to conflicting incentives which could be detrimental to the functioning of equity capital markets.428 Varying approaches within the institutional view complicate a general and well structured presentation of the key propositions.429 Therefore, in the following sections, the transaction cost, property rights and principal agency theories will be presented as three prominent doctrines within the institutional paradigm which are most relevant to insider ownership analysis.430 4.1.2.2 Transaction Cost Theory A special characteristic of the institutional view is the abolishment of the perfect markets assumption and thereby the explicit emphasis on transaction costs.431 The origin of transaction costs goes back to C OASE (1937) who examines the nature of firms in general and concludes that “[t]he main reason why it is profitable to establish a firm would seem to be that there is a cost of using the price mechanism [i.e. the market].”432 Following the argument of C OASE (1960), transaction costs could be classified into search costs, information costs, negotiation and contracting costs and inspection, enforcement and control costs. Keeping in mind these various costs, one might be tempted to ask why a significant share of property rights is coordinated via market mechanisms anyhow? Apparently, numerous institutions help to lower these transaction costs: For example, stock exchanges, consultancy services and standardized means of payment lower the search and information costs. Similarly, various corporate forms with different asignments of property rights and different types of contracts aim at reducing negotiation and contracting costs. Finally, securities, insurance contracts, future markets, guarantees, trademarks and contract penal427

428 429 430

431 432

The acquisition of a share finally also represents a contractual relationship as the buyer derives certain rights (e.g. voting rights, dividends and the right to resale) from the buying of shares. See Wolff (2000), pp. 15-16. See Wolff (2000), p. 14. Cf. Nathusius (2005), pp. 44-47. Other institutional theories include e.g. the incomplete contract theory, public choice theory and new organizational economics. See Becht (1997), pp. 16-17; Richter and Furubotn (1999), pp. 35-38. An example of another non-institutional approach would be the resource based view which e.g. explains family firm performance by a unique bundle of resources and capabilities (“familiness”). See Habbershon, Williams and MacMillan (2003); Chrisman, Chua and Litz (2003). See Furubotn and Pejovich (1972), p. 45. Coase (1937), p. 390.

96

Insider Ownership and Performance

ties help to lower the control costs and to assure the recoverability of property rights.433 Alternatively, transaction costs could be subdivided into two groups: costs of coordination and costs of motivation. While the former especially refer to the costs of gathering and processing information as well as the contracting costs, the latter indicate a lack of incentive for one party to enter a transaction in case it cannot completely oversee the transaction.434 Transaction costs exist because market participants do not share homogenous expectations in reality as the degree of uncertainty regarding future events is very high and information is asymmetrically distributed.435 However, these costs do not only occur in market transactions but also within a hierarchic organizational form, e.g. the corporation where contracts are not eliminated but usually greatly reduced.436 Therefore, transaction cost theory would deal with analyzing which institutional settings allow transactions to take place at minimal costs and thereby simultaneously enable allocative efficiency.437 Even though the concept of transaction costs is hardly directly transferrable to questions regarding the ownership structure of public stock corporations, it is mentioned here for at least two reasons: First, it constitutes one of the crucial and most known concepts within the institutional view which must not be neglected for the sake of completeness. Second, transaction costs accrue when property rights, the basic element of the next presented institutional approach, are transferred.438 4.1.2.3 Property Rights Theory The property rights theory as well as the principal agent theory can be classified into the group of incentive based theories. The property rights theory cannot be seen as a classical theory because it lacks a cohesive system of premises and stringent formalization. Rather, it can be seen as a contract theoretical view, which focuses on the economic reasoning of institutions and the effects of legal 433 434 435 436

437 438

See Sch¨uller (1983), pp. 158-160. Cf. Coase (1960), p. 15. See Wolf (1999), p. 9. See Wolff (2000), p. 19. See Coase (1937), p. 391. Cf. Hart and Moore (1990) who analyze in which cases transactions should be carried out within the firm or through the market. See Williamson (1975), p. 26; Sch¨uller (1983), p. 153. Cf. Kr¨usselberg (1983), pp. 45-78. See Sch¨uller (1983), p. 151.

Insider Ownership and Performance

97

regulations.439 Under the property rights approach, the neoclassical black box perspective of the firm is altered towards a nexus of contracts view. According to this, FAMA (1980) stresses that “[. . . ] ownership of capital should not be confused with ownership of the firm. Each factor in a firm is owned by somebody. The firm is just the set of contracts covering the way inputs are joined to create outputs and the way receipts from outputs are shared among inputs.”440 In contrast to neoclassical theory, the property rights approach does not refer to simple goods — or the production or consumption of them — but to individuals’ property rights on material and intangible goods.441 Specifically, property rights refer to a bundle of rights, which consists of the individual rights to i) use the resource, ii) change the resource, iii) be the residual claimant on the resource and iv) sell off these rights to a third party.442 For example with respect to firms, these rights would equal the right to use the corporations’ assets to produce goods, the right to invest (i.e. to change financial assets into physical or intangible assets) or employ laborers in exchange for wage, the right to participate in the corporations’ profits (e.g. to receive dividends) and the right to sell off assets or business segments to other market participants. While in the solely owner-managed firm of J ENSEN AND M ECKLING (1976) one single individual owns the whole bundle of rights, in the diffuse ownership corporation the use rights (i.e. i) and ii)) and the benefits rights (i.e. iii) and iv)) have been separated.443 Therefore, “[s]hareholders may be viewed as parties that have explicitly contracted for a relatively well-defined subset of rights as prescribed under the corporate charter.”444 Alternatively, it could be argued that shareholders transfer the use rights of their contributed capital via employment contracts

439

440 441 442 443 444

See Wolff (2000), pp. 8, 18. Furubotn and Pejovich (1972), p. 1137, surveying property rights literature state that “[t]he contributions here are quite diverse in style and content but are characterized by a common emphasis on certain basic ideas concerning the interconnectedness of ownership rights, incentives, and economic behavior.” Fama (1980), p. 290. See Wolff (2000), p. 17. See Alchian and Demsetz (1972), p. 783; Richter and Furubotn (1999), p. 82. See Kang and Sorensen (1999), p. 125. Kang and Sorensen (1999), p. 126.

98

Insider Ownership and Performance

to the management in order to realize specialization gains which could not be achieved if both parties acted separately.445 Thus, the “[. . . ] key idea in the analysis is that different property rights assignments lead to different penalty-reward structures and, hence, decide the choices that are open to decision makers.”446 In the diffuse ownership corporation, one central issue is (minority) shareholders’ inability to capture many of the benefits rights while owning only limited use rights which are largely delegated to outside managers. The delegation of rights is costly as it provides managers with discretionary scope and the transaction costs of (complete) information and control of the manager’s behavior might be too large.447 As ownership in corporations with limited liability shields investors’ personal wealth, the legal power of shareholders is generally limited in scope.448 Furthermore, the level of transaction costs will determine the degree to which property rights are likely to be exercised. In contrast to a diffuse ownership corporation, in a (partially) ownermanaged firm insider ownership can be interpreted as a specific assignment of a company’s property rights where the four separate types of property rights become (partially) reunified in the hands of one party, i.e. the owner-manager. If one argues — as claimed by F URUBOTN AND P EJOVICH (1972) — that it is “[. . . ] the crucial task for the new property rights approach [. . . ] to show that the content of property rights affects the allocation and use of resources in specific and predictable ways [. . . ]”449 the analysis of the relationship between insider ownership and performance could be seen as such an attempt. 4.1.2.4 Principal Agent Theory The principal agent theory and the property rights approach are highly complementary even though both concepts developed independently from each other.450 Indeed, the principal agent theory could be seen as a finer specification of the transactions cost or property rights approaches as elements of both theories are incorporated. While both concepts examine the distribution of property rights and the existence of (institutional) transaction costs in individual con445 446 447 448 449 450

See Iber (1987), p. 35. Furubotn and Pejovich (1972), p. 1138. See Iber (1987), p. 35. See Kang and Sorensen (1999), p. 127. Furubotn and Pejovich (1972), p. 1139. See Jensen and Meckling (1976), p. 308. Cf. Reb´erioux (2003), p. 68.

Insider Ownership and Performance

99

tracting settings respectively, the principal agent theory focuses on the special problems arising from agency.451 The Economic Theory of Agency was first developed by ROSS (1973) who defines agency settings rather universally: “[. . . ] an agency relationship has arisen between two (or more) parties when one, designated as the agent, acts for, on behalf of, or as representative for the other, designated the principal, in a particular domain of decision problems.”452 More specifically, contractual arrangements between employers and employees as well as between a state and its citizens are listed as examples for such typical agency relationships.453 According to the property rights perspective, the principal transfers certain property rights, i.e. the use rights, to an agent who is contractually mandated to pursue the principal’s objectives. Thereby, the asymmetric information distribution between principal and agent constitutes one key principal agent theory underlying assumption:454 Adverse selection refers to asymmetric information before contract conclusion which motivates the party with superior knowledge to engage in opportunistic behavior before the contract is closed;455 moral hazard relates to asymmetric information afterwards, i.e. the better informed party is motivated to behave opportunistically after an agreement has been concluded. The problem of moral hazard can be further differentiated into problems arising from hidden information and hidden action. The former implies that the agent may intentionally withhold information from the principal to obtain a personal advantage, e.g. by concealing his own (bad) performance. The latter refers to the fact that the principal can usually only observe the results and not the individual activities of the agent. As the final outcome is usually driven by both the agent’s actions and at least one exogenous random factor (i.e. uncertainty), the principal has difficulties to attribute good or bad performance to either one of these two aspects.456 451 452

453

454 455 456

See Wolff (2000), p. 20. Ross (1973), p. 134. As stated in section 2.2.2, Smith ([1994] 1776), p. 800, already indirectly pointed to the same problem. Kang and Sorensen (1999), p. 128, list further examples: clients and lawyers, political party members and party leaders, rulers and state officials. Cf. Leland and Pyle (1977). The most prominent example in this context is Akerlof (1970)’s seminal analysis of the market for used cars. See Furubotn and Pejovich (1972), p. 196.

100

Insider Ownership and Performance

Generally, the principal has two options to influence the behavior of the agent: incentives and control.457 For example, the principal might try to make the agent act in his best interest by including accordant provisions in the contract which align the preferences of both parties. Alternatively, the principal might acquire costly, additional information about the agent’s actions and thereby limit the agent’s discretionary scope. As a consequence of the agency setting, so-called agency costs arise which consist in three elements: incentive or bonding costs, monitoring costs and the residual loss. Incentive costs are born by the principal to make the agent act in his interest while in some situations “[. . . ] it will pay the agent to expend resources (bonding costs) to guarantee that he will not take certain actions which would harm the principal or to ensure that the principal will be compensated if he does take such actions.”458 Monitoring costs refer to costs incurred by the principal to acquire and process information about the agent’s activities or to enforce his contractual rights in case of malpractice by the agent. Finally, the incentives and bonding as well as monitoring efforts determine the divergence between the agent’s decisions and those maximizing the principal’s welfare. The resulting difference in the principal’s welfare is referred to as the residual loss.459 Finally, agency-theory examines how total agency costs can be minimized through contractual arrangements.460 After having generally analyzed the efficiency of such agency settings, ROSS (1973) — without any emphasis — claims that: “[. . . ] in a world of true uncertainty where adequate contingent markets do not exist, the manager of the firm is essentially an agent of the shareholders.”461 Three years later, this observation is formally analyzed by J ENSEN AND M ECK LING (1976) in their seminal work Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. They relate the concept of agency costs to the issue of separation of ownership and control which was raised some forty

457 458 459 460 461

Cf. section 3.4.3. Jensen and Meckling (1976), p. 308 See Jensen and Meckling (1976), p. 308. See Wolff (2000), p. 22. Cf. Barnea, Haugen and Senbet (1985), pp. 1-3. Ross (1973), p. 138.

Insider Ownership and Performance

101

years earlier by B ERLE AND M EANS (1932).462 As discussed in section 3.1, their analysis departs from the case of an entirely owner-managed firm (i.e. a firm where the manager owns 100% of the residual claims) and shows how socalled agency costs of outside equity arise when a portion of the residual claims is sold to outsiders. As the partial owner-manager will only have to bear a disproportionate share of the costs, while enjoying all non pecuniary benefits, a suboptimal mix of pecuniary and non pecuniary benefits will be consumed.463 As in diffuse ownership corporations464 virtually all claims have been sold to the public, the consumption of corporate resources in the form of perquisites by the management might be especially severe. Consequently, a higher degree of monitoring becomes desirable for all shareholders. Nevertheless, the more serious problem might be that: “[. . . ] as the manager’s ownership claim falls, his incentive to devote significant effort to creative activities such as searching out new profitable ventures falls. He may in fact avoid such ventures simply because it requires too much trouble or effort on his part to manage or to learn about new technologies. Avoidance of these personal costs and the anxieties that go with them also represent a source of on the job utility to him and it can result in the value of the firm being substantially lower than it otherwise could be.”465 Therefore, in corporations with significant levels of insider ownership, agency problems may generally play a comparably smaller role which in turn might have a positive effect on firm value.466 Thus, the key question of this study is

462

463

464

465

466

See section 3.1. Another important work about agency theory is Fama (1980)’s seminal article Agency Problems and the Theory of the Firm. See Jensen and Meckling (1976), p. 312. Cf. Demsetz (1983), pp. 380-383. In contrast, the stewardship theory by Davis, Schoorman and Donaldson (1997), pp. 25-26, claims that generally “[. . . ] the behaviors of the executive are aligned with the interests of the principals.” Alternatively, instead of diffuse ownership one could also speak of fragmented property rights. See Kang and Sorensen (1999), p. 128. Jensen and Meckling (1976), p. 313. Note that agency problems do not only exist between shareholders and management in public stock corporations but also in a large variety of other organizations as partnerships, financial mutuals or non profit organizations. See Fama and Jensen (1983b), pp. 311-321. For the case of venture capital companies and portfolio companies cf. Fingerle (2005), p. 104. See section 5.3.

102

Insider Ownership and Performance

how agency costs are related to insider ownership in the modern public stock corporation.467 4.2

Empirical Evidence on Insider Ownership and Performance

Having presented a theoretical top-down view of insider ownership in the preceding section, an overview of the most relevant literature with respect to insider ownership — and especially the relationship between insider ownership and performance — will now be presented. As the amount of relevant literature is enormous, only selected works closely related to the empirical study carried out in chapter 6 will be cited. Primarily, the endogeneity issue involved in empirical studies analyzing the ownership-performance relationship will be reviewed (section 4.2.1). Secondly, a comprehensive selection of studies will be presented in a summary table (section 4.2.2). Third, the most relevant and famous studies will be explicitly described according to the corresponding corporate governance systems (section 4.2.3).468 Fourth, the shortcomings of research regarding insider ownership in Germany will be summarized (section 4.2.4). 4.2.1

Determinants and Effects of Insider Ownership

The empirical observations by B ERLE AND M EANS (1932), which imply a positive relationship between ownership concentration and performance, as well as the principal agent analysis by J ENSEN AND M ECKLING (1976), which posits a positive linear impact of insider ownership on performance, remained largely unquestioned for a long time. The seminal works of D EMSETZ (1983) and D EMSETZ AND L EHN (1985) are among the first to cast doubt on these predictions. The former generally views “[. . . ] the ownership structure of the firm as an endogenous outcome of a maximizing process [. . . ]”469 while the latter argue that ownership structures are optimally chosen with respect to value maximization and find empirical evidence that the large differences in owner-

467

468

469

Note that the remarks primarily apply to the so-called agency costs of equity. Besides, so-called agency costs of debt exist. See Nassauer (2000), pp. 12-17. See section 2.3. Cf. La Porta, Lopez-de Silanes, Shleifer and Vishny (2002), p. 1161, for the classification of various countries into civil law and common law countries. Demsetz (1983), p. 377.

Insider Ownership and Performance

103

ship structures are explained by factors as firm size, control potential (proxied by the instability of the profit rate) and regulation.470 The value maximizing firm size varies within and among industries. The larger the value maximizing firm size, the higher the value of a given fraction of ownership. Consequently, higher values will lead to a lower degree of ownership concentration. Moreover, a lower fraction of voting rights is needed to gain effective control the larger the firm is. Consequently, both effects imply that the larger a firm the more diffused its ownership structure (“risk-neutral effects of size on ownership”). With respect to control potential the following considerations play a decisive role: The markets of corporate control and managerial labor align management’s interests with those of the shareholders. However, since the use of both markets is costly (i.e. transaction costs exist), variations in ownership structures can be seen as responses to these costs. Firm-specific uncertainty is one factor influencing the control potential because the monitoring of those firms requires higher information costs. Thus, instability can be expected to be associated with higher levels of ownership concentration. Finally, systematic regulation limits owners’ discretion over the use of their property rights and hence reduces the control potential. Furthermore, a subsidized monitoring will be performed by the regulators. Both effects are likely to decrease the degree of ownership concentration.471 For a sample of 511 large U.S. corporations D EMSETZ AND L EHN (1985) measure ownership concentration by the percentage of equity owned by the five largest shareholders, the 20 largest shareholders as well as an approximated Herfindahl index.472 In their regression analysis they find evidence for the hypothesized determinants of ownership concentration. The results are robust to the use of the three different measures of ownership concentration as dependent variables and the classification of owners into different groups (i.e. all investors, family and individuals as well as institutional investors).473 Unlike B ERLE AND M EANS (1932), D EMSETZ AND L EHN (1985) do not expect any positive correlation between ownership concentration and perfor470 471 472 473

See Demsetz and Lehn (1985), pp. 1155-1158. See Demsetz and Lehn (1985), pp. 1156-1161. This issue will be dealt with later on. See footnote 631. See Demsetz and Lehn (1985), pp. 1162-1173.

104

Insider Ownership and Performance

mance since higher costs of loosening control over management might be offset by lower capital costs derived from a diffuse shareholder structure. Using five year averages of accounting profit rates as dependent variables, they do not find evidence for a positive relationship between ownership concentration and performance.474 Even though their study exclusively applies to ownership concentration, one could easily transfer the main considerations to the case of owner identity or more specifically insider ownership. Without evaluating the arguments in favor of or against the endogeneity of ownership structures — this will be done later on in section 6.9.4.2 — it should be noted that some of the empirical studies presented in the following qualitatively or quantitatively address this issue while others, and especially the early studies, do not. These differences may help to explain the partially large differences amongst empirical findings even within an individual examined country (or corporate governance system). Thus, the interpretation and comparison of these results demands special caution. 4.2.2

Overview of Empirical Studies on Insider Ownership

The overview of empirical studies will be separated into two parts: First, primarily descriptive studies referring to the German market (or several markets including the German market) will be presented (section 4.2.2.1). It should be noted that these studies often do not explicitly address insiders as an owner type; however, it can be reasonably assumed that the majority of the frequently used owner category individuals/ families refers to managerial ownership as defined in this study.475 Second, studies investigating the insider ownershipperformance relationship will be listed (section 4.2.2.2). Thereby, selected studies focusing on family ownership, which usually represents a special case of insider ownership, are included as well.476 Finally, it should be noted that 474 475

476

See Demsetz and Lehn (1985), pp. 1173-1176. The descriptive results of the empirical study in chapter 6 show that e.g. for the 2003 sub-sample an average of 32.0% of the control rights can be attributed to insider individuals while only 0.9% can be attributed to outsider individuals. Thus, insiders represent 97.3% of all shareholders classified as insider and outsider individuals. See figures 6-2 and 6-7. Of course, many more German and international studies examine the relationship between family ownership and performance or related topics. Recent examples amongst others are: McConaughy, Walker, Henderson and Mishra (1998); McConaughy, Matthews and Fialko (2001); Anderson and Reeb (2003); Schiereck and Jaskiewicz (2004); Ehrhardt, Nowak and Weber (2006); Sraer and Thesmar (2004); Kuklinski, Lowinski and Schiereck (2003); Barth, Gulbrandsen and Schøne (2005); Hillier and McColgan (2005); Jaskiewicz,

Insider Ownership and Performance

105

the overview cannot be exhaustive due to the enormous attention ownershipperformance relationships have attracted in research. As a consequence, the selection of the most relevant studies will be subject to the author’s discretion.

Gonzal´ez, Men´endez and Schiereck (2005); Corstjens, Peyer and Van der Heyden (2005); Maury (2006); Ben-Amar and Andr´e (2006).

Descriptive Studies on the German Market

Sample(s)

Germany (1996) First ever official cross-section of voting blocks disclosed to BAWe/BaFin; 430 firms

Germany (1997/ 1999) 418/ 422 listed corporations respectively

Germany (1992) 105 listed and 51 unlisted firms of the 200 largest (measured by turnover) non-financial firms

Author(s) (Year) Journal / Book Title

Becht and B¨ohmer (2003) International Review of Law and Economics

Bott (2002) Aktion¨arsstruktur, Kontrolle und Erfolg von Unternehmen

Edwards and Nibler (2000) Economic Policy

Exercised voting rights on shareholders’ general meetings (control rights)

Database major holdings of voting rights in officially listed companies (BAWe/BaFin)

Database major holdings of voting rights in officially listed companies (BAWe/ BaFin)

Source(s) of Ownership Data (Germany)

Concentration as sum of the n-largest voting blocks with n = 1 and 2

Concentration as sum of the n-largest voting blocks with n = 1, 2 and 3

Concentration as sum of the n-largest voting blocks with n = 1, 3, 5 and all known voting blocks

Owner Classes (Concentration)

7 classes: individuals/ families, non-bank firms, three big banks, other banks, state, foreigners, others

11 classes: individuals, foreign state, banks, non-financial firms, associations/ GbRs/ workers/ family pools, state, holding companies, investment firms, bank related investment firms, foundations, insurance companies 6 classes: individuals, non-financial firms, financial firms, state, foreigners, holding companies

Owner Classes (Identity)

Table 4-1: Empirical Studies on German Shareholder Structures

4.2.2.1

All votes exercised at shareholders’ general meetings

Direct and indirect stakes > 5%

Direct and indirect stakes > 5%

Disclosure Requirements (Germany)

Individuals/ families as second largest owner of voting equity (only listed firms): 13% (mean ownership stake: 20%)

Individuals/familes as largest owner of voting equity (only listed firms): 40% (mean ownership stake: 57%)

Percentage of individuals as largest shareholder: 18.0%

Percentage of direct shareholdings by individuals: 33.1%

Mean size of voting blocks owned by individuals: 27.2%

Voting blocks owned by individuals in percent of total voting blocks: 27.4%

Main Result(s) (Germany)

continued on next page . . .

Ownership is highly concentrated but this must not have a positive effect on performance; control activities are likely to differ among and within different shareholder identity groups; the analysis of announced changes in shareholders’ identity largely does not yield significant results Focus on bank ownership and concentration; ownership concentration is important but banks do not play a role in the governance of large German firms while other types of large shareholders do so

Business groups — involving less regulated legal forms as intermediate layer — can substantially reduce transparency of German shareholder structures; banks that are exempted to disclose proxy voting rights constitute another disclosure gap; German capital market is dominated by a few firms and concentration of control is very high

Study Focus and Conclusion(s) (Germany)

106 Insider Ownership and Performance

Germany (1963/ 1973/ 1983) n/a

Iber (1987) Entwicklung der Aktion¨arsstruktur b¨orsennotierter deutscher Aktiengesellschaften

Franks and Mayer (2001) The Review of Financial Studies

Germany and 12 other Western European countries (1996-1999) 704 listed and unlisted German firms (total: 5,232) Germany (1990) 171 listed industrial and commercial companies

Faccio and Lang (2002) Journal of Financial Economics

. . . continued from previous page Van der Elst (2000) Germany, BelFinancial Law Ingium, France, stitute - Working Spain, Italy, Paper and U.K. (1990/ 1999) 171/ 501 companies respectively (for Germany)

Statistische Bundesamt; Depot statistics of Deutsche Bundesbank

1990 data: Franks and Mayer (1997) 1999 data: Hoppenstedt Aktienf¨uhrer (2000); Database major holdings of voting rights in officially listed companies (BAWe/BaFin) Commerzbank “Wer geh¨ort zu wem?” (1997); FT “Extel Financial Company Resarch CD-Rom” (1997); Worldscope (1998) Hoppenstedt Aktienf¨uhrer (1990)

Concentration as sum of the n-largest voting blocks with n = 1 and 4; majority shareholders (> 50%)

Closely held (at least one shareholder owns > 25%) and widely held firms

Percentage of firms controlled by different controlling owners at the 20% threshold

Size of largest shareholding; major shareholders; majority control (size and identity)

9 classes: individuals/ families, domestic firms, foreign firms, insurance companies, trusts/ institutional investors, banks, state, other domestic public authorities, others 6 classes: private households, foreigners, state, banks, insurance companies, corporations

7 classes: family, state, widely held corporation, widely held financial, miscellaneous, cross-holdings, widely held

9 classes: individuals, non-financial firms, state, (foundations), banks, insurance companies, pension funds, investment funds, foreigners

Only shares deposited by banks (roughly 60% of total share capital); see Portfolio Management (1988)

Direct and indirect stakes > 25%

Direct and indirect stakes > 5%

Direct and indirect stakes (> 25% for 1990; > 5% for 1999)

Identity of majority shareholders is private household: 47.8% (1963), 37.5% (1973) and 34.5% (1983)

Percentage of shares ownership (at market values) by private households: 26.0% (1963), 24.3% (1973) and 16.9% (1983)

Portion of companies with a familygroup shareholder in excess of: i) 25%: 20.5% ii) 50%: 16.4% iii) 75%: 5.3%

Portion of companies which are controlled by families: 64.6% (2nd highest after France) Portion of companies which are widely held (no controlling owners): 10.4% (lowest ratio of all countries)

Percentage of companies which is majority controlled in 1999: 49% (thereof: 22% by individuals/ families)

Percentage of companies where individuals own stakes > 25%: 21% (1990) and 40% (1999)

continued on next page . . .

Focus on board turnover and performance; high level of concentration; little evidence that principals with large share stakes, such as families, have greater incentives to monitor poorly performing management than agents, such as banks or insurance companies Concentration raised significantly from 1963 to 1983 but slowed down afterwards; private households’ importance declined while corporations extended their power; the position of minority shareholders weakened

Focus on cross-country comparison of ownership structures in 13 Western European countries; dual share classes and pyramid structures are used to enhance the control of the largest shareholder

In Continental European countries the blockholder system with powerful controlling families still exists; not only ownership concentration but also the nature of the shareholding concentration (= shareholder identy) should be taken into account

Insider Ownership and Performance 107

Germany and 26 other countries (1995) Two subsamples: i) top 20 firms by market capitalization ii) smallest 10 firms with market capitalization > USD 500m Germany (1986) 61.4% of domestic issued share capital (i.e. the portion deposited by banks and held directly by banks)

La Porta, Lopezde Silanes and Shleifer (1999) Journal of Finance

Portfolio Management (1988) Aktion¨arsstruktur und Anteilsbesitz in der Bundesrepublik

Germany (19871994) 946 listed and unlisted corporations (4,433 firm years)

K¨oke (2004) Journal of Corporate Finance

. . . continued from previous page K¨oke (1999) Germany (1994Working Paper 1998) 1,207 observations during the five year period on listed AGs

Authors’ survey; Depot Statistics of Deutsche Bundesbank

Hoppenstedt Aktienf¨uhrer (1997)

Annual reports published by Bayerische Hypothekenund Wechselbank

Hoppenstedt Konzernstrukturdatenbank (KSD)

n/a

n/a

Concentration as sum of the n-largest voting blocks with n = 1, 2 and 3; Herfindahl index; Cubbin and Leech (1983) index Concentration as sum of the n-largest voting blocks with n = 1; Herfindahl index

8 classes: individuals, non-profit organizations, investment funds, insurance companies, non-financial firms, banks, state, foreigners

6 classes: families, state, widely held nonfinancial firm, widely held financial firm, dispersed shareholdings, others

6 main classes: individuals (and thereof: insiders), non-financial firm, financial firms, state, dispersed shareholdings

6 classes: individual, non-financial firms, financial firms, state, foreigners, dispersed shareholdings

Direct and indirect stakes > 25%

Direct and indirect stakes > 5%

Percentage of shareholdings by insiders for listed firms: 22.5%

Direct stakes > 5%

Percentage ownership by individuals (readjusted for free-float portion): 26.6%

Mean ownership stake of families in medium-sized listed firms: 40%

Mean ownership stake of families in large listed firms: 10%

Mean ownership stake by insiders for non-listed firms: 12.0%

Percentage of shares owned by individuals (1998, only listed AGs): 10.8% Percentage of individuals as largest shareholder (1998, only listed AGs): 10.6%

Direct and indirect stakes > 5%

Infrastructure of the German financial market is underdeveloped compared to other developed nations but improvements are on their way; main problem might be the backward oriented regulatory policy (e.g. discrimination of stock corporation in tax law) continued on next page . . .

Focus on causes and consequences of control changes; insiders try to hinder changes in control; tight shareholder control (i.e. a high degree of ownership concentration) acts as a substitute for disciplinary control changes; firms not experiencing a change of control tend to show higher insider ownership levels Except in economies with very good shareholder protection, relatively few large and medium-sized firms are widely held; the controlling shareholders typically have power significantly in excess of their cash flow rights

Study of all legal forms of corporations, not only listed ones; ownership is not constant as commonly assumed; ownership is measured also at higher levels but pyramid structures are relatively rare; crossownership is of minor relevance

108 Insider Ownership and Performance

Database major holdings of voting rights in officially listed companies (BAWe/BaFin)

Concentration as sum of the n-largest voting blocks with n = 1; Herfindahl index

Oriented at ECGN owner classes: individuals/ families, financial sector (big three and foreign banks, other domestic banks, insurance companies, other financial firms), non-financial firms, state, others

Direct and indirect stakes > 5%

Source: Author’s demonstration. Owner identity classes have been (partially) standardized by the author.

. . . continued from previous page W´ojcik (2003) Germany (1997Environment and 2001) Planning A All companies listed on the official market (“Amtlicher Handel”) (415-463) Blockholdings by individuals and families in listed non-financial firms: 38.8% (42.2%) of all blockholdings in 1997 (2001)

Strengthening position of individuals and families and retreat of financial firms (especially banks) between 1997 and 2001; network of German crossholdings exhibits persistence but change is on its way; acceleration of this process can be expected

Insider Ownership and Performance 109

Studies on Insider Ownership and Performance

U.S. (1991) 326 firms of the Fortune 500 manufacturing firms

Cho (1998) Journal of Financial Economics

Percentage of shares owned by officers and directors (Proxy statements, Value Line Investment Survey)

Percentage of shares owned by officers and directors (Institutional Shareholder Services)

U.S. (1990) Cross section of 369 non-financial, non-regulated S&P 500 firms

Barnhart, Marr and Rosenstein (1994) Managerial and Decision Economics

Insider Ownership Measure(s) (Source)

Percentage of shares owned by officers and directors (Disclosure CD-Rom)

Sample(s)

I. Studies on the U.S. and U.K. markets Agrawal and KnoeU.S. (1987) ber (1996) Cross section of Journal of Finan383 Forbes 800 cial and Quantitafirms tive Analysis

Author(s) (Year) Journal

Tobin’s Q

Market-to-book ratio

Tobin’s Q

Performance Measure(s)

OLS-regression (piecewise linear); 2SLSregression (simultaneous equation system)

OLS-regression; instrumental variable (IV) regression

OLS-regression; 2SLS-regression (simultaneous equation system)

Method(s)

Corporate investments, R&D ratio, cash flow, business risk, size, market value of equity, industry dummies

Block and institutional ownership, outside board members, number of insiders, CEO characteristics, listing, leverage, acquisitions, business risk, R&D and advertising ratio, size, diversification, cash flow return, industry dummies Institutional ownership, outside board members, board size, R&D and advertisting ratio, size, leverage, industry dummies

Control Variables

Table 4-2: Empirical Studies on Insider Ownership and Performance

4.2.2.2

Focus on board composition; OLS-results indicate significant curvilinear relation between insider ownership and performance (as McConnell and Servaes (1990)); board composition and insider ownership may both be endogenous but IV-results are very sensitive to the choice of instruments Focus on ownership structure, investment and corporate value; OLS-results suggest increasing-decreasing-increasing impact of insider ownership on corporate investment; according to 2SLS-results investment affects corporate value which, in turn, affects ownership structure but not the reverse; therefore endogeneity significantly affects the results and must be regarded

Simultaneous analysis of the performance effects of seven control mechanisms; evidence of interdependence in the use of control mechanisms; OLS-regressions yield significant positive bell-shaped relationship between insider ownership and performance; in the simultaneous equation system the effects of insider ownership become insignificant

Main Result(s)

Ownership may not be an effective incentive mechanism to induce managers to make value-maximizing investment decisions; ownership structure is equilibrium output from a market process continued on next page . . .

Alignment of interest hypothesis; entrenchment hypothesis

Optimal use of control mechanisms; no relationship between insider ownership and performance can be observed

Explanation(s)

110 Insider Ownership and Performance

U.S. (1996, 1998) Cross section of 1,147 firm years of high R&D firms listed on the NYSE, AMEX and NASDAQ

U.K. (1995) Cross section of 802 non-financial firms

Davies, Hillier and McColgan (2005) Journal of Corporate Finance

U.S. (1983-1987) 94 thrift institutions converting from mutual to stock ownership U.S. (1993-2000) 8,576 firm years

Cui and Mak (2002) Journal of Corporate Finance

Coles, Lemmon and Meschke (2003) Working Paper

Cole and Mehran (1998) Journal of Financial Economics

. . . continued from previous page Chung and Pruitt U.S. (1986) (1996) 404 firms Journal of Banking & Finance

Percentage of shares owned by i) all officers and directors, ii) the CEO; ratio of total insider ownership to the number of these insiders (Proxy statements) Percentage of shares owned by company management (only stakes > 0.5%) (MacMillan London Stock Exchange Yearbook)

Percentage of shares owned by the CEO (Execucomp)

i) Percentage and ii) dollar value of shares owned by the CEO (The Directory of America’s Corporate Elite; published by Business Week in 1987) Largest percentage ownership stake by a single officer or director

Tobin’s Q

Tobin’s Q; return on assets

Tobin’s Q

Annual stock returns

Tobin’s Q

OLS-regression; 2SLS-regression

OLS-regression (piecewise linear); 2SLSregression

(Pooled-) OLSregression; 2SLS-regression (simultaneous equation system); fixed-effects panel data model

Difference in mean; OLSregression

OLS-regression; 3SLS-regression (simultaneous equation system)

Block ownership, R&D and capital expenditure ratio, leverage, market value of equity, size, adjusted profits, industry dummies

Size, leverage, R&D ratio, asset growth rate, fixed assets ratio, industry and time dummies

Institutional-, noninstitutional and ESOP ownership, pre-change stock returns, time dummies Size, leverage, R&D and advertising ratio, industry dummies

Size, CEO (tenure, age, founder, human capital), company age, industry dummies

Non-linear relationship between insider ownership and corporate value due to conflicting managerial incentives; quintic structure results in double-hump-curve; insider ownership and corporate value are codeterministic, i.e. performance influences corporate ownership and vice versa

Focus on changes of ownership structure from mutual to stock ownership in the thrift industry; changes in performance are positively associated with changes in ownership by managers OLS-results indicate hump-shaped positive relation between insider ownership and performance; fixed-effects and 2SLS-analysis yield insignificant results; in empirical settings, the endogeneity problem of ownership and performance is substantial and difficult to correct using IVs, fixed-effects and simultaneous equations; standard approaches to endogeneity typically fail to provide a solution Focus on high R&D firms; non-linear Wshaped relationship between insider ownership and firm performance; opposite results for use of return on assets as performance measure; robust to 2SLS-estimation; endogeneity of insider ownership is not an issue

CEO ownership and performance are highly positively correlated (in both directions); market value of equity, executive ownership and executive compensation are jointly determined

continued on next page . . .

Ownership structure is equilibrium output from a market process; alignment of interest hypothesis

Performance and insider ownership vary together endogenously, as their underlying determinants, marginal productivity of investment and effort, vary in the cross-section and through time Alignment of interest hypothesis; entrenchment hypothesis; industry specifics explain deviations from prior results

Firms optimally establish their managerial compensation plans in response to both its environment and the specific characteristics of its CEO; alignment of interest hypothesis Alignment of interest hypothesis

Insider Ownership and Performance 111

Percentage of i) cash flow and ii) control rights owned by officers and directors (Proxy statements; 10-Ks)

U.S. (1994-2001) 1,282 firm years of dual share class companies

U.S. (1992-1997) 7,134 firm years from 1,307 firms of the S&P Super Composite Index (unbalanced panel)

Gompers, Ishii and Metrick (2004) National Bureau of Economic Research - Working Paper

Habib and Ljungqvist (2003) Working Paper

Percentage of shares owned by the CEO (Execucomp)

Aggregated, average (1976-1980) shareholdings owned by i) top management, ii) the CEO and iii) all board members (Corporate Data Exchange) Percentage of shares owned by officers and directors (Financial statements; Extel Financial)

. . . continued from previous page Demsetz and VillaU.S. (1976-1981) longa (2001) Cross section of Journal of Corpo223 firms; ranrate Finance dom sub-sample from the original Demsetz and Lehn (1985) study Faccio and Lasfer U.K. (1996(1999) 1997) Working Paper 1,650 nonfinancial firms listed on the LSX

Tobin’s Q

Stochastic frontier models using maximum likelihood; OLS-regression

Median FamaMacBethand median pooled regression

OLS-regression

Tobin’s Q (industry adjusted); return on assets; return on equity

Tobin’s Q

OLS-regression; 2SLS-regression

Tobin’s Q (average 1976-1980); accounting profit rate (average 1976-1980)

CEO incentive variables (stock options, capital- and product market pressure, board size), size, R&D, capital expenditure and advertising ratio, leverage, cost of capital, industry growth, coverage by financial analysts

Block ownership, number of directors, outside board members, CEO/chairman split, non-executive chairman, size, leverage Size, index membership, age, industry and time dummies

Shareholder concentration, R&D, advertising and investment ratio, leverage, concentration, market risk, stock price volatility, size, industry dummies

The observed 16% shortfall of actual Tobin’s Q to a hypothetical value-maximizing Q is related to the incentives provided to CEOs: apparently too few shares and too many options are granted CEOs by the boards; endogeneity does not appear to be a problem; corporate boards adjust incentives dynamically

Focus on dual share class companies; relation of insider cash flow rights to firm value is positive and concave while the relation of insider control rights to firm value is negative and convex; robust to the use of operating performance measure (e.g. sales growth) as dependent variable; not robust to use of net profit margin and return on equity as performance measures

Managers become entrenched if they hold more than 12% of the shares in their companies by controlling the board; however, insider ownership does not create or destroy value, i.e. the relationship is weak or non-existent.

If ownership is modeled as an endogenous variable no evidence is found that certain ownership structures influence performance in any significant way; coefficients of single equation models of the effect of ownership structure on performance are biased

continued on next page . . .

Managers who are not residual claimants fail to maximize value; companies operate under suboptimal incentives; further analysis necessary to judge whether an equilibrium exists or not

Alignment of interest hypothesis; entrenchment hypothesis

Entrenchment hypothesis; effectiveness of U.K. external corporate governance system might effectively prevent entrenchment

Ownership structure is equilibrium output from a market process

112 Insider Ownership and Performance

U.S. (1971, 1974, 1977, 1980, 1983) 142 firms listed on the NYSE

U.S. (1982-1992) Random sample of listed Compustat firms; unbalanced panel declining from 398 (1982) to 293 (1992) firms

U.S. (1935, 1995) 1,500 listed firms in 1935 and a comparable sample of 4,200 firms in 1995

Hermalin and Weisbach (1991) Financial Management

Himmelberg, Hubbard and Palia (1999) Journal of Financial Economics

Holderness, Kroszner and Sheehan (1999) The Journal of Finance

. . . continued from previous page Halpern, KiU.S. (1981-1986) eschnick and 126 LBO and 126 Rotenberg (2005) non-LBO acquiThe Quarterly Resitions view of Economics and Finance

Percentage of shares owned by officers and directors (SEC Section 16 report (1935); Compact Disclosure (1995))

i) Total and ii) average percentage of shares owned by toplevel managers (Proxy statements)

Percentage of shares owned by current CEO and all former CEOs still on the board (Proxy statements)

Percentage of shares owned by management prior to acquisition (SEC-filings; proxy statements)

Market-to-book ratio

Tobin’s Q; return on assets

Tobin’s Q

Market-to-book ratio

OLS-regression (piecewise linear)

OLS-regression; IV-regression; fixed-effects panel data model

OLS-regression (piecewise linear); IVregression

OLS-regression

Size, leverage, industry dummies

R&D, advertising, investment and cash flow ratio, stock price volatility, size, industry and time dummies

R&D and advertising ratio, size

n/a

Focus on determinants of insider ownership which has almost doubled from an average of 13% in 1935 to 21% in 1995; 1935: nonmonotonic relation (increase in performance 0-5%, decrease 5-25% (both significant), decrease beyond 25% (not significant)); 1995: similar but less significant results; the breakpoints in the non-monotonic relationship do not change from 1935 to 1995

Insider ownership is far more widespread in small firms (32.0%) than in large firms (13.4%); some evidence of bell-shaped relation between insider ownership and performance; controlling for fixed firm effects changes in insider ownership have no significant impact on performance

Focus on board composition; noticeable reduction in agency costs resulting from increasing insider ownership at low levels of ownership; insider ownership seems to be endogenous; pooled results are similar to single year results

Non-linear relationship between management’s pre-transaction equity position and firm value prior to acquisition; robust to control for industry effects

continued on next page . . .

Alignment of interest hypothesis; entrenchment hypothesis

Insider ownership is explained by unobserved firm heterogeneity (i.e. differences in the contracting environment); firms choose among alternative mechanisms for minimizing agency costs

Alignment of interest hypothesis; entrenchment hypothesis

Segmentation hypothesis (misalignment of insider and outsider shareholders’ interests if the insider ownership stake is either too low or too high)

Insider Ownership and Performance 113

Mehran (1995) Journal of Financial Economics

McConnell and Servaes (1995) Journal of Financial Economics

Loderer and Martin (1997) Journal of Financial Economics McConnell and Servaes (1990) Journal of Financial Economics

U.S. (1976, 1986, 1988) Cross sections of 1,173 firms in 1976, 1,093 in 1986 and 826 firms in 1988 U.S. (1979-1980) Cross section of 153 randomly selected manufacturing firms

U.S. (1976, 1986) Cross sections of 1,173 firms in 1976 and 1,093 in 1986

U.S. (1978-1988) 867 acquisitions

. . . continued from previous page Howe, Vogt and He U.S. (1979-1993) (2003) Four sub-samples The Financial Re(n): dividend view initiations (293), dividend increases (6,534), dividend decreases (997) and share repurchases (124)

Percentage of shares owned by i) the CEO, ii) top executives and iii) all officers and directors (Proxy statements)

Percentage of shares owned by officers and directors (Disclosure)

Percentage of shares owned by officers and directors (Proxy statements) Percentage of shares owned by officers and directors (Value Line Investment Survey)

Percentage of shares owned by insiders (i.e. officers, directors and beneficial owners) (only stakes > 5% at year end in the year prior to the event date) (CDA/Spectrum)

Tobin’s Q; return on assets

Tobin’s Q

CAR; Tobin’s Q; accounting profit rate Tobin’s Q; accounting profit rate

Short-term (3 days) and long run (36 month) CARs; Tobin’s Q; accounting profit rate

OLS-regression

OLS-regression

OLS-regression; 2SLS-regression (simultaneous equation system) OLS-regression

Event study; OLS-regression

R&D and fixed assets ratio, leverage, size, business risk

Leverage, blockholders, institutional investors, R&D and advertising ratio, size

Block ownership, R&D and advertising ratio, leverage, size, industry dummies

Block ownership, acquisition payment, stock price volatility

Tobin’s Q, free cash flow, capital expenditure ratio

Firm performance is positively related to the percentage of equity held by managers and to the percentage of their compensation that is equity-based; insider ownership and equitybased compensation substitute for each other; robust to use of return on assets as performance measure

Insider ownership is higher in high growth than in low growth firms; significant nonmonotonic bell-shaped relation between insider ownership and firm value; this effect is slightly larger in low-growth firms

Relation between insider ownership and Tobin’s Q slopes upward until insider ownership reaches approx. 45% and then slopes slightly downward; results are robust to inclusion of control variables and the use of the accounting profit rate as performance measure

Firms grouped in three equally sized groups according to low, medium and high insider ownership; short-term announcement effects are significantly higher for the high insider ownership group (compared to the low insider ownership group) for all four event-types; long run results only yield significant outperformance for high insider ownership firms for the dividend increase and dividend decrease sample; results are robust to other ownership grouping techniques No evidence is found that larger managerial shareholdings lead to better performance

continued on next page . . .

Alignment of interest hypothesis; incentive hypothesis

Stulz (1988) model: increasing insider ownership decreases the probability of a successful tender offer while the premium in case of a tender offer increases Equity ownership structure matters but differs between firms with many and those with few positive net present value projects

Ownership structure is equilibrium output from a market process

Insider ownership is an important factor in explaining market reactions to cash flow announcements; signaling framework of Leland and Pyle (1977)

114 Insider Ownership and Performance

U.S. (1994-2000) 2,808 firm years from 508 Fortune 500 firms

Edwards and Nibler (2000) Economic Policy

Germany (1992) 105 listed and 51 unlisted firms of the 200 largest non-financial firms

II. Studies on the German market Edwards and WeiGermany (1992) chenrieder (2004) 97 listed firms (of German Economic the Nibler (1998) Review sample, which consists of 158 of the 200 largest non-financial firms)

Villalonga and Amit (2006) Working Paper

. . . continued from previous page Morck, Shleifer U.S. (1980) and Vishny (1988) Cross section of Journal of Finan371 Fortune 500 cial Economics firms

Family ownership (not insider ownership); control rights on shareholders’ meetings; cash flow rights as percentage of total dividends (Sample of Nibler (1998)) Family ownership (not insider ownership); largest shareholder conditional on i) being an individual/ family with minority stake or ii) being an individual or family with majority stake

Percentage of shares owned by i) all board members and ii) the top two officers (only stakes > 0.2%) (Corporate Data Exchange) Percentage of shares owned by officers and directors (Proxy statements)

Market-to-book ratio (only listed firms)

Market-to-book ratio

Market-to-book ratio; Tobin’s Q

Tobin’s Q; accounting profit rate

OLS-regression

OLS-regression

OLS-regression; difference in mean; fixed- and random effects panel data model

OLS-regression (piecewise linear)

Largest, second and third largest shareholder’s stake by identity, size, beta, codetermination, leverage, pension and other provisions, industry dummies

Leverage, pension and other provisions ratio, beta, sales growth, size, industry dummies

Family ownership, block ownership, control enhancing device, family CEO, governance index, dividends, leverage, market risk, size, R&D and investment ratio, diversification, age, industry and time dummies

R&D and advertising ratio, leverage, size, growth of labor force, industry dummies

Focus on ownership concentration; family control rights are detrimental to firm value on a significant level; family cash flow rights are significantly positively related to firm value; estimated percentage change of the market-to-book ratio when ownership changes from complete dispersion to full ownership is largest (102.1%) for the type of family shareholders Focus on effects of bank ownership and concentration; the largest shareholder — conditional on being an individual or family with minority (majority) stake — has a (not) significant positive effect on performance; ownership is highly concentrated but banks do not play an important role in the governance of large German firms while other types of large shareholders do

Significant non-monotonic relation (increase in performance 0-5%, decrease 5-25%, increase beyond 25%); robust to different breakpoints, exclusion of large firms, split between outsider and insider board ownership; similar results for top two officers; not robust to accounting profit rate as performance measure Family ownership creates value if the founder serves as the CEO or Chairman with a hired CEO; firm value is destroyed when descendants serve as CEO

continued on next page . . .

Efficient monitoring hypothesis (but effectiveness depends on blockholder’s identity); expropriation hypothesis

Expropriation by majority shareholders (control rights); alignment of interest hypotheses (cash flow rights)

Owner-manager conflict in non-family firms is more costly than conflict between family and nonfamily shareholders in founder-CEO firms

Alignment of interest hypothesis; entrenchment hypothesis (cf. Chen, Hexter and Hu (1993) for similar results)

Insider Ownership and Performance 115

Chen, Cheung, Stouraitis and Wong (2005) Pacific-Basin Finance Journal

Hong Kong (1995-1998) Sample of 412 firms (1,648 firm years) publicly listed on the Stock Exchange of Hong Kong

Percentage of shares owned by the controlling family (aggregated shareholdings of directors with the same surname) (Annual reports; Financial Times)

Voting rights’ concentration of family shareholders (not insider ownership) in the first 3 years following the IPO (Diverse: Hoppenstedt Aktienf¨uhrer, Commerzbank “Wer geh¨ort zu wem?” etc.) Gorton and Schmid Germany (1989Percentage of control (2002) 1993) rights owned by Federal Reserve 5 cross sections management, other Bank of St. Louis with a total of employees and (doWorking Paper 902 observations mestic) families on the largest (Hoppenstedt (fornon-financial merly: Saling) firms Aktienf¨uhrer) III. Studies on other markets (single-country studies) Bøhren and ØdeNorway (1989Percentage of shares gaard (2004) 1997) owned by i) all offiWorking Paper Between 129 to cers and directors, ii) 217 non-financial aggregated and iii) the firms (1,069 firm largest insider years) listed on (Insider transactions the Oslo Stock reported to the Oslo Exchange (OSE) Stock Exchange (OSE))

. . . continued from previous page Ehrhardt and Germany (1970Nowak (2003a) 1990) Journal of Small 105 IPOs of Business Managefounding family ment owned firms

OLS-regression; GMM and fixed-effects regression; 3SLS regression (simultaneous equation system)

Tobin’s Q; return on assets; market return on stock

OLS-regression (piecewise linear); firm fixed-effects model

Locally weighted regression (LOESS)

Market-to-book ratio

Market-to-book ratio; return on assets; return on equity

Difference in mean; difference in median

Long run (36 months) buy-andhold abnormal returns

Ownership concentration measures, blockholders by owner type, board size, fraction of voting shares, leverage, dividends, investments to income, stock price volatility, stock turnover, beta, market value of equity CEO duality, number and type of directors, audit committee, size, leverage, sales growth, industry and time dummies

Codetermination, bank- and government ownership, block ownership, size, industry and time dummies

Sample divided in four groups according to voting proportion of founding family stake 3 years after IPO: i) 0-25%, ii) 25-50%, iii) 50-75% and iv) 75-100%

Corporate governance matters for economic performance: insider ownership matters the most, direct ownership is superior to indirect ownership and performance decreases with board size, leverage, dividend payout, nonvoting shares and outside ownership concentration; partial analysis makes sense since the results persist in single-equation models; simultaneous equation results are very sensitive to the choice of instruments Focus on ownership concentration and dividend policy; no positive (significant) relationship between family ownership and performance can be observed; negative effect on market-to-book ratio if CEO also serves as chairman of the board; little relationship between family ownership and dividend policy

Study focus on performance effects of codetermination; insider ownership (control rights) has significant positive effect on performance in all years except 1989 (and in the pooled cross section as well); ownership is exogenous since control changes are rare

Influence of founding family persists for at least 10 years following the IPO; nonmonotonic relation between family ownership and performance; if concentration remains high (group iv) significant negative effect on performance; negative but not significant effects for groups i) and iii); positive but not significant effect for group ii)

continued on next page . . .

Ownership structure is equilibrium output from a market process or due to unobserved firm heterogeneity

Alignment of interest hypothesis; entrenchment hypothesis

Alignment of interest hypothesis

Private benefits of control for family shareholders are so large that it seldom pays out to sell the firm completely; negative effects mainly attributed to issuance of nonvoting shares

116 Insider Ownership and Performance

Spain (19901999) 1,233 firm years of non-financial companies

Norway (1996) Cross section of 120 non-financial listed firms

Denmark (19982001) Between 425 firm years of non-financial companies

De Miguel, Pindado and De la Torre (2004) Strategic Management Journal

Mishra, Randøy and Jenssen (2001) Journal of International Financial Management and Accounting

Rose (2005) European Management Journal

. . . continued from previous page Chen, Guo and Japan (1987Mande (2003) 1995) Pacific-Basin 1,040 to 1,094 Finance Journal firm years for 123 Japanese firms

Percentage of shares owned by all insiders (Annual accounts; Copenhagen Stock Exchange)

Founding family ownership (not insider ownership); percentage of shares owned by the founding family and its relatives (Annual reports; secondary sources)

Percentage of shares owned by all insiders (CNMV Spanish Security Exchange Commission “Significant shares for all quoted companies”)

Percentage of shares owned by the board of directors (Nikkei NEEDS; Micro database)

Tobin’s Q

Tobin’s Q

Market value of shares divided by the replacement value of total assets

Tobin’s Q

OLS-regression; 3SLS-regression

OLS-regression

IV-estimator: Generalized Method of Moments (GMM)

OLS-regression; 2SLS-regression (simultaneous equation system); fixed-effects model

Insider compensation, size, capital expenditures, trust ownership, outside board members, supervisory board member age, industry dummy

Size, sales growth, financial leverage, return on assets, assets tangibility, age, board (size, independence, stock ownership), CEO tenure, industry dummies

Ownership concentration, leverage, size, ratio of intangible assets to total assets

Institutional and bank ownership, R&D ratio, size, leverage, return on assets, market value of equity, dividends, stock price volatility, industry dummies

In OLS-regressions the coefficients of the insider ownership variable are positive but significant (0.01 level) only for low levels of insider ownership; in a 3SLS simultaneous equation system this relationship disappears and firm performance seems to influence insider ownership (reverse causality); mean insider ownership for Danish firms is relatively small about 8.6%

Founding family control measured by 4 different measures (one of them ownership); positive relation between founding family control and firm value; founding family control is more valuable among younger firms (only for the Founder-CEO specification) and firms with small board size; relationship dilutes when multiple share classes are present

OLS-results indicate a negative impact of insider ownership on performance at low levels of insider ownership and a positive impact at high levels; controlling for firm fixedeffects performance increases monotonically with managerial ownership; robust to treatment of ownership and performance as endogenous variables Ownership concentration has a non-linear (bell-shaped) effect on Spanish firm values while in other countries that effect is linear; for insider ownership the non-linear Morck, Shleifer and Vishny (1988) relationship can be observed but the entrenchment-interval (3570%) is higher than in the U.S. (5-25%or 1241%); differences can be explained by differences in the corporate governance systems

continued on next page . . .

Managerial ownership is not crucial in corporate governance

Founding family involvement reduces agency costs and allows improved monitoring

Alignment of interest hypothesis; entrenchment hypothesis

Alignment of interest hypothesis

Insider Ownership and Performance 117

Krivogorsky (2006) The International Journal of Accounting

Gugler, Mueller and Yurtoglu (2004) Working Paper

U.S., nonU.S. AngloSaxon Countries; European civil law countries (1996 - 2000) Between 1,924 and 16,543 firm years 9 Continental European countries incl. Germany (2000-2001) 81 non-financial firms with dual listing on the NYSE

IV. Multi-country studies Barontini and 11 ContinenCaprio (2005) tal European ECGI - Working countries incl. Paper Germany excl. U.K. and Ireland (1999-2001) 1,852 firm years

. . . continued from previous page Schmid (2003) Switzerland Working Paper (2002) Cross section of 145 listed firms

Percentage of shares owned board members; dummy variable if CEO is founder of the firm (Proxy statements; 20-F)

Identity of the ultimate largest shareholder and its voting and cash flow rights (Stock market registers; annual reports; Worldscope; Extel; Osiris; national annual directories) Percentage and value of shares owned by insiders (Compact Disclosure; Global Vantage; Compustat)

Percentage of shares owned by officers and directors (Annual reports)

Market-to-book ratio; return on assets; return on equity

Tobin’s Q (average; marginal)

Tobin’s Q; return on assets

Tobin’s Q

OLS-regression; GLS regression

OLS-regression

Random effects OLS-regression; robust regression biweight estimator (M-estimator)

OLS-regression; 3SLS-regression (simultaneous equation system)

Size, age, asset growth, leverage, block ownership, institutional ownership, board structure, industry dummies

Size, institutional shareholdings, R&D to sales ratio

Size, sales growth, leverage, industry dummies

Block ownership, leverage, outside board members, size, stock price volatility, beta, sales growth, age, dividends, intangibles, industry dummies

Strong positive impact of relational investors and independent directors on performance; insider ownership has no influence on performance; only if the CEO is founder of the company, it has a positive impact in the marketto-book equation; small sample size (81 companies from 9 different countries) yields regression models with relatively low explanatory power

Separation of wealth and entrenchment effect: the first is significantly positively the latter significantly negatively related to performance across all countries and control categories; the wealth effect is stronger in the U.S. than in other countries

Family control is positive for European corporations at both the founder and the descendants stage; these benefits are partially wasted by inefficient control strategies (i.e. the separation of control and cash flow rights)

Statistically significant positive influence of shares owned by officers and directors on firm value; similar results for a constructed managerial incentive index; no significant impact for ratio of stock-based to cash compensation; governance mechanisms are significantly related to each other

continued on next page . . .

Continental European countries are stakeholder oriented and insider dominated; maximization of share price is not necessarily management’s main objective

Alignment of interest hypothesis; Entrenchment hypothesis

Alignment of interest hypothesis

Alignment of interest hypothesis

118 Insider Ownership and Performance

U.S., U.K., Germany, Japan (1998-1999) 4 cross sections of 2,198, 674, 319, and 1,015 firms respectively

Family ownership (not insider ownership): four dummy variables to measure family control by the largest shareholder (Faccio and Lang (2002); Worldscope) Percentage of shares owned by insiders (Compact Disclosure, www.hemscott.co.uk, Hoppenstedt Aktienf¨uhrer, Japan Company Handbook) Tobin’s Q

Tobin’s Q; return on assets

OLS-regression; 2SLS-regression (separate analysis for each country)

Fixed-effects OLS-regression

Block ownership, institutional ownership, leverage, capital expenditures, cash flow, size, business risk, sales growth, industry dummies

Size, sales growth, capital expenditures, leverage, industry dummies

No consistent relationship between insider ownership and performance across the four countries, but significant associations within the four countries (negative effect in U.S. and U.K. vs. positive effect in Germany and Japan); unclear results on endogeneity of ownership; cross country comparisons difficult because of data restrictions

Family ownership improves profitability if at least one of the top two officer positions is held by family (especially at higher control levels); firm valuation is only positively effected at lower control levels (nonmajority firms); passive family ownership does not effect profitability Ownership matters depending on the specific CG-system; ownership structure is equilibrium output from a market process

Alignment of interest hypothesis; entrenchment hypothesis

Source: Author’s demonstration. A comprehensive overview about empirical studies on ownership structures and performance is also provided by Mathiesen (2005).

Seifert, Gonenc and Wright (2005) Journal of Multinational Financial Management

. . . continued from previous page Maury (2006) 13 European Journal of Corpocountries incl. rate Finance Germany (1998) 1,672 nonfinancial firms

Insider Ownership and Performance 119

120

Insider Ownership and Performance

4.2.3

Selected Studies on Insider Ownership and Performance

4.2.3.1 Results from Capital Market Oriented Systems The study of M ORCK , S HLEIFER AND V ISHNY (1988) is among the first to empirically test the effects of insider ownership on performance as predicted by agency theory. In a 1980 cross-section of 371 U.S. Fortune 500 firms, they find evidence of a nonmonotonic relationship between insider ownership and firm value as measured by Tobin’s Q. Specifically, ownership stakes of the board of directors seem to be value enhancing in the 0-5% and 25-100% range but value decreasing in the interval from 5-25%. This pattern is explained by antagonizing alignment of interest and entrenchment effects. While at low and high levels of insider ownership the alignment of interest effects seem to dominate, entrenchment effects prevail at intermediate levels of managerial shareholdings. Since it is a priori not possible to predict which effect will dominate at different levels of insider ownership, the exact position of the inflection points in the ownership-performance relationship remains an empirical issue. As theory on the exact thresholds of value increasing (decreasing) levels of insider ownership is poor, later studies arbitrarily use different points of inflection.477 To test the robustness of their results, M ORCK , S HLEIFER AND V ISHNY (1988) also use accounting profit rates as an alternative performance measure. As a result, they obtain similar yet hardly significant results.478 Two years later, M C C ONNELL AND S ERVAES (1990) provide new evidence on the relationship between insider ownership and performance by examining a sample of 1,173 and 1,093 non-financial U.S. firms in 1976 and 1986 respectively. They find a positive relationship between performance as measured by Tobin’s Q and insider ownership only for insider ownership levels up to 49.4% (1976) and 37.6% (1986) respectively. Above this threshold the relationship becomes slightly negative. This curvilinear pattern is often referred to as a bellshaped or inverse U-shaped relationship. The results are robust to the use of a set of control variables to address the issue of spurious correlation between insider ownership, performance and potentially omitted variables. Using return on assets as an alternative accounting based performance measure, they get similar results. Even though the results are qualitatively comparable to those of 477 478

Cf. Hermalin and Weisbach (1991); Chen, Hexter and Hu (1993); Davies, Hillier and McColgan (2005). See Morck, Shleifer and Vishny (1988).

Insider Ownership and Performance

121

M ORCK , S HLEIFER AND V ISHNY (1988), they do not find evidence for their piecewise-linear relationship between insider ownership and performance.479 The next trend-setting study was carried out by AGRAWAL AND K NOEBER (1996) who examine seven different corporate governance mechanisms to control managerial agency problems individually and simultaneously. Besides insiders’ shareholdings, shareholdings of institutions, large blockholders, the percentage of outside directors, debt policy and the market for managerial labor and corporate control are treated as interdependent mechanisms to align management’s interests with those of the shareholders. Out of the Forbes 800 firms in 1987480 , they analyze 383 U.S. firms with complete data and find that insider ownership only has a positive impact on performance as measured by Tobin’s Q in the individual OLS-regression model. If insider ownership is included in an OLS-regression model together with the other six governance mechanisms or treated as an endogenous variable in a 2SLS-regression, the effect on performance is not significantly different from zero anymore. Similarly, the percentage of outside directors, debt policy and the corporate control market show significant coefficients in the individual OLS-regressions but only the percentage of outside directors preserves its puzzling negative impact in the 2SLSregression. Thus, the findings of AGRAWAL AND K NOEBER (1996) largely contradict to those of the two previously presented studies. However, they are consistent with the view of D EMSETZ (1983) that firms’ ownership structures are the endogenous outcome of an equilibrium process and, therefore, control mechanisms are optimally chosen.481 Regarding the relationship between insider ownership and performance, H IM MELBERG , H UBBARD AND PALIA (1999) analyze unbalanced U.S. panel data ranging from 398 firm observations with available ownership data in 1982 to 293 in 1992. They employ a fixed-effects model and instrumental variables to control for unobserved firm heterogeneity. First, they show that observable characteristics in firms’ contracting environment influence the level of managerial ownership. Examples for such characteristics are firm size (measured by 479 480

481

See McConnell and Servaes (1990). The Forbes magazine lists the 500 largest U.S. firms as measured by four different key figures: sales, total assets, market value of equity and profits. A total of 800 firms make it into at least one of the four lists. See Agrawal and Knoeber (1996), p. 385. See Agrawal and Knoeber (1996). Cf. Demsetz (1983), p. 377.

122

Insider Ownership and Performance

sales), the scope for discretionary spending (measured by the capital to sales ratio or the ratios of R&D and advertising spending to capital) or managerial risk aversion (measured by the standard deviation of stock prices). Second, analyzing the link between managerial ownership and performance they cannot find evidence that changes in insider ownership affect performance if they control for both observed firm characteristics and firm fixed-effects. Their findings are in line with the notion that firms optimally choose corporate governance mechanisms in the long run to reduce agency costs.482 D EMSETZ AND V ILLALONGA (2001) examine the relationship between ownership structure and performance for a sub-sample of 233 U.S. firms of the earlier D EMSETZ AND L EHN (1985) study which contained ownership data on 511 corporations in 1980. Criticizing that most prior studies exclusively relied on measures of ownership concentration or managerial shareholdings, they argue that both measures should be examined simultaneously to account for the complexity of interests represented by a given ownership structure. Accounting for endogeneity by comparing OLS- and 2SLS-regression results with performance (Tobin’s Q) and insider ownership as dependent variables, none of the two ownership variables turns out to be statistically significant in the performance equation. Hence, they find no evidence that variations in firms’ observed ownership structures cause systematic variations in observed firm performance. Instead, ownership structures seem to evolve endogenously in a market process which removes any predictable relationship between observed ownership structures and performance. Differences in ownership structures — diffused vs. concentrated or low vs. high levels of insider ownership — are originated in the firms’ environment (e.g. regulation or scale economics).483 Finally, two recent studies analyzing family rather than insider ownership will be presented. First, A NDERSON AND R EEB (2003) analyze the link between founding family ownership and firm performance for a sample of 403 nonfinancial/ non-utility firms during the period between 1992 and 1999. They find that family ownership in the U.S. is significant in 35% of the S&P 500 companies and that total family ownership on average reaches around 18% of the outstanding equity. Using Tobin’s Q and return on assets as performance 482 483

See Himmelberg, Hubbard and Palia (1999). See Demsetz and Villalonga (2001).

Insider Ownership and Performance

123

measures, they find that family firms perform better than non-family firms, especially if a family member serves as CEO. The relationship between family ownership and performance is non-linear: By including a quadratic term of the family ownership variable, they find that family ownership appears to be value increasing up to ownership stakes of about 30% and becomes detrimental to performance beyond this threshold. No evidence is found that minority shareholders are adversely affected if families own considerable stakes in listed companies.484 Second, V ILLALONGA AND A MIT (2006) analyze U.S. Fortune 500 firms in the seven year period ranging from 1994 to 2000. Using Tobin’s Q as dependent variable, they find a positive, linear and significant impact of family ownership stakes (and also of a family ownership dummy variable). They refine their analyses to differentiate between various family firm characteristics and conclude that family ownership only has a positive impact on performance if the founder serves as CEO or takes the position of a chairman with a hired CEO. Descendants serving as CEOs are detrimental to firm value. The authors assume that the conflict between management and the shareholders in non-family firms is more costly than the conflict between family and non-family shareholders in firms where the founder still serves as CEO. Contrarily, in descendant-CEO firms, the conflict between family and non-family shareholders dominates.485 4.2.3.2 Results from Corporate Law Oriented Systems 4.2.3.2.1

Germany

For the case of Germany, a positive relationship between insider ownership and performance is documented by G ORTON AND S CHMID (2002). Analyzing the performance effects of codetermination in Germany, they find as a corollary that insider ownership486 has a significant positive impact on performance while parity-codetermination has a significant negative impact. Examining a sample of the 250 largest non-financial companies and using market-to-book ratios as dependent variable, they find a significant impact in four out of five sample years ranging from 1989 to 1993. This result is also robust to treating the sample 484 485 486

See Anderson and Reeb (2003). See Villalonga and Amit (2006). Insider ownership in the context of this study is defined as the percentage of equity control rights owned by management, other employees and families. See Gorton and Schmid (2002), p. 13.

124

Insider Ownership and Performance

as an unbalanced panel with time fixed-effects. They also find evidence for a positive effect of (outsider) blockholding: The variable measuring the degree of concentration, i.e. the percentage of shares owned by the largest shareholder, enters all regression equations on a significant level as well.487 In the context of family firms, E HRHARDT AND N OWAK (2003 A ) analyze 105 German IPOs of founding family owned firms in the period ranging from 1970 to 1990. Ten years after the IPO, the mean ownership stake of the founding family decreases to 40.4%, i.e. majority control has usually been given up. Furthermore, they classify IPOs into four groups according to the stake still held by the founding family three years after the IPO. In doing so, they find that the long run abnormal performance as measured by buy-and-hold stock returns is only positive if the founding family stake lies in the 25-50% quartile. For the other three groups (i.e. 0-25%, 50-75% and 75-100%), the abnormal performance is negative but only the coefficient of the group with family ownership stakes above 75% is statistically significant different from zero. The results indicate that involvement of the founding family is value increasing only within a certain ownership range. E HRHARDT AND N OWAK (2003 A ) explain their results by a combination of alignment of interest, signaling and entrenchment considerations.488 Focusing on ownership concentration and share valuation, E DWARDS AND W EICHENRIEDER (2004) examine a sample of 97 German companies in 1991. They try to disentangle the beneficial effects of concentration (e.g. increased monitoring by blockholders, less expropriation due to greater cash flow rights) from the harmful effects (e.g. private benefits of control) by differentiating between control rights and cash flow rights.489 Thereby, they find that the concentration of cash flow rights has a positive impact on the market value of equity for all but one type of largest shareholder with public sector shareholders being the only exception. Moreover, the positive effect is largest and highly significant if a family or individual represents the largest shareholder. In contrast, the concentration of control rights exerts a negative impact on firm value for all types of largest shareholders. Even though their classification of family shareholders 487 488 489

See Gorton and Schmid (2002). See Ehrhardt and Nowak (2003a). Cf. section 6.3.1.1.

Insider Ownership and Performance

125

is not identical to that of insiders applied in the empirical study (chapter 6), both groups are likely to overlap considerably. Thus, the results of E DWARDS AND W EICHENRIEDER (2004) can also be seen as evidence regarding the insider ownership-performance relationship. 490 4.2.3.2.2

Other Nations

In Norway, B ØHREN AND Ø DEGAARD (2004) analyze a sample of 1,069 firm years of companies listed on the Oslo Stock Exchange (OSE) in the period ranging from 1989 to 1997. They conclude that insider ownership creates value if it does not rise above a certain level (inflection point around insider ownership levels of 61%). As the results are robust to the use of a wide range of single–equation models, it appears that governance mechanisms are independent and, hence, may be examined individually rather than only in an integrated system as claimed by some researchers.491 The findings seem to depend on the choice of the performance measure and especially on the choice of instruments in the simultaneous equation models. Thus, the inability to detect significant relationships in models allowing for endogeneity may not indicate the optimal use of corporate governance mechanisms but rather might be the result of an underdeveloped theory of how adequate instruments for the ownership variable can be found.492 The Japanese market is examined by C HEN , G UO AND M ANDE (2003) who study the relationship between insider ownership and firm value as measured by Tobin’s Q for 123 firms in the nine year period between 1987 and 1995. The prevalence of managerial ownership in Japan — a mean of 2.0% — is rather low compared to other corporate law countries (and even compared to the U.S.493 ). They find a positive (negative) relationship between insider ownership and performance at high (low) levels of insider ownership in their single equation models. However, controlling for firm fixed-effects the relationship becomes monotonic and positive. The result that managerial ownership has a 490 491 492

493

See Edwards and Weichenrieder (2004). See Agrawal and Knoeber (1996); Schmid (2003). See Bøhren and Ødegaard (2004). Cf. Mishra, Randøy and Jenssen (2001) who analyze the influence of Norwegian founding family ownership on firm value. Rose (2005) analyzes the link between managerial ownership and firm performance for listed Danish firms. For example, Himmelberg, Hubbard and Palia (1999), p. 362, report average managerial ownership stakes of 13.4% even for the largest sub-group of their sample. Cf. section 4.2.3.1.

126

Insider Ownership and Performance

positive impact on firm value also persists if managerial ownership is treated as an endogenous variable in a simultaneous equation system.494 In Spain, D E M IGUEL , P INDADO AND D E LA T ORRE (2004) analyze an unbalanced panel data set of 135 non-financial listed companies for the period ranging from 1990 to 1999. They find a cubic relationship between the level of insider ownership and firm value: For insider ownership levels below 35% and above 70%, the effect on firm value is positive. In the intermediate range from 35-50% the impact is negative. They explain their results with a combination of convergence of interest and entrenchment arguments. While their results are generally in line with (earlier) results from the U.S., it seems that Spanish insiders get entrenched at higher levels of ownership than their U.S. peers. Furthermore, they argue that expropriation in Spain might be promoted by the comparable illiquidity of the domestic capital market as minority shareholders are limited in their ability to sell stakes if they feel exploited by controlling shareholders.495 In the case of Switzerland, a study of 145 listed firms performed by S CHMID (2003) at the end of 2003 reveals a positive and statistically significant influence of managerial shareholdings on firm value as measured by Tobin’s Q. Besides the percentage of shares owned by the members of the board, two additional measures of the officers’ and directors’ commitment to the firm are examined: the percentage of cash to share based compensation as well as a proprietary constructed compensation index taking into account ten different aspects of compensation. Thereby, only the latter confirms the positive effect on firm value on a statistically significant level. The positive relationship between insider ownership and performance persists if four additional governance mechanisms (concentrated outside shareholdings, debt policy, board size and fraction of outside board members) are incorporated in a simultaneous equation system. The results are also highly supportive for the hypothesis that different corporate governance mechanisms (e.g. managerial shareholdings and outside directors of the board) are substituted for each other.496 494 495 496

See Chen, Guo and Mande (2003). For comparable results regarding the Taiwanese market cf. Chen (2005). See De Miguel, Pindado and De la Torre (2004). See Schmid (2003).

Insider Ownership and Performance 4.2.4

127

Shortcomings of German Research

As can be seen from the overview of empirical studies on insider ownership (section 4.2.2), the majority of studies focusing on the German market is rather descriptive in nature. Typically, these studies describe either generally the ownership structures of German listed companies or more specifically the identity of their (largest) blockholder(s). However, far less research was performed on the economic impact of these control arrangements (i.e. the shareholder structures) for the German market even though comparable international evidence is rather rich.497 Traditionally, corporate governance related empirical studies focus on the special role of banks in the German market498 while a broader discussion about corporate governance just evolved recently. This might explain why relatively few empirical studies analyze the determinants and effects of ownership structures of German listed stock corporations. This is surprising as precisely the prevalence of bank financing as a particularity of the German market suggests that the role of certain shareholder structures might be different from that in other countries and corporate governance systems respectively. Moreover, the comparatively few studies analyzing German shareholder structures often suffer from a lack of data availability and poor data quality. Before 1995, only shareholdings above 25% had to be disclosed thereby seriously limiting (or rather prohibiting) the chance to gain an authentic picture on the shareholder structures of German listed companies.499 Thereafter, the duty to report shareholdings, which passed or fell below the thresholds of 5%, 10%, 25%, 50% and 75%, to the BaFin improved the availability and quality of shareholder structure data. Nevertheless, certain disclosure gaps remain and evidence on the effective enforcement of disclosure regulation is still lacking.500 As a consequence, ownership panel data — necessary for applying advanced econometric estimation techniques — is just in the process of becoming available. It

497 498

499

500

See Becht and B¨ohmer (2003), p. 26. Cf. Wenger and Kaserer (1998a); Schr¨oder and Schrader (1998); Edwards and Nibler (2000); Gorton and Schmid (2000); Andreani (2003). As a result, especially earlier studies rely on rather incomplete data sources as depot statistics which only incorporated the shareholdings deposited at financial institutions. Cf. Iber (1987); Portfolio Management (1988). See Becht and B¨ohmer (2003), pp. 15-26; Bott (2002), pp. 172-240.

128

Insider Ownership and Performance

is likely that data quality will further improve in the future when the disclosure of additional voting thresholds becomes mandatory.501 At this point in time, the accuracy of available shareholder data does not suggest that the construction of a high frequency (e.g. yearly) panel data set constitutes a promising effort.502 Therefore, the shareholder structures of an unbalanced low frequency sample of German listed corporations at the end of 1993, 1998 and 2003 will be examined in the empirical part of this study (chapter 6). The detailed classification of shareholders along the dimensions of shareholders’ identity and shareholders’ concentration is rather unique and especially adequate for the purpose of corporate governance studies. The examination of similarly constructed samples in the future, when data quality has further improved and finally also smaller shareholdings have become more transparent, suggests to be a promising field of research.

501 502

For plans to alter the disclosure thresholds of voting blocks see footnote 602. Cf. sections 6.2.2 and 6.10. An analysis in the empirical part (chapter 6) shows that in 56% of all cases the insider ownership stake remains largely unchanged over a ten year period. See figure 6-17.

Hypotheses on Insider Ownership

5

129

Hypotheses on Causes and Effects of Insider Ownership

This chapter will develop hypotheses related to the determinants of insider ownership (section 5.1), the effects of insider ownership on corporate control (section 5.2) and the effects of insider ownership on corporate performance (section 5.3). The formulated hypotheses will be based on theoretical considerations as well as empirical evidence and will be tested in the empirical study (chapter 6). All hypotheses will be formulated and tested as non-directional hypotheses. Non-directional hypotheses, in contrast to directional hypotheses, only posit the existence of a relationship but do not specify the direction (i.e. positive or negative relationship). They are preferred to directional hypotheses because of two reasons: First, for some hypotheses (e.g. H-5.1.2 or H-5.2.7) it can be argued in favor of a relationship in both directions making the formulation of a directional hypothesis impossible. Second, non-directional hypotheses provide more conservative estimates and allow a uniform use of test statistics.503 5.1

Determinants of Insider Ownership

In their famous study, D EMSETZ AND L EHN (1985) refer to three general forces affecting ownership structures and claim that the value-maximizing firm size is the most obvious. A negative effect of firm size on ownership concentration can be expected because of two reasons: First, an increasing firm size usually implies that the market value of a given fraction of ownership rises which in turn usually reduces the degree of ownership concentration. Second, a given degree of control can generally be achieved with a smaller ownership share in larger firms where ownership is typically more dispersed.504 Since high insider ownership levels are generally only a particular occurrence of ownership concentration, it can be expected that insiders — and especially founder-owners — reduce their shareholdings along with an increasing firm size. The assumption of a negative relationship between firm size and insider ownership is supported by a survey on 400 small firms in the U.S. performed by N ORTON (1991). The results show that agency considerations play a compa503 504

Cf. Laatz (1993), pp. 519-521. See Demsetz and Lehn (1985), pp. 1157-1158. Demsetz (1983), p. 388, already reported that managerial ownership rose from 2.1% for the ten largest U.S. Fortune 500 firms in 1975 to 20.4% for the ten smallest, and to 32.5% for randomly selected non-Fortune 500 firms. Cf. Himmelberg, Hubbard and Palia (1999), p. 364.

130

Hypotheses on Insider Ownership

rable minor role in the choice of the capital structure. Instead, a pecking order running from using internal financing, external short-term and long-term debt to issuing common equity as a last device seems to describe the funding behavior most adequately.505 Hence, growing firms would have to issue common equity from time to time and thereby dilute the erstwhile owner-managers’ stakes in the company. As a consequence, large firms will show lower levels of insider ownership than smaller firms.506 Therefore, the following hypothesis is proposed: Hypothesis 5.1.1 Firm size has an effect on the level of insider ownership. There are diverse explanations that firm specific risk has either a positive (A) or negative (B) effect on the level of insider ownership. For the former (hypothesis H-5.1.2 A), D EMSETZ AND L EHN (1985) argue as follows: If interests of managers and shareholders are perfectly aligned through the markets for corporate control and managerial labor, control potential would not help to explain ownership structures.507 However, typical market frictions, e.g. transaction costs, “[. . . ] impose a specific identity and control potential on firms. Alterations in the structure of corporate ownership [. . . ] can be understood as a response to these costs.”508 The costs of monitoring management are lower for firms operating in a relatively stable environment (low firm specific risk) while in less predictable environments (high firm specific risk), monitoring becomes more difficult and opportunities for moral hazard rise.509 Hence, the control potential rises with increasing uncertainty which in turn would increase the effects of an alignment of interests between shareholders and management through high insider ownership levels. Thus, higher firm specific risk may be expected to lead to higher insider ownership levels.510 505 506 507

508 509

510

See Myers (1984), pp. 581-585. See Norton (1991), pp. 290-291. Cf. Faccio and Lang (2002), p. 381. See Demsetz and Lehn (1985), pp. 1158-1159. Control potential is defined as the wealth gain achievable through the monitoring of managers’ decision making by firm’s owners. Demsetz and Lehn (1985), p. 1159. See Demsetz and Lehn (1985), pp. 1158-1159. Similarly, Black, Jang and Kim (2006), p. 675, argue that the riskier the firm, the stronger the governance needed. Of course the direction of the relationship in this context is debatable: One might plausibly argue that hired managers will adapt a lower risk level since unusual high profits may increase shareholders’ expectations while losses or disappointed expectations might put the manager’s job at risk. See Thonet and Poensgen (1979), p. 24.

Hypotheses on Insider Ownership

131

In favor of a negative relation between firm specific risk and the level of insider ownership (hypothesis H-5.1.2 B) H IMMELBERG , H UBBARD AND PALIA (1999) claim that the optimal level of insider ownership is the lower, the higher the firm specific risk is, because ceteris paribus higher insider ownership levels mean less diversification of managers’ personal wealth.511 From a macroeconomic view, H OLDERNESS , K ROSZNER AND S HEEHAN (1999) find that lower stock volatility and greater hedging opportunities, which are generally associated with the development of financial markets, are important factors in explaining the increase in U.S. insider ownership from 1935 to 1995.512 As D EMSETZ AND V ILLALONGA (2001) claim: “It seems likely that [. . . ] complex consequences of risk influence ownership structure, but the complexity of this influence makes it difficult to pre-sign regression coefficients.”513 Hypothesis 5.1.2 Firm specific risk has an effect on the level of insider ownership. Other potential determinants of insider ownership as proposed by existing literature, e.g. R&D expenditures, the ratio of capital expenditures to sales, the number of operating divisions of the firm or non debt tax shields arising from depreciations and investment tax credits,514 are not elaborated since such data is not readily available for German firms and, consequently, related hypotheses could not be tested in the empirical analysis. From a multi-country perspective, one might argue that insider ownership should be lower (higher) in countries with strong (weak) legal protection of minority shareholders.515 5.2

Insider Ownership and Corporate Control

The supervisory board, or the U.S. equivalent board of directors, was already described as an important mechanism of internal control in section 2.2.3. Insiders owning significant shares of the company might influence the composition of the supervisory board through their voting at the shareholders’ meeting and, 511 512 513 514

515

See Himmelberg, Hubbard and Palia (1999), p. 365. Cf. Mueller (2002), pp. 6, 18-19. See Holderness, Kroszner and Sheehan (1999), pp. 441, 455-466. Demsetz and Villalonga (2001), p. 222. See Himmelberg, Hubbard and Palia (1999), pp. 363-365; Jensen, Solberg and Zorn (1992), p. 249; Chaplinski and Niehaus (1993), p. 58. See Burkart, Panunzi and Shleifer (2003), pp. 2174-2183. Cf. La Porta, Lopez-de Silanes and Shleifer (1999).

132

Hypotheses on Insider Ownership

in case of a three-quarters majority, might also be able to alter its size by changing the statutes.516 Recalling that supervisory board members usually are jointly and severally liable and each member has the right to call a supervisory board meeting, owner-managers might prefer small board sizes to limit the diversity of opinions and, hence, the control capability of the supervisory board517 : “CEOs have the same insecurities and defense mechanisms as other human beings; few will accept, much less seek, the monitoring and criticism of an active and attentive board.”518 The idea that supervisory board size represents a peculiar control mechanism and therefore may influence corporate performance was first proposed by L IP TON AND L ORSCH (1992) and J ENSEN (1993).519 Both argue that large boards are less likely to function effectively or that the number of board members should be limited to seven or eight persons.520 Consequently, owner manages — i.e. both management and supervisory board members with significant shareholdings — have at least two incentives to prefer smaller sized supervisory boards: On the one hand, as managers they are likely to fear the complex and lengthy decision making in and negotiation with large (oversized) supervisory boards.521 On the other hand, as owners they might not be willing to spend corporate resources on ineffectively large supervisory boards.522 Thus, the following hypothesis is posited: Hypothesis 5.2.1 The level of insider ownership has an effect on the size of the supervisory board. The size of the management board determines the number of persons which conjointly are in charge of governing the operating business and deciding on strategic issues. While larger companies with several segments and geographically diverse business activities usually need to adopt a certain management board size to divide tasks among a sufficiently high number of individuals, smaller 516 517 518 519 520

521 522

See §§ 101, 179 AktG. For the rights and duties of supervisory board members see section 3.2.2. Jensen (1993), p. 863. See Beiner, Drobetz, Schmid and Zimmermann (2006), p. 256. See Lipton and Lorsch (1992), pp. 59-77; Jensen (1993), p. 865. Cf. Ben-Amar and Andr´e (2006), pp. 534536. Cf. Cromme (2002a), p. 12-14. Cf. Weber and Dudney (2003), p. 404.

Hypotheses on Insider Ownership

133

companies often are characterized by small board sizes. Besides firm size, the level of insider ownership can be expected to have an impact on management board size as well.523 As J ENSEN AND M ECKLING (1976) claim, the consumption of private benefits rises along with an increasing share of equity sold to outsiders because the associated costs only have to be borne disproportionately. Managers not owning any equity do not directly bear any costs of the private benefits they consume. Hence, owner-managers, aware of this problematic, might fear to engage pure outside managers to assist them in running their business and rather prefer smaller sized management boards. Therefore, the next hypothesis is suggested: Hypothesis 5.2.2 The level of insider ownership has an effect on the size of the management board. A corporation issuing equity to the public provides shareholders usually with both cash flow and voting rights. However, voting rights might be — within the legal boundaries already mentioned — proportionately (“one-share-onevote principle”) or disproportionately assigned to the income claims, i.e. cash flow rights, which constitutes the most common case of voting restrictions.524 Since one of the primary ways of monitoring management is to exercise voting rights at the shareholders’ meeting, owner-managers might have strong interests to disproportionately accumulate voting rights. This restricts the voting rights of other shareholders and helps owner-managers to evade outside shareholders’ scrutiny. The extent to which disproportionate voting rights can be assigned to any party (not necessarily insiders) depends on the ratio of non-voting to voting equity. The larger the ratio, the higher is the divergence of control and cash flow rights. Besides owning disproportionate portions of voting equity themselves, owner-managers might also try to place voting equity by outside investors which are expected to show a supportive attitude towards the incumbent management. While the next but one hypothesis exclusively relates to management’s control rights in excess of its cash flow rights, the following hypothesis proposes that

523

524

Cf. Heaney, Naughton, Truong, Davidson, Fry and McKenzie (2006) who claim that past performance also influences the size and stability of the officers and directors group. Cf. Grossman and Hart (1988), pp. 176-178; Harris and Raviv (1988b), pp. 203-235; Easterbrook and Fischel (1983), pp. 408-410.

134

Hypotheses on Insider Ownership

the mere existence of voting and non-voting equity, as well as its ratio, is driven by the level of insider ownership: Hypothesis 5.2.3 The level of insider ownership has an effect on the presence of voting restrictions. The most widespread instrument which enables owner-managers to accumulate voting rights in excess of cash flow rights is the issuance of non-voting preference shares.525 As G ROSSMAN AND H ART (1988) state, security benefits (cash flow rights) refer to the corporation’s total market value while private benefits of control refer to the benefits the management or large shareholders obtain for themselves.526 If control is valuable beyond the cash flow rights attached to equity ownership, e.g. because of associated private benefits, there are strong incentives to hold a disproportionate share of control rights.527 Original owners, initially often bearing all costs and benefits of control as sole ownermanagers, might be reluctant to cede control and accumulate disproportionate control rights.528 Thus, the following hypothesis is proposed: Hypothesis 5.2.4 The level of insider ownership has an effect on the difference between insiders’ control and cash flow rights. Besides insider ownership, monitoring by large blockholders is another way that can potentially reduce agency problems and, thereby, increase the value of the firm.529 Blockholders with sufficiently large shareholdings may have both the incentive and the power to effectively monitor managers.530 While the relation of insider ownership and corporate performance will be elaborated in the next section, this observation has also interesting implications on the corporate control setting. At high levels of (cash flow) insider ownership interests of management and outside shareholders are rather aligned and agency problems are comparatively small. Therefore, the costs of maldiversification, which blockholders usually have to bear, are likely to outweigh the benefits of increased 525

526 527 528 529 530

Pyramidal ownership is another means of achieving disproportionate control rights. See amongst others K¨oke (1999), pp. 6-8; Franks and Mayer (2001), pp. 946-952. See Grossman and Hart (1988), p. 177. See Denis and McConnell (2003), p. 23. See Gompers, Ishii and Metrick (2004), p. 20. See Seifert, Gonenc and Wright (2005), p. 172. See Shleifer and Vishny (1997), pp. 754-755; Becht (1997), p. 23.

Hypotheses on Insider Ownership

135

control. Consequently, the aggregated level of outside blockholdings can be expected to be lower. Similarly, at low levels of insider ownership the benefits of active control might dominate, leading to an inverse relation between the level of insider ownership and the level of blockholdings.531 As it can reasonably be stated that interests of different blockholder types are diverse,532 it would be preferable to distinguish between those blockholder types in the empirical analyses. However, a lack of data accuracy and the resulting decrease in the number of instances per blockholder type prevents analyses with regard to different blockholder types. This is especially true for rather small sample sizes. Consequently, it will not be differentiated between different blockholder types and blockholders will be treated as a homogenous group. As not just the level of blockholdings but also the number of blockholdings will be effected, the following two hypotheses are posited: Hypothesis 5.2.5 The level of insider ownership has an effect on the level of outside blockholdings. Hypothesis 5.2.6 The level of insider ownership has an effect on the number of outside blockholdings. There are several arguments to believe that insider ownership either has a positive (A) or negative (B) effect on firms’ leverage. For the former (hypothesis H5.2.7 A) J ENSEN AND M ECKLING (1976) claim that diversified shareholders have strong incentives to expropriate the providers of debt capital by investing in riskier projects with higher expected returns (“asset substitution”). Contrary, undiversified shareholders with substantial wealth at risk and a typically longterm investment perspective can be expected to mitigate agency problems with providers of debt capital since firms usually are in need to re-enter debt markets on a regular basis. As a consequence, high insider ownership firms might be characterized by higher degrees of leverage.533 Similarly, L ELAND AND P YLE (1977) argue that because of informational asymmetries, borrowers must indirectly transfer information about the quality of their project (or company) by observable actions, e.g. by investing in their 531 532 533

Cf. Harjoto and Garen (2005), p. 677; footnote 757. See sections 3.3.2 and 3.3.3. See Anderson, Mansi and Reeb (2003), p. 264. Cf. Jensen and Meckling (1976), pp. 333-343.

136

Hypotheses on Insider Ownership

own project. This might be seen as a signal to the lending market that reveals the true quality of the project.534 Hence, firms with higher levels of insider ownership can be assumed to show higher debt levels. A NDERSON , M ANSI AND R EEB (2003) propose that family firms535 face lower costs of debt and therefore adopt higher debt levels. For these firms, the agency costs of debt arising from the asset substitution or risk shifting problem, can be expected to be smaller for at least two reasons: On the one hand, founding families, whose wealth is usually rather undiversified invested in their firms, are especially interested in firm survival since they want to pass “[. . . ] the firm as a going concern to their heirs rather than merely passing their wealth.”536 Hence, they are less likely to increase risk at debtholders’ expenses and more likely to maximize firm value rather than shareholder value compared to other large shareholders. On the other hand, because long-lasting business relationships and reputation concerns play a major role for family firms, exploitive actions on the part of the family appear less likely.537 Finally, dominating inside shareholders might be influenced in their financing decision by their desire to avoid outside influence which would arise from the issuance of voting equity to outside investors. Recalling that non-voting equity can only be issued up to a certain limit in Germany,538 those firms are likely to exploit their debt capability before raising outside equity.539 As a consequence, comparatively higher debt levels may be characteristic for firms with high levels of insider ownership.540 Arguments for a negative effect of the level of insider ownership on firms’ leverage (hypothesis H-5.2.7 B) can be found in the inverse interpretation of the control hypothesis of free cash flow proposed by J ENSEN (1986): Since corporate managers (i.e. agents without significant levels of ownership stakes in the firm) 534 535

536 537 538 539

540

See Leland and Pyle (1977), p. 371. Cf. Ross (1977a). As already discussed, family firms and firms with high levels of insider ownership can be assumed to be very similar. See section 4.2. Anderson, Mansi and Reeb (2003), p. 267. See Anderson, Mansi and Reeb (2003), pp. 266-268. See table 6-9. See Norton (1991), pp. 290-291; Myers (1984), pp. 581-585; Harijono and Tanewksi (2004), pp. 7-8. Cf. Kim and Sorensen (1986). Cf. Graham and Harvey (2001), pp. 209-232. This is also in line with the pecking order theory of Myers (1984), pp. 581-585.

Hypotheses on Insider Ownership

137

have incentives to grow their firms beyond the optimal size (“overinvestment”), “[. . . ] debt creation, without retention of the proceeds of the issue, enables managers to effectively bond their promise to pay out future cash flows.”541 Thus, firms with high levels of insider ownership, where managers take both the role of agents and principals, are likely to adopt lower debt levels as compared to firms with low levels of insider ownership. Alternatively, one could argue that insiders with significant ownership stakes might prefer more discretionary freedom, which is narrowed by the disciplining effect of debt, and hence adapt lower debt levels. Another noteworthy aspect is the effect of leverage on risk. As debt levels increase, the claims attached to shares become more risky and insiders, owning large stakes in their companies, have to bear relatively more risk regarding their investment in their firms, i.e. there are also costs to increasing debt ratios. Recalling that owner-managers usually hold a rather undiversified portfolio, there are clear incentives for them to install lower leverage ratios.542 Based on all these arguments the following hypothesis is suggested:543 Hypothesis 5.2.7 The level of insider ownership has an effect on the level of firms’ leverage. Other potential effects of insider ownership on the corporate control setting as proposed by existing literature, e.g. corporate takeovers, are not examined since reliable data is limited for German firms and, consequently, related hypotheses could not be tested in the empirical investigation.544 5.3

Insider Ownership and Corporate Performance

In their pathbreaking work J ENSEN AND M ECKLING (1976) hypothesize that as insider ownership declines, agency costs, i.e. the costs of deviation from value-maximization, rise since managers only have to bear a disproportionately small share of these costs. As a consequence, managers might be less vigilant 541 542 543

544

Jensen (1986), p. 324. Cf. Jensen, Solberg and Zorn (1992); Friend and Lang (1988); Stulz (1990). See Stulz (1988), p. 41; McConaughy, Matthews and Fialko (2001), p. 36. For the sake of completeness, it should be noted that there exist arguments which deny a direct relationship between insider ownership and leverage but claim that both share common determinants instead. See Chaplinski and Niehaus (1993); Crutchley and Hansen (1989). See amongst others Williams (1990); Grossman and Hart (1980); Brickley, Lease and Smith (1988); Weston, Siu and Johnson (2001).

138

Hypotheses on Insider Ownership

in dealing with corporate resources, focus on the consumption of perquisites or engage in shirking. To turn the argument on its head, that means the conflict between management and shareholders and, hence, possibly harmful effects on shareholder value and corporate performance, can be moderated by a higher participation of the company’s management in its own company.545 Referring to managers as agents and shareholders as principals,546 J ENSEN AND M ECK LING (1976) put it in the following way: “[. . . ] in some situations it will pay the agent to expend resources (bonding costs) to guarantee that he will not take certain actions which would harm the principal [. . . ].”547 According to this view, management’s investment in its own company constitutes a bonding action, to signal shareholders their commitment to the firm. Later on, M ORCK , S HLEIFER AND V ISHNY (1988) pick up the idea that market value should increase along with management ownership and call it the convergence of interest hypothesis.548 Empirical studies examining either the direct relationship between the equity share of owner-managers and agency costs549 or the indirect relationship to performance are numerous and will be presented later on in section 4.2.2. An alternative theory of the relationship between insider ownership and firm valuation, which focuses on the takeover process, is provided by S TULZ (1988). According to this theory, potential acquirers must pay higher control premiums for firms with higher levels of insider ownership, since insiders are more likely to refuse tendering shares because of the private benefits of control they derive from their current employment. At low insider ownership levels this might increase firms’ ex-ante value as potential targets.550 Even though S TULZ 545 546 547 548 549

550

See Jensen and Meckling (1976), pp. 308-330; Morck, Shleifer and Vishny (1988), p. 293. See section 4.1.2.4. Jensen and Meckling (1976), p. 308. See Morck, Shleifer and Vishny (1988), p. 293. Ang, Cole and Lin (2000), p. 82, report that agency costs are higher in firms not solely owned by their managers and that these costs rise as the level of managerial ownership falls. Specifically, two conflictive effects of the level of insider ownership (α) are hypothesized by Stulz (1988), p. 26: “On the one hand, a higher α adversely affects the value of the target as it decreases the probability of a hostile takeover attempt. On the other hand, the premium offered if such an attempt is made increases with α. [. . . ] therefore, [. . . ] the value of the firm increases or falls when α increases depending on whether α is small or large, so that there is a unique value of α that maximizes the value of the firm.” Cf. Mikkelson and Partch (1989) whose empirical analysis shows that the likelihood of successful acquisitions of firms is unrelated to the level of insider ownership.

Hypotheses on Insider Ownership

139

(1988)’s prediction of a positive relationship between insider ownership and firm value at lower insider ownership levels is in line with the argumentation of J ENSEN AND M ECKLING (1976), it is worthwhile to note the different explanations: The latter argue in favor of converging interests between management and shareholders while the former claims that management’s more effective opposition to takeovers leads to higher firm valuations.551 If insiders own significant shares of their firm, they often become relatively undiversified as both their human capital and large parts of their financial capital are invested into a singe company. The firm’s investments will be chosen not to endanger the owner-manager’s personal wealth and usually a long-term horizon is used in evaluating investment decisions. Hence, also investment projects which become valuable not until a decade can be pursued and the yearly risk declines.552 As already discussed, such long-term investments are likely to be left aside by managers in widely held public stock corporations.553 Thus, the following hypothesis is suggested: Hypothesis 5.3.1 The level of insider ownership has an effect on corporate performance (“Convergence of Interest Hypothesis”). At low insider ownership levels, market discipline (e.g. the product or managerial labor market)554 might suffice to force management to maximize shareholder value. Contrarily, at high insider ownership levels management might own enough shares and voting power to ensure employment with the firm at attractive terms even if they do not always act in the best interest of the shareholders.555 In this context, K ASERER AND WAGNER (2004) show that for the case of Germany standardized management board compensation is significantly positive influenced by the free float portion which might indicate a lack of control.556 Remarking that managers with virtual unlimited control over their 551

552 553 554 555

556

See Stulz (1988), pp. 25-27; Harris and Raviv (1988a), pp. 55-75. Cf. Morck, Shleifer and Vishny (1988), p. 313; Hubbard and Palia (1995), p. 791; Roosenboom and Schramade (2005), p. 271. Apart from agency theory, in the context of family firms sometimes the unique resource pool (“familiness”) is mentioned as a source of better performance. See Habbershon, Williams and MacMillan (2003); Chrisman, Chua and Litz (2003). See B¨uhler, Fueglistaller and Zellweger (2005), pp. 16-17. See section 3.4.2.1 where the time preference problem was outlined. Cf. Hart (1983), pp. 366-382; Fama (1980), pp. 288-307. See Belkaoui and Pavlik (1992), p. 345. For example, Weston (1979) notices that no firm with an insider ownership level above 30% had ever been acquired in a hostile takeover. Cf. Qiu (2003), p. 13. See Kaserer and Wagner (2004), pp. 15-16.

140

Hypotheses on Insider Ownership

company might be primarily guided by their own preferences at the expense of value-maximizing behavior, M ORCK , S HLEIFER AND V ISHNY (1988) formulate the so-called entrenchment hypothesis which “[. . . ] predicts that corporate assets can be less valuable when managed by an individual free from checks on his control.”557 The authors concede that it is hard to predict which of the two effects, i.e. convergence of interest vs. entrenchment, dominates but presume, that entrenchment might prevail “[. . . ] for some range of high [insider] ownership stakes.”558 While J ENSEN AND M ECKLING (1976) derive a monotonic, positive relationship between insider ownership and firm value in their formal analysis, S TULZ (1988)’s theory predicts that above a certain, unique value-maximizing level of insider ownership, further stakes owned by management might be detrimental to firm value. The explanation for this is that the positive effect of a higher takeover premium mentioned earlier is more than offset by a decreasing probability that a corporate control event will actually take place at high insider ownership levels.559 Another objection is raised by S ANDERS (2001) who claims that insider ownership could rather increase than decrease agency costs as company executives could avoid necessary risks because of asymmetrical risk properties of stock ownership compared to stock option pay.560 Alternatively, not ownership per se might effect firm value negatively above a certain threshold but the extent to which managers are able to control the (supervisory) board at discretion (which in turn would depend on its shareholdings). For example, the positions of chairman of the board and CEO are combined frequently in U.S. corporations. As a consequence, monitoring on the part of the 557 558

559 560

Morck, Shleifer and Vishny (1988), p. 294; see also pp. 293-294. Morck, Shleifer and Vishny (1988), p. 294. Hubbard and Palia (1995), pp. 784-786, explain this relationship with the “diversification-control hypothesis” which states that managers of acquiring firms are likely to overpay at rather low and rather high levels of managerial ownership: At low levels of insider ownership, high bid premiums are accepted to enable additional perquisite consumption. At high levels of insider ownership gains from diversification and the acquisition of private benefits of control increase the bid premium (inverse bell-shaped relationship). See footnote 555. Cf. Williams (1990), pp. 1638-1639; Qiu (2003), p. 25. See Sanders (2001), p. 489. Nevertheless, the author concludes that stock options are rather “[. . . ] motivational ‘carrots’ that lack the complementary disciplinary ‘stick’ [. . . ]” and hence may not serve as an adequate substitute for stock ownership from a corporate governance/incentive based view.

Hypotheses on Insider Ownership

141

board can be expected to suffer under this constellation.561 Thus, the following hypothesis is proposed: Hypothesis 5.3.2 The level of insider ownership above a certain threshold has an effect on corporate performance (“Entrenchment Hypothesis”). As reported by G OMPERS , I SHII AND M ETRICK (2004), one constraint of the above named hypotheses is that the underlying separate forces, i.e. incentives and control rights stemming from equity ownership, are usually aggregated in only one variable, namely the level of insider ownership.562 However, if the link between cash flow incentives and voting control is broken, e.g. through the issuance of voting common shares and non-voting preference shares, both effects can be disentangled. Hence, insiders’ control rights refer only to their stake in voting shares while insiders’ cash flow rights refer to their weighted stake of both voting and non-voting shares.563 If management owns excessive control rights, i.e. control rights above its cash flow rights, agency costs and the expected consumption of private benefits are expected to be especially large since the potential for entrenchment rises with increasing control rights while the incentives from cash flow ownership comparatively fall.564 H ARVEY, L INS AND ROPER (2004) refer to this as “cash flow rights leverage”.565 In the context of IPOs, ROOSENBOOM AND S CHRAMADE (2005) claim that management’s post-IPO cash flow ownership is one effective bonding mechanism to mitigate agency problems. Contrary, high post-IPO control rights by the owner-managers will lead investors to anticipate larger agency costs and, hence, to demand larger discounts because “[a] controlling owner may treat himself preferentially at the expense of other investors, especially if his voting power is significantly larger than his cash flow ownership.”566 561 562 563

564

565 566

See Faccio and Lasfer (1999), pp. 10-11. See Gompers, Ishii and Metrick (2004), p. 2. Prior U.S. evidence reported by DeAngelo and DeAngelo (1985), p. 45, shows that in dual-class companies officers own nearly triply as much shares in the superior voting stock class (59.2%) compared to the inferior voting stock class (20.8%). See Lins (2003), p. 160. Cf. Claessens, Djankov, Fan and Lang (2002), pp. 2741-2744; Barontini and Caprio (2005), pp. 14; Harris and Raviv (1988b). Harvey, Lins and Roper (2004), p. 7. Roosenboom and Schramade (2005), p. 273; see pp. 271-273.

142

Hypotheses on Insider Ownership

Similarly, B EBCHUK (1999), developing a rent-protection theory of corporate ownership, states that owner-managers may keep a lock on control after an IPO if private benefits of control are large because otherwise rival management would be tempted to gain control over the company to extract these large private benefits of control. Moreover, this lock on control is even more favorable for the owner-manager because it reduces his costs of holding a large fraction of the cash flow rights, which usually implies that his personal wealth becomes rather undiversified.567 However, since in such settings agency costs can be expected to be very large, the following hypothesis is suggested: Hypothesis 5.3.3 The difference between insiders’ control and cash flow rights has an effect on corporate performance. Finally, one should refer to those who deny any causal relationship between insider ownership and corporate performance. Most notably, D EMSETZ (1983) hypothesized that the ownership structure might be an endogenous outcome of a value-maximizing equilibrium process (“natural selection hypothesis”). Later on, several empirical tests, e.g. by D EMSETZ AND L EHN (1985) or H IMMEL BERG , H UBBARD AND PALIA (1999), seem to confirm the assumption that insider ownership and corporate performance share common determinants and, hence, any observed relation might be spurious.568 While these results almost exclusively rely on studies about the U.S. market, it is argued that the situation for Germany is different and endogeneity of insider ownership appears less plausible.569

567 568

569

See Bebchuk (1999), pp. 25-27. See Demsetz (1983), pp. 377, 383-390; Demsetz and Lehn (1985), pp. 1173-1176; Himmelberg, Hubbard and Palia (1999), pp. 354-356. See section 6.9.4.2.

Empirical Analysis of Insider Ownership

6

143

Empirical Analysis of Insider Ownership in German Public Stock Corporations

Having presented the underlying theory and an overview of existing empirical research on the role and the effects of insider ownership, this section shows the results of an empirical examination of the German public equity market during the ten-year period ranging from 1993 to 2003. The explicit and extensive coverage of insider ownership in Germany in the descriptive statistics and regression models constitutes the main novelty and distinguishing feature of this empirical corporate governance analysis. 6.1

Study Design

It has become prevailing opinion that corporate governance studies, especially those dedicated to the partial analysis of selected corporate governance mechanisms, at the current state of research primarily make sense as single country studies. The differences in corporate governance systems are too large570 and the distortions from distinctions in data quality and availability across countries are immense.571 Given the existing research gap concerning insider ownership in Germany (see section 4.2.4), this study targets exclusively the German market.572 Even though it is widely acknowledged that the multiple governance mechanisms are highly interrelated573 , a partial analysis of insider ownership as one selected mechanism will be performed as more research on the individual mechanisms is needed before an integrated approach appears sensible. Furthermore, it is limited to the case of publicly listed companies where the separation 570

571

572

573

See e.g. Seifert, Gonenc and Wright (2005) who hint to the problems involved in an empirical comparative corporate governance study of the U.S., the U.K., Germany and Japan. Finally, they come to the conclusion that their results of the individual countries are not apt to make comparisons among them. Furthermore, La Porta, Lopez-de Silanes, Shleifer and Vishny (2002), pp. 1160-1161, show that valuation levels differ significantly between civil law and common law countries. Denis and McConnell (2003), p. 18, survey international corporate governance literature and conclude with respect to the role of ownership structures: “The evidence from around the world indicates that the relation between ownership structure and firm performance varies — both by country and by blockholder identity.” Cf. Berrar (2001), p. 41; Mayer (1998); Krivogorsky (2006). For differences in public access to ownership information among European countries see Becht (1997), pp. 40-42. In the following the “German market” is defined by two selection criteria: First, referring only to listed firms (see footnote 574), companies must be listed in the official or regulated market of the Frankfurter Wertpapierb¨orse (“Frankfurt Stock Exchange”). Second, companies must also be incorporated in Germany, hence, excluding foreign companies with a (second) listing in Germany. See section 6.2.1. See Agrawal and Knoeber (1996); Bhagat and Jefferis (2002); Schmid (2003); Beiner, Drobetz, Schmid and Zimmermann (2006); Berry, Fields and Wilkins (2006).

144

Empirical Analysis of Insider Ownership

of ownership and control is most obvious and agency problems are expected to be largest.574 The first part of the empirical study performs a descriptive analysis of shareholder structures — with a special focus on insider ownership — of German listed corporations. Thereby an insight on the actual constellations of interests in German stock corporations can be gained.575 Shareholder structures shape the variety and relative weight of potentially diverse interests and determine the control potential of individual owner groups. The aim of the second part of this study is to test the hypotheses developed in section 5 on the role of insider ownership as a corporate governance mechanism in the German market.576 Generally, most corporate governance studies in this field fall into one of the following two main categories: event studies as well as cross-sectional and panel data. The former focus on the dynamic aspects of ownership, i.e. changes in ownership (“the event”) and examine the effects on stock prices after such events.577 The latter use more static data to usually explain corporate performance or success in a regression framework. Given the lack of studies on insider ownership in Germany, no standard methodology has evolved yet. Another widespread type of ownership studies refers to the IPO setting and analyzes changes in ownership after the IPO and corresponding performance effects.578 This study will apply regression analyses based on cross-sectional and panel data for a variety of reasons: First, the majority of international studies on insider ownership uses regression analyses hence making the results of this study 574

575

576 577

578

As discussed in chapter 3, the separation of ownership and control in public companies is the main source of agency problems making corporate governance mechanisms more important. Closely-held private companies by definition are largely untroubled by those problems and, hence, are not subject of this study. Cf. Schmidt, Drukarczyk, Honold, Prigge, Sch¨uler and Tetens (1997), p. 47. In contrast the INTES — Komission Governance Kodex f¨ur Familienunternehmen (2004), p. 4, states that corporate governance does not play a minor role in non-listed companies. Of course, the aggregation across companies in the calculation of descriptive statistics (e.g. average shareholding of a certain group) causes a loss of information. In contrast case studies which analyze in detail single corporations are adequate for the exploration of ideas and hypotheses. Since the theoretical framework concerning insider ownership is already highly developed (see chapters 3 and 4) the verification of such hypotheses can only be carried out on the basis of a large representative sample. See Bott (2002), p. 293. For a compact methodical review of event studies see Weston, Siu and Johnson (2001), pp. 171-185; Bodie, Kane and Marcus (1999), pp. 338-341. The concept of event studies was initially developed by Fama, Fisher, Jensen and Roll (1969) in their seminal article The Adjustment of Stock Prices to New Information. See Ehrhardt and Nowak (2003a); Kuklinski, Lowinski and Schiereck (2003); Jaskiewicz, Gonzal´ez, Men´endez and Schiereck (2005).

Empirical Analysis of Insider Ownership

145

comparable to existing research. Second, ownership data are rather static in Germany and the quality of data does often not allow to specify a clear event date.579 Third, the dynamic aspect of ownership can — and will — also be integrated into a regression framework by including changes of ownership as an explanatory (independent) variable. Fourth, changes in ownership structure may come along with strategic changes which also might influence company performance. Last and maybe most important, companies experiencing large ownership events are certainly not a random sample.580 The applied regression models are geared at existing (predominantly U.S.) studies applying modifications due to the peculiarities of the German corporate governance system.581 Instead of examining a single cross section of German listed corporations, this study consists of three cross sections covering the ten-year period between 1993 and 2003. This offers at least three advantages over the use of a single cross section: First, the German public equity market has experienced serious changes during the last decade. While only a minority of the German population directly or indirectly owned stocks up to the mid nineties and the universe of listed firms was rather small, the stock market has catched up during the last decade.582 It will be interesting to observe whether and how this development will be reflected in a comparison between 1993 and 2003 data. Second, and associated with the first argument, the quality and quantity of available shareholder structure data increased. Therefore, the comparison of 1993 and 2003 data will show whether the increased transparency yields new insights into the nature of German shareholder structures. And third, a longitudinal approach also allows to analyze changes in insider ownership over time and the use of 579

580

581

582

According to §§ 21, 25 WpHG up to 16 calendar days can pass until an obligatory announcement over certain changes in the shareholder structure is to be published in a newspaper authorized by the German Stock Exchanges. Therefore, it can be assumed that the actual announcement date is hard to detect and, hence, any measured announcement effect is likely to be diluted. See Thomsen, Pedersen and Kvist (2006), pp. 250-251. For example, Filbeck and Chadwell (1993) analyze insider ownership in the context of dividend increases in the banking industry. Because of the very particular study object, the findings should only interpreted with caution. For a comprehensive confrontation of event studies and cross-sectional and panel data see Bott (2002), pp. 239-298. Bott (2002) argues in favor of event studies to examine performance effects of certain ownership characteristics but her empirical study largely fails to detect significant abnormal returns which are dependent on the identity of the selling and buying shareholder respectively. It is often stated that the partial privatization of “Telekom AG” (the former state-owned telecom monopolist) in 1996 was the starting signal for an evolving equity culture in Germany. In fact the number of direct and indirect share owners (i.e. owners of investment funds) rose by 99% between 1997 and 2003. See Deutsches Aktieninstitut e.V. (DAI) (2006), 08.3-Zahl-D; Deutsches Aktieninstitut e.V. (DAI) (2000), p. 2. Cf. footnote 82.

146

Empirical Analysis of Insider Ownership

lagged variables opening up new research perspectives. However, instead of collecting data for all the years, only three cutoff dates, i.e. the year-end of 1993, 1998 and 2003, will be examined. This proceeding is attributed to the fact that ownership structures in Germany are traditionally quite stable and the lack of transparency does not allow to detect minor changes respectively.583 Therefore, the applied method could also be described as an unbalanced low frequency panel approach since the same corporations are examined over time, a procedure that will be more clearly described in the following section 6.2.1.584 Given the uncertainty regarding the correct specification of regression models linking ownership to performance and the adequate specification of insider ownership, a rather broad approach is taken applying several variations in the specification of the regression models and various robustness tests. While the robustness tests are supposed to make the tentative predications of the individual regression models more convincing, some limitations of the study will persist. These limitations will be outlined later on in section 6.10. The complexity of and problems involved in empirical corporate governance studies are admittedly non-trivial and the conclusions and interpretations deserve a cautious view on the methodical problems involved.585 6.2

Sample and Data

6.2.1

Sample Selection

The universe for this study are all publicly listed German companies. In Germany, companies can strive for a listing at six different stock exchanges named according to their domiciles in Berlin-Bremen, D¨usseldorf, Hamburg-Hanover, Munich, Stuttgart and Frankfurt. While the former five are often referred to as regional stock exchanges, the latter Frankfurter Wertpapierb¨orse (FWB) (“Frankfurt Stock Exchange”) is by far the most important market place in Germany. It is operated by the private company Deutsche B¨orse AG and, includ583

584

585

Reviewing the study of Himmelberg, Hubbard and Palia (1999), Zhou (2001), p. 560, concludes that “[i]n panel data with firm fixed-effects it would be hard to find a meaningful relationship between ownership and performance even if one existed [. . . ]” because ownership usually changes slowly from year to year within a company. As will be shown, certain characteristics of the sample data limit the effectiveness of panel approaches in the field of shareholder structure research. Cf. section 6.10. For a comprehensive overview of the econometric problems involved in corporate governance studies see B¨orsch-Supan and K¨oke (2002).

147

Empirical Analysis of Insider Ownership Transparency Standard / Market Segment

General Standard

Transparency Standard Prime Standard

Both Standards

Panel A: Share classes from all companies (domestic and foreign) Official market (“Amtlicher Markt”) Regulated Market (“Geregelter Markt”) Regulated unofficial market a (“Freiverkehr”) Total

177

252

429

243

208

451

-

-

6,130

420

460

7,010

183

172

355

227

175

402

-

-

191

410

347

948

Panel B: Share classes from German companies Official market (“Amtlicher Markt”) Regulated Market (“Geregelter Markt”) Regulated unofficial market a (“Freiverkehr”) Total a

No differentiation between General and Prime Standard since membership in these transparency standards requires a listing in either the official or the regulated market. Source: Author’s illustration. Data from the website of Deutsche B¨orse (www.deutsche-boerse.com as of 09.09.2005).

Table 6-1: Listings at the Frankfurt Stock Exchange 2005 ing its electronic trading platform XETRA, it cumulates a 97% market share in the dealing of domestic stocks (July 2005).586 Regarding its dominant role, the sample population was restricted to companies listed at the FWB.587 Furthermore, companies had to be listed in one of the two most regulated segments either the official market (“Amtlicher Markt”) or the regulated market (“Geregelter Markt”) hence excluding companies from the regulated unofficial market (“Freiverkehr”), where primarily foreign companies are listed. As can be seen in table 6-1, a distinct majority of roughly 79.9% (= (355+402)/948) of all German share classes listed at the FWB are listed in the official or regulated market. 586 587

See Deutsche B¨orse (2005b). Since only small companies are solely listed at one or more of the regional exchanges but not at the FWB, this shifts the study focus slightly towards larger listed companies. In 1999, of the 1,043 domestic listed companies 795 (or 76.2%) were listed at the FWB. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 02-2.

148

Empirical Analysis of Insider Ownership 1993

1998

2003

Number of share classes represented in the CDAX as of 31.12.YYYYa

380

520

719

Number of dual listings (i.e. more than one share class is listed)

42

60

38

Number of companies represented in the CDAX as of 31.12.YYYY

338

460

681

Number of financial companies (i.e. investment companies, mortgage banks, credit banks, insurance companies)

63

80

29

Number of non-financial companies represented in the CDAX as of 31.12.YYYY

275

380

652

Number of companies without continuous listing over at least one five year period (1993-1998, 1998-2003)

39

22

362

Total number of sample companies

236

358

290

a

The CDAX includes the shares of all domestic companies listed in Prime Standard and General Standard. The index represents the German equity market in its entirety, i.e. all companies listed on the FWB. Source: Author’s illustration. CDAX data from Deutsche B¨orse.

Table 6-2: Sample Selection 1993-2003

All German share classes listed in the official or regulated market at the FWB are included in the so-called Composite-DAX (CDAX) stock index (see last column of Panel B in table 6-1). Therefore, the CDAX at the end of 1993, 1998 and 2003 represents the basis for the sample selection process. As shown in table 6-2, from the total number of share classes, the number of secondary share classes (or dual-share classes) had to be subtracted to derive the number of listed companies. Secondary share classes usually exist if a company has issued both voting ordinary and non-voting preference shares.588

588

Note that German preference shares are not completely comparable to U.S. preference shares. For a description of the characteristics of German preference shares see Daske and Ehrhardt (2000), pp. 2-7.

Empirical Analysis of Insider Ownership

149

Then, all financial companies, i.e. investment companies, mortgage banks, credit banks and insurance companies, were deducted because of their limited comparability to non-financial firms.589 From the remaining non-financial companies all companies that were members of the CDAX during at least two out of the three cutoff dates (end of December 1993, 1998 and 2003) were selected. This restriction was introduced to enable the analysis of ownership structures over time for at least one five-year period.590 The remaining number of nonfinancial companies, 236 (1993), 358 (1998) and 290 (2003) is used for the descriptive statistics on shareholder structures in section 6.4.1. The effects of this exclusion restriction are illustrated in figure 6-1. A total number of 168 companies is represented in all three sub-samples (1993, 1998 and 2003). Another 68 firms are members of the CDAX at the end of 1993 and 1998 but ceased to be listed in 2003. Similarly, 122 companies were CDAX members at the end of 1998 and 2003 but not at the end of 1993; the majority of these companies had their initial public offering during the period ranging from 1993 to 1998 even though a small number has experienced changes in the respective stock market segment, e.g. from the unregulated to the regulated market.591 In the regression analyses, the sample size had to be further reduced due to a lack of accounting and market data. This resulted in final sample sizes of 132-133 (1993), 212-220 (1998) and 235-245 (2003) companies for the three different cross-sections and a pooled cross-section consisting of 579-598 firm years.592 The sample captures between 36.0%-37.6% (2003), 48.0%-48.4% 589

590

591

592

This restriction is applied in most empirical studies. See Morck, Shleifer and Vishny (1988), p. 295; McConnell and Servaes (1990), p. 600; Schmid (2003), p. 10; Gugler, Mueller and Yurtoglu (2004), p. 10. This criterion might introduce a possible sample selection bias into the analysis. The findings of Iber (1987), p. 176, indicate that the probability of delistings for family firms is comparatively low and, hence, any bias should not be severe. Furthermore, Jostarndt (2006), p. 21, reports that management’s ownership stake around financial distress events remains rather constant. Therefore, the level of insider ownership in case of bankruptcy should not be systematically different from the usual level. Possible effects will be discussed in section 6.10. Similarly, K¨oke (2004), p. 60, also selects sample companies which have at least four years of continuous data (sample II) and finds that for listed firms the only systematic difference compared to a sample selection without this restriction (sample I) is related to firm size measured by total assets. Firm size is larger for sample I, i.e. smaller firms appear to be more often subject to delistings, acquisitions or bankruptcies. The final sample size varies for the individual analyses according to the varying availability of the three different (dependent) performance variables. The numbers here refer to the base case regression models (see section 6.9.3). As discussed in section 6.1, the sample can also be described as an unbalanced low frequency panel.

150

Empirical Analysis of Insider Ownership

400

Number of sample companies

358

290

300

122

236 122

200

68

68

168

168

168

1993

1998

2003

100

0

Group: 1993, 1998 and 2003

Group: 1993 and 1998 only

Group: 1998 and 2003 only

Source: Author’s findings.

Figure 6-1: Sample Composition 1993-2003 (1993) and 55.8%-57.9% (1998) of all non-financial CDAX companies. The comparably small percentage in 2003 is attributed to the IPO-Boom during the new economy (1999-2001). These newly listed companies are not included in the sample because of the mentioned exclusion restriction. Consequently, the sample rather represents the more traditional listed companies and not the companies which went public during the heyday of the new economy. 6.2.2

Data and Sources

Data on insider ownership and shareholder structures in Germany, which constitute the center of this empirical study, are rare and fairly troubled. The absence of extensive disclosure provisions concerning insider ownership — comparable to the Williams Act in the U.S.593 — make empirical studies for Germany difficult and less informative respectively.594 This study is based on shareholder data published in the Hoppenstedt Aktienf¨uhrer (“Hoppenstedt Stock Guide”) which 593

594

The Williams Act of 1968 is a federal law that amended the Securities and Exchange Act of 1934. It requires mandatory disclosure of information regarding cash tender offers and the filing and public disclosure with the SEC if anyone acquires more than 5% of the outstanding shares of a listed corporation. See Becht and B¨ohmer (2003), pp. 1, 4-7; Becht (1997), pp. 28-34.

Empirical Analysis of Insider Ownership

151

is released on a semiannual basis. The first issues of 1994, 1999 and 2004 are used to survey the shareholder structure data on all sample companies at the end of 1993, 1998 and 2003 respectively.595 The Hoppenstedt Aktienf¨uhrer asserts to make use of all publicly available information concerning shareholder structures — and other corporate, accounting and stock market data as well — for the majority of all publicly listed companies in Germany and the entire universe of those listed in the exchanges’ official and regulated markets.596 The nature of the original sources of such ownership data gives rise to diverse intricacies and problems, which limit their usefulness and ask for a cautious interpretation of all results based on those data; especially, but not only in a scientific context.597 Hence, only the key characteristics of the most important sources of ownership information will be summarized shortly.598 In Germany, several statutory provisions exist side by side which regulate the disclosure of ownership stakes of stock corporations. Even if publicly listed stock corporations were willing to completely disclose their shareholder structures, most of them would not be able to do so since so-called Inhaberaktien (“bearer stocks”) are far more widespread than so-called Namensaktien (“registered shares”) in Germany.599 As bearer stocks can be — and usually are — sold without notification of the company, the public firms normally do not know the identity of (all) their shareholders. Consequently, legal regulations are necessary to force the owners of shares — at least beyond certain thresholds — to reveal their shareholdings. Those regulations can be found in a variety of acts in the fields of corporate law (e.g. the Stock Corporation Act), accounting law,

595

596

597

598

599

Note that up to 1996 the Hoppenstedt Aktienf¨uhrer was named “Saling Aktienf¨uhrer”. In the rare cases where no shareholder structure data were available in the 1994-I, 1999-I and 2003-I editions of Hoppenstedt Aktienf¨uhrer ownership data were taken from the previous or next issue. Data, especially in the earlier 1994-I issue, are often imprecise. For statements as “above (below) 25%” the relevant percentage is increased (decreased) by 0.01% to 25.01% (24.99%). For statements as “around” or “approximately” the given percentage is used. See Daske and Ehrhardt (2000), p. 23. The limitations of this study caused by poor quality and insufficient availability of shareholder data will be explained later on in section 6.10. For a detailed analysis of the explanatory power of information concerning shareholder structures in Germany see Bott (2002), pp. 172-240. She presumes that the lacking availability of accurate shareholder structure data might represent a bottleneck for empirical research in this field. For example, in the 2003 sample of this study, out of the 270 listed ordinary share classes only 6.7% were registered shares the rest being bearer stocks. For a description of both stock types cf. Achleitner and Thommen (2003), pp. 504-506.

152

Empirical Analysis of Insider Ownership

stock exchange law, competition law and most important capital market law.600 In capital market law the Wertpapierhandelsgesetz (WpHG) (“Securities Trading Act”) contains several provisions regarding the disclosure of voting rights. Most important §§ 21-29 WpHG, which regulate the duty of notification of the share owner and the duty of publication of the notified company, became effective as of January 1, 1995 after the “European Council Directive 88/627/EEC of 12 December 1988 on the information to be published when a major holding in a listed company is acquired or disposed of” (“Transparency Directive”) required EU member countries to establish national disclosure standards concerning changes in shareholder structures.601 Everybody whose voting rights in a publicly listed company exceed, reach or fall below the thresholds of 5%, 10%, 25%, 50% and 75% must notify the listed company as well as the BaFin within at most seven calendar days about the size of his precise voting stake according to § 21 WpHG.602 Besides direct share ownership, § 22 WpHG requires to include the shares of certain related parties (e.g. controlled subsidiaries or family members with pooled-voting) in the calculation of the size of the relevant voting block. According to § 25 WpHG, the listed company must publish this notification within a period of at most nine calendar days in a newspaper authorized by the German stock exchanges. Changes within any interval marked by two adjacent thresholds fall not under these regulations and are usually not disclosed. Consequently, the inaccurateness of registered blockholdings is likely to increase over time as long as no other relevant threshold is passed. Furthermore, since July 1, 2002 insiders are obliged to notify their company and the BaFin about dealings in their company’s

600

601

602

See Bott (2002), p. 230; Becht and B¨ohmer (1999), pp. 26-37. For the effects which the choice of different data sources on share ownership (in the U.S.) can have on the results in empirical studies see Kole (1995); Anderson and Lee (1997). Note that §§ 21-29 WpHG refer only to the disclosure of voting rights and not share capital stakes which might differ for companies which also have non-voting preference share capital outstanding. Cf. Becht (1997), pp. 114-118. The original thresholds defined in the EU Directive are 10%, 20%, 33%, 50% and 66% but the 20% and 33% thresholds might be substituted by a 25% threshold. Cf. Van der Elst (2000), p. 15. A current draft of the Transparency Directive Ratification Act (“Transparenzrichtlinie-Umsetzungsgesetz” (“TUG”)) proposes to implement 3%, 15%, 20% and 30% as additional thresholds. See section 2.4.2.

Empirical Analysis of Insider Ownership

153

stocks or derivatives based on those stocks within five working days according to §15a (1) WpHG.603 Before 1995, shares of the capital stock — not voting rights — only larger than 25% had to be disclosed under corporate law (§§ 20-21 AktG).604 Therefore, this change in regulation is also expected to cause systematic differences between the 1993 and 1998 or 2003 samples. The decrease of the disclosure threshold increased significantly the transparency of voting blocks below the blocking minority but also limits the comparability of the three cross sections.605 These regulations largely determine the scope of publicly available information on shareholder structures. Since ownership data in the Hoppenstedt Aktienf¨uhrer are not limited to voting rights disclosed due to WpHG, also data on total ownership on the capital stock (so-called cash flow rights) are available. As presumed, the number of known shareholders increased considerably during the sample period — by 75.0% between 1993 and 2003 — due to changes in disclosure legislation as it is shown in table 6-3. Nevertheless, the mean number of known shareholders (or better registered blockholders) of 1.6 in 1993 and 2.8 in 2003 still indicates a low degree of transparency in shareholder structures of German publicly listed companies.606 Besides shareholder structures, further corporate governance related firm specific information was extracted from the Hoppenstedt Aktienf¨uhrer: With respect to the supervisory board, the number of members, the number of employee representatives, and the nationality of the chairman (as a dichotomous variable German/ non-German) were researched. Similarly, the number of members of the management board was surveyed. Finally, possible restrictions on voting rights, i.e. any deviation from the one-share-one-vote principle were recorded.607 603

604

605 606

607

According to § 15a (2, 3) WpHG, insiders are defined as members of the management and supervisory board and other individually liable shareholders. The duty of notification applies also to spouses, civil unions and first-grade relatives. Besides the higher initial threshold (25% vs. 5%), only a certain group, namely enterprises, had to disclose their ownership of German listed corporations. See Becht and B¨ohmer (2003), pp. 15-26; Bott (2002), p. 198. These observations are roughly in line with Becht and B¨ohmer (2003), p. 7, who state that in 1996 only 20% of firms have more than two registered blockholders. Besides the issuance of non-voting preference shares, limitations of the exercise of voting rights (e.g. at most 20% of capital stock) or multiple voting rights for certain share classes (eventually limited to decisions concerning certain provisions of the company’s articles) are common forms of deviation from the one-share-

154

Empirical Analysis of Insider Ownership 1993

1998

2003

Total number of sample companies

236

358

290

Total number of shareholder entries in Hoppenstedt Aktienf¨uhrer

605

1.130

1.113

Average number of shareholder entries per sample companya

2.6

3.2

3.8

Average number of shareholder entries per sample company without free-float portionb

1.6

2.2

2.8

a b

Including a mandatory free-float portion for each company so that shareholdings add up to 100%. Excluding the mandatory free-float portion for each company. Calculated as ”Average number of shareholder entries per sample company” less 1. Source: Author’s illustration.

Table 6-3: Shareholder Structure Data 1993-2003

From the data in table 6-4 some clear trends in the corporate governance setting of the sample companies can be inferred. With respect to the supervisory board, the average number of supervisory board members decreased from 10.9 in 1993 to 8.6 in 2003. This might reflect a reaction to recent criticism which attributed the lack of effective control through the supervisory board partially to oversized and hence ineffective supervisory boards.608 Furthermore, an even larger decrease in the number of labor representatives in the supervisory board can be observed. As a result, the ratio of labor representatives to all supervisory board members shrinks from 40.4% in 1993 to 38.4% in 2003. Acknowledging that this ratio is mainly prescribed by codetermination laws609 , this development might be an outcome of a changing industry distribution of the sample companies.610 As a second trend, the internationalization of German corporations becomes apparent by locking at the ratio of foreign supervisory board chairmen

608 609 610

one-vote principle in Germany (e.g. Volkswagen AG, Lindner Holding KGaA, CEWE Color Holding AG or Muehlbauer Holding AG.). See Blum (2005), pp. 157-160; Baums (2003), pp. 174-175; Cromme (2002b), pp. 123-124. See section 2.4.3. As can be seen in table 6-18, e.g. the portion of — usually codetermined — utilities companies among the sample companies decreased from 10.4% in 1993 to 7.0% in 2003.

155

Empirical Analysis of Insider Ownership 1993

1998

2003

Total number of sample companies

236

358

290

Average number of supervisory board members (Mode)

10.9 (6)

9.1 (6)

8.6 (6)

Average number of labor representatives in the supervisory board (Mode)

4.4 (2)

3.5 (0)

3.3 (0)

Average ratio of labor representatives to all supervisory board members in % (Mode)

40.4 (50.0)

38.9 (33.3)

38.4 (33.3)

Number of foreign supervisory board chairmena

3

16

16

Ratio of foreign supervisory board chairmen to all chairmen in %

1.3

4.4

5.5

Average number of management board members (Mode)

4.1 (3)

3.5 (3)

3.3 (2)

a

Only incidences where the foreign nationality is stated in the Hoppenstedt Aktienf¨uhrer could be accounted for. Source: Author’s findings; Data from Hoppenstedt Aktienf¨uhrer 1994, 1999 and 2004.

Table 6-4: Board Characteristics 1993-2003

to all chairmen. That ratio increased form 1.3% in 1993 to 3.3% in 2003.611 On the one side, foreign corporations acquiring (majority) stakes in domestic listed companies might have increasingly sent representatives to the supervisory board of German companies. On the other side, the search for higher qualified and adequate board members could also have fostered that phenomenon.612 With regard to management board members, a similar trend as for supervisory board members can be observed. The average number declined from 4.0 in 1993 to 3.3 in 2003. Similarly to the situation in supervisory board, criticism

611

612

The percentage of shares held by non-resident investors in 2000 was estimated to amount to 19.9% and thereby was among the lowest ratios in a comparison of 13 European countries. See Coto (2002), p. 40. Cf. Blum (2005), pp. 160-161.

156

Empirical Analysis of Insider Ownership

arose about management boards’ tendency to reach consensus among all board members which slows down or hinders fast decision making.613 Market data as stock prices, dividend payments etc. are obtained by Datastream (Thomson Financial). Accounting data as total assets, sales or debt are obtained via Datastream from the Worldscope Database. Finally, a modified version of the industry classification scheme applied by Deutsche B¨orse to build sectorindices is used to divide the sample into eight industry groups. Details on all data used will be presented in the next section. 6.3

Definition of Variables

6.3.1

Shareholder Structure Variables

6.3.1.1 Introductory Remarks As in the preceding theoretical analysis, shareholder structures in the empirical part will be analyzed along two dimensions, shareholder identity and shareholder concentration.614 Further, it will be distinguished between control rights and cash flow rights. The former only refer to the distribution of ordinary (voting) shares while the latter refer to the consolidated distribution of both voting ordinary shares and non-voting preference shares.615 In this case the respective ownership stakes of voting and non-voting shares are weighted according to the ratio of total voting to total non-voting share capital.616 Last, it will be distinguished between direct ownership and ultimate ownership. Direct ownership refers to the first level owner as indicated in Hoppenstedt Aktienf¨uhrer who has an immediate stake in the company. Since a significant number of these direct owners are firms, which are in turn owned by other shareholders, ownership is also measured at the ultimate owner level. Therefore, corporations as direct owners which have majority shareholders (> 50%) themselves will be classified 613

614 615

616

Cf. Blum (2005), pp. 172-173. For example, Gerhard Cromme, the chairman of the current German corporate governance code commission, pleads for a reduction of membership in supervisory boards. See N.N. (2003), p. 14. See section 3.3. Cf. Grossman and Hart (1988), p. 176, who view “[. . . ] the corporation’s security-voting structure as a mechanism for shifting control to a superior rival [to incumbent management], if such a team exists.” For example, a company has ordinary shares outstanding which account for 60.0% or the share capital the rest being non-voting preference shares. The members of the management board (MGMT) own 20.5% of the ordinary shares and 5.0% of the preference shares. Hence, MGMT’s control rights are 20.5% and MGMT’s cash flow rights are 20.5% x 60.0% + 5.0% x 40.0% = 14.3%. See appendix A. Cf. DeAngelo and DeAngelo (1985), pp. 45-46; Edwards and Weichenrieder (2004), p. 152.

Empirical Analysis of Insider Ownership

157

according to the identity of the majority shareholder.617 For example, the stake of an intermediate (holding) company which is fully controlled by a management board member will be classified as management board ownership on the ultimate owner level. As not indicated otherwise, the variables always represent the percentage ownership of the respective group. In this section all basic ownership variables will be presented while modifications to these variables, which are only used in a single model, will be introduced later. 6.3.1.2 Shareholder Identity To analyze the distribution of ownership stakes among different shareholder groups, all shareholder entries from the Hoppenstedt Aktienf¨uhrer had to be classified according to shareholders’ identity. Thereby, it was differentiated between insider owner types and outsider owner types as it can be seen in table 6-5. Further, the 16 classifications on the sub-category level were condensed to six main categories to allow for different levels of aggregation in the analyses.618 The six main owner categories are management (MGMT), non-financial corporations (CORP), financial institutions (FINC), government (GOV), miscellaneous (MISC), and all non-registered shareholders, i.e. the free float portion (FF). It is important to realize that FF may contain shareholders of the other categories as long as their ownership stake is not publicly known (usually if the shareholding is below the 5% threshold).619 The shareholdings of these owner categories always add up to 100% since FF is calculated as 100% less the sum of all other known ownership stakes. The main distinction from other more frequently used owner typologies is that this classification explicitly lists management as a separate owner type whilst others (e.g. the Bundesbank scheme620 ) classify management more generally as individuals. As will be shown, the vast 617

618

619 620

If necessary, the ultimate owner will be traced trough several layers of ownership until no majority shareholder can be found. This procedure is similar to that used by Franks and Mayer (2001), p. 948; K¨oke (1999), p. 7. See appendix A. Cf. Gorton and Schmid (2000), pp. 34-37. Gleisberg (2003), p. 98, argues that on the one hand, a detailed classification is necessary to account for individual characteristics of certain owner groups. On the other hand, it is necessary to aggregate owner groups to make an empirical analysis feasible. For a summary of German shareholder disclosure regulations see section 6.2.2. The classification scheme of Deutsche Bundesbank differentiates between non-financial sector (private households, corporations, public authorities), financial sector (banks, insurances and pension funds, investment funds and other financial institutions) and foreign countries. See Deutsche Bundesbank (1997), p. 29.

158

Empirical Analysis of Insider Ownership

majority of these individuals belongs to the management group as defined in the context of this study.621 The main category MGMT is further divided into three sub-categories: active members of the management board (MB), active members of the supervisory board (SB) and former board members of both board (FBM).622 For all three groups, shares of the respective family members are also included since it can be reasonably assumed that they are usually acting in concert with management. The decision to classify former board members as quasi-insiders pays attention to the characteristic peculiarity of German companies that former board members with large ownership stakes often execute considerable influence on “their” former companies without being officially in charge.623 Because of the novelty of this study, no standard approach of how to realize insider ownership in the context of German companies has evolved. Therefore, the descriptive statistics and regression analyses will use both the sub-category measures (MB, SB and FBM) individually as well as the aggregated insider ownership measure (MGMT).624 Non-financial corporations (CORP) are cut into two sub-groups according to the size of their ownership stake: Ownership stakes up to 50% are labeled as corporations (COR) while ownership stakes above this threshold are called business groups (BGR). This differentiation makes sense since the nature of a strategic investment of a company into another company below the 50% threshold can be expected to be different from that where the mother company has majority control over a (listed) subsidiary company.625 Because the relation between both parties in a corporate group can be expected to be different from that in a strategic investment, COR and BGR will be analyzed separately and aggregated (CORP). 621 622

623

624

625

Cf. figures 6-2 and 6-7. Since the sub-categories are not mutually exclusive the categories of active board members (i.e. MB and SB) dominate the category of former board members (FBM). Therefore, a management board member who quits the management board joining the supervisory board will always be classified as SB even though he also constitutes a former member of the management board (FBM). See Haller, Jaskiewicz, May and Schiereck (2004) who discuss the role of active and non-active shareholders/partners in German companies. Also in the U.S. researchers have not come to a uniform conclusion, e.g. whether to include directors in the measurement of insider ownership or not. See Ricardo-Campbell (1983); Demsetz and Villalonga (2001), p. 214. Of course also ownership stakes below 50% may allow for substantial control but the majority threshold appears to be the most logical cut.

Empirical Analysis of Insider Ownership MainSubCategory Category

159

Description

Insider Owner Types MGMT MB SB FBM

Management, i.e. active and former management and supervisory board members Active member of the management board (“Vorstand”) including family members Active member of the supervisory board (“Aufsichtsrat”) including family members Former member of the management or supervisory board including family members

Outsider Owner Types CORP BGR COR FINC BANK INSR INVC INST

GOV MISC

INST-F GOV INVD EMP OTH TRE

FF

FF

INTM

INTM

Non-financial corporation Business group (owning more than 50% of voting rights) Corporation (owning less than 50% of voting rights) Financial institutions Bank, i.e. mortgage, credit or investment bank (for own account) Insurance company Investment company, i.e. venture capital or buyout company Domestic institutional investor, i.e. asset management company, pension fund, bank (for third party account) Foreign institutional investor (see above) Government, i.e. German municipal, state or federal government Miscellaneous Outside individual, i.e. private persons who are not insiders Employee(s), i.e. employees of the company excluding members of the boards Others, i.e. all shareholders which cannot be assigned to another category Treasury shares, i.e. shares hold by the company itself (limited to 10% due to § 71 (2) AktG) Free float, i.e. dispersed portion of the shares calculated as 100% less sum of the shareholdings of all other categories Intermediate (holding) company, which is attributed to the other ownership categories at the ultimate owner level

Source: Author’s illustration.

Table 6-5: Shareholder Identity Classification Scheme Financial institutions (FINC), even though excluded as a study subject, represent an important shareholder group. On the sub-category level, it is divided into banks (BANK), i.e. mortgage, credit or investment banks, which own stakes for its own account, insurance companies (INSR), investment companies (INVC), i.e. venture capital or buyout companies, domestic institutional

160

Empirical Analysis of Insider Ownership

investors (INST), i.e. asset management companies, pension funds, banks for third party account, and foreign institutional investors (INST-F). The differentiation between BANK and INST is not always clear-cut since e.g. the largest domestic mutual fund companies belong to commercial banks or insurance companies.626 However, since shareholdings of these companies are usually classified as Sonderverm¨ogen (“separate estate”), which belongs to the investors, the ultimate owner rule is not applied and shareholdings of these companies will be classified as INST.627 Even though legally obliged to serve the interests of its investors, conflicts of interest between the superordinate bank and the original investors may arise of course. All public authorities belonging to municipal, state or federal government will be classified into the government (GOV) shareholder group. State-owned banks, e.g. the Kreditanstalt f¨ur Wiederaufbau (KfW), will be classified as GOV according to the ultimate owner concept.628 No sub-categories exist for this main category. Last, the miscellaneous (MISC) group consists of outside individuals (INVD), i.e. natural private persons not being insiders, the company’s employees (EMP), i.e. employees of the company excluding members of the boards, and other shareholders not belonging to any other group (OTH). Own shares held by the company, so-called treasury shares (TRE), were also included in the scheme even though they do not represent a shareholder group in the original sense.629 At the direct owner level, an additional owner group, intermediate (holding) companies (INTM), is introduced. In this group companies without operating business whose primary purpose is to hold participations in other companies are 626

627 628

629

The six largest mutual fund companies — cumulating 84.9% market share in the German open mutual fund market — are dominated by banks or insurance companies. In the following, they are listed with their superordinate company and market shares as of 30.06.2005: DWS-group (Deutsche Bank, 24.0%), Deka-group (Sparkassen-Finanzgruppe, 17.3%), Union-group (Volks- und Raiffeisenbanken, 17.1%), Allianz Dresdner Global Investors-group (Allianz, 15.8%), Cominvest-group (Commerzbank, 5.4%), and Activest-group (HypoVereinsbank (HVB), 5.3%). See BVI Bundesverband Investment und Asset Management e.V. (2005). Cf. Wenger and Kaserer (1998b), p. 61; Helmis (2002), p. 63. See § 1 InvG (formerly §§ 1, 6 KAGG). The KfW belongs to 80% to federal and to 20% to state government. Consequently, e.g. the sale of partially privatized Telekom AG shares from federal government to the KfW resembles more a “parking” solution to improve public budget than a real change of control. In this study, these cases will be classified as GOV. According to § 71b AktG, the company is not entitled to any rights (including voting and dividend rights) which result from the acquisition of treasury shares which changes the nature of these shareholdings significantly. Furthermore, the acquisition of treasury stocks is limited to 10% of total share capital due to § 71 (2) AktG.

Empirical Analysis of Insider Ownership

161

included. These entries are imperatively attributed to one of the other ownership categories at the ultimate owner level.630 6.3.1.3 Shareholder Concentration With respect to shareholder concentration three types of variables, which are not mutually exclusive, are constructed as can be seen in table 6-6. The largest shareholder variable (C1) measures the shareholding of the largest registered shareholder.631 All outside shareholders owning more than 5% of the voting rights are classified as blockholders or block owners (BLOCK O).632 Insider blockholders are not included since it is expected that outside block owners and insider ownership act as substitutional governance mechanisms.633 Blockholders are divided into four sub-categories according to the size of their blockholding: For example, blockholders with stakes between 5.0% and 24.9% are labeled as blockholder type 1 (BLOCK O T1) while blockholders with stakes between 75% and 100% are labeled as blockholder type 4 (BLOCK O T4).634 Finally, the number of outside blockholders is used to measure among how many different blockholders the stake of all outside blockholders, i.e. BLOCK O, is distributed. As not indicated otherwise, all concentration measures refer to control rights and not cash flow rights since shareholder concentration refers to the power of certain shareholder groups which in turn is determined by their respective portion of voting rights.635

630 631

632

633

634

635

See footnote 617. In the context of describing the concentration of shareholder structures it is common use to label the size of the n-largest shareholders as C-n. Hence, C1 indicates the shareholding of the largest shareholder, C3 indicates the shareholdings of the three largest shareholders. See e.g. Leech and Leahy (1991), p. 1419. Since the average number of registered shareholders in this study is between 1.6 and 2.8 (see table 6-3) only the use of C1 seems appropriate. Cf. Pound (1988), p. 248; Thomsen, Pedersen and Kvist (2006), p. 248; Barclay and Holderness (1991), p. 862. Closely related to blockholders is the only loosely defined concept of relational investors. Bhagat, Black and Blair (2003) define relational investors as those outside shareholders who hold 10% of a company’s shares for at least four years. Therefore, it also requires a long-term orientation of the shareholders in addition to a defined minimum stake. Cf. Franks and Mayer (2001), pp. 946-947, or Emmons and Schmid (1998), p. 31, who also use 25%, 50% and 75% as thresholds. An alternative approach to measure ownership concentration is the construction of a Herfindahl index. See Gorton and Schmid (2000), p. 39; Demsetz and Lehn (1985), p. 1163. In the context of this study, the construction of a Herfindahl index does not appear promising as the average number of shareholder entries per sample company varies between only 1.6 (1993) and 2.8 (2003). See table 6-3.

162

Empirical Analysis of Insider Ownership

MainCategory

SubCategory

Description

C1

Largest shareholder (inside and outside owner types)

BLOCK O

Blockholder (outside owner type); size of relevant blockholding between 5% and 100%. Blockholder type 1 (outside owner type); size of the relevant blockholding between 5.0% and 24.9% Blockholder type 2 (outside owner type); size of the relevant blockholding between 25.0% and 49.9% Blockholder type 3 (outside owner type); size of the relevant blockholding between 50.0% and 74.9% Blockholder type 4 (outside owner type); size of the relevant blockholding between 75.0% and 100.0%

BLOCK O T1 BLOCK O T2 BLOCK O T3 BLOCK O T4

BLOCK NO

Number of all blockholders (outside owner type); size of the relevant blockholding between 5% and 100%.

Source: Author’s illustration.

Table 6-6: Shareholder Concentration Classification Scheme In the descriptive statistics on shareholder structures in section 6.4.1, shareholder concentration will also be analyzed along shareholder identities, i.e. the identity of blockholders will be examined. To clarify the process how the shareholder structure of a single company is classified, appendix A illustrates the classification of the shareholder structure of an exemplary company. 6.3.2

Performance Variables

Three performance measures will be used in the main analysis of insider ownership and corporate performance: historic buy-and-hold stock returns (BAHR), market-to-book values (MTBV) and return on assets (ROA). Alternative performance measures will be used as robustness checks later on in section 6.9.4.1. Historic stock returns are measured as discrete buy-and-hold total returns (BAHR) over a period of 60 months. By using total returns it is assumed that all potential dividend payments would be reinvested. Hence, for the 2003 cross section the BAHRs are measured during the 60 months period from December 1998 to December 2003.636 Analogously, the stock price returns for the 636

It would be desirable to measure BAHRs in the 60 months period following (and not preceding) the cross section of ownership data. However, this is not possible for the 2003 cross section. Regarding the rather

Empirical Analysis of Insider Ownership

163

1998 and 1993 sub-samples are measured over the period from December 1993 to December 1998 and December 1988 to December 1993 respectively.637 A total of six outliers with values above 500% are excluded from the analysis.638 Since four of the six excluded cases show a MGMT-share of over 40% (average: 35.2%) the analysis of the remaining observations would rather understate than overstate any positive relation between insider ownership and performance.639   RIt BAHRt = − 1 × 100 RIt−60 Where BAHRt = Buy-and-hold returns over the preceding five years in % t = t ∈ [31.12.1993, 31.12.1998, 31.12.2003] RIt = Return index on month-end t t RIt = Rt−1 × PPt +D t−1 with RIt−1 = Return index on previous month-end = Price on ex-date Pt Pt−1 = Price on previous date = Dividend payment associated with ex-date t Dt It should be noted that this approach, in a certain sense, is more conservative than the MTBV approach, or the Tobin’s Q approach, frequently used in the U.S. literature. This should be illustrated by the following example: Assuming

637

638

639

inert nature of German ownership data (see section 6.9.4.2) it is assumed that the used ownership structures are an acceptable proxy for five-year lagged ownership structures. The number of valid cases of the BAHR variable is lower than the total sample size since total return prices where not available for the whole 60 months period for several companies. This could result either from a primary listing during that period or a lack of availability in the Datastream database. See section 6.2.1. Excluded BAHRs: 0 (1993), 4 (1998) and 2 (2003). Cf. Maury (2006), p. 325, who caps the dependent variables Tobin’s Q, ROA and ROE at the 5th and 95th percentiles to reduce the weight of extreme values. McConnell and Servaes (1990), p. 609, report that their results do not change if the distributions of Tobin’s Q and the insider ownership variable are truncated at their 5th and 95th percentiles. Similarly, Gugler, Mueller and Yurtoglu (2004), p. 10, caps all basic variables at the 1st and 99th percentile. Cf. Schwalbach (2004), p. 174. For example for the year 2003, the inclusion of the two outliers, with BAHRs of +982% (MGMT: 0.50) and +1,117% (MGMT: 0.00), in a regression of BAHRs on insider ownership variables (MB, SB, FBM) decreases the R2 from 0.406 to 0.244 and the t-statistics for the insider ownership variables fall by factors of two to three. Since the standardized residuals for the two outliers lie above +7-times standard deviation they could be excluded from the analysis. The regression coefficients show the same signs but lower t-statistics.

164

Empirical Analysis of Insider Ownership

that for whatever reason there is a positive relationship between insider ownership and firm performance, stock prices would react accordingly right in the moment when the ownership structure becomes public if the market is completely aware of this relationship and can be assumed to be at least semi-strong efficient.640 Hence, as long as no changes in the ownership structure occur no under- or outperformance would be observable, even though insider controlled companies would be economically successful. Under these conditions, the stock price approach would not be able to detect any relation between insider ownership and firm performance.641 However, if the market does not fully reflect the benefits of insider control right from the beginning, stock price returns would convey partial information about the market’s assessment of the benefits of insider ownership.642 It seems plausible that the market is affected by such learning effects, especially if longer periods are taken into consideration. However, the longer the period of observation, the more likely it is that even a rather sticky ownership variable is subject to changes and, hence, the stock price movement would be affected by such changes.643 For that reason an observation period of five years seems appropriate, being sufficiently long in order to account for the market’s learning effects, but sufficiently short not to be too much affected by changes in the insider ownership structure.644 The ratio of the market value of equity to the book value of equity is essentially the same as Tobin’s Q.645 MTBVs are measured at the end of the respective year and are calculated as follows: M T BVt =

640 641 642 643

644

645

M V (Equity)t BV (Equity)t

The original idea of the efficient market hypothesis was put down by Fama (1965). Cf. Bhagat and Jefferis (2002), p. 18. Cf. Drobetz, Schillhofer and Zimmermann (2004), p. 278. Cf. Gompers, Ishii and Metrick (2003), p. 108, who apply a similar long-horizon approach and call it a “long-run event study”. Examples for studies using stock returns as performance measure: Wolf (1999); Ehrhardt and Nowak (2003a); Drobetz, Schillhofer and Zimmermann (2004); Goergen (1998); Ehrhardt, Nowak and Weber (2006); Kaserer and Moldenhauer (2005). For problems regarding the use of stock returns as long-term performance measure cf. Barber and Lyon (1997); Lyon, Barber and Chih-Ling (1999). See Gorton and Schmid (2000), p. 44. Tobin’s Q will be used as an alternative performance measure later on in section 6.9.4.1.

Empirical Analysis of Insider Ownership

165

Where = Market-to-book value at t M T BVt t = t ∈ [31.12.1993, 31.12.1998, 31.12.2003] M V (Equity)t = Market value of equity at t, measured as the sum of the market values of all share classes at t BV (Equity)t = Book value of equity capital and reserves at t The market value of equity, i.e. the nominator of the fraction, by definition cannot become negative, while the book value of equity, i.e. the denominator, can do so. In those cases the MTBV cannot be interpreted. Similarly, the MTBV becomes very large, if the denominator approaches zero even though the nominator might be very small. Hence, negative, zero and MTBVs above 15 were excluded from further analysis.646 Because of the problems involved in the calculation of Tobin’s Q, which mainly arise from German historical cost accounting, MTBVs are a frequently used proxy for Tobin’s Q as performance measure in German corporate governance studies.647 Next, the return on assets (ROA) measures the overall profitability of the firm. Even though the return on equity (ROE) appears to be a more suitable measure if shareholder wealth maximization is considered as the fundamental business objective, the deficiencies of this measure are severe if used in econometric testing.648 Hence, the more robust ROA is used instead. Since the return, i.e. the nominator of the fraction, represents an accounting figure, biases through accounting policy might be introduced. However, accounting based performance measures are less affected by market moods than, e.g. stock returns or Tobin’s Q.649 The ROA is calculated as follows:650

646

647

648 649 650

Excluded negative/zero MTBVs: 3 (1993), 7 (1998) and 14 (2003). Excluded MTBVs above 15: 3 (1993), 17 (1998) and 3 (2003). Cf. Drobetz, Schillhofer and Zimmermann (2004), p. 277; footnote 638. However, problems in the calculation of both measures might be problematic only to the degree to which they introduce systematic biases into the analysis. Examples for studies using MTBV as performance measure: Thonet and Poensgen (1979); Edwards and Nibler (2000); Gorton and Schmid (2000); Gorton and Schmid (2002); Edwards and Weichenrieder (2004); Drobetz, Schillhofer and Zimmermann (2004). See Phani, Reddy, Ramachandran and Bhattacharyya (2005), p. 10. Cf. section 6.9.4.1. See Bhagat and Jefferis (2002), p. 19. One outlier (value > 100%) was excluded from further analysis: 0 (1993), 0 (1998) and 1 (2003). This case shows an insider ownership level of 0%. Cf. footnote 638.

166

Empirical Analysis of Insider Ownership 

ROAt =



P ATt + Interestt × (1 − T ax) −1  (T otal Assets)t BV

× 100

Where Return on assets (ROA) at t in % t ∈ [1993, 1998, 2003] Published after tax profit during year t Total interest charges during year t Tax rate during year t. Calculated as income taxes divided by the sum of income taxes and PAT  (T otal Assets)t = Average of book value of total assets at begining and BV end of year t

ROAt t P ATt Interestt T axt

= = = = =

The nominator of the fraction is the published after tax profit plus interest charges times one minus the tax rate.651 The denominator represents the average of total assets for the respective year.652 6.3.3

Control and Other Variables

Besides ownership and performance variables, various variables regarding firm characteristics and the corporate control setting of companies are introduced. While some hypotheses will be tested in univariate analyses, others — especially those related to the relationship between insider ownership and corporate performance — will be primarily examined in multivariate models, where control variables adjust e.g. for the influence of firm size on corporate performance.653 Firm Size (LN ASSETS), measured by the natural logarithm of total assets654 , is included to account for the fact that insider ownership in very large corporations is expected to be less widespread. Moreover, governance mechanisms might be different in large firms. Firm specific risk (FIRM RISK) mea651 652

653

654

See Maury (2006), p. 325. Examples for studies using ROA as performance measure: Demsetz and Lehn (1985); Wolf (1999); Cui and Mak (2002); Ehrhardt, Nowak and Weber (2006); Maury (2006). Without the inclusion of control variables any observed correlation between a performance measure and insider ownership might be a spurious result of a correlation between these two variables and a third omitted variable. See Morck, Shleifer and Vishny (1988), p. 605. Cf. section 6.10. As the variation in total assets across firms is enormous, it has become standard to use the natural logarithm of total assets or total sales as a proxy for firm size. See amongst others Agrawal and Knoeber (1996), p. 388; Himmelberg, Hubbard and Palia (1999), p. 367; Drobetz, Schillhofer and Zimmermann (2004), p. 288.

Empirical Analysis of Insider Ownership

167

sures the unsystematic, undiversifiable portion of companies’ total risk. It is measured as the residuals’ sum of squares (SSE) from a regression of individual stock returns on the returns of the market (CDAX) over the preceding 60 months.655 The financial structure (DEBT RATIO), measured as total debt to firm value (total debt + market value of equity), reflects the disciplining effect of higher interest burdens on management behavior.656 The growth potential (SALES G), which is expected to be captured in the market valuation of equity, is proxied by the average annual sales growth over the past three years.657 For example, sales growth would be measured as the annual growth of total sales during the period from 2001 to 2003 in the 2003 sub-sample. It is included in the analysis to differentiate higher market valuations arising from higher growth potential from those that might be the result of lower agency costs due to the alignment of interest among management and shareholders.658 The dummy variable dividends (DIV) indicates whether dividends have been paid during the respective year. Dividend payments generally could be seen as a positive signal on part of the management to outside investors about the company’s prospects.659 In addition to the control variables, six variables relating to the corporate control situation of listed companies are introduced. Management board size (MB NO) and supervisory board size (SB NO) measure the degree to which decision or control power is split up between several persons. The dummy variable deviation form the one-share-one-vote principle (VOTE) indicates whether any voting restrictions (e.g. non-voting preference shares, multiple voting rights, voting caps) exist.660 The dispersion of insiders’ control and cash flow rights (MGMT CRmCF), calculated as MGMT’s control rights minus MGMT’s cash flow rights, measures to which degree insiders own control rights in excess of cash flow rights. This usually can be achieved by owning comparatively more 655

656 657 658

659

660

See Demsetz, Marc and Philip (1996), p. 6; Brown and Kapadia (2005), p. 8. One outlier (value > 15.0) was excluded from further analysis: 0 (1993), 0 (1998) and 1 (2003). During the period from January 1999 to December 2003 the stock price of this small new economy company first rose by 1,744% and the fell by 99.90% leading to a SSE of 42.9. The next highest value was 12.1. See Demsetz and Villalonga (2001), p. 221. See Schmid (2003), p. 9; Beiner, Drobetz, Schmid and Zimmermann (2006), p. 253. One outlier (value > 10) was excluded from further analysis: 0 (1993), 1 (1998) and 0 (2003); cf. Maury (2006), p. 326, who also caps control variables at the 5th and 95th percentiles to reduce the weight of extreme values. Cf. Beiner, Drobetz, Schmid and Zimmermann (2006) p. 260; Edwards and Weichenrieder (2004), pp. 155156. Cf. Gorton and Schmid (2000), pp. 42-43.

168

Code

Empirical Analysis of Insider Ownership

Description

Panel A: Control Variables (Firm Characteristics) LN ASSETS FIRM RISK

DEBT RATIO

SALES G DIV

Size of the company, measured as the natural logarithm of total assets Firm specific risk, measured as the sum of squared residuals (SSE) from a regression of individual stock returns on market returns (CDAX) over the preceding 60 months Debt ratio, proxied as the ratio of book value of total debt divided by the sum of book value of total debt and market value of equity Annual sales growth, measured over the preceding three years Dummy variable: 1, if the company paid dividends during the year and 0 otherwise

Panel B: Other Variables MB NO SB NO VOTE

MGMT CRmCF

BLOCK O BLOCK NO

Number of all management board members Number of all supervisory board members Level of voting restrictions; 0 if no non-voting preference shares are issued and 1 divided by the ratio of ordinary share capital to preference share capital if non-voting preference shares are outstanding Difference between control rights and cash flow rights for the aggregate insider ownership measure MGMT (see section 6.3.1) Aggregated ownership by all outside blockholders, owning at least 5% (see section 6.3.1) Number of all outside blockholders, owning at least 5% (see section 6.3.1)

Source: Author’s illustration.

Table 6-7: Overview of Control and Other Variables ordinary shares than preference shares. Finally, two shareholder structure variables, already introduced in section 6.3.1, are also expected to play an important role in the corporate control setting: Block ownership (BLOCK O), the cumulative shareholdings of all outside shareholders owning more than 5% of the voting rights, and the number of blockholders (BLOCK NO) measure the concentration of shareholders and hence the monitoring power they might exercise on management. To control for performance differences across industries — e.g. due to cyclicality — the sample companies were categorized into eight industry groups.

169

Empirical Analysis of Insider Ownership Deutsche B¨orse 1993 / 1998

Deutsche B¨orse 2003

Modified Classification 1993 / 1998 / 2003

Automobile

Automobile

Automobile

Chemicals

Chemicals Pharma & Healthcare

Chemicals

Construction

Construction

Construction

Consumers

Consumers Retail

Consumers

Electronics

Media Software Technology Telecommunication

Electronics

Breweries

Food & Beverages

Food & Beverages

Mechanical engineering Paper Iron and steel Textile

Industrial Basic Resources

Industrial

Transportation Utilities

Transportation & Logistics Utilities

Utilities & Transportation

Mortgage banks Commercial banks

Banks

Financial companya

Investment companies

Financial Services

Financial companya

Insurance

Insurance

Financial companya

a

All types of financial companies were excluded from the sample. For details see section 6.2.1. Source: Author’s illustration.

Table 6-8: Industry Classification Scheme

As illustrated in table 6-8, the classification is geared at the industry classification scheme used by Deutsche B¨orse to classify companies listed at the FWB. Since the industry scheme changed during the sample period, some modifications were necessary.661 This resulted in the following eight industries: Automobile, Chemicals, Construction, Consumers, Electronics, Food & Beverages,

661

Currently, the scheme used by Deutsche B¨orse AG classifies Prime Standard companies into 18 different industries 15 of those belonging to the non-financial sector. The number of industry categories was reduced by grouping from 15 to eight non-financial categories in order to increase the number of cases in each category.

170

Empirical Analysis of Insider Ownership

Industrial, and Utilities & Transportation.662 In the regression analyses industry effects were incorporated by the inclusion of industry dummy variables. 6.4

Descriptive Results

This section presents summary statistics for the previously defined variables following the applied differentiation between shareholder structure, performance and control variables. Thereby, in line with the focus of this study the main focus will rest on descriptive statistics on shareholder structures and especially insider ownership. The unique shareholder classification scheme (see table 6-5) allows for a new view on the role of insider ownership in German listed stock corporations. According to the differentiation in section 3.3, the aspects of shareholder identity and shareholder concentration will be analyzed separately as well as combined. 6.4.1

Shareholder Structures Variables

6.4.1.1 Shareholder Identity The descriptive statistics on the identity of shareholders will distinguish between control rights and cash flow rights and between the ultimate owner and direct owner level (see section 6.3.1). For the sake of clarity, not the entire universe of possible combinations (control vs. cash flow rights, ultimate vs. direct owner level, main vs. sub-category level) can be presented. Instead, those statistics which impart the most important information or which enter the causal analyses later on are selected at the author’s discretion. All data refer to the ultimate owner level, if not indicated otherwise.663 6.4.1.1.1

Control Rights

Based on 236, 358 and 290 firm observations for the years 1993, 1998, and 2003 respectively (see table 6.2.1), figure 6-2 illustrates the distribution of unweighted mean control rights among the six main owner categories, i.e. their mean ownership stake if equal weight is attributed to all observations independent of the corresponding firm size. As can be seen, insiders or manage662 663

Financial industries are not included since they were excluded from the sample as object of investigation. Direct ownership seems less meaningful than ultimate ownership from an economic point of view. Therefore, direct owner data will only be presented in section 6.4.1.1.4 to illustrate the importance of the ultimate owner concept. Cf. Franks and Mayer (2001), p. 948; K¨oke (1999), p. 7.

Empirical Analysis of Insider Ownership

171

ment (MGMT) constitute the most important owner group cumulating a relatively stable portion of average control rights between 32.0% (2003) and 33.7% (1998).664 The median of MGMT is considerably lower in 1993 (1.5%) and moderately lower in 1993 (22.6%) and 2003 (22.5%) indicating a positive skewness of the distribution of insider ownership. The very low value for 1993 is caused by the fact that smaller stakes were hardly disclosed in 1993.665 These findings are roughly in line with the 27.2% mean ownership stake of individuals ¨ (2003) in their 1996 analysis of BaFin notifound by B ECHT AND B OHMER 666 However, the distribution of MGMT is highly positively skewed, fications. i.e. a large percentage — between 39.7% in 2003 and 50.0% in 1993 — does not own any or only very small stakes and the median values usually lie significantly below the average.667 This pattern is very common for most ownership variables — especially those on insider ownership — and should be kept in mind in the interpretation of the presented findings.668 Apart from free float (FF), the owner group corporations (CORP) ranks second but its share decreased from 25.8% in 1993 to 20.0% in 2003. On the one hand, this could reflect the gradual disintegration of the so-called Deutschland AG (“Germany, Inc.”), a term which refers to the network of cross holdings among German listed stock corporations.669 On the other hand, the figures could (partially) simply represent a statistical illusion: The changes in disclosure regulations in 1995 (see section 6.2.2) increased the transparency of ownership structures by making also smaller shareholdings, especially those between 5% and 25 public. This causes an extraordinary increase in the mean shareholdings of those owner groups which traditionally owned primarily smaller stakes. Vice versa, owner groups with generally larger shareholdings (e.g. CORP) show a relative decrease even though the actual sit664

665 666 667

668

669

For historical (e.g. 1931-1937, 1960-1998) shareholdings of slightly different shareholder types cf. Ernst (2001), pp. 7-9. See footnote 670. See Becht and B¨ohmer (2003), p. 13. Seifert, Gonenc and Wright (2005), p. 182, analyzing 319 German companies also report a mean insider ownership level of 19.2% in the year 2000 while the median is equal to 0.0%. Cf. Himmelberg, Hubbard and Palia (1999), p. 363. See Jensen and Warner (1988), pp. 5-6 and Chen, Cheung, Stouraitis and Wong (2005), p. 435, who find very similar patterns for insider and family ownership in U.S. and Hong-Kong firms respectively. See H¨opner and Krempel (2005), pp. 10-11; W´ojcik (2003). Cf. Wenger and Kaserer (1998a), pp. 503-511; section 2.4.3; footnote 129.

172

Empirical Analysis of Insider Ownership

40

Mean Control Rights (in %)

35

33.7 32.7

31.8 31.3 31.9

32.0

30 25.8 24.3

25

20.0

20 15 12.0

10 5.6

6.8

5

3.0 2.5

1.2

1.1 1.4

2.9

0 MGMT

CORP

FINC

1993

GOV

1998

MISC

FF

2003

The medians for MGMT are: 1.5% (1993), 22.6% (1998) and 22.5% (2003). The medians for FF are: 25.0% (1993), 27.9% (1998) and 27.0% (2003). All other medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-2: Control Rights 1993-2003

uation might not have changed at all.670 Nevertheless, this cannot explain the decrease between 1998 and 2003 and is in line with comparable research.671 The largest absolute change can be observed in the group of financial institutions (FINC) whose average stake increased by 114.3% from 5.6% in 1993 to 12.0% in 2003. Even though this observation might be partially due to the mentioned size-effect of disclosure changes, it can also be seen as a trend towards financial intermediation. This development could indicate that many people, especially those new to the stock market, might prefer investment funds to direct

670

671

The described effect of changes in disclosure regulations will be referred to as “size-effect of disclosure changes” and, as a general problem, might bias the comparisons over time for all presented results. Cf. table 6-10. An alternative explanation might be that the disclosure standards introduced in 1995 only became fully effective with a lag of three or more years.

Empirical Analysis of Insider Ownership

173

investments in the stock market.672 This presumption will be supported by the subsequent analysis on the sub-category level (see figure 6-6). The decline of state holdings (GOV) from 3.0% (1993) to 1.2% (2003) reflects the proceeding privatizations efforts of German public authorities.673 The increase in MISC can only be interpreted on the sub-category level for practical purposes. Finally, it should be stressed that the mean free float (FF) reaches only between 31.3% (1998) and 31.9% (2003).674 Therefore, the observations of B ERLE AND M EANS (1932), who found out that shareholder structures of U.S. listed companies around 1930 were highly dispersed, do not reflect the German situation at the end (beginning) of the 20th (21th) century.675 Furthermore, this underlines that the importance of the capital market in Germany is still different from that in capital market oriented countries as the U.S., where the FF is considerably higher.676 Taking a closer look at MGMT on the sub-category level reveals the distribution of ownership stakes among management board (MB), supervisory board (SB) and former board members (FBM). As can be seen in figure 6-3, MB represents the most important sub-group with mean ownership stakes between 11.6% (2003) and 16.1% (1998) followed by SB with average stakes ranging from 9.6% (1998) to 11.7% (1993). While no clear pattern over time can be observed for MB and SB, the FBM-portion shows a distinct increase from 7.1% in 1993 to 9.6% in 2003. This could be explained by the gradual exit of the founder generation from active management, which interestingly seems to not being necessarily associated with a sale of their respective shares. This ob-

672

673

674

675 676

This observation is in line with the results of a study conducted by Deutsches Aktieninstitut e.V. (DAI) (2006): In the period 1998-2003, the percentage of citizens above 14 years with direct shareholdings in relation to the total German population increased slightly by 9.9% from 7.1% to 7.8% while the percentage of investment funds owners increased by 154% from 5.0% to 12.7%. See Deutsches Aktieninstitut e.V. (DAI) (2006), p. 08.3-Zahl-D. The 1993 (1998) value of 3.2% (2.5%) is rather close to the 3.2% (1.9%) reported by Ernst (2001), p. 9, for 1993 (1998). Similarly, K¨oke (1999), p. 16, reported a mean percentage of dispersed shareholding of 36.1% for listed AGs in 1998. Cf. Taussig and Barker (1925), pp. 11-18. For example, McConnell and Servaes (1990), pp. 600-601, report that the mean insider ownership share for 1976 (1986) is 13.9% (11.8%) and the average ownership of all blockholders is 32.4% (25.6%). Hence, the remaining free float portion would be 53.7% in 1976 and 62.7% in 1986.

174

Empirical Analysis of Insider Ownership

20

Mean Control Rights (in %)

16.1

15

14.0 11.6

11.7 10.7 9.6

10

9.6 8.0 7.1

5

0 MB

SB

1993

1998

FBM

2003

All medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-3: Control Rights 1993-2003 (MGMT) servation strengthens the argument to include FBM as quasi-insiders as they constitute an important owner group like MB or SB.677 To examine whether the phenomenon of insider ownership is concentrated in certain industries, figure 6-4 plots the 2003 MGMT-portion across industries. There are no remarkable differences among six of the eight industry categories. However, the insider ownership pattern in the Food & Beverages as well as in the Utilities & Transportation industry is quite different from other industries. In fact, the former has an unusual high insider ownership share of 64.2%, while in the latter, the opposite is true showing an insider share of only 4.8%. Presumably, this result is driven by a size effect and small group sizes of the Food & Beverages (n = 13) and Utilities & Transportation industry (n = 20). Moreover, it should be noted that utilities in Germany usually are formerly state owned companies. Anyhow, it can be stated that insider ownership is a widespread 677

In the context of family firms, Villalonga and Amit (2006), p. 386, argue that it is important to differentiate among three elements in the definition of family firms: management (i.e. family members in the management board), control (i.e. family members as representatives in the supervisory board) and ownership (i.e. family ownership). Cf. Kuklinski, Lowinski and Schiereck (2003), pp. 103-104.

175

Empirical Analysis of Insider Ownership

70

64.2

Mean Control Rights (in %)

60 50 39.3

40

32.0

34.8 30.3

30.1

30

27.9

29.6

20 10

4.8

El ec tr on Fo ic s( od 49 + ) Be ve ra ge s( 13 ) In du st U ri til al iti (6 es 7) + Tr an sp A .( ll 20 In ) du st ri es (2 90 )

er s( 65 )

(2 9)

um

n C on s

tr uc t io

C on s

ic m he C

A

ut

om

ob

al s

ile

(2 8)

(1 9)

0

Numbers in parentheses indicate the number of relevant companies per industry group. The results for 1998 and 1993 do not considerably differ from those in 2003. For a definition of the industries see table 6-8. Source: Author’s findings.

Figure 6-4: Control Rights in Industries 2003 (MGMT) phenomenon in German non-financial listed companies and not a particularity of a specific industry. As shown in figure 6-5, in the main category CORP on the sub-category level majority owned business groups (BGR) with mean ownership stakes between 15.4% (2003) and 21.1% (1998) dominate by far the minority stakes of corporations (COR) which account only for an average of 3.2% (1998) to 5.4% (1993).678 Hence, listed subsidiaries are more frequent than corporations’ strategic investments into other listed companies without majority control. The fact that the average BGR-portion, which by definition was subject to mandatory disclosure during the whole sample period, declined from 21.1% in 1998 to 15.4% in 2003 testifies the proceeding disintegration of the Deutschland AG beyond possible biases caused by changes in disclosure regulations.679 678

679

By definition each individual stake classified as majority owned business group must exceed 50%, but the average values for the whole sample are considerably smaller due to the skewness of the distribution. See footnote 129.

176

Empirical Analysis of Insider Ownership

25 21.1 20.4

Mean Control Rights (in %)

20

15.4

15

10

5.4

5

4.7 3.2

0 COR 1993

BGR 1998

2003

All medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-5: Control Rights 1993-2003 (CORP) In the FINC-category, several interesting patterns can be observed (see figure 6-6). First, banks (BANK) account only for moderate average ownership stakes between 2.6% (1998) and 3.8% (1993). Hence, the often claimed power of banks in the German corporate governance systems seems to stem more from proxy voting rights and supervisory board mandates than from direct ownerships stakes.680 Second, insurance companies (INSR) account for small and stable mean ownership stakes between 1.1% (1993) and 1.2% (1998). Third, investment companies (INVC) showed the highest relative increase at all. Virtually non existing with a 0.2% ownership stake in 1993, they reached a mean portion of control rights of 4.9% in 2003. This is in line with the perception that private equity investors started targeting the German market especially during the last decade.681 Private equity companies usually do not directly invest into listed companies but they may retain considerable stakes in their former portfolio companies after having sold them to the public via an IPO. Fourth, the 680 681

See Baums (1993), pp. 44-45; Edwards and Nibler (2000), pp. 245-251. Cf. section 2.4.3. See Bundesverband Deutscher Kapitalbeteiligungsgesellschaften (BVK) (2003), pp. 5-6. Cf. Achleitner and Kl¨ockner (2005), pp. 9-10; Bottazzi and Da Rin (2003), pp. 3-6.

177

Empirical Analysis of Insider Ownership

6 4.9

Mean Control Rights (in %)

5

4

3.8

3.1

3

2.6 2.0

2 1.2

1.1

1.2

1.2 0.9

1

0.8

0.8

0.5 0.2

0.0

0 BANK

INSR 1993

INVC 1998

INST

INST-F 2003

All medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-6: Control Rights 1993-2003 (FINC) trend towards financial intermediation is mirrored in the increasing importance of institutional investors (INST and INST-F) as an owner group. However, the rise from 0.5% in 1993 to 2.8% in 2003 might be significantly driven by the size-effect of disclosure changes since INST are usually prohibited from owning stakes larger than 10% in other companies and, hence, those stakes were normally not disclosed prior to 1995.682 Last, foreign institutional investors (INST-F) only started to appear as an owner group in 1998, indicating increased interest in the German capital market by foreign investors. Looking at the MISC-category in figure 6-7 allows for three general statements: First, outsider individuals (INVD) account only for a tiny portion of ownership stakes by all individuals, i.e. insider and outsider persons. Putting the mean 0.9% ownership stake of INVD in 2003 in relation to the 32.0% stake of 682

Most INST are organized as “Kapitalanlagegesellschaften” (“Investment companies”) which are subject to special regulation in Germany (esp. the “Investmentgesetz” (InvG) (“Investment Act”) or until 2004 the “Gesetz u¨ ber Kapitalanlagegesellschaften” (KAGG) (“Investment Company Act”)). According to § 64 (2) InvG (formerly § 8a (3) KAGG) investment companies must not own more than 10% of the voting rights in any listed corporation. Therefore, it can be expected that INST is understated, especially in 1993 but also to a minor degree in 1998 and 2003.

178

Empirical Analysis of Insider Ownership

MGMT, it becomes obvious that outside individuals only represent about 2.7% of all — i.e. inside and outside — individual shareholders.683 Second, employees’ (EMP) participation in their companies is low between 0.1% (1993) and 0.5% (1998). These low percentages are likely to understate employees’ real shareholdings as the majority of their shareholdings will be classified into the FF-category. Nevertheless, these figures show that employees cannot be seen as an important organized shareholder group. Finally, the evolution of treasury shares (TRE) in 2003 could be explained twofold: On the one hand, companies might have tried to profit from their comparatively low market valuations in 2003 by acquiring shares in their own companies.684 On the other hand, this might reflect companies’ preference to indirectly distribute profits back to shareholders — primarily driven by tax considerations685 — instead of pursuing investment projects. However, it should be noted that the hitherto presented figures are unweighted means and, hence, systematic differences in small and large companies are not taken into account. In fact, the picture changes substantially, if means weighted by the market capitalization of equity are calculated. This is shown in figure 6-8.686 The average MGMT-stake ranges from 10.9% (1993) to 14.0% (1998), ¨ (2003) i.e. 58-65% lower than in the unweighted case. B ECHT AND B OHMER find, for 1996 blockholder data of German listed companies, that the mean market weighted equity ownership by individuals is 13.1% and thereby 52% below their unweighted portion of 27.2%.687 The owner categories CORP and GOV deserve attention in 1998. The comparatively lower share of CORP in 1998 indicates that the market values of companies, where other companies are important shareholders, were suppressed in a downward-spiral.688 With regard to financial institutions (FINC) their weighted 683

684 685

686

687 688

It should be emphasized once more that these figures only refer to registered shareholdings. Of course, FF contains the mass of dispersed individual shareholdings. Cf. section 6.10. See Wiebe (2003), p. 32. Share buy backs enjoy tax privileges compared to dividend payouts because increases in stock prices (capital gains) are tax free for the majority of investors as long as the shares are held longer than one year. Cf. N.N. (1999), p. 4; G¨obel and Kormaier (2006), pp. 65-70. The shareholdings of the different owner types are weighted by the mean of month-end equity market capitalizations in the relevant year. See Becht and B¨ohmer (2003), p. 13. As the DAX declined by 18.9% from its year-high of 6,171.43 (20.07.1998) to 5,002.39 (31.12.1998), valuations of companies holding significant stakes in other listed companies might suffer disproportionately as

179

Empirical Analysis of Insider Ownership

Mean Control Rights (in %)

2

1.1

1

0.9 0.7

0.6

0.5

0.5 0.4

0.5 0.3

0.1 0.0

0.0

0 INVD

EMP 1993

OTH 1998

TRE 2003

All medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-7: Control Rights 1993-2003 (MISC) share increases by the factor 1.4 to 1.8, e.g. from 5.6% (see figure 6-2) to 9.1% in 1993.689 On the other side, market values of partially state-owned companies (GOV) showed comparatively high valuations which might be influenced by the partial privatization of Deutsche Telekom AG in 1996.690 If the companies are weighted by their market capitalization, the average FFportion increases up to 45.3% in 2003 or 55.9% in 1993.691 Evidently, managerial ownership is the more relevant the smaller the market capitalization of a company. This is also shown in figure 6-9, where sample companies are

689

690

691

the value of the held portfolios declines. Similarly, in the aftermath of the new economy share prices of corporates with large holdings in other listed companies (e.g. Allianz AG, M¨unchner R¨uck AG) suffered exceptionally. See Alich, Dohmen and Sommer (2002), p. 1. For example, Drukarczyk, Honold and Sch¨uler (1995), p. 20, report higher absolute shares for banks, insurance companies and mutual funds in 1993 and 1994 but also find that the market weighted share of these shareholder groups surpass the unweighted ones by similar factors ranging from 1.4 to 1.8. The value of the shares of Deutsche Telekom AG increased by 68% from their IPO on November, 18, 1996 to year end 1998. Source: Datastream. For the 30 largest German listed companies (DAX), Sommer (2005), p. 29, reports even an unweighted mean free float portion of 81.5% (72.4%) in 2004 (2001).

180

Empirical Analysis of Insider Ownership

60

55.9

55 50.4

50 Mean Control Rights (in %)

45.3

45 40 35 30 25 17.9

20 14.0 12.5

15 10.9

10

16.2

15.4 12.4 9.3

12.1

9.1 6.4

5.4

5

4.2 0.8 1.9

0 MGMT

CORP

FINC

1993

GOV 1998

MISC

FF

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-8: Control Rights Weighted by Market Value 1993-2003 grouped according to their size as measured by total assets.692 As can be seen, MGMT’s mean control rights decrease (almost) monotonically from the 41.1% (2003) to 48.4% (1998) range for the smallest size quintile to a 9.6% (1993) to 18.1% (1998) range for the largest companies.693 Although comparable figures are not available for the U.S. or the U.K., it can be safely assumed that the mean free float weighted by market capitalization would be substantially higher than the 62.7% reported above.694 In fact, H IMMEL BERG , H UBBARD AND PALIA (1999) find average total managerial ownership stakes of 13.4% for companies whose sales exceed USD 188m while smaller 692

693

694

Size quintiles are builded on basis of total assets for the relevant years. The 1st quintile comprises of the 20% smallest, the 5th quintile of the 20% largest companies as measured by total assets. The results do not considerably differ if the market value of equity is used as size measure instead of total assets (the coefficient for the correlation of total assets and market value of equity is 0.82). A similar analysis was conducted by Taussig and Barker (1925) for U.S. corporations at the beginning of the 20th century. Arranging their surveyed companies according to size they already observe for large corporations that “[. . . ] the participation of the managers in ownership usually decreases as size increases, but with no regularity.” Taussig and Barker (1925), pp. 16-17. Cf. pp. 11-18; Belkaoui and Pavlik (1992), p. 346. See footnote 676.

181

Empirical Analysis of Insider Ownership

55 48.4

50

Mean Control Rights (in %)

45

47.5 44.2 41.1

40

38.2

38.7 36.9

37.2 34.7

35

30.1

30

28.7 26.1

25 18.1

20

15.0

15 9.6

10 5 0 1st Quintile

2nd Quintile

3rd Quintile

1993

1998

4th Quintile

5th Quintile

2003

Source: Author’s findings.

Figure 6-9: Control Rights in Size Groups 1993-2003 (MGMT) companies show considerable higher insider ownership stakes between 25.4% (USD 22m < sales < USD 188m) and 32.0% (sales < USD 22m).695 Although these results do not include external blockholdings, it can be expected that this clear size-effect exists also for outside blockholdings. 6.4.1.1.2

Cash Flow Rights

Switching the perspective from control to cash flow rights should enable the detecting of systematic differences between the ownership of ordinary voting and non-voting preference shares. Most obvious, MGMT’s cash flow stake lies in the 27.8% (1993) to 29.6% (1998) range (see figure 6-10) and, hence, is 3.0% to 4.9% lower than the respective control stake. The reverse is true for FF, which reaches 36.0% (2003) to 38.2% (1993). This is not surprising since preference shares are predominantly designed for dispersed ownership, as possible blockholders of preference shares would not gain any control potential, at least as long as the company meets the promised minimum payment of its mandatory 695

See Himmelberg, Hubbard and Palia (1999), p. 362.

182

Empirical Analysis of Insider Ownership

45 38.2

40 Mean Cash Flow Rights (in %)

36.5 36.0

35 29.6 29.0

30

27.8 24.8

25

23.6 19.4

20 15

11.7

10 5.4 6.6

5

3.0 2.5 1.2

0.9 1.3

2.7

0 MGMT

CORP 1993

FINC

GOV 1998

MISC

FF

2003

The medians for MGMT are: 1.5% (1993), 17.8% (1998) and 21.1% (2003). The medians for FF are: 35.9% (1993), 38.3% (1998) and 35.5% (2003). All other medians are 0. For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-10: Cash Flow Rights 1993-2003 (Main Categories) preferred dividend.696 On the sub-category level — apart from the MGMTportion which is further analyzed in the next section — no major changes can be observed. 6.4.1.1.3

Control and Cash Flow Rights

Figure 6-11 illustrates the differences between control and cash flow rights for the six main categories; a positive value indicates that the respective party on average owns control rights in excess of cash flow rights while a negative value indicates the opposite. As already discussed, the cash flow rights relating to FF exceed the associated control rights while all other parties show an opposite pattern. Most remarkably, the MGMT-group owns control rights which are on average between 3.0% (2003) and 4.9% (1993) larger than its cash flow rights. This is in line with results of G OMPERS , I SHII AND M ETRICK (2004) who 696

According to § 140 (2) AktG the voting rights reinstate in those cases when a company fails to meet the preferred dividend payments on two consecutive dividend dates (dividend rights are accruing).

183

Difference between Control and Cash Flow Rights (in %)

Empirical Analysis of Insider Ownership

10

5

4.9 4.1 2.9 1.0 0.7

0.7

0 MGMT

CORP

0.2 0.2 0.3

0.0 0.0 0.0

0.2 0.2 0.2

FINC

GOV

MISC

FF

-4.2

-5 -5.2 -6.4

-10 1993

1998

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-11: Excess Control Rights 1993-2003 find for a U.S. sample of dual share class companies that managers’ and directors’ mean voting rights reach 50.7% while the respective cash flow rights are substantially lower (26.7%).697 Furthermore, also corporate shareholders, and less financial institutions, seem to own excessive control rights even though to a smaller extent.698 Analyzing the disproportionate control rights of the MGMT-group on the subcategory level, reveals further information about the nature of this phenomenon. Management board members (MB) claim the largest part of these excessive control rights followed by members of the supervisory board (SB) as can be seen in figure 6-12. This intuitively makes sense since active board members profit most from excessive control rights since these help to shelter active management from outside investors’ scrutiny. As in the aggregate (MGMT), MB and SB 697

698

See Gompers, Ishii and Metrick (2004), p. 12. This effect, not surprisingly, is much larger than the observed 3.0-4.9%, since only dual share class companies were examined. A similar observation has been made by Gorton and Schmid (2000), p. 70, who claim that “[. . . ] several institutional features, aside from a small stock market, suggest, that the link between cash-flow rights and control rights is somewhat uncoupled.”

Difference between Control and Cash Flow Rights (in %)

184

Empirical Analysis of Insider Ownership

4

3.1

3

2.0

2 1.6

1.0

1

1.2

1.0

1.1 0.8

0.2

0 MB

SB

1993

1998

FBM

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-12: Excess Control Rights 1993-2003 (MGMT) show a distinct decrease over time from 3.1% to 1.0% (i.e. a decrease of 67.7%) and 1.6% to 0.8% (i.e. a decrease of 50.0%) from 1993 to 2003 respectively. For former board members (FBM), this pattern is overcompensated by the general increased importance of this group (see figure 6-3). Hence, it can be stated that management tries to preserve its discretionary scope and to seal itself off from capital market control even though this tendency seems to have faded during the sample period. The presumption that MGMT uses the issuance of non-voting preference shares to escape capital market pressure, is also reflected in the sample statistics. As illustrated in table 6-9, the percentage of companies with dual share classes decreased from 29.2% in 1993 to 17.9% in 2003 in this sample.699 Furthermore, the ratio of ordinary to preference share capital increased from 3.2 to 5.5 over the same period. This development is also reflected in the percentage of compa699

Similarly, Jenkinson and Ljungqvist (2001), p. 407, find that defensive share structures (voting caps, only non-voting listed shares, limited share transferability, deviations from one-share-one-vote) exist in 20.6% of all German listed companies as of 1992.

185

Empirical Analysis of Insider Ownership 1993

1998

2003

Total number of sample companies

236

358

290

Number of companies with dual-class shares

69

79

52

Companies with dual-class shares in % of total

29.2

21.9

17.9

Average ratio of ordinary to preference share capitala

3.2

3.1

5.5

Number of companies with any deviation from the one-share-one-vote principleb

92

106

60

Companies with any deviation from the one-shareone-vote principle in % of total

39.0

29.6

20.7

a b

For companies with dual-class shares. § 139 (2) AktG states that non-voting preferred shares may be issued only up to one-half of the share capital which translates into a minimum value of 1. Besides dual-class shares, other deviations form the one-share-one-vote principle include e.g. voting restrictions or non-listed preference shares with multiple voting rights. Source: Author’s findings.

Table 6-9: Incidences of Control Enhancing Devices 1993-2003 nies which show any deviation from the one-share-one-vote principle. Besides the issuance of non-voting preference shares, this includes statutory caps on the voting rights of each individual shareholder700 or multiple voting rights for preference shares with regard to decisions defined in certain paragraphs of the company’s articles.701 While 39.0% of all sample companies showed such deviations in 1993, the percentage has fallen to 20.7% in 2003. This indicates an increased orientation towards capital markets’ expectations. 6.4.1.1.4

Ultimate and Direct Ownership

Similar to the differences between control and cash flow rights, the differences between ultimate ownership, presented hitherto, and direct ownership can be analyzed. Figure 6-13 illustrates total control rights on the direct owner level 700

701

The number of companies with statutory voting restrictions has decreased. In 2003, only one sample company, i.e. Volkswagen AG, limited the voting power of a single shareholder to 20%. In 1993, at least Phoenix AG, Dyckerhoff AG, AVA AG and Rosenthal had similar voting restrictions in place. Most restrictive was the cap of AVA AG which limited the voting power of a single shareholder to 1/1,000 of the share capital. For an overview of companies with voting restrictions see cf. Gorton and Schmid (2000), p. 43; Goergen, Manjon and Renneboog (2004), pp. 4-5. For example in 2003: Lindner Holding KGaA, CEWE Color Holding AG or Muehlbauer Holding AG, Robert Cordier AG among others.

186

Empirical Analysis of Insider Ownership

Mean Control Rights - Direct Ownership (in %)

35 31.9 29.0

30 25

31.3 31.9

23.9

25.1

26.3 22.5

20.2

20 15

13.4 9.1 9.6

9.6

10 4.0 5.2

5

1.5 1.5 0.4

0.7 1.1

2.1

0 MGMT

CORP

FINC

1993

GOV

MISC

1998

FF

INTM

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-13: Direct Ownership Stakes 1993-2003 while figure 6-14 depicts the differences to ultimate level ownership. In contrast to the ultimate owner data, on the direct owner level also the category intermediate (INTM) appears. Since these stakes must be attributed to one of the other groups on the ultimate owner level, it can be seen which group collates most of these control rights on the ultimate owner level. While all owner groups — apart from CORP702 — claim parts of the INTM-stake, the dominance of the MGMT-group is striking. The reasons for MGMT’s motivation to hold its stakes through intermediate holding companies might be tax motivated since corporate taxes are usually lower than income taxes in Germany, hence, protecting income from higher tax burdens.703 Nevertheless, the process of data collection and shareholder classification shows that sometimes INTMs appear to serve as a camouflage for the real (ultimate) owners: Because of the comparatively low transparency standards re702

703

Since corporations, as INTM, can be dominated by other owner groups on the ultimate level it also shows a higher ownership stake at the direct owner level and, hence, a negative value in figure 6-14. For example, the majority of capital gains and dividend incomes from investments in corporations are tax free on the corporate level. See § 8b KStG (K¨orperschaftsteuergesetz (“Corporation Tax Law”).

187

Empirical Analysis of Insider Ownership

Difference between Control Rights - Direct and Indirect Ownership (in %)

15 11.8

10

8.8 8.7

5 2.4

CORP

1.6 1.6

1.5

1.0 0.8

0.5 0.3 0.8

0 MGMT

-5

FINC -3.2

GOV

INTM

MISC

-2.0 -2.4

-10

-9.1 -9.6 -13.4

-15 1993

1998

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-14: Ultimate and Direct Control Rights 1993-2003 garding the disclosure of shareholdings in Germany — as regulated primarily in the WpHG and enforced by the BaFin704 — MGMT can try to hide its (majority) stakes in the company in order not to prevent minority shareholders from investing into their company.705 This is in line with the observations of F RANKS AND M AYER (2001) who observe that families and banks are more prominent on the ultimate than on the direct shareholder level.706 6.4.1.2 Shareholder Concentration The shareholder concentration will be analyzed along the three concentration measures presented in section 6.3.1: largest shareholder, blockholder and number of blockholders.707 704 705

706 707

See section 6.2.2. For example, shares held via non-listed firms must only be attributed to their ultimate owner if he holds a majority in this non-listed firm. Similarly, family control may not be disclosed if no formal pool contract exists. See Becht and B¨ohmer (2003), pp. 16-17, 21-22. See Franks and Mayer (2001), p. 948. Cf. K¨oke (1999), p. 20. See table 6-6.

188

Empirical Analysis of Insider Ownership

6.4.1.2.1

Largest Shareholder

Analyzing the concentration of shareholder structures along the largest shareholder measure gives rise to two questions: First, which identity has the largest shareholder and, second, what is the size of the corresponding stake. Figure 615 illustrates the frequency distribution of the identity of the largest shareholder. Once more, MGMT turns out to represent the largest fraction of all largest shareholders: Corporate insiders represent the largest shareholder in between 47.1% (1993) and 51.1% (1998) of all cases. This is well in line with the findings of E DWARDS AND N IBLER (2000), who report that individuals/families are the largest shareholder in 40.0% of all cases in their sample of 105 listed German firms in 1992.708 Hence, in roughly half of all sample companies a person (or family) out of the MGMT-group is the most powerful shareholder. Recalling the prominence of management ownership from section 6.4.1.1, this result appears less surprising. Similar to the distribution of control rights on the aggregate level709 , CORP ranks second followed by FINC and GOV.710 While 4.2% of all companies had no identifiable shareholder in 1993, this fraction has diminished to 0.3% in 2003. Again, this reflects the increased transparency in shareholder structures during the sample period. Complementary to figure 6-15, table 6-10 reveals the identity of the largest owner on the sub-category level as well as the mean size of the corresponding stakes. As can be seen from the last row, the size of the average largest stake decreased from 59.0% in 1993 to 56.5% in 1998 and 52.1% in 2003. This is ¨ ¨ (2003), K OKE (1999) and in line with the findings of B ECHT AND B OHMER E DWARDS AND W EICHENRIEDER (2004), who report that the average size of the largest shareholding is 58.9%, 57.7% and 56.2% respectively.711 Similarly, FACCIO AND L ANG (2002) report 54.5% as the size of the largest shareholding for their German sub-sample in the period from 1996 to 1999, which is also very 708 709 710

711

See Edwards and Nibler (2000), p. 247. See figure 6-2. The result that banks represent the largest shareholder in 3.8% (2003) to 5.5% (1993) of all companies (see table 6-10) is somewhat lower than the 12% observed by Gorton and Schmid (2000) in their 1986 large company sample. They find non-financial firms to represent the largest ultimate shareholder in 42% of all companies. See Gorton and Schmid (2000), p. 39. See Becht and B¨ohmer (2003), p. 8. Since the averages for the three, five and all largest shareholders only rise to 68.7%, 70.0% and 70.2% respectively, the largest shareholder rarely faces competition for voting rights from other blockholders. See K¨oke (1999), p. 11; Edwards and Weichenrieder (2004), p. 154.

189

Empirical Analysis of Insider Ownership

Largest Shareholder's Identity (in % of Total)

55 50

51.1 50.8 47.1

45 40 34.4

35

31.0

30

26.2

25 20

16.7

15 8.4

10

10.9 3.7

5.1 3.4

5

2.4

0.8

1.7

4.2 2.0 0.3

0 MGMT

CORP

FINC

1993

GOV

1998

MISC

None

2003

For a definition of the owner categories see table 6-5. Source: Author’s findings.

Figure 6-15: Largest Shareholder’s Identity 1993-2003 close to the 56.5% found in this analysis.712 Finally, also the 65% (1997) and ´ (2003) are well in line with the figures reported 60% (2001) found by W OJCIK 713 above. However, the findings for the size of the largest shareholdings in this study are higher than the 46.1% to 47% found by E DWARDS AND N IBLER (2000) and VAN DER E LST (2000).714 The size of the largest shareholding is largest if the owner’s identity falls into the CORP-group, where the mean largest stake varies between 65.5% (1993) and 73.1% (1998). Slightly smaller are the stakes of the MGMT-group with averages ranging from 50.7% (2003) to 66.2% (1993).715 Whilst the size of the largest shareholding is decreasing over time for most owner groups, FINC shows a different pattern, i.e. the average stake increased from 29.2% (1993) 712 713 714

715

See Faccio and Lang (2002), p. 392. See W´ojcik (2003), pp. 1442-1443. See Edwards and Nibler (2000), p. 247; Van der Elst (2000), p. 33. Becht and B¨ohmer (2003), p. 9, presume that these deviations stem from a focus of both studies on larger firms. Cf. Becht and B¨ohmer (2003), p. 3. Edwards and Nibler (2000), p. 247, in their 1992 sample find a comparable mean of 57% for the size of the largest shareholder if he is classified as individual or family.

190

Empirical Analysis of Insider Ownership

Identitya

1993

1998

2003

CORP

34.4 (65.5) 26.3 (75.9) 8.1 (31.5)

31.0 (73.1) 25.7 (81.3) 5.3 (33.4)

26.2 (66.6) 17.9 (82.6) 8.3 (31.9)

8.4 (29.2) 5.5 (28.6) 2.1 (25.7) 0.4

10.9 (25.4) 4.2 (25.0) 2.0 (22.0) 2.2 (27.0) 0.8 (54.0) 1.7 (14.3)

16.7 (42.1) 3.8 (34.7) 2.1 (22.2) 5.9 (64.8) 2.8 (45.1) 2.1 (6.8)

BGR COR FINC BANK INSR INVC

b

INST

0.4 b

INST-F GOV

5.1 (48.3)

3.4 (45.8)

2.4 (38.4)

MGMT

47.1 (66.2) 19.1 (70.5) 17.4 (62.0) 10.6 (65.2)

51.1 (56.5) 25.4 (53.0) 14.0 (55.7) 11.7 (64.8)

50.8 (50.7) 19.7 (47.8) 16.6 (51.3) 14.5 (53.9)

0.8 0.4

1.7 (37.9) 0.6

3.7 (29.6) 1.0

b

b

b

0.3

0.7 (18.5) 1.7 (42.2) 0.3

MB SB FBM MISC

b

INVD EMP

b

OTH

0.4 b

0.8 (39.4)

TRE

b

None TOTAL a b

4.2

2.0

0.3

100.0 / (59.0)

100.0 / (56.5)

100.0 / (52.1)

Distribution of largest shareholding as per owner identity in percent. The average size in percent of the largest shareholding in percent is shown in parentheses. For a definition of the owner categories see table 6-5. Average size of largest shareholding is not shown if mean would be based on less than three instances. Source: Author’s findings.

Table 6-10: Overview of Largest Shareholder Identity 1993-2003

Empirical Analysis of Insider Ownership

191

and 25.4% (1998) to 42.1% (2003). Notably, this increase was driven by the rising importance of INVC, in both number and size, which own an average stake of 64.8% in 2003 in those cases, where they represent the largest shareholder. These numbers match the notion that private investors started heavily to target the German market from 1998 on.716 6.4.1.2.2

Blockholders

Blockholders: Overview While the previous section only reported distribution characteristics of the largest shareholding, this section focus on all blockholders, i.e. all outside shareholders owning at least 5% of the voting rights. In table 6-11, an overview of the distribution of blockholders’ identities at the three cut-off dates is given. The observed pattern is quite similar to that of the largest shareholding: While 57.5% of all blockholders in 1993 belong to CORP, this portion decreased to 35.5% in 2003. This once more emphasizes the dismantling of the Deutschland AG, which was constituted by a dense net of corporate cross holdings.717 In contrast, FINC clearly gained importance raising their respective share from 29.4% in 1993 to 51.8% in 2003. While BANK actually lost shares, all other sub-category types, and especially INST and INST-F, experienced increases. GOV faced the largest relative decline from representing 7.9% of all blockholders in 1993 to only 3.3% in 2003, which reflects German privatization efforts during the last decade. With respect to the size of the blockholdings, a sharp decline can be observed throughout the sample period. The average size of all blockholdings decreased from 38.3% in 1993 to 28.3% in 2003. As already discussed, parts of this decline might be attributed to the size-effect of disclosure changes during the sample period.718

716

717 718

The portfolio of German private equity investments between 1997 and 2003 rose by a yearly average of 36.1%. See Bundesverband Deutscher Kapitalbeteiligungsgesellschaften (BVK) (2003), pp. 5-6. Bottazzi and Da Rin (2003), pp. 3-4, claim that European venture capital fundraising had increased by the factor ten from 1995 to 2000. Cf. Achleitner and Kl¨ockner (2005), pp. 9-10; Kamp and Krieger (2005), pp. 10-14. Cf. Wenger and Kaserer (1998b), pp. 51-61; H¨opner and Krempel (2005), pp. 9-14; Busse (2005), p. 2. See footnote 670.

192

Empirical Analysis of Insider Ownership

Year/ Identitya

1993

1998

2003

CORP

57.5 (49.5) 30.4 (73.5) 27.1 (22.5)

44.6 (58.4) 28.7 (80.3) 15.9 (18.9)

35.5 (46.0) 15.3 (80.1) 20.3 (20.2)

29.4 (19.9) 20.1 (20.3) 7.5 (16.4) 1.4 (15.0) 0.5

40.7 (15.3) 15.9 (17.4) 8.9 (12.7) 7.3 (16.6) 1.2 (34.9) 7.3 (9.0)

51.8 (18.1) 17.3 (15.8) 9.3 (15.1) 12.6 (33.7) 5.6 (8.1) 7.0 (7.4)

BGR COR

FINC BANK INSR INVC INST

b

INST-F

GOV

7.9 (38.1)

8.6 (28.9)

3.3 (32.1)

MISC

5.1 (19.1) 2.8 (14.3)

9.3 (16.8) 4.7 (13.6) 0.7

100.0 (28.3)

OTH

2.3 (24.8)

6.1 (14.6) 3.4 (11.4) 0.9 (10.4) 1.8 (22.5)

TOTAL

100.0 (38.3)

100.0 (35.7)

INVD EMP

a b

b

4.0 (21.4)

Distribution of outside blockholders as per owner identity in percent. The average size in percent of the blockholdings is shown in parentheses. For a definition of the owner categories see table 6-5. Average size of blockholdings is not shown if mean would be based on less than three instances. Source: Author’s findings.

Table 6-11: Overview of Blockholder Identity 1993-2003

Empirical Analysis of Insider Ownership

193

Besides the overview provided in table 6-11, tables 6-12 to 6-14 reveal more detailed information by distinguishing between both blockholders’ identity and blockholdings’ size — as categorized in table 6-6 — for each sub-sample. Blockholders: 1993 In 1993 (see table 6-12), blockholders owning stakes between 5% and 24.9% (i.e. type 1) account for 40.2% (= 86/214) of all blockholdings and hence represent the largest of the four blockholder type groups. Out of the 236 sample companies, 69 (= 30+39) or 29.0% had a majority outside blockholder.719 Furthermore, it can be seen that FINC- and MISC-blockholders predominantly fall into the blockholder type 1 or 2 categories while CORP (and GOV) blockholders are roughly equally split between the blockholder type 1/2 and 3/4 categories.720 Finally, it should be noted that the average size of the blockholdings for blockholder type 2 and 3 are 30.8% and 57.5% and, therefore, lie clearly below the mean value of the respective blockholding intervals ((25%+50%)/2 = 37.5% and (50%+75%)/2 = 62.5%). As holdings above the relevant voting thresholds (e.g. 25% or 50%) do not ease expropriation of minority shareholders nor confer additional decision rights721 , this finding underlines that blockholders seem to choose the size of their blockholdings with regard to control thresholds established by the German Stock Corporation Act.722 This suggests that blockholdings are primarily held for control purposes as voting threshold should not play any role if the stakes would be held as passive investments. Alternatively, this finding might only reflect the positive skewness typical for most ownership variables.723 Blockholders: 1998 The pattern for 1998 (see table 6-13) is very similar to that in 1993. Most notably, type 1 blockholders could strengthen their dominance by increasing their share of all blockholdings from 40.2% in 1993 to 54.4% (= 178/327) in 719

720

721 722 723

The value is lower than e.g. the 72% reported by Jenkinson and Ljungqvist (2001), p. 404, or the 42% reported by Edwards and Nibler (2000), p. 246, as these figures from 1991 and 1992 also contain insider majority blockholders. For example, domestic institutional investors are usually prevented from owning large stakes in a single company by the InvG. See sections 3.3.2.3 and 3.3.2.5. See Holderness and Sheehan (1988), p. 325. See e.g. §§ 103, 122, 133, 179 and 182 AktG. Cf. Becht and B¨ohmer (2003), p. 10; K¨oke (1999), p. 12. For a comparable analysis of blockholdings based on the 1993 issue of Hoppenstedt/Saling Aktienf¨uhrer see Gorton and Schmid (2002), p. 42.

194 Type/ Identitya CORP BGRc COR

FINC BANK INSR INVC

Empirical Analysis of Insider Ownership Type 1b

Type 2b

Type 3b

Type 4b

All Tpyes

36 (14.8) 1 (10.0) 35 (15.0)

27 (33.7) 4 (31.6) 23 (34.1)

24 (57.1) 24 (57.1)

36 (90.8) 36 (90.8)

123 (49.5) 65 (73.5) 58 (22.5)

36 (11.8) 22 (11.8) 12 (12.2) 2 (10.0)

26 (29.0) 21 (29.2) 4 (29.2) 1 (25.1)

1 (75.0)

1 (75.0)

63 (19.9) 43 (20.3) 16 (16.4) 3 (15.0) 1 (75.0)

1 (98.2)

17 (38.1)

1 (75.1)

11 (19.1) 6 (14.3)

1 (75.1)

5 (24.8)

39 (90.2)

214 (38.3)

INST INST-F

GOV

6 (15.3)

4 (25.5)

MISC

8 (10.6) 4 (8.9)

2 (25.1) 2 (25.1)

INVD

6 (59.3)

EMP OTH

4 (12.2)

TOTAL

86 (13.2)

a b c

59 (30.8)

30 (57.5)

Number of relevant blockholders. The average size of the blockholdings in percent is shown in parentheses. For a definition of the owner categories see table 6-5. Blockholder types are defined by size of the relevant blockholding: Type 1 = 5.0% - 24.9%, Type 2 = 25.0% 49.9% , Type 3 = 50.0% - 74.9%, Type 4 = 75.0% - 100.0%. See table 6-6. Business groups (i.e. non-financial corporations with shareholdings greater than 50%) can be classified as type 1 or type 2 blockholders if the majority is achieved cumulatively through several blockholdings (on different levels of the company). Source: Author’s findings.

Table 6-12: Blockholder Identity and Types 1993

Empirical Analysis of Insider Ownership

195

1998. The absolute number of type 1 blockholders stayed constant for CORP with 36 instances in 1993 and 1998 while the corresponding number for FINC rose by 208.3% from 36 to 111. As in the other sub-samples, the mean size of all CORP-blockholdings (58.4%) is significantly larger than for the rest. This ¨ is in line with the findings of B ECHT AND B OHMER (2003) who report that industrial firms show the highest mean blockholding size (i.e. 61.2%).724 In 1998, for the first time a considerable number of institutional blockholders can be observed. Interestingly, the vast majority of these institutional investors (24/(24+4) = 85.7%) are foreigners, hence classified as (INST-F), indicating that domestic institutional investors only rarely held stakes above 5% in 1998. Blockholders: 2003 The development of blockholders’ characteristics from 1998 to 2003 corroborates the development towards smaller blockholdings already observed from 1993 to 1998: In 2003, 63.4% (= 199/314) of all blockholdings (vs. 40.2% in 1993 and 54.4% in 1998) fall into the type 1 category, i.e. are between 5.0% and 24.9%. In the CORP-category, a relative shift from larger (type 3 and 4) blockholdings towards smaller (type 1 and 2) blockholdings can be observed. While type 3 and 4 blockholdings accounted for 61.6% of all CORP-blockholdings in 1998, this portion diminished to 42.5% (= (14+37)/120) in 2003. One reason for that development might be the abolishment of capital gains taxes on sales of large stakes held by corporations in 2002.725 This enabled tax free realizations of capital gains by selling off blockholdings which often had market values (far) above their book values.726 Besides, domestic institutional investors (INST) gained importance by increasing their stake of all blockholdings from 0.5% in 1998 to 5.4% (= 17/314) in 2003. 6.4.1.2.3

Number of Blockholders

The number of blockholders per company increased during the ten year sample period: While each company had 0.90 (= 214/236) outside blockholders on average in 1993, this ratio rose to 0.91 (= 327/358) in 1998 and 1.08 (= 314/290) 724 725 726

See Becht and B¨ohmer (2003), pp. 13-14. See Becht and B¨ohmer (2003), p. 4. The so-called Tax Reduction Law (“Gesetz zur Senkung der Steuers¨atze und zur Reform der Unternehmensbesteuerung” (Steuersenkungsgesetz - StSenkG)) was introduced on October 23, 2000.

196

Empirical Analysis of Insider Ownership Type 1b

Type 2b

Type 3b

Type 4b

All Tpyes

36 (13.0)

29 (61.4) 29 (61.4)

61 (91.8) 61 (91.8)

36 (13.0)

20 (34.2) 4 (41.5) 16 (32.3)

146 (58.4) 94 (80.3) 52 (18.9)

111 (11.2) 41 (12.8) 26 (10.6) 20 (12.2) 1 (10.0) 23 (8.1)

19 (31.6) 10 (32.6) 3 (31.1) 3 (33.3) 2 (25.2) 1 (30.4)

2 (54.6) 1 (55.2)

1 (79.4)

133 (15.3) 52 (17.4) 29 (12.7) 24 (16.6) 4 (34.9) 24 (9.0)

GOV

13 (13.0)

11 (31.7)

2 (56.8)

MISC

18 (10.1) 10 (8.9) 3 (10.4) 5 (12.1)

1 (36.5) 1 (36.5)

1 (74.2)

178 (11.6)

51 (32.7)

Type/ Identitya CORP BGRc COR

FINC BANK INSR INVC INST INST-F

INVD EMP OTH

TOTAL a b c

1 (54.1) 1 (79.4)

2 (89.9)

20 (14.6) 11 (11.4) 3 (10.4) 6 (22.5)

1 (74.2) 34 (61.1)

28 (28.9)

64 (91.6)

327 (35.7)

Number of relevant blockholders. The average size of the blockholdings in percent is shown in parentheses. For a definition of the owner categories see table 6-5. Blockholder types are defined by size of the relevant blockholding: Type 1 = 5.0% - 24.9%, Type 2 = 25.0% 49.9% , Type 3 = 50.0% - 74.9%, Type 4 = 75.0% - 100.0%. See table 6-6. Business groups (i.e. non-financial corporations with shareholdings greater than 50%) can be classified as type 1 or type 2 blockholders if the majority is achieved cumulatively through several blockholdings (on different levels of the company). Source: Author’s findings.

Table 6-13: Blockholder Identity and Types 1998

197

Empirical Analysis of Insider Ownership Type/ Identitya

Type 1b

Type 2b

Type 3b

Type 4b

All Tpyes

CORP

45 (13.3%) 4 (14.1%) 41 (13.2%)

24 (35.2%) 4 (37.7%) 20 (34.7%)

14 (59.3%) 14 (59.3%)

37 (92.4%) 37 (92.4%)

120 (47.4%) 58 (75.5%) 62 (21.1%)

126 (9.8%) 42 (10.5%) 24 (9.6%) 23 (13.1%) 16 (6.9%) 21 (7.4%)

17 (31.6%) 8 (29.3%) 2 (33.9%) 6 (34.5%) 1 (27.0%)

3 (60.6%) 1 (55.6%) 2 (63.1%)

10 (86.5%) 1 (92.0%)

156 (18.1%) 52 (15.8%) 28 (15.1%) 38 (33.7%) 17 (8.1%) 21 (7.4%)

GOV

5 (14.6%)

3 (30.3%)

MISC

23 (11.4%) 12 (11.7%) 2 (11.3%) 9 (11.1%)

3 (26.9%) 2 (25.3%)

2 (63.3%)

1 (30.0%)

2 (63.3%)

199 (10.9%)

47 (33.0%)

19 (59.9%)

BGRc COR

FINC BANK INSR INVC INST INST-F

INVD EMP OTH

TOTAL a b c

9 (85.8%)

2 (78.8%)

10 (32.1%) 28 (16.8%) 14 (13.6%) 2 (11.3%) 12 (21.4%)

49 (90.6%)

314 (29.6%)

Number of relevant blockholders. The average size of the blockholdings in percent is shown in parentheses. For a definition of the owner categories see table 6-5. Blockholder types are defined by size of the relevant blockholding: Type 1 = 5.0% - 24.9%, Type 2 = 25.0% 49.9% , Type 3 = 50.0% - 74.9%, Type 4 = 75.0% - 100.0%. See table 6-6. Business groups (i.e. non-financial corporations with shareholdings greater than 50%) can be classified as type 1 or type 2 blockholders if the majority is achieved cumulatively through several blockholdings (on different levels of the company). Source: Author’s findings.

Table 6-14: Blockholder Identity and Types 2003

198

Empirical Analysis of Insider Ownership

in 2003.727 These figures are fairly below the 1.76 (= 755/430) blockholders ¨ per company reported by B ECHT AND B OHMER (2003). However, subtracting blockholdings of individuals and family pools, which usually would represent insiders in the context of this study and hence are not included in this statistic, this ratio declines to 1.23, which is roughly in line with the findings of this study. The portion of companies without any outside or inside blockholder in this study decreased along with higher disclosure standards from 4.7% (1993) to 1.7% (2003) which is once again well in line with the 1.6 % reported by ¨ (2003).728 B ECHT AND B OHMER Because one company by definition cannot have more than one type 3 or type 4 blockholder, the number of these blockholders reveals also the total number of sample companies, which are majority controlled by outside blockholders. This ratio declined from 29.2% in 1993 (= (30+39)/236), to 27.4% in 1998 (= (34+64)/358) and 23.5% in 2003 (= (19+49)/290). These figures are significantly lower than the 48.5% (1999) found by VAN DER E LST (2000), the 57.3% (1990) reported by F RANKS AND M AYER (2001) and the 64.7% (1996) ¨ found by B ECHT AND B OHMER (2003).729 The 18.7% increase in the number of blockholdings from 1998 to 2003 and the respective 17.1% decrease in their average size, are clear signs of a deconcentration of shareholder structures. These results are well in line with those of ´ W OJCIK (2003), who reports a 25% increase in the number of and a 10% decrease in the size of voting blocks for listed German corporations between 1997 and 2001. His results reveal that deconcentration has taken place irrespective of firm size, industry and location and, hence, should not be attributed to particular parts of the economy.730 6.4.2

Performance Variables

The descriptive statistics on shareholder structure variables are based on the total number of sample companies, i.e. 236 in 1993, 358 in 1998 and 290 in 2003, to get an overview of shareholder structures in German listed companies as comprehensive as possible. However, two factors demanded a reduction of 727 728 729 730

See tables 6-12 to 6-14. See Becht and B¨ohmer (2003), pp. 8, 12-13. See Van der Elst (2000), p. 39; Franks and Mayer (2001), p. 947; Becht and B¨ohmer (2003), p. 10. See W´ojcik (2003), pp. 1442-1444.

Empirical Analysis of Insider Ownership

199

the final sample used in the regression analyses. On the one hand, complete accounting or market data is not available for all companies in the Datastream and Worldscope databases. On the other hand, for the 1993 and 1998 subsamples complete stock return data for the whole period of 60 months preceding the cut-off date are partially not available since some of these companies just went public during that period. Since stock market performance varies largely over the five year periods,731 the inclusion of these firms into the sample (based on a shorter period, for which stock return data could be obtained) might have biased the results. The mean (median) five year buy-and-hold stock returns (BAHR) over the five year period are 77.1% (51.2%) for the period 1989-1993, 23.2% (3.3%) for the period from 1993-1998 and -7.1% (-22.6%) for the period from 1999-2003 (see table 6-15). Even though the CDAX rose by 96.0% during the period ranging from 1993 to 1998, the sample mean is significantly lower because the sample only includes those companies, which had their IPO prior to 31.12.1993. Hence, the new economy firms, which significantly contributed to the sharp rise of the CDAX, are not included in this sample. This could be interpreted as a new listing bias and underlines that the results of this study refer to the more traditional German listed firms, and are less influenced by the rather unusual times of the new economy.732 Interestingly, a somewhat different pattern can be observed for firm value, as measured by the market-to-book-value ratio (MTBV) as defined in section 6.3.2. It reaches an average of 2.80 in 1993, then increases to 3.05 in 1998 and finally declines to 2.02 in 2003 reflecting the overall trends in market valuations during that time frame. Notably, during the heyday of the new economy, the maximum absolute value of MTBV is reached in 1998 with 14.81 reflecting the high valuations at that time. Finally, the temporal development of return on assets (ROA) resembles that of MTBV. The highest mean value (5.4%) is reached

731

732

The mean annual total returns on the CDAX during the five year periods 1989-1993, 1994-1998 and 19992003 were 13.1%, 15.5% and 0.3% while the standard deviations reached 26.4%, 17.7% and 33.3% respectively. Source: Datastream. See section 6.2.1. This development is also reflected in the difference between mean and median returns. Since the median values lie significantly under the mean values the mean is positively biased by a small number of companies with returns far above the mean.

200

Empirical Analysis of Insider Ownership

Sample Period

1993

1998

2003

236

358

290

240 23.184 3.316 84.529 -97.482 478.920

285 -7.140 -22.584 87.251 -99.873 424.392

319 3.054 2.257 2.340 0.482 14.809

238 2.024 1.537 1.671 0.233 11.212

281 5.449 4.854 9.097 -34.440 81.629

251 2.747 3.695 10.317 -50.880 40.002

Total number of sample companies

Panel A: 5 year Buy-and-Hold Returns (BAHR) (in %) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum

165 77.105 51.176 96.123 -68.657 481.335

Panel B: Market-to-Book-Value (MTBV) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum

201 2.795 2.436 1.615 0.720 13.388

Panel C: 5 Return on Assets (ROA) (in %) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum

202 3.056 3.320 6.560 -32.802 27.857

For the definition of the performance measures see section 6.3.2. Source: Author’s illustration. Stock return and accounting data are from the Datastream database.

Table 6-15: Descriptive Statistics of Performance Variables in the 1998 sub-sample, followed by the 1993 sub-sample (3.1%) and the 2003 sub-sample (2.7%).733 6.4.3

Control and other Variables

The summary statistics of firm characteristics used in the further analyses are shown in table 6-16. Mean firm size — measured by the natural logarithm of total assets (LN ASSETS) — is highest in 2003 (12.88 or EUR 392m)734 and remains relatively constant during the three sub-sample periods. Firm specific 733

734

Similarly, Maury (2006), p. 326, reports a mean ROA value of 5.9% for his 1998 sample of European companies. However, the real mean of total assets is much larger (i.e. EUR 4,711m in 2003) as the logarithmic calculus largely eliminates the skewness of the original variable’s distribution.

Empirical Analysis of Insider Ownership

201

risk (FIRM RISK) is lowest in 1993 and seems to increase over time.735 The mean leverage ratio (DEBT RATIO) ranges from 20.4% in 1998 to 31.8% in 2003.736 Mean annual sales growth (SALES G) is highest in 1998 (16.9%) and becomes negative in 2003 (-0.9%) along with the economic downturn. While 81% of all sample companies pay dividends (DIV) in 1993, this ratio decreases to 69% in 1998 and 60% in 1993.737 Table 6-17 shows summary statistics for those variables relating to the corporate control situation of the sample companies. The mean size of the supervisory board (SB NO) falls from 10.9 members in 1993 to 8.6 members in 2003, indicating a general trend towards smaller board sizes. Similarly, the average size of the management board (MB NO) declines from 4.1 members in 1993 to 3.3 in 2003. The level of voting restrictions (VOTE) decreases from 0.17 in 1993 to 0.12 in 2003 reflecting the already mentioned decline of non-voting preference shares. In line with this development, the mean insiders’ control rights in excess of their cash flow rights (MGMT CRmCF) decrease from 4.9% in 1993 to 2.9% in 2003.738 Hence, it appears that the attention towards corporate governance issues had increased during the sample period.739 Finally, the blockholder statistics for the sample companies are similar to those presented in table 6-11 and section 6.4.1.2.3. The mean ownership share of outside blockholders (BLOCK O) varies between 32.0% (2003) and 34.7% (1993) while the mean number of blockholders (BLOCK NO) ranges from 0.91 (1993) to 1.08 (2003).740 735

736

737

738

739

740

The large difference between 1993/1998 and 2003 is driven by a few outliers in 2003, where the maximum reaches a value three times as high as in the other periods. Similarly, Richter and Gr¨oniger (2000), p. 305, find a mean debt ratio at market values for CDAX companies at the end of 1997 of 23%. Since the debt ratio is measured in market values, the increase from 1998 to 2003 probably results from the suppressed market values of equity at the end of 2003. Seifert, Gonenc and Wright (2005), p. 182, find a mean leverage ratio of 21.0% for three-year averages (1997-1999). N.N. (2006) reports that in 2004 a total of 60% — and especially smaller listed companies — did not pay any dividends. The 1998 mean values of firm size, debt ratio and sales growth are very similar to those reported by Maury (2006), p. 326, for his 1998 sample of European firms. The distribution of MGMT CRmCF is negatively skewed since the median is 0.0%. Since excess control rights only exist if preference shares are outstanding, the degree of excess control rights is much higher in those cases. Cf. Harvey, Lins and Roper (2004), p. 9, who report that average control rights exceed management’s cash flow rights by more than the factor two in emerging countries. Of course, this result might also be effected e.g. by changes in the industry composition of the sample (see 6-18). Cf. section 6.2.2. This is significantly lower than e.g. the mean level of blockholder ownership of 58.1% found for German companies by Thomsen, Pedersen and Kvist (2006), p. 256, in a sample of 74 (666) firm-(years) between 1990-1998.

202 Sample Period Total number of sample companies

Empirical Analysis of Insider Ownership 1993

1998

2003

236

358

290

214 12.854 12.721 1.775 8.350 18.280

349 12.560 12.274 1.900 8.540 18.720

261 12.875 12.514 2.036 7.000 18.990

165 0.359 0.293 0.332 0.047 3.365

244 0.536 0.356 0.624 0.005 4.740

286 1.504 0.731 1.954 0.015 12.060

206 0.216 0.176 0.187 0.000 0.828

342 0.204 0.152 0.201 0.000 0.832

253 0.318 0.280 0.269 0.000 0.939

194 0.059 0.014 0.265 -0.331 2.627

327 0.169 0.650 0.566 -1.000 6.105

260 -0.009 -0.022 0.368 -1.000 5.138

329 0.690 1.000 0.463 0.000 1.000

258 0.600 1.000 0.491 0.000 1.000

Panel A: Firm Size (LN ASSETS) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum Panel B: Firm Risk (FIRM RISK) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum Panel C: Leverage (DEBT RATIO) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum Panel D: Sales Growth (SALES G) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum

Panel E: Dividends (DIV) (Dummy Variable) Number of valid cases (n) Mean Median Standard deviation Minimum Maximum

198 0.810 1.000 0.395 0.000 1.000

For the definition of the control variables see section 6.3.3. Source: Author’s illustration. Accounting and market data are from the Datastream/Worldscope databases.

Table 6-16: Descriptive Statistics of Control Variables I

203

Empirical Analysis of Insider Ownership Sample Period Total number of sample companies

1993

1998

2003

236

358

290

9.140 6.000 5.440 3.0-21.0

8.630 6.000 5.280 3.0-20.0

Panel A: Number of Supervisory Board Members (SB NO) Mean Median Standard deviation Min. - Max.

10.850 12.000 5.472 3.0-22.0

Panel B: Number of Management Board Members (MB NO) Mean Median Standard deviation Min. - Max.

4.050 4.000 2.172 1.0-17.0

3.500 3.000 1.894 1.0-17.0

3.270 3.000 1.782 1.0-14.0

0.139 0.000 0.305 0.0-1.0

0.119 0.000 0.292 0.0-1.0

Panel C: Voting Restrictions (VOTE) Mean Median Standard deviation Min. - Max.

0.173 0.000 0.319 0.0-1.0

Panel D: Insiders’ Control above Cash Flow Rights (MGMT CRmCF) Mean Median Standard deviation Min. - Max.

0.049 0.000 1.113 0.0-0.5

0.041 0.000 0.114 0.0-0.5

0.029 0.000 0.097 0.0-0.5

0.326 0.150 0.374 0.0-1.0

0.320 0.174 0.366 0.0-1.0

0.910 1.000 1.043 0.0-6.0

1.080 1.000 1.188 0.0-6.0

Panel E: Block Ownership (BLOCK O) Mean Median Standard deviation Min. - Max.

0.347 0.249 0.372 0.0-1.0

Panel F: Number of Blockholders (BLOCK NO) Mean Median Standard deviation Min. - Max.

0.910 1.000 1.040 0.0-6.0

For the definition of the control variables see section 6.3.3. Source: Author’s illustration. Ownership and board data are from the Hoppenstedt Aktienf¨uhrer.

Table 6-17: Descriptive Statistics of Control Variables II

204

Empirical Analysis of Insider Ownership

Total number of sample companies

236

358

290

6.4 8.9 12.3 19.1 7.6 7.2 29.7 8.9

5.9 8.9 10.1 20.7 15.1 5.6 25.1 8.7

6.6 9.7 10 22.4 16.9 4.5 23.1 6.9

100.0

100.0

100.0

Sample Companies per Industry (in %) Automobile Chemicals Construction Consumers Electronics Food & Beverages Industrial Utilities & Transportation Totalb

For the definition of the industries see section 6.3.3. Numbers do not always add up to 100.0% due to rounding. Source: Author’s illustration. Industry classifications are modified from Deutsche B¨orse.

Table 6-18: Descriptive Statistics of Industry Split With regard to the industry split, table 6-18 reveals some moderate shifts between the sub-sample periods. Most strikingly, the share of sample companies belonging to the Electronics industry rises from 7.6% in 1993 to 16.9% in 2003 while the share of Consumer companies increases from 19.1% to 22.4% over the same period. Contrary, the shares of Industrial, Utilities & Transportation and Food & Beverages shrink from 29.7% to 23.1%, 8.9% to 6.9% and 7.2% to 4.5% respectively. 6.5

Overview of Hypotheses

Table 6-19 summarizes the hypotheses on the determinants and effects of insider ownership developed in chapter 5. In addition to the hypotheses themselves, it also lists the expected signs of the relationships.

205

Empirical Analysis of Insider Ownership Number

Hypothesis

Expected Sign

Panel A: Determinants of Insider Ownership H 5.1.1

Firm size has an effect on the level of insider ownership.

−

H 5.1.2

Firm specific risk has an effect on the level of insider ownership.

A: + / B: −

Panel B: Insider Ownership and Corporate Control H 5.2.1

The level of insider ownership has an effect on the size of the supervisory board.

−

H 5.2.2

The level of insider ownership has an effect on the size of the management board.

−

H 5.2.3

The level of insider ownership has an effect on the presence of voting restrictions.

+

H 5.2.4

The level of insider ownership has an effect on the difference between insiders’ control and cash flow rights.

+

H 5.2.5

The level of insider ownership has an effect on the level of outside blockholdings.

−

H 5.2.6

The level of insider ownership has an effect on the number of outside blockholdings.

−

H 5.2.7

The level of insider ownership has an effect on the level of firms’ leverage.

A: + / B: −

Panel C: Insider Ownership and Corporate Performance H 5.3.1

The level of insider ownership has an effect on corporate performance.

+

H 5.3.2

The level of insider ownership above a certain threshold has an effect on corporate performance.

−

H 5.3.3

The difference between insiders’ control and cash flow rights has an effect on corporate performance.

−

For a detailed description of the hypotheses see chapter 5. Source: Author’s illustration.

Table 6-19: Overview of Hypotheses

206 6.6

Empirical Analysis of Insider Ownership Methodology

In this study, three sub-samples, i.e. three cross sections at year end 1993, 1998 and 2003, as well as the pooled sample of German listed companies are used to examine the causes and effects insider ownership. Though being aware of the problems arising from the use of primarily cross-sectional data, this choice appears plausible if the following reasons are taken into account: First, it seems necessary to better understand shareholder structures at large and to learn more about the appropriate measurement of insider ownership before going into a deeper, more sophisticated analysis since insider ownership in Germany experienced little attention in research until now. Second, since the historical availability of shareholder structure data in Germany is rather limited741 and data quality is poor742 , the construction of a large and comprehensive panel data set faces an enormous effort. Furthermore, such an effort would probably not be rewarded because the mentioned problem regarding data quality poses natural limits to the examination of high frequency (e.g. yearly) shareholder structure data. Third, insider ownership tends to be rather sticky for the case of Germany (as will be shown in section 6.9), limiting the insights from a panel data analysis. The research question is addressed in a three step analysis as illustrated in figure 6-16. In a first step, a cause analysis of insider ownership is conducted. Examining the determinants of insider ownership in more detail will provide for a better understanding of its implications. Therefore, simple and multiple ordinary least square (OLS) regression analyses as well as difference in means tests are performed. The corresponding findings are presented in section 6.7, also labeled as “Results I”. Throughout the empirical results, only condensed summary findings will be presented in the text section while more detailed results can be found in the corresponding appendices. In a second step, sections 6.8 and 6.9 investigate the effects of insider ownership on the corporate control setting (“Results II”) and corporate performance (“Results III”) respectively. The former primarily illustrates why insider ownership plays an important role in the corporate control setting. As claimed by S HORT (1994), the lack of research regarding the relationship between (insider) owner741 742

See Edwards and Weichenrieder (2004), p. 165. Cf. K¨oke (2002), pp. 139-140.

207

Empirical Analysis of Insider Ownership

Corporate Control

Results II Firm Characteristics • Determinants of insider ownership • Firm size (LN_ASSETS) • Firm specific risk (FIRM_RISK) • Other firm characteristics • Sales growth (SALES_G) • Dividends (DIV) • Leverage (DEBT_RATO) • Industry

Insider Ownership Results I

• Main category level • All insider types (MGMT) • Sub-category level • Management board members (MB) • Supervisory board members (SB) • Former board members (FBM)

Results III

• Size of the supervisory board (SB_NO) • Size of the management board (MB_NO) • Level of voting restrictions (VOTE) • Difference between MGMT‘s control and cash flow rights (MGMT_CRmCF) • Level of outside blockholdings (BLOCK_O) • Number of outside blockholdings (BLOCK_NO) • Leverage (DEBT_RATIO)

Corporate Performance • Market-to-book value (MTBV) • Buy and hold stock returns (BAHR)

Results III

• Tobin’s Q (Q) • Return on assets (ROA) • Return on equity (ROE)

Cause analysis of insider ownership

Effect analysis of insider ownership

Source: Author’s findings.

Figure 6-16: Overview of Methodology ship and control constitutes a serious omission of much of the previous work.743 Therefore, the same statistical standard procedures, as in the cause analysis, are also applied in the analysis of the effects of insider ownership. The latter addresses the primary research question of this study whether insider ownership has an effect on corporate performance or not. Since this question has attracted a lot of attention in prior — primarily U.S. — research, the regression analyses will be more extensive, testing for different effects of the insider ownership variables, different model specifications and different time periods. In the pooled sample, the possibility of non-independent observations is addressed by estimating fixed-effects models which include two dummy variables to account for possible differences between the sub-samples.744 Furthermore, the reliability of the results is stressed by a series of robustness tests, which examine the sensitivity of the results to the use of alternative performance measures or whether endogeneity or multi-collinearity are likely to having biased the results. Finally, as the complexity of empirical corporate governance studies is non trivial, sec743 744

See Short (1994), p. 227. See Sanders (2001), p. 484; Anderson and Reeb (2003), p. 1315.

208

Empirical Analysis of Insider Ownership

tion 6.10 will summarize the limitations of this study and discuss how they might have influenced the presented results.745 6.7

Results I: Determinants of Insider Ownership

This section presents the first part of the empirical results, which investigates the causes of insider ownership as illustrated in figure 6-16. Therefore, the influence of firm size (LN ASSETS) and firm specific risk (FIRM RISK) on the level of insider ownership — measured on the aggregate level (MGMT) — will be analyzed. 6.7.1

Simple Regression Analysis

As a first analysis, simple linear regressions (OLS) are performed to examine the isolated influence of the two proposed firm characteristics on the level of insider ownership. The level of insider ownership in sections 6.7 to 6.9 always refers to cash flow rights if not indicated otherwise.746 The signs of the coefficients of firm size (LN ASSETS) are negative and highly significant (see table 6-20) indicating that larger firms tend to show lower levels of insider ownership. Regarding the goodness of fit of the regression models, firm size is able to explain between 9.3% (2003) and 14.1% (1993) of the variation of insider ownership (see R2 in table B-1 of Appendix B747 ). In 2003, an increase (decrease) of total assets by 10% (30%) — departing from the mean748 — results in an average decrease (increase) of insider ownership by -0.5% (+0.4%). The signs of the coefficients of firm specific risk (FIRM RISK) are positive and significant in 1998 and the pooled sample . Hence, riskier firms show on average higher levels of insider ownership. It should be noted that the explanatory power of the regression models is relatively small, i.e. they explain at most 2.9% 745 746

747

748

Cf. Bhagat and Jefferis (2002), pp. 9-42; B¨orsch-Supan and K¨oke (2002). This convention is introduced to ensure the comparability of the results. Furthermore, the focus of interest (i.e. insider ownership and performance) is theoretically derived from alignment of interest considerations which apply for cash flow and not control rights. Peterson (1965), p. 18, argues that cash flow rights matter more than control rights: “[. . . ] [T]he main strength of the stockholder’s position lies in the economics of it, not in the votes that support it.” Not that table 6-20 shows only the values of R2 adj. while the tables in the appendices show both R2 and R2 adj. The mean for LN ASSETS in 2003 is 12.875, which represents an average of total assets of EUR 390.4m. See table 6-16; footnote 734.

209

Empirical Analysis of Insider Ownership Sample Period

1993

1998

Independent Variable

Dependent Variable: MGMT

LN ASSETSb n R2 adj.

-0.066 214 0.137

FIRM RISKb n R2 adj.

0.110 165 0.009

a b

Pooleda

2003

***

-0.058 349 0.120

***

-0.046 261 0.089

0.088 244 0.025

***

0.009 286 0.000

***

-0.055 824 0.113

***

0.017 695 0.009

*

The pooled models includes two dummy variables to account for differences between the sub-samples. OLS-regression coefficients; *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table B-1 of Appendix B. Source: Author’s findings.

Table 6-20: Results I: Simple Regression (1998) and at least as few as 0.4% (1993) of the variation of the level of insider ownership from its mean. The results of a correlation analysis are not presented since Bravais-Pearson correlation coefficients are standardized coefficients from a simple regression of one variable on another. Hence, the results are essentially the same as those in the simple regression analysis. With respect to the correlation between firm size and firm specific risk a significant negative relationship — correlation coefficients between -0.367 (2003) and -0.247 (1993) — can be observed. This appears intuitively plausible as smaller firms are assumed to usually bear higher (unsystematic) risk. 6.7.2

Difference in Means

Besides the preceding simple regression analysis, an independent sample t-test of differences in means is conducted.749 Therefore, the sub-samples and the pooled sample are each grouped into two equally sized sub-groups according to firm size (see table 6-21) and firm specific risk (see table 6-22). Then, the means of the level of insider ownership (MGMT) of the smaller and larger as well as 749

Theoretically, this test would be adequate e.g. to test the hypothesis that smaller (larger) firms show higher (lower) levels of insider ownership which would represent basically an alternative but directional formulation of H-5.1.1. It was applied, to show that the results are robust to the type of hypotheses (effect vs. difference). Cf. Laatz (1993), pp. 525-529.

210

Empirical Analysis of Insider Ownership

Sample Period

1993

1998

2003

Pooled

11.456 14.251

11.080 14.049

11.305 14.457

11.236 14.237

0.387 *** 0.205

0.374 *** 0.203

0.383 *** 0.196

***

13,424 35,832 9,582 *** 25,244

20,154 *** 14,037

200,150 *** 139,750

***

Grouping Variable LN ASSETS

Mean (Small) Mean (Large)

Panel A: T-Test for Independent Samples MGMT

Mean (Small) Mean (Large)

0.375 0.181

Panel B: Mann-Whitney U-Test MGMT

Rank sum (Small) Rank sum (Large)

*, ** and *** denote differences in means significant on the 0.10, 0.05 and 0.01 level (2-sided). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table B-2 of Appendix B. Source: Author’s findings.

Table 6-21: Results I: Difference in Means I the less risky and riskier sub-groups are compared. Last, it is examined whether potential differences between the sub-groups are significant from a statistical point of view. While Panel A shows results from the prevalent t-test for independent samples750 , Panel B lists results from the less known Mann-Whitney U-test, which does not assume normally distributed data as the t-test. However, the non-parametric Mann-Whitney U-test represents the theoretically appropriate test as for all but the 1993 data of LN ASSETS the null hypothesis of normality can be rejected on the 0.05 level on basis of a Kolmogorov-Smirnov goodness of fit test.751 With respect to firm size the means of the small and the large firm size subgroups vary heavily. In the pooled sample, the mean value for LN ASSETS of the small size group is 11.236 (representing total assets of EUR 75.8m) while the value of the large size group reaches 14.237 (representing total assets of EUR 1,524.2m, i.e. a twentyfold higher value). The corresponding average 750

751

Theoretically, for the pooled case the observations are not fully independent, hence, limiting the expressiveness of the test. Cf. Mosler and Schmid (2004), pp. 254-262. The Mann-Whitney U-test does not provide a real comparison in means but examines whether the sample sub-groups (i.e. small and large or riskier and less risky firms) stem from a population with different distribution attributes. Cf. Laatz (1993), p. 524.

211

Empirical Analysis of Insider Ownership Sample Period

1993

1998

2003

Pooled

0.186 0.533

0.209 0.864

0.397 2.612

0.245 1.542

0.205 0.303

0.270 0.309

0.210 0.309

***

108,902 132,958

***

Grouping Variable FIRM RISK

Mean (Low risk) Mean (High risk)

Panel A: T-Test for Independent Samples MGMT

Mean (Low risk) Mean (High risk)

0.164 0.268

**

6,278 7,417

13,733 ** 16,158

**

Panel B: Mann-Whitney U-Test MGMT

Rank sum (Low risk) Rank sum (High risk)

19,252 ** 21,790

*

*, ** and *** denote differences in means significant on the 0.10, 0.05 and 0.01 level (2-sided). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table B-3 of Appendix B. Source: Author’s findings.

Table 6-22: Results I: Difference in Means II

level of insider ownership in the small size sub-group reaches 38.3% while the larger size group only reaches 19.6%. Hence, larger firms tend to show lower insider ownership levels. This difference remains relatively constant across the sub-samples and is always significant on the 0.01 level. Similarly, the MannWhitney U-test rejects the hypothesis that both sub-groups stem from the same population on the 0.01 level for all four sub-samples. Regarding firm specific risk, the large difference in means of the low (0.397) and high risk (2.612) sub-groups in the 2003 sub-sample attracts attention. Recalling that FIRM RISK is measured as the sum of squared residuals (SSE) from a regression of individual stock returns on market returns over a period of 60 months, it appears plausible that the period from January 1999 to December 2003 shows the highest variation, since this period first included a sharp rise and then a deep fall in stock prices (burst of the new economy). Looking at the mean levels of insider ownership in the low and high risk sub-groups it becomes obvious that riskier firms show on average higher levels of insider ownership. For the pooled model, the high risk sub-group shows an average of 30.9% while the low risk sub-group only reaches 21.0%. The differences are

212

Empirical Analysis of Insider Ownership

Sample Period

1993

1998

Independent Variables

Dependent Variable: MGMT

Constantb

1.030

***

LN ASSETSb

-0.063

FIRM RISKb n R2 adj. a b

Pooleda

2003

0.869

***

0.896

***

0.957

***

*** -0.049

*** -0.047

***

-0.052

***

0.000

0.049

-0.003

0.001

151 0.146

238 0.095

258 0.088

647 0.110

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients; *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table B-4 of Appendix B. Source: Author’s findings.

Table 6-23: Results I: Multiple Regression significant at least on the 0.05 level for all but the 2003 sub-sample. This finding is also confirmed by the Mann-Whitney U-test. Hence, insider ownership tends to be higher in riskier firms even though the average difference is not as large as between smaller and larger firms. By splitting up the samples in a higher number of sub-groups (e.g. quartiles: very small, small, large, and very large firm size) a sensitivity analysis was performed. A one-factor analysis of variance (ANOVA) leads to comparable results as the two group case, indicating that the differences depend not on the choice of grouping. Since the ANOVA-analysis does not provide new insights, the corresponding results are not presented. 6.7.3

Multiple Regression Analysis

While the previous two sections examined the isolated influence of firm size and firm specific risk on the level of insider ownership, this section analyzes their combined influence in a multiple regression framework (see table 6-23). The pooled sample additionally includes two dummy variables to account for differences in the level of insider ownership between the sub-samples. In the multiple regression analyses firm size and firm specific risk have the same signs as in the simple regressions, i.e. firm size has a negative effect on the level of insider ownership while firm specific risk has a positive one. Notably,

Empirical Analysis of Insider Ownership

213

the regression coefficients of firm size remain significant on the 0.01 level for all sub-samples and the pooled model, while those of firm specific risk are not significant in any of the multiple regression models. The explanatory power of the multiple regression models ranges from 9.5% in 2003 to 15.7% in 1993 and, hence, is only slightly higher than in the simple regressions of firm size on insider ownership (see R2 in table B-4 of Appendix B). 6.7.4

Discussion of Results

To summarize the knowledge gained about determinants of insider ownership, table 6-24 consolidates the results with regard to the proposed hypotheses. With respect to firm size the results are pretty clear and convincing. In line with hypothesis H-5.1.1, firm size has a highly significant negative effect on the level of insider ownership.752 This appears intuitively logical, if the typical life cycle of firms is taken into account: The entrepreneur, typically owning 100% of the equity in the beginning, aims at growing the business and, hence, is in need of external funding. Assuming further that debt can only be obtained up to a certain debt to equity ratio, the owner must raise outside equity at some point and, therefore, reduce his ownership stake in the firm below 100%. As a consequence, this pattern usually reduces the level of insider ownership as firms grow.753 Similar findings of a strong negative impact of firm size on insider ownership are also found by J ENSEN , S OLBERG AND Z ORN (1992). Their simultaneous equation analysis yields a highly significant relationship between firm size and insider ownership while debt ratio, dividend payout, business risk, R&D expenditures and the number of operating divisions fail to consistently explain insider ownership levels on a significant level.754 Finally, also G RAHAM AND H ARVEY (2001) report a highly significant negative correlation between firm size and management ownership.755 752

753 754

755

For example, a similar relationship is also detected by Mueller (2002), p. 19, for the case of private German companies. Cf. Jensen and Meckling (1976); section 4. See Jensen, Solberg and Zorn (1992), pp. 257-259. Only R&D expenditures and the number of operating divisions as independent variables show significant coefficients in one out of the two sample periods. See Graham and Harvey (2001), pp. 194-195. Similarly, insider ownership is also perceived to decrease with the IPO-age (and therefore along with firm size): Berry, Fields and Wilkins (2006), p. 4, analyzing U.S. IPOs report that CEO ownership declines from 11.3% just after the IPO to 3.6% 11 years later. Similarly, Harjoto and Garen (2005), p. 671, analyzing U.S. IPOs find that mean insider ownership decreases from 42.2% at the IPO to 28.7% in the third year after the IPO. For Germany, Ehrhardt, Nowak and Weber (2006), pp. 12-13, report that founding family ownership decreases rather slowly over time.

214

Empirical Analysis of Insider Ownership

Number

Hypothesis

Expected Influence

H-5.1.1

Firm size has an effect on the level of insider ownership.

−

H-5.1.2

Firm specific risk has an effect on the level of insider ownership.

A: + / B: −

Empirical Results

A:

 

/ B:

denotes that based on the empirical results the hypothesis should not be rejected; ( ) denotes that based on the empirical results the hypothesis should not be rejected - even though the evidence is weaker than for ; denotes that the hypothesis should be rejected based on the empirical results which lack significance on traditional levels. Source: Author’s findings.

Table 6-24: Results I: Review of Hypotheses

The case of firm specific risk appears to be more ambiguous. Firm specific risk has a significant positive influence on insider ownership in the simple regression analysis and high risk firms show higher mean insider ownership levels in the difference in means comparison. These results are in favor of interpretation A of hypothesis H-5.1.2: Insiders in firms with high firm specific risk — and therefore a larger scope for moral hazard — signal their commitment to other shareholders by holding larger stakes in their companies. However, the positive effect of firm specific risk is weaker than that of firm size and disappears in the multiple regression analysis. Remembering that firm size and firm specific risk are negatively correlated on the 0.01 significance level, the relation between firm risk and insider ownership disappears once a correction for firm size is taken into account as it has been done in the multiple regression analysis. Hence, the influence of firm specific risk on insider ownership just might be spurious. As firm age or industry affiliation, firm specific risk is significantly related to firm size which underlines the problem of finding appropriate determinants of insider ownership which are not simultaneously related to firm size.756 As a consequence, of the two examined firm characteristics only firm size turns out to monotonically determine insider ownership.

756

Cf. Williams (1990), p. 1635.

Empirical Analysis of Insider Ownership 6.8

215

Results II: Insider Ownership and Corporate Control

Having presented evidence on the determinants of insider ownership in the previous section, the second part of the empirical results investigates the effects of insider ownership on corporate control (see figure 6-16). Therefore, seven different aspects related to the corporate control setting of a company, namely the size of the supervisory board (SB NO) and management board (MB NO), the level of voting restrictions (VOTE), the difference between insiders’ control and cash flow rights (MGMT CRmCF), the level of outside blockholdings (BLOCK O), the number of outside blockholdings (BLOCK NO) and the leverage (DEBT RATIO) are examined. 6.8.1

Simple Regression Analysis

First, simple linear regressions (OLS) are performed to examine the isolated influence of insider ownership on the proposed characteristics relating to the corporate control setting. With respect to the size of the supervisory board (SB NO), the regression coefficients are negative for all sub-samples and highly significant (0.01 level) for all except the 1993 sub-sample (see table 6-25). In 2003, a 10% higher insider ownership share — the mean of MGMT’s cash flow rights is 29.0% — would on average be associated with a reduction in the size of the supervisory board by 0.58 persons (departing from a mean of 8.63 persons). Insider ownership is able to explain 11.5% of the variation of SB NO (see table C-1 of Appendix C). Regarding the size of the management board (MB NO), the regression coefficients are negative and highly significant (0.01 level) for all except the 1993 sub-sample. In 2003, a 10% higher insider ownership share would come along with an average reduction in the size of the management board by 0.10 persons (departing from a mean of 3.27 persons). However, besides the minor economic significance, insider ownership is only able to explain 3.2% of the variation of MB NO (see table C-1 of Appendix C). Analyzing the effects of insider ownership on the presence of voting restrictions (VOTE), the regression coefficients are positive and significant on the 0.05 level in 1993 and the pooled sample. Hence, higher levels of insider ownership result in an increased presence of voting restrictions. However, the rather low explanatory power of the simple regression models ranging from 0.1% in 2003 to 2.7% in 1993 does not suggest further interpretation. Contrary, the similar

216 Sample Period

Empirical Analysis of Insider Ownership 1993

1998

Independent Variable MGMTb n R2 adj.

Pooleda

2003

Dependent Variable: SB NO -0.100 236 -0.004

-7.438 356 0.182

***

-5.828 284 0.112

***

-4.914 876 0.102

***

-0.779 873 0.036

***

0.076 884 0.007

**

Dependent Variable: MB NO MGMTb n R2 adj.

0.150 236 -0.004

-1.208 354 0.037

***

-1.034 283 0.029

***

Dependent Variable: VOTE MGMTb n R2 adj.

0.167 236 0.023

**

0.055 357 0.000

0.024 290 -0.003

Dependent Variable: MGMT CRmCF MGMTb n R2 adj.

0.129 236 0.126

***

0.078 358 0.043

***

0.063 290 0.037

***

0.087 884 0.065

***

-0.848 884 0.506

***

Dependent Variable: BLOCK O MGMTb n R2 adj.

-0.874 235 0.548

***

-0.861 358 0.518

***

-0.808 290 0.457

***

Dependent Variable: BLOCK NO MGMTb n R2 adj.

-1.992 236 0.364

***

-1.883 358 0.318

***

-1.964 290 0.254

***

-1.939 884 0.308

***

Dependent Variable: DEBT RATIO MGMTb n R2 adj. a b

0.002 206 -0.050

-0.043 341 0.002

0.077 253 0.004

0.006 801 0.048

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients; *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to tables C-1 to C-3 in Appendix C. Source: Author’s illustration.

Table 6-25: Results II: Simple Regression

Empirical Analysis of Insider Ownership

217

measure related to the difference between insiders’ control and cash flow rights (MGMT CRmCF) yields clearer insights: The coefficients are positive and always significant on the 0.01 level. In 1993, a 10% higher insider ownership level is associated with on average 1.29% higher excessive control rights and the corresponding R2 is 12.9% (see table C-2 of Appendix C). The results for the other sub-samples are similar but small in economic significance and model fit. With respect to outside blockholdings both, the level of outside blockholdings (BLOCK O) as well as the number of outside blockholdings (BLOCK NO), are negatively related to the level of insider ownership. In 2003, a 10% higher insider ownership share results in an 8.08% lower level of outside blockholdings representing almost an 1:1 substitution rate (R2 = 45.9%).757 Simultaneously, the number of blockholders decreases on average by 0.19 (R2 = 25.7%). The coefficients are comparable across the different sub-samples (i.e. 1993, 1998 and the pooled sample), suggesting a rather stable relationship between insider ownership and outside blockholdings. Finally, the impact of the level of insider ownership on firms’ leverage (DEBT RATIO) is examined. The coefficients lack economic and statistical significance in all sub-samples and the model fit is rather low. Hence, no evidence is found that the level of insider ownership effects firms’ financial structure in an observable manner. 6.8.2

Correlation Analysis

Besides the simple regression analysis, a correlation analysis provides further interesting insights into firms’ corporate control setting as it also shows the relationships between the individual variables. In the first column of table 6-26, the correlation coefficients between insider ownership (MGMT) and the seven corporate control variables show the same signs as in the simple regression an757

The regression of the level of insider ownership on the level of outside blockholdings is not unproblematic since both are partially dependent as the sum of all ownership shares must always add up to 100%. Hence, a large insider ownership share of e.g. 80% (α) limits the maximum value for the level of outside blockholdings to 20% (1 − α). In the two group case the correlation among both measures (i.e. α and 1 − α) must be -1 and hence regression results would be redundant. However, since the shareholder classification applied in this study differentiates among more categories and the level of outside blockholdings also can be 0, the analysis makes sense, even though results should be interpreted with caution.

218

Empirical Analysis of Insider Ownership

Sample Period

Pooled Sample

Covariates

MGMT

SB NO

−−−

MB NO

−−−

+++

++

(+)

(+)

MGMTCRmCF

+++

(+)

++

+++

BLOCK O

−−−

+++

(−)

−−−

−−−

BLOCK NO

−−−

+++

++

−−−

−−−

+++

(+)

+

(+)

+++

+++

−−−

VOTE

DEBTRATIO

SB NO

MB NO

VOTE

MGMTCRmCF

BLOCK O

BLOCKNO

+

Signs of Bravais-Pearson correlation coefficients; +, ++ and + + + denote significant positive correlations coefficients on the 0.10, 0.05 and 0.01 level (2-sided); −, −− and − − − denote significant negative correlations on the 0.10, 0.05 and 0.01 significance level (2-sided); (+) and (−) denote positive/ negative correlation coefficients which are not significant on traditional levels. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table C-4 in Appendix C. Source: Author’s findings.

Table 6-26: Results II: Correlation Analysis alyses from the previous chapter.758 The negative relation between MGMT and BLOCK O is highly significant and underlines that high levels of insider ownership may reduce the need for other costly monitoring mechanisms.759 As the second column of table 6-26 shows, the number of supervisory board members (SB NO) is positively and significantly (0.01 level) correlated with the number of management board members (MB NO) and with the level (BLOCK O) and number of outside blockholdings (BLOCK NO). The latter 758 759

See section 6.7.1. Contrary, Krivogorsky (2006), p. 189, finds a positive correlation between insider ownership and block ownership.

Empirical Analysis of Insider Ownership

219

might reflect the wants of outside blockholders to send representatives to the supervisory board, which in turn might induce firms to increase the size of their supervisory boards. Finally, the correlation coefficients of the two measures which measure the presence of voting restrictions or excessive control rights, i.e. VOTE and MGMT CRmCF, with BLOCK O and BLOCK NO as well as with DEBT RATIO appear interesting. The former are negative, significant on the 0.01 level, indicating that outside blockholders avoid to invest in companies where management owns excessive control rights. The latter are positive and significant on the 0.01 level, confirming the presumption that founder-managers try to minimize outside influence by first increasing debt levels and then, if further financing is needed, by issuing non-voting equity.760 6.8.3

Difference in Means

Next, a difference in means analysis is conducted, applying both the parametric t-test for independent samples and the non-parametric Mann-Whitney U-test.761 Therefore, the sub-samples where divided into two equally sized groups, according to the levels of insider ownership (low insider ownership (Low IO) vs. high insider ownership (High IO)). It is noteworthy that the mean ownership levels in the low insider ownership groups vary between 0.0% (1993) and 2.1% (2003) while they range from 55.5% (1993) to 57.6% (1998) in the high insider ownership sub-groups (see table 6-27). As in the previous section, for all data the null hypothesis of normal distribution must be rejected on the 0.05 significance level. Hence, Panel B of table 6-27 represents the statistically correct test. However, the results of the more common T-test in Panel A are almost identical. With respect to the number of supervisory board members and the number of management board members, the results strengthen the findings of the simple regression analysis: Firms with a high level of insider ownership show significantly smaller board sizes. While the influence of the level of insider ownership on the presence of voting restrictions in the simple regressions turned out to be significant only in the 1993 sub-sample and in the pooled sample, the difference in means comparison shows that high insider ownership firms on average 760 761

See section 5.2. For details on the tests refer to section 6.7.2.

220

Empirical Analysis of Insider Ownership

Sample Period

1993

1998

2003

Pooled

0.000 0.555

0.016 0.576

0.021 0.559

0.012 0.566

(−)

−−−

−−−

−−−

(+)

−−−

−−

−−−

+++

+++

+

+++

+++

+++

+++

+++

−−−

−−−

−−−

−−−

−−−

−−−

−−−

−−−

(+)

(−)

(+)

(+)

(−)

−−−

−−−

−−−

(+)

−−−

−

−−

+++

+++

+

+++

+++

+++

+++

+++

−−−

−−−

−−−

−−−

−−−

−−−

−−−

−−−

(+)

(−)

(+)

(+)

Grouping Variable MGMT

Mean (Low IO) Mean (High IO)

Panel A: T-Test for Independent Samples SB NO

Low IO High IO MB NO Low IO High IO VOTE Low IO High IO MGMT CRmCF Low IO High IO Low IO BLOCK O High IO Low IO BLOCK NO High IO Low IO DEBT RATIO High IO Panel B: Mann-Whitney U-Test SB NO

Low IO High IO Low IO MB NO High IO VOTE Low IO High IO MGMT CRmCF Low IO High IO Low IO BLOCK O High IO Low IO BLOCK NO High IO Low IO DEBT RATIO High IO

+, ++ and + + + denote that the value for the high insider ownership sub-group is significantly higher than for the low insider ownership sub-group on the 0.10, 0.05 and 0.01 level; −, −− and − − − denote that the value for the high insider ownership sub-group is significantly lower than for the low insider ownership subgroup on the 0.10, 0.05 and 0.01 level; (+) and (−) denote higher/lower values for the high insider ownership sub-group which are not significant on traditional levels. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to tables C-5 and C-6 in Appendix C. Source: Author’s findings.

Table 6-27: Results II: Difference in Means

Empirical Analysis of Insider Ownership

221

are characterized by significantly higher levels of voting restrictions (0.05 significance level in the 2003 sub-sample and 0.01 level in the other sub-samples). In line with previous analyses, firms with higher insider ownership levels are characterized by significantly (0.01 level) larger insiders’ excess control rights as well as lower levels of and fewer outside blockholdings. While the excess control rights for the low insider ownership group are below 0.1% for all subsamples, the high insider ownership group shows mean excess control rights between as much as 9.9% in 1993 and 5.9% in 2003 (see table C-5 in Appendix C). Therefore, it seems that insiders’ excess control rights on average decreased during the decade from 1993 to 2003. As assumed, outside blockholdings are significantly larger and more frequent in the low insider ownership sub-groups than in the high insider ownership subgroups. In 2003, the average level of blockholdings in the low insider ownership sub-group amounts to 56.9% while it reaches only 7.2% in the high insider ownership sub-group. This underlines the widespread existence of outside ¨ blockholdings and is in line with the evidence found by B ECHT AND B OHMER (2003) that 82.3% of listed German firms have a minority blockholder who controls more than 25% and 64.7% of listed firms are even majority controlled by blockholders.762 Similarly, the number of outside blockholders is significantly (0.01 level) smaller for the high insider ownership sub-groups throughout the sub-samples and the pooled sample. Thus, it seems that outside block ownership is a substitute to insider ownership and, hence, both ownership phenomenons should be taken into account in the analysis.763 Finally, with regard to differences in the degree of leverage, high insider ownership companies tend to show slightly higher debt levels with exception of the 1998 sub-sample. However, the differences are relatively small, not significant and hence might represent a random result without any explanatory power.

762 763

See Becht and B¨ohmer (2003), p. 8. For U.S. firms e.g. Villalonga and Amit (2006), p. 396, report that equity ownership by outside blockholders is significantly higher in non-family firms than in family firms.

222

Empirical Analysis of Insider Ownership

Number

Hypothesis

Expected Influence

H-5.2.1

The level of insider ownership has an effect on the size of the supervisory board.

−

H-5.2.2

The level of insider ownership has an effect on the size of the management board.

−

H-5.2.3

The level of insider ownership has an effect on the level of voting restrictions.

+

H-5.2.4

The level of insider ownership has an effect on the difference between insiders’ control and cash flow rights.

+

H-5.2.5

The level of insider ownership has an effect on the level of outside blockholdings.

−

H-5.2.6

The level of insider ownership has an effect on the number of outside blockholdings.

−

H-5.2.7

The level of insider ownership has an effect on the level of firms’ leverage.

A: + / B: −

Empirical Results

 

A:

/ B:

denotes that based on the empirical results the hypothesis should not be rejected; ( ) denotes that based on the empirical results the hypothesis should not be rejected - even though the evidence is weaker than for ; denotes that the hypothesis should be rejected based on the empirical results which lack significance on traditional levels. Source: Author’s findings.

Table 6-28: Results II: Review of Hypotheses 6.8.4

Discussion of Results

To consolidate the findings about the effects of insider ownership on the corporate control setting, table 6-28 reports the acceptance or rejection of the proposed hypotheses based on the presented empirical results. With regard to supervisory board size and management board size strong evidence is found in favor of hypotheses H-5.2.1 and H-5.2.2 that the level of insider ownership has a significantly negative effect on board sizes. With regard to the supervisory board, this might reveal insiders preference to limit the external control potential, which in turn usually grows if an increasing number of potentially diverse interests is represented in the supervisory board. Com-

Empirical Analysis of Insider Ownership

223

parisons with other studies are difficult since the relationship between insider ownership and board size was rarely examined. Regarding the size of the management board, insiders with large ownership stakes apparently avoid dividing the power amongst too many individuals. Even though the acceptance of the proposed hypotheses appears plausible from a theoretical perspective, it should be acknowledged that the findings also may be spurious results related to firm size: Larger firms — which show lower levels of insider ownership — are also expected to have larger boards due to larger operations (e.g. international, several segments) or higher staff. Also the second pair of control variables, the presence of voting restrictions and the difference between insiders’ control and cash flow rights, suggests that insiders owning significant shares in their companies try to minimize the capability of outside control. For the former (H-5.2.3), only weak evidence is found that the level of insider ownership has a positive effect on the presence of deviations from the one-share-one-vote principle. For the latter (H-5.2.4), strong evidence is found that insiders feel discomfort to pass on voting rights to outside shareholders if they are in need of external funding.764 The hypothesized substitutional relationship between insider ownership and outside blockholdings (H-5.2.5 and H-5.2.6) is strongly corroborated by the empirical results. It seems that both, high levels of insider ownership as well as the presence of outside blockholders are means of either aligning the interests of outside shareholders and management or limiting management’s scope for discretionary action by monitoring. How each of these instruments is related to corporate performance will be the subject of the next section. Finally, the hypothesis that insider ownership has an impact on debt levels (H5.2.7) must be rejected on the results at hand. This is in line with the findings of C HAPLINSKI AND N IEHAUS (1993) that leverage and inside equity are not explained by the same variables.765 The findings are opposed to K IM AND S ORENSEN (1986), M EHRAN (1992), AGRAWAL AND K NOEBER (1996) or B ERGER , O FEK AND Y ERMACK (1997) whose results indicate a signifi764

765

Cf. Andres (2006), p. 10, who reports that family firms’ affinity to issue non-voting preferred stock is significantly higher. Cf. Chaplinski and Niehaus (1993), pp. 55-63.

224

Empirical Analysis of Insider Ownership

cant positive relationship between insider ownership and debt ratios766 and to J ENSEN , S OLBERG AND Z ORN (1992) who find a negative relationship between insider ownership and leverage.767 6.9

Results III: Insider Ownership and Corporate Performance

After having presented evidence on the determinants of insider ownership and its effects on the corporate control setting, the third and most extensive empirical part examines the effects of insider ownership on corporate performance (see figure 6-16). As corporate performance is influenced by a large number of different factors, more emphasis is put on multiple regression analysis, as e.g. simple regression analysis does not yield any meaningful results.768 Thus, only a brief look at the correlation matrix and the difference in means tests is taken before the multiple regression models, which take the center stage of the analysis, will be presented. 6.9.1

Correlation Analysis

As can be seen in table 6-29, all three performance measures, i.e. buy-and-hold returns (BAHR), market-to-book values (MTBV) and return on assets (ROA), are positively correlated.769 While the correlation coefficients between BAHR and MTBV (0.344) and BAHR and ROA (0.242) are significant on the 0.01 level, that of MTBV and ROA (0.015) is close to zero and not significant on the 0.1 level. Looking at the correlations between performance and the insider ownership variables yields rather ambiguous results. Out of the five bivariate correlation coefficients which turn out to be significant at least on the 0.05 level, four show positive signs (MTBV/MB, MTBV/MGMT, ROA/SB, ROA/MGMT) and only one shows a negative sign (BAHR/MB). The remaining seven correlation coefficients are not significantly different from zero. Hence, even though the 766

767 768

769

See Kim and Sorensen (1986), pp. 134-140; Mehran (1992), pp. 551-555; Agrawal and Knoeber (1996), p. 388; Berger, Ofek and Yermack (1997), p. 1419. It should be noted that this study uses CEO rather than officers’ and directors’ ownership as explanatory variable. See Jensen, Solberg and Zorn (1992), pp. 256, 259-260. McConnell and Servaes (1990), p. 605, express concerns that in such cases any observed relation between insider ownership and performance might be a spurious correlation between these two variables and an omitted variable. The reported correlations are based on the pooled sample. The results for the individual sub-samples, not presented for the sake of clarity, do not differ in a meaningful way from those reported.

225

Empirical Analysis of Insider Ownership Sample Period

Pooled Sample

Covariates

BAHR

MTBV

+++

ROA

+++

(+)

MB

−−−

+++

SB

(+)

(−)

FBM

(−)

(−)

(+)

−−−

−−−

MGMT

(−)

++

+++

+++

+++

+++

+++

(+)

(−−)

−−−

−−−

−−−

BLOCK O

MTBV

ROA

MB

SB

FBM

MGMT

(+) +++ −−−

−−−

Signs of Bravais-Pearson correlation coefficients; +, ++ and + + + denote significant positive correlations coefficients on the 0.10, 0.05 and 0.01 level (2-sided); −, −− and − − − denote significant negative correlations on the 0.10, 0.05 and 0.01 significance level (2-sided); (+) and (−) denote positive/ negative correlation coefficients which are not significant on traditional levels. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table D-1 in Appendix D. Source: Author’s findings.

Table 6-29: Results III: Correlation Analysis I performance measures are positively correlated, the relationship between the different performance measures and insider ownership seems to be unclear. As other factors besides insider ownership are also likely to have an impact on corporate performance, the explanatory power of univariate analyses (as bivariate correlations) is limited. Regarding blockholdings, only the positive correlation between BAHR and BLOCK O (0.106) turns out to be significant (0.01 level). The highly negative correlations between insider ownership and blockholdings have already been discussed in the previous section.770 The assumption that univariate analyses of insider ownership and corporate performance are inadequately complex is supported by the highly significant correlation coefficients between the performance and control variables.771 As illustrated in table 6-30, 13 out of the 15 bivariate correlation coefficients for the three performance variables and the five non-ownership control variables 770 771

See section 6.8. As already stated, the control variables have mainly been derived from comparable prior research.

226

Empirical Analysis of Insider Ownership

Sample Period

Pooled Sample

Covariates

BAHR

MTBV

+++

ROA

+++

(+)

LN ASSETS

+++

−−

++

FIRM RISK

−−−

(+)

−−−

−−−

DEBT RATIO

−−−

−−−

−−−

+++

+++

SALES G

+++

+++

+++

(−)

(−)

(−)

DIV

+++

(−)

+++

+++

−−−

−−−

MTBV

ROA

LNFIRMASSETS RISK

DEBT- SALES G RATIO

(+)

Signs of Bravais-Pearson correlation coefficients; +, ++ and + + + denote significant positive correlations coefficients on the 0.10, 0.05 and 0.01 level (2-sided); −, −− and − − − denote significant negative correlations on the 0.10, 0.05 and 0.01 significance level (2-sided); (+) and (−) denote positive/ negative correlation coefficients which are not significant on traditional levels. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table D-2 in Appendix D. Source: Author’s findings.

Table 6-30: Results III: Correlation Analysis II (LN ASSETS, FIRM RISK, DEBT RATIO, SALES G and DIV) are significantly different from zero at least on the 0.05 level. For the case of BAHR, all control variables are significantly correlated on the 0.01 level. With regard to the bivariate correlation coefficients among control variables, only three out of ten coefficients show significant values above/ below 0.2. This concerns the already mentioned relationship between LN ASSETS and FIRM RISK (-0.301) and the correlations of DIV and LN ASSETS (0.366) and DIV and FIRM RISK (-0.400).772 The coefficients have the expected signs since it can reasonably be expected that larger firms are less risky and more likely to pay dividends while higher indebted firms are less likely to pay dividends. As problems of multi-collinearity can exist if independent variables in multiple regression models are correlated on higher levels, this issue will be 772

See table D-2 in Appendix D.

Empirical Analysis of Insider Ownership

227

addressed later on in section 6.9.4.3.773 Finally, also the positive correlations between DEBT RATIO and FIRM RISK as well as the negative correlation between DEBT RATIO and DIV are in line with economic reasoning: Higher leverage leads to an increase in the default risk and the payout of dividends appears less likely if simultaneously additional debt is taken on. 6.9.2

Difference in Means

Analogously to the analysis of insider ownership and corporate control, a difference in means comparison is conducted using both the t-test for independent samples and the Mann-Whitney U-test (see table 6-31).774 Therefore, the samples are divided into two equally sized groups of high (High IO) and low insider ownership companies (Low IO). Then, the sub-groups’ means for the three performance measures are compared. If BAHR are used as performance measure, the high insider ownership sub-group (High IO) shows lower stock returns in all three sub-samples as well as in the pooled sample, even though only the difference in the pooled sample is significant in on the 0.1 level (see Panel A). Contrarily, using MTBV or ROA as performance measures yields uniformly higher means for the high insider ownership sub-groups. The results for the 2003 sub-sample are not significant in any case with exception of the MannWhitney U-test of differences if BAHR is used as performance measure. In addition to the three main performance measures, the difference in means test is also performed for two control variables, namely SALES G and DIV. Companies with higher insider ownership levels show higher growth rates in all four sub-samples but the difference is significant (0.05 level) in only two sub-samples.775 The results for differences in the payment of dividends are ambiguous and not significant at all.776 To sum up, the difference in means tests do not allow a prudent judgement whether the performance of companies with higher levels of insider ownership is superior to those with lower levels of insider ownership. 773

774 775

776

Eckey, Kosfeld and Dreger (2001), p. 90, state that as a rule of thumb correlation coefficients of 0.8 and higher might indicate problems with collinearity. Cf. sections 6.7.2 and 6.8.3. Similarly, Villalonga and Amit (2006), p. 394, state that family firms show significantly higher growth rates than non-family firms. Cf. McConaughy, Matthews and Fialko (2001), pp. 40-46.

228

Empirical Analysis of Insider Ownership

Sample Period

1993

1998

2003

Pooled

0.000 0.555

0.016 0.576

0.021 0.559

0.012 0.566

Grouping Variable MGMT

Mean (Low IO) Mean (High IO)

Panel A: T-Test for Independent Samples BAHR

Low IO High IO

(−)

(−)

(−)

−

Low IO High IO

(+)

++

(+)

++

Low IO High IO

++

+++

(+)

+++

SALES G Low IO High IO

(+)

++

(+)

++

(+)

(+)

(−)

(−)

MTBV

ROA

DIV

Low IO High IO

Panel B: Mann-Whitney U-Test BAHR

MTBV

ROA

Low IO High IO

(−)

(−)

−

−−

Low IO High IO

(+)

(+)

(−)

(−)

Low IO High IO

+++

+++

(+)

+++

+

+++

(+)

+++

(+)

(−)

(+)

(−)

SALES G Low IO High IO DIV

Low IO High IO

+, ++ and + + + denote that the value for the high insider ownership sub-group is significantly higher than for the low insider ownership sub-group on the 0.10, 0.05 and 0.01 level; −, −− and − − − denote that the value for the high insider ownership sub-group is significantly lower than for the low insider ownership subgroup on the 0.10, 0.05 and 0.01 level; (+) and (−) denote higher/lower values for the high insider ownership sub-group which are not significant on traditional levels. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to tables D-3 and D-4 in Appendix D. Source: Author’s findings.

Table 6-31: Results III: Difference in Means

Empirical Analysis of Insider Ownership 6.9.3

229

Multiple Regression Analysis

Because of the novelty of this study, little is known about how to analyze effects of insider ownership on corporate performance in a multiple regression framework for the case of Germany. Hence, the following section 6.9.3.1 will develop a base case model for the sub-sample 2003. Section 6.9.3.2 will test alternative specifications derived from literature, before the base case model is applied to the sub-samples of 1998 and 1993 (sections 6.9.3.3 and 6.9.3.4). Finally, the pooled sample is examined (section 6.9.3.5) and the effects of changes in the level of insider ownership will be addressed in section 6.9.3.6.777 6.9.3.1 Base Case (Cross Section 2003) In addition to the three different performance measures (BAHR, MTBV and ROA), two different specifications of the level of insider ownership are used. This results in a total number of six different regression equations (labeled model 1 to model 6 as shown in table 6-32). As more knowledge must be gained about the appropriate measurement of insider ownership in Germany, the regression analyses are carried out using the three sub-category insider ownership variables as separate regressors (MB, SB and FBM) in models 1, 3 and 5 as well as the aggregated insider ownership variable (MGMT) in models 2, 4 and 6. For the 2003 sub-sample, OLS-regression results are presented in table 6-32, where models 1 and 2 use stock returns (BAHR) as dependent variable, whereas models 3 to 6 use market-to-book values (MTBV) and return on assets (ROA) respectively. As already discussed, no distinct relationship between insider ownership and corporate performance could be detected in the univariate analysis because a number of other factors is expected to influence both the level of insider ownership as well as corporate performance. Consequently, 14 control variables (thereof seven industry dummy variables) are included to isolate the partial effect of insider ownership on performance. The level of outside blockholdings (BLOCK O), i.e. the cumulative shareholdings of all outsiders which individually hold at least 5% of the voting rights, is included to account for possible substitutional effects between insider owner777

Similarly, Chen, Guo and Mande (2003) also present their Japanese results for nine sub-samples (i.e. the individual years from 1987 to 1995) and a pooled sample.

230

Empirical Analysis of Insider Ownership

ship, as an instrument to align shareholders’ and management’s interests, and outside block ownership, as an instrument to ensure effective monitoring.778 In addition, the number of outside blockholders (BLOCK NO) takes into account that the potential for effective monitoring decreases if the control rights are split up among an increasing number of outside parties. Strong arguments to include both, the fraction of shares owned by management and a variable measuring shareholder concentration, in regression analyses is made by D EMSETZ AND V ILLALONGA (2001): “The variation in the importance of these two types of owners, the five largest shareholders and the management, correlate positively across the firms in our sample, but not so much so as to allow a claim that one of these measures is redundant if the other is used. [. . . ] Therefore, other things being equal, a study that uses both measures to account for the complexity of interests represented by a given ownership structure should give a more accurate picture of the ownership–performance relation than those that rely on only one of the two measures.”779 Firm Size (LN ASSETS), measured by the natural logarithm of total assets, is included to account for the fact that insider ownership in very large corporations is less widespread.780 Firm risk (FIRM RISK) measures the unsystematic, diversifiable portion of a firm’s total risk. Unsystematic firm risk was chosen instead of systematic risk because beta factors — the common measure for firms’ systematic risk — turn out to be rather fluctuating over time if measured over a relatively long period of 60 months. As the resulting regression estimates of beta factors are too noisy, it is assumed that the systematic risk is captured by a constant and consequently can be neglected in the regression models.781 Hence, 778

779

780 781

Zingales (2000), p. 1641, questioning the theory of the firm notes that such monitoring by outside blockholders might not work at “all types of firms”. For example, the presence of blockholders also might protect underperforming firms from the takeover market as discussed when Porsche AG acquired a stake in Volkswagen AG in 2005. See Hussla (2005a), p. 37. Contrary, Agrawal and Mandelker (1990), pp. 153-155, find support for the monitoring benefits of blockholders. Demsetz and Villalonga (2001), p. 215. Cf. Anderson, Mansi and Reeb (2003), p. 265; Faccio and Lasfer (1999), p. 12; Lehmann and Weigand (2000), p. 162; Lins (2003), p. 172. See Graham and Harvey (2001), p. 195. The regression coefficients of the corresponding beta factors (see section 6.3.3) are significant at least on the 0.1 level in only 66.9% of all cases. Furthermore, the R2 of the corresponding models are relatively low and vary between 11.4% (2003) and 24.6% (1993). If the beta factor is included in the base case models as an additional independent variable — as illustrated in tables D-15 and D-16 of Appendix D — it turns out to be

Empirical Analysis of Insider Ownership

231

in the following only the unsystematic risk is considered as an independent variable. The inclusion of leverage (DEBT RATIO) in the regression model takes the disciplining effect of higher interest burdens on management’s behavior into account.782 The growth potential (SALES G), which is expected to be captured in the market valuation of equity, is proxied by the average annual sales growth over the past three years.783 It is included in the analysis to differentiate higher market valuations arising from higher growth potential from those that might be the result of lower agency costs due to the alignment of interest among management and shareholders. The dummy variable dividends (DIV) indicates whether dividends have been paid during the respective year.784 Finally, seven industry dummy variables are used — but not presented in the results for the sake of clarity — to account for heterogeneity among eight different industries.785 As can be seen from the last but one row of table 6-32, between 235 to 247 complete data sets are available for the total sample size of 290 companies in 2003 depending on the choice of the respective performance measure. In contrast to the univariate analysis in sections 6.9.1 and 6.9.2, in the multivariate analyses the signs of all but one (the exception is FBM in model 5) insider ownership coefficients in models 1 to 6 are positive, indicating a positive impact of insider ownership on firm performance. For the stock return models 1 and 2, all insider ownership coefficients turn out to be significantly different from zero at least on the 0.05 level. For the MTBV models 3 and 4, three of the four coefficients are significant at least on the 0.1 level and for the ROA models 5 and 6 only supervisory board ownership (SB) shows a significant coefficient (0.01 level). With regard to the explanatory power of the models it should be noted that the adjusted R2 is equal to 36.4% (average models 1 and 2) if BAHR are used as

782

783

784

785

significant only in the 2003 sub-sample (and therefore weakly significant in the pooled-sample). However, the coefficients of the insider ownership variables remain largely unchanged. See Demsetz and Villalonga (2001), p. 221. Cf. Short (1994), p. 228; Rennneboog (2000), p. 1985; Harijono and Tanewksi (2004). See Maury (2006), p. 325; Edwards and Weichenrieder (2004), p. 156. Cf. McConnell and Servaes (1995), pp. 135-137. Cf. Beiner, Drobetz, Schmid and Zimmermann (2006), p. 256; Edwards and Weichenrieder (2004), pp. 155156. For a more detailed description of the variables included refer to sections 6.3.1, 6.3.2 and 6.3.3. The use of size, leverage, sales growth and industry as control variables is also described as ‘standard in literature’ by Barontini and Caprio (2005), pp. 9-10.

247 0.362

n R2 adj.

*** * ***

*** *** ***

247 0.366

Yes

*** * ***

*** *** ***

***

* ** *

235 0.171

Yes

1.107 ** -0.126 0.175 ** 0.231 *** -1.978 *** 0.333 * 0.220

0.797 0.859 1.421

-0.685

(3)

**

235 0.174

Yes

1.128 -0.123 0.186 *** 0.222 ** -2.024 *** 0.302 * 0.187

1.080

-0.822

(4)

MTBV

***

246 0.186

Yes

0.773 0.734 0.794 ** -0.219 -11.533 *** 3.996 2.253

3.859 9.189 -0.581

-6.214

(5)

(6)

246 0.170

Yes

0.653 0.736 0.734 ** -0.188 -11.027 *** 3.970 2.400 *

3.979

-5.112

ROA

Table 6-32: Results III: Multiple Regression - Base Case (2003)

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the respective t-statistics refer to tables D-5 and D-6 in Appendix D. Source: Author’s findings.

Yes

93.867 -10.985 9.738 4.726 -80.448 26.629 60.492

93.402 -11.152 9.415 5.175 -79.910 28.175 61.620

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies

82.590

** ** **

67.875 91.900 82.465

-140.738

MB SB FBM MGMT

*

-137.945

Intercept

(2)

(1)

BAHR

Model No.

Dependent Variable

232 Empirical Analysis of Insider Ownership

Empirical Analysis of Insider Ownership

233

dependent variable, 17.3% (average models 3 and 4) if MTBV and 17.8% (average of models 5 and 6) if ROA are used as dependent variable. The lower explanatory power of the models using the latter two performance measures is in line with the view that accounting performance measures are rather noisy for German companies.786 Since the BAHR models have the highest explanatory power, BAHR will be used as the primary performance measure to test alternative specifications of the relationship between insider ownership and performance in the next section. The individual measures of insider ownership (MB, SB, FBM) do not convey any information superior to that derived from the use of the aggregated measure of insider ownership (MGMT) as the coefficients are similar in signs, size and significance.787 Thus, only the latter specification (i.e. model 2) will be discussed in more detail. The insider ownership coefficient of 82.6 — significant at the 0.01 level — states that on average an increase in insider ownership by 100 basis points results in an increase of 83 basis points of the five year stock price performance. Among the control variables, firm size, firm risk, growth potential and dividend payments have a positive effect on stock returns while leverage turns out to have a negative impact. While the positive effects of sales growth and dividend payments appear intuitively plausible, the other effects deserve further discussion. A positive relationship between firm size and performance can be expected if size is related to market shares and, hence, efficiency effects are present. Alternatively, if diversification improves performance, a positive relationship might be explained to the extent that size is related to diversification.788 One possible explanation for the negative effect of leverage might be that small and highly leveraged firms experienced more serious devaluations in their stock prices during the market downturn from 2001 to 2003.789 The positive sign of the coefficient of firm risk could signify that better performing firms, which managed to recover from the general drop in market values of equity between 2000 and 2003, necessarily show higher return variations than those which did not. 786 787

788

789

Cf. Edwards and Weichenrieder (2004), p. 152. The coefficients for MGMT tend to show higher significance levels presumably because the number of instances is higher for the aggregate than for the individual measures. See Gugler, Mueller and Yurtoglu (2004), p. 8. However, this is not in line with the usually observed sizeeffect that smaller firms are characterized by better performance. A negative effect of leverage on firm performance as measured by Tobin’s Q is found by Villalonga and Amit (2006), p. 400.

234

Empirical Analysis of Insider Ownership

Finally, the results strongly support the presumption that board ownership and outside blockholdings are substitutes to each other. In fact, according to model 2, the marginal rate of substitution is equal to 82.6/93.9 = 0.88. Hence, a change in insider ownership by 100 basis points must be accompanied by an offsetting change of 88 basis points in external blockholdings in order not to have any impact on firm performance. In a very strict sense, it follows from this that external blockholdings are more effective in terms of value creation. However, given the variance in the data one should not stress this result.790 As a corollary, it is interesting to note that the coefficient on the number of blockholders variable is significantly (0.01 level) negative. This is in line with the view that the benefits of outside control decrease the more dispersed blockholdings are.791 Out of the seven industry dummy variables, four (including the intercept) enter the equation on a significant level (0.1 significance level). 6.9.3.2 Base Case Extended (Cross Section 2003) After having assumed a pure linear specification of the impact of insider ownership on performance in the previous section, now the possibility of alternative specifications is investigated . The curvilinear relationship between insider ownership and performance found by M C C ONNELL AND S ERVAES (1990) is emulated by including both the normal and the squared term of insider ownership, labeled as MGMT SQ, in model 7 (see table 6-33).792 As a result, the coefficient for MGMT becomes negative but not on a significant level. The coefficient of MGMT SQ is positive and significant at the 0.05 level. Thus, the bell-shaped relationship found by M C C ONNELL AND S ERVAES (1990), where insider ownership above a certain threshold becomes value destroying, cannot be observed in this sample.793 Contrary, according to the coefficients estimated in model 7, insider ownership ceteris paribus starts to positively affect perfor790

791

792 793

Cf. Gorton and Schmid (2000), pp. 63-64. For example, Thomsen, Pedersen and Kvist (2006) find a significant negative relationship between blockholder ownership and firm value/accounting returns for Continental European firms. Contrary, Guti´errez and Trib´o (2004), pp. 1-38, examining a sample of Spanish firms from 1996 to 2000 find a positive relationship between both blockholders’ ownership stake as well as their number and performance. See McConnell and Servaes (1990). Cf. Halpern, Kieschnick and Rotenberg (2005), pp. 782-784. The reliability of results including higher terms of insider ownership as independent variables appears not unproblematic because of the arising multi-collinearity. In the 2003 sub-sample, the variance inflation factors (VIFs) for MGMT and MGMT SQ reach 16.5 and 12.6 respectively indicating presence of multicollinearity. No procedure to deal with this problem can be found in McConnell and Servaes (1990). A detailed discussion of the problem of multi-collinearity will follow later in section 6.9.4.3.

Empirical Analysis of Insider Ownership

235

mance above 20.4%. This is quite interesting, as the result is not in accordance with the view that large insider stakes are harmful to outside shareholders because of their expropriation via the consumption of private benefits by insiders.794 However, as the MGMT-variable is not significant, this result cannot be seen as reliable evidence for such a relationship. Next, it is checked whether it is possible to replicate the piecewise-linear relationship between insider ownership and corporate performance first found by M ORCK , S HLEIFER AND V ISHNY (1988) and later on confirmed by C HO (1998). Dividing MGMT in three sub-variables — one for low (MGMT 0to5), medium (MGMT 5to25) and high (MGMT 25to100) insider ownership stakes — using the thresholds of 5% and 25% as proposed by M ORCK , S HLEIFER AND V ISHNY (1988), only the coefficient for insider ownership above 25% turns out to be positive and significant (0.01 level). This can be seen from the results of model 8 in table 6-33. As in the previous model, in this specification insider ownership seems to have a positive effect only at rather high levels of insider ownership, i.e. above 25%. Even by looking at different thresholds it is not possible to get more significant results.795 Hence, the linear relationship between insider ownership and firm performance, as used in model 2, still represents the most convincing specification, particularly given that non-linear specifications with two or more turning points appear somewhat arbitrary and theory on the relevant turning points is poor or less persuasive respectively. Suggestions to alter the insider ownership variable to reflect the concentration of insider ownership were implemented in models 9a and 9b.796 In model 9a, the coefficient for the average ownership share per board member, calculated as MGMT divided by the sum of management and supervisory board members (i.e. MGMT AV = MGMT/(MB NO+SB NO)), is positive but less significant than in the base case. Nevertheless, this result has to be interpreted with caution because of a methodological issue involved: Since it is not possible to obtain the 794

795

796

It should be noted that also higher terms of MGMT were included as done by Davies, Hillier and McColgan (2005), pp. 647-649, 653, without obtaining more promising results than those found in the base case model 2. For example, Chen, Hexter and Hu (1993) use 5-7% and 10-12% as alternative thresholds and obtain results similar to Morck, Shleifer and Vishny (1988). Cui and Mak (2002) find a W-relationship with 10%, 30% and 50% as turning points. Cf. Hermalin and Weisbach (1991); De Miguel, Pindado and De la Torre (2004); Davies, Hillier and McColgan (2005). See Cui and Mak (2002), p. 320.

-99.631

-77.025 188.738

Intercept

MGMT MGMT MGMT MGMT MGMT MGMT MGMT MGMT MGMT

Yes

247 0.388

Industry Dummies

n R2 adj.

**

247 0.373

Yes

[. . .]

-92.188 -52.321 138.756

-109.954

(8)

***

247 0.375

Yes

[. . .]

548.477

-149.507

(9a)

**

*

BAHR

247 0.372

Yes

[. . .]

-17.013

101.220

-120.091

(9b)

**

***

247 0.331

Yes

[. . .]

-0.004

-73.145

(10)

247 0.334

Yes

[. . .]

58.170

-69.909

(11)

Table 6-33: Results III: Multiple Regression - Base Case Extended

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). Control variables are not reported for the sake of clarity. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the respective t-statistics refer to tables D-7 and D-8 in Appendix D. Source: Author’s findings.

[. . .]

[. . .]

SQ 0to5 5to25 25to100 AV NO EUR CRmCF

(7)

Model No.

Dependent Variable

236 Empirical Analysis of Insider Ownership

Empirical Analysis of Insider Ownership

237

number of all former board members (nor this would be especially useful), the divisor of the average insider ownership variable contains the share of all active and former board members while the denominator is only based on the number of all active board members. In model 9b a different approach is taken to account for the concentration of insider ownership: Besides the cumulative shareholdings of insiders (MGMT), the number of those registered insider shareholders (MGMT NO) is included as an additional explanatory variable. The result is similar to that previously found for the case of blockholders: While MGMT is positive, MGMT NO is negative (both significant at the 0.01 and 0.05 level respectively) indicating that the positive effect declines, if the insider ownership share is spread across an increasing number of insiders.797 Even though the results of model 9b appear as plausible as the base case specification in model 2, the base case specifications of section 6.9.3.1 will be preferred in the following analyses, because of the advantages associated with dealing with only one — and not two — possible endogenous insider variables.798 Next, in model 10, insider ownership is measured in terms of the euro-value instead of the percentage share.799 It can be argued that shareholders’ wealth loss in an over-paying acquiring firm, which represents a typical case of empirebuilding, is proportional to the value of their holdings and not their percentage share.800 The accordingly defined variable MGMT EUR is calculated as MGMT times the market value of equity measured as the average of monthly observations during the respective year. However, MGMT EUR turns out to be both statistically and economically insignificant as the coefficient is very close to zero (-0.004). Consequently, insiders’ cumulative percentage ownership stakes seem to be the most appropriate measure of insider ownership. Finally, the effects of the difference between insiders’ cash flow and control rights are examined in model 11. Therefore, the variable MGMT CRmCF, which measures the absolute difference between insiders’ control and cash flow rights, is included instead of the variable MGMT. The coefficient turns out to be positive contrary to the expected negative sign. However, since the coefficient 797

798 799 800

Contrarily, Agrawal and Knoeber (1996), p. 386, argue that an increasing number of insiders should reduce the costs resulting from holding under-diversified portfolios. For the problem of possible endogeneity cf. section 6.9.4.2. See Sanders (2001), p. 482. See Gugler, Mueller and Yurtoglu (2004), pp. 3-4.

238

Empirical Analysis of Insider Ownership

lacks statistical significance this result should not be interpreted.801 To summarize, it does not seem that any of the variations of the insider ownership variable discussed in this section generates more convincing results than the simple insider ownership measures MB, SB and FBM as well as MGMT used in the base case specifications of section 6.9.3.1. 6.9.3.3 Base Case (Cross Section 1998) After having tested different specifications of the insider ownershipperformance relationship in the previous section, this and the following sections will examine whether the highly significant impact of insider ownership on corporate performance found in the 2003 base case models can also be detected in the other sub-samples. In the 1998 sub-sample, the number of available observations ranges from 212 (MTBV) to 220 (BAHR) as can be seen in table 6-34. The signs of all but one insider ownership coefficients — the exception is FBM in model 14 — in models 12 to 17 are positive, confirming that the positive impact of insider ownership on firm performance is also present in the 1998 sub-sample. In the stock return models 12 and 13, three of the four insider ownership coefficients are significantly different from zero at least on the 0.05 level. For the aggregate insider ownership measure MGMT the coefficient is 69.3 and, therefore, slightly smaller than in the 2003 sub-sample (82.6). In the MTBV models 14 and 15, only two of the four coefficients (i.e. MB and SB) are significant at least on the 0.1 level. Unlike in the 2003 sub-sample, in the ROA models 16 and 17 all insider ownership coefficients are positive and significant at least on the 0.1 level. The explanatory power of the models is somewhat smaller than in the 2003 sub-sample. For the BAHR models, the adjusted R2 is equal to 33.7% (average models 12 and 13), for the MTBV models it reaches 13.0% (average of model 14 and 15) and for the ROA models (models 16 and 17) it is 8.6%. With respect to the control variables, the sign of BLOCK NO changes from the 2003 to the 1998 sub-sample but the positive coefficient in the latter ceases to be significant. Similarly, the influence of FIRM RISK becomes negative but is not significant at traditional levels. To sum up, the base case specifications devel801

The results are qualitatively similar for the other sub-samples and the use of different performance measures (results not presented).

220 0.334

n R2 adj.

** * ***

***

*

220 0.339

Yes

** * ***

***

*

***

**

0.841

0.615

(15)

212 0.142

Yes

212 0.117

Yes

0.558 0.481 0.064 0.058 0.212 *** 0.184 ** 0.712 0.774 -1.967 *** -1.998 *** -0.178 -0.149 -0.277 -0.219

1.849 * 1.282 * -0.178

0.126

(14)

MTBV

215 0.086

Yes

0.691 0.059 0.483 2.518 -3.305 2.386 1.115

5.556 * 11.217 * 7.773 **

-7.750

(16)

215 0.085

Yes

0.766 0.114 0.541 2.560 -3.897 2.287 1.114

8.630

-8.281

(17)

ROA

***

Table 6-34: Results III: Multiple Regression - Base Case (1998)

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the respective t-statistics refer to tables D-9 and D-10 in Appendix D. Source: Author’s findings.

Yes

33.046 4.072 11.524 -3.605 -57.973 33.495 41.592

32.341 3.994 11.192 -3.239 -57.188 33.828 41.967

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies

69.304

*** **

56.734 70.968 75.251

-122.079

MB SB FBM MGMT

**

-117.284

Intercept

(13)

(12)

BAHR

Model No.

Dependent Variable

Empirical Analysis of Insider Ownership 239

240

Empirical Analysis of Insider Ownership

oped in the 2003 sub-sample can be successfully cross-checked with the data of the 1998 sub-sample, indicating that the results are relatively stable over the periods ranging from 1999 to 2003 (2003 sub-sample) and 1994 to 1998 (1998 sub-sample). 6.9.3.4 Base Case (Cross Section 1993) Having found evidence for a positive impact of insider ownership on corporate performance in the 2003 and 1998 sub-samples, last the 1993 sub-sample is examined. The results of the six base case models (models 18 to 23), as illustrated in table 6-35, show that only one of the 12 insider ownership coefficients, namely MB in the ROA model 22 turns out to be significant. However, eight of these coefficients, including the significant one, show positive signs indicating a positive relationship between insider ownership and performance if anything. As already mentioned, this conclusion should be regarded with caution as the results also could simply be a random outcome. Three explanations why results of the 1993 sub-sample differ from those of the 2003 and 1998 sub-samples appear plausible: First, it might of course be possible that there is not any positive/negative relationship between the level of insider ownership and corporate performance during the 1993 sample period and, hence, such a relationship cannot be detected in the data. Second, another compelling reason might be that the smaller sample sizes — 131 to 132 observations in the 1993 sub-sample vs. 212 to 247 in the 2003 and 1998 sub-samples — do not suffice to effectively isolate a relationship between insider ownership and corporate performance. Third, and related to the second argument, shareholder structure data for 1993 is less detailed and less accurate and especially smaller ownership stakes of insiders were hardly disclosed.802 Consequently, measurement errors in the explanatory insider ownership variables might prevent significant results. The last argument is also supported by the fact that also the significant positive impact of blockholdings ceases to exist in the 1993 sub-sample.803 To conclude, the results of the 1993 sub-sample seem to nei802

803

The average number of shareholder entries in the Hoppenstedt Aktienf¨uhrer is 1.6 for 1993 vs. 2.2 for 1998 and 2.8 for 2003. See table 6-3. Only the coefficients of BLOCK O in the MTBV equations (models 20 and 21) are positive and significant on the 0.1 level.

133 0.288

n R2 adj.

133 0.294

Yes

**

0.304

2.615

(21) **

132 0.180

Yes

132 0.187

Yes

0.802 * 0.752 * -0.183 -0.199 * 0.043 0.030 0.196 0.159 -2.172 *** -2.190 *** 0.535 0.598 * -0.771 ** -0.689 *

0.734 0.661 -0.081

2.363

(20)

MTBV

131 0.192

Yes

-2.457 0.063 0.421 1.547 -5.887 ** 0.539 0.653

7.274 ** -2.624 1.604

-4.291

(22)

(23)

131 0.149

Yes

-3.05 -0.024 0.241 0.905 -5.177 * 0.318 1.083

0.455

-2.029

ROA

Table 6-35: Results III: Multiple Regression - Base Case (1993)

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the respective t-statistics refer to tables D-11 and D-11 in Appendix D. Source: Author’s findings.

Yes

-14.306 -2.965 17.883 *** 115.632 ** -136.452 *** 58.818 ** 88.830 ***

-11.022 -2.166 18.711 *** 118.104 ** -136.653 *** 55.234 ** 83.540 ***

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies

-1.636

27.584 16.791 -23.864

***

MB SB FBM MGMT

*** -282.907

-297.691

Intercept

(19)

(18)

BAHR

Model No.

Dependent Variable

Empirical Analysis of Insider Ownership 241

242

Empirical Analysis of Insider Ownership

ther corroborate nor contradict the previous findings from the 2003 and 1998 sub-samples. 6.9.3.5 Base Case (Pooled Sample) After having examined the six base case equations for the three sub-samples separately, table 6-36 shows the regression results if the three sub-samples are pooled to one aggregated sample. To account for potential time fixed-effects, in addition to the known set of control variables two time dummy variables are included.804 Through the pooling of the sub-samples, the number of available firm observations for the different models rises to 579 for the MTBV models and 600 for the BAHR models. In the pooled sample, all insider ownership coefficients are positive and nine out of 12 are significant at least on the 0.05 level. Notably, the coefficients of all aggregate measures of insider ownership (MGMT) in models 25, 27 and 29 are highly significant. The lack of significance for some individual ownership measures (FBM in the MTBV model 26, MB and FMB in the ROA model 28) might partly be driven by the relative small number of occurrences for each of the three insider ownership sub-categories.805 With regard to the control variables, the positive impact of the level of outside blockholdings is sustained for the equations with BAHR and MTBV as dependent variables. In the same models, the negative impact of the number of blockholders still can be found but lacks significance. Eight of the 12 time dummy variables are significant at least on the 0.1 level806 , suggesting that corporate performance as measured by the three performance measures differed clearly across the three different sub-sample periods.807 To summarize, the results from the pooled sample corroborate the presented evidence since the results are rather robust, i.e. irrespective of the performance 804

805

806

807

Anderson and Reeb (2003), pp. 1315, 1325, investigating founding-family ownership and firm performance also use a two-way fixed-effects (time, industry) model for their regressions analysis. They show that more advanced econometric techniques such as random-effects panel data regressions, pooled, time-series average regressions and Fama and MacBeth (1973) regressions yield results quantitatively and qualitatively similar to their two-way fixed-effects regression results. Cf. Andres (2006), p. 12. For the 2003 sub-sample (n = 290), the aggregated insider ownership variable MGMT has 92 non-zero instances while MB, SB and FBM have 29, 47 and 35 non-zero instances respectively. The coefficients of the time dummy variables YEAR 1998 and YEAR 1993 are only shown in tables D-13 and D-14 in appendix D. For the mean values of the performance variables in the three sub-samples see table 6-15 and section 6.4.2.

600 0.355

n R2 adj.

*** *** ***

***

***

600 0.356

Yes

*** *** ***

***

***

***

*** ** *** 0.957

-0.044

(27)

***

579 0.160

Yes

579 0.159

Yes

0.857 *** 0.826 *** -0.057 -0.062 0.169 *** 0.157 *** 0.244 * 0.257 ** -1.985 *** -1.968 *** 0.196 0.221 -0.127 -0.090

1.364 1.079 0.670

-0.214

(26)

MTBV

** 4.201

-4.984

(29)

**

592 0.119

Yes

592 0.116

Yes

-0.773 -0.832 0.194 0.208 0.601 *** 0.593 *** -0.118 -0.086 -8.127 *** -8.237 *** 2.992 *** 3.005 *** 1.556 * 1.622 *

3.574 6.731 2.190

-5.369

(28)

ROA

Table 6-36: Results III: Multiple Regression - Base Case (Pooled)

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the respective t-statistics refer to tables D-13 and D-14 in Appendix D. Source: Author’s findings.

Yes

46.095 -3.980 11.701 4.239 -95.907 36.685 56.078

45.943 -4.105 11.554 4.330 -92.056 36.915 56.340

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry/ Time Dummies

57.806

** *** **

49.357 66.855 53.560

MB SB FBM MGMT

-157.855

(25)

-156.967 ***

(24)

BAHR

Intercept

Model No.

Dependent Variable

Empirical Analysis of Insider Ownership 243

244

Empirical Analysis of Insider Ownership

measure used and the time period under investigation. Hence, as far as the German capital market is concerned, a positive relationship between insider ownership and corporate performance is likely to exist. 6.9.3.6 Models of Dynamic Ownership In the context of the choice and measurement of the performance variables (section 6.3.2), it has been mentioned that the use of stock returns as performance measure represents a rather conservative approach. If the market was aware of a positive relationship between insider ownership and performance, prices would react immediately to reflect such information. In other words, stock prices would only react if changes in the level of insider ownership occur. The most widespread approach to investigate such stock price reactions is the event study methodology developed by FAMA , F ISHER , J ENSEN AND ROLL (1969). Here, a different approach is taken because of the different nature of the data set: Possible reactions of stock prices on new information, i.e. changes in the level of insider ownership, will be examined by including an insider ownership variable which measures the change in the level of ownership (MGMT CHANGE) in the previously reported base case regression models.808 MGMT CHANGE reflects the absolute change of the variable MGMT from 1998 to 2003 and 1993 to 1998.809 For the 1993 sub-sample, no observations are available since MGMT has been measured at the end of 1993 for the first time and, consequently, only two period of changes in insider ownership levels are on-hand. Summary results of the regression models with MGMT CHANGE as explanatory variable are presented in table 6-37. With BAHR as dependent variable, the dynamic insider ownership variable has positive and significant (at least 0.05 level) coefficients in the 2003 sub-sample and the pooled sample. A 10% increase in the level of insider ownership between 1998 and 2003 would be accompanied by on average 8.2% higher buy-and-hold stock returns. In the 1998 sub-sample, the coefficient of MGMT CHANGE is negative but lacks signifi808

809

Cf. K¨oke (2004), pp. 65-75, or Cole and Mehran (1998), pp. 301-307, who analyze changes of corporate control for the case of Germany or changes of ownership in the U.S. thrift industry. For the 2003 sub-sample, the mean change in MGMT (from 1998 to 2003) is -2.6%. For the 1998 subsample, the mean change in MGMT (from 1993 to 1998) is -2.8%. These values do not equal the differences in figure 6-10 because, e.g. the 1998 MGMT-share of 29.6% does not only contain the firms from the 2003 sub-sample, but also firms which are CDAX member at the end of 1993 but not at the end of 2003. Cf. section 6.2.1.

245

Empirical Analysis of Insider Ownership Sample Period

1993

Independent Variable MGMT CHANGEb Control Variables n R2 adj.

1998

Pooleda

2003

Dependent Variable: BAHR

n/a

-15.529 [. . . ] 204 0.347

81.532 [. . . ] 247 0.382

***

45.599 [. . . ] 451 0.352

**

Dependent Variable: MTBV MGMT CHANGEb Control Variables n R2 adj.

n/a

-0.148 [. . . ] 195 0.168

0.459 [. . . ] 235 0.161

0.349 [. . . ] 430 0.157

Dependent Variable: ROA MGMT CHANGEb Control Variables n R2 adj. a b

n/a

4.408 [. . . ] 198 0.181

-0.296 [. . . ] 246 0.164

0.743 [. . . ] 444 0.105

The pooled models includes one dummy variable to account for differences between the two sub-samples. OLS-regression coefficient. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the coefficients of the control variables refer to tables D-17 to D-19 in Appendix D. Source: Author’s illustration.

Table 6-37: Results III: Multiple Regression - Dynamic Ownership cance. The explanatory power of the models (R2 from 37.5% (1998) to 42.0% (2003)) is at levels comparable to those in the equations including insider ownership as measured by the absolute fraction of shares owned by insiders. If MTBV or ROA are used as dependent variables, the dynamic insider ownership coefficients are positive in four out of six cases but all coefficients lack statistical significance. From a theoretical point of view, there are some good arguments to favor changes in the level of insider ownership as the more appropriate subject of investigation. Nevertheless, the following three reasons might explain, why the results of the dynamic ownership models yield more ambiguous results than the base case specifications: First, both the change in the level of insider ownership and performance — at least in the case of BAHR as dependent variable — are measured over a relatively long time period of five years. Hence, the ef-

246

Empirical Analysis of Insider Ownership

fect, i.e. a stock market reaction, cannot be isolated as accurately as in an event study design, which usually examines daily abnormal returns around a specified event date.810 Second, changes in the level of insider ownership — and especially increases — might not represent a real accumulation of ownership stakes by management but also could be an artificial result of changes in disclosure (or even disclosure requirements).811 Last, the specification of MGMT CHANGE assumes that e.g. a decrease in the level of insider ownership from 40% to 20% has on average the same effect as a decrease from 20% to 0%. This just might be an unsatisfying realization of the insider ownership-performance relationship as the incentives structures in both cases are likely to be very different indeed. 6.9.4

Robustness Tests

As the question whether insider ownership and corporate performance are related is at the very core of this study, the results of the previous sections will be subject to special scrutiny. Besides the effects of using two alternative performance measures (“APM”) (section 6.9.4.1), the econometric issues of endogeneity (section 6.9.4.2) and multi-collinearity (section 6.9.4.3) will be addressed to stress the reliability of the previously presented findings. 6.9.4.1 Alternative Performance Measures It was already shown that the positive impact of insider ownership on firm performance can be detected for three different performance measures, namely five year buy-and-hold returns (BAHR), market-to-book values (MTBV) and return on assets (ROA). Among those measures BAHR delivered the most significant results. As the sensitivity of empirical findings to the used performance measures has already been documented812 , this section examines whether the results can be replicated if two additional performance measures, Tobin’s Q (Q) and return on equity (ROE), are used.813

810 811 812 813

See Fama, Fisher, Jensen and Roll (1969), pp. 3-7. See section 6.2.2. See e.g. Bøhren and Ødegaard (2001), p. 25. Examples for studies using these alternative performance measures: I) Tobin’s Q: Wolf (1999); Cui and Mak (2002); Beiner, Drobetz, Schmid and Zimmermann (2004). II) ROE: Thonet and Poensgen (1979); Wolf (1999); Gorton and Schmid (2000); Gleisberg (2003).

Empirical Analysis of Insider Ownership

247

Firm value and, hence, expected performance can be measured as Tobin’s Q (Q). As suggested by several researches814 , here the so-called simple Tobin’s Q is used.815 Hence, Q is calculated as the market value of the firm divided by its book value. The market value is calculated as the sum of market value of equity plus the book value of debt as market values of debt are rarely available for German companies. The book value of the firm is approximated by the book value of total assets.816 Qt =

M V (Equity)t + BV (Debt)t BV (T otal Assets)t

Where = Tobin’s Q at t = t ∈ [31.12.1993, 31.12.1998, 31.12.2003] = Market value of equity at t, measured as the sum of the market values of all share classes at t = Book value of total debt at t BV (Debt)t BV (T otal Assets)t = Book value of total assets at t

Qt t M V (Equity)t

It should be noted that according to the definition of Tobin’s Q the correct denominator should be the replacement cost of total assets. Following other studies we proxy the unobservable replacement costs of total assets by their book value. However, there are strong objections against this procedure in a Continental European accounting context, where historical cost accounting is still prevalent.817 Because of the drawbacks involved by using Tobin’s Q to measure performance for German corporations, it is likely that models using Q as dependent variable have lower explanatory power. Nevertheless, it appears worthwhile to present also selected results with Tobin’s Q as performance indicator 814

815

816

817

See amongst others Chung and Pruitt (1994); Agrawal and Knoeber (1996); Beiner, Drobetz, Schmid and Zimmermann (2006). Chung and Pruitt (1994) show that in a series of regressions at least 96.6% of the variability of a more theoretically correct model of Tobin’s Q is explained by their simple approximation of Tobin’s Q. DeDalt, Donaldson and Garner (2003) also argue in favor of the simple Q estimator. Contrarily, Perfect and Wiles (1994) stress the effects of different calculations of Tobin’s Q on empirical results. Data on the market value of equity, the book value of total debt and the book value of shareholders’ equity and reserves are taken from the Datastream/Worldscope databases. Cf. Edwards and Weichenrieder (2004), p. 152; B¨orsch-Supan and K¨oke (2002), p. 318.

248

Empirical Analysis of Insider Ownership

to facilitate the comparison with international — and especially U.S. — studies which predominantly rely on Tobin’s Q as an indicator of firm value.818 Finally, the return on equity (ROE) measures the return which shareholders yield on their equity. Since shareholders’ equity and reserves, i.e. the denominator of the fraction, shows larger variations than total assets, it is expected to be a more volatile performance measure. As shareholders’ equity and reserves even can become negative, those cases do not allow for a reasonable analysis and are excluded from further analysis.819 Similarly, if the denominator approaches zero, extreme high values of ROE can be observed even though profits (i.e. the nominator) are relatively low. Therefore, those cases will be treated as outliers.820 It is calculated as follows:   P ATt ROEt = − 1 × 100  BV (Equity)t Where ROEt t P ATt  (Equity)t BV

= = = =

Return on equity (ROE) at t in % t ∈ [1993, 1998, 2003] Published after tax profit during year t Average of equity capital and reserves at begining and end of year t

Table 6-38 shows descriptive statistics for both alternative performance measures. Because of the methodical problems described above, the number of available observations for ROE is significantly smaller than for the three main performance variables.821 For both Q and ROE the means are highest in the 1998 sub-sample as it has been for MTBV and ROA. The mean values of Q varying from 0.953 (2003) to 1.190 (1998) appear to be at the lower end while 818

819 820

821

One value below 0.1 (1993) and 19 above 5.0 were excluded as outliers: 0 (1993), 18 (1998) and 1 (2003). Since 14 of the 19 upper outliers showed a MGMT-share of over 40% (average: 46.7%) a bias in account of a positive relationship between insider ownership and performance might be introduced if any. Cf. footnote 638. See Phani, Reddy, Ramachandran and Bhattacharyya (2005), p. 10. Excluded ROEs because of negative denominator: 52 (1993), 51 (1998) and 90 (2003). Excluded ROEs above 100: 0 (1993), 3 (1998) and 2 (2003). Cf. footnote 638. See table 6-15.

249

Empirical Analysis of Insider Ownership

the mean values for ROE ranging from 11.9% (1993) to 17.4% (1998) appear to be at the higher end of expectations.822 Sample Period

1993

1998

2003

236

358

290

206 1.026 0.864 0.634 0.241 4.905

325 1.190 0.886 0.918 0.100 4.987

254 0.953 0.740 0.658 0.192 4.144

235 17.373 14.558 13.676 0.844 83.590

172 14.775 12.088 12.998 0.685 91.288

Total number of sample companies Panel A: Tobin’s Q (Q) Number of valid cases (n) Mean Median Standard Deviation Minimum Maximum

Panel B: Return on Equity (ROE) (in %) Number of valid cases (n) Mean Median Standard Deviation Minimum Maximum

155 11.912 9.099 10.856 0.000 94.835

For the definition of the performance measures see section 6.9.4.1. Source: Author’s illustration. Stock return and accounting data are from the Datastream/Worldscope database.

Table 6-38: Results III: Descriptive Statistics of APM A summary of the insider ownership regression coefficients of the resulting 16 equations (two insider ownership specifications, four sub-samples and two performance measures) is shown in table 6-39. All coefficients show positive signs even though only MB is significant in the 1993 and the pooled sample if Tobin’s Q is taken as performance measure. It is exclusively significant in the 1993 sub-sample if ROE is used as dependent variable. The R2 for the Tobin’s Q models range from 21.4% to 35.9% and, hence, are comparable to those of the BAHR models in the previous sections. For ROE as dependent variable, the explanatory power in most cases is considerably lower with values between 13.5% and 16.9% (the exception is the 1993 sub-sample with R2 of 44.8% and 44.1% respectively). 822

This might be partially caused by the exclusion of certain extreme cases as described above. Seifert, Gonenc and Wright (2005), p. 182, get a similar low mean Tobin’s Q (1.286) for three year averages from 1997-1999.

250 Sample Period

Empirical Analysis of Insider Ownership 1993

2003

Pooleda

0.132 [. . . ] 223 0.158

0.142 [. . . ] 248 0.316

0.145 [. . . ] 604 0.215

0.000 0.296 0.034 [. . . ] 223 0.157

0.248 0.124 0.097 [. . . ] 248 0.312

0.241 0.169 0.072 [. . . ] 604 0.214

1998

Panel A: Alternative Performance Measure - Tobin’s Q MGMTb Control Variables n R2 adj.

0.162 [. . . ] 133 0.209

MBb SBb FBMb Control Variables n R2 adj.

0.824 0.119 0.055 [. . . ] 132 0.245

**

*

Panel B: Alternative Performance Measure - Return on Equity (ROE) MGMTb Control Variables n R2 adj.

7.099 [. . . ] 99 0.340

MBb SBb FBMb Control Variables n R2 adj.

11.063 3.694 9.226 [. . . ] 99 0.332

a b

**

6.888 [. . . ] 172 0.061

4.117 [. . . ] 170 0.082

5.344 [. . . ] 441 0.101

9.430 2.606 10.299 [. . . ] 172 0.060

3.752 6.984 0.630 [. . . ] 170 0.076

5.976 3.813 6.694 [. . . ] 441 0.098

The pooled models includes one dummy variable to account for differences between the two sub-samples. OLS-regression coefficient. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results including the coefficients of the control variables refer to tables E-1 to E-4 in Appendix E. Source: Author’s illustration.

Table 6-39: Results III: Multiple Regression - APM

To conclude, the results by using the two alternative performance measures as dependent variable neither strengthen nor contradict the findings from the previous sections. Firstly, the methodical issues involved in the calculation of ROEs reduce sample sizes which might have led to less significant results. Secondly, the same issues might also limit the usefulness of ROE as a performance measure. For example, companies with low earnings can show huge ROEs if shareholders equity is close to zero (which mainly represents an accounting ef-

Empirical Analysis of Insider Ownership

251

fect).823 Similarly, the problems involved in the use of Tobin’s Q — mainly that no replacement costs of assets are available because of historic cost accounting — may have limited the expressiveness of the results.824 6.9.4.2 Endogeneity One important source of possible errors in the presented regression results is the previously mentioned issue of endogeneity.825 In simple words, it is not clear whether firms with high levels of insider ownership perform well because of reductions in agency costs or insiders increase their ownership stakes if they expect their firms to perform well. Similarly, unobserved firm characteristics could simultaneously impact both the level of insider ownership and performance.826 Consequently, an observed relationship between insider ownership and corporate performance might indeed be a case of spurious correlation with an omitted variable and OLS-regression estimates would be inconsistent.827 Following the argumentation of E DWARDS AND W EICHENRIEDER (2004) and E DWARDS AND N IBLER (2000), a restricted panel sample was examined.828 As described in section 6.2.1, complete shareholder structure data for all three sub-samples (1993, 1998 and 2003) are available for 168 firms.829 Because of other missing data (i.e. performance or control variables), complete data sets as needed for the base case regression specification (model 2) were available for only 86 companies for all three sub-samples. Therefore, the final sample size amounted to 86 companies or 258 (= 86x3) firm year observations. Changes in the level of insider ownership could occur during the first period (1993-1998) and/or during the second period (1998-2003). A major change in insider ownership is defined as a change of at least 10% of MGMT from its previous level, to disregard minor shifts which also might stem from measure823

824

825 826

827

828 829

In those cases, the resulting measurement errors in the endogenous variable (ROE) do not cause estimated coefficients to be biased as this additional variation is included in the error term. However, estimation results are weakened because of a low signal-to-noise ratio. See B¨orsch-Supan and K¨oke (2002), p. 317. For objections against the use of Tobin’s Q for the case of Germany because of historic cost accounting see Edwards and Weichenrieder (2004), p. 152. See section 4.2.1. B¨orsch-Supan and K¨oke (2002), p. 297, refer to both as cases of reverse causality and label the former structural reverse causality and the latter spurious correlation. See Leland and Pyle (1977), pp. 371-387; Cho (1998), pp. 105-106; Himmelberg, Hubbard and Palia (1999), pp. 357-360; Demsetz and Villalonga (2001), pp. 215-217; Edwards and Weichenrieder (2004), pp. 161-162. See Edwards and Weichenrieder (2004), p. 162; Edwards and Nibler (2000), p. 252. See figure 6-1.

252

Empirical Analysis of Insider Ownership

INCREASE (n = 5) INCREASE (n = 14)

NO CHANGE (n = 2) DECREASE (n = 7) INCREASE (n = 4)

MGMT’93 (n = 86)

NO CHANGE (n = 54)

NO CHANGE (n = 48) DECREASE (n = 2) INCREASE (n = 8)

DECREASE (n = 18)

NO CHANGE (n = 6) DECREASE (n = 4)

1993

1998

2003

Source: Author’s illustration.

Figure 6-17: Changes of Insider Ownership 1993-2003 ment errors. Hence, a company was classified as a no change company, if the level of MGMT remains in a 90-110% interval based on its previous value. As illustrated in figure 6-17, out of the 86 companies 48 (or 55.8%) did not experience major changes in the degree of insider ownership in both periods. This underlines once more the stickiness of insider ownership previously claimed by E DWARDS AND N IBLER (2000).830 The base case specification (model 2 in 2003 and model 25 in the pooled sample) is then applied to the restricted panel and a sub-sample. The sub-sample includes all companies that did not experience major changes in insider ownership during both periods (n = 48). Summary results for the two resulting regression models are shown in table 6-40. As in the base case models, the coefficients of MGMT are positive and significant at least on the 0.1 level in both cases. The coefficients of MGMT — 57.9 in the complete sample and 66.2 830

See Edwards and Nibler (2000), p. 252. Similarly, Kaserer and Moldenhauer (2005) show that a five year lagged insider ownership variable adds perceivable explanatory power to a regression model with current level of insider ownership as dependent variable and is highly significant. See Kaserer and Moldenhauer (2005), p. 19. Cf. Thomsen, Pedersen and Kvist (2006), p. 251.

253

Empirical Analysis of Insider Ownership

Dependent Variable

BAHR

BAHR

Model No.

(55) Complete

Intercept

-157.599

***

-167.437

**

MGMT

57.938

**

66.156

*

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998 YEAR 1993 Industry Dummies n R2 adj.

30.113 -6.448 13.217 18.339 -147.260 26.423 50.480 -4.001 43.021

(56) Sub-sample

* *** * *** ** *** ***

18.770 -4.661 14.328 13.154 -162.421 19.518 59.651 -9.145 36.833

Yes

Yes

258 0.329

144 0.292

*** *** *** *** **

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For detailed results refer to table E-5 in Appendix E. Source: Author’s findings.

Table 6-40: Results III: Multiple Regression - Endogeneity (Pooled) in the sub-sample — are at a comparable level as the coefficient (57.8) in the pooled base case regression (model 25). Since the positive relationship between insider ownership and performance persists even if no changes in insider ownership occur (sub-sample), it appears reasonable to treat insider ownership as exogenous or, at least, predetermined with respect to performance.831 The fact that the significance level of the MGMTcoefficient (i.e. 6.1% in the sub-sample) is smaller compared to the base case might be driven by smaller sample sizes and hence must not necessarily be evidence for a weaker relationship between insider ownership and performance in this sub-group. 831

For a similar argumentation cf. Gorton and Schmid (2000), p. 51; Gorton and Schmid (2002), p. 12.

254

Empirical Analysis of Insider Ownership

Especially in U.S. studies, the problem of endogeneity is often addressed by building simultaneous equation systems.832 Here, this approach will not be pursued as it is well known that two-stage least square (2SLS) estimators are quite sensitive to the specification of the equation system. The theory for choosing instrumental variables is poor and variations in the choice of instruments significantly effect the results.833 This is a severe problem of all empirical governance studies dealing with simultaneous equation systems. As pointed out by H IM MELBERG , H UBBARD AND PALIA (1999): “[. . . ] instrumental variables for managerial ownership are difficult to find. The basic problem is that for any variable that plausibly determines the optimal level of managerial ownership, it is also possible to argue that the same variable might plausibly affect Tobin’s Q [as a measure for corporate value].”834 Empirical evidence on the dimension of problems related to the use of (inadequate) instrumental variables in corporate governance studies is given by BARNHART, M ARR AND ROSENSTEIN (1994). Investigating the relationship between board composition and firm performance in an instrumental variable approach, they find that small changes in the instruments used in first-stage regressions induce major changes in the estimates of the second-stage model. Thus, those results should be interpreted with extreme caution.835 Similarly, B ØHREN AND Ø DEGAARD (2004) — examining the relationship between corporate governance and performance — show that the sign and significance of an insider ownership variable varies largely according to nine different sets of plausible instruments used in a simultaneous equation system.836 Hence, it is argued that endogeneity is not primarily a question of how the results of an ordinary OLS-equation compare to the results of an appropriate 2SLS-estimation.837 It is more a question of economic and empirical reason832

833 834 835 836 837

See e.g. for the U.S.: Agrawal and Knoeber (1996); Cho (1998). For Switzerland: Cho (1998); Maury (2006). For Germany: Kaserer and Moldenhauer (2005) build a simultaneous equation system for a similar sample as used in this study. For the concept of simultaneous equation models see Wooldridge (2002), pp. 209-245. Only few studies explicitly test the validity of their instruments. See e.g. Palia (2001), p. 751. Himmelberg, Hubbard and Palia (1999), p. 379. Cf. Harvey, Lins and Roper (2004), p. 10. See Barnhart, Marr and Rosenstein (1994), pp. 335-338. See Bøhren and Ødegaard (2004), pp. 16-18. Cf. Coles, Lemmon and Meschke (2003), p. 32. Cf. Agrawal and Knoeber (1996), p. 391.

Empirical Analysis of Insider Ownership

255

ing. Given that insider ownership is a rather inert variable, at least for the case of Germany838 , endogeneity in this study is perceived less immanent than in the U.S. data.839 There, insider ownership is much more related to firm performance, as insider ownership is largely related to compensation contracts.840 This is still very different from the German situation. This conclusion is also supported by the empirical results of G UGLER AND W EIGAND (2002) who find that insider ownership is at least partially endogenously determined in the U.S. while large shareholders — ubiquitous in Germany — affect firm performance exogenously.841 6.9.4.3 Multi-Collinearity The linear regression model assumes that regressors are not linear dependent. This issue will be addressed in this section as the coefficients estimates from OLS-regressions become less reliable the higher the degree of multi-collinearity is. The consequences of multi-collinearity are twofold: On the one hand, the presence of multi-collinearity signals redundancy in the data. On the other hand, and more severely, the information derived by coefficients estimates cannot definitely be attributed to an individual variable. As a consequence, the explanatory power can be significant even though none of the coefficients is significant. Alternatively, the coefficients estimates can vary largely depending on whether individual variables are included in the model or not.842 One indicator for the presence of multi-collinearity are bivariate correlation coefficients of the independent variables near the absolute value of one. As can be seen in tables D-1 and D-2 of Appendix D, this is not the case. There is only one relatively high correlation between MGMT and BLOCK O (-0.713), which already has been discussed.843 However, since correlation coefficients only re-

838

839 840 841 842 843

Mikkelson and Partch (1989), p. 287, state that “[t]here is considerable variation in managers’ voting stakes across firms, but little variation over time for most firms”. This statement strikingly depicts the highly positive cross-sectional skewness and the inertia of the insider ownership data found in this study. Cf. Barontini and Caprio (2005), p. 29; Denis and McConnell (2003), p. 18. See Palia (2001), pp. 735-764; Monsen and Downs (1965), p. 226. See Gugler and Weigand (2002), pp. 483–486. See Backhaus, Erichson, Plinke and Weiber (2003), pp. 88-89. This is still below the critical value of 0.8 for correlation coefficients between independent variables which is stated as a rule of thumb by Eckey, Kosfeld and Dreger (2001), p. 90.

256

Empirical Analysis of Insider Ownership

flect bivariate correlations, a so-called high-grade multi-collinearity might still exist.844 One method to detect multi-collinearity are so-called variance inflation factors (VIFs). They express the degree to which collinearity among explanatory variables affects the precision of regression estimates. The VIF can be calculated for each independent variable by performing a regression of the examined (independent) variable on all other independent variables. The transformation of the resulting R2 to the reciprocal value of 1 − R2 yields the corresponding VIF. The higher the VIFs, the higher the degree of multi-collinearity. Even though there are different opinions on which level of VIFs is still acceptable, one common interpretation is that values above four or five indicate possible problems with multi-collinearity while values above ten indicate strong multi-collinearity.845 Table 6-41 replicates the results of the three base case regression models with aggregated insider ownership MGMT as independent and BAHR, MTBV and ROA as dependent variables (i.e, models 2, 4 and 6) and reports in addition the corresponding VIFs for all independent variables. The VIFs for all independent variables are below the critical value of 4.0 and excluding the seven industry dummy variables even below the value of 2.6. Not surprisingly, the values are highest for the two interdependent ownership variables MGMT and BLOCK O.846 The VIFs for the other models — with the exception of model 7847 — are at comparable levels and hence indicate that multi-collinearity is not a serious problem in the specified regression equations. 6.9.5

Discussion of Results

The findings regarding the proposed hypotheses on the effects of insider ownership on corporate performance are summarized in table 6-42. While evidence is found for the main hypothesis (H-5.3.1), i.e. insider ownership has a positive impact on performance, the empirical results do not support the hypotheses that insider ownership above a certain threshold (H-5.3.2) and that the differ844 845

846 847

See Backhaus, Erichson, Plinke and Weiber (2003), pp. 89-90. Cf. Besley, Kuh and Welsch (1980), pp. 112-117. For the general problem of multicollinearity cf. Greene (2003), pp. 56-59; Hocking (1983), pp. 221, 224-225. For the reasoning to include both variables in the regression equations see section 6.9.3.1. See footnote 793.

2.538

n/a

VIF

*** 2.438 *** 1.632 *** 1.919 1.522 *** 1.230 * 1.033 *** 1.602

***

*

BAHR

**

2.495

n/a

VIF

235 0.174

Yes

1.128 2.392 -0.123 1.656 0.186 *** 1.885 0.222 ** 1.435 -2.024 *** 1.236 0.302 * 1.028 0.187 1.559

1.080

-0.822

(4)

MTBV

246 0.170

Yes

0.653 0.736 0.734 -0.188 -11.027 3.970 2.400

3.979

-5.112

(6)

*

***

**

ROA

2.422 1.644 1.916 1.529 1.217 1.031 1.608

2.523

n/a

VIF

Table 6-41: Results III: Multiple Regression - Multi-Collinearity (2003)

OLS-regression coefficients. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). All variance inflation factors (VIFs) for the industry dummy variables are below 4.0. The maximum values for industry dummy variables are: 3.717 (model 2), 3.896 (model 4) and 3.744 (model 6). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. For t-statistics refer to tables D-7 and D-8. Source: Author’s findings.

247 0.366

93.867 -10.985 9.738 4.726 -80.448 26.629 60.492

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

n R2 adj.

82.590

MGMT

Yes

-140.738

Intercept

Industry Dummies

(2)

Model No.

Dependent Variable

Empirical Analysis of Insider Ownership 257

258

Empirical Analysis of Insider Ownership

ence between insiders’ control and cash flow rights (H-5.3.3) are detrimental to corporate performance. Whereas the results of the univariate analyses are rather ambiguous, the multiple regression analyses provide comprehensive evidence in favor of the existence of a positive effect of insider ownership on corporate performance. The base case specifications, which assume a simple linear relationship between the insider ownership variables and the performance measures, seem to capture best the hypothesized effects. Non-monotonic relationships as piecewise linear or bellshaped interrelations could not be detected in the context of this study. The positive effect on performance persists independently of the performance measure used. The evidence is strongest if stock price returns are used as performance measure. While there are some methodological objections against the use of stock price returns as performance measure848 , market-to-book ratios and return on assets are subject to severe accounting distortions. These distortion would be unproblematic as long as no systematic biases with respect to the examined parameters exist. However, accounting-based performance measures generally might reveal only unsatisfactory information about corporate performance and, hence, might simply represent an inferior performance measure. With regard to the realization of insider ownership, the aggregated measure (MGMT) turns out to be the most appropriate one. In the equations containing the individual insider ownership measures MB, SB and FBM, the signs of the three insider ownership coefficients are almost always identical to the coefficient of the aggregate measure. However, the coefficients of the individual insider ownership variables are usually less significant, which is probably caused by fewer instances for each individual measure. Since the ownership stakes of the management board (MB), supervisory board (SB) and former board members (FBM) all are positively related to performance, the aggregation into the combined measure MGMT seems justifiable and reasonable. Generally, the evidence is strongest — i.e. the coefficients of the insider ownership variable are most significant and the explanatory power of the regression models is highest — for the 2003 sub-sample and the pooled sample. In 1998, the evidence is only slightly weaker but hardly any significant positive relation848

See section 6.3.2.

259

Empirical Analysis of Insider Ownership Number

Hypothesis

Expected Influence

H-5.3.1

The level of insider ownership has an effect on corporate performance.

+

H-5.3.2

The level of insider ownership above a certain threshold has an effect on corporate performance.

−

H-5.3.3

The difference between insiders’ control and cash flow rights has an effect on corporate performance.

−

Empirical Results

denotes that based on the empirical results the hypothesis should not be rejected; ( ) denotes that based on the empirical results the hypothesis should not be rejected - even though the evidence is weaker than for ; denotes that the hypothesis should be rejected based on the empirical results which lack significance on traditional levels. Source: Author’s findings.

Table 6-42: Results III: Review of Hypotheses

ship between insider ownership and performance can be detected in the 1993 sub-sample. While it is not possible to eliminate the alternative that simply no positive relationship exists during the 1993 sample period, it appears more plausible that data restrictions conceal such a relationship: First, disclosure requirements demanded only the publication of corporate ownership stakes above 25%. Second, the availability (and quality) of capital market and accounting data for the period between 1989 and 1993 is clearly restrained for the German market, which in turn reduces the number of available observations and increases possible measurement errors. Besides insider ownership, also outsider block ownership appears to be an effective corporate governance mechanism. In spite of the diversity of outside blockholders interests, which was discussed in section 3.3.2, the absolute level of aggregated outside blockholdings is (rather) uniformly positively related to performance in the regression models. The fact that the number of outside blockholders is negatively related to performance at the same time, strengthens the presumption that blockholders might act as effective monitors indeed: If the control potential (i.e. the aggregated level of outside blockholdings) is split up among an increasing number of parties, the positive effect on perfor-

260

Empirical Analysis of Insider Ownership

mance diminishes. Furthermore, the analyses indicate that insider ownership and outside block ownership might act as substitutes. As a comprehensive overview of studies examining the relationship between insider ownership and corporate performance was already given in section 4.2, here affirmative and contradicting studies will only be listed as a reference. For Germany, G ORTON AND S CHMID (2002) find a positive impact of insider ownership on performance as a corollary in a study about German co-determination. Further consenting evidence is provided by studies on family ownership which usually are highly related but not necessarily identical to those on insider ownership. In this context, E DWARDS AND W EICHENRIEDER (2004), E HRHARDT AND N OWAK (2003 A ) and E DWARDS AND N IBLER (2000) find a positive relationship between insider ownership and performance for all or certain levels of insider ownership (e.g. up to 75%) as well as in certain settings (e.g. if the largest shareholder is classified as an individual/family). E HRHARDT, N OWAK AND W EBER (2006) find that family firms outperform non-family firms in operating performance in the long run but investors cannot profit from this advantage because stock returns are not significantly higher. In contrast K UKLINSKI , L OWINSKI AND S CHIERECK (2003) report five-year negative abnormal returns for shareholders of German family firms’ IPOs.849 For Switzerland, S CHMID (2003) finds a positive influence of the fraction of shares owned by officers and directors on firm value. In the Norwegian market, B ØHREN AND Ø DEGAARD (2004) and M ISHRA , R ANDØY AND J ENSSEN (2001) show that firm performance is positively related to insider ownership and founding family ownership respectively. In Japan C HEN , G UO AND M ANDE (2003) validate that performance increases monotonically along with the level of managerial ownership. As Japan and Germany both are often labeled as bank-dominated corporate governance systems the similarity of the results should not surprise.850 The positive effects of insider ownership on per-

849 850

Cf. Jaskiewicz, Gonzal´ez, Men´endez and Schiereck (2005). Note that mean managerial ownership for the sample of 123 Japanese firms was only 2.01% compared to the 27.8% (1993) to 29.6% (1998) found for Germany in this study. See Chen, Guo and Mande (2003), p. 271. See figure 6-10.

Empirical Analysis of Insider Ownership

261

formance for Germany and Japan are also confirmed by S EIFERT, G ONENC AND W RIGHT (2005) in their multi-country study.851 Especially earlier studies on the U.S. market also claim the presence of a positive, causal relationship between insider ownership and corporate performance. Among the most known studies which focused on insider ownership and performance are those of M ORCK , S HLEIFER AND V ISHNY (1988), M C C ONNELL AND S ERVAES (1990) and M EHRAN (1995). Contrary, later U.S. studies predominantly show that the level of insider ownership is the optimal output of an equilibrium process and endogenous to performance. Examples are the studies of AGRAWAL AND K NOEBER (1996), C HO (1998), H IMMELBERG , H UB BARD AND PALIA (1999) and D EMSETZ AND V ILLALONGA (2001). To conclude, as it repeatedly arose from the preceding analysis, the functionality of corporate governance seems to vary largely across different countries or more generic across different corporate governance systems. Since the results of this study on the German market are by and large supported by findings from other corporate law countries, it is fair to say that there is more evidence in favor of than against a positive impact of insider ownership on corporate performance in Germany. As a simple linear effect of insider ownership on performance is detected in the empirical analyses, the incentive aspect of insider ownership, i.e. the alignment of shareholders’ and managers’ interests, seems to dominate negative effects of potential expropriation or insufficient diversification of management’s personal wealth. Despite the rather consistent results of this study, any conclusion based on these findings should be regarded with caution because of the problems involved in this study — which largely apply to most other corporate governance studies as well — are not negligible. 6.10

Limitations of the Empirical Study

Having presented comprehensive evidence on the determinants and effects of insider ownership, now unresolved problems and limitations of the study will be discussed. The non trivial problems of (empirical) corporate governance studies are summarized by B ØHREN AND Ø DEGAARD (2004) in the following statement: 851

For Europe, Maury (2006) finds a positive impact of family control — measured similarly as insider ownership in this study — on performance in a multi-country study.

262

Empirical Analysis of Insider Ownership “Corporate governance is a young academic field characterized by partial theories, limited access to high-quality data, inconsistent empirics, and unresolved methodological problems.”852

An excellent survey of the methodological issues typically present in empirical corporate governance studies — with a special focus on the German market — ¨ ¨ is provided by B ORSCH -S UPAN AND K OKE (2002). They identify four main categories of econometric problems: Endogeneity (or reverse causality), sample selectivity, missing variables and measurement errors.853 While the problem of endogeneity has already been addressed in sections 4.2.1 and 6.9.4.2, the extent of the other three problems in the context of this study will be discussed in the next paragraphs.854 The sample selectivity issue can generally be subdivided into two separate problems which might cause a so-called sample selection bias: First, the sample is focused on large listed corporations and therefore does not contain particular types of firms (e.g. private corporations, financial companies). Second, because performance is measured over a certain period companies which lack a performance history over the whole sample period (e.g. delisting, bankruptcy, takeover, IPO) are eliminated from the sample (sample attrition).855 The first problem is considered to play a minor role since agency problems are expected to be much more significant in listed than in unlisted companies, where ownership is typically less dispersed. Therefore, the selection criterion that companies must be publicly listed appears plausible. Nevertheless, the decision to go public is likely to depend on performance and, hence, “[. . . ] assuming that performance is irrelevant for the listing decision certainly is a strong assumption.”856 Consequently, the results only apply to the universe of public firms, which only represent a sub-group of the universe of all German firms. The results do not apply to the financial sector because financial companies, whose valuation ratios and accounting profits are not comparable to those 852 853 854

855 856

Bøhren and Ødegaard (2004), p. 18. See B¨orsch-Supan and K¨oke (2002), pp. 295-326. Another overview of econometric problems related to empirical ownership studies is provided by Short (1994), pp. 224-227. See B¨orsch-Supan and K¨oke (2002), pp. 305-311. B¨orsch-Supan and K¨oke (2002), p. 307.

Empirical Analysis of Insider Ownership

263

of non-financial companies, are excluded from the sample.857 With regard to the second problem (sample attrition), a bias would only be introduced if firms eliminated from the sample show systematically biased insider ownership levels. While this cannot be disproved (otherwise the companies would have been included in the sample), there are no compelling reasons why such a bias should exist.858 The problem of missing variables (or omitted variables) can only be avoided if all theoretically relevant governance mechanisms are included in the empirical models. Besides the ownership structure, financial structure, board structure, product market competition and past takeover activity might represent such mechanisms.859 While some of these mechanisms (i.e. ownership structure and financial structure) are included in the regression models, others (e.g. product market competition and past takeover activity) could not be included because of data constraints.860 One specific example of a probably misspecified variable might be the outside blockholder variable (BLOCK O), which does not differentiate among different blockholder types. Hence, it can be claimed that variables measuring the blockholdings of various owner types are missing. While this study found evidence that the level of outside blockholdings generally has a positive impact on corporate performance, it could not disentangle the effects of different types of external blockholdings. As discussed in section 3.3.2, different parties of shareholders are likely to have diverging interests and, therefore, might be variably effective in monitoring management.861 It was also tested within the scope of this study whether differences amongst the effects of different outside shareholder groups on performance can be detected but the number of instances for each blockholder group was too small.862 Hence, the interpretation of a positive impact of outside blockholdings on performance 857

858

859 860

861 862

Cf. Maury (2006), pp. 336-339. He shows that the inclusion of financial firms leads to insignificant results but does not change the sign of family ownership coefficients. The findings of Iber (1987), p. 176, indicate that the probability of delistings for family firms is comparatively low and hence any bias should not be severe. See section 2.2. See B¨orsch-Supan and K¨oke (2002), pp. 311-316. Furthermore, it is argued that more partial analyses of individual corporate governance mechanisms in Germany are necessary before an integrated model makes sense. Cf. Pound (1988), pp. 237-266; Barclay and Holderness (1991), p. 870. As no useful insights have been gained from this analysis, the results are not reported.

264

Empirical Analysis of Insider Ownership

should be regarded with caution because it may not apply to all owner types to the same extent.863 Finally, measurement errors in variables can lead to biased results. Measurement errors are obviously present in the case of ownership variables.864 As described in section 6.2.2, ownership disclosure requirements changed during the sample periods making comparisons amongst the three sub-samples difficult. Measurement errors also exist within each sub-sample, since notification according to the WpHG is necessary only if the thresholds of 5%, 10%, 25% and 50% are passed but not if e.g. an ownership stake rises from 25% to 49%, which in fact represents a significant change. Moreover, certain exceptions from the notification and publication rules concerning share ownership introduce further measurement errors: For example, individual funds of investment companies (“Kapitalanlagegesellschaften (KAGs)”) must only report stakes greater than 10% in contrast to the 5% threshold valid for other shareholders. Since funds usually do not hold stakes exceeding 10%, stakes owned by KAGs are essentially not disclosed.865 Finally, there is few evidence on how market participants comply with the notification and publication rules since enforcement of these standards is vague.866 As discussed in this section natural limits are posed to the interpretation of the results. The next chapter will outline which implications for theory and practice can be derived from the empirical results.

863 864

865

866

Cf. Short (1994), pp. 221, 228. See B¨orsch-Supan and K¨oke (2002), pp. 316-320. For the U.S., Dlugosz, Fahlenbrach, Gompers and Metrick (2006) document the measurement error of blockholdings if certain ownership databases are used. Cf. Kole (1995); Anderson and Lee (1997). See Becht and B¨ohmer (2003), p. 16. § 32 (2) InvG (formerly § 10 (1a) KAGG) specifies the limitation to § 21 WpHG. According to § 39 (2, 4) deliberate and airy offences against § 21 WpHG are regulatory offences which can be fined with up to EUR 200,000. For an overview of how ownership disclosure can be legally avoided see Becht and B¨ohmer (2003), pp. 15-26. Cf. Bott (2002), pp. 201-229.

Conclusion and Implications

7 7.1

265

Conclusion and Implications Conclusion

The subject of insider ownership — or, more generally, ownership structures — have been discussed from different perspectives during the course of this study: The insights from the theoretical analysis of shareholders’ interests, the survey of empirical evidence and the empirical study on the German market, make “[. . . ] the ‘managerialist’ assumption that ownership structure does not matter[. . . ]”867 hard to believe. Nevertheless, the main finding that insider ownership has a positive impact on corporate performance should be regarded with caution because of the mentioned limitations involved in the empirical analysis. Evidence for the importance of ownership structures for corporate performance is still rare and ambiguous for the case of Germany. However, the body of international evidence is growing that “[. . . ] firm value is not independent of who owns the firm. [. . . ] Thus, the modern public corporation is like any other asset in that its value depends in part on the skills of its owners.”868 The empirical and theoretical analyses both suggest that insiders who own significant stakes in their companies may be more successful in managing their businesses. This observation applies to operating performance (as measured by the return on assets) as well as to stock market performance (as measured by total stock returns and market-to-book ratios). The positive relationship between insider ownership and corporate performance seems to be linear, i.e. it holds for low, intermediate and high levels of insider ownership. In contrast to insider dominated firms, corporate governance problems may be more severe in diffuse ownership corporations. This leads J ENSEN (1989) to the conclusion that: “[t]he publicly held corporation, the main engine of economic progress in the United States for a century, has outlived its usefulness in many sectors of the economy and is being eclipsed. [. . . ] Takeovers, corporate breakups, divisional spin-offs, leveraged buyouts, and going-private transactions are the most visible manifestations of a massive organizational change in the economy.”869 867 868

869

Kang and Sorensen (1999), p. 140. Barclay and Holderness (1991), p. 878. This statement was given in the context of block trades but also seems to preserve its validity with regard to insider ownership. Jensen (1989), p. 61.

266

Conclusion and Implications

In contrast to this rather drastic conclusion, Z OBEL (2002) derives a more moderate conclusion regarding the future of diffuse and concentrated ownership corporations: “In the West, the pendulum seems to swing between the model of the tightly controlled family company and that of the widely held public company. I feel strongly that between these two extremes lies the most sustainable model.”870 The diffuse ownership corporation is rather the exception than the rule as the empirical analysis of the German equity market shows. Furthermore, the findings indicate that ownership structures of German listed companies are currently subject to significant changes. It can be expected that ownership structures will change even more in the future because of a variety of reasons: advance of funded pension schemes, financial liberalization, increasing investments by foreign investors, privatization efforts and establishment of stronger shareholder protections standards.871 As presently insider dominated (family) firms constitute a substantial portion of future IPO candidates, corporate insiders can be expected to maintain their role as a decisive shareholder group in the German market. To conclude this study, the next section will summarize its implications for research, theory and practice. 7.2 Implications 7.2.1

For Research and Theory

Future empirical, as well as, theoretical research could benefit from the insights gained from the analyses carried out in this study. With respect to research designs, the issue concerning the potential endogeneity of German ownership data deserves further attention.872 Therefore, event studies and panel data analyses appear to be promising approaches. Event studies873 could be used to examine performance effects in the context of arguably exogenous changes in ownership — e.g. due to changes in tax policy or regulation — and thereby avoid 870 871 872 873

Zobel (2002), p. 133. See Gillan and Starks (2003), p. 17. Cf. K¨oke (2002), pp. 138-139; B¨orsch-Supan and K¨oke (2002), pp. 297-305. The drawbacks of event studies for the purpose of examining the insider ownership-performance relationship have already been discussed in section 6.1.

Conclusion and Implications

267

the problem of endogeneity.874 However, first of all, the quality and quantity of ownership data would have to improve and increase respectively. Otherwise, an accurate measurement of changes in the level of insider ownership around certain event dates or across time cannot be achieved. The fact that insider ownership appears to be a rather inert variable (see section 6.9.4.2) imposes natural limits on the results which could be derived from panel data analyses. Nevertheless, new research approaches might be extremely valuable for analyzing in more detail why or in which cases insider ownership leads to better performance of German public stock corporations. As discussed, the availability and quality of ownership data has already amended and is likely to further improve in the future. Thereby, new research perspectives will be offered. Meanwhile, it would be interesting to estimate the magnitude of measurement errors introduced by the deficiencies of the currently available data sources. A comparison of the most frequently used public and private data sources — e.g. the database of major holdings of voting rights in officially listed companies (BaFin), Hoppenstedt Aktienf¨uhrer, “Wer geh¨ort zu wem?” (Commerzbank), Deutsche B¨orse and Commercial Registers (shareholders meetings’ vote casts) — is likely to result in large differences on the company level. An analysis of the causes of such differences and their effects on economic analysis would yield interesting insights into the nature of German ownership data, assist data providers with improving the quality of their data and finally ease the selection of the appropriate data source for forthcoming research.875 The appropriate realization of ownership structures represents another challenge to all related research. Both dimensions of ownership structures — i.e. shareholders’ identity and shareholders’ concentration — should be examined to provide a comprehensive picture on the control setting in public companies. For any individual company, the control setting can be effectively analyzed by looking at the size of the registered shareholdings and the identity of the respective owners. However, it remains unclear how both dimensions — 874 875

See Himmelberg, Hubbard and Palia (1999), p. 382. For the U.S., Anderson and Lee (1997) compare ownership data of five sources and conclude that the choice of the source might affect economic inferences drawn from any analysis based on these sources. In contrast, Kole (1995) concludes in her comparison of three sources of managerial ownership data that differences among these data sources cannot explain contradictory empirical evidence.

268

Conclusion and Implications

and the corresponding interdependencies — can be realized simultaneously into one or more ownership variables to enable large scale economic analyses. As suggested by the theoretical analysis, different types of blockholders (or more generally shareholders) are likely to have different auxiliary interests. Therefore, it appears desirable to distinguish among different blockholder groups in empirical analyses as well.876 Next, managerial ownership should be analyzed in the broader framework of managerial compensation as the rationale of the effects of insider ownership on corporate performance is primarily based on incentive arguments. While some U.S. studies examine executive compensation along with executive ownership877 , this aspect remains largely unresearched for the case of Germany. For example, the question of whether variable compensation represents an adequate (or even better) substitute for insider ownership would yield new interesting insights into the compensation system of German public corporations.878 Such analyses became possible just recently when the transparency of management board compensation increased due to new regulations.879 The combined analysis of management board members’ insider ownership stakes and compensation packages is likely to provide a more complete picture of managers’ economic incentives than the partial analysis of insider ownership. The same argument may apply for the case of supervisory board members as well. While the theory on the effects of insider ownership on corporate performance is well developed, less (empirical) research exists on how insider ownership affects business operations. A detailed analysis of how corporate ownership structures influence the investment, R&D, personnel and strategic choices of corporations might corroborate or disprove frequently cited vague presumptions.880 The role of these firm characteristics for insider dominated companies or family firms remains largely unresearched with personnel aspects being the only ex876 877 878

879 880

See section 3.3; Becht and B¨ohmer (2003), p. 13. Cf. Andres (2006). See amongst others Mehran (1995); Chung and Pruitt (1996); Weber and Dudney (2003). For a U.S. survey on option-based compensation see Muurling and Lehnert (2004). Cf. Kaserer and Wagner (2004). See section 2.4.2. Hasler (2004), pp. 6-7, rather arbitrarily lists the following characteristics of family firms: less risky investments, long-term planning, focus on niche markets, continuous consolidation of market position, focus on long-term not short-term value creation, comprehensive business know-how, high commitment and entrepreneurial spirit, managers loyal to the company, fast decision making and less bureaucracy, value rather than growth orientation, family values determine corporate values.

Conclusion and Implications

269

ception. For example, the question of whether family or outside managers lead to better performance has attracted some (primarily non-German) research.881 Besides an examination of the functional aspects of insider ownership, it would be interesting to observe its role at different stages of firms’ life-cycle (e.g. IPO, expansion, mergers and acquisitions, delistings and bankruptcies). Finally, such analyses might be able to clarify whether a real difference between insider ownership and family ownership exists or whether both terms refer basically to the same underlying concept. As already discussed, the interdependencies among different corporate governance mechanisms are likely to be enormous while the theoretical foundations for a structural model of how corporate governance works are still lacking. Some researchers argue that individual corporate governance mechanisms should always be analyzed together with other — potentially substitutional — corporate governance mechanisms882 while others claim that current theory cannot properly capture how performance relates to a wider set of corporate governance mechanisms.883 Nevertheless, also the former group partially admits that the simultaneous choice of individual governance mechanisms still needs to be modeled. Thus, from both perspectives more research on this issue appears necessary.884 With respect to the simultaneous choice of the respective governance mechanisms, more comparative analysis of corporate law oriented corporate governance systems (e.g. Germany) and capital market oriented systems (e.g. the U.S.) may offer promising research perspectives. While the differences among corporate governance systems may be too large to enable empirical analyses of corporate governance mechanisms in multi-country studies, a detailed examination of why a particular mechanism only seems to work in certain corporate governance systems is likely to yield interesting insights. For example, the question of why insider ownership seems to be an effective corporate governance mechanism only in corporate law oriented systems but not in capital market oriented systems still has to be answered. While it has already been argued that this dif881

882 883 884

See Barth, Gulbrandsen and Schøne (2005); Anderson and Reeb (2003). For the U.S., Aggarwal and Samwick (2006) also analyze the relationship between managerial incentives and corporate investments. See amongst others Agrawal and Knoeber (1996); Beiner, Drobetz, Schmid and Zimmermann (2006). See Bøhren and Ødegaard (2004); Kaserer and Moldenhauer (2005). See Himmelberg, Hubbard and Palia (1999), pp. 382-383.

270

Conclusion and Implications

ference might stem from the different nature of insider ownership — i.e. insider ownership is primarily founder-driven in Germany vs. compensation-driven in the U.S.885 — more evidence on this hypothesis is desirable. Similar differences can be expected to exist for other corporate governance mechanisms as well. Finally, some general remarks on promising research on the ownershipperformance relationship will conclude this section. Some researchers call for a broader perspective and more historical and interdisciplinary (e.g. sociology or social psychology) research as the problems of principal agent settings are not new and the results from conventional analyses are rather unclear and ambiguous.886 Furthermore, corporate governance research should always pay appropriate attention to the legal environment as it considerably shapes the choice and effectiveness of the individual corporate governance mechanisms.887 Last, one might critically ask whether the traditional theoretical arguments — as well as the corresponding empirical results — still hold, when the nature of firms has fundamentally changed. Motivational and control aspects in modern, human capital-oriented service firms might be fundamentally different from those in traditional asset-based production companies.888 7.2.2

For Practice

The results of the empirical analysis show that shareholder structures do matter for firms’ corporate governance situation. The level of insider ownership affects board sizes and the extent of management’s excess control rights which can be seen as corporate governance related firm characteristics. Having discussed possible positive and negative effects of insider ownership on corporate performance, the empirical findings suggest that higher insider ownership levels translate into better operating performance (as measured by the return on assets) which in turn translates into higher market valuations (as measured by buy-and-hold returns and market-to-book ratios). Hence, it seems that investors — or more general capital market participants — should pay (more) attention 885 886

887

888

See section 6.9.4.2. See Weber (1998), p. 168, who compares the problem to motivate and control managers in modern corporations to that of companies sending sailing ships to East Asia in the 18th century which had to motivate and control their captains as well. See Thomsen, Pedersen and Kvist (2006), p. 266. Cf. La Porta, Lopez-de Silanes, Shleifer and Vishny (2002). See Zingales (2000); Kang and Sorensen (1999), p. 140.

Conclusion and Implications

271

to the shareholder structures of public stock corporations. Some first endeavors can already be observed: For example, the three big international rating agencies (Standard & Poor’s (S&P), Moody’s and FitchRatings) meanwhile all offer corporate governance ratings that, amongst others, explicitly incorporate aspects of corporations’ ownership structures.889 Another example would be the German Entrepreneurial Index (GEX), a stock market index introduced by Deutsche B¨orse AG, the operator of the largest German stock exchange, in the beginning of 2005.890 The GEX is the first known, official stock index which uses shareholder structures as a selection criterion. Thereby a new type of style-index is created.891 Primarily, stock indices should provide information on the aggregated development of the underlying shares (“information function”) and serve as a benchmark (“benchmark function”). Besides, stock indices also bear an investment and trading function.892 With regard to the investment function, passive investors can replicate the index portfolio and/or even might treat the entrepreneurial firms of the GEX as a separate asset class. Even though some preliminary evidence exists, the question of whether insider dominated firms represent an independent asset class still has to be explored. With respect to the trading function, the GEX enables private individuals to invest into a diversified group of insider dominated firms by buying derivative financial products on the index.893 Besides the example of the GEX, the case of Deutsche B¨orse AG shows that not only the company’s management but also potent outside shareholders may exert significant influence on decisions affecting corporate value as acquisitions, dividends or share buy backs.894 The growing awareness about the importance of corporate governance can also be derived from the increased percentage of voting rights present at German shareholders’ meetings895 or the

889 890 891 892 893 894 895

See Brink and Romeike (2005), pp. 135-139. See section 1.1; figure 1-1. See Deutsche B¨orse (2005a); section 1.1. See Schmitz-Esser (2001), pp. 15-98. See Kaserer, Achleitner, Moldenhauer and Ampenberger (2006), pp. 13-17. See Eckert and Zsch¨apitz (2005). See footnote 323.

272

Conclusion and Implications

rising number of cases where investors actively opposed management’s proposals.896 From a manager’s perspective, the results of this study suggest that an entrepreneurial involvement in his firm, i.e. the willingness to considerably participate in his company’s chances and risks, finally pays off for shareholders. Therefore, managers might try to acquire (significant) stakes in their company or increase their current shareholdings to signal their confidence with respect to the company’s development to outsider shareholders. Supervisory board members should explicitly take managers’ (lacking) ownership stakes into account by determining management board members’ compensation packages. The same argument also applies to the compensation of supervisory board members, which in turn is determined by the shareholders’ meeting or the corporate charter. Finally, human resource departments might recognize the incentives provided by share ownership also for lower level employees. If shareholder structures matter indeed, policy makers should try to further increase the transparency of shareholdings of listed companies. As mentioned, current sources of ownership data are problematic with respect to both data quality and quantity. This is (partially) driven by comparatively lax disclosure requirements or a lack of enforcement respectively.897 Therefore, it appears desirable to extend the disclosure regulations regarding ownership stakes in public stock corporations and to enforce such regulations effectively.898 As for other corporate governance mechanisms as well — e.g. the remuneration of management and supervisory board members — a corporation’s ownership structure must be transparent to enable its effectiveness as a corporate governance mechanism. Especially minority shareholders should be aware of threatening expropriation by blockholders and might in turn increase their monitoring activities. Another compelling argument for transparent ownership structures is put forward by B ECHT (1997):

896

897 898

For example, capital raises with exclusion of subscription rights for fractional amounts were recently denied by shareholders at the shareholders’ meetings of Continental AG and Hypo Real Estate AG. See Hussla (2006), p. 25. See sections 6.2.2 and 6.10. See Becht (1997), p. 54.

Conclusion and Implications

273

“The best argument for increased disclosure is that there are no good arguments against disclosure.”899 Even though this statement appears somewhat oversimplified, as the administrative costs of additional regulations for both the reporting and the supervising party can turn out to be significant, the key message is noteworthy. Furthermore, one might argue that an extensive disclosure of share ownership is even more important in the case of Germany as the degree of ownership concentration is high.900 While the transparency of directors’ dealings has increased significantly trough the introduction of § 15a WpHG on July 1, 2002 (and the corresponding amendments of the AnSVG on October 30, 2004), the total share ownership of individual insiders is only available from the corporations’ financial statements or — with adverse effects — from diverse public (e.g. BaFin’s database of voting rights in officially listed companies according to § 21 WpHG) or private sources (e.g. Hoppenstedt Aktienf¨uhrer). In contrast, the U.S. Securities and Exchange Commission (SEC) provides detailed information on the absolute level and changes in share ownership by all officers and directors online through their Electronic Data Gathering, Analysis, and Retrieval System (EDGAR) (see appendix F).901 Finally, one might argue that comprehensive disclosure strengthens outside control by the capital market as “[. . . ] with adequate disclosure and transparency standards in place, it is ultimately the capital market which rewards good governance practices and punishes bad ones.”902 Concerning an effective control through the capital market, the findings of the empirical study suggest that the German equity market is still underdeveloped. Therefore, the promotion of the capital market should be accelerated. Financial markets have become increasingly integrated so that German companies which aim at issuing large volumes of equity have to play by the rules of the international — Anglo-Saxon oriented — financial markets.903 The large diffusion of family control and the means by which it is exerted seem to be less a problem 899 900 901

902 903

See Becht (1997), p. 60. See section 6.4.1.2; Becht (1997), p. 60. See Bundesanstalt f¨ur Finanzdienstleistungsaufsicht (BaFin) (2006); Securities and Exchange Commission (SEC) (2006). Drobetz, Schillhofer and Zimmermann (2004), p. 291. See Matthes (2000), p. 39. Cf. Coffee and Berle (1999) for a perspective of global convergence in corporate governance and its implications.

274

Conclusion and Implications

than the general conditions.904 Numerous efforts to adopt international financial market standards, e.g. rules regarding mandatory takeover offers or consistent accounting standards, have taken place in Germany in the past decade. Nevertheless, other aspects relevant for well-functioning capital markets — e.g. the effective prosecution of insider trading905 , the abolishment of the tax discrimination of equity investments or the provision of sufficient risk capital for young companies — still remain underdeveloped.906 The lacking breadth of the German capital market induces high market impact costs of larger investment volumes which are typical for institutional investors or control seeking parties.907 It should be noted that a vicious circle (or chicken-egg problem) between market liquidity and the inertia of ownership structures exists: On the one hand, an illiquid market prevents shareholders from selling their stakes without (significant) market impact costs. On the other hand, without more trading activity the stock market will not become more liquid. This is an important issue since the underdevelopment of the German capital market indirectly prevents other governance mechanisms, e.g. takeovers, from exercising their power.

904 905 906

907

See Barontini and Caprio (2005), p. 23. See footnote 127. Cf. Monopolkommission (1998), p. 122. For example, the Law on Private Equity Companies (“Gesetz u¨ ber Unternehmensbeteiligungsgesellschaften” (“UBGG”)) largely failed to improve the conditions for private equity investments. See Monopolkommission (1998), p. 26.

Appendix

Appendix

275

277

Appendix

A

Exemplary Shareholder Structure Classification

STEP 1: COLLECT SOURCE DATA (Company A) Management board Supervisory board John Q. Public, Mr. Unknown Mrs. X, Mr. Y, Mr. Z Share capital EUR 25,000,000.- share capital - EUR 15,000,000.- ordinary share capital - EUR 10,000,000.- preference share capital Shareholders Ordinary Shares (in %) Company B 35.01 John Q. Public 8.50 12.00 JQP Asset Mgmt. GmbH (100% owned by John Q. Public) Money Investment Fund Ltd. 5.50 Freefloat 38.99 TOTAL 100.00

Preference Shares (in %) 0.00 0.00 5.00 0.00 95.00 100.00

STEP 2a: SHAREHOLDER IDENTITY Shareholders Company B John Q. Public JQP Asset Management GmbH Money Investment Fund Ltd. Freefloat

Direct Owner-Type COR MB INTM INST-F FF

Ultimate Owner-Type COR MB MB INST-F FF

STEP 2b: SHAREHOLDER IDENTITY - SAMPLE VARIABLES Ownership at Ultimate Level MB COR INST-F FF TOTAL

Control Rights (in %) 20.50 35.01 5.50 38.99 100.00

Cash Flow Rights (in %) 14.30 21.01 3.30 61.39 100.00

STEP 3: SHAREHOLDER CONCENTRATION - SAMPLE VARIABLES Concentration Variable C1 BLOCK O BLOCK O T1 BLOCK O T2 BLOCK O T3 BLOCK O T4 = BLOCK NO

2

Control Rights (in %) 35.01 40.51 5.50 35.01 n/a

Owner Type COR n/a INST-F COR n/a

279

Appendix

B

Results I: Determinants of Insider Ownership

Sample Period

1993

1998

Independent Variable LN ASSETS

FIRM RISK

Dependent Variable: MGMT

Constant

1.128 (7.766) LN ASSETS -0.066 (-5.907) YEAR 1998 YEAR 1993 -

*** 1.019 *** 0.882 (-9.691) (7.553) *** -0.058 *** -0.046 (-6.954) (-5.140) -

1.001 (14.085) *** -0.055 (-10.364) -0.012 (-0.450) -0.010 (0.424)

n R2 R2 adj. F-Statistic

214 0.141 0.137 34.895

349 0.122 0.120 *** 48.353

261 0.093 0.089 *** 26.424

824 0.116 0.113 *** 35.973

Constant

0.176 (5.102) 0.110 (1.552) -

*** 0.207 (7.652) 0.088 (2.691) -

*** 0.275 *** 0.264 (12.039) (11.554) *** 0.009 0.017 (1.008) (1.905) -0.054 (-1.680) -0.019 (-0.655)

FIRM RISK YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic a

Pooleda

2003

165 0.015 0.009 2.498

**

244 0.029 0.025 7.239

286 0.004 0.000 *** 1.016

***

695 0.014 0.009 3.196

*** ***

*** *** * *

***

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients. t-statistics reported in parentheses. *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table B-1: R I: Simple Regression

280

Appendix

Sample Period

1993

1998

2003

Pooled

11.456 14.251

11.080 14.049

11.305 14.457

11.236 14.237

0.387 0.205 *** (5.672)a 0.000

0.374 0.203 *** (4.658)b 0.000

0.383 0.196 *** (9.052)a 0.000

35,832 25,244 *** (-5.744) 0.000

20,154 14,037 *** (-5.082) 0.000

200,150 139,750 *** (-9.227) 0.000

Grouping Variable LNASSETS

Mean (Small) Mean (Large)

Panel A: T-Test for Independent Samples MGMT

Mean (Small) Mean (Large) (t-statistic) Significance

0.375 0.181 (4.754)a 0.000

***

Panel B: Mann-Whitney U-Test MGMT

a b

Rank sum (Small) Rank sum (Large) (Z-statistic) Significance

13,424 9,582 (-4.517) 0.000

***

Different variances assumed according to Levene-test of equal variances (0.05 significance level). Equal variances assumed according to Levene-test of equal variances (0.05 significance level). *, ** and *** denote differences in means significant on the 0.10, 0.05 and 0.01 level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table B-2: R I: Difference in Means I Sample Period

1993

1998

2003

Pooled

0.186 0.533

0.209 0.864

0.397 2.612

0.245 1.542

Grouping Variable FIRMRISK

Mean (Low) Mean (High)

Panel A: T-Test for Independent Samples MGMT

Mean (Low) Mean (High) (t-statistic) Significance

0.164 0.268 (-2.241)a 0,026

0.205 0.303 ** (-2.369)a 0.019

0.270 0.309 ** (-1.063)a 0.289

0.210 0.309 (-4.255)b 0.000

13,733 16,158 (-2.392) 0.017

19,252 21,790 (-1.875) 0.061

108,902 132,958 (-4.919) 0.000

***

Panel B: Mann-Whitney U-Test MGMT

a b

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

6,278 7,417 (-2.249) 0.250

**

**

*

***

Different variances assumed according to Levene-test of equal variances (0.05 significance level). Equal variances assumed according to Levene-test of equal variances (0.05 significance level). *, ** and *** denote differences in means significant on the 0.10, 0.05 and 0.01 level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table B-3: R I: Difference in Means II

281

Appendix Sample Period

1993

1998

Independent Variables Constant LN ASSETS FIRM RISK YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic a

Pooleda

2003

Dependent Variable: MGMT 1.030 (6.015) -0.063 (-5.085) 0.000 (0.005) -

***

0.869 (5.649) *** -0.049 (-4.387) 0.049 (1.324) -

***

0.896 *** 0.957 *** (6.790) (10.808) *** -0.047 *** -0.052 *** (-4.886) (-8.283) -0.003 0.001 (-0.224) (0.114) -0.067 ** (-2.128) -0.030 (-1.092)

151 0.157 0.146 13.777

238 0.102 0.095 *** 13.370

258 0.095 0.088 *** 13.361

647 0.115 0.110 *** 20.920

***

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients. t-statistics reported in parentheses. *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table B-4: R I: Multiple Regression

283

Appendix

C

Results II: Insider Ownership and Corporate Control

Sample Period

1993

1998

Independent Variable MGMT

Constant MGMT YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic

Pooleda

2003

Dependent Variable: SB NO 10.880 (22.871) -0.100 (-0.089) -

***

11.343 (31.643) -7.438 (-8.943) -

236 0.000 -0.004 0.008

356 0.184 0.182 79.971

*** ***

***

10.312 (25.437) -5.828 (-6.066) 284 0.115 0.112 36.802

*** ***

***

10.048 *** (28.923) -4.914 *** (-8.763) 2.168 *** (4.753) 0.548 (1.331) 876 0.105 0.102 34.171

***

Dependent Variable: MB NO MGMT

Constant MGMT

4.005 (21.220) 0.150 (0.333)

***

3.859 (28.354) -1.208 (-3.837)

*** ***

3.567 (24.812) -1.034 (-3.052)

*** ***

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic a

236 0.000 -0.004 0.111

354 0.040 0.037 14.724

***

283 0.032 0.029 9.313

***

3.492 *** (26.954) -0.779 *** (-3.736) 0.771 *** (4.542) 0.240 (1.561) 873 0.040 0.036 12.003

***

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients. t-statistics reported in parentheses. *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table C-1: R II: Simple Regression I

284

Appendix

Sample Period

1993

1998

Independent Variable MGMT

Constant MGMT

Pooleda

2003

Dependent Variable: VOTE 0.126 (4.621) 0.167 (2.571)

*** **

0.123 (5.541) 0.055 (1.074)

***

0.112 (4.715) 0.024 (0.426)

***

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic

236 0.027 0.023 6.612

357 0.003 0.000 1.153

**

290 0.001 -0.003 0.181

0.096 (4.770) 0.076 (2.314) 0.055 (2.064) 0.020 (0.851) 884 0.011 0.007 3.165

*** ** **

**

Dependent Variable: MGMT CRmCF MGMT

Constant MGMT

0.014 (1.487) 0.129 (5.893)

***

0.018 (2.274) 0.078 (4.148)

** ***

0.011 (1.448) 0.063 (3.469)

***

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic a

236 0.129 0.126 34.732

***

358 0.046 0.043 17.203

***

290 0.040 0.037 12.035

***

0.004 (0.596) 0.087 (7.691) 0.021 (2.282) 0.012 (1.394) 884 0.068 0.065 21.340

*** **

***

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients. t-statistics reported in parentheses. *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table C-2: R II: Simple Regression II

285

Appendix Sample Period

1993

1998

Independent Variable MGMT

Constant MGMT

Dependent Variable: BLOCK O 0.590 (27.154) -0.874 (-16.894)

***

0.581 (30.712) *** -0.861 (-19.610)

***

0.555 (25.456) *** -0.808 (-15.641)

0.566 *** (32.675) *** -0.848 *** (-30.127) 0.016 (0.707) 0.010 (0.499)

235 0.549 0.548 285.400

358 0.519 0.518 *** 384.555

290 0.459 0.457 *** 244.636

884 0.508 0.506 *** 303.054

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic

Pooleda

2003

***

***

Dependent Variable: BLOCK NO MGMT

Constant MGMT

1.460 (20.295) -1.992 (-11.633)

***

1.471 (23.460) *** -1.883 (-12.933)

236 0.366 0.364 135.317

358 0.320 0.318 *** 167.265

***

1.653 (19.910) *** -1.964 (-9.981)

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic

***

290 0.257 0.254 99.629

***

1.646 *** (27.167) *** -1.939 *** (-19.719) -0.201 ** (-2.515) -0.158 ** (-2.202) 884 0.310 0.308 *** 131.931

***

Dependent Variable: DEBT RATIO MGMT

Constant MGMT

0.215 (12.408) 0.002 (0.041)

***

0.216 (14.466) -0.043 (-1.236)

***

0.296 (12.771) 0.077 (1.401)

YEAR 1998 YEAR 1993 n R2 R2 adj. F-Statistic a

206 0.000 -0.050 0.002

341 0.004 0.002 1.527

253 0.008 0.004 1.962

***

0.317 (20.089) 0.006 (0.228) -0.103 (-4.929) -0.115 (-6.237) 801 0.051 0.048 14.384

***

*** ***

***

The pooled models include two dummy variables to account for differences between the sub-samples. OLS-regression coefficients. t-statistics reported in parentheses. *, ** and *** denote significance on the 0.10, 0.05 and 0.01 level. For a definition of the variables see tables 6-5 to 6-7. Source: Author’s findings.

Table C-3: R II: Simple Regression III

-0.284 0.000 876

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

SB NO

MB NO

VOTE

MGMT CRmCF

BLOCK O

BLOCK NO

DEBT RATIO

***

***

***

**

***

***

0.058 0.100 801

0.243 0.000 876

0.174 0.000 876

0.009 0.800 876

0.014 0.680 876

0.594 0.000 872

SB NO

*

***

***

***

0.015 0.682 799

0.084 0.013 873

-0.021 0.539 873

0.075 0.028 873

0.050 0.142 873

MB NO

**

**

0.121 0.001 801

-0.152 0.000 884

-0.115 0.001 884

0.764 0.000 884

VOTE

***

***

***

***

Pooled Sample

0.106 0.003 801

-0.246 0.000 884

-0.289 0.000 884

MGMTCRmCF

***

***

***

-0.140 0.000 801

0.557 0.000 884

BLOCK O

***

***

0.053 0.131 801

BLOCKNO

Table C-4: R II: Correlation Analysis

*, ** and *** denote Bravais-Pearson correlation coefficients significant on the 0.10, 0.05 and 0.01 significance level (2-sided). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

0.009 0.798 801

-0.552 0.000 884

-0.713 0.000 884

0.249 0.000 884

0.076 0.023 884

-0.127 0.000 873

MGMT

Covariates

Sample Period

286 Appendix

287

Appendix Sample Period

1993

1998

2003

Pooled

0.000 0.555

0.016 0.576

0.021 0.559

0.012 0.566

Grouping Variable MGMT

Mean (Low) Mean (High)

Corporate Control Variables SB NO

MB NO

VOTE

MGMTCRmCF

BLOCKO

BLOCKNO

DEBTRATIO

a b

Mean (Low) Mean (High) (t-statistic) Significance

11.020 10.690 (0.463)b 0.644

11.35 6.920 (8.419)a 0.000

10.46 6.770 *** (6.302)a 0.000

10.98 7.880 *** (8.733)a 0.000

Mean (Low) Mean (High) (t-statistic) Significance

3.970 4.120 (-0.509)b 0.611

3.800 3.200 (2.979)a 0.003

3.510 3.010 *** (2.382)a 0.018

3.77 3.38 (2.956)a 0.003

Mean (Low) Mean (High) (t-statistic) Significance

0.085 0.261 (-4.403)a 0.000

0.076 0.203 *** (-4.009)a 0.000

0.089 0.148 *** (-1.735)a 0.084

Mean (Low) Mean (High) (t-statistic) Significance

0.000 0.099 (-7.442)a 0.000

0.001 0.082 *** (-7.178)a 0.000

0.000 0.059 *** (-5.394)a 0.000

0.000645 0.078602 *** (-11.456)a 0.000

Mean (Low) Mean (High) (t-statistic) Significance

0.630 0.065 (17.889)a 0.000

0.599 0.053 *** (20.163)a 0.000

0.569 0.072 *** (15.801)a 0.000

0.599571 0.059912 *** (31.571)a 0.000

Mean (Low) Mean (High) (t-statistic) Significance

1.470 0.340 (10.005)a 0.000

1.460 0.370 *** (11.578)a 0.000

1.610 0.550 *** (8.497)b 0.000

1.15 0.42 *** (17.142)a 0.000

Mean (Low) Mean (High) (t-statistic) Significance

0.201 0.230 (1.139)b 0.256

0.209 0.198 (0.504)b 0.614

0.301 0.336 (-1.050)b 0.295

0.235 0.251 (-0.948)b 0.343

**

*

0.0851 0.1978 (-5.588)a 0.000

***

***

***

***

***

***

Different variances assumed according to Levene-test of equal variances (0.05 significance level). Equal variances assumed according to Levene-test of equal variances (0.05 significance level). *, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table C-5: R II: Difference in Means (T-Test)

288

Appendix

Sample Period

1993

1998

2003

Pooled

7,021 20,945

16,110 48,151

10,585 31,610

97,903 293,267

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

14,173 13,793 (-0.370) 0.711

39,243 24,303 (-7.651) 0.000

24,065 16,405 *** (-5.452) 0.000

223,747 160,380 *** (-8.388) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

13,578 14,388 (-0.784) 0.433

33,892 28,943 (-2.627) 0.009

21,334 18,853 *** (-1.740) 0.082

199,902 181,599 (-2.445) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

12,243 15,724 (-4.131) 0,000

29,654 34,608 *** (-3.487) 0.000

20,091 22,105 *** (-2.109) 0.035

180,593 210,577 ** (-5.392) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

11,387 16,579 (-7.159) 0.000

27,683 36,579 *** (-6.759) 0.000

18,745 23,450 *** (-5.330) 0.000

168,107 223,064 ***(-10.953) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

19,848 8,118 (-11.665) 0.000

45,583 18,678 ***(-14.242) 0.000

29,373 12,822 ***(-11.919) 0.000

276,966 114,204 ***(-22.196) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

18,791 9,175 (-9.797) 0.000

42,958 21,304 ***(-11.846) 0.000

27,324 14,871 *** (-9.184) 0.000

259,269 131,901 ***(-17.847) 0.000

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

10,237 11,085 (-0.991) 0.322

29,926 28,727 (-0.281) 0.779

15,618 16,514 (-1.098) 0.272

159,118 162,084 (-0.882) 0.378

Grouping Variable MGMT

Rank sum (Low) Rank sum (High)

Corporate Control Variables SB NO

MB NO

VOTE

MGMTCRmCF

BLOCKO

BLOCKNO

DEBTRATIO

*

***

**

***

***

***

***

*, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table C-6: R II: Difference in Means (Mann-Whitney U-Test)

Appendix

D

Results III: Insider Ownership and Corporate Performance

289

0.344 0.000 605

Correlation Significance n

Correlation Significance n

Correlation -0.110 Significance 0.004 n 690

Correlation Significance n

Correlation -0.026 Significance 0.493 n 690

Correlation -0.058 Significance 0.125 n 690

Correlation Significance n

MTBV

ROA

MB

SB

FBM

MGMT

BLOCK O

***

***

***

***

0.011 0.758 758

0.076 0.036 758

-0.009 0.803 758

-0.017 0.631 758

0.130 0.000 758

0.015 0.690 689

MTBV

**

***

-0.093 0.012 734

0.119 0.001 734

0.014 0.702 734

0.101 0.006 734

0.052 0.156 734

ROA

**

***

***

-0.403 0.000 884

0.514 0.000 884

-0.165 0.000 884

-0.184 0.000 884

MB

***

***

***

***

Pooled Sample

-0.319 0.000 884

0.462 0.000 884

-0.140 0.000 884

SB

***

***

***

-0.284 0.000 884

0.443 0.000 884

FBM

***

***

-0.713 0.000 884

MGMT

***

Table D-1: R III: Correlation Analysis I

*, ** and *** denote correlation coefficients significant on the 0.10, 0.05 and 0.01 significance level (2-sided). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

0.106 0.005 690

0.046 0.225 690

0.242 0.000 611

BAHR

Covariates

Sample Period

290 Appendix

0.344 0.000 605

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

Correlation Significance n

MTBV

ROA

LN ASSETS

FIRM RISK

DEBT RATIO

SALES G

DIV

0.015 0.690 689

0.001 0.979 610

0.183 0.000 725

*** -0.015 0.685 727

***

*** -0.337 0.000 758

***

*** -0.087 0.016 758

***

***

MTBV

-0.119 0.003 616

0.073 0.047 734

0.204 0.000 712

*** 0.134 0.000 719

*** -0.224 0.000 725

**

ROA

***

***

***

***

**

0.366 0.000 785

-0.050 0.164 781

0.117 0.001 801

-0.301 0.000 647

LNASSETS

-0.019 0.635 636

0.170 0.000 638

FIRMRISK

*** -0.400 0.000 622

***

***

Pooled Sample

***

***

-0.143 0.000 767

-0.049 0.180 763

DEBTRATIO

***

0.007 0.850 750

SALES G

Table D-2: R III: Correlation Analysis II

*, ** and *** denote correlation coefficients significant on the 0.10, 0.05 and 0.01 significance level (2-sided). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

0.422 0.000 616

0.172 0.000 630

-0.332 0.000 632

-0.317 0.000 689

0.280 0.000 641

0.242 0.000 611

BAHR

Covariates

Sample Period

Appendix 291

292

Appendix

Sample Period

1993

1998

2003

Pooled

0.000 0.555

0.016 0.576

0.021 0.559

0.012 0.566

Mean (Low) Mean (High) (t-statistic) Significance

82.283 69.339 (0.847)b 0.398

23.818 22.282 (0.138)b 0.890

-2.221 -12.094 (0.955)b 0.340

29.698 15.932 (1.910)b 0.057

Mean (Low) Mean (High) (t-statistic) Significance

2.680 2.910 (-1.009)b 0.314

2.752 3.369 (-2.371)b 0.018

2.018 2.030 (-0.055)b 0.956

2.507 2.822 (-2.148)a 0.032

Mean (Low) Mean (High) (t-statistic) Significance

1.870 4.153 (-2.503)b 0.013

3.966 7.173 (-2.988)b 0.003

2.307 3.198 *** (-0.684)b 0.495

2.864 4.926 (-3.091)a 0.002

-0.038 0.022 (-1.318)b 0.189

0.042 0.125 (-2.565)a 0.011

0.610 0.590 (0.355)b 0.723

0.690 0.690 (-0.061)b 0.951

Grouping Variable MGMT

Mean (Low) Mean (High)

Performance Variables BAHR

MTBV

ROA

**

**

*

**

***

Control Variables SALESG

DIV

a b

Mean (Low) Mean (High) (t-statistic) Significance

0.044 0.074 (-0.784)b 0.434

0.101 0.241 (-2.254)b 0.025

Mean (Low) Mean (High) (t-statistic) Significance

0.810 0.810 (-0.054)a 0.957

0.670 0.710 (-0.915)b 0.361

**

**

Different variances assumed according to Levene-test of equal variances (0.05 significance level). Equal variances assumed according to Levene-test of equal variances (0.05 significance level). *, ** and *** denote differences in means significant on the 0.10, 0.05 and 0.01 level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-3: R III: Difference in Means (T-Test)

293

Appendix Sample Period

1993

1998

2003

Pooled

7,021 20,945

16,110 48,151

10,585 31,610

97,903 293,267

Rank sum (Low) 8,669 Rank sum (High) 5,026 (Z-statistic) (-1.503) Significance 0.133

17,355 11,565 (-0.688) 0.491

21,602 19,153 (-1.657) 0.097

139,866 98,529 (-3.029) 0.002

MTBV

Rank sum (Low) 9,813 Rank sum (High) 10,488 (Z-statistic) (-0.696) Significance 0.486

25,280 25,760 (-0.972) 0.331

14,363 14,078 (-0.043) 0.965

144,011 143,650 (-0.696) 0.487

ROA

Rank sum (Low) 8,476 Rank sum (High) 12,027 (Z-statistic) (-3.299) Significance 0.001

19,132 20,489 *** (-3.179) 0.001

15,155 16,471 *** (-1.473) 0.141

126,483 143,262 (-4.202) 0.000

23,944 29,684 (-4.030) 0.000

16,331 17,600 *** (-1.477) 0.140

143,603 161,768 (-3.688) 0.000

26,593 27,693 (-0.915) 0.360

17,016 16,395 (-0.355) 0.722

152,328 156,178 (-0.061) 0.951

Grouping Variable MGMT

Rank sum (Low) Rank sum (High)

Performance Variables BAHR

*

***

***

Control Variables SALESG

DIV

Rank sum (Low) Rank sum (High) (Z-statistic) Significance

8,755 10,160 (-1.797) 0.072

Rank sum (Low) 9,239 Rank sum (High) 10,463 (Z-statistic) (-0.055) Significance 0.956

*

***

*, ** and *** denote significance on the 0.10, 0.05 and 0.01 significance level. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-4: R III: Difference in Means (Mann-Whitney U-Test)

294

Appendix

Dependent Variable Model No.

BAHR (1)

(2)

Intercept

-137.945 (-1.728)

*

MB

67.875 (2.265) 91.900 (2.3376) 82.465 (2.383)

**

SB FBM

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

93.402 (3.367) -11.152 (-2.885) 9.415 (2.645) 5.175 (0.494) -79.910 (-4.359) 28.1745 (1.9238) 61.620 (3.940)

-140.738 (-1.747)

*

**

*** *** ***

*** * ***

(3)

(4)

-0.685 (-0.673)

-0.822 (0.830)

0.797 (1.516) 0.859 (1.965) 1.421 (1.912)

**

MGMT BLOCK O

MTBV

82.590 (2.622)

***

93.867 (3.342) -10.985 (-2.837) 9.738 (2.708) 4.726 (0.452) -80.448 (-4.227) 26.629 (1.804) 60.492 -3.907

*** *** ***

*** * ***

** * 1.080 (2.205)

1.107 ** 1.128 (2.383) (2.433) -0.126 -0.123 (-1.544) (-1.502) 0.175 ** 0.186 *** (2.450) (2.703) 0.231 *** 0.222 ** (2.708) (2.564) -1.978 *** -2.024 *** (-4.838) (-5.153) 0.333 * 0.302 * (1.947) (1.745) 0.22 0.187 (0.952) (0.835)

Yes

Yes

Yes

Yes

247 0.406 0.362 9.213

247 0.404 0.366 10.448

235 0.231 0.171 3.843

235 0.227 0.174 4.285

***

***

**

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-5: R III: Multiple Regression - Base Case (2003) I

295

Appendix

Dependent Variable Model No.

ROA (5)

(6)

Intercept

-6.214 (1.135)

-5.112 (-0.908)

MB

3.859 (0.934) 9.189 (2.842) -0.581 (-0.184)

SB FBM

***

MGMT BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

3.979 (1.317) 0.773 (0.392) 0.734 (1.141) 0.794 (2.234) -0.219 (-0.376) -11.533 (-3.766) 3.996 (1.605) 2.253 (1.597)

**

***

0.653 (0.320) 0.736 (1.151) 0.734 (2.036) -0.188 (-0.304) -11.0274 (-3.501) 3.970 (1.582) 2.4 (1.684)

Yes

Yes

246 0.234 0.186 4.301

246 0.221 0.170 4.338

***

**

***

*

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-6: R III: Multiple Regression - Base Case (2003) II

296

Appendix

Dependent Variable Model No.

BAHR (7)

(8)

(9a)

Intercept

-99.631 (-1.488)

-109.954 (-1.547)

-149.507 (-1.900)

*

MGMT

-77.025 (-1.104) 188.738 (2.056)

548.477 (2.359)

**

83.594 (3.346) -11.307 (-2.850) 11.224 (2.912) 2.448 (0.285) -84.098 (-4.803) 27.635 (1.848) 57.904 (4.206)

***

MGMT SQ

**

MGMT 0to5

-92.188 (-0.227) -52.321 (-0.502) 138.756 (2.707)

MGMT 5to25 MGMT 25to100

***

MGMT AV BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

72.449 (2.981) -9.102 (-2.323) 7.849 (2.444) 4.714 (0.515) -80.980 (-4.556) 27.965 (1.869) 66.068 (4.169)

*** ** **

*** * ***

76.192 (2.922) -8.817 (-2.107) 8.312 (2.485) 5.233 (0.523) -82.698 (-4.596) 26.560 (1.808) 65.364 (3.946)

*** ** **

*** * ***

Yes

Yes

Yes

247 0.427 0.388 10.728

247 0.417 0.373 9.619

247 0.413 0.375 10.833

***

***

*** ***

*** * ***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-7: R III: Multiple Regression - Base Case Extended (2003) I

297

Appendix

Dependent Variable Model No.

BAHR (9b)

(10)

(10)

Intercept

-120.091 (1.533)

-73.145 (-1.125)

-69.909 (-1.082)

MGMT

101.220 (3.023) -17.013 (-2.135)

MGMT NO

*** **

MGMT EUR

-0.004 (1.335)

MGMT CRmCF BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

58.170 (1.030) 82.979 (2.954) -9.267 (-2.243) 8.805 (2.482) 6.156 (0.589) -83.212 (-4.407) 25.790 (1.776) 62.384 (4.018)

*** ** **

*** * ***

53.585 (2.893) -15.622 (-3.292) 7.705 (2.373) 4.922 (0.413) -79.858 (-3.853) 29.212 (1.964) 61.269 (3.537)

*** *** **

*** ** ***

57.892 (3.080) -14.681 (-3.138) 6.965 (2.292) 5.109 (0.425) -79.369 (-3.758) 29.184 (2.002) 59.184 (3.343)

Yes

Yes

Yes

247 0.413 0.372 10.111

247 0.373 0.331 9.123

247 0.374 0.334 9.208

***

***

*** *** **

*** ** ***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-8: R III: Multiple Regression - Base Case Extended (2003) II

298

Appendix

Dependent Variable Model No. Intercept MB SB FBM

BAHR (12) -117.284 (-2.182) 56.734 (1.554) 70.968 (2.579) 75.251 (2.259)

(13) **

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

32.341 (1.839) 3.994 (0.826) 11.192 (3.672) -3.239 (-0.500) -57.188 (-2.314) 33.828 (1.927) 41.967 (3.228)

-122.079 (-2.308)

**

**

*

***

** * ***

(14)

(15)

0.126 (0.780)

0.615 (0.403)

1.849 * (1.706) 1.282 * (1.658) -0.178 (-0.329)

***

MGMT BLOCK O

MTBV

69.304 (2.580)

***

33.046 (1.887) 4.072 (0.846) 11.524 (3.955) -3.605 (-0.567) -57.973 (-2.220) 33.495 (1.898) 41.592 (3.279)

*

***

** * ***

0.841 (1.299) 0.558 0.481 (1.255) (1.090) 0.064 0.058 (0.435) (0.404) 0.212 *** 0.184 ** (2.641) (2.461) 0.712 0.774 (1.167) (1.252) -1.967 *** -1.998 *** (-3.495) (-3.381) -0.178 -0.149 (-0.627) (-0.511) -0.277 -0.219 (-0.872) (-0.665)

Yes

Yes

Yes

Yes

220 0.386 0.334 7.457

220 0.384 0.339 8.490

212 0.211 0.142 3.054

212 0.180 0.117 2.684

***

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-9: R III: Multiple Regression - Base Case (1998) I

299

Appendix

Dependent Variable Model No.

ROA (16)

(17)

Intercept

-7.750 (-0.995)

-8.281 (-1.054)

MB

5.556 (1.695) 11.217 (1.947) 7.773 (2.210)

SB FBM

* * **

MGMT BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

8.630 (2.632) 0.691 (0.292) 0.059 (0.112) 0.483 (1.356) 2.518 (0.669) -3.305 (-0.956) 2.386 (1.401) 1.115 (0.656)

0.766 (0.321) 0.114 (0.221) 0.541 (1.488) 2.560 (0.639) -3.897 (-1.134) 2.287 (1.260) 1.114 (0.646)

Yes

Yes

215 0.159 0.086 2.183

215 0.149 0.085 2.323

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-10: R III: Multiple Regression - Base Case (1998) II

300

Appendix

Dependent Variable Model No.

BAHR (18)

Intercept

-297.691 (-3.439)

MB

27.584 (0.561) 16.791 (0.369) -23.864 (-0.498)

SB FBM

(19) ***

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

-282.907 (-3.407)

(20) ***

2.363 (-1.988)

(21) **

-1.636 (-0.043) -11.022 (-0.371) -2.166 (-324) 18.711 (3.871) 118.104 (2.064) -136.653 (-3.916) 55.234 (2.069) 83.540 (-4.117)

2.615 (2.254)

**

0.734 (0.778) 0.661 (0.989) -0.081 (-0.105)

MGMT BLOCK O

MTBV

-14.306 (-0.491) -2.965 (-0.441) *** 17.883 (3.925) ** 115.632 (2.000) *** -136.452 (-3.823) ** 58.818 (2.154) *** 88.830 (-4.319)

0.304 (0.552)

*** ** *** ** ***

0.802 * (1.876) -0.183 (-1.644) 0.043 (0.627) 0.196 (0.271) -2.172 *** (-3.477) 0.535 (1.479) -0.771 ** (-2.155)

0.752 (1.792) -0.199 (-1.725) 0.030 (0.464) 0.159 (0.225) -2.190 (-3.668) 0.598 (1.739) -0.689 (-1.911)

Yes

Yes

Yes

Yes

133 0.379 0.288 4.135

133 0.374 0.294 4.665

132 0.286 0.180 2.691

132 0.28 0.187 3.007

***

***

***

* *

*** * *

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-11: R III: Multiple Regression - Base Case (1993) I

301

Appendix

Dependent Variable Model No.

ROA (22)

(23)

Intercept

-4.291 (-0.828)

-2.029 (-0.382)

MB

7.274 (2.317) -2.624 (-0.967) 1.604 (0.762)

SB FBM

**

MGMT BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

0.455 (0.219) -2.457 (-1.304) 0.063 (0.148) 0.421 (1.436) 1.547 (0.746) -5.887 (-2.246) 0.539 (0.275) 0.653 (0.434)

**

-3.05 (-1.598) -0.024 (-0.056) 0.241 (0.823) 0.905 (0.407) -5.177 (-1.868) 0.318 (0.174) 1.083 (0.732)

Yes

Yes

131 0.298 0.192 2.819

131 0.247 0.149 2.512

***

*

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-12: R III: Multiple Regression - Base Case (1993) II

302

Appendix

Dependent Variable Model No. Intercept MB SB FBM

BAHR (24)

(25)

-156.967 (-3.043)

***

49.357 (2.342) 66.855 (2.875) 53.560 (2.537)

**

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998 YEAR 1993

Industry Dummies n R2 R2 adj. F-Statistic

45.943 (2.873) -4.105 (-1.404) 11.554 (4.729) 4.330 (0.490) -92.056 (-6.284) 36.915 (3.266) 56.340 (-5.855) 2.664 (0.340) 51.050 (5.212)

-157.855 (-3.062)

***

**

***

***

*** *** ***

***

(26)

(27)

-0.214 (-0.276)

-0.044 (-0.059)

1.364 (2.422) 1.079 (3.020) 0.670 (1.570)

***

MGMT BLOCK O

MTBV

57.806 (3.082)

***

46.095 (2.871) -3.980 (-1.357) 11.701 (4.838) 4.239 (0.474) -95.907 (-6.342) 36.685 (3.195) 56.078 (-5.862) 2.698 (0.343) 51.169 (5.189)

***

***

*** *** ***

***

** ***

0.957 (2.906)

0.857 *** 0.826 *** (3.067) (2.986) -0.057 -0.062 (-0.876) (-0.965) 0.169 *** 0.157 *** (3.799) (3.783) 0.244 * 0.257 ** (1.954) (2.010) -1.985 *** -1.968 *** (-6.694) (-6.615) 0.196 0.221 (1.181) (1.314) -0.127 -0.090 (-0.745) (-0.555) 0.462 ** 0.462 ** (2.518) (2.509) 0.775 *** 0.773 *** (4.440) (4.449)

Yes

Yes

Yes

Yes

600 0.375 0.355 18.327

600 0.374 0.356 20.471

579 0.187 0.160 6.773

579 0.183 0.159 7.410

***

***

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-13: R III: Multiple Regression - Base Case (Pooled) I

303

Appendix

Dependent Variable Model No. Intercept MB SB FBM

ROA (28)

(29)

-156.967 (-3.043)

***

49.357 (2.342) 66.855 (2.875) 53.560 (2.537)

**

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998 YEAR 1993

Industry Dummies n R2 R2 adj. F-Statistic

45.943 (2.873) -4.105 (-1.404) 11.554 (4.729) 4.330 (0.490) -92.056 (-6.284) 36.915 (3.266) 56.340 (-5.855) 2.664 (0.340) 51.050 (5.212)

***

57.806 (3.082)

***

46.095 (2.871) -3.980 (-1.357) 11.701 (4.838) 4.239 (0.474) -95.907 (-6.342) 36.685 (3.195) 56.078 (-5.862) 2.698 (0.343) 51.169 (5.189)

***

*** **

MGMT BLOCK O

-157.855 (-3.062)

***

***

*** *** ***

***

Yes

Yes

600 0.375 0.355 18.327

600 0.374 0.356 20.471

***

***

*** *** ***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-14: R III: Multiple Regression - Base Case (Pooled) II

304

Appendix

Dep. Variable Model No.

BAHR (1993) (18b)

Intercept

-291.286 (-3.340)

MB

31.680 (0.653) 27.067 (0.606) -15.870 (-0.344)

SB FBM

(19b) ***

BLOCK NO LN ASSETS FIRM RISK BETA DEBT RATIO SALES G DIV

Ind. Dummies n R2 R2 adj. F-Statistic

-277.939 (-3.326)

(12b) ***

-119.931 (-2.084) 52.564 (1.562) 66.935 (2.207) 70.788 (2.209)

MGMT BLOCK O

BAHR (1998) (13b) **

-1.410 (-0.044) -4.762 (-0.660) *** 15.789 (3.045) * 101.932 (1.861) 25.837 (0.964) *** -135.228 (-3.876) ** 60.752 (2.114) *** 91.970 (4.268)

*** *

*** ** ***

27.977 (1.443) 3.907 (0.829) 11.966 (3.078) -1.832 (-0.239) -8.383 (-0.338) -56.968 (-2.278) 34.364 (1.911) 42.296 (3.269)

**

65.017 (2.423)

**

** **

6.993 (0.190) 1.325 (0.041) -4.002 (-0.556) 16.524 (2.973) 104.167 (1.918) 25.543 (0.940) -134.996 (-3.962) 57.044 (2.038) 86.957 (4.059)

-124.698 (-2.206)

***

** * ***

28.607 (1.466) 3.986 (0.849) 12.308 (3.336) -2.168 (-0.283) -8.511 (-0.341) -57.778 (-2.190) 34.042 (1.884) 41.936 (3.317)

Yes

Yes

Yes

Yes

133 0.384 0.287 3.945

133 0.379 0.293 4.423

220 0.386 0.331 7.029

220 0.385 0.337 7.945

***

***

***

***

** * ***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-15: R III: Multiple Regression - Base Case (Beta) I

305

Appendix

Dep. Variable Model No. Intercept MB SB FBM

BAHR (2003) (1b) -146.667 (-1.958) 42.942 (1.561) 62.346 (1.857) 54.582 (1.723)

(2b) *

BLOCK NO LN ASSETS FIRM RISK BETA DEBT RATIO SALES G DIV

62.021 (2.508) -11.396 (-2.890) 12.711 (3.387) 12.508 (1.108) -39.613 (-3.124) -93.406 (-5.551) 23.559 (1.683) 57.952 (4.148)

-148.980 (-1.978)

(24b) **

*

** *** ***

*** *** * ***

54.478 (2.019)

**

62.062 (2.508) -11.265 (-2.850) 13.003 (3.418) 12.225 (1.077) -40.029 (-3.078) -93.968 (-5.412) 22.271 (1.579) 57.011 (4.126)

** *** ***

*** ***

***

YEAR 1998 YEAR 1993

Ind. Dummies n R2 R2 adj. F-Statistic

-162.107 (-3.173) 38.030 (1.955) 54.864 (2.469) 41.980 (2.141)

*

MGMT BLOCK O

BAHR (Pooled)

32.733 (2.145) -3.942 (-1.349) 13.129 (4.738) 8.397 (0.865) -19.422 (-1.747) -96.708 (-7.130) 36.610 (3.146) 55.030 (6.037) 6.192 (0.744) 59.592 (5.156)

(25b) ***

-162.973 (-3.193)

***

46.040 (2.679)

***

32.785 (2.148) -3.821 (-1.304) 13.280 (4.827) 8.336 (0.848) -19.555 (-1.748) -97.556 (-7.204) 36.389 (3.081) 54.773 (6.058) 6.250 (0.748) 59.765 (5.131)

**

* ** **

**

***

* *** *** ***

***

Yes

Yes

Yes

Yes

247 0.435 0.391 9.757

247 0.434 0.394 11.015

600 0.380 0.359 17.742

600 0.379 0.360 19.707

***

***

***

***

* *** *** ***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-16: R III: Multiple Regression - Base Case (Beta) II

306

Dependent Variable Model No.

Appendix

BAHR

MTBV

ROA

(30)

(31)

(32)

Intercept

-53.157 (-1.012)

0.247 (0.275)

-1.580 (-0.360)

MGMT CHANGE

81.532 (2.765)

***

0.459 (1.233)

-0.296 (-0.109)

BLOCK O

70.627 (3.518) -14.537 (-3.317) 6.444 (2.384) 4.996 (0.505) -82.660 (-4.695) 29.328 (2.134) 60.549 (-4.116)

***

0.707 (1.845) -0.173 (-2.032) 0.145 (2.136) 0.227 (2.572) -2.024 (-5.161) 0.328 (2.004) 0.194 (0.860)

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

*** **

*** ** ***

* ** ** ** *** **

-1.286 (-0.712) 0.533 (0.862) 0.599 * (1.853) -0.172 (-0.259) -10.874 *** (-3.427) 4.082 (1.633) 2.455 * (1.686)

Yes

Yes

Yes

247 0.420 0.382 11.149

235 0.214 0.161 3.983

246 0.215 0.164 4.202

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-17: R III: Multiple Regression - Dynamic Ownership (2003)

307

Appendix

Dependent Variable Model No.

BAHR

MTBV

ROA

(33)

(34)

(35)

Intercept

-66.094 (-1.430)

0.115 (0.088)

-5.994 (-0.783)

MGMT CHANGE

-15.529 (-0.845)

-0.148 (-0.313)

4.408 (1.331)

BLOCK O

-4.353 (-0.299) 1.221 (0.270) 10.109 (3.604) 4.247 (0.448) -95.465 (-3.701) 9.431 (0.726) 46.113 (3.872)

0.336 (1.013) -0.120 (-1.161) 0.234 (3.740) 1.513 (2.062) -2.509 (-4.062) -0.214 (-0.701) 0.101 (0.290)

-2.087 (-1.001) -0.804 (-1.109) 0.441 (1.131) 7.801 (1.368) -5.518 (-1.425) 1.579 (0.889) 3.068 (1.630)

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

***

***

***

*** ** **

Yes

Yes

Yes

204 0.395 0.347 8.187

195 0.232 0.168 3.614

198 0.181 0.114 2.690

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-18: R III: Multiple Regression - Dynamic Ownership (1998)

308

Dependent Variable Model No.

Appendix

BAHR

MTBV

ROA

(36)

(37)

(38)

Intercept

-59.397 (-1.384)

0.618 (0.811)

-1.205 (-0.308)

MGMT CHANGE

45.599 (2.253)

**

0.349 (1.165)

0.743 (0.317)

BLOCK O

36.534 (2.577) -7.458 (-2.172) 7.797 (3.691) 3.963 (0.406) -82.268 (-5.384) 24.280 (2.514) 55.191 (4.896) 2.108 (0.265)

***

0.475 (1.829) -0.127 (-1.981) 0.155 (3.282) 0.329 (2.326) -2.050 (-6.097) 0.126 (0.534) 0.000 (0.002) 0.499 (2.571)

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998

Industry Dummies n R2 R2 adj. F-Statistic

** ***

*** ** ***

* ** *** ** ***

**

-2.492 (-1.864) 0.042 (0.097) 0.464 (1.889) 0.219 (0.311) -8.987 (-3.632) 2.904 (2.067) 1.759 (1.554) 1.236 (1.171)

Yes

Yes

Yes

451 0.375 0.352 16.245

430 0.189 0.157 6.010

444 0.137 0.105 4.232

***

***

*

*

*** **

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table D-19: R III: Multiple Regression - Dynamic Ownership (Pooled)

Appendix

E

Results III: Insider Ownership and Corporate Performance - Robustness

309

310

Appendix

Dependent Variable Model No.

Q (39)

Intercept

1.206 (3.570)

MB

0.248 (1.097) 0.124 (0.661) 0.097 (0.560)

SB FBM

(40) ***

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

1.237 (3.751)

***

(41)

(42)

3.065 (0.336)

3.816 (0.397)

3.752 (0.695) 6.984 (1.407) 0.630 (0.126)

MGMT BLOCK O

ROE

0.142 (0.896) 0.170 (1.109) -0.015 (-0.539) -0.016 (-0.765) 0.095 ** (2.364) -0.975 *** (-7.020) 0.496 *** (5.276) 0.183 ** (2.207)

0.166 (1.081) -0.016 (-0.587) -0.018 (-0.948) 0.098 (2.421) -0.967 (-7.112) 0.507 (5.287) 0.193 (2.302)

4.117 (0.944)

** *** *** **

6.492 * (1.808) -1.216 (-1.095) 0.642 (1.342) 3.968 *** (2.776) -8.313 (-1.420) 0.077 (0.032) 0.403 (0.139)

6.535 * (1.783) -1.195 (-1.049) 0.606 (1.208) 4.018 *** (2.796) -8.103 (-1.346) 0.005 (0.002) 0.539 (0.197)

Yes

Yes

Yes

Yes

248 0.359 0.312 7.577

248 0.357 0.316 8.605

170 0.169 0.076 1.815

170 0.163 0.082 2.004

***

***

**

**

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table E-1: R III: Multiple Regression - APM (2003)

311

Appendix

Dependent Variable Model No.

Q (43)

Intercept

2.089 (4.679)

MB

-0.000 (-0.001) 0.269 (1.178) 0.034 (0.132)

SB FBM

(44) ***

BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

2.078 (4.729)

***

(45)

(46)

1.220 (0.104)

0.853 (0.075)

9.430 (1.124) 2.606 (0.428) 10.299 (1.252)

MGMT BLOCK O

ROE

0.132 (0.654) -0.265 * (-1.683) 0.051 (1.131) -0.082 *** (-3.234) -0.027 (-0.401) -0.718 *** (-3.577) 0.084 (1.024) 0.090 (0.987)

6.888 (1.078)

-0.262 * (-1.649) 0.054 (1.205) -0.080 *** (-3.242) -0.025 (-0.412) -0.752 *** (-3.623) 0.083 (1.016) 0.090 (0.984)

6.760 (1.364) -1.247 (-1.196) 1.020 (1.518) 6.258 (1.601) -2.143 (-0.284) 6.015 * (1.945) -0.736 (-0.176)

6.758 (1.367) -1.297 (-1.260) 0.976 (1.521) 6.588 * (1.664) -1.263 (-0.164) 6.001 * (1.913) -0.601 (-0.140)

Yes

Yes

Yes

Yes

223 0.222 0.157 3.433

223 0.215 0.158 3.787

172 0.154 0.060 1.644

172 0.144 0.061 1.746

***

***

*

**

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table E-2: R III: Multiple Regression - APM (1998)

312

Appendix

Dependent Variable Model No.

Q (47)

(48)

Intercept

2.281 (5.892)

***

MB

0.824 (2.373) 0.119 (0.575) 0.055 (0.361)

**

SB FBM MGMT BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV

Industry Dummies n R2 R2 adj. F-Statistic

ROE

2.540 (6.390)

***

(49)

(50)

-3.169 (-0.262)

-1.614 (-0.123)

11.063 (2.049) 3.694 (0.733) 9.226 (0.876)

**

0.162 (1.014) -0.018 (-0.142) 0.043 (1.243) -0.084 *** (-3.415) -0.244 (-1.471) -0.928 *** (-5.039) 0.079 (0.642) -0.162 ** (-2.110)

7.099 (1.404)

-0.083 (-0.622) 0.031 (0.858) -0.102 *** (-4.177) -0.302 * (-1.812) -0.881 *** (-4.666) 0.095 (0.744) -0.092 (-1.185)

0.822 (0.236) -0.213 (-0.243) 1.039 (1.510) 18.986 (1.610) -27.207 *** (-3.688) 24.176 *** (4.934) -2.547 (-0.551)

0.368 (0.102) -0.164 (-0.173) 0.921 (1.199) 18.124 (1.454) -27.409 *** (-3.550) 24.590 *** (5.183) -3.184 (0.666)

Yes

Yes

Yes

Yes

133 0.342 0.245 3.513

133 0.299 0.209 3.329

99 0.448 0.332 3.862

99 0.441 0.34 4.359

***

***

***

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. The calculations of the t-statistics are based on the SPSS macro written by Hayes and Cai (2005). *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table E-3: R III: Multiple Regression - APM (1993)

313

Appendix

Dependent Variable Model No.

Q (51)

(52)

Intercept

1.700 (6.973)

***

MB

0.241 (1.645) 0.169 (1.297) 0.072 (0.592)

*

SB FBM MGMT BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998 YEAR 1993

Industry Dummies n R2 R2 adj. F-Statistic

ROE

1.739 (7.302)

***

(53)

(54)

2.121 (0.343)

1.983 (0.317)

5.976 (1.386) 3.813 (1.157) 6.694 (1.290) 0.145 (1.388)

-0.064 (-0.682) 0.020 (0.962) -0.050 *** (-3.561) 0.047 (1.394) -0.898 *** (-8.842) 0.223 * (1.892) 0.123 ** (2.304) -0.064 (-1.094) -0.064 (-1.128)

5.344 (1.578)

-0.070 (-0.751) 0.018 (0.893) -0.053 *** (-3.934) 0.051 (1.495) -0.890 *** (-8.838) 0.230 * (1.899) 0.131 ** (2.460) -0.064 (-1.090) -0.064 (-1.128)

5.939 (2.351) -1.297 (-2.066) 0.774 (2.125) 4.398 (3.733) -7.609 (-1.885) 4.317 (1.325) -0.417 (-0.176) 3.350 (2.258) -1.572 (-0.960)

** ** ** *** *

**

5.919 (2.353) -1.304 (-2.051) 0.769 (2.118) 4.373 (3.842) -7.474 (-1.836) 4.358 (1.331) -0.471 (-0.200) 3.326 (2.212) -1.615 (-0.985)

Yes

Yes

Yes

Yes

604 0.239 0.214 9.636

604 0.237 0.215 10.705

441 0.137 0.098 3.508

441 0.135 0.101 3.898

***

***

***

** ** ** *** *

**

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table E-4: R III: Multiple Regression - APM (Pooled)

314

Dependent Variable

Appendix

BAHR

BAHR

Model No.

(55) Complete

Intercept

-157.599 (-3.299)

***

-167.437 (-2.136)

**

MGMT

57.938 (2.525)

**

66.156 (1.887)

*

BLOCK O BLOCK NO LN ASSETS FIRM RISK DEBT RATIO SALES G DIV YEAR 1998 YEAR 1993

Industry Dummies n R2 R2 adj. F-Statistic

30.113 (1.625) -6.448 (-1.769) 13.217 (4.929) 18.339 (1.763) -147.260 (-6.387) 26.423 (2.185) 50.480 (5.264) -4.001 (-0.409) 43.021 (3.715)

(56) Sub-sample

* *** * *** ** ***

***

18.770 (0.800) -4.661 (-0.855) 14.328 (2.742) 13.154 (0.965) -162.421 (-4.567) 19.518 (2.605) 59.651 (4.029) -9.145 (-0.720) 36.833 (2.505)

Yes

Yes

258 0.373 0.329 8.406

144 0.371 0.292 4.682

***

***

*** *** ***

**

***

OLS-regression coefficients. Heteroskedasticity robust White (1980) t-statistics reported in parentheses. *, ** and *** indicate significance on the 0.10, 0.05 and 0.01 level (2-tailed). “Complete” (model 55) refers to 86 companies for which complete data sets were available at all three sub-samples. “Sub-sample” (model 56) refers to 41 companies for which MGMT did not changed during all three sub-samples. For a definition of the variables see tables 6-5 to 6-7 and section 6.3. Source: Author’s findings.

Table E-5: R III: Multiple Regression - Endogeneity (Pooled)

Exemplary Ownership Filings in the U.S. and Germany

Figure F-1: SEC Filing of Beneficial Ownership (U.S.)

Source: Securities and Exchange Commission (SEC) (2006).

Figure represents an exemplary extract.

F

Appendix 315

Figure F-2: BaFin Database of Directors’ Dealings (Germany)

Source: Bundesanstalt f¨ur Finanzdienstleistungsaufsicht (BaFin) (2006).

Figure represents an exemplary extract.

316 Appendix

References

317

References Achleitner, A.-K. and Bassen, A. (2002): Entwicklungsstand des Shareholder Value-Ansatzes in Deutschland — Empirische Befunde. In: H. Siegwart, J. Mahari and M. Ruffner, eds., Meilensteine des Managements, vol. IX: Corporate Governance, Shareholder Value and Finance. Basel et al., pp. 611–635. Achleitner, A.-K., Kaserer, C. and Moldenhauer, B. (2005): German Entrepreneurial Index (GEX) – Ein Style Index zur Performance eigent¨umergef¨uhrter Unternehmen. Finanz-Betrieb, vol. 7 (2): pp. 118–126. Achleitner, A.-K. and Kl¨ockner, O. (2005): Employment contribution of private equity and venture capital in Europe. EVCA Research Paper. Achleitner, A.-K. and Moldenhauer, B. (2006): Potential des GEX wird noch nicht ausgesch¨opft — Ergebnisse einer CEFS-Befragung anl¨asslich des einj¨ahrigen Bestehens des neuen Aktienindex. GoingPublic, vol. 7: pp. 78–79. Achleitner, A.-K. and Thommen, J.-P. (2003): Allgemeine Betriebswirtschaftslehre - Umfassende Einf¨uhrung aus managementorientierter Sicht. Wiesbaden, 4th edn. Adams, M. (1999): Cross holdings in Germany. Journal of Institutional and Theoretical Economics, vol. 155 (1): pp. 80–109. Aggarwal, R. K. and Samwick, A. A. (2006): Empire-builders and shirkers: Investment, firm performance, and managerial incentives. Journal of Corporate Finance, vol. 12 (3): pp. 489–515. Aglietta, M. and Breton, R. (2003): Financial systems and modes of corporate control. In: R. Cobbaut and J. Lenoble, eds., Corporate Governance — An Institutionalist Approach. Kluwer Law International, The Hague, pp. 117–138. Agrawal, A. and Knoeber, C. R. (1996): Firm performance and mechanisms to control agency problems between managers and shareholders. Journal of Financial and Quantitative Analysis, vol. 31 (3): pp. 377–197. Agrawal, A. and Mandelker, G. N. (1990): Large shareholders and the monitoring of managers: The case of antitakeover charter amendments. Journal of Financial and Quantitative Analysis, vol. 25 (2): pp. 143–161. Akerlof, G. (1970): The market for lemons. The Quarterly Journal of Economics, vol. 84 (3): pp. 488–500.

318

References Albers, M. (2002): Corporate Governance in Aktiengesellschaften: Entscheidungsprozesse und Wirkungsanalyse zum Gesetz zur Kontrolle und Transparenz im Unternehmensbereich (KonTraG), vol. 2844 of Europ¨aische Hochschulschriften: Reihe 5, Volks- und Betriebswirtschaft. Peter Lang — Europ¨aischer Verlag der Wissenschaften, Frankfurt am Main. Alchian, A. A. (1950): Uncertainty, evolution, and economic theory. Journal of Political Economy, vol. 58 (3): pp. 211–221. Alchian, A. A. and Demsetz, H. (1972): Production, information costs, and economic organization. The American Economic Review, vol. 62 (5): pp. 777–795. Alich, H., Dohmen, C. and Sommer, U. (2002): DAX rutscht auf tiefsten Stand seit f¨unfeinhalb Jahren. Handelsblatt 19.19.2002, (181): p. 1. Allen, F. and Gale, D. (2000): Corporate governance and competition. In: X. Vives, ed., Corporate Governance — Theoretical and Empirical Perspectives. Cambridge University Press, Cambridge et al., pp. 23–94. Allianz (2006): FAQ — Europ¨aische Gesellschaft: Allianz SE. URL http://www.allianz.com/azcom/dp/cda/0,,1099900-49,00.html, (viewed: 06.07.2006). Amihud, Y. (2002): Illiquidity and stock returns: Cross-section and time-series effects. Journal of Financial Markets, vol. 5 (1): pp. 31– 56. Anderson, R. C. and Lee, D. S. (1997): Ownership studies: The data sources does matter. Journal of Financial and Quantitative Analysis, vol. 32 (3): pp. 311–329. Anderson, R. C., Mansi, S. A. and Reeb, D. M. (2003): Founding family ownership and the agency cost of debt. Journal of Financial Economics, vol. 68 (2): pp. 263–285. Anderson, R. C. and Reeb, D. M. (2003): Founding-family ownership and firm performance: Evidence from the S&P 500. The Journal of Finance, vol. 58 (3): pp. 1301–1327. Andreani, E. (2003): Corporate control and the financial system in Germany: Recent changes in the role of banks. Universit¨at Rostock Th¨unen Series of Applied Economic Theory - Working Paper No. 37. Andres, C. (2006): Family ownership as the optimal organizational structure? Working Paper. Ang, J. S., Cole, R. A. and Lin, J. W. (2000): Agency costs and ownership structure. The Journal of Finance, vol. 55 (1): pp. 81–106.

References

319

Astrachan, J. H., Klein, S. B. and Smyrnios, K. X. (2002): The F-PEC scale of family influence: A proposal for solving the family business definition problem. Family Business Review, vol. 15 (1): pp. 45–58. Backhaus, K., Erichson, B., Plinke, W. and Weiber, R. (2003): Multivariate Analysemethoden. Springer, Berlin et al., 10th edn. Barber, B. and Lyon, J. (1997): Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of Financial Economics, vol. 43 (3): pp. 341–372. Barclay, M. J. and Holderness, C. G. (1989): Private benefits from control of public corporations. Journal of Financial Economics, vol. 25 (2): pp. 371–395. Barclay, M. J. and Holderness, C. G. (1991): Negotiated block trades and corporate control. The Journal of Finance, vol. 46 (3): pp. 861– 878. Barnea, A., Haugen, R. A. and Senbet, L. W. (1985): Agency Problems and Financial Contracting. Prentice-Hall, Englewood-Cliffs. Barnhart, S. W., Marr, M. W. and Rosenstein, S. (1994): Firm performance and board composition: Some new evidence. Managerial and Decision Economics, vol. 15 (4): pp. 329–340. Barontini, R. and Caprio, L. (2005): The effect of ownership structure and family control on firm value and performance. Evidence from Continental Europe. European Corporate Governance Institute (ECGI) - Finance Working Paper No. 88/2005. Barth, E., Gulbrandsen, T. and Schøne, P. (2005): Family ownership and productivity: The role of owner-management. Journal of Corporate Finance, vol. 11 (1-2): pp. 107–127. Bassen, A. (2002): Institutionelle Investoren und Corporate Governance. Deutscher Universit¨ats-Verlag, Wiesbaden. Bassen, A., Kleinschmidt, M., Prigge, S. and Z¨ollner, C. (2005): Corporate Governance und Unternehmenserfolg — Empirische Befunde zur Wirkung des Deutschen Corporate Governance Kodex. Working Paper. Baums, T. (1993): Macht von Banken und Versicherungen — Stellungnahme f¨ur den Wirtschaftsausschuß des Deutschen Bundestages. Working Paper No. 10/1993. Baums, T. (2002): Requests to legislators for forward looking corporate governance. In: G. Cromme, ed., Corporate Governance Report 2002 — Vortr¨age und Diskussionen der 1. Konferenz Deutscher Corporate Governance Kodex. Sch¨affer-Poeschel Verlag, Stuttgart, pp. 131–139.

320

References Baums, T. (2003): Podium discussion on “corporate governance in Europe” — Statement. In: G. Cromme, ed., Corporate Governance Report 2003 — Vortr¨age und Diskussionen der 2. Konferenz Deutscher Corporate Governance Kodex. Sch¨affer-Poeschel Verlag, Stuttgart, pp. 174–176. Baums, T. and Fraune, C. (1995): Institutionelle Anleger und Publikumsgesellschaft: Eine empirische Untersuchung. Die Aktiengesellschaft, vol. 40 (3): pp. 97–112. Bebchuk, L. A. (1999): A rent-protection theory of corporate ownership and control. National Bureau of Economic Research - Working Paper. Becht, M. (1997): Strong blockholders, weak owners and the need for European mandatory disclosure. European Corporate Governance Network - Executive Report. Becht, M. and B¨ohmer, E. (1999): Transparency of ownership and control in Germany. Working Paper. Becht, M. and B¨ohmer, E. (2003): Voting control in German corporations. International Review of Law and Economics, vol. 23 (1): pp. 1–29. Beiner, S., Drobetz, W., Schmid, M. M. and Zimmermann, H. (2004): Corporate Governance und Unternehmensbewertung in der Schweiz. Working Paper. Beiner, S., Drobetz, W., Schmid, M. M. and Zimmermann, H. (2006): An integrated framework of corporate governance and firm valuation. European Financial Management, vol. 12 (2): pp. 249–283. Belkaoui, A. and Pavlik, E. (1992): The effects of ownership structure and diversification strategy on performance. Managerial and Decision Economics, vol. 13: pp. 343–352. Ben-Amar, W. and Andr´e, P. (2006): Separation of ownership from control and acquiring firm performance: The case of family ownership in Canada. Journal of Business Finance & Accounting, vol. 33 (3): pp. 517–543. Benders, R. and Drost, F. (2006): Sparkassen wollen Depotstimmrecht str¨arken. Handelsblatt 24./25./26.02.2006, (40): p. 23. Benston, G. J. (1985): The self-serving management hypothesis: Some evidence. Journal of Accounting & Economics, vol. 7 (1-3): pp. 67– 84. Berger, P. G., Ofek, E. and Yermack, D. L. (1997): Managerial entrenchment and capital structure decisions. The Journal of Finance, vol. 52 (4): pp. 1411–1438. Berle, A. A. (1965): The impact of the corporation on classical economic theory. The Quarterly Journal of Economics, vol. 79 (1): pp. 25–40.

References

321

Berle, A. A. and Means, G. C. (1932): The Modern Corporation and Private Property. Harcourt, Brace and World, New York. Berndt, M. (2002): Global Differences in Corporate Governance Systems — Theory and Implications for Reforms. Gabler Edition Wissenschaft. Deutscher Universit¨ats-Verlag. Berrar, C. (2001): Die Entwicklung der Corporate Governance in Deutschland im internationalen Vergleich. No. 174 in Wirtschaftsrecht und Wirtschaftspolitik. Nomos Verlagsgesellschaft, BadenBaden. Berry, T. K., Fields, L. P. and Wilkins, M. S. (2006): The interaction among multiple governance mechanisms in young newly public firms. Journal of Corporate Finance, vol. 12 (3): pp. 449–466. Besley, D. A., Kuh, E. and Welsch, R. E. (1980): Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. John Wiley, New York. Bhagat, S., Black, B. S. and Blair, M. (2003): Relational investing and firm performance. Journal of Financial Research, vol. 27 (1): pp. 1–30. Bhagat, S. and Jefferis, Jr., R. H. (2002): The Econometrics of Corporate Governance Studies. The MIT Press, Cambridge, London. Bischoff, J. (1994): Das Shareholder-Value Konzept — Darstellung Probleme - Handlungsm¨oglichkeiten. Deutscher Universit¨ats-Verlag, Wiesbaden. Black, B. S. (1992): Institutional investors and corporate governance: The case for institutional voice. Journal of Applied Corporate Finance, vol. 5: pp. 19–32. Black, B. S., Jang, H. and Kim, W. (2006): Predicting firms’ corporate governance choices: Evidence from Korea. Journal of Corporate Finance, vol. 12 (3): pp. 660–691. Blum, A. (2005): Corporate Governance — Ein Vergleich zwischen den gesetzlichen Regelungen in Deutschland und den Forderungen des Kapitalmarkts. Josef Eul Verlag, Lohmar. Bodie, Z., Kane, A. and Marcus, A. J. (1999): Investments. Irwin McGraw-Hill, Boston et al., 4th edn. B¨ohmer, E. (2001): Germany. In: K. Gugler, ed., Corporate Governance and Economic Performance. Oxford University Press, pp. 96–120. B¨ohmer, E. and Ljungqvist, A. (2004): On the decision to go public: Evidence from privately-held firms. Discussion Paper 16/2004, Deutsche Bundesbank, Frankfurt am Main. Bøhren, Ø. and Ødegaard, B. A. (2001): Corporate governance and economic performance in Norwegian listed firms. Working Paper.

322

References Bøhren, Ø. and Ødegaard, B. A. (2004): Governance and performance revisted. European Corporate Governance Institute (ECGI) - Finance Working Paper No. 28/2003. Bolton, P. and Thadden, E.-L. v. (1998): Liquidity and control: A dynamic theory of corporate ownership structure. Journal of Institutional and Theoretical Economics, vol. 154 (1): pp. 177–211. Borokhovich, K. A., Parrino, R. and Trapani, T. (1996): Outside directors and CEO selection. Journal of Financial and Quantitative Analysis, vol. 31: pp. 337–355. B¨orsch-Supan, A. (1998): Capital’s contribution to productivity and the nature of competition. Brookings Papers on Economic Activity: pp. 205–244. B¨orsch-Supan, A. and K¨oke, J. (2002): An applied econometricians’ view of empirical corporate governance studies. German Economic Review, vol. 3 (3): pp. 295–326. Bott, C. (2002): Aktion¨arsstruktur, Kontrolle und Erfolg von Unternehmen. Gabler Edition Wissenschaft. Deutscher Universit¨atsVerlag. Bottazzi, L. and Da Rin, M. (2003): Financing entrepreneurial firms in Europe: Facts, issues, and research agenda. CESifo - Working Paper No. 958. Brealey, R. A. and Myers, S. C. (2000): Principles of Corporate Finance. Irwin McGraw-Hill, 6th edn. Brick, I. E., Palmon, O. and Wald, J. K. (2006): CEO compensation, director compensation, and firm performance. Journal of Corporate Finance, vol. 12 (3): pp. 403–423. Brickley, J. A., Lease, R. C. and Smith, C. W. J. (1988): Ownership structure and voting on antitakeover amendments. Journal of Financial Economics, vol. 20: pp. 267–291. Brink, G. J. v. d. and Romeike, F. (2005): Corporate Governance und Risikomanagement im Finanzdienstleistungsbereich. Sch¨afferPoeschel Verlag, Stuttgart. Brown, G. and Kapadia, N. (2005): Firm-specific risk and equity market development. Working Paper. ¨ Buhler, P., Fueglistaller, U. and Zellweger, T. (2005): Schweizer Familienunternehmen an der B¨orse. Working Paper. ¨ Finanzdienstleistungsaufsicht (BaFin) (2004): Bundesanstalt fur Jahresbericht 2003. Bonn, Frankfurt am Main. ¨ Finanzdienstleistungsaufsicht (BaFin) (2006): Bundesanstalt fur Datenbank — Meldungen nach § 15a WpHG (Directors’ Deal-

References

323

ings). URL http://www.bafin.de/datenbanken/p15a.html, (viewed: 30.07.2006). Bundesministerium der Finanzen (2005): Beteiligungsbericht 2005. URL http://www.bundesfinanzministerium.de/lang de/DE/Service/ Downloads/Abt VIII/Beteiligungsbericht 202005,templateId=raw, property=publicationFile.pdf, (viewed: 27.07.2006). Bundesministerium der Finanzen (2006): Geld und Kredit. URL http: //www.bundesfinanzministerium.de, (viewed: 20.05.2006). Bundesregierung (2004): Corporate Governance — Unternehmensf¨uhrung - Unternehmenskontrolle - Modernisierung des Aktienrechts. URL http://www.bundesregierung.de/Politikthemen/ Steuern-und-Finanzen/,7124/Was-ist-Corporate-Governance.htm, (viewed: 01.08.2005). Bundesverband Deutscher Kapitalbeteiligungsgesellschaften (BVK) (2003): BVK Statistik 2003. Burkart, M., Gromb, D. and Panunzi, F. (1997): Large shareholders, monitoring, and the value of the firm. The Quarterly Journal of Economics, vol. 112 (3): pp. 693–728. Burkart, M., Gromb, D. and Panunzi, F. (2000): Agency conflicts in public and negotiated transfers of corporate control. The Journal of Finance, vol. 55 (2): pp. 647–677. Burkart, M., Panunzi, F. and Shleifer, A. (2003): Family firms. The Journal of Finance, vol. 58 (5): pp. 2167–2201. Burrough, J. B. and Helgar, J. S. (1990): Barbarians at the Gate. Random House, London. Busse, C. (2005): Bye-bye Deutschland AG — Kehraus bei den Finanzkonzernen. Handelsblatt 13.01.2005, (9): p. 2. Buz, D. A. (1994): How do large minority shareholders wield control? Managerial and Decision Economics, vol. 15 (4): pp. 291–298. BVI Bundesverband Investment und Asset Management e.V. (2005): BVI-Gesamtstatistik (Stichtag: 30.06.2005). URL http://www.bvi. de/downloads/gfs05q2.pdf, (viewed: 30.09.2005). Carlin, W. and Mayer, C. (2000): How do financial systems affect economic performance? In: X. Vives, ed., Corporate Governance — Theoretical and Empirical Perspectives. Cambridge University Press, Cambridge et al., pp. 137–168. Chaplinski, S. and Niehaus, G. (1993): Do inside ownership and leverage share common determinants? Quarterly Journal of Business and Economics, vol. 32 (4): pp. 51–65.

324

References Chen, C. R., Guo, W. and Mande, V. (2003): Managerial ownership and firm valuation: Evidence from Japanese firms. Pacific-Basin Finance Journal, vol. 11 (3): pp. 267–283. Chen, H., Hexter, J. L. and Hu, M. Y. (1993): Management ownership and corporate value. Managerial and Decision Economics, vol. 14 (4): pp. 335–346. Chen, M.-Y. (2005): Group affiliation, identity of managers, and the relation between managerial ownership and performance. International Review of Financial Analysis, vol. 14 (5): pp. 533–558. Chen, Z., Cheung, Y.-L., Stouraitis, A. and Wong, A. W. S. (2005): Ownership concentration, firm performance, and dividend policy in Hong Kong. Pacific-Basin Finance Journal, vol. 13 (4): pp. 431–449. Cheung, Y.-L., Stouraitis, A. and Wong, A. W. (2005): Ownership concentration and executive compensation in closely held firms: Evidence from Hong Kong. Journal of Empirical Finance, vol. 11: pp. 511–532. Chiang, R. and Venkatesh (1988): Insider holdings and perceptions of information asymmetry: A note. The Journal of Finance, vol. 43 (4): pp. 1041–1048. Chirinko, R., van Ees, H., Garretsen, H. and Sterlen, E. (2004): Investor protections and concentrated ownership: Assessing corporate control mechanisms in the Netherlands. German Economic Review, vol. 5 (2): pp. 119–138. Cho, M.-H. (1998): Ownership structure, investment, and the corporate value: An empirical analysis. Journal of Financial Economics, vol. 47: pp. 103–121. Chrisman, J. J., Chua, J. and Sharma, P. (1996): A Review and Annotated Bibliography of Family Business Studies. Kluwer Academic Publishers, Boston, Dordrecht, London. Chrisman, J. J., Chua, J. H. and Litz, R. (2003): A unified systems perspective of family firm performance: An extension and integration. Journal of Business Venturing, vol. 18 (4): pp. 467–472. Chung, K. H. and Pruitt, S. W. (1994): A simple approximation of Tobin’s q. Financial Management, vol. 23 (3): pp. 70–74. Chung, K. H. and Pruitt, S. W. (1996): Executive ownership, corporate value, and executive compensation: A unifying framework. Journal of Banking & Finance, vol. 20 (7): pp. 1135–1159. Claassen, J. (2003): Summary of the discussion “Corporate governance in Europe”. In: G. Cromme, ed., Corporate Governance Report 2003 — Vortr¨age und Diskussionen der 2. Konferenz Deutscher Corporate Governance Kodex. Cromme, Gerhard, Stuttgart, pp. 177–181.

References

325

Claessens, S., Djankov, S., Fan, J. P. H. and Lang, L. H. P. (2002): Disentangling the incentive and entrenchment effects of large shareholdings. The Journal of Finance, vol. 57 (6): pp. 2741–2771. Clement, W. (2004): Corporate Governance — Bedeutung f¨ur den Wirtschaftsstandort Deutschland. Speech of the Federal Minister of Economics and Labor at the conference “Corporate Government Kodex” on 23.06.2004, Berlin. Coase, R. H. (1937): The nature of the firm. Economica, vol. 4 (16): pp. 386–405. Coase, R. H. (1960): The problem of social cost. Journal of Law & Economics, vol. 3: pp. 1–44. Coase, R. H. (1992): The institutional structure of production. American Economic Review, vol. 82 (4): pp. 713–719. Coase, R. H. (1998): The new institutional economics. American Economic Review, vol. 88 (2): pp. 72–74. Coffee, J. C. (2001): The rise of dispersed ownership: The role of law in the separation of ownership and control. The Yale Law Journal, vol. 111 (1): pp. 1–80. Coffee, J. C. and Berle, A. A. (1999): The future as history: The prospects for global convergence in corporate governance and its implications. Northwestern University Law Review, vol. 93: pp. 641– 707. Cole, R. A. and Mehran, A. (1998): The effect of changes in ownership structure on performance: Evidence from the thrift industry. Journal of Financial Economics, vol. 50 (3): pp. 291–317. Coles, J. L., Daniel, N. D. and Naveen, L. (2006): Managerial incentives and risk-taking. Journal of Financial Economics, vol. 79: pp. 431–468. Coles, J. L., Lemmon, M. L. and Meschke, J. F. (2003): Structural models and endogenity in corporate finance: The link between managerial ownership and corporate performance. EFA 2003 Annual Conference Paper No. 693. Commission of the European Communities (2002): SMEs in Europe, including a first glance at EU Candidate Countries. URL http://ec.europa.eu/enterprise/enterprise policy/analysis/ doc/smes observatory 2002 report2 en.pdf, (viewed: 27.07.2006), Observatory of European SMEs No. 2/2002. Commission of the European Communities (2003a): Access to finance of small and medium-sized enterprises. URL http://ec.europa.eu/ enterprise/entrepreneurship/financing/docs/com(2003)713 en.pdf,

326

References (viewed: 20.09.2005), Communication from the Comission to the Council and the European Parliament. Commission of the European Communities (2003b): Modernisierung des Gesellschaftsrechts und Verbesserung der Corporate Governance in der Europ¨aischen Union — Aktionsplan. URL http://eur-lex.europa.eu/LexUriServ/site/de/com/2003/ com2003 0284de01.pdf, (viewed: 27.07.2006). Commission of the European Communities (2004): Commission recommendation on fostering an appropriate regime for the remuneration of directors of listed companies. URL http://www.europa.eu.int/comm/internal market/company/ docs/director-remun/2004-recommendation de.pdf, (viewed: 20.09.2005). Cornell, B. and Shapiro, A. C. (1987): Corporate stakeholders and corporate finance. Financial Management, vol. 16 (1): pp. 5–14. Corstjens, M., Peyer, U. and Van der Heyden, L. (2005): Stock market performance of family firms. Working Paper. Coto, D. G. (2002): Share ownership structure in Europe. Economics & Statistics Subcommittee, Federation of European Securities Exchanges (FESE). Cromme, G. (2002a): Aktuelle Fragen der Corporate Governance. Zentrum f¨ur Europ¨aisches Wirtschaftsrecht - Vortr¨age und Berichte No. 133. Cromme, G. (2002b): Corporate Governance in Germany — Aspiration and Reality. In: G. Cromme, ed., Corporate Governance Report 2002 — Vortr¨age und Diskussionen der 1. Konferenz Deutscher Corporate Governance Kodex. Sch¨affer-Poeschel Verlag, Stuttgart, pp. 115–129. Cromme, G. (2003): Development of corporate governance in Germany. In: G. Cromme, ed., Corporate Governance Report 2003 — Vortr¨age und Diskussionen der 2. Konferenz Deutscher Corporate Governance Kodex. Sch¨affer-Poeschel Verlag, Stuttgart, pp. 139–154. Crutchley, C. E. and Hansen, R. S. (1989): A test of the agency theory of managerial ownership, corporate leverage, and corporate dividends. Financial Management, vol. 18 (4): pp. 36–46. Cubbin, J. and Leech, D. (1983): The effect of shareholding dispersion on the degree of control in Britisch companies: Theory and evidence. The Economic Journal, vol. 93 (370): pp. 351–369. Cui, H. and Mak, Y. T. (2002): The relationship between managerial ownership and firm performance in high R&D firms. Journal of Corporate Finance, vol. 8 (4): pp. 313–336.

References

327

Dann, L. Y. and DeAngelo, H. (1988): Corporate financial policy and corporate control: A study of defensive adjustments in asset and ownership structure. Journal of Financial Economics, vol. 20: pp. 87– 127. Daouk, H., Lee, C. M. C. and Ng, D. (2006): Capital market governance: How do security laws affect market performance. Journal of Corporate Finance, vol. 12 (3): pp. 560–593. Daske, S. and Ehrhardt, O. (2000): Der Kurs- und Renditeunterschied von Stamm- und Vorzugsaktien: Eine Untersuchung am deutschen Kapitalmarkt. Working Paper. Davies, J. R., Hillier, D. and McColgan, P. (2005): Ownership structure, managerial behavior and corporate value. Journal of Corporate Finance, vol. 11 (4): pp. 645–660. Davis, J. H., Schoorman, F. D. and Donaldson, L. (1997): Toward a stewardship theory of management. Academy of Management Review, vol. 22 (1): pp. 20–47. De Miguel, A., Pindado, J. and De la Torre, C. (2004): Ownership structure and firm value: New evidence from Spain. Strategic Management Journal, vol. 25: pp. 1199–1207. DeAngelo, H. and DeAngelo, L. (1985): Managerial ownership of voting rights: A study of public corporations with dual classes of common stock. Journal of Financial Economics, vol. 14: pp. 33–69. DeDalt, P. J., Donaldson, J. R. and Garner, J. L. (2003): Will any q do? The Journal of Financial Research, vol. 26 (4): pp. 535–551. Working Paper. Demsetz, H. (1983): The structure of ownership and the theory of the firm. Journal of Law & Economics, vol. 26 (2): pp. 375–390. Demsetz, H. (1986): Corporate control, insider trading, and rates of return. American Economic Review, vol. 76 (2): pp. 313–316. Demsetz, H. and Lehn, K. (1985): The structure of corporate ownership: Causes and consequences. Journal of Political Economy, vol. 93 (6): pp. 1155–1177. Demsetz, H. and Villalonga, B. (2001): Ownership structure and corporate performance. Journal of Corporate Finance, vol. 7 (3): pp. 209–233. Demsetz, R., Marc, S. and Philip, S. (1996): Banks with something to lose: The disciplinary role of franchise value. Economic Policy: pp. 1–14. Denis, D. J., Hanouna, P. and Sarin, A. (2006): Is there a dark side to incentive compensation. Journal of Corporate Finance, vol. 12 (3): pp. 467–488.

328

References Denis, D. K. and McConnell, J. J. (2003): International corporate governance. European Corporate Governance Institute (ECGI) - Finance Working Paper No. 05/2003. Deutsche Bank (2002): Governance in Deutschland – Perspektiven der Wissenschaft, Konferenzbericht 02-3. Deutsche Bank Research. Deutsche B¨orse (2005a): Guide to the German Entrepreneurial Index (GEX) — Version 1.0. URL http://deutsche-boerse. com/dbag/dispatch/en/binary/gdb content pool/imported files/ public files/10 downloads/50 informations services/30 Indices Index Licensing/21 guidelines/10 share indices/gex guide.pdf, (viewed: 21.04.2006). Deutsche B¨orse (2005b): Monatsums¨atze an deutschen Wertpapierb¨orsen im Juli 2005. URL http://deutsche-boerse.com/dbag/ dispatch/de/kir/gdb navigation/private investors/50 Statistics/ 21 Spot Market Order Book Statistics/50 Monthly Turnovers, (viewed: 31.10.2005). Deutsche Bundesbank (1997): Monatsbericht Januar 1997. Frankfurt am Main. Deutsche Bundesbank (2004a): Monatsbericht April 2004. Frankfurt am Main. Deutsche Bundesbank (2004b): Monatsbericht Juli 2004. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2000): Zahl der Aktion¨are in Deutschland u¨ berschreitet 5 Millionen, DAI-Kurzstudie 1/2000. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2001): Zahl der Aktion¨are und Besitzer von Aktienfondsanteilen steigt weiter an, DAI-Kurzstudie 3/2001. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2005a): Dax-Aktion¨arsstrukturen. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2005b): Verhalten und Pr¨aferenzen deutscher Aktion¨are, Studie des Deutschen Aktieninstituts, Heft 29. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2005c): Vertrauen in die Aktien w¨achst, DAI-Kurzstudie 3/2005. Frankfurt am Main. Deutsches Aktieninstitut e.V. (DAI) (2006): DAI-Factbook 2006. Frankfurt am Main. Dietl, H. M. (1998): Capital Markets and Corporate Governance in Japan, Germany and the United States — Organizational response to market inefficiencies. Routledge, London, New York.

References

329

Dlugosz, J., Fahlenbrach, R., Gompers, P. and Metrick, A. (2006): Large blocks of stock: Prevalence, size, and measurement. Journal of Corporate Finance, vol. 12 (3): pp. 594–618. Dostert, E. (2004): Die B¨orse liebt Familien — Viele Unternehmen der deutschen Clans schlagen den DAX. S¨uddeutsche Zeitung 30.06.2004: p. 21. Drobetz, W., Schillhofer, A. and Zimmermann, H. (2004): Corporate governance and expected stock returns: Evidence from Germany. European Financial Management, vol. 10 (2): pp. 267–293. Drost, F. M. (2005): Anlegersch¨utzer fordern Managerhaftung. Handelsblatt 16.10.2005, (199): p. 38. Drukarczyk, J. (1993): Theorie und Politik der Finanzierung. M¨unchen, 2nd edn. ¨ Drukarczyk, J., Honold, D. and Schuler, A. (1995): Shareholders and intensity of control. Regensburger Diskussionsbeitr¨age zur Wirtschaftswissenschaft No. 284. Dyck, A. and Zingales, L. (2004): Private benefits of control: An international comparison. The Journal of Finance, vol. 59 (2): pp. 537– 600. Easterbrook, F. H. and Fischel, D. R. (1982): Corporate control transactions. The Yale Law Journal, vol. 91 (4): pp. 698–737. Easterbrook, F. H. and Fischel, D. R. (1983): Voting in corporate law. Journal of Law & Economics, vol. 26 (2): pp. 395–427. Easterbrook, F. H. and Fischel, D. R. (1991): The Economic Structure of Corporate Law. Harvard University Press. Eckert, D. and Zsch¨apitz, H. (2005): Aktion¨arsstruktur wird zum Investmentkriterium. Die Welt 08.03.2005. ¨ Eckey, H.-F., Kosfeld, R. and Dreger, C. (2001): Okonometrie: Grundlagen - Methoden - Beispiele. Gabler, Wiesbaden, 2nd edn. Edwards, J. S. S. and Nibler, M. (2000): Corporate governance in Germany: The role of banks and ownership concentration. Economic Policy, vol. 15 (31): pp. 239–267. Edwards, J. S. S. and Weichenrieder, A. J. (2004): Ownership concentration and share valuation. German Economic Review, vol. 5 (2): pp. 143–171. Ehrhardt, O. and Nowak, E. (2003a): The effect of IPOs on German family-owned firms: Governance changes, ownership structure, and performance. Journal of Small Business Management, vol. 41 (2): pp. 222–232.

330

References Ehrhardt, O. and Nowak, E. (2003b): Private Kontrollrenten in deutschen Unternehmen. Die Betriebswirtschaft, vol. 63 (4): pp. 363–377. Ehrhardt, O., Nowak, E. and Weber, F.-M. (2006): Running in the family — The evolution of ownership, control, and performance in German family-owned firms 1903-2003. Swiss Finance Institute Research Paper No. 06-13. Elston, J. A. (2004): Bank influence, firm performance and survival: Empirical evidence from Germany 1970-1986. Corporate Ownership & Control, vol. 1 (2): pp. 65–70. Emmons, W. R. and Schmid, F. A. (1998): Universal banking, control rights, and corporate finance in Germany. Review July/August 1998, Federal Reserve Bank of St. Louis. ergo Kommunikation (2005): Corporate Governance Survey 2005. K¨oln, Frankfurt am Main. Ernst, C. (2001): Die Aktion¨arsstruktur deutscher und amerikanischer Unternehmen — Ein internationaler Vergleich. Reihe Wirtschaft und Kommunikation. Verlag Ulmer Manuskripte. Faccio, M. and Lang, H. P. L. (2002): The ultimate ownership of Western European corporations. Journal of Financial Economics, vol. 65 (3): pp. 365–395. Faccio, M. and Lasfer, M. A. (1999): Managerial ownership, board structure and firm value: The UK evidence. Working Paper. Fama, E. F. (1965): The behaviour of stock market prices. Journal of Business, vol. 38 (1): pp. 34–105. Fama, E. F. (1980): Agency problems and the theory of the firm. Journal of Political Economy, vol. 88 (2): pp. 288–307. Fama, E. F., Fisher, L., Jensen, M. C. and Roll, R. (1969): The adjustment of stock prices to new information. International Economic Review, vol. 10 (1): pp. 1–21. Fama, E. F. and Jensen, M. C. (1983a): Agency problems and residual claims. Journal of Law & Economics, vol. 26: pp. 327–349. Fama, E. F. and Jensen, M. C. (1983b): Separation of ownership and control. Journal of Law & Economics, vol. 26: pp. 304–325. Fama, E. F. and MacBeth, J. (1973): Risk, return, and equilibrium: Empirical tests. Journal of Political Economy, vol. 81: pp. 607–636. Farinha, J. (2003): Corproate governance: A survey of the literature. Working Paper. Felton, R. F. (2002): A new era in corporate governance. The McKinsey Quarterly, (2): pp. 28–41.

References

331

Felton, R. F. and Watson, M. (2002): Change across the board. The McKinsey Quarterly, (4): pp. 31–45. Filbeck, G. and Chadwell, P. A. (1993): Insider ownership and dividend signals: Additional evidence from the banking industry. Journal of Economics and Finance, vol. 17 (2): pp. 109–117. Fingerle, C. H. (2005): Smart Money — Influence of Venture Capitalists on High Potential Companies, vol. 7 of Entrepreneurial and Financial Studies. Verlag Wissenschaft & Praxis, Sternenfels. Fisher, I. (1930): The theory of interest — As determined by impatience to spend income and opportunity to invest it. Macmillan, New York. Fockenbrock, D. (2005a): Corporate Governance auf dem Pr¨ufstand. Handelsblatt 07.03.2005, (46): p. 14. Fockenbrock, D. (2005b): Transparenz: Unbefriedigend. Handelsblatt 05.04.2005, (65): p. 2. Fockenbrock, D. (2005c): Zweite Karriere als Kontrolleur. Handelsblatt 23.06.2005, (119): p. 16. Fockenbrock, D. (2006): Konzerne umgehen Cromme-Kodex. Handelsblatt 03.01.2006, (2): p. 1. Fockenbrock, D. and Stratmann, K. (2006): Netzwerk der Deutschland AG d¨unnt aus. Handelsblatt 06.07.2006, (128): p. 14. Franke, D. (2004): B¨orse: Zwischen Delisting und neuer Hoffnung. Die Bank, (2): pp. 136–139. Franks, J. and Mayer, C. (1990): Capital markets and corporate control: A study of France, Germany and the UK. Economic Policy, vol. 5 (10): pp. 189–231. Franks, J. and Mayer, C. (1997): Corporate ownership and control in U.K., Germany, and France. Journal of Applied Corporate Finance, vol. 9 (4): pp. 30–45. Franks, J. and Mayer, C. (2001): Ownership and control of German corporations. The Review of Financial Studies, vol. 14 (4): pp. 943– 977. Friend, I. and Lang, L. H. P. (1988): An empirical test of the impact of managerial self-interest on corporate capital structure. The Journal of Finance, vol. 43 (2): pp. 271–281. Furubotn, E. G. and Pejovich, S. (1972): Property rights and economic theory: A survey of recent literature. Journal of Economic Literature, vol. 10 (4): pp. 1137–1162. Gillan, S. L. (2006): Recent developments in corporate governance: An overview. Journal of Corporate Finance, vol. 12 (3): pp. 381–402.

332

References Gillan, S. L. and Starks, L. T. (2003): Corporate governance, corporate ownership, and the role of institutional investors: A global perspective. Journal of Applied Finance, vol. 13 (2): pp. 14–22. Gillmann, B. (2004): Streit um Offenlegung von Top-Geh¨altern. Handelsblatt 20.09.2004, (182): p. 3. Gleisberg, R. (2003): B¨orsengang und Unternehmensentwicklung — der Einfluss ver¨anderter Eigentums- und Kontrollstrukturen, vol. 4 of GoingPublic Forschungsreihe. Going Public Media AG, Wolfratshausen. G¨obel, S. and Kormaier, B. (2006): Bilanzierung des Aktienr¨uckkaufs im internationalen Vergleich. Praxis der internationalen Rechnungslegung (PiR), vol. 2 (5): pp. 65–70. G¨ocmen, V. and Meyer, T. (2004): Mehr Wert dank Eigent¨umerUnternehmen — Grossaktion¨are bringen “ihre” Unternehmen voran. Neue Z¨uricher Zeitung 29.06.2004, (148): p. 25. Goergen, M. (1998): Insider retention and long-run performance in German and UK IPOs. School of Management, UMIST (Manchester) Working Paper. Goergen, M., Manjon, M. C. and Renneboog, L. D. R. (2004): Recent developments in German corporate governance. Working Paper No. 123/2004. Gompers, P., Ishii, J. and Metrick, A. (2003): Corporate governance and equity prices. The Quarterly Journal of Economics, vol. 118 (1): pp. 107–155. Gompers, P., Ishii, J. and Metrick, A. (2004): Incentives vs. control: An analysis of U.S. dual-class companies. National Bureau of Economic Research - Working Paper. Gorton, G. and Schmid, F. (2002): Class struggle inside the firm: A study of German codetermination. The Federal Reserve Bank of St. Louis - Working Paper 2000-025B. Gorton, G. and Schmid, F. A. (2000): Universal banking and the performance of German firms. Journal of Financial Economics, vol. 58 (1– 2): pp. 29–80. Graham, J. and Harvey, C. (2001): The theory and practice of corporate finance: Evidence from the field. Journal of Financial Economics, vol. 60: pp. 187–243. Gravelle, H. and Rees, R. (1984): Microeconomics. Modern Economics. Longman, New York. Greene, W. H. (2003): Econometric Analysis. Prentice Hall, Upper Saddle River, 5th edn.

References

333

Gr¨oneweg, S. (2004): Verwandtschaft an der B¨orse. S¨uddeutsche Zeitung 02.07.2004, (150): p. 22. Grossman, S. J. and Hart, O. D. (1980): Takeover bids, the free-rider problem, and the theory of the corporation. The Bell Journal of Economics, vol. 11 (1): pp. 42–64. Grossman, S. J. and Hart, O. D. (1986): The costs and benefits of ownership: A theory of vertical and lateral integration. Journal of Political Economy, vol. 94: pp. 691–719. Grossman, S. J. and Hart, O. D. (1988): One share-one vote and the market for corporate control. Journal of Empirical Finance, vol. 20 (1–2): pp. 175–202. Gugler, K., Mueller, D. C. and Yurtoglu, B. B. (2004): The effects of insider ownership on investment performance: An international comparison. Working Paper. Gugler, K. and Weigand, J. (2002): Is ownership really endogenous? Applied Economic Letters, vol. 10 (8): pp. 483–486. Guti´errez, M. and Trib´o, J. A. (2004): Private benefits extraction in closely-held corporations: The case for multiple large shareholders. European Corporate Governance Institute - Working Paper No. 52/2004. Habbershon, T. G., Williams, M. and MacMillan, I. C. (2003): A unified systems perspective of family firm performance. Journal of Business Venturing, vol. 18 (4): pp. 451–465. Habib, M. A. and Ljungqvist, A. (2003): Firm value and managerial incentives: A stochastic frontier approach. Wharton Financial Institutions Center - Working Paper. Hajek, S. (2004): Kritische Stimme — Dominante Großaktion¨are sind besser als ihr Ruf. Wirtschaftswoche, (17): pp. 114–117. Haller, M. J., Jaskiewicz, P., May, P. and Schiereck, D. (2004): Bedeutung von nicht aktiven Gesellschaftern in Familienunternehmen. EBS - Working Paper. Halpern, P., Kieschnick, R. and Rotenberg, W. (2005): Managerial shareholdings, firm value, and acquired corporations. The Quarterly Review of Economics and Finance, vol. 45 (4-5): pp. 781–795. Harijono, M. A. and Tanewksi, G. A. (2004): The impact of family control of firms on leverage: Australian evidence. Working Paper. Harjoto, M. and Garen, J. (2005): Inside ownership beyond the IPO: The evolution of corporate ownership concentration. Journal of Corporate Finance, vol. 11 (4): pp. 661–679. Harris, A. B. and Kramer, A. S. (2002): Corporate governance: PreEnron, Post-Enron. Working Paper.

334

References Harris, M. and Raviv, A. (1988a): Corporate control contests and capital structure. Journal of Financial Economics, vol. 20: pp. 55–86. Harris, M. and Raviv, A. (1988b): Corporate governance: Voting rights and majority rules. Journal of Financial Economics, vol. 20: pp. 203– 235. Hart, O. D. (1980): Perfect competition and optimal product differentiation. Journal of Economic Theory, vol. 22 (2): pp. 2279–312. Hart, O. D. (1983): The market mechanism as an incentive scheme. The Bell Journal of Economics, vol. 14 (2): pp. 366–382. Hart, O. D. and Moore, J. (1990): Property rights and the nature of the firm. Journal of Political Economy, vol. 98: pp. 1119–1158. Harvey, C. R., Lins, K. V. and Roper, A. H. (2004): The effect of capital structure when expected agency costs are extreme. Journal of Financial Economics, vol. 74 (1): pp. 3–30. Hasler, P.-T. (2004): Die Performance familiengef¨uhrter Unternehmen. Equity Research, HypoVereinsbank (HVB). Hayes, A. F. and Cai, L. (2005): Using heteroskedasticity-consistent standard error estimators in OLS regression: An introduction and software implementation. Working Paper. Heaney, R., Naughton, T., Truong, T., Davidson, S., Fry, T. and McKenzie, M. (2006): The link between performance and changes in the size and stability of a firm’s officers and directors. Journal of Multinational Financial Management. Forthcomming. Helmis, S. (2002): Corporate Governance in Deutschland — Eigentumsund Kontrollstrukturen und rechtliche Rahmenbedingungen in der “Deutschland AG”. Institute for Mergers & Acquisitions (IMA) Working Paper No. 03/2002. Helmis, S. (2003): Fluch oder Segen f¨ur den deutschen Kapitalmarkt: Eine empirische Analyse der Squeeze-outs in deutschen Aktiengesellschaften. Institute for Mergers & Acquisitions (IMA) - Working Paper. Henkel (2006): Corporate Governance Henkel. URL http: //www.henkel.de/int henkel/ir de/binarydata/pdf/Corporate Governance Modell Henkel 2006 Stand 04 de.pdf, (viewed: 27.05.2006). Hermalin, B. E. and Weisbach, M. S. (1991): The effects of board composition and direct incentives on firm performance. Financial Management, vol. 20 (4): pp. 101–112. Hillier, D. and McColgan, P. (2005): Firm performance, entrenchment and managerial succession in family firms. Working Paper.

References

335

Himmelberg, C. P., Hubbard, R. G. and Palia, D. (1999): Understanding the determinants of managerial ownership and the link between ownership and performance. Journal of Financial Economics, vol. 53 (3): pp. 353–384. Hirschman, A. O. (1970): Exit Voice and Loyalty: Responses to Decline in Firms, Organizations, and States. Harvard University Press. Hocking, R. R. (1983): Developments in linear regression methodology: 1959-1982. Technometrics, vol. 25 (3): pp. 219–230. Holderness, C. G., Kroszner, R. S. and Sheehan, D. P. (1999): Were the good old days that good? Changes in managerial stock ownership since the great depression. The Journal of Finance, vol. 54 (2): pp. 435–469. Holderness, C. G. and Sheehan, D. P. (1985): Raiders or saviors? The evidence on six controversial investors. Journal of Financial Economics, vol. 14 (4): pp. 555–580. Holderness, C. G. and Sheehan, D. P. (1988): The role of majority shareholders in publicly held corporations. Journal of Financial Economics, vol. 20: pp. 317–346. Holl, P. (1975): Effect of control type on the performance of the firm in the U.K. The Journal of Industrial Economics, vol. 23 (4): pp. 257–271. Hommel, U. and Schneider, H. (2003): Financing the German Mittelstand. EIP Papers, vol. 8 (2): pp. 52–90. H¨opner, M. and Jackson, G. (2001): An emerging market for corporate control? — The Mannesmann takeover and German corporate governance. MPIfG - Discussion Paper No. 04/2000. H¨opner, M. and Krempel, L. (2005): Ein Netzwerk in Aufl¨osung: Wie die Deutschland AG zerf¨allt. In: MPIfG-Jahrbuch 2003/04. MaxPlanck-Institut f¨ur Gesellschaftsforschung, K¨oln, pp. 9–14. Hopt, K. J. and Leyens, P. C. (2004): Board models in Europe. Recent developments of internal corporate governance structures in Germany, the United Kingdom, France, and Italy. European Corporate Governance Institute (ECGI) - Working Paper. Howe, K. M., Vogt, S. and He, J. (2003): The effect of managerial ownership on the short- and long-run response to cash distributions. The Financial Review, vol. 38: pp. 179–196. Hubbard, R. G. and Palia, D. (1995): Benefits of control, managerial ownership, and the stock returns of acquiring firms. The RAND Journal of Economics, vol. 26 (4): pp. 782–792. Hussla, G. A. (2005a): Abwehrstrategien schaden Anlegern. Handelsblatt 11.10.2005, (196): p. 37.

336

References Hussla, G. A. (2005b): Aktion¨are proben den Aufstand. Handelsblatt 06.04.2005, (66): p. 31. Hussla, G. A. (2006): Konzerne aktivieren Aktion¨are. Handelsblatt 11.05.2006, (91): p. 25. Iber, B. (1987): Entwicklung der Aktion¨arsstruktur b¨orsennotierter deutscher Aktiengesellschaften — Eine theoretische und empirische Analyse f¨ur den Zeitraum 1963-1983. Kieler Schriften zur Finanzwirtschaft. Wissenschaftsverlag Vauk Kiel. ¨ Familienunternehmen INTES — Komission Governance Kodex fur (2004): Governance Kodex f¨ur Familienunternehmen. URL http://www.intes-online.de/kodex/Governance Kodex.pdf, (viewed: 13.04.2005). Januszewski, S. I., K¨oke, F. J. and Winter, J. K. (1999): Product market competition, corporate governance and firm performance: An empirical analysis for Germany. Discussion Paper No. 63/1999. Jaskiewicz, P., Gonzal´ez, V. M., Men´endez, S. and Schiereck, D. (2005): Long-run IPO performance analysis of German and Spanish family-owned businesses. Family Business Review, vol. 18 (3): pp. 179–202. Jenkinson, T. and Ljungqvist (2001): The role of hostile stakes in German corporate governance. Journal of Corporate Finance, vol. 7: pp. 397–446. Jensen, G. R., Solberg, D. P. and Zorn, T. S. (1992): Simultaneous determination of insider ownership, debt and dividend policies. Journal of Financial and Quantitative Analysis, vol. 27 (2): pp. 247–263. Jensen, M. C. (1986): Agency costs of free cash flow, corporate finance, and takeovers. The American Economic Review, vol. 76 (2): pp. 323– 329. Jensen, M. C. (1989): The eclipse of the public corporation. Harvard Business Review, vol. 67: pp. 61–70. Jensen, M. C. (1993): The modern industrial revolution, exit, and the failure of internal control systems. The Journal of Finance, vol. 48 (3): pp. 831–880. Jensen, M. C. (2000): A Theory of the Firm: Governance, Residual Claims, and Organizational Forms. Havard University Press, Cambridge, London. Jensen, M. C. and Meckling, W. H. (1976): Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, vol. 3 (4): pp. 305–360.

References

337

Jensen, M. C. and Meckling, W. H. (1979): Rights and production functions: An application to labor-managed firms and codetermination. Journal of Business, vol. 52: pp. 469–506. Jensen, M. C. and Ruback, R. S. (1983): The market for corporate control: The scientific evidence. Journal of Financial Economics, vol. 11: pp. 5–50. Jensen, M. C. and Warner, J. B. (1988): The distribution of power among corporate managers, shareholders, and directors. Journal of Financial Economics, vol. 20: pp. 3–24. Jostarndt, P. (2006): Financial distress, corporate control, and management turnover — A German panel analysis. Working Paper. ¨ Jurgens, U., Rupp, J. and Vitols, K. (2000): Corporate Governance and Shareholder Value in Deutschland. Ver¨offentlichungsreihe der Abteilung Regulierung und Arbeit des Forschungsschwerpunktes Technik-Arbeit-Umwelt des Wissenschaftszentrum Berlin f¨ur Sozialforschung - FS II 00-202. Kahn, C. and Winton, A. (1998): Ownership structure, speculation, and shareholder intervention. The Journal of Finance, vol. 53 (1): pp. 99– 129. Kamp, L. and Krieger, A. (2005): Die Aktivit¨aten von Finanzinvestoren in Deutschland — Hintergr¨unde und Orientierungen. Hans-B¨ockler Stiftung - Working Paper. Kang, D. L. and Sorensen, A. B. (1999): Ownership organization and firm performance. Annual Review of Sociology, vol. 25: pp. 121–44. Karpoff, J. M. (1999): Corporate finance and corporate control: An empirical research agenda. Journal of Institutional and Theoretical Economics, vol. 155 (1): pp. 239–246. Kaserer, C. (2001): Corporate Governance in der Schweiz — Eine o¨ konomisch orientierte Bestandsaufnahme. In: R. Schauer and G. Aschinger, eds., Probleme der Finanzm¨arkte: W¨ahrungs- und Finanzkriesen / Aspekte der Unternehmensfinanzierung. Trauner Universit¨ats Verlag, pp. 175–196. Kaserer, C. (2005): Wie die Chefl¨ohne zu kontrollieren sind. Der Bund 21.06.2005: p. 29. Interview of Christoph Kaserer by Hansueli Sch¨ochli. Kaserer, C., Achleitner, A.-K. and Moldenhauer, B. (2005): Entrepreneurial Firms: Einfluss der Eigent¨umerstuktur auf Unternehmensperformance und Kapitalmarktf¨ahigkeit. BetriebsBerater, vol. 60 (5): pp. 2–8. BB-Spezial.

338

References Kaserer, C., Achleitner, A.-K., Moldenhauer, B. and Ampenberger, M. (2006): Ein Jahr “German Entrepreneurial Index (GEX)” - Bestandsaufnahme und Analyse. Finanz-Betrieb, vol. 8 (1): pp. 10–17. Kaserer, C. and Moldenhauer, B. (2005): Insider ownership and corporate performance: Evidence from Germany. CEFS - Working Paper No. 01/2005. Kaserer, C. and Wagner, N. (2004): Determinanten der Vorstandsverg¨utung in Deutschland: Paradigmenwechsel oder Versagen in der Unternehmenskontrolle? CEFS - Working Paper No. 06/2004. Kemper, O. (2001): Going Privates in Deutschland. Institute for Mergers & Acquisitions (IMA) - Arbeitsbericht No. 02/2001. Kewes, T. and Herz, C. (2006): Endlich Nummer eins. Handelsblatt 17.05.2006, (95): p. 15. Kim, W. S. and Sorensen, E. H. (1986): Evidence on the impact of the agency costs of debt on corporate debt policy. Journal of Financial and Quantitative Analysis, vol. 21 (2): pp. 131–144. Klein, B. (1983): Contracting costs and residual claims: The separation of ownership and control. Journal of Law & Economics, vol. 26: pp. 367–374. Klein, S. B., Astrachan, J. H. and Smyrnios, K. X. (2005a): The FPEC scale of family influence: Construction, implication for theory. Entrepreneurship Theory and Practice, vol. 29 (3): pp. 321–339. Klein, S. B., Pieper, T. M. and Jaskiewicz, P. (2005b): Antecedents for agency and stewardship orientation in corporate governance — The role of culture. Research proposal submitted for consideration for Family Enterprise Research Conference (FERC). Koenen, J. (2004): DAX-Vorst¨ande f¨urchten Imageverlust — Fast die H¨alfte der deutschen Manager spricht sich f¨ur einen individuellen Ausweis der Geh¨alter aus. Handelsblatt 19.08.2004, (160): p. 2. K¨oke, J. (1999): New evidence on ownership structures in Germany. ZEW Discussion Paper No. 60/1999. K¨oke, J. (2002): Corporate Governance in Germany — An Empirical Investigation, vol. 17 of ZEW Economics Studies. Physica-Verlag. K¨oke, J. (2004): The market for corporate control in a bank-based economy: A governance device? Journal of Corporate Finance, vol. 10 (1): pp. 53–80. Kole, S. R. (1995): Measuring managerial equity ownership: A comparison of sources on ownership data. Journal of Corporate Finance, vol. 1 (3-4): pp. 413–435. Kr¨amer, H.-J. (2002): Das Unternehmensinteresse als Verhaltensmaxime der Leitungsorgane einer Aktiengesellschaft im Rahmen

References

339

der Organhaftung — In Abgrenzung zum Gesellschaftsinteresse und unter Ber¨ucksichtigung US-amerikanischer Literatur, vol. 18 of Juristische Reihe TENEA. Berlin. Kretschmann, A. (1976): Der Einfluss der Aktion¨arsstruktur auf die Kursbildung. Krivogorsky, V. (2006): Ownership, board stucture, and performance in continental Europe. The International Journal of Accounting, vol. 41 (2): pp. 176–197. ¨ Krusselberg, H.-G. (1983): Property Rights-Theorie und Wohlfahrts¨okonomik. In: A. Sch¨uller, ed., Property Rights und o¨ konomische Theorie. Vahlen, M¨unchen, pp. 45–78. Kuklinski, J., Lowinski, F. and Schiereck, D. (2003): Familienunternehmen als Underperformer: Zur B¨orsenbewertung mittelst¨andischer Unternehmen. In: J.-A. Meyer, ed., Unternehmensbewertung und Basel II in kleinen und mittleren Unternehmen — Jahrbuch der KMU-Forschung und Praxis. Josef Eul Verlag, Lohmar, pp. 103–120. La Porta, R., Lopez-de Silanes, F. and Shleifer, A. (1999): Corporate ownership around the world. The Journal of Finance, vol. 54 (2): pp. 471–517. La Porta, R., Lopez-de Silanes, F., Shleifer, A. and Vishny, R. W. (1998): Law and finance. Journal of Political Economy, vol. 106 (6): pp. 1113–1155. La Porta, R., Lopez-de Silanes, F., Shleifer, A. and Vishny, R. W. (2000): Investor protection and corporate governance. Journal of Financial Economics, vol. 58 (1): pp. 1–25. La Porta, R., Lopez-de Silanes, F., Shleifer, A. and Vishny, R. W. (2002): Investor protection and corporate valuation. The Journal of Finance, vol. 57 (3): pp. 1147–1170. Laatz, W. (1993): Empirische Methoden — Ein Lehrbuch f¨ur Sozialwissenschaftler. Verlag Harri Deutsch, Thun, Frankfurt am Main, 7th edn. Leech, D. and Leahy, J. (1991): Ownership structure, control type classifications and the performance of large British companies. The Economic Journal, vol. 101 (2): pp. 1418–1437. Lehmann, E. and Weigand, J. (2000): Does the governed corporation perform better? Governance structures and corporate performance in Germany. European Finance Review, vol. 4 (2): pp. 157–195. Leland, H. E. and Pyle, D. H. (1977): Informational asymmetries, financial structure, and financial intermediation. The Journal of Finance, vol. 32 (2): pp. 371–387.

340

References Lenoble, J. (2003): From an incentive to a reflexive approach to corporate governance. In: R. Cobbaut and J. Lenoble, eds., Corporate Governance — An Institutionalist Approach. Kluwer Law International, The Hague, pp. 17–63. Lewellen, W. G. (1969): Management and ownership in the large firm. The Journal of Finance, vol. 24 (2): pp. 299–322. Lins, K. V. (2003): Equity ownership and firm value in emerging markets. Journal of Financial and Quantitative Analysis, vol. 38 (1): pp. 159–184. Lipton, M. and Lorsch, J. W. (1992): A modest proposal for improved corporate governance. Business Lawyer, vol. 48 (1): pp. 59–77. Loderer, C. and Martin, K. (1997): Executive stock ownership and performance: Tracking faint traces. Journal of Financial Economics, vol. 45 (2): pp. 223–255. Luo, S., Courtenay, S. and Hossain, M. (2006): The effect of voluntary disclosure, ownership structure and proprietary cost on the returnfuture earnings relation. Pacific-Basin Finance Journal. Forthcoming. Lyon, J., Barber, B. and Chih-Ling, T. (1999): Improved methods for tests of long-run abnormal stock returns. The Journal of Finance, vol. 54 (1): pp. 165–201. Mahoney, W. F. (1993): The Active Shareholder. John Wiley & Sons, Inc., New York. Maisch, M. and Teschler, D. (2005): Großes Theater — Inszenierte Aktion¨arsdemokratie — Eine Rezension zur Hauptversammlung der Deutschen Bank. Handelsblatt 19.05.2005, (95): p. 12. Mann, A. (2003): Corporate Governance Systeme — Funktion und Entwicklung am Beispiel von Deutschland und Großbritannien, vol. 46 of Schriften zur wirtschaftswissenschaftlichen Analyse des Rechts. Duncker & Humbolt, Berlin. Mathiesen, H. (2005): Empirical studies on ownership structure and performance — Document version: 2.3. URL http://www.encycogov. com/A5OwnershipStructures/OwPerfStudies/Table Ow AB.asp, (viewed: 20.08.2005), encycogov.com — The encyclopedia about Corporate Governance. Matthes, J. (2000): Das deutsche Corporate-Governance-System: Wandel von Stakeholder-Orientierung zum Shareholder-Value-Denken, vol. 259 of Beitr¨age zur Wirtschafts- und Sozialpolitik. Deutscher Instituts Verlag, K¨oln.

References

341

Maug, E. (1998): Large shareholders as monitors: Is there a trade-off between liquidity and control? The Journal of Finance, vol. 53 (1): pp. 65–98. Maug, E. (2001): Ownership structure and the life-cycle of the firm: A theory of the decision to go public. European Finance Review, vol. 5 (3): pp. 167–200. Maury, B. (2006): Family ownership and firm performance: Empirical evidence from Western European corporations. Journal of Corporate Finance, vol. 12 (2): pp. 321–341. Mayer, C. (1996): Corporate governance, competition and performance. OECD No. 27. Mayer, C. (1998): Financial systems and corporate governance: A review of the international evidence. Journal of Institutional and Theoretical Economics, vol. 154 (1): pp. 144–165. McConaughy, D. L., Matthews, C. H. and Fialko, A. S. (2001): Founding family controlled firms: Performance, risk, and value. Journal of Small Business Management, vol. 39 (1): pp. 31–49. McConaughy, D. L., Walker, M. C., Henderson, Jr., G. V. and Mishra, C. S. (1998): Founding family controlled firms: Efficiency and value. Review of Financial Economics, vol. 7 (1): pp. 1–19. McConnell, J. J. and Servaes, H. (1990): Additional evidence on equity ownership and corporate value. Journal of Financial Economics, vol. 27 (2): pp. 595–612. McConnell, J. J. and Servaes, H. (1995): Equity ownership and the two faces of debt. Journal of Financial Economics, vol. 39 (1): pp. 131–157. Means, G. C. (1931): The separation of ownership and control in American industry. The Quarterly Journal of Economics, vol. 46 (1): pp. 68–100. Mehran, H. (1992): Executive incentive plans, corporate control, and capital structure. Journal of Financial and Quantitative Analysis, vol. 27 (4): pp. 539–560. Mehran, H. (1995): Executive compensation structure, ownership, and firm performance. Journal of Financial Economics, vol. 38: pp. 163– 184. Mello, A. S. and Parsons, J. E. (1998): Going public and the ownership structure of the firm. Journal of Financial Economics, vol. 49 (1): pp. 79–109. Mikkelson, W. H. and Partch, M. M. (1989): Managers’ voting rights and corporate control. Journal of Financial Economics, vol. 25: pp. 236–290.

342

References Milbourn, T. T. (1998): Financial systems and corporate governance: A review of the international evidence — Comment. Journal of Institutional and Theoretical Economics, vol. 154 (1): pp. 170–176. Mishra, C. S., Randøy, T. and Jenssen, J. I. (2001): The effect of founding family influence on firm value and corporate governance. Journal of International Financial Management and Accounting, vol. 12 (3): pp. 235–260. Moldenhauer, B. (2003): Strukturierungen von Privatisierungen. European Business School (ebs), University Schloß Reichartshausen - Unpublished diploma thesis. Monopolkommission (1998): Ordungspolitische Leitlinien f¨ur ein funktionsf¨ahiges Finanzsystem, Sondergutachten 26. Baden-Baden. Monopolkommission (2003): Netzwettbewerb durch Regulierung, Hauptgutachten 2000/2001. Baden-Baden. Monopolkommission (2005): Wettbewebspolitik im Schatten “Nationaler Champions”, Hauptgutachten 2002/2003. Baden-Baden. Monsen, R. Joseph, J. and Downs, A. (1965): A theory of large managerial firms. The Journal of Political Economy, vol. 73 (3): pp. 221– 236. Morck, R., Shleifer, A. and Vishny, R. W. (1988): Management ownership and market valuation: An empirical analysis. Journal of Financial Economics, vol. 20: pp. 293–315. Morck, R., Shleifer, A. and Vishny, R. W. (1989): Alternative mechanisms for corporate control. American Economic Review, vol. 79 (4): pp. 842–853. Mosler, K. and Schmid, F. (2004): Wahrscheinlichkeitsrechnung und schließende Statistik. Springer, Berlin, Heidelberg. Mueller, Elisabeth abd Spitz, A. (2002): Managerial ownership and firm performance in German small and medium-sized entreprises. Working Paper. Munzig, P. G. (2003): Enron and the economics of corporate governance. Working Paper. Murali, R. and Welch, J. B. (1989): Agents, owners, control and performance. Journal of Business Finance & Accounting, vol. 16 (3): pp. 385–389. Muurling, R. and Lehnert, T. (2004): Option-based compensation: a survey. The International Journal of Accounting, vol. 39 (4): pp. 365– 401. Myers, S. C. (1984): The capital structure puzzle. The Journal of Finance, vol. 39 (3): pp. 575–592.

References

343

Nassauer, F. (2000): Corporate Governance und die Internationalisierung von Unternehmungen, vol. 2671 of Europ¨aische Hochschulschriften: Reihe 5, Volks- und Betriebswirtschaft. Peter Lang — Europ¨aischer Verlag der Wissenschaften, Frankfurt am Main. Nathusius, E. (2005): Syndizierte Venture-Capital-Finanzierung, vol. 6 of Entrepreneurial and Financial Studies. Verlag Wissenschaft & Praxis, Sternenfels. Nestor, S. (2001): International efforts to improve corporate governance: Why and how. OECD. Nibler, M. (1998): Bank control and corporate performance in Germany: The evidence. University of Cambridge - Unpublished PhD thesis. N.N. (1999): Steuerexperte Herzig erwartet Boom beim Aktienr¨uckkauf durch Gesellschaften. Handelsblatt 08.06.1999, (108): p. 4. N.N. (2003): Corporate-Governance-Experten verlangen grundlegende ¨ Anderungen. Handelsblatt 27.06.2003, (121): p. 14. N.N. (2004): Analysten empfehlen Familienunternehmen – Grossaktion¨are sorgen f¨ur stabile Kursentwicklung. Handelsblatt 01.07.2004, (125): p. 35. N.N. (2005a): Dax-Vorst¨ande halten besser Maß. Handelsblatt 21.09.2005, (183): p. 12. N.N. (2005b): Die 50 gr¨oßten privaten Unternehmen Europas. Handelsblatt 16.08.2005, (157): p. 11. N.N. (2006): Aktion¨are r¨ugen Dividendenpolitik. Handelsblatt 01.06.2006, (105): p. 27. North, D. C. N. (1983): Comment on Stigler and Friedland, “The literature of economics: The case of Berle and Means”. Journal of Law & Economics, vol. 26: pp. 269–271. Norton, E. (1991): Capital structure and small public firms. Journal of Business Venturing, vol. 6 (4): pp. 287–303. Nowak, E., Rott, R. and Mahr, T. G. (2004): Wer den Kodex nicht einh¨alt, den bestraft der Kapitalmarkt? Johann Wolfgang Goethe Universit¨at Frankfurt am Main - Working Paper Series: Finance & Accounting No. 143. OECD (1998a): Shareholder value and the market in corporate control in OECD countries. Financial Market Trends 69, Organisation for Economic Co-operation and Development. OECD (1998b): Structural and regulatory developments in OECD countries. Financial Market Trends 71, Organisation for Economic Cooperation and Development.

344

References OECD (2004): OECD principles of corporate governance. Organisation for Economic Co-operation and Development. ¨ Ordemann, T., Muller, S. and Brackschulze, K. (2005): Handlungsempfehlungen f¨ur mittelst¨andische Unternehmen bei der Kreditfinanzierung. Betriebs-Berater, vol. 60 (5): pp. 19–24. BBSpezial. Palia, D. (2001): The endogeneity of managerial compensation in firm valuation: A solution. The Review of Financial Studies, vol. 14 (3): pp. 735–764. Peasnell, K. V., Pope, P. F. and Young, S. (2003): Managerial equity ownership and the demand for outside directors. European Financial Management, vol. 9 (2): pp. 231–250. Perfect, S. B. and Wiles, K. W. (1994): Alternative constructions of Tobin’s q: An empirical comparision. Journal of Empirical Finance, vol. 1 (3–4): pp. 313–341. Permanent Subcommittee on Investigations (2002): The role of the board of directors in Enron’s collapse. United States Senate. Peterson, S. (1965): Corporate control and capitalism. The Quarterly Journal of Economics, vol. 79 (1): pp. 1–24. Phani, B. V., Reddy, V. N., Ramachandran, N. and Bhattacharyya, A. K. (2005): Insider ownership, corporate governance and corproate performance. Working Paper. Portfolio Management (1988): Aktion¨arsstruktur und Aktienbesitz in der Bundesrepublik. PM-Studie. Potthoff, E. (1995): Ein Kodex f¨ur den Aufsichtsrat. Der Betrieb, vol. 48: pp. 163–164. Pound, J. (1988): Proxy contests and the efficiency of shareholder oversight. Journal of Financial Economics, vol. 20: pp. 237–265. Qiu, L. (2003): Which institutional investors monitor. Working Paper. Raettig, L. (2004): Eine Neuauflage des Neuen Marktes ist nicht notwendig. Frankfurter Allgemeine Zeitung 16.11.2004, (268): p. 23. Rajan, R. G. and Zingales, L. (2000): The governance of the new entreprise. In: X. Vives, ed., Corporate Governance — Theoretical and Empirical Perspectives. Cambridge University Press, Cambridge et al., pp. 201–232. Rappaport, A. (1999): Shareholder Value. Sch¨affer Poeschel, Stuttgart, 2nd edn. Reb´erioux, A. (2003): The micro-foundations of shareholder value: A critical survey of contractarian approaches. In: R. Cobbaut and J. Lenoble, eds., Corporate Governance — An Institutionalist Approach. Kluwer Law International, The Hague, pp. 65–85.

References

345

Rennneboog, L. (2000): Ownership, managerial control and the governance of companies listed on the Brussels stock exchange. Journal of Banking & Finance, vol. 24 (12): pp. 1959–1995. Ricardo-Campbell, R. (1983): Comments on the structure of ownership and the theory of the firm. Journal of Law & Economics, vol. 26: pp. 391–393. Richter, F. and Gr¨oniger, B. (2000): Kapitalmarktorientierte Steuerung mit Rendite- und Wachstumszielen. Sch¨affer-Poeschel Verlag, pp. 289–320. Richter, R. and Furubotn, E. G. (1999): Neue Institutionen¨okonomik. Mohr Siebeck, T¨ubingen, 2nd edn. Roberts, J. and Van den Steen, E. (2000): Shareholder interests, human capital investments and corporate governance. Stanford University Graduate School of Business - Working paper No. 1631. Roe, M. J. (1990): Political and legal constraints on ownership and control of public companies. Journal of Financial Economics, vol. 27 (1): pp. 7–41. Roe, M. J. (1994): Strong Managers, Weak Owners: The Political Roots of American Corporate Finance. Princeton University Press, Princeton. Roosenboom, P. and Schramade, W. (2005): The price of power: Valuing the controlling position of owner-managers in French IPO firms. Journal of Corporate Finance, vol. 12 (2): pp. 270–295. Rose, C. (2005): Managerial ownership and firm performance in listed Danish firms: In search of the missing link. European Management Journal, vol. 23 (5): pp. 542–553. Rosenstein, S. and Wyatt, J. G. (1990): Outside directors, board independence, and shareholder wealth. Journal of Financial Economics, vol. 26 (2): pp. 175–191. Ross, S. A. (1973): The economic theory of agency: The principal’s problem. The American Economic Review, vol. 63 (2): pp. 134–139. Ross, S. A. (1977a): The determination of financial structure: The incentive-signaling approach. The Bell Journal of Economics, vol. 8 (1): pp. 23–40. Ross, S. A. (1977b): Informational asymmetries, financial structure, and financial intermediation: Discussion. The Journal of Finance, vol. 32 (2): pp. 412–415. Roth, G. H. and W¨orle, U. (2004): Die Unabh¨angigkeit des Aufsichtsrats — Recht und Wirklichkeit. Zeitschrift f¨ur Unternehmens- und Gesellschaftsrecht (ZGR), vol. 33 (5): pp. 565–630.

346

References Sanders, G. (2001): Behavioral responses of CEOs to stock ownership and stock option pay. Academy of Management Journal, vol. 44 (3): pp. 477–492. Scheerer, M. (2006): EU st¨arkt Aktion¨arsrechte. Handelsblatt 11.01.2006, (8): p. 23. Schiereck, D. and Jaskiewicz, P. (2004): Familienunternehmen zeigen unterdurchschnittliche Performance. Going Public, (8): pp. 12–14. Schmid, M. (2003): Managerial incentives and firm valuation: Evidence from Switzerland. Working Paper. ¨ Schmidt, H., Drukarczyk, J., Honold, D., Prigge, S., Schuler, A. and Tetens, G. (1997): Corporate Governance in Germany, vol. 31 of Ver¨offentlichungen des HWWA-Institut f¨ur Wirtschaftsforschung Hamburg. Nomos Verlagsgesellschaft, Baden-Baden. Schmidt, K. (1986): Gesellschaftsrecht. K¨oln. Schmidt, R. H. (2003): Corporate governance in Germany: An economic perspective. Working Paper No. 118. Schmitz-Esser, V. (2001): Aktienindizes im Portfoliomanagement — Funktionen, Merkmale und Indexeffekte. Uhlenbruch Verlag, Bad Soden. Schmoller, G. v. (1900): Grundriß der Allgemeinen Volkswirtschaftslehre. M¨unchen. Schneider, H. and Heidenhain, M. (2000): The German Stock Corporation Act, vol. 16 of Series of Legislation in Translation. C.H. Beck and Kluwer Law International, M¨unchen et al., 2nd edn. Schr¨oder, U. and Schrader, A. (1998): The changing role of banks and corporate governance in Germany: Evolution towards the market? In: S. W. Black and M. Moersch, eds., Competition and Convergence in Financial Markets — The German and Anglo-American Models. Elsevier, Amsterdam et al., pp. 17–34. Schubert, W. and Hommelhoff, P. (1985): Hundert Jahre modernes Aktienrecht — Eine Sammlung von Texten und Quellen zur Aktienrechtsreform 1884 mit zwei Einf¨uhrungen. Walter de Gruyter, Berlin, New York. ¨ Schuller, A. (1983): Property Rights, Theorie der Firma und wettbewerbliches Marktsystem. In: A. Sch¨uller, ed., Property Rights und o¨ konomische Theorie. Vahlen, M¨unchen, pp. 145–183. Schwalbach, J. (2004): Ownership concentration and share valuation: Discussion. German Economic Review, vol. 5 (2): pp. 173–176. Scott, K. (1999): Institutions of corporate governance. Journal of Institutional and Theoretical Economics, vol. 155 (1): pp. 3–13.

References

347

Securities and Exchange Commission (SEC) (2006): Electronic data gathering, analysis, and retrieval system (edgar). URL http://www. sec.gov/edgar.shtml, (viewed: 30.07.2006). Seifert, B., Gonenc, H. and Wright, J. (2005): The international evidence on performance and equity ownership by insiders, blockholders, and institutions. Journal of Multinational Financial Management, vol. 15 (2): pp. 171–191. Sherman, H. C. and Kaen, F. R. (1997): Die deutschen Banken und ihr Einfluß auf Unternehmensentscheidungen. ifo Schnelldienst, vol. 50 (23): pp. 3–20. Shleifer, A. and Vishny, R. W. (1989): Management entrenchment — The case of manager-specific investments. Journal of Financial Economics, vol. 25 (1): pp. 123–139. Shleifer, A. and Vishny, R. W. (1997): A survey of corporate governance. The Journal of Finance, vol. 52 (2): pp. 737–783. Short, H. (1994): Ownership, control, financial structure and the performance of firms. Journal of Economic Surveys, vol. 8 (3): pp. 203– 249. Smith, A. ([1994] 1776): An Iquiry into the Nature and Causes of the Wealth of Nations (Reprint). The Modern Library, New York. ¨ Sommer, U. (2005): H¨oherer Streubesitz erleichtert Ubernahmen im Dax. Handelsblatt 19./20./21.08.2005, (160): p. 29. Sommer, U. (2006): Aufsichtsr¨ate gehen mit 67 noch nicht in Rente. Handelsblatt 20.03.2006, (56): p. 16. Sprenger, C. (2002): Ownership and corporate governance in Russian industry: A survey. European Bank for Reconstruction and Development - Working paper No. 70. Sraer, D. and Thesmar, D. (2004): Performance and behavior of family firms: Evidence from the French stock market. CEPR Discussion Paper No. 4520. Statistisches Bundesamt (2005): Umsatzsteuerstatistik 2003. URL http://www.destatis.de/download/d/fist/fistdow3.xls, (viewed: 12.08.2005). Steiger, M. (1998): Institutionelle Investoren und Corporate Governance — Eine empirische Analyse. ZEW - Dokumentation No. 05/1998. Steinbeis, M. (2004): Gr¨une drohen mit Gesetz f¨ur Aufsichtsr¨ate. Handelsblatt 26.08.2004, (165): p. 6. Stigler, G. (1958): The economies of scale. Journal of Law & Economics, vol. 1: pp. 54–71.

348

References Stigler, G. J. and Friedland, C. (1983): The literature of economics: The case of Berle and Means. Journal of Law & Economics, vol. 26: pp. 237–268. Stiglitz, J. (1985): Credit markets and the control of capital. Journal of Money, Credit and Banking, vol. 17: pp. 133–152. Stulz, R. M. (1988): Managerial control of voting rights: Financing policies and the market for corporate control. Journal of Financial Economics, vol. 20: pp. 25–54. Stulz, R. M. (1990): Managerial discretion and optimal financing policies. Journal of Financial Economics, vol. 26 (1): pp. 3–27. Taussig, F. W. and Barker, W. S. (1925): American corporations and their executives: A statistical inquiry. The Quarterly Journal of Economics, vol. 40 (1): pp. 1–51. Theisen, M. R. (2003): Herausforderung Corporate Governance. Die Betriebswirtschaft, vol. 63 (4): pp. 441–464. Thomsen, S., Pedersen, T. and Kvist, H. K. (2006): Blockholder ownership: Effects on firm value in market and control based governance systems. Journal of Corporate Finance, vol. 12: pp. 246–269. Thonet, P. J. and Poensgen, O. H. (1979): Managerial control and economic performance in Western Germany. The Journal of Industrial Economics, vol. 28 (1): pp. 23–37. Tirole, J. (2001): Corporate governance. Econometrica, vol. 69 (1): pp. 1–35. Ulen, T. S. (1993): The Coasean firm in law and economics. Journal of Corporation Law, vol. 18 (2): pp. 301–331. Valc´arcel, S. (2002): Theorie der Unternehmung und Corporate Governance — Eine vertrags- und ressourcenbezogene Betrachtung, vol. 101 of Beitr¨age zur betriebswirtschaftlichen Forschung. Deutscher Universit¨ats-Verlag, Wiesbaden. Van der Elst, C. (2000): The equity markets, ownership structures and control: Towards an international harmonisation. Universit¨at Gent Working Paper No. 04/2000. Veblen, T. (1924): The Engineers and the Price System. Viking, New York. Villalonga, B. and Amit, R. (2006): How do family ownership, control and management affect firm value? Journal of Financial Economics, vol. 80 (2): pp. 385–417. Vives, X. (2000): Corporate governance: Does it matter? In: X. Vives, ed., Corporate Governance — Theoretical and Empirical Perspectives. Cambridge University Press, Cambridge et al., pp. 1–21.

References

349

von Heusinger, R. (2002): Ackermann s¨agt sich seinen Stuhl — Die neue F¨uhrungsstruktur der Deutschen Bank passt nicht. Die Zeit 24.01.2002, (5): p. 26. von Heusinger, R. (2004): Rettet den Kapitalismus: Die B¨orse zerst¨ort jede langfristige Unternehmensstrategie — Zeit sich vom Shareholder-Value zu verabschieden. Die Zeit 28.10.2004, (45): pp. 35–36. ¨ von Rosen, R. (2003): Uberregulierung behindert Kapitalm¨arkte und Wachstum. Frankfurter Allgemeine Zeitung 14.08.2003. von Rosen, R. (2004): Kapitalmarkt und Corporate Governance unter Einbeziehung der Rechnungslegung. Der Konzern, (5): pp. 325–332. Wahl, S. J. (2004): Private Debt — Private Fremdkapital- und Mezzanine-Kapital-Platzierungen bei institutionellen Investoren, vol. 4 of Entrepreneurial and Financial Studies. Verlag Wissenschaft & Praxis, Sternenfels. Weber, M. (1998): Financial systems and corporate governance: A review of the international evidence — Comment. Journal of Institutional and Theoretical Economics, vol. 154 (1): pp. 166–169. Weber, M. and Dudney, D. (2003): A reduced form coefficients analysis of executive ownership, corporate value, and executive compensation. The Financial Review, vol. 38: pp. 399–413. Weisbach, M. S. (1988): Outside directors and CEO turnover. Journal of Financial Economics, vol. 20: pp. 431–460. Weishaupt, G. (2005): M¨achtige Strippenzieher. Handelsblatt 11.10.2005, (196): p. 24. Weishaupt, G. (2006): Die Haniels — Wandel und Werte. Handelsblatt 16.05.2006, (94): p. 35. Wenger, E. and Kaserer, C. (1998a): German banks and corporate governance: A critical view. In: K. J. Hopt, H. Kanda, M. J. Roe, E. Wymeersch and S. Prigge, eds., Comparative Corporate Governance - The State of the Art and Emerging Research. Oxford University Press, pp. 499–536. Wenger, E. and Kaserer, C. (1998b): The German system of corporate governance — A model which should not be imitated. In: S. W. Black and M. Moersch, eds., Competition and Convergence in Financial Markets — The German and Anglo-American Models. Amsterdam et al., pp. 41–78. Weston, J. F. (1979): The tender takeover. Mergers and Acquisitions, vol. 14: pp. 74–82.

350

References Weston, J. F., Siu, J. A. and Johnson, B. A. (2001): Takeovers, Restructuring & Corporate Governance. Prentice Hall Finance Series. Prentice Hall, New Jersey, 3rd edn. White, H. (1980): A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, vol. 48: pp. 817–838. Wiebe, F. (2003): Bewertung so niedrig wie Anfang der neunziger Jahre. Handelsblatt 14.03.2003, (52): p. 32. Wiedeking, W. (2005): Geiz ist eine Tods¨unde. Die Zeit 28.04.2005, (18): pp. 17–20. Interview of Wendelin Widedeking by Hanns-Bruno Kammert¨ons and Stephan Lebert. Williams, V. B. (1990): Managerial share ownership and the stock price effects of antitakeover amendment proposals. The Journal of Finance, vol. 45 (5): pp. 1627–1640. Williamson, O. E. (1975): Markets and Hierarchies: Analysis and Antitrust Implications. The Free Press, New York. Winkens, W. (2000): Unternehmensfinanzierung und unvollst¨andige Vertr¨age. No. 15 in Reihe: Finanzierung, Kapitalmarkt und Banken. Josef Eul Verlag, Lohmar, K¨oln. Wirtschaftsrat der CDU e.V. (2005): Herausforderungen f¨ur deutsche Unternehmen: Corporate Governance in Europa. Wittkowski, B. (2006): Der Aktion¨ar soll eigenverantwortlich werden. B¨orsen-Zeitung 27.01.2006, (19): p. 4. Woidtke, T. (2002): Agents watching agents? Evidence from pension fund ownership and firm value. Journal of Financial Economics, vol. 63 (1): pp. 99–131. W´ojcik, D. (2003): Change in the German model of corporate governance: Evidence from blockholdings 1997-2001. Environment and Planning A, vol. 35 (8): pp. 1431–1458. Wolf, J. B. (1999): The Effects of Agency Problems on the Financial Behavior, Performance, and Efficiency of German Industrial Stock Corporations, vol. 2424 of European University Studies. Peter Lang — Europ¨aischer Verlag der Wissenschaften, Frankfurt am Main et al. Wolff, U. (2000): Beteiligungsbesitz und Corporate Governance - Eine Effizienzanalyse institutioneller Finanzierungsbedingungen. Gabler Edition Wissenschaft. Deutscher Universit¨ats-Verlag, Wiesbaden. Wooldridge, J. M. (2002): Econometric Analysis of Cross Section and Panel Data. The MIT Press, Cambridge, London. Yermack, D. (2005): Flights of fancy: Corporate jets, CEO perquisites, and inferior shareholder returns. Working Paper.

References

351

Zhou, X. (2001): Understanding the determinants of managerial ownership and the link between ownership and performance: Comment. Journal of Financial Economics, vol. 62 (3): pp. 559–571. Ziener, M., Riecke, T., Hussla, G. A. and Fockenbrock, D. (2006): Das Pendel schl¨agt zur¨uck. Handelsblatt 03.08.2006, (148): p. 2. Zingales, L. (2000): In search of new foundations. The Journal of Finance, vol. 55 (4): pp. 1623–1353. Zobel, J. (2002): A case for the family owned conglomerate. The McKinsey Quarterly, (4): pp. 128–137. Interview of Jaime Zobel by Ken Gibson.

Entrepreneurial and Financial Studies Herausgegeben von: Prof. Dr. Dr. Ann-Kristin Achleitner, Prof. Dr. Christoph Kaserer Band

1:

Band

2:

Band

3:

Band

4:

Band

5:

Band Band

6: 7:

Band

8:

Band 9:

Ronald Engel: Seed-Finanzierung wachstumsorientierter Unternehmensgründungen, 2003. Ann-Kristin Achleitner, Simon Wahl: Corporate Restructuring in Deutschland, 2003. Tobias Popoviü: Customer Capital – Die Wertschöpfung von E-Commerce-Unternehmen und ihre zweckadäquate Bewertung aus Perspektive des Aktienresearch, 2004. Simon J. Wahl: Private Debt – Private Fremdkapital- und MezzanineKapital-Platzierungen bei institutionellen Investoren, 2004. Moritz Hagenmüller: Investor Relations von Private-Equity-Partnerships, 2004. Eva Nathusius: Syndizierte Venture-Capital-Finanzierung, 2005. Christian H. Fingerle: Smart Money ņ Influence of Venture Capitalists on High Potential Companies, 2005. Jens Mittnacht: Die Kapitalmarktbewertung von Desinvestitionen – eine Ereignisstudie über Selloffs und Unit Buyouts in Kontinentaleuropa, 2006. Benjamin Moldenhauer: Insider Ownership, Shareholder Structures and Corporate Governance, 2007.