The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry

Jan-Peer Laabs The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry GABLE...

Author: Jan-Peer Laabs

58 downloads 703 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Jan-Peer Laabs The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry

GABLER EDITION WISSENSCHAFT EBS Forschung Schriftenreihe der European Business School (EBS) International University · Schloss Reichartshausen Herausgegeben von Univ.-Prof. Ansgar Richter, PhD

Band 71

Die European Business School (EBS) – gegründet im Jahr 1971 – ist Deutschlands älteste private Wissenschaftliche Hochschule für Betriebswirtschaftslehre im Universitätsrang. Dieser Vorreiterrolle fühlen sich ihre Professoren und Doktoranden in Forschung und Lehre verpflichtet. Mit der Schriftenreihe präsentiert die European Business School (EBS) ausgewählte Ergebnisse ihrer betriebs- und volkswirtschaftlichen Forschung.

Jan-Peer Laabs

The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry With a foreword by Prof. Dr. Dirk Schiereck

GABLER EDITION WISSENSCHAFT

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

Dissertation European Business School, International University Schloss Reichartshausen, Oestrich-Winkel, 2009 D1540

1st Edition 2009 All rights reserved © Gabler | GWV Fachverlage GmbH, Wiesbaden 2009 Editorial Office: Claudia Jeske / Britta Göhrisch-Radmacher Gabler is part of the specialist publishing group Springer Science+Business Media. www.gabler.de No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the copyright holder. Registered and/or industrial names, trade names, trade descriptions etc. cited in this publication are part of the law for trade-mark protection and may not be used free in any form or by any means even if this is not specifically marked. Cover design: Regine Zimmer, Dipl.-Designerin, Frankfurt/Main Printed on acid-free paper Printed in Germany ISBN 978-3-8349-1693-8

Foreword It is precisely the kind of highly variable success that Schaeffler KG was forced to endure during its acquisitions of FAG Kugelfischer and Continental, which provides a telling account of the myriad potential consequences of mergers and acquisitions in the automotive supply industry. The global competitive landscape for automotive suppliers is wholly unique, and not just because its consumers and key customers, the automotive manufacturers, have joined together in an ever-narrowing oligopoly over the last twenty years, nor because the current focus on suppliers is already helping restore an increasingly stable balance of power. Since automotive manufacturers serve international markets, the competitive environment for suppliers has also always been transnational. This becomes all the more relevant when we consider recent shifts in the automotive manufacturing value chain. Meanwhile, some suppliers are producing a higher value contribution during construction of a car than the actual manufacturer. In these situations, international M&A transactions seem especially attractive, and should, in actual fact, generate positive reactions on the capital markets. But is that really the case? Studies looking specifically into the long-term success of international M&A transactions have been a rare commodity until now, the current state of knowledge on the success of acquisitions in the automotive supplier industry is even more limited, and the transferability of evidence based on experiences in other sectors is highly questionable. There was, therefore, an urgent need to analyze the success factors for M&A in this industry This paper seizes on this considerable research gap, with great attention to detail and maximum possible accuracy. The primary objective was to use capital market data and financial accounting information to investigate the success of international M&A transactions in the automotive supply industry, and to identify the key factors in this success. Thus can we achieve an objective state of knowledge, which can be used as a

VI

Foreword

basis for deriving well-founded, recommended actions to be implemented in industry practice. Mr. Laabs is in the best possible position to fulfill his self-imposed objectives in terms of this dissertation. The paper contains a wealth of very interesting results, and is written in such a way that the reader will thoroughly enjoy studying this entire work from cover to cover. I wish this paper the extensive circulation it deserves.

Professor Dr. Dirk Schiereck

Preface Writing the following doctoral thesis has been an inspiring and shaping personal experience. However, the progress of my research would not have been as advanced, focused, and exciting without the support of many people whom I would like to acknowledge and thank with the following lines. They made a successful completion of my work possible. First, I would like thank my doctoral supervisor, Prof. Dr. Dirk Schiereck, for his academic guidance and encouragement in preparing this thesis. Besides providing valuable advice and instant feedback, he always managed to balance academic "judgment calls" with an extraordinarily constructive and collaborative working atmosphere. I appreciated the discussions and will remember our joint journey to an academic conference in Nashville as one of the highlights of these past two years. In addition, I would like to thank Prof. Dr. Ronald Gleich for readily agreeing to assume the role of secondary advisor to my thesis. His strategic perspective on the implications of my work represented an enriching addition to my finance-driven research approach. Second, I am thankful for the support of a number of friends and colleagues. First and foremost, I would like to acknowledge the exceptional help and support of my friend Nils Steiner. Besides being an inspirational discussion partner, he contributed significantly with his technical and modeling experience. Without his "hands-on" practical advice, some of my calculations might have taken considerably longer, or even, would not have found their way into the final version of this thesis. In addition, I would like to thank my colleagues and friends from the Frankfurt Office "Fellow Community" for the joint times we shared, both by discussing on or off topic or by playing numerous rounds of table soccer. All other friends I would like to thank for spurring my motivation by continuously asking for my work's progress. This book is the answer to these questions.

VIII

Preface

Last, but definitely not least, I would like to thank my family and, in particular, my parents. Their unconditional support and encouraging mindset gave me the freedom I needed to pursue my goals and the certainty of their helping advice and opinion when required. Finally, I am especially grateful to my long-time girl-friend Nadine and our daughter Lina-Marie. They not only managed to cope with the additional time constraints inherent in writing a doctoral thesis, but also seemed to outbalance any upcoming doubts on my side with an increased level of encouragement and support. Their support, patience, and love have been invaluable sources of inspiration and motivation to finish this thesis. I dedicate this book to them.

Jan-Peer Laabs

Content Overview List of Figures................................................................................................................XV List of Tables .............................................................................................................. XVII List of Abbreviations ................................................................................................... XIX 1

Introduction .............................................................................................................. 1

2

Research Foundations............................................................................................... 9

3

Study 1: Determinants of Capital Market Performance ......................................... 23

4

Study 2: Does Operating Performance Meet Market Expectations?...................... 93

5

Study 3: How a Good Bidder Becomes a Good Target – The Case of Continental AG Acquiring Siemens VDO ........................................................... 133

6

Conclusion............................................................................................................ 173

Appendix ...................................................................................................................... 181 References .................................................................................................................... 189

Table of Contents List of Figures................................................................................................................XV List of Tables .............................................................................................................. XVII List of Abbreviations ................................................................................................... XIX 1

Introduction .............................................................................................................. 1 1.1 Problem Definition and Objectives .................................................................. 1 1.2 Course of Analysis ........................................................................................... 5

2

Research Foundations............................................................................................... 9 2.1 The Automotive Supply Industry ..................................................................... 9 2.1.1 Definition and Scope ............................................................................ 9 2.1.2 Current Trends and Challenges .......................................................... 10 2.2 Mergers and Acquisitions............................................................................... 15 2.3 Measuring Success of Mergers and Acquisitions........................................... 16 2.3.1 Time Horizons .................................................................................... 16 2.3.2 Research Approaches ......................................................................... 19

3

Study 1: Determinants of Capital Market Performance ......................................... 23 3.1 Introduction .................................................................................................... 23 3.2 Literature Review and Hypotheses................................................................. 25 3.2.1 Related Literature ............................................................................... 25 3.2.2 Hypotheses ......................................................................................... 29 3.2.2.1 The Overall Effect ............................................................... 29 3.2.2.2 The Impact of Transaction Characteristics.......................... 29 3.2.2.3 The Impact of Acquirer Characteristics............................... 31 3.3 Data and Methodology ................................................................................... 33 3.3.1 Identifying Merging Companies......................................................... 33 3.3.2 Portfolio of Matching Firms ............................................................... 35 3.3.3 Econometric Strategy ......................................................................... 36 3.3.3.1 Short-term Methodology ..................................................... 36 3.3.3.2 Long-term Methodology...................................................... 40 3.4 Empirical Results............................................................................................ 43 3.4.1 The Overall Effect .............................................................................. 43 3.4.2 The Impact of Transaction Characteristics......................................... 47 3.4.2.1 Geographical Expansion...................................................... 47

XII

Table of Contents

3.4.2.2 Product Diversification........................................................ 52 3.4.2.3 Transaction Size .................................................................. 57 3.4.3 The Impact of Acquirer Characteristics.............................................. 61 3.4.3.1 Product Groups .................................................................... 61 3.4.3.2 Acquisition Experience........................................................ 67 3.5 Robustness Cheques and Cross-Sectional Regressions.................................. 72 3.5.1 Regression of Short-term CAARs ...................................................... 73 3.5.2 Regression of Long-term BHARs ...................................................... 83 3.6 Conclusion...................................................................................................... 90 4

Study 2: Does Operating Performance Meet Market Expectations?...................... 93 4.1 Introduction .................................................................................................... 93 4.2 Literature Review and Hypotheses................................................................. 95 4.2.1 Related Literature ............................................................................... 95 4.2.2 Hypotheses ......................................................................................... 98 4.2.2.1 The Overall Effect ............................................................... 98 4.2.2.2 The Impact of Transaction and Acquirer Characteristics.. 100 4.2.2.3 The Correlation of Accounting- and Event-Study Results 100 4.3 Data and Methodology ................................................................................. 101 4.3.1 Identifying Merging Companies....................................................... 101 4.3.2 Portfolio of Matching Firms ............................................................. 103 4.3.3 Econometric Strategy ....................................................................... 105 4.4 Empirical Results.......................................................................................... 108 4.4.1 The Overall Effect ............................................................................ 108 4.4.2 Determinants of Profitability............................................................ 113 4.5 The Correlation of Accounting- and Event-Study Results........................... 121 4.6 Conclusion.................................................................................................... 129

5

Study 3: How a Good Bidder Becomes a Good Target – The Case of Continental AG Acquiring Siemens VDO ........................................................... 133 5.1 Introduction .................................................................................................. 133 5.2 Literature Review and Industry Overview ................................................... 135 5.2.1 Related Literature ............................................................................. 135 5.2.2 Overview of the Automotive Supply Industry ................................. 139 5.3 Case Study Background ............................................................................... 142 5.3.1 The Transaction Partners .................................................................. 142 5.3.1.1 Continental AG.................................................................. 142 5.3.1.2 Siemens VDO Automotive................................................ 146 5.3.2 Transaction Motives ......................................................................... 149 5.3.3 The Acquisition Event ...................................................................... 153

Table of Contents

XIII

5.4 Acquisition Performance .............................................................................. 156 5.4.1 The Capital Market Perspective ....................................................... 156 5.4.2 Performance Analysis....................................................................... 164 5.5 Discussion and Conclusion........................................................................... 170 6

Conclusion............................................................................................................ 173

Appendix ...................................................................................................................... 181 References .................................................................................................................... 189

List of Figures Figure 2.1: M&A Transactions in the Automotive Supply Industry.............................. 13 Figure 3.1: CAARs to Acquirers, Targets, and Combined Entities ............................... 45 Figure 5.1: Transaction Volume in the Automotive Supply Industry .......................... 140 Figure 5.2: Share Price Development of Continental AG and Siemens AG................ 158 Figure 5.3: Long-term Relative Share Price Development of Continental AG............ 161 Figure 5.4: Revenue Split of Continental AG across Divisions ................................... 169

List of Tables Table 3.1:

Overview of Hypotheses and Predicted Value Impact on Acquirers ........ 30

Table 3.2:

Overview of the Transaction Sample – Descriptive Statistics .................. 35

Table 3.3:

CAARs to Acquirers.................................................................................. 44

Table 3.4:

BHARs to Acquirers.................................................................................. 46

Table 3.5:

Abnormal Returns (FF3F) to Acquirers .................................................... 47

Table 3.6:

Acquirer CAARs – Differences by Geographical Scope .......................... 49

Table 3.7:

Acquirer CAARs – Diferences by Continental Scope............................... 50

Table 3.8:

Acquirer BHARs – Differences by Geographical Scope .......................... 51

Table 3.9:

Abnormal Returns to Acquirers – Differences by Geographical Scope.... 51

Table 3.10: Acquirer CAARs – Differences by Product Scope ................................... 54 Table 3.11: Acquirer BHARs– Differences by Product Scope .................................... 55 Table 3.12: Abnormal Returns to Acquirers – Differences by Product Scope............. 56 Table 3.13: Acquirer CAARs – Differences by Transaction Size................................ 58 Table 3.14: Acquirer BHARs – Differences by Transaction Size................................ 60 Table 3.15: Abnormal Returns to Acquirers – Differences by Transaction Size ......... 61 Table 3.16: Acquirer CAARs – Differences by Product Group ................................... 63 Table 3.17: Acquirer BHARs – Differences by Product Group ................................... 65 Table 3.18: Abnormal Returns to Acquirers – Differences by Product Group ............ 67 Table 3.19: Acquirer CAARs – Differences by Acquisition Experience ..................... 69 Table 3.20: Acquirer CAARs – Differences for Bidder Champions............................ 70 Table 3.21: Acquirer BHARs – Differences by Acquisition Experience ..................... 71 Table 3.22: Abnormal Returns to Acquirers – Differences by Acquirer Experience... 72 Table 3.23: Regression of Short-term CAARs to Acquirers ........................................ 77 Table 3.24: Step-wise Regression of Short-term CAARs to Acquirers ....................... 78 Table 3.25: Regression of Long-term BHARs to Acquirers ........................................ 86 Table 3.26: Step-wise Regression of 36-month BHARs to Acquirers ......................... 87 Table 4.1:

Overview of the Transaction Sample – Descriptive Statistics ................ 103

Table 4.2:

Average Performance Following M&A Transactions............................. 109

XVIII

List of Tables

Table 4.3:

Abnormal Performance Changes in the Automotive Supply Industry.... 111

Table 4.4:

Regression of Median Post-Merger Abnormal Performance .................. 112

Table 4.5:

Abnormal Performance Change – Differences by Product Scope .......... 115

Table 4.6:

Abnormal Performance Change – Differences by Geographical Scope . 116

Table 4.7:

Stepwise Regression of Abnormal Performance Changes ...................... 119

Table 4.8:

BHARs to Acquirers Applied in the Accounting Study.......................... 122

Table 4.9:

Regression of 36-month BHARs on Abnormal Performance Changes .. 123

Table 4.10: Relationship between Event- and Accounting-Study Results ................. 127 Table 5.1:

Milestones in the Continental/Siemens VDO Transaction...................... 154

Table 5.2:

Abnormal Announcement Returns to Continental AG and Siemens AG 159

Table 5.3:

Buy-and-Hold Abnormal Returns to Continental AG............................. 162

Table 5.4:

Abnormal Returns to Continental AG versus Different Peer Groups ..... 163

Table 5.5:

Quarterly Balance Sheet/Income Statement Items of Continental AG ... 165

Table 5.6:

Unadjusted Quarterly Performance Indicators of Continental AG ......... 167

Table 5.7:

Abnormal Quarterly Performance Indicators of Continental AG ........... 168

List of Abbreviations Adj.R2

Adjusted Coefficient of Determination

AG

Aktiengesellschaft (Stock Corporation)

Avg.

Average

BHAR

Buy-and-Hold Abnormal Return

BHR

Buy-and-Hold Return

bn

Billion

BVA

Book Value of Assets

CAAR

Cumulative Average Abnormal Return

CAGR

Compound Annual Growth Rate

CAR

Cumulative Abnormal Return

CF

Operating Cash Flow

COGS

Cost of Goods Sold

Conti.

Continental AG

Corp.

Corporation

CO2

Carbon Dioxide

DAX

Deutscher Aktien Index (German Stock Index)

DWS

Durbin-Watson Statistic

EBIT

Earnings Before Interest and Taxes

EBITDA

Earnings Before Interest, Taxes, Depreciation, and Amortization

EMH

Efficient Market Hypothesis

EUR

Euro

FDI

Foreign Direct Investment

FF3F

Fama-French-3-Factor

FTSE

Financial Times Stock Exchange

GHz

Gigahertz

GmbH

Gesellschaft mit beschränkter Haftung (Limited Liability Corporation)

XX

List of Abbreviations

GST

Generalized Sign Test

HML

High Minus Low

i.e.

that is

IFRS

International Financial Reporting Standard

IPO

Initial Public Offering

M&A

Mergers and Acquisitions

MDN

Median

NOPLAT

Net Operating Profit Less Adjusted Taxes

OEM

Original Equipment Manufacturer

OLS

Ordinary Least Squares

R&D

Research and Development

SDC

Securities Data Corporation

SG&A

Selling, General and Administrative Expense

SIC

Standard Industrial Classification

SMB

Small Minus Big

S&P

Standard & Poor's

S&P 500

Standard & Poor's 500 Index

US

United States

USD

United States Dollar

US-GAAP

United States – Generally Accepted Accounting Principles

VDO

Vereinigte Deutsche Tachometerwerke und OSA Apparate GmbH

vs.

versus

1

Introduction

1.1

Problem Definition and Objectives As a result of economic and structural changes, the automotive supply industry

has been facing significant consolidation activity over the last twenty years. The pressure to produce better equipped and less expensive automobiles created a growing trend towards specialization and internationalization among automotive suppliers. For many players, mergers and acquisitions (M&A) became a common strategic response to these trends inducing increasingly dominated product ranges and a truly global competition. Consequently, the number of existing suppliers has been continuously decreasing: In Europe, for example, the number of direct suppliers dropped from 10,000 in the early 1970s to 3,000 in 1995 and to an estimated 500 in the year 2000 (Sadler (1999)). Between 1991 and 1999, horizontal M&A transactions with more than USD 50 million transaction value steadily increased both in total numbers as well as in total volume. While the influence of this potential merger wave appeared to weaken after the year 2000, the multi-billion dollar transactions involving German Continental AG in 2007 and 2008 have recently revived consolidation discussions and potentially indicate the start of a next major consolidation wave. The unique competitive situation within the automotive supply industry promotes this on-going consolidation. By the early 21st century, automotive suppliers are facing increasing pressure from both sides of the automotive value chain. On the one side, the Original Equipment Manufacturers (OEMs) as the suppliers' main customers consolidate and, at the same time, become more and more sophisticated. While the globalization of OEMs forces suppliers to costly follow abroad, a change in the manufacturers' sourcing preferences additionally strains the suppliers' profit situation. Since OEMs now prefer to source full systems of components from a limited number of suppliers, they transfer an increasing degree of production and development tasks down the value chain (Sadler (1999), Von Corswant and Fredriksson (2002)). On the other side of

2

1 Introduction

the value chain, increasing raw material prices for crude oil, natural rubber as well as many metals compile a threat for the remaining supplier profits. Consequently, competition among suppliers remains fierce and is still growing. Where a few international players dominate the product range, strong barriers to entry exist. Where specialization does not create a competitive edge, the comparably high number of existing suppliers encourages a very strong rivalry, as apparent in the leather and tire industries (Aktas, de Bodt, and Derbaix (2004)). Due to these specific industry characteristics, previous research on the value creation potential of M&A among automotive suppliers identifies the industry as an outlier: capital markets generally seem to appreciate M&A as a valuable response strategy by granting significant positive abnormal announcement returns to acquirers. These positive announcement returns represent the industry-specific synergy and efficiency potentials underlying the transactions as perceived by investors and capital markets (Mentz and Schiereck (2008)). Although preceding M&A literature on the general announcement effect of M&A also concludes that M&A create value for the combined entities, positive announcement returns to acquiring companies are clearly an exception. Comparing a variety of different studies, Bruner (2002) finds announcement returns for acquirers to be essentially zero, while Loughran and Vijh (1997) even conclude them to be overall negative. Consequently, the outstanding competitive characteristics of the automotive supply industry appear to also translate into an outstanding value creation potential for suppliers engaging in takeover activities. However, given that announcement returns represent a short-term assessment of value creation building on investors' expectations about future performance, the question remains whether acquirers are in fact able to sustain their exceptional positive returns beyond a short-term perspective. Previous literature on the general long-term capital market performance of acquirers provides a rather negative outlook. The majority of studies including the early work of Mandelker (1974) and the methodological milestones of Franks, Harris and Titman (1991) and Loughran and Vijh (1997) point to-

1.1 Problem Definition and Objectives

3

wards significant value losses for acquirers over three to five years following an M&A announcement. In the light of these consistently negative findings, it becomes even more relevant to assess the long-term performance of acquirers in the automotive supply industry. If suppliers were able to sustain their exceptional positive announcement returns in the long-run, the industry could truly be regarded as a consistent positive outlier representing a unique investment opportunity for industry in- and outsiders. If, however, acquirers were not able to sustain their positive announcement returns, this result would reveal their inability to fulfill capital market expectations about the inherent synergy potentials. In the light of on-going consolidation, such an inability poses a challenge to the management of merging companies and calls for a comprehensive analysis of determinants for the observed value effect. Besides the negative outlook delivered by preceding literature, a number of theoretical considerations makes the assessment of long-term M&A success generally more challenging than determining short-term announcement returns. Firstly, the application of longer time horizons requires well-specified methodological approaches to overcome inherent statistical biases. As a result, studies assessing long-term performance usually apply either statistically advanced methodological approaches or a combination of different methodologies to arrive at valid conclusions (Barber and Lyon (1997)). Secondly, long-term success of M&A can be determined based on different data sources. Besides the capital market perspective as expressed in abnormal share returns, publicly available accounting information also frequently serves as the basis for analyzing the long-term post-merger performance of acquirers. While accounting studies focus exclusively on historic data, both the resulting general effect of M&A on acquirer performance as well as the correlation of results under the two approaches remain ambiguous (Healy, Palepu, and Ruback (1992), Fridolfsson and Stennek (2005)). Consequently, a combination of different approaches and data bases is required to generate a comprehensive view on the long-term success of M&A and to explore potential determinants behind long-term performance.

4

1 Introduction

In general, it becomes apparent that both the uniquely competitive situation within the industry as well as the theoretical considerations on the measurement of longterm M&A success significantly influence any approach assessing long-term performance within the automotive supply industry. Consequently, this industry represents a particularly relevant research object in order to assess the sustainability of positive abnormal announcement returns in the long-term and its underlying determinants. Overall, the following objectives serve as the basis for the upcoming empirical analysis: 1. The study empirically analyzes the long-term value creation potential of M&A in the automotive supply industry. In addition to arriving at an overall judgment on the long-term value creation potential, the question whether acquirers in the automotive supply industry are able to sustain their extraordinary positive short-term returns over a long-term time horizon is addressed in particular. Therefore, a combination of different methodological approaches including long-term event- and accounting-study methodologies is applied on a base sample of 230 M&A transactions within the global automotive supply industry between 1981 and 2007. 2. Based on the overall assessment of the value creation potential, determinant variables of long-term performance are analyzed for the magnitude and direction of their respective value impact. In this regard, variables identified to influence short- and long-term performance in preceding general finance literature are challenged against their effect within the specific competitive environment of the automotive supply industry. By identifying a comprehensive list of potential value drivers, this study does not only advance previous literature by an additional perspective on the effect of the various determinants, but also allows senior management of automotive suppliers to identify valuemaximizing strategies for future takeovers. 3. With regard to the described theoretical considerations, this study focuses on creating a comprehensive and differentiated perspective on the acquirers'

1.2 Course of Analysis

5

post-merger performance. For this purpose, the last objective of this study is to verify the derived effects between the different available methodologies. The long-term abnormal capital market returns are challenged against the abnormal performance of acquirers as published in their accounting statements. Thereby, a differentiation between genuine value impact and the subjective assessment of capital markets becomes feasible. A case study on a successful takeover from the industry provides an additional opportunity to verify the findings from the general event studies.

1.2

Course of Analysis This chapter defines and explains the overall objectives of this thesis and sets

the course of analysis for the following elaborations. Chapter 2 provides a number of research foundations including a definition of 'automotive supplier', 'mergers and acquisitions', and different research approaches. In addition, it provides a description of the structure and competitive environment within the automotive supply industry and of the different theoretical considerations in measuring long-term post-merger performance. Thereby, the chapter intends to create an understanding for the competitive pressures affecting automotive suppliers and the various methodological approaches available to determine long-term success of takeover activities. Chapters 3 to 5 each represent a self-contained empirical study addressing different research questions related to the overall topic. In sum, they compose the main empirical analyses of this thesis and are developed in subsequent order. Chapter 3 examines the short- and long-term value effects of horizontal mergers and acquisitions on the capital market performance of acquirers in the automotive supply industry. It answers to the questions whether acquirers in this industry are actually able to sustain their extraordinary short-term share returns in the long-run and what the potential determinants for the observed short- and long-term value effects are.

6

1 Introduction

For this purpose, the significant positive announcement returns determined by preceding research are at first updated and validated against a global sample of 230 takeover announcements between 1981 and 2007. Then, this chapter challenges the short-term announcement returns against the long-term capital market performance of the respective acquirers. By applying the Fama-French-3-Factor-model (FF3F-model) in calendar-time and the control-firm approach in event-time, a comprehensive perspective on the post-merger capital market performance is created revealing whether acquirers are able to sustain their positive short-term abnormal returns. After this, Chapter 3 analyzes a number of determining variables for both the short- and long-term capital market performance and tests for their significance using a cross-sectional regression model of the determined long-term Buy-and-Hold Abnormal Returns (BHARs). Chapter 4 addresses the research question whether the observed long-term capital market returns to acquirers are consistent with their post-merger operating performance. Besides creating an overall view on the accounting performance, this chapter also addresses the question how the results under both approaches correlate, i.e. to what extent the results under both approaches are consistent and could therefore serve as substitutes. For this purpose, the post-merger change in performance of acquirers is firstly determined based on a total of six performance indicators. Afterwards, the performance is normalized against a sample of non-merging peers from the same industry. As with the capital market performance in Chapter 3, Chapter 4 also analyzes the observed operating performance impact for determinant variables and tests for their significance with a cross-sectional regression model. In line with the primary objective of this thesis, it intends to develop a consistent perspective on the value creation potential of M&A across different data bases. In a later step, Chapter 4 addresses its second research question by determining a correlation between accounting- and event-study results in order to advance the methodological discussion on the substitutability of both methodologies. This section also differentiates results between different performance indicators and analyzes potential capital market premiums awarded for a correctly anticipated longterm profitability development.

1.2 Course of Analysis

7

Chapter 5 validates the derived empirical findings of Chapters 3 and 4 against a case study from the industry, namely the acquisition of Siemens VDO by Continental AG in July 2007. The main questions addressed include whether an industry leader of significant size can complete a consistently successful transaction across multiple analytical approaches and what characteristics contribute to Continental's present situation. In a first step, this chapter examines the motivation behind Continental's takeover of Siemens VDO and evaluates the motives against the present industry trends and transaction background. Afterwards, it assesses the short- and long-term post-merger wealth effects of this transaction on Continental's capital market and operating performance. Besides assessing the absolute value impact, the chapter focuses on comparing Continental's performance against a number of different peer benchmarks. By replicating the methodological approaches of the preceding chapters, this case study afterwards confirms the main performance drivers previously determined in the cross-sectional regressions. Overall, it provides an additional perspective on which M&A strategy potentially leads to long-term post-merger success within the automotive supply industry. Chapter 6 consolidates the main findings from the three empirical studies, identifies interrelations and determines key success factors for long-term success of M&A within the automotive supply industry. The study concludes with an outlook and presents potential areas for further academic research.

2

Research Foundations The following chapter provides a brief overview of the research foundations re-

lating to the three main components of this study's title, namely the automotive supply industry, mergers and acquisitions, and the measurement of M&A success. Besides providing fundamental definitions, this chapter also intends to define the scope of the following analyses and to create a basic understanding not only for the current situation within the automotive supply industry but also for the theoretical discussion around the measurement of long-term value creation through M&A transactions. Overall, it becomes apparent that the automotive supply industry provides a uniquely competitive environment in which automotive suppliers are increasingly seeking M&A activity as a response strategy. For determining M&A success, especially the consideration of a longer time horizon requires a differentiated use of various available methodologies.

2.1

The Automotive Supply Industry

2.1.1 Definition and Scope Although preceding literature comprises a number of different attempts to define the term 'automotive supplier,' an unambiguous definition has only recently been developed. According to Mentz (2006), an 'automotive supplier' is defined as any economic entity directly or indirectly delivering products and/or services to car producers, socalled OEMs, in order to be included in the production process of automobiles or eventually become part of the automobile itself. Since the focus of this thesis likewise lies in analyzing a comprehensive and global sample of M&A transactions in the automotive supply industry, adopting this definition for its purpose carries two major advantages. On the one hand, it extends the scope of the automotive supply industry to service companies enabling this study to create an even more comprehensive perspective on the specific long-term value creation potential within this industry. On the other hand, it

10

2 Research Foundations

confines the scope of the industry to genuine suppliers. Consequently, OEMs and the raw materials industries are excluded under the applied definition; both represent their respective ends of the observed automotive value chain. Based on the depth of the respective value added, the differentiation of automotive suppliers into different supplier tiers has become common and relates back to the pyramid-shaped supply chains developed mainly in the Japanese automotive and electronics industry. First-tier suppliers produce the highest value added by delivering subassembled units or modules (e.g. complete seats or transmissions) to OEMs. Second-tier suppliers produce less complex components usually delivered to first-tier suppliers for inclusion into their sub-assembled units. Third-tier suppliers produce standardized or basic products requiring a minimum of additional logistical services or planning (Von Corswant, Wynstra, and Wetzels (2003)). While this pyramid structure enables automotive producers to simplify their material flows, it also decreases the coordination efforts for producers towards their suppliers: first-tier suppliers coordinate their second-tier suppliers, second-tier suppliers the third-tier, and so on (Kamath and Liker (1994)). Consequently, manufacturers are able to reduce their direct relations to a few first-tier suppliers while the suppliers coordinate themselves down the supply chain (Von Corswant and Fredriksson (2002)). 2.1.2 Current Trends and Challenges This section outlines trends and challenges influencing automotive suppliers over the last twenty years. It aims to create an understanding for the competitive pressures promoting the continuous consolidation within the automotive supply industry. As the competitive environment for suppliers is by definition strongly related to their customers, the main set of trends relates to the customer-supplier relationship and the sourcing behavior of car producers. These trends include globalization, outsourcing, shorter product life cycles and an increasing degree of sophistication in the customersupplier relationship.

2.1 The Automotive Supply Industry

11

Although recent literature finds that the transfer of production and product development activities by car producers into an increasing number of countries has recently slowed down, car producers are still enforcing globalization by other means (Von Corswant and Fredriksson (2002)). Main examples include the global use of common automotive platforms in different models as well as a growing interest of OEMs to source supplies from the same supplier on a world-wide basis (Sadler (1999)). In order to meet their customers' demands regarding just-in-time delivery as well as local regulative requirements including customs and in-country quotas, the pressure on automotive suppliers to create a costly global presence is growing. As a result, the average number of countries in which a sample of suppliers maintains production facilities increased from 5.8 in 1988 to 15.0 in 2003. Product development activities of the same group took place in 5.9 countries in 2003 as opposed to 2.6 in 1988 (Von Corswant and Fredriksson (2002)). Wherever financially affordable, it appears that automotive suppliers are following the internationalization of their customers either by geographical expansion or cross-border acquisitions (Abrenica (1998), Sadler (1999)). A second trend frequently addressed in academic literature is the growing degree of outsourcing (Mercer (1995), McIvor, Humphreys and McAleer (1998)). Automotive producers increasingly seek to outsource parts of their production facilities and to purchase full systems of components from their suppliers rather than individual parts (Sadler (1999)). Von Corswant and Fredriksson (2002) show that OEMs were expected to acquire more than 65 per cent of their total turnover in 2003 as purchased materials. While exposing them to the risk of losing control over the car as a whole, OEMs still benefit from outsourcing a significant degree of their production to a limited number of suppliers. As a result, they are able to focus their coordination efforts on a few relationships with first-tier suppliers (McIvor et al. (1998)). For the suppliers, however, coordination efforts likewise increase as first-tier suppliers also outsource a growing part of their activities. In the end, first-tier suppliers are increasingly required not only to manage their customer demands up-stream, but also a growing number of lower-tier suppliers down the value chain.

12

2 Research Foundations

At the same time, product life-cycles within the automotive industry become shorter. As they converge towards a minimum product life-cycle length for the producers, the challenge for suppliers increases. On the one hand, suppliers are required to take over the product renewal responsibility and to upgrade their systems and components regularly. On the other hand, automotive producers are also increasingly transferring full product development tasks. While the producers' share of total product development resources averaged at around 70% in 1988, it dropped to approximately 60% ten years later (Von Corswant and Fredriksson (2002)). As a result, suppliers are acquiring valuable production expertise and product development capabilities in order to be capable of delivering innovative products at high frequency. The closer in the supply chain a supplier is situated to the automobile producer, the larger the degree of actual product development activity it shoulders (Von Corswant et al. (2003)). A last trend stems from a generally increasing degree of sophistication within the customer-supplier relationship. While reducing the number of relationships to direct suppliers, OEMs try to establish a collaborative partnership with a limited number of first-tier suppliers. As a result, the supplier is required not only to fulfill its customers' demands but also to actively engage in subcontracting, adhere to just-in-time logistics, and meet legal and warranty requirements. As the first-tier suppliers are expected to exhibit constant cost-reductions on a year-to-year basis, lower-tier suppliers are increasingly struggling to adhere to cost and contractual requirements (McIvor et al. (1998)). Besides these trends concerning the sourcing behavior of OEMs, the intensity of competition among suppliers also increases. As producers try to source complete systems from a limited number of first-tier suppliers, supply companies develop an increasing tendency towards specialization on particular products or segments. Some leading firms have even become inseparably connected with particular systems or technologies (Sadler (1999)). One relevant example for such a connection includes the speedometers of Siemens VDO, whose success story started with their prominent placement into the Volkswagen Beetle in 1939. Where this specialization strategy is successful, the level of competition is fairly low, due to the limited number of rivaling companies dominating

2.1 The Automotive Supply Industry

13

the product range (e.g., only a few players provide very complex products such as Xenon headlights). As a result, high barriers to entry exist almost ruling out new market entries into these product ranges. Where specialization is not able to create a competitive edge, the particularly high number of remaining suppliers creates a very strong competitive environment and fierce rivalry. This is the case, for example, in the leather and tire industries (Aktas et al. (2004)). Figure 2.1: M&A Transactions in the Automotive Supply Industry 25

3 2.5

20

2 15 1.5 10 1 5

Transaction Value (inflation adj., in USD billion)

2007

2005

2003

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

0

0.5 0

Avg. Transaction Value

Source: Thomson One Banker, Bloomberg M&A database, own calculations

Lastly, automotive suppliers are also facing decreasing profit margins. While OEMs expect constant cost reductions decreasing the suppliers' sales volume (McIvor et al. (1998)), increasing prices for raw materials raise their costs of goods sold. Although the slowing down US economy and the weak US Dollar offset some of the price effects on suppliers, especially for European and Asian companies, the majority of raw materials reached record prices in 2007: For example, the average price of crude oil rose by 11%, the price of processed metals increased by 9% and of natural rubber by 10%. The majority of other raw materials including copper, steel, and nickel experienced similar

14

2 Research Foundations

price increases with different volatilities (Continental (2007b)). In connection with the increasing pressure from customers and competitors, these costs for raw materials put a strong strain on the profit situation of automotive suppliers. As a result, many suppliers have realized losses or significant profit reductions over the first years of this century. Between 2000 and 2002, for example, many suppliers suffered from significant profit reductions of up to 50% (Fitzgerald (2002)). Given these challenging industry conditions, mergers and acquisitions represent a common response strategy for automotive suppliers in order to offset some of the challenges described. Although consolidation still remains far lower among automotive suppliers than among OEMs, consolidation activity has significantly increased among suppliers over the last two decades, especially in Europe, Northern America and Japan/South Korea. It is estimated that the number of direct suppliers in Europe, for example, dropped from 10,000 in the early 1970s to 3,000 in 1995 and to about 500 in the year 2000 (Sadler (1999)). Figure 2.1 emphasizes this consolidation activity and shows how a strong merger wave affected the automotive supply industry during the 1990s. Between 1991 and 1999, significant M&A transactions (with a transaction value of more than USD 50 million) steadily increased both in total number as well as in total inflation-adjusted transaction volume. A closer look at the average transaction values reveals that they peaked three times over the last 30 years: once during an early consolidation wave in the 1980s, once during the merger wave of the 1990s and once just recently with the USD 15.7 billion transaction of Continental AG. Consequently, it is likewise reasonable to assume that, instead of just one merger wave in the 1990s, consolidation in the automotive supply industry follows a continuing activity pattern with the 2007 Continental acquisition of VDO potentially representing the starting point of the next significant consolidation wave.

2.2 Mergers and Acquisitions

2.2

15

Mergers and Acquisitions Especially in US-American academic literature, 'mergers and acquisitions' have

developed into a widely-used collective term representing all corporate transactions in the course of which ownership and control of a firm are transferred from one hand to another (see Weston, Chung, and Siu (1990)). Jensen and Ruback (1983) describe the resulting market as a "market for corporate control", in which managers and management teams are actively competing for their right to control corporate resources. Alongside corporate expansions, these transactions in a broader sense also include corporate disposals as well as structural changes in the corporate control, corporate ownership or governance of a firm (Copeland and Weston (1988)). Corporate expansions are regarded as mergers and acquisitions in a narrower sense comprehending all transactions forming one economic unit from two or more previous ones (Weston et al. (1990)). As the purpose of this thesis is to determine long-term success of M&A as an attempt to realize synergy potentials in the automotive supply industry, this study focuses on expanding transactions (M&A in a narrower sense). Mergers and acquisition in a narrower sense can be distinguished based on the future legal status of the target and the degree of codetermination on the target side.1 In a merger, the management boards of both sides agree to combine their firms and jointly prepare a respective proposal to their shareholders. After the shareholders' agreement, one of the merging firms usually ceases to exist as a legal entity and becomes part of the other. As a special sub-form, a consolidation represents a merger in which both companies cease to exist and shareholders of both sides receive shares in a newly found corporate entity. By contrast, the legal status of the target usually remains unaltered in a corporate acquisition. Acquisitions are carried out either as share deals, in which the acquirer directly addresses the shareholders of the target to tender their shares, or asset deals, in which one firm acquires the assets of another. While a tender offer usually bypasses the management board of the target, no formal agreement of the shareholders is 1

The following description is based on Damodaran (2004).

16

2 Research Foundations

needed. In an asset deal, however, the shareholders of the target have to formally approve the acquisition. Once the assets are transferred, the target firm can eventually be liquidated. For the purpose of this thesis, the described focus on expanding transactions is again restricted to only those transactions involving the transfer of majority control, i.e. where the acquirer gains at least 50% of the outstanding shares or private equity. These transactions represent the most comprehensive form of mergers and acquisitions and are expected to have the most profound impact measurable in form of a capital market reaction. For the remainder of this study, the terms 'mergers and acquisitions,' 'corporate acquisitions,' 'consolidation activity,' 'takeover transactions,' 'deals,' and 'corporate combinations' are used interchangeably and all refer to the described expanding transactions involving the transfer of majority control.

2.3

Measuring Success of Mergers and Acquisitions

2.3.1 Time Horizons One challenge in measuring the success of mergers and acquisitions lies in addressing success over different time horizons. While preceding literature distinguishes between short-term announcement returns and long-term abnormal returns, different statistical and methodological issues influence the observable performance under both time horizons. In addition, the past only produced common agreement on the measurement methodology assessing short-term announcement returns (Brown and Warner (1980)). The methodology to assess long-term abnormal returns is constantly refined and discussed in recent publications (see Mitchell and Stafford (2000)). Since this study focuses on the potentially more challenging measurement of long-term success, the following section briefly describes the main available methodologies for both time horizons and their inherent problems.

2.3 Measuring Success of Mergers and Acquisitions

17

While finding frequent application in the past, the event-study methodology usually serves as the means for measuring short-term M&A success. It determines to what extent the share price performance of a transaction partner is considered abnormal, i.e. different from what is expected (Brown and Warner (1980)). This approach builds on and, at the same time, challenges the Efficient Market Hypothesis (EMH), which assumes new information to be incorporated promptly and correctly into return information (Bruner (2002)). One common approach to determine expected returns lies in the two-factor capital asset pricing model, a market model estimating expected returns on the basis of a market portfolio and a risk premium (Black (1972)). Since this eventstudy methodology in connection with the market model is superior to a number of more sophisticated approaches (Brown and Warner (1980)) and robust against ignoring autocorrelation and variance changes in daily data (Brown and Warner (1985)), preceding literature establishes this approach as the predominant methodology for measuring short-term success. Besides methodological agreement, the determined value creation potential in short-term event studies has also converged towards a common opinion. Corporate control transactions in form of mergers or acquisitions are in general associated with significant wealth creation for the combined entity. However, this value creation is unevenly distributed between the shareholders of the target and the acquirer. While targets realize significant positive announcement returns of up to thirty per cent (Jensen and Ruback (1983)), the returns to acquirers are essentially zero (Bruner (2002)); Loughran and Vijh (1997) even conclude that acquirer returns are overall negative or, at most, insignificantly positive. Given the synergy potential in the automotive supply industry, this industry is clearly an outlier as acquirers are able to realize significant positive abnormal returns of +1.6% in the 11-day event period surrounding the announcement date (Mentz and Schiereck (2008)). By contrast, the measurement of long-term M&A success is subject to a number of theoretical pitfalls originating from analyzing longer time periods, the choice of vari-

18

2 Research Foundations

ous performance benchmarks, and the applied statistics. As abnormal performance requires measurement against a benchmark, a longer time period automatically raises concerns about a potential new listing and rebalancing bias within the benchmark (Barber and Lyon (1997)). If abnormal performance is determined against a benchmark portfolio of non-merging firms, the results may be skewed, auto-correlated or exposed to heteroscedasticity. In the past, research focused on applying a combination of event-time and calendar-time approaches in order to overcome these various biases and problems. Buy-and-hold Abnormal Returns (BHARs) represent the most commonly-used methodology in determining long-term performance in event-time (see, for example, Ritter (1991), Barber and Lyon (1997), and Lyon, Barber, and Tsai (1999)). The BHARs are determined by normalizing the investment experience of an investor in an acquiring firm against the investment in a non-merging benchmark. The benchmarks are commonly determined by a character-based matching procedure based on market capitalization and market-to-book ratios (Lyon et al. (1999)). By carefully matching acquiring firms with reference portfolios or control-firms and by applying advanced test statistics on a sufficiently large sample, this method can overcome its inherent statistical pitfalls such as the new listing bias, the rebalancing bias and skewness (Lyon et al., (1999)). However, it is still exposed to significant cross-correlation, especially due to the application of not randomly selected matching samples. The Fama-French-3-Factor-model (FF3F) as a common representative of the calendar-time approaches eliminates the cross-correlation problem (Fama and French (1993)), while at the same time being exposed to other potential pitfalls such as heteroscedasticity. It determines abnormal returns of acquiring companies by regressing a time series of acquirers' excess returns (return less risk-free rate) on the time series of market excess returns, the time series of the difference in returns of small and big companies, and the time series of the difference in returns of companies with high and low marketto-book values (Fama and French (1993)). All in all, it becomes apparent that a comprehensive assessment of the long-term capital market performance of acquirers in the

2.3 Measuring Success of Mergers and Acquisitions

19

automotive supply industry requires a differentiated and sophisticated methodological approach in order to arrive at valid conclusions about the general value-creationpotential within the industry. 2.3.2 Research Approaches Besides considering different time horizons in assessing the success of mergers and acquisitions, the variety of available research approaches also contributes to the complexity of long-term M&A analysis. Bruner (2002) describes four main research approaches for determining a perspective on M&A success and profitability: event studies, accounting studies, executive surveys, and clinical studies. The author also concludes that a comprehensive assessment of M&A success must look for patterns of confirmation across approaches and studies. Consequently, this study makes use of a variety of different approaches as it claims to provide a comprehensive view on the situation within the automotive supply industry. The following section provides an overview of the most frequently used research approaches within the field. Event studies represent the most commonly used research approach that dominates the field since the 1970s (Bruner (2002)). Event studies build on share return information and examine the abnormal returns to shareholders around or after the announcement of a transaction. As pointed out in the previous section, event studies can either have a short-term or a long-term focus. Abnormal returns are determined by comparing the returns of an acquirer or target to a benchmark return, usually the expected returns as determined by the capital asset pricing model in short-term event studies or the benchmark returns of a matching firm in long-term event studies. Event studies are generally forward-looking and build on the efficient market hypothesis: it is assumed that share prices represent the present value of expected future cash flows to shareholders (Bruner (2002)). However, given the subjectivity in assessing future cash flows, event studies are also exposed to the risk that some information may not be correctly incorporated into

20

2 Research Foundations

the share price (Eberhart, Maxwell, and Siddique (2004)). In addition, short-term eventstudy results are sensitive to the event and estimation periods chosen; while abnormal returns are usually mitigated in larger event-windows, larger estimation periods are likely contaminated by other confounding events (Rhoades (1994)). The results obtained in long-term event studies, however, are often subject to the respective long-term methodology and the benchmark chosen. As pointed out in the previous section, the later requires a comprehensive methodological approach to overcome inherent statistical biases. Accounting studies represent the second research approach frequently applied in assessing value creation through M&A. These studies examine reported financial results and answer the question whether acquirers outperformed their non-merging peers. While focusing on performance indicators or general balance sheet structure, accounting studies are therefore primarily backward-looking and build on accredited published accounting statements. While published data carries credibility to the reader, a general comparability of published data across different years and reporting standards does not always apply (Bruner (2002)). To overcome the shortcomings inherent in the two broad research approaches, preceding literature commonly applies case or clinical studies as a third complementary approach. By focusing on a single transaction, they facilitate a more detailed analysis of outstanding transaction phenomena that usually exceed the scope of large-sample event or accounting studies (Kaplan, Mitchell, and Wruck (1997)). Case studies generally follow a structured approach inducing new insights from a detailed qualitative description of a real transaction and an evaluation of its performance along a number of pre-defined criteria that can be either quantitative or qualitative in nature (Eisenhardt (1989)). As a last research approach, Bruner (2002) presents executives surveys in which generalizations from a limited number of executive questionnaires are drawn. Although the respective studies are generally based on standardized questionnaires, a similar sampling bias applies as present in executive interviews (Kaplan et al. (1997)). As mainly

2.3 Measuring Success of Mergers and Acquisitions

21

acquirers in successful transactions are willing to generously provide information, either in interviews or in questionnaires, this study refrains from applying this last approach. Instead, the following chapters will focus on providing a comprehensive overview of the long-term value creation through M&A in the automotive supply industry using a comprehensive mix of event-study, accounting-study, and case-study methodologies.

3

Study 1: Determinants of Capital Market Performance2

3.1

Introduction Over the last decades, wealth creation through mergers and acquisitions has

been extensively discussed in empirical finance research. With a variety of different approaches and foci, authors focusing on short-term announcement effects unambiguously conclude that mergers and acquisitions seem to create value. However, as this short-term value creation potential is mostly attributed to the shareholders of the target companies (Bradley, Desai, and Kim (1988), p.31), a closer look at the returns to acquiring firms reveals a different pattern: While short-term announcement returns for acquiring companies average at around zero (Bruner (2002)), long-term post-merger returns even indicate significant value losses on the acquirer side.3 In the light of ongoing merger activity and consolidation, these negative reactions pose a challenge to the management of merging companies and call for a comprehensive list of determinants for the direction and magnitude of the experienced value effect. Towards the turn of the last century, many industries including the automotive supply industry were facing increasing merger activity. The pressure to produce better equipped and less expensive automobiles created a growing trend towards specialization and internationalization of the industry. While some product ranges such as braking components are now dominated by very few players, acquisitions and foreign direct investments also led to geographical expansion of players across borders and across continents (Sadler (1999)). In the light of these specific industry characteristics, previous research in the automotive supply industry shows that acquiring companies realize significant positive short-term returns as an expression of the global synergy and efficiency potential underlying the transactions (Mentz and Schiereck (2008)). 2

3

An excerpt of this chapter has already been published in the Journal of Economics and Finance; see Laabs and Schiereck (2009). See, for example, Agrawal, Jaffe, and Mandelker (1992), who report a significant value loss of about 10% to acquiring companies over a five-year period following merger completion.

24

3 Study 1: Determinants of Capital Market Performance

However, the question, whether acquirers are able to sustain these exceptional positive returns beyond a short-term announcement window, remains open. Due to various methodological difficulties associated with studies on long-term abnormal performance, evidence from other industries remains scarce or narrows its focus on either one of the time horizons (short- vs. long-term) or on a single deal characteristic (method of payment, cross-border). In contrast to previous research, this study determines the shortand long-term performance of acquiring firms in the automotive supply industry, it uses a combination of Buy-and-Hold Abnormal Returns and expected returns based on the Fama-French-3-Factor-model to determine statistically reliable indications of long-term performance, and it analyzes a comprehensive list of transaction and acquirer characteristics for their respective impact on the short- and long-term wealth effect. The objective of this study is twofold: After updating previously published announcement effects on acquirers in the automotive supply industry, the short-term perspective is expanded by long-term abnormal returns based on event-time (BHARs) and calendar-time (FF3F). Then the observed return patterns are examined to detect and categorize determinant variables. The main focus lies in determining an effect based on the differences in geography, product range, transaction size, and bidder experience. To support the findings, a regression analysis is used to determine correlations and to test for statistical significance. The remainder of this study is organized as follows: Section 3.2 provides a brief overview of the relevant literature and outlines the derived hypotheses for the short- and long-term value creation effects on acquiring companies in the automotive supply industry. Section 3.3 presents the applied methodology as well as the sample selection procedure. The following Section 3.4 contains the empirical results including the overall short-term announcement effects and the long-term performance as well as the analyses concerning different value drivers. Section 3.5 complements these results with a crosssectional analysis before Section 3.6 summarizes the findings and concludes.

3.2 Literature Review and Hypotheses

3.2

25

Literature Review and Hypotheses

3.2.1 Related Literature A significant number of studies have examined the short-term announcement effect and the long-term capital market performance of acquiring firms. For the short-term announcement effect, previous M&A research has found that corporate control transactions in form of mergers or acquisitions are in general associated with significant wealth creation. Jensen and Ruback (1983) summarize early findings from the 1950s to 1970s with positive returns to bidders and targets: On average, they find positive cumulative abnormal returns (CARs) of +29.09% to target shareholders and +3.81% to bidder shareholders, both within the month of a successful merger or tender offer announcement. However, research has also shown that positive returns to acquirers are decreasing over time: Bradley et al. (1988) describe how the average abnormal return for bidders falls from a significant +4.1% in the 1960s, to +1.3% in the 1970s, and to a significant 2.9% in the 1980s. More recent studies support these findings and demonstrate a more negative effect for acquiring companies: using a sample of 4,256 merger events between 1973 and 1998, Andrade, Mitchell and Stafford (2001) find a negative but insignificant -0.7% return for acquiring companies within a three-day event-window surrounding the announcement date. Bruner (2002) provides a comprehensive overview of 44 studies investigating abnormal returns to acquiring companies and summarizes abnormal returns to be essentially zero; Loughran and Vijh (1997) conclude that acquirer returns are overall negative or, at most, insignificantly positive. In the light of this conclusion, previous findings from the automotive supply industry clearly identify the industry as an outlier. Using a sample of 201 M&A transactions in the automotive supply industry between 1981 and 2004, Mentz (2006) finds a significant positive abnormal return to acquiring companies of +1.6% in the eleven-day event period surrounding the announcement date. This result stands in clear contrast not

26

3 Study 1: Determinants of Capital Market Performance

only to the general results presented above but also to other event studies with industry focus.4 The author argues this finding to be the result of industry-specific synergy potentials perceived by capital markets: for the automotive supply industry, mergers and acquisitions seem to be a feasible measure to realize synergy- and efficiency-gains. Mentz (2006) also analyzes a number of underlying variables influencing the positive returns to bidders and concludes that bidder returns for automotive suppliers are positively influenced by a number of characteristics: an acquirer located in Northern America (vs. Asia and Europe), a transaction date between 1993 and 1998, a national transaction (vs. cross-border), deal financing without stocks and fully in cash, a public target, an acquirer of above average size, and an acquirer which has not completed another deal in the three years preceding the regarded transaction. These findings are discussed in more detail in the following section in order to derive appropriate hypotheses for the results of this study. The long-term post-acquisition performance of acquiring firms has been analyzed since the 1970s and can be characterized by three major research phases.5 Phase 1 contains the earlier work of the 1970s and 1980s, in which the analysis of long-term performance is mostly treated as a side-note to short-term event studies. Among the first to analyze long-term performance was Mandelker (1974): Based on a merger sample of 241 mergers, the author determines insignificant negative long-term abnormal returns of -1.4% over a 40-month period. By using the standard market model normally applied in short-term studies, Malatesta (1983) finds a significant -7.6% for 256 mergers in a 12month period following the merger event. Since the appearance of significant long-term abnormal returns contradicts the commonly-held Efficient Market Hypothesis (EMH), 4

5

Other event studies focusing on single industries mainly report negative abnormal returns to acquiring companies, for example -0.6% for electric utilities (Berry (2000)), -0.3% for telecommunications (Akdo÷u (2009)), and -0.1% for banks (Beitel, Schiereck, and Wahrenburg (2004)). Agrawal and Jaffe (2000) originally distinguish two major phases: Phase 1 represents the earlier work of the 1970s and 1980s and phase 2 the later, more advanced methodological research of the 1990s. However, since more recent studies apply additionally advanced methodologies, a third phase starting at the end of the 1990s is added to distinguish between the more recent work on methodologies and the methodological foundations of phase 2.

3.2 Literature Review and Hypotheses

27

the interest in long-term return behavior increased in the 1980s. Other studies following in this period include Barnes (1984), Bradley and Jarrell (1988), and Franks and Harris (1989). Agrawal and Jaffe (2000) provide a comprehensive overview of these various early studies. Franks et al. (1991) represent the turning point in the analysis of long-term postmerger performance and the start of phase 2. Their study was the first to be exclusively devoted to the long-term performance of acquiring firms. By introducing sophisticated benchmarks and by combining calendar- and event-time approaches, the authors clearly set themselves apart from the simple statistical approaches employed during phase 1. Nevertheless, based on their calendar-time approach, they do not find evidence for abnormal returns to acquirers and conclude that previous findings are the result of misspecification in the appropriate benchmarks. About a year later, Agrawal et al. (1992) find that the non-existence of abnormal returns is specific to the time period examined by Franks et al. (1991): Overall, Agrawal et al. find a significant negative abnormal return of 10% in the five years following a merger. Similar results are presented by Loderer and Martin (1992), Kennedy and Limmack (1996), Gregory (1997), and Rau and Vermaelen (1998): All find negative abnormal returns between -1% and -18% within two to five years after the transactions. In general, the studies of phase 2 can be characterized by an advanced methodology to arrive at abnormal returns; most of the studies employ character-basedmatching approaches based on size, risk and market-to-book ratios or expected-return models such as the ten-factor-model employed by Franks et al. (1991). Although these models are generally more reliable and sophisticated than the ones employed in phase 1, they still fail to address some of the problems that usually afflict long-run abnormal performance. For example, Barber and Lyon (1997) find that test statistics calculated on the basis of reference portfolios are subject to the rebalancing and new listing biases; they also document positive skewness in CARs and in BHARs, a finding that inhibits any inference on the basis of a normality assumption. Mitchell and Stafford (2000) ar-

28

3 Study 1: Determinants of Capital Market Performance

gue that the bootstrapping methodology used to infer from skewed abnormal returns is inappropriate. Since acquisitions are often clustered by industry, the underlying assumption of event independence is breached. Starting in the late 1990s, these methodological challenges are explicitly addressed by the advanced BHAR-approaches applied in the studies of phase 3. The first study associated with phase 3 is Loughran and Vijh (1997): Using a sample 947 transactions between 1970 and 1989, the authors find negative BHARs of -6.5% within a fiveyear time period following the transaction. They also document a difference between merger transactions and acquisitions: While mergers lead to negative returns of -15.9%, acquisitions yield positive returns of +43%. Likewise they find that stock-financed transactions yield negative returns while cash-financed transactions provide positive long-term abnormal returns to acquirers. Overall, Loughran and Vijh were the first to apply the advanced BHAR methodology in which they determine abnormal returns of acquirers on the basis of control-firms matched by market value and market-to-bookratio. Mitchell and Stafford (2000) calculate BHARs for a sample of 2,068 US transactions between 1961 and 1993 and find insignificant negative returns of -1% (equalweighted) and significant abnormal returns of -3.8% (value-weighted) over a three-year period. Since additional calculations using a 3-Factor model only yield significant returns with equal-weights, they conclude that their BHAR-results might be overstated. Bouwman, Fuller, and Nain (2009) show that significant long-term returns depend on the valuation of the market. While acquirers in highly-valued markets experience significant positive abnormal short-term returns and negative long-term returns, acquirers in lowly-valued markets experience short-term insignificant positive reactions and longterm insignificant positive returns. However, the majority of the more recent studies in this field point to negative abnormal returns. Significant abnormal returns are mostly derived using BHAR-approaches while calendar-time approaches yield mixed results.

3.2 Literature Review and Hypotheses

29

Other examples of negative long-term returns are provided by Black, Carnes, and Jandik. (2001), Sinha (2004), and Gregory and Matatko (2004). 3.2.2

Hypotheses

3.2.2.1 The Overall Effect Previous research reports significant positive short-term announcement effects to acquirers in the automotive supply industry and thereby identifies the industry as a strong outlier compared to other industries (Mentz and Schiereck (2008)). Since this study is based on transactions within the same industry, it is also expected to determine positive short-term announcement returns. From a long-term perspective, the majority of preceding studies points to negative abnormal returns, even if the particular industry benefits from regulated and protected profit margins such as the telecommunications industry (Ferris and Park (2002)). Therefore, it is hypothesized that acquirers in the automotive supply industry realize negative long-term abnormal returns as well. Given the positive short-term return behavior, the long-term effect is assumed to be slightly negative and insignificant. H3.1: Acquirers in the automotives supply industry experience positive announcement effects, which carry over as an insignificant long-term underperformance. 3.2.2.2 The Impact of Transaction Characteristics The following two sections contain a number of hypotheses concerning the impact of transaction and acquirer characteristics on both the short-term announcement effect as well as the post-merger capital market performance of acquirers. Since one focus of this study lies on detecting long-term performance in the automotive supply industry and since long-term performance is often argued to be a function of long-term post-merger management and synergy realization, this section focuses on variables assumed to directly influence the available synergy potential, namely geographic expan-

30

3 Study 1: Determinants of Capital Market Performance

sion, product diversification, transaction size, product sectors, and bidding experience. Other variables such as timing, public status of the target, and acquirer continent are also analyzed but omitted from the following univariate analysis. Table 3.1 provides an overview of the derived hypotheses and the assumed incremental effects. Table 3.1: Overview of Hypotheses and Predicted Value Impact on Acquirers Predicted Value Impact On Acquiring Companies Overall Effect Incremental Effect of Transaction Characteristics Geographical Expansion Cross-border deal Cross-continental deal Diversification Transaction Size Incremental Effect of Acquirer Characteristics Product Sector Engine Drive Components Electrical Components Acquisition Experience Multi Acquirer

Hypothesis 3.1

ShortTerm Positive (significant)

LongTerm Negative (insignificant)

None Negative Negative Positive

Negative Negative Negative Positive

Negative Negative Negative

Negative Negative Negative

Negative

Positive

3.2

3.3 3.4 3.5

3.6

A large number of studies focus on the effect of internationalization on shortterm value creation for acquirers, but yield mixed results. While some conclude that geographical expansion in form of cross-border transactions yield positive returns for acquirers (Goergen and Renneboog (2004)), others argue that these effects are characteristic to certain industries and countries (Dewenter (1995), Kiymaz and Mukherjee (2000)). For the automotive supply industry, Mentz and Schiereck (2008) do not find significant short-term return differences between national and cross-border transactions for acquirers while transcontinental deals have a negative influence. The literature on long-run performance is less ambiguous and unanimously points to long-run negative returns to acquiring companies in cross-border transactions due to the more challenging post-merger integration and imperfect information (Conn, Cosh, Guest, and Hughes (2005), Aw and Chatterjee (2004), and Black et al. (2001)).

3.2 Literature Review and Hypotheses

31

H3.2: Cross-border transactions yield no significant impact on short-term announcement returns, but will significantly decrease long-term performance. In the past, activity diversification has normally been followed by a negative impact on long-term performance (Agrawal et al. (1992), Ferris and Park (2002)). These observed return patterns can be argued to be the result of a more difficult synergy realization process in diversifying mergers and acquisitions. Given the complexity and competition patterns in the automotive supply industry, where various suppliers produce a wide range of products from high-tech electrical components to tires, it is assumed that synergy- and efficiency-gains can be realized more easily in focusing transactions. Diversification into other product lines will have a negative impact on returns. H3.3: Product diversification negatively impacts short- and long-term returns to acquiring companies. The effect of transaction size on value creation for acquiring companies is regarded less extensively in prior research. Ferris and Park (2002) argue that larger transactions are more likely to result in economies of scale in research and production facilities and present corresponding supportive results for a long-term study in the telecommunications industry. As the automotive supply industry can primarily be regarded as a production industry, it is assumed that it shows similar reactions as previously determined for the telecommunications industry: The larger the transaction, the better the economies of scale and the stronger the positive influence on returns. H3.4: Larger transactions positively influence short- and long-term returns. 3.2.2.3 The Impact of Acquirer Characteristics The automotive supply industry contains a number of different product segments. Competitiveness within each segment strongly depends on the individual prod-

32

3 Study 1: Determinants of Capital Market Performance

uct: As few suppliers deliver very complex products, for example Xenon-headlights, competition in this particular segment can be regarded as low. In other segments such as the tire or leather supply segment, competition is significantly higher due to a higher number of producers (Aktas et al. (2004)). Furthermore, some industry segments including suppliers of tires, forgings, and bearings suffer from over-capacity which in turn increases the degree of rivalry experienced within these segments (Carr (1993)). Therefore, it is inferred that the effects of M&A activity on acquirers within the automotive supply industry exhibit different magnitudes depending on the observed product sector of operations. Acquirers in product segments that are exposed to stronger competition and overcapacity are assumed to realize synergy gains more efficiently than other acquirers. Therefore, they are expected to experience a more positive impact on short- and long-term returns. H3.5: Takeovers involving acquirers from the exterior, interior or tires product segments yield positively influenced short- and long-term returns. The last acquirer characteristic to be analyzed here is the acquisition experience of the acquirer. Haleblian and Finkelstein (1999) determine a positive correlation between the squared number of completed transactions of an acquirer and the magnitude of the cumulative abnormal returns experienced in short-term event-windows. They argue that acquirers gain target-integrating experience that they can leverage in additional transactions. Fuller, Netter, and Stegemoller (2002), however, provide evidence that acquirers with more than five transactions within a three-year period before the observed transaction yield significant negative returns. For the automotive supply industry, acquirers without any bidding experience within the preceding three years of a transaction experience significantly higher positive announcement returns than experienced acquirers. Whether this observation can be attributed to management hubris or other explanations remains unanswered (Mentz (2006)).

3.3 Data and Methodology

33

Previous findings from the long-term perspective support the argument led by Haleblian and Finkelstein: Antoniou and Zhao (2004) analyze a sample of 179 transactions by UK acquirers between 1991 and 1998 and find a significant positive influence of multi-bidders on post-merger capital market returns. While multi-bidders outperform the original sample average, single-bidders underperform with respect to the original sample performance. Therefore, it is assumed that the short-run outperformance of single-bidders in the automotive supply industry as described by Mentz (2006) holds true for the announcement effects, but fails to persist in the long-run. Over three years, the integration experience of multi-bidders will yield more positive returns and outperform the single-bidders. H3.6: Multi-bidders experience lower short-term announcement effects, but outperform single-bidders in the long-run.

3.3

Data and Methodology

3.3.1

Identifying Merging Companies The sample of mergers and acquisitions for the empirical event study is drawn

from the Securities Data Company (SDC)/Thomson One Banker Deals database and the Bloomberg M&A database. It includes all takeover events announced between January 1st, 1981, and September 1st, 2007. The total number of M&A deals is reduced to yield only those transactions meeting the following criteria.

At the time of the takeover, the target and acquirer company both possess active operations in the automotive supply industry.

The acquirer intends to purchase 50% or more of the outstanding shares or of the private equity, for publicly traded and privately held targets respectively.

The total transaction value accumulates to at least USD 50 million.

34

3 Study 1: Determinants of Capital Market Performance

The acquiring company is located in one of the following geographic regions: Europe, North and South America, and Asia. In addition, all targets and acquirers are double-checked by a press research us-

ing the Factiva database to ensure that all transactions are horizontal and that announcement dates are correct as provided by the databases. All non-horizontal deals as well as deals involving financial investment companies are excluded. The described selection criteria result in a final takeover sample of 230 events in the automotive supply industry between the years 1981 and 2007. Table 3.2 provides an overview of the frequency distribution over time and reveals a strong concentration of events between the years 1995 and 2000. Since other recent research provides evidence of significant merger clustering within certain industries during the 1990s (see Mulherin and Boone (2000), p.123), and since Andrade and Stafford (2004) find that fifty percent of all mergers within an industry occur within a five-year time period, the distribution over time is assumed to be representative and valid. The average transaction value ranges from USD 63.0 million to USD 2,819.0 million, whereas the later is mainly driven by the USD 15.8 billion acquisition of Siemens VDO by Continental AG in 2007. This transaction likewise represents the largest transaction volume in the sample. In order to determine short- and long-term abnormal returns, daily and monthly adjusted stock prices for all public targets and acquirers are downloaded from the Thomson Datastream database.6

6

To reflect the influence of dividend payments as well as share issuances or repurchases on return data, the adjusted stock prices denoted by data type “RI” were selected.

3.3 Data and Methodology

35

Table 3.2: Overview of the Transaction Sample – Descriptive Statistics7 Year

Transactions

(%)

Avg.Trans.Value (USD million)

1981 1984 1985 1986 1987 1988 1989 1990 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Total

1 3 1 8 2 6 9 5 4 4 5 12 19 26 27 30 11 13 5 8 9 9 7 6 230

0.4 1.3 0.4 3.5 0.9 2.6 3.9 2.2 1.7 1.7 2.2 5.2 8.3 11.3 11.7 13.0 4.8 5.7 2.2 3.5 3.9 3.9 3.0 2.6 100.0

63.0 99.2 597.8 323.7 375.0 813.8 351.5 200.6 188.1 438.3 161.5 265.4 383.1 285.0 680.0 620.0 338.0 316.2 249.9 174.2 280.6 219.1 553.4 2819.0 208.18

Acquirer Region - Number of Transactions Americas Europe Asia Other 1 3 1 6 0 4 4 1 2 4 5 8 16 17 19 16 5 7 3 3 4 4 3 2 138

2 1 1 4 4 2

3 2 9 8 10 5 4 1 1 4 2 1 2 66

1 1 1

1 1

3 1 2 1 3 1 3 3 2 22

1

1

4

3.3.2 Portfolio of Matching Firms Both methods for determining long-term abnormal returns as employed in this study require a set of comparable, non-merging firms from the same industry. Since there is no broad global index available for the automotive supply industry which comforts the purpose of the following analyses, this study determines a portfolio of matching firms from the constituents' lists of broader country-specific indices. This portfolio

7

For the remainder of the document, Asian-Pacific deals outside the automotive triad (North America, Europe, Japan/South Korea) will be presented as "Other Regions" (namely India and Australia).

36

3 Study 1: Determinants of Capital Market Performance

is then used to create an artificial industry index that serves as an input for the FF3Fapproach. The first step in constructing the portfolio of matching firms includes screening the constituents’ lists underlying the country-specific industry indices for companies in the automotive supply industry. Downloading all available constituents’ lists of indices in the automotive parts industry as supplied and constructed by the Thomson Datastream database results in 42 lists of international automotive suppliers.8 In a second step, all 42 constituents' lists are aggregated and compared to the respective lists from prior time periods. In a third step, the initial set of 109 companies derived from the Datastream database is cross-referenced to a list of the top 100 OEMs in the automotive supply sector as published on a yearly basis by the Automotive News (see Automotive News (2003)). Another 21 publicly traded automotive supply companies are added to yield a total number of 130 matching firms. The final portfolio contains 49 Asian, 46 American, and 35 European publicly traded companies. For each firm, daily and monthly stock returns as well as monthly market values and yearly market-to-book ratios are downloaded from the Thomson Datastream database. 3.3.3 Econometric Strategy 3.3.3.1 Short-term Methodology Short-term announcement returns are assessed using the event-study methodology in connection with the standard market model as described by Brown and Warner (1985). Although the choice of the applied model in a short-term context does not significantly influence derived results (see Brown and Warner (1980)), applying the standard model ensures direct comparability to the results from other industries: Both, Mentz (2006) for the automotive supply industry and the majority of other industry-

8

Constituents’ Lists are denominated by the industry-specific MNEMONIC code “LAUPRT” plus a two-digit country-specific suffix. For example, a list of U.S. automotive suppliers constituting the automotive parts industrial index in the US can be derived using the code “LAUPRTUS.”

3.3 Data and Methodology

37

related event studies make use of this approach (see Berry (2000) and Akdo÷u (2009)). Returns under the market model are determined on the basis of Equation (3.1).

Rit = α i + β i ⋅ Rmt + ε it

(3.1)

Rit and Rmt represent the returns in period t of security i and of the market portfolio respectively. İit stands for the zero-mean disturbance term, which is commonly referred to as the abnormal return. On the basis of this model, abnormal returns are determined as described by Equation (3.2).

ARit = Rit − (α i + β i ⋅ Rmt )

(3.2)

As the return of the market portfolio within the model generally refers to a market index associated with the given securities over time, local indices are determined for each country represented in the takeover sample (see Coutts, Mills, and Roberts (1994)). For example, the S&P 500 Composite return data applies for U.S. American companies, return data of the FTSE-All Shares index for companies located in Great Britain, and the DAX 30 index for German companies within the sample. Using different indices for each represented country accounts for regional differences in industry-returns and country-specific risk profiles. The market models are estimated by using Ordinary Least Squares (OLS) regression over a 200-trading-day period starting at trading day t = -250 relative to the earliest announcement date of the M&A event. On the basis of these estimated market model parameters, abnormal returns for all target and acquiring companies are derived.9 Cumulative abnormal returns are calculated as defined by Equation 9

In order to develop a perspective on total shareholder impact, the combined effect on an artificially combined entity is derived on the basis of market-value weighted returns: AR combined

;t

=

AR Acq , t * MV

Acq , t

+ AR Tar , t * MV Tar , t

MV

Acq , t

+ MV Tar , t

38

3 Study 1: Determinants of Capital Market Performance

(3.3). The longest event-window is 41 days: t=[-20;+20], with t = 0 being the announcement date of the transaction.

m2

CAREventWindow ( m1, m 2) =

(3.3)

¦ AR

it

t = m1

1 n ¦ CARP,i n i =1 t= σ (CARP ) / n

(3.4)

To test for statistical significance of the short-term abnormal returns, this study employs three test statistics. The first represents a simple parametric t-test as described by Equation (3.4). The sample average CAR is divided by the standard deviation across the individual company returns over the square root of the number of observations in the sample. This test statistic follows the student's t-distribution for a sample or subsample below 30 observations and a normal distribution when regarding a larger sample (see Barber and Lyon (1996), p. 373f.). The cross-sectional test as proposed by Boehmer, Musumeci, and Poulsen (1991) accounts for a potential event-induced increase in standard deviation by combining variance information from the event and the estimation period. Equation (3.5) describes the corresponding test statistic and Equation (3.5a) the respective standardization procedure to standardize abnormal returns with the standard deviation of returns in the estimation period.10

(3.5)

10

t=

1 N

N

¦ SAR

i ,t

i =1

N SAR § 1 i ,t ¨ SAR − ¦ ¦ i ,t N ( N − 1) ¨ N i =1 ©

· ¸¸ ¹

2

For a more detailed description see Boehmer et al. (1991), as well as Mikkelson and Partch (1988).

3.3 Data and Methodology

(3.5a)

SARi ,t =

39

CARi ,e1 −e2 SDE ,i ,t

;with

SDE ,i ,t = SDi ,t

§ · ¨ ¸ 2 − ( R R ) 1 ¨ ¸ M ,t M 1+ + ¨ T ¸ T 2 ¨¨ ¦ ( RM , j − RM ) ¸¸ © j =1 ¹

With: SARi,t = Standardized abnormal return for company i on day t SDE,i,t = Standard error of the event period for company i on day t SDi,t = Standard error of the estimation period for company i

T = Number of days in the estimation period RM,t = Market return on day t of the event period RM,j = Market return on day j of the estimation period RM = Average return of the index in the event period Since prior research provides some evidence that non-parametric test statistics can be more powerful than parametric t-statistics (Barber and Lyon (1996), p.360), the non-parametric Generalised Sign Test (GST) as proposed by Cowan, Nandkumar, and Singh (1990) completes the test statistics applied in the short-term study. The GST tests whether the percentage of positive observations (p) within a sample is greater than the expected percentage of positive observations (p') (Equation (3.6)). The expected percentage of positive observations is derived on the basis of the abnormal returns in the estimation period (Equation (3.7)).11

(3.6)

11

z=

N ∗ p − E ( N ∗ p' ) N ∗ p'∗(1 − p' )

; with:

In addition, differences in means of two subsamples are tested for statistical significance with the methodology applied by Baradwaj, Dubofsky, and Fraser (1992) and Siems (1996).

40

3 Study 1: Determinants of Capital Market Performance

N = Number of companies E = "Expectation Operator" P = Percentage of positive observations p' = Expected percentage of positive observations

(3.7)

p' =

1 n 1 E200 ¦ ¦ S i ,t ; with: n i =1 200 t = E1

Si,t = 1 for AR i,t >0 and AR i,t = 0 otherwise.

3.3.3.2 Long-term Methodology

To address the various statistical problems associated with the different methods of determining long-term capital market performance, a combination of the two most accepted methodologies is applied. BHARs represent the most commonly-used methodology in determining long-term returns in event-time (see, for example, Ritter (1991), Barber and Lyon (1997), and Lyon et al. (1999)). However, it is exposed to significant cross-correlation, especially due to the application of not randomly selected matching samples. The FF3F as a common representative of the calendar-time approaches eliminates the cross-correlation problem (Fama and French (1993)), while at the same time being exposed to other potential pitfalls such as heteroscedasticity. The BHARs are determined using a character-based matching approach, which individually matches non-merging counter-parts to the acquiring firms on the basis of market values and market-to-book ratios.12 The matching procedure follows the approach proposed by Lyon et al. (1999):

For each public acquiring firm in the takeover sample, the relevant size (market value) and market-to-book ratio is determined. Market values are determined as

12

Market values and market-to-book-ratios have been frequently used in prior research and appear to be the dominant matching criteria in the field. Examples include Lyon et al. (1999), Mitchell and Stafford (2000), Rau and Vermaelen (1998), and Loughran and Vijh (1997).

3.3 Data and Methodology

41

the last quote in the last June preceding the transaction, market-to-book ratios are derived for the last completed fiscal year before the transaction.13

Likewise, for each year, market values at the end of June and market-to-book ratios are downloaded for the full list of matching firms. For each acquisition, the total list of potential matches is then reduced to yield only those firms for which both descriptive data fields as well as monthly return data are available.

In a second step, the list of potential matches is reduced to those companies with a market value in the range of 70% to 130% of the acquirer's market value.

Finally, from the list of companies with a market value between 70% and 130% of the acquirer's market value, the one with the smallest absolute difference in market-to-book ratio is selected as the control-firm for the analysis. Based on this procedure, control-firms for each transaction and each acquiring

company between 1980 and 2004 are determined. These control-firms represent the benchmark in determining abnormal returns to acquirers. Equation (3.8) shows how the abnormal returns (BHARs) are derived as the difference between the Buy-and-Hold Return (BHR) of an investor in the acquiring company and the BHR of an investor in the control-firm. BHARs are determined for a period of 36 months following the takeover announcement. The average BHARs for the total sample as well as for the respective subsamples are calculated as an equal-weighted average and as a value-weighted average where the respective market value at the end of the preceding June serves as the weight. Statistical significance is tested using standard t-statistic.

13

All information is downloaded from the Thomson Datastream database.

42

3 Study 1: Determinants of Capital Market Performance

s +T

(3.8)

s +T

BHARi ,T = ∏ (1 + Ri ,t ) −∏ (1 + Rcontrolfirm,t ) ; with: t =s

t =s

BHARi ,T = BHAR of company i over T months Ri ,t = Monthly return of company i in month t Rcontrolfirm ,t = Monthly return of the control-firm in month t

The second model employed relates to the most commonly used calendar-time approach originally developed by Fama and French (1993). The "Fama-French-3Factor-model" (FF3F) determines abnormal returns of acquiring companies by regressing a time series of acquirers' excess returns (return less risk-free rate) on the time series of market excess returns, the time series of the differences in returns of small and big companies (SMB), and the time series of differences in returns of companies with high and low (HML) market-to-book values (see Equation (3.9)).

(3.9)

R p ,t − R f ,t = a p − b p ( RM ,t − R f ,t ) + s p SMB + h p HML + e p ,t

The acquirers' return Rp,t is determined as the average return of a continuously changing acquirer portfolio: Each acquirer is part of the acquirer portfolio for 36 months after the announcement of its respective acquisition, the portfolio in month t therefore contains all acquirers which were active in the preceding 36-month period. The average return across the portfolio is calculated with equal-weights as well as value-weights and yields two time series of acquirer returns in calendar-time. The one-to-three-year US treasury rate serves as an approximation for the risk-free-rate. The market return in the automotive supply industry RM,t is calculated as the artificial index return of the matching portfolio as described in Section 3.3.2. SMB represents the monthly differences in returns between small and big companies in the matching portfolio; HML determines

3.4 Empirical Results

43

the return differences between companies form the matching portfolio with high and with low market-to-book ratios. The time series of SMB and HML are determined using the procedure proposed by Fama and French (1993): In a first step, all matching companies are ranked based on their market value at the end of June of each year. Using the median as a separator, two portfolios are created, one with the small (S) and one with the big (B) companies. Likewise, the sample is ranked by market-to-book ratios and divided into three portfolios with the 30% highest (H), the 30% lowest (L), and the remaining 40% middle (M) market-to-book ratio companies. The derived portfolios are combined to yield a total of six portfolios (S/L, S/M, S/H, B/L, B/M, and B/H). The return differences between small and big companies are calculated as the difference between the simple averages of the three portfolios containing big companies (B/L, B/M, and B/H) and the three portfolios containing small companies (S/L, S/M, and S/H). For the individual portfolio returns, value-weighted averages are applied. The time series of HML are determined in a similar way, the monthly return differences are represented by the difference in simple means of the two portfolios with high market-to-book ratios and the two portfolios with low market-to-book ratios. The factors ap, bp, sp, hp, and ep,t are the results of the regression analysis; the intercept ap represents the average monthly abnormal return to the observed acquirers over the 36 months following an acquisition. Statistical significance of these factors is determined using standard t-statistic.

3.4

Empirical Results

3.4.1 The Overall Effect

Table 3.3 reports the short-term announcement effect of mergers and acquisitions on the total sample of acquiring firms in the automotive supply industry. Upon the immediate announcement of a transaction, acquirers earn a highly significant 0.98% in the [-1,0] event-window. In the 31 and 41 days around the announcement day, the value

44

3 Study 1: Determinants of Capital Market Performance

gain increases to approximately 2%. The majority of returns are highly significant at the 1%-level. On the one hand, this finding confirms prior research in this industry and, on the other hand, the exceptional position of the industry: Unlike results from the majority of other industries, positive short-term returns to acquirers represent the capital market's perception of extraordinary synergy potentials in the automotive supply industry. As a result, Figure 3.1 shows that all participating deal parties gain from this perception and realize positive abnormal returns upon the announcement of a deal.14 Table 3.3: CAARs to Acquirers Acquirers (n=206) Event-

t-Test

z-Test

Generalized Sign Test

Window CAAR t-value p-value z-value p-value P z-Value p-value [-20,20] 2.23% 1.94 0.03 ** 2.28 0.01 ** 52% 1.76 0.04 ** [-20,10] 2.03% 1.99 0.02 ** 2.37 <0.01 *** 53% 2.18 0.01 ** [-10,10] 1.84% 2.13 0.02 ** 2.28 0.01 ** 50% 1.20 0.11 [-5,5] 2.32% 3.20 <0.01 *** 3.80 <0.01 *** 55% 2.60 <0.01 *** [-1,1] 1.76% 3.95 <0.01 *** 4.92 <0.01 *** 59% 3.86 <0.01 *** [-1,0] 0.98% 3.09 <0.01 *** 3.50 <0.01 *** 54% 2.46 <0.01 *** [0] 0.94% 3.48 <0.01 *** 4.65 <0.01 *** 56% 2.88 <0.01 *** [0,1] 1.73% 4.12 <0.01 *** 5.73 <0.01 *** 59% 3.72 <0.01 *** [0,5] 2.05% 3.78 <0.01 *** 4.72 <0.01 *** 58% 3.44 <0.01 *** [0,10] 1.27% 2.13 0.02 ** 2.50 <0.01 *** 51% 1.48 0.07 * [0,20] 1.46% 1.86 0.03 ** 2.13 0.02 ** 49% 0.92 0.18 This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in mergers and acquisitions in the automotive supply industry. It contains all public acquirers whose trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-test, the cross-sectional test as proposed by Boehmer et al. (1991) (z-Test), and the Generalized Sign Test as proposed by Cowan et al. (1990). P stands for the percentage of transactions in the sample that resulted in a positive abnormal return.

14

In order to gain a comprehensive view on the value creation in the automotive supply industry and to validate prior findings from Mentz (2006), Appendix 1, 2, and 3 contain an overview of abnormal announcement returns to targets and the combined entities. All results support prior findings and are not discussed further in the scope of this study.

3.4 Empirical Results

45

Figure 3.1: CAARs to Acquirers, Targets, and Combined Entities 18% 16% 14%

CAAR

12% 10% 8% 6% 4% 2% 0% -20 -18 -16 -14 -12 -10 -8 -6 -4 -2

0

2

4

6

8 10 12 14 16 18 20

Days relative to announcement date Acquirers (n=206)

Targets (n=111)

Combined Entities (n=101)

Table 3.4 provides an overview of the long-term BHARs to acquiring companies in the automotive supply industry. Overall, BHARs to acquirers are consistently negative across both equal- and value-weighted calculation methods and across all three event periods. In addition, they increase over time from an insignificant -2.8% and -1.2% within a 12-month period to a significant -16.7% and -17.4% over the 36 months following the transaction (significant at the 5%-level).15 While negative returns of about 15 to 20% generally correspond to previous findings on long-term BHARs (see Black et al. (2001), Gregory and Matatko (2004)), they also indicate that the original short-term over-performance cannot be sustained in a longterm perspective. Instead, the results suggest that long-term returns strongly 15

Previous research finds less significant value-weighted BHARs than equal-weighted BHARs indicating a potential influence of smaller companies in their applied samples (Mitchell and Stafford (2000) and André, Kooli, and L'Her (2004)).

46

3 Study 1: Determinants of Capital Market Performance

underperform the matching portfolio and yield significant value losses for acquiring companies. Table 3.4: BHARs to Acquirers Total Sample BHARs (n=164) Equal-weighted BHARs

Value-weighted BHARs

Event-Window

BHAR

t-value

p-value

BHAR

t-value

p-value

12 months

-2.79%

-0.70

0.49

-1.16%

-0.29

0.77

24 months

-8.50%

-1.39

0.17

-6.59%

-1.08

0.28

-16.68% -2.18 0.03 ** -17.38% -2.28 0.02 ** 36 months This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table differentiates between a 12-, 24-, and 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratios). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test.

The results of the calendar-time approach generally support these findings. Table 3.5 shows that applying a Fama-French-3-Factor-model results in equal- and valueweighted negative abnormal returns to acquirers. While the overall model and the individual factors are strongly significant, the average monthly abnormal returns as depicted by the intercept 'a' only yield a significant -0.6% under the application of equalweighted portfolios; value-weighted portfolios lose an insignificant -0.3%. Over a 36month horizon, these monthly averages roughly translate to an approximate compounded value loss of 19% and 10%, for the value-weighted and equal-weighted results respectively. These results can be regarded as fairly consistent with the BHARs presented in Table 3.4. The results support the short-term prediction of Hypothesis 3.1. While capital markets evaluate mergers and acquisitions in this particular industry positively in the short-run, this exceptional position cannot be sustained over a three-year holding period. In the long-run, acquirers in the automotive supply industry lose significant value and underperform their non-merging matching firms. Furthermore, in contrast to the major-

3.4 Empirical Results

47

ity of previous studies, the automotive supply industry exhibits consistently significant negative abnormal returns which are very robust across different applied approaches. Table 3.5: Abnormal Returns (FF3F) to Acquirers Fama-French-3-Factor-Model (Total Sample) Equal-Weighted Panel Total Sample t-values Stat. Quality

Value-Weighted

a

b

s

h

a

b

s

h

-0.006**

0.927***

0.483***

-0.087

-0.003

0.937***

0.172***

-0.273***

-2.19

23.38

5.92

-1.05

-1.07

24.22

2.21

-3.46

F = 196.23***

R2adj. = 0.66

F = 212.21***

R2adj. = 0.68

This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply industry. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2004 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test. In addition, statistical quality of the model is indicated by the adjusted determination coefficient (adj. R2) and the corresponding F-statistic.

3.4.2 The Impact of Transaction Characteristics 3.4.2.1 Geographical Expansion

Hypothesis 3.2 predicts no significant impact of geographical expansion on short-term announcement returns in the automotive supply industry. From Table 3.6, it becomes apparent that both national and cross-border transactions yield significant positive abnormal returns particularly in the short event-windows around a takeover announcement. While cross-border transactions are followed by a significant +1.58% in the [-1,0] event-window, national transaction yield a significant +1.93% over the same two-day time period. The difference of approximately 0.3% is not statistically significant; the original hypothesis seems to hold. However, the longer the observed eventwindow becomes, the bigger the return difference becomes between cross-border and national transactions. While all event-windows of 11 or more days yield an insignificant approximate of +1% for cross-border transactions, acquirers almost consistently gain significant positive returns of above 2% in national transactions. In these event-

48

3 Study 1: Determinants of Capital Market Performance

windows, the differences in means also become significant. Furthermore, the medians of cross-border transactions frequently turn to negative returns, another indicator that cross-border transactions could be a determinant of less positive or even negative abnormal short-term returns to acquirers. Table 3.7 reveals a similar but clearer pattern for cross-continental deals. While again all CAARs to acquirers are consistently positive in both panels, the difference in means already becomes significant in the three-day event-window [-1,1] with a difference of 0.84% between cross-continental and within-continent transactions. Therefore, it is concluded that the findings are consistent with those of Mentz (2006), capital markets seem to value transactions within a single continent more than cross-continental transactions. The results are more profound and significant for cross-continental transactions than for cross-border transactions. One potential explanation for this difference lies in the increasing number of European deals, where open borders and the unified European legislation simplify cross-border transactions. As a result, the previously observed larger integration efforts in cross-border deals decrease at least in European transactions. Table 3.8 shifts perspectives and presents findings about the impact of geographical expansion on long-term abnormal returns. BHARs over a 36-month holding period to acquirers in cross-continent, cross-border and national takeover deals remain negative in total. The only significant result can be obtained for value-weighted BHARs in cross-border transactions. In direct comparison, the -20.95% (significant at the 5%level) represent a stronger effect than the -17.38% on the total sample (also significant at the 5%-level). As the equal-weighted return is almost identical to the total sample, the resulting BHARs are obviously influenced by a size effect. Consistent with the findings of Beitel et al. (2004) for European bank M&A, national and cross-continent transactions appear to perform better than the total transaction sample. However, given their statistical insignificance, these returns can only serve as a weak indicator for a positive effect of national and cross-continent takeover deals.

3.4 Empirical Results

49

Table 3.6: Acquirer CAARs – Differences by Geographical Scope Acquirers CAAR Mean Median t-value (t) z-value (B) z-value (GST) 1.06% -0.77% 0.71 1.21 0.50 1.22% 0.38% 0.91 1.28 1.10 1.21% -0.78% 1.05 1.28 0.30 0.97% -0.58% 1.06 1.69 ** 0.30 1.58% 0.50% 2.22 ** 3.28 *** 1.89 ** 0.82% 0.15% 1.74 ** 2.12 ** 1.10 0.71% 0.05% 1.75 ** 2.55 *** 1.89 ** 1.48% 0.71% 2.26 ** 3.63 *** 2.09 ** 1.38% 0.10% 1.75 ** 2.43 *** 1.10 0.61% -0.72% 0.73 0.99 0.11 0.46% -0.84% 0.43 0.89 0.30 3.37% 2.22% 1.94 ** 1.98 ** 1.98 ** National 2.83% 3.79% 1.84 ** 2.06 ** 1.98 ** Transactions (n=104) 2.46% 1.20% 1.91 ** 1.94 ** 1.39 * 3.64% 2.61% 3.28 *** 3.71 *** 3.36 *** 1.93% 1.11% 3.58 *** 3.74 *** 3.56 *** 1.13% 0.67% 2.66 *** 2.84 *** 2.38 *** 1.17% 0.18% 3.25 *** 4.05 *** 2.18 ** 1.96% 1.58% 3.74 *** 4.56 *** 3.16 *** 2.70% 1.54% 3.64 *** 4.42 *** 3.75 *** 1.91% 1.28% 2.27 ** 2.69 *** 1.98 ** 2.45% 0.58% 2.11 ** 2.16 ** 1.00 CAAR Means Absolute Panel 1 Panel 2 Difference t-value [-20,20] 1.06% 3.37% 2.30% 2.91 *** Mean [-20,10] 1.22% 2.83% 1.61% 2.15 ** Difference [-10,10] 1.21% 2.46% 1.25% 1.82 ** Test [-5,5] 0.97% 3.64% 2.68% 4.27 *** [-1,1] 1.58% 1.93% 0.35% 0.71 [-1,0] 0.82% 1.13% 0.32% 0.77 [0] 0.71% 1.17% 0.45% 1.18 [0,1] 1.48% 1.96% 0.48% 1.01 [0,5] 1.38% 2.70% 1.33% 2.44 *** [0,10] 0.61% 1.91% 1.30% 2.28 ** [0,20] 0.46% 2.45% 1.99% 3.04 *** This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by geographical scope. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Statistical significance of the differences in means of the two subsamples is derived following a t-test as presented in Siems (1996). Panel Cross-Border Transactions (n=102)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

50

3 Study 1: Determinants of Capital Market Performance

Table 3.7: Acquirer CAARs – Differences by Continental Scope Acquirers CAAR Mean Median t-value (t) z-value (B) z-value (GST) 0.62% -0.93% 0.43 0.57 0.30 0.83% 0.78% 0.61 0.58 0.99 0.67% -0.63% 0.59 0.48 0.07 0.47% -0.74% 0.44 0.60 -0.38 1.23% 0.81% 1.49 * 1.81 ** 1.67 ** 0.71% -0.05% 1.31 * 1.36 * 0.53 0.72% 0.05% 1.64 * 2.27 ** 1.90 ** 1.24% 0.91% 1.67 ** 2.43 *** 2.13 ** 1.23% 0.41% 1.38 * 1.66 ** 1.22 0.70% 0.22% 0.78 0.51 0.76 0.49% 0.02% 0.46 0.47 0.76 3.18% 1.85% 1.97 ** 2.35 *** 2.00 ** Within 2.74% 1.59% 1.94 ** 2.47 *** 2.00 ** Continent 2.54% 0.43% 2.11 ** 2.42 *** 1.47 * Transactions (n=129) 3.42% 2.20% 3.59 *** 4.27 *** 3.59 *** 2.07% 1.06% 4.06 *** 4.80 *** 3.59 *** 1.13% 0.55% 2.91 *** 3.37 *** 2.70 *** 1.08% 0.14% 3.11 *** 4.10 *** 2.17 ** 2.02% 1.21% 4.02 *** 5.33 *** 3.06 *** 2.54% 1.01% 3.73 *** 4.63 *** 3.41 *** 1.61% 0.15% 2.03 ** 2.73 *** 1.29 * 2.05% -0.61% 1.88 ** 2.26 ** 0.58 CAAR Means Absolute Panel 1 Panel 2 Difference t-value [-20,20] 0.62% 3.18% 2.56% 3.05 *** Mean [-20,10] 0.83% 2.74% 1.91% 2.41 *** Difference [-10,10] 0.67% 2.54% 1.88% 2.57 *** Test [-5,5] 0.47% 3.42% 2.95% 4.40 *** [-1,1] 1.23% 2.07% 0.84% 1.60 * [-1,0] 0.71% 1.13% 0.42% 0.96 [0] 0.72% 1.08% 0.36% 0.87 [0,1] 1.24% 2.02% 0.77% 1.52 * [0,5] 1.23% 2.54% 1.31% 2.26 ** [0,10] 0.70% 1.61% 0.91% 1.49 * [0,20] 0.49% 2.05% 1.56% 2.23 ** This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by continental scope. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Statistical significance of the differences in means of the two subsamples is derived following a t-test as presented in Siems (1996). Panel CrossContinental Transactions (n=77)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

3.4 Empirical Results

51

Table 3.8: Acquirer BHARs – Differences by Geographical Scope Cross-Border vs. National Subsamples – 36-month BHARs Equal-weighted BHARs

Value-weighted BHARs

Panel

BHAR

t-value

p-value

BHAR

t-value

Cross-Continental

-6.23%

-0.55

0.58

-15.23%

-1.35

p-value 0.18

Cross-Border

-16.88%

-1.61

0.11

-20.95%

-2.00

0.05 **

-14.90% -1.33 0.19 -11.01% -0.98 0.33 National This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by geographical scope of the deal. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table contains BHARs over a 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratios). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test.

Table 3.9: Abnormal Returns to Acquirers – Differences by Geographical Scope Fama-French-3-Factor-Model (Subsamples by geographical scope) Equal-Weighted Panel

a

b

Cross-Continental

-0.004

t-values

-1.04

Statistical Quality Cross-Border t-values

0.963*** 0.479*** 18.03

F = 121.41***

4.33

Value-Weighted h

19.78

F = 137.83***

National

-0.005

t-values

-1.30

h

0.125

-0.228**

-0.81

1.27

-2.29

-0.68

R2adj. = 0.58

5.65

s

-0.003 0.954***

0.913*** 0.578*** -0.185* 18.43

b

0.36

R2adj. = 0.60

4.27

a

0.037

-0.007** 0.939*** 0.417*** -0.067 -2.00

Statistical Quality

s

-1.79

19.49

F = 136.24***

R2adj. = 0.57

-0.002 0.987***

0.085

-0.287***

-0.58

0.88

-2.96

20.73

F = 151.11*** 0.000 -0.05

R2adj. = 0.60

0.940*** 0.290*** -0.391*** 18.34

2.80

-3.75

F = 125.61*** R2adj. = 0.56 Statistical Quality F = 125.17*** R2adj. = 0.55 This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply, differentiated by scope of the underlying deals. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2007 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test. In addition, statistical quality of the models is indicated by the adjusted determination coefficient (adj. R2) and the corresponding F-statistic.

As shown by Table 3.9, the abnormal returns in calendar-time exhibit a similar pattern. The only significant result is derived for equal-weighted portfolios in cross-

52

3 Study 1: Determinants of Capital Market Performance

border transactions and points to a slightly more negative monthly return of -0.7% as opposed to the -0.6% of the total sample. The value-weighted abnormal return for crossborder transactions remains insignificant at -0.2%. For national transactions, a valueweighted abnormal return of 0.0% indicates a positive effect of national takeovers on the long-term performance of acquirers. However, due to the lack of statistical significance, it is again only an indication for a potential effect to be tested by the following regression analyses. Overall, the findings of a negative effect of cross-border transactions on long-term returns correspond to previous research in the field (see Conn et al. (2005)). However, a positive counter-effect of national transactions cannot be determined. 3.4.2.2 Product Diversification

Besides geographical expansion, the diversification of activities and product sectors is frequently found to be a determinant of M&A performance. As presented in Chapter 2.1, the automotive supply industry contains a number of different product groups requiring diverse technical and organizational skills. One common approach to determine product similarities within one industry is the comparison of four-digit Standard Industrial Classification (SIC) codes.16 However, since SIC codes generally depend on their data source and may change over time, this study follows an innovative approach: In order to test for the incremental effect of product diversification, all target and acquiring companies are clustered according to their operational focus. Companies that operate in more than one of the following product sectors are assigned to each of the first three sectors stated either by the Automotive News or the respective company home page. The following six overarching product groups are developed on the basis of the business descriptions applied by the Automotive News (Automotive News (2003)):

16

For example, Comment and Jarrell (1995) use industry classifications via four-digit SIC codes to identify firms with similar focus of operations within the same industry; Eckbo (1992) uses SIC codes to describe horizontal and vertical mergers within a given sample.

3.4 Empirical Results

53

Exterior Body Work and Chassis – All metal and plastic components used for

building the load-bearing frame and outer appearance of the car, including general metal components, doors, hoods, bumpers, and car windows.

Engine and Combustion Related Components – Parts and components directly

used to build the engine, including full engine systems, oil filtration systems, cylinders, fuel injection pumps, and pistons.

Driving Components – All components applied in the driving process while not

qualifying as pure electrical components or part of the engine. Among others, this group includes suppliers of suspensions, shock absorbers, braking systems, steering systems, and transmissions.

Electrical Parts – Includes all electrical parts and components applied to a car.

The full range covers suppliers of complex systems and of basic electrical parts. Examples include windshield wipers, airbag and security systems, automatic door-locks, alarm systems, switches, relays, and cable harnesses.

Interior Components and Textiles – All plastic and textile components used for

lining the interior of a car including leather and seats, dash boards, wall and roof coverings.

Tires – This group includes suppliers of rubber and tire products as well as of

steel and aluminum rims.

54

3 Study 1: Determinants of Capital Market Performance

Table 3.10: Acquirer CAARs – Differences by Product Scope Acquirers CAAR Mean Median t-value (t) z-value (B) z-value (GST) 3.34% 2.88% 2.58 *** 2.32 ** 3.17 *** 2.85% 2.91% 2.53 *** 2.22 ** 2.99 *** 2.29% 1.75% 2.29 ** 2.11 ** 1.88 ** 2.37% 1.08% 3.12 *** 3.14 *** 2.62 *** 1.77% 1.10% 3.52 *** 3.70 *** 2.99 *** 0.86% 0.24% 2.16 ** 2.25 ** 1.88 ** 1.07% 0.13% 2.96 *** 3.91 *** 2.99 *** 1.98% 1.15% 3.96 *** 4.85 *** 3.35 *** 2.56% 1.11% 3.81 *** 4.51 *** 3.54 *** 1.98% 1.95% 2.62 *** 2.83 *** 2.43 *** 2.46% 1.15% 2.48 *** 2.65 *** 1.70 ** 0.71% -2.34% 0.34 0.96 -1.00 Diversifying 0.90% -1.95% 0.48 1.21 -0.13 Transactions (n=87) 1.22% -2.37% 0.80 1.14 -0.35 2.25% 0.08% 1.64 * 2.25 ** 0.94 1.74% 0.60% 2.17 ** 3.25 *** 2.45 *** 1.13% 0.55% 2.20 ** 2.72 *** 1.59 * 0.78% 0.04% 1.88 ** 2.64 *** 0.94 1.38% 0.65% 1.92 ** 3.18 *** 1.80 ** 1.34% 0.08% 1.50 * 2.08 ** 1.16 0.30% -0.77% 0.31 0.65 -0.57 0.10% -3.24% 0.08 0.26 -0.57 CAAR Means Absolute Panel 1 Panel 2 Difference t-value [-20,20] 3.34% 0.71% 2.63% 3.26 *** Mean [-20,10] 2.85% 0.90% 1.95% 2.56 *** Difference [-10,10] 2.29% 1.22% 1.07% 1.53 * Test [-5,5] 2.37% 2.25% 0.11% 0.17 [-1,1] 1.77% 1.74% 0.03% 0.07 [-1,0] 0.86% 1.13% 0.27% 0.63 [0] 1.07% 0.78% 0.29% 0.73 [0,1] 1.98% 1.38% 0.59% 1.21 [0,5] 2.56% 1.34% 1.22% 2.19 ** [0,10] 1.98% 0.30% 1.68% 2.88 *** [0,20] 2.46% 0.10% 2.36% 3.52 *** This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by product scope of the transaction. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Statistical significance of the differences in means of the subsamples is derived following a t-test as presented in Siems (1996). Panel Intra Industry Transactions (n=119)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

3.4 Empirical Results

55

All transactions in which the acquirer and the target are both part of the same cluster are treated as 'within product group/intra-industry,' all others represent 'diversifying transactions.' Table 3.10 presents the CAARs derived for these two subsamples. Similar to the results in the previous section, short-term announcement returns are nearly the same in the very short event-windows. In the [-1,1] event-window, transactions within product groups yield a significant +1.77% (1% significance-level) and diversifying transactions yield a significant +1.74% (1% significance-level); the differences are almost non-existent and not significant. In the longer event-windows, only focusing transactions yield significant positive returns up to 3.3% in the 41 days surrounding the transaction. The differences in means are highly significant and negative medians for diversifying transaction again indicate potential value destruction in the short-term announcement effect for diversifying transactions. On the basis of these results, it can be inferred that upon immediate announcement capital markets react similarly to both types of transactions. Including the run-up-, information leakage- and informationdelay-phase, the return patterns indicate a favorable valuation of focusing deals. Table 3.11: Acquirer BHARs– Differences by Product Scope Intra-Industry vs. Diversifying Subsamples – 36-month BHARs Equal-weighted BHARs Panel

BHAR

t-value

Intra-Industry Transactions

-10.28%

-1.07

Value-weighted BHARs

p-value

BHAR

t-value

p-value

0.29

-10.24%

-1.07

0.29

-24.86% -2.01 0.05 ** -26.32% -2.13 0.04 ** Diversifying Transactions This table shows the average Buy-and-Hold Abnormal returns to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by product scope of the transaction. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table contains BHARs over a 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratios). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test.

The BHARs as presented in Table 3.11 strongly support the results of the shortterm assessment as well as the negative impact assumed in the original hypothesis. Over

56

3 Study 1: Determinants of Capital Market Performance

a 36-month holding period, diversifying transactions yield significant negative BHARs at the 5%-significance-level: -24.86% for the equal-weighted calculations and -26.32% for the value-weighted returns. Intra-industry transactions show insignificant and lower negative returns than diversifying deals. Therefore, short-term announcement effects as well as long-term BHARs seem to support Hypothesis 3.3: While CAARs positively reward focusing transactions and long-term BHARs indicate a significant underperformance of diversifying transactions, focusing activities in M&A transactions appears to carry a greater chance to realize synergy potentials. Table 3.12: Abnormal Returns to Acquirers – Differences by Product Scope Fama-French-3-Factor-Model (Intra-Industry vs. Diversifying Subsamples) Equal-Weighted Panel

a

b

s

Value-Weighted h

a

Intra-Industry -0.009*** 0.896*** 0.431*** -0.065 -0.004 t-values

-2.83

Stat. Quality

20.76

F = 153.27***

Diversifying

-0.002

t-values

-0.43

4.86

-0.72

R2adj. = 0.60

1.003*** 0.561*** -0.093 16.32

4.61

-0.77

-1.43

b

s

h

0.930***

0.147*

-0.213**

22.02

1.73

-2.48

F = 175.84*** 0.001 0.24

R2adj. = 0.63

0.962*** 0.312** 15.26

2.59

-0.424*** -3.57

Stat. Quality F = 98.15*** R2adj. = 0.52 F = 88.10*** R2adj. = 0.49 This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply, differentiated by scope of the underlying deals. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2007 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test. In addition, statistical quality of the model is indicated by the adjusted determination coefficient (adj. R2) and the corresponding F-statistic.

However, the derived return patterns are not fully confirmed by the results of the calendar-time approach. As described in Table 3.12, the FF3F finds significant monthly negative returns of -0.9% (equal-weighted, 1% significance-level) for intra-industry transactions, all other returns stay insignificant. Intra-industry transactions appear to underperform the total sample, which yields -0.6% respectively. Diversifying transactions show an insignificant negative return of -0.2% for equal-weighted portfolios and

3.4 Empirical Results

57

an also insignificant positive 0.1% for value-weighted portfolios. However, since both returns stay statistically insignificant, their interpretation potential is limited. In general, calendar-time approaches are often described as more robust and reliable than BHARs, especially when the analysis is exposed to industry clustering. The results here indicate that firms that do not diversify significantly underperform the market. However, in the light of previous findings in other industries and the short-term effects derived above, it is assumed that the BHAR-analysis represents a more realistic view on long-term performance in the automotive supply industry and that diversification has a value destroying effect. These findings are again challenged in the multiple regression models following in Section 3.5. 3.4.2.3 Transaction Size

In order to test for the incremental effect of transaction size on the performance of acquirers, the total transaction sample is ranked by underlying transaction volume in US dollars and grouped into size-dependent clusters. In addition, the 20 largest and 20 smallest transactions are separated into independent groups to determine the effect of size at the extreme ends of the scale. Table 3.13 reveals the short-term effect measured around the announcement dates. In general, it becomes apparent that larger transactions constantly outperform the total sample averages as presented above. Over the [-1,0] event-window, the largest 20 transactions gain a significant 2.98% (significant at the 5%-level), which corresponds to three times the average of the total sample (0.98%). Although this value gain decreases to a significant 1.7% for the largest third of the transactions, all averages in these upper-size clusters remain above the total average.

58

3 Study 1: Determinants of Capital Market Performance

Table 3.13: Acquirer CAARs – Differences by Transaction Size Acquirers/ Panel Largest 20 Transactions (n=18)

Smallest 20 Transactions (n=18)

Largest 33% (n=70)

Middle 33% (n=67)

Smallest 33% (n=69)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

CAAR Mean 5.20% 5.14% 3.06% 6.04% 5.97% 2.98% 2.82% 5.81% 7.31% 4.22% 4.27% 3.21% 2.15% 2.59% 1.97% 0.28% -0.84% 0.00% 1.12% 1.11% 1.20% 2.27% 5.30% 5.61% 5.26% 4.45% 3.08% 1.70% 1.67% 3.06% 4.10% 3.52% 3.21% -0.82% -1.13% -1.26% 1.17% 0.64% 0.55% 0.41% 0.50% 0.38% -1.23% -0.92% 2.07% 1.47% 1.39% 1.28% 1.50% 0.66% 0.73% 1.56% 1.59% 1.41% 2.01%

Median 3.21% 6.00% 1.37% 3.94% 4.21% 0.60% 0.99% 3.41% 5.73% 3.12% 1.91% 2.88% 2.06% 2.31% 0.93% -0.37% -0.38% -0.10% -0.19% -0.02% 0.57% -3.08% 3.21% 4.05% 2.55% 2.37% 1.81% 0.71% 0.32% 2.01% 2.03% 3.28% 1.83% -0.61% -2.22% -2.72% -0.08% 0.24% 0.10% 0.05% 0.09% -0.13% -2.12% -2.27% -0.80% 1.44% 0.63% 0.30% 0.48% 0.24% 0.08% 0.54% 0.41% 0.15% 0.02%

t-value (t) 1.11 1.30 * 0.91 1.84 ** 1.94 ** 1.61 * 1.77 ** 2.07 ** 2.45 *** 1.44 * 1.12 0.86 0.97 1.33 * 1.12 0.41 -1.15 -0.01 1.32 * 0.78 0.98 0.80 2.23 ** 2.62 *** 2.80 *** 2.77 *** 3.12 *** 2.51 *** 2.68 *** 3.31 *** 3.65 *** 3.15 *** 2.24 ** -0.50 -0.75 -1.12 1.19 0.95 1.20 1.19 0.80 0.49 -1.32 -0.77 1.15 0.99 1.11 1.24 2.71 *** 1.40 * 1.95 ** 2.96 *** 1.93 ** 1.46 * 1.43 *

z-value (B) 1.21 1.43 * 0.83 2.40 *** 3.26 *** 2.17 ** 3.10 *** 4.02 *** 4.14 *** 2.15 ** 1.58 * 0.47 0.53 0.53 0.58 0.07 -0.73 0.32 1.08 0.03 -0.06 0.05 2.57 *** 2.87 *** 3.11 *** 3.48 *** 3.78 *** 2.43 *** 3.55 *** 4.67 *** 4.78 *** 3.59 *** 2.73 *** -0.11 -0.20 -0.65 1.42 * 1.64 * 1.59 * 1.64 * 1.55 * 0.97 -0.93 -0.58 1.28 * 1.17 1.19 1.53 * 3.11 *** 2.00 ** 2.74 *** 3.70 *** 2.28 ** 1.58 * 1.54 *

z-value (GST) 1.13 1.13 0.66 1.61 * 2.08 ** 0.66 1.13 2.55 *** 2.55 *** 1.13 0.66 0.92 0.92 1.39 * 1.39 * -0.50 -0.50 -0.03 -0.50 -0.03 0.92 -0.50 2.00 ** 2.48 *** 2.48 *** 2.72 *** 3.92 *** 1.76 ** 2.00 ** 3.68 *** 4.16 *** 3.44 *** 1.76 ** 0.29 -0.20 -1.67 0.53 1.27 1.02 1.27 0.78 0.53 -1.92 -1.18 0.75 1.47 * 1.23 1.23 1.47 * 1.47 * 1.71 ** 1.96 ** 1.23 0.99 0.99

This table shows the CAARs to acquirers in the automotive supply industry, differentiated by transaction size. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively; tests include a standard t-Test (t), the cross-sectional test (B), and the Generalized Sign Test (GST).

3.4 Empirical Results

59

On the other end of the scale, the observed patterns indicate two main conclusions: Firstly, short-term returns to acquirers in smaller transactions are consistently lower than announcement effects experienced after large deals. The smallest 20 transactions gain an insignificant 3.21 % over the 41 days around the takeover announcement and lose an insignificant 0.8% in the [-1,0] event-window. Secondly, the smallest transactions are not performing the worst. The acquirers in the middle tier realize +0.55% (10% significant) in the [-1,0] event-window, in which they underperform the smallest third tier by about 0.1%. In addition, they lose approximately 1% in the longest event period underperforming the takeovers in the smallest 33% as well as the 20 smallest transactions. Consequently, short-term performance in the automotive supply industry might not be a linear function of transaction size as previously assumed by Mentz (2006), but could be assumed to have two effects: Superior resources and experience enable large acquirers in large transactions to realize above-average short-term returns, while acquirers in very small transactions benefit from their relative size and reduced complexity of their targets. Takeovers in between the two extremes benefit the least as the increasing complexity cannot be offset with sufficient resources and experience. The BHARs as provided by Table 3.14 generally support a significant underperformance of medium-sized takeover deals. The medium third of deals lose a significant -34.96% equal-weighted and -32.14% value-weighted over a 36-month holding period, both significant at the 5%-level. While medium-sized deals underperform in comparison to the other panels, it appears that larger deals yield less negative returns while smaller deals even lead to positive long-term averages. One possible explanation lies in the relative size of acquirers to targets: Since large acquirers are usually associated with large takeovers, large available resources weaken the overall negative effect on long-term performance. For very small takeovers, the reduced complexity of the targets seems to weigh more and even allows for positive returns to acquirers. As the returns remain statistically insignificant, their interpretation can only serve as an indication.

60

3 Study 1: Determinants of Capital Market Performance

Table 3.14: Acquirer BHARs – Differences by Transaction Size Subsamples by Size of Transaction – 36-month BHARs Equal-weighted BHARs

Value-weighted BHARs

Panel

BHAR

t-value

p-value

BHAR

t-value

p-value

Largest 20 Transactions

-12.31%

-0.73

0.47

-17.61%

-1.05

0.31

Smallest 20 Transactions

15.56%

0.61

0.55

16.01%

0.63

0.54

Largest 33%

-16.75%

-1.33

0.19

-19.43%

-1.54

0.13

Middle 33%

-34.96%

-2.28

0.03 **

-32.14%

-2.10

0.04 **

2.02% 0.18 0.86 4.97% 0.45 0.65 Smallest 33% This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by transaction size. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table contains BHARs over a 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratios). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test.

The only significant result obtained from the calendar-time approach as presented in Table 3.15 is a -1.1% loss for the smallest third transactions (equal-weighted, significant at the 5%-level). Overall, the equal-weighted returns seem to increase with size, the largest 20 transactions yield an insignificant +0.1%, the largest third experience monthly average returns of -0.3%. The value-weighted returns, though, are less indicative and altogether insignificant. In combination with the BHARs and the short-term effects, it can therefore be concluded that larger transactions have a consistent positive effect on short- and long-term performance: Hypothesis 3.4 seems to hold. In addition, there are also indications that very small transactions are also a determinant for positive effects on overall performance, but the results are not consistently observable across the different approaches applied.

3.4 Empirical Results

61

Table 3.15: Abnormal Returns to Acquirers – Differences by Transaction Size Fama-French-3-Factor-Model (Subsamples by transaction size) Equal-Weighted

Value-Weighted

Panel Largest 20 Transactions

a

b

s

h

a

b

s

h

0.001

1.076***

0.330

0.369*

0.010

1.172***

-0.038

0.024

t-values

0.21

10.31

1.61

1.91

1.39

10.86

-0.19

0.13

Stat. Quality Smallest 20 Transactions

-0.008

t-values

-1.35

Stat. Quality

F = 40.67***

-0.003

t-values

-0.95

Stat. Quality

-0.003

t-values

-0.96

-2.50

4.53

19.41

4.05

4.85

0.85

1.029*** 0.382** 12.16

F = 55.94***

2.25

-0.590*** -3.49

R2adj. = 0.39

-0.004 0.958***

0.070

-0.122

-1.12

0.74

-1.31

19.21

F = 135.69***

R2adj. = 0.60

-0.120

-0.005 0.926*** 0.245**

-1.31

-1.50

R2adj. = 0.65

4.56

0.005

R2adj. = 0.40

0.82

0.853*** 0.570*** -0.226* 14.13

F = 44.97***

0.082

R2adj. = 0.60

1.017*** 0.451*** 21.73

-1.85

R2adj. = 0.35

0.995*** 0.410***

F = 167.19***

Smallest Third -0.011** t-values

10.93

F = 136.61***

Middle Third

Stat. Quality

0.856*** 0.730*** -0.300*

F = 46.96***

Largest Third

R2adj. = 0.37

-1.80

18.56

F = 128.10***

2.55

-0.370*** -3.92

R2adj. = 0.58

-0.004 0.931*** 0.388*** -0.556*** -0.83

13.65

2.98

-4.41

F = 72.33*** R2adj. = 0.46 Stat. Quality F = 73.74*** R2adj. = 0.44 This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply industry, differentiated by transaction size. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2007 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test. In addition, statistical quality of the model is indicated by the adjusted determination coefficient (adj. R2) and the corresponding F-statistic.

3.4.3 The Impact of Acquirer Characteristics 3.4.3.1 Product Groups

After the previous section has focused on deal characteristics and their impact on short- and long-term capital market performance of acquirers, the following section highlights two acquirer characteristics that are presumably influencing post-takeover performance. The first characteristic to be analyzed is the affiliation of the acquirers

62

3 Study 1: Determinants of Capital Market Performance

with different product groups within the automotive supply industry. As described in Chapter 2.1, a variety of different products supplied to automotive producers exist and can be characterized by different levels of competition and varying degrees of consolidation. In order to test for the impact of these different product groups, the same clusters as applied under the 'product diversification hypothesis' (Section 3.4.2.2) are individually analyzed in this section. Table 3.16 exhibits the short-term announcement effects to acquirers across the different product groups. In general, it becomes apparent that the observed return patterns vary from event-window to event-window. For example, in the [-1,0] eventwindow, all product groups except the 'electrical components' yield significant positive announcement returns of about 1%. In the eleven days around the announcement day, 'engine' and 'drive components' yield the strongest and most significant positive returns of about 2.6% while all others remain positive but below 2%. In this event-window, suppliers of tires and rubber products represent the negative outlier and lose an insignificant -0.05%. In the longest event-window of 41 days, the product groups 'exterior/chassis' and 'interior' are the only groups to yield significant positive and above average returns of around 3%. All others are insignificant while the 'tires' product group again yields negative returns.

3.4 Empirical Results

63

Table 3.16: Acquirer CAARs – Differences by Product Group Acquirers Panel Exterior/ Chassis (n=66)

Engine (n=51)

Drive Components (n=105)

Electrical Components (n=71)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

CAAR Mean Median 3.40% 2.65% 2.89% 1.87% 1.33% -0.69% 1.98% 0.33% 2.05% 1.42% 1.29% 0.21% 1.09% 0.07% 1.85% 0.87% 2.05% 0.50% 1.16% -0.12% 1.66% -0.55% 1.31% -3.26% 1.41% -1.44% 2.44% -2.41% 2.66% 2.59% 1.61% 1.26% 1.11% 1.18% 0.85% 0.22% 1.35% 0.90% 2.02% 1.33% 0.06% 0.22% -0.04% -0.63% 1.74% -0.73% 2.00% 1.23% 1.70% -0.46% 2.63% 0.76% 1.37% 0.81% 0.89% 0.09% 0.81% 0.13% 1.28% 0.77% 2.13% 0.99% 1.18% -0.01% 0.93% -0.73% 1.51% 1.11% 1.75% 0.29% 1.26% 0.43% 1.57% 0.76% 0.37% 0.10% -0.04% -0.08% 0.08% 0.05% 0.50% 0.65% 1.52% 0.70% 0.92% 0.17% 0.68% -0.20%

t-value(t) 1.74 ** 1.62 * 0.94 1.45 * 1.98 ** 1.80 ** 1.83 ** 1.98 ** 1.64 * 0.96 1.10 0.47 0.57 1.09 1.45 * 2.08 ** 1.70 ** 1.44 * 1.87 ** 2.41 *** 0.06 -0.03 1.05 1.33 * 1.35 * 2.30 ** 2.44 *** 2.04 ** 2.14 ** 2.47 *** 2.89 *** 1.43 * 0.86 1.01 1.34 * 1.23 1.83 ** 0.64 -0.11 0.25 0.92 2.11 ** 1.18 0.65

z-value (B) z-value (GST) 1.60 * 2.15 ** 1.60 * 1.90 ** 0.70 0.42 1.45 * 1.90 ** 2.61 *** 2.39 *** 2.47 *** 1.65 ** 2.59 *** 1.16 2.65 *** 1.65 ** 1.90 ** 1.16 0.64 0.42 0.74 0.17 0.53 -0.50 0.60 0.06 1.09 0.06 1.89 ** 1.75 ** 2.61 *** 2.03 ** 2.35 *** 2.03 ** 2.39 *** 2.03 ** 2.56 *** 1.75 ** 2.90 *** 2.31 ** 0.42 0.91 0.23 -0.22 1.29 * 0.91 1.81 ** 2.08 ** 1.75 ** 0.91 3.18 *** 2.28 ** 3.28 *** 2.08 ** 2.63 *** 1.30 * 3.28 *** 2.28 ** 3.69 *** 2.48 *** 4.01 *** 2.28 ** 2.13 ** 0.91 1.10 0.32 1.27 0.87 1.42 * 0.63 1.54 * 1.11 2.19 ** 2.06 ** 0.69 1.35 * -0.41 -0.08 0.05 1.58 * 1.11 1.82 ** 2.77 *** 1.82 ** 1.66 ** 0.87 1.28 0.39

64

3 Study 1: Determinants of Capital Market Performance

Table 3.16: Acquirer CAARs – Differences by Product Group (2/2) Acquirers EventCAAR Mean Median Window t-value (t) z-value (B) z-value (GST) [-20,20] 2.54% 2.28% 1.47 * 1.62 * 1.44 * [-20,10] 2.10% 4.09% 1.37 * 1.41 * 1.73 ** [-10,10] 1.03% -1.37% 0.76 0.95 0.03 [-5,5] 1.96% 0.74% 1.75 ** 1.64 ** 1.73 ** [-1,1] 1.67% 1.49% 2.29 ** 2.14 ** 2.58 *** [-1,0] 0.97% 0.65% 1.83 ** 1.64 * 2.01 ** [0] 1.04% 0.18% 2.20 ** 2.58 *** 2.30 ** [0,1] 1.74% 1.15% 2.45 *** 2.77 *** 1.44 * [0,5] 1.93% 0.77% 2.02 ** 2.21 ** 1.73 ** [0,10] 0.78% -0.39% 0.67 0.88 0.31 [0,20] 1.23% -1.01% 0.81 1.18 -0.26 [-20,20] -1.16% -1.05% -0.51 -0.13 0.43 Tires/Rubber (n=32) [-20,10] -0.90% -3.19% -0.39 -0.22 -0.28 [-10,10] -1.04% -2.57% -0.59 -0.62 -0.63 [-5,5] -0.05% -0.99% -0.05 0.23 -0.28 [-1,1] 1.87% 1.09% 2.76 *** 2.98 *** 2.20 ** [-1,0] 0.86% 0.43% 1.78 ** 1.89 ** 0.79 [0] 0.80% -0.14% 1.75 ** 2.27 ** 0.08 [0,1] 1.81% 0.98% 2.61 *** 3.28 *** 1.49 * [0,5] 0.60% 0.32% 0.78 0.74 0.79 [0,10] 0.02% -1.09% 0.02 -0.21 0.08 [0,20] -0.24% -0.95% -0.17 -0.05 0.08 This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by product group of the acquirer. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Panel Interior/ Textile (n=50)

These patterns point to some remarkable conclusions. First of all, producers of tires and rubber products almost consistently represent the outlier across product groups and across event-windows. This pattern could be the result of diverging industry characteristics present for this particular product group. As their return pattern resembles the negative returns typically observed for other industries (see Berry (2000), Akdo÷u (2009)), it could be assumed that especially tire producers are not fully dependent on the automotive industry and find customers for rubber products in other industries. Therefore, capital markets assume that they cannot realize synergy potentials from M&A in the same way other automotive suppliers might be able to. Another reason for their out-

3.4 Empirical Results

65

lying position might be the presence of large tire producers. Since they represent the smallest group among the six and since they contain suppliers with high market capitalization, the impact of a single takeover could be treated differently in the form of negative abnormal returns. A second conclusion on the basis of the observed short-term announcement pattern is that there is no consistent positive outlier across the different event-windows. However, if the longest event-window is considered, the original hypothesis seems to hold: Product groups with less competition and less overcapacity (engines, drive and electrical components) have a consistent negative effect on short-term returns. In the exterior and interior segment, the reduction of overcapacity through mergers and acquisitions as well as the increased market share after a takeover are positively valued and lead to significantly higher positive announcement returns of 2.5% and 3.4% respectively. The tire product group again represents an outlier as described above. Table 3.17: Acquirer BHARs – Differences by Product Group Subsamples by Product Segment – 36-month BHARs Equal-weighted BHARs

Value-weighted BHARs

Panel

BHAR

t-value

p-value

BHAR

t-value

p-value

Exterior

-19.81%

-1.71

0.09 *

-31.78%

-2.74

<0.01 ***

Engine

-7.59%

-0.58

0.56

0.13%

0.01

Drive Components

-27.06%

-2.48

0.02 **

-22.73%

-2.08

0.04 **

Electric Components

-25.73%

-2.08

0.04 **

-17.97%

-1.45

0.15

Interior/Textile

-33.65%

-1.86

0.07 *

-41.22%

-2.28

0.03 **

0.99

19.02% 1.05 0.30 9.14% 0.50 0.62 Tires/Rubber This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by product group of the acquirer. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table contains BHARs over a 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratios). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test.

The long-term BHARs as presented in Table 3.17 partially support these findings. At least for the two segments 'drive components' and 'electric components', the less

66

3 Study 1: Determinants of Capital Market Performance

positive short-term announcement returns translate into significantly negative long-term BHARs of below -20%. Acquirers producing drive components realize a significant loss of -27.06% (5%-level) equal-weighted as well as -22.73% value-weighted (5%-level). The calendar-time approach as presented in Table 3.18 also yields significant negative returns below the total sample average: both product groups yield monthly negative returns between -0.8% and -1.0% (all significant at the 5%-level). Therefore, it can be concluded that acquirers active in one of these two product groups significantly underperform in realizing synergy potentials from takeover activities. While the long-run performance consistently point to an underperformance, this development appears to be anticipated in below-average but insignificant positive short-term announcement returns. Therefore, Hypothesis 3.5 holds concerning the below-average synergyrealization in product sectors with lower competition and no overcapacity. The opposite effect for product sectors with high competition and overcapacity remains inconsistent. For the producers of 'exterior' and 'interior' components, the calendar-time approach yields negative but insignificant returns. The BHAR-analysis finds highly significant and below average negative returns for both product groups. Following the BHAR-analysis, the argument states that the positive effect observed for shortterm abnormal returns does not carry over into the long-term perspective but rather reverses into a negative effect. The calendar-time returns, however, support the argument of Hypothesis 3.5. Overall, the evidence remains inconclusive. Therefore, Hypothesis 3.5 holds only partially true: Product sectors without high competition and no overcapacity yield lower short-term announcement returns and fail to realize synergies in the post-merger integration, as becomes apparent in significant long-term underperformance. A positive effect for other industry segments is not detected.

3.4 Empirical Results

67

Table 3.18: Abnormal Returns to Acquirers – Differences by Product Group Fama-French-3-Factor-Model (Subsamples by Product Segment) Equal-Weighted Panel

a

b

Exterior

-0.006

t-values

-1.51

Stat. Quality

-0.005

t-values

-1.14

Stat. Quality Drive Comp. t-values

Electric Comp. t-values

-0.72

R2adj. = 0.54

0.934*** 0.532*** -0.046 17.11

18.17

F = 118.86***

4.74

-0.40

R2adj. = 0.51

4.85

-1.27

R2adj. = 0.54

-0.010** 0.839*** 0.543*** -0.096 16.19

F = 97.53***

Interior/Textile

-0.004

t-values

-1.04

Stat. Quality

4.63

-0.009** 0.892*** 0.490*** -0.131

-2.58

Stat. Quality

17.19

F = 107.22***

-2.33

Stat. Quality

Value-Weighted h

0.953*** 0.510*** -0.079

F = 108.15***

Engine

s

0.001

t-values

0.29

-0.89

R2adj. = 0.49

1.007*** 0.581*** -0.081 15.69

F = 92.10***

Tires/Rubber

5.10

4.59

-0.66

R2adj. = 0.52

1.095*** 0.551*** -0.103 16.86

4.32 2

-0.82

a

b

-0.005 -1.40

16.34

F = 101.39*** -0.001 -0.25

s

h

0.905*** 0.320*** -0.459*** 3.01

0.936*** 0.340*** 16.72

F = 109.27***

-4.37

R2adj. = 0.53

3.02

-0.219* -1.93

R2adj. = 0.52

-0.008** 0.832*** 0.345*** -0.346*** -2.18

17.08

F = 112.75***

3.52

-3.49

R2adj. = 0.53

-0.008** 0.857*** 0.240** -0.261*** -2.17

17.34

F = 112.14***

2.42

-2.61

R2adj. = 0.53

0.002

1.043***

0.243*

-0.352***

0.44

15.46

1.90

-2.84

F = 89.67***

R2adj. = 0.52

0.003

1.134***

0.152

-0.464***

0.63

15.94

1.13

-3.54

R2adj. =

0.53 F = 90.72*** Stat. Quality F = 103.31*** R adj. = 0.57 This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply industry, differentiated by product group of the acquirer. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2007 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively and is tested using a standard t-Test. In addition, statistical quality of the model is indicated by the adjusted determination coefficient (adj. R2) and the corresponding F-statistic.

3.4.3.2 Acquisition Experience

As the final acquirer characteristic, this section analyzes the influence of acquisition experience on the acquirers' post-merger capital market performance. For this purpose, the total sample of transactions is divided into three subgroups: one group for single bidders, i.e. acquirers that were active once between 1980 and 2007, one group

68

3 Study 1: Determinants of Capital Market Performance

for multiple bidders and a subsample of the multiple-bidders for those which acquired more than five targets over the observed time period (so-called 'bidder champions'). In general, 65% of the 206 transactions completed by public acquirers qualify as multibidder transactions; one-time acquirers engage in 72 takeovers. Four bidder champions (Lear Corporation, Johnson Controls, Dana Corporation, and Continental AG) complete a total of 30 transactions. In general, the short-term announcement effects provided by Table 3.19 confirm earlier findings in the automotive supply industry. While multi-bidders earn a significant 1.10% in the [-1,1] event-window, single bidders earn a significant 2.98%. The difference of nearly 2% is highly significant and remains significant in most of the event-windows. Therefore, capital markets seem to react more positively to a takeover announcement by an inexperienced bidder than to one made by a multi-bidder. As this result confirms earlier findings of Mentz (2006), it also contradicts findings of other studies (Haleblian and Finkelstein (1999)). For the automotive supply industry, shortterm announcement effects seem to rather reflect positive expectations in strengthening of smaller, inexperienced acquirers than in the learning effect of multiple bidders. A similar pattern holds for the short-term effect on bidder champions. Table 3.20 shows that their return difference remains between 0.67% and 3.76% with varying significance levels. This finding concurs with earlier work of Fuller et al. (2002): Capital markets appear to value acquirers who focus on the integration of few or only one potential target and supposedly realize a greater part of the synergies associated with the takeover. Alternatively, the element of surprise and relative size could also be a factor of influence. As returns to bidder champions stay small, this potential absence of abnormal returns could also be the result of build-in acquisition expectations. Since bidder champions frequently engage in takeovers, future takeovers are already expected and preemptively priced into the share price. For inexperienced bidders, the announcement might be surprising and thus leading to a more significant short-term value reaction.

3.4 Empirical Results

69

Table 3.19: Acquirer CAARs – Differences by Acquisition Experience Acquirers CAAR Mean Median t-value (t) z-value (B) z-value (GST) 1.47% 1.05% 1.15 1.36 * 1.29 * 1.78% 1.50% 1.50 * 1.68 ** 1.98 ** 1.48% 0.00% 1.44 * 1.62 * 0.77 1.89% 0.48% 2.16 ** 2.32 ** 2.15 ** 1.10% 0.52% 2.58 *** 2.11 ** 2.67 *** 0.39% -0.06% 1.25 0.51 0.60 0.38% 0.06% 1.50 * 1.20 1.98 ** 1.09% 0.71% 2.79 *** 2.74 *** 2.33 *** 1.54% 0.46% 2.82 *** 2.85 *** 1.81 ** 1.00% 0.31% 1.52 * 1.62 * 1.29 * 0.69% -0.35% 0.82 0.98 0.42 3.63% 0.17% 1.61 * 1.97 ** 1.23 Single 2.49% 0.22% 1.29 * 1.72 ** 0.99 Bidders (n=72) 2.51% 0.08% 1.59 * 1.64 * 0.99 3.12% 2.16% 2.42 *** 3.24 *** 1.46 * 2.98% 1.88% 3.03 *** 5.28 *** 2.89 *** 2.06% 1.23% 3.04 *** 4.84 *** 3.36 *** 1.99% 0.80% 3.33 *** 5.74 *** 2.18 ** 2.91% 1.71% 3.09 *** 5.79 *** 3.13 *** 2.99% 1.56% 2.56 *** 3.99 *** 3.36 *** 1.77% -0.27% 1.48 * 1.97 ** 0.75 2.91% -0.01% 1.79 ** 2.15 ** 0.99 CAAR Means Absolute Panel 1 Panel 2 Difference t-value [-20,20] 1.47% 3.63% 2.16% 2.49 *** Mean 1.78% 2.49% 0.71% 0.87 Difference [-20,10] [-10,10] 1.48% 2.51% 1.03% 1.37 * Test [-5,5] 1.89% 3.12% 1.23% 1.77 ** [-1,1] 1.10% 2.98% 1.89% 3.55 *** [-1,0] 0.39% 2.06% 1.67% 3.73 *** [0] 0.38% 1.99% 1.60% 3.89 *** [0,1] 1.09% 2.91% 1.82% 3.53 *** [0,5] 1.54% 2.99% 1.44% 2.45 *** [0,10] 1.00% 1.77% 0.77% 1.24 [0,20] 0.69% 2.91% 2.22% 3.11 *** This table shows the cumulative average abnormal returns (CAARs) to acquiring companies in the automotive supply industry, differentiated by multi vs. single acquirers. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. Statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Statistical significance of the differences in means of the subsamples is derived following a t-test presented in Siems (1996). Panel MultiBidders (n=134)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

70

3 Study 1: Determinants of Capital Market Performance

Table 3.20: Acquirer CAARs – Differences for Bidder Champions Acquirers CAAR Mean Median t-value (t) z-value (B) z-value (GST) -0.62% 0.04% -0.35 -0.40 0.33 -1.18% 1.15% -0.69 -0.93 0.69 0.17% -0.38% 0.13 -0.22 0.33 1.10% -0.01% 0.94 0.45 0.33 0.94% 0.85% 1.17 0.57 2.15 ** 0.26% -0.11% 0.39 -0.51 -0.41 0.26% 0.08% 0.50 -0.12 0.69 0.93% 1.18% 1.33 * 1.01 1.42 * 1.44% 0.72% 1.43 * 1.16 1.79 ** 0.70% 0.68% 0.62 0.44 0.33 1.27% 0.16% 0.82 0.87 0.69 2.71% 0.30% 2.07 ** 2.55 *** 1.77 ** Other 2.58% 1.24% 2.23 ** 2.85 *** 2.08 ** Bidders (n=176) 2.13% 0.10% 2.16 ** 2.50 *** 1.17 2.53% 0.77% 3.06 *** 3.88 *** 2.68 *** 1.90% 0.98% 3.78 *** 5.06 *** 3.29 *** 1.10% 0.54% 3.12 *** 3.99 *** 2.83 *** 1.06% 0.11% 3.48 *** 5.06 *** 2.83 *** 1.86% 0.90% 3.92 *** 5.75 *** 3.44 *** 2.15% 0.80% 3.52 *** 4.61 *** 2.98 *** 1.37% 0.16% 2.03 ** 2.51 *** 1.47 * 1.50% -0.67% 1.69 ** 1.95 ** 0.71 CAAR Means Absolute Panel 1 Panel 2 Difference t-value [-20,20] -0.62% 2.71% 3.33% 2.11 ** Mean -1.18% 2.58% 3.76% 2.52 *** Difference [-20,10] [-10,10] 0.17% 2.13% 1.95% 1.42 * Test [-5,5] 1.10% 2.53% 1.42% 1.13 [-1,1] 0.94% 1.90% 0.96% 0.98 [-1,0] 0.26% 1.10% 0.84% 1.00 [0] 0.26% 1.06% 0.80% 1.04 [0,1] 0.93% 1.86% 0.93% 0.97 [0,5] 1.44% 2.15% 0.71% 0.65 [0,10] 0.70% 1.37% 0.67% 0.58 [0,20] 1.27% 1.50% 0.23% 0.18 This table shows the cumulative average abnormal returns (CAAR) to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by bidder champions vs. other bidders. It contains all public acquirers for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. Statistical significance is tested using a standard t-Test (t), the cross-sectional test as proposed by Boehmer et al. (1991) (B), and the Generalized Sign Test (GST) as proposed by Cowan et al. (1990). Statistical significance of the differences in means of the subsamples is derived following a t-test presented in Siems (1996). Panel Bidder Champions (n=30)

EventWindow [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20] [-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

3.4 Empirical Results

71

The long-term performance as determined by the BHARs in Table 3.21 supports the original hypothesis. Single bidders are the only subgroup for which significant negative abnormal returns are determined. Over 36 months following the transactions, single bidders realize a value loss of -36.49% equal-weighted and -37.47% valueweighted (both significant at the 5%-level). Thereby, they clearly underperform the more experienced multi-bidders and bidder champions, for which only insignificant returns are obtained. Based on this BHAR-analysis, it can be inferred that single bidders are lacking integration experience and fail to realize the original synergies underlying their short-term market expectations. Table 3.21: Acquirer BHARs – Differences by Acquisition Experience Subsamples by acquisition experience – 36-month BHARs Equal-weighted BHARs Panel

BHAR

Multi-bidders Bidder champions

t-value

p-value

Value-weighted BHARs BHAR

t-value

p-value

-8.48%

-1.02

0.31

-12.05%

-1.45

0.15

-10.34%

-0.66

0.52

-14.73%

-0.93

0.36

-36.49% -2.22 0.03 ** -37.47% -2.28 0.03 ** Single bidders This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry, differentiated by bidding experience of the acquirer. Multi bidders have completed more than one merger or acquisition between 1980 and 2007, bidder champions more than five. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table contains BHARs over a 36-month holding period. All acquirers between 1980 and 2004 are included for which the relevant matching information is available (market values and market-to-book ratio). Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-test.

Table 3.22 indicates that the 'lacking experience' assumption underlying the BHAR-analysis is also supported by the results of the FF3F-model. This model also determines negative and significant returns for multi-bidders of -0.5% value-weighted. However, as this value loss is less negative than the loss experienced by the single bidders, it is assumed that multiple bidders can better leverage their prior takeover knowledge to realize synergy gains. For bidder champions, the calendar-time approach finds no evidence of significant value losses at all. Overall, these findings point to a slightly different conclusion than previously assumed in Hypothesis 3.6. While the element of

72

3 Study 1: Determinants of Capital Market Performance

surprise in combination with a potentially higher chance of realizing the basic synergies leads to a significant positive effect on short-term announcement returns for inexperienced single bidders, long-term performance is positively influenced by the acquisition experience and the inherent knowledge of post-merger integration and synergy realization. Firms that have acquired competitors in the past appear to have a higher chance to make another acquisition a success and realize efficiency gains in the long-run. Table 3.22: Abnormal Returns to Acquirers – Differences by Acquirer Experience Fama-French-3-Factor-Model (Subsamples by Bidding Experience) Equal-Weighted Sample

a

b

Multi bidders

-0.002

t-values

-0.78

Stat. Quality

F = 193.94***

Bidder champions

0.000

t-values

-0.05

Stat. Quality Single bidders t-values

6.43

h

a

b

-0.086

-0.005*

0.905***

-1.02

-1.90

20.44

R2adj. = 0.68

0.990*** 0.877*** -0.266** 15.94 F = 104.17***

-0.011*** -2.61

Stat. Quality

0.986*** 0.550*** 22.94

Value-Weighted

s

7.10

R2adj. = 0.56

0.907*** 0.296*** 16.58

F = 94.86***

-2.20

2.64

F = 161.35*** 0.001

s

h

0.219** -0.236*** 2.58

-2.92

R2adj. = 0.66

0.940*** 0.599*** -0.422***

0.17

11.44

F = 58.53***

3.90

-2.95

R2adj. = 0.49

-0.018

-0.002

0.961***

-0.071

-0.245**

-0.16

-0.43

16.35

-0.61

-2.08

R2adj. = 0.50

F = 91.34***

R2adj. = 0.49

This table shows the results of a long-term event-study, based on the Fama-French-3-Factor-model, for the total sample of mergers and acquisitions in the automotive supply industry, differentiated by acquisition experience of the acquirer. Multi bidders have completed more than one merger or acquisition between 1980 and 2007, bidder champions more than five. The intercept 'a' stands for the average monthly abnormal return to the acquiring companies (experienced over a period of three years following the transaction). Results are differentiated between an equal-weighted and a value-weighted approach. All public acquirers between 1980 and 2007 are included. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test. In addition, statistical quality of the model is indicated by the adjusted determination coefficient (adj. R2) and the corresponding Fstatistic.

3.5

Robustness Cheques and Cross-Sectional Regressions

After the preceding univariate analyses have provided first insights into the determinants of short-term announcement returns and long-term capital market performance in the automotive supply industry, the following section focuses on determining

3.5 Robustness Cheques and Cross-Sectional Regressions

73

significant relationships between transaction or acquirer characteristics and the measured performance patterns. Using a multivariate regression analysis, this analysis attempts to find a number of determinants that in turn can be used to make inferences about the acquirers' performance in future takeover deals. Since the previous sections distinguish between short- and long-term capital market performances, the following regression analysis also focuses on the two different horizons: Section 3.5.1 contains a multivariate regression analysis to determine transaction and acquirer characteristics that significantly impact short-term announcement returns, Section 3.5.2 repeats the procedure and analyzes the impact of the same characteristics on the long-term capital market performance.17 3.5.1 Regression of Short-term CAARs

In order to gain further insights into potential dependencies, a cross-sectional regression is performed on the cumulative average abnormal returns to the acquirers using the model presented in Equation (3.10). In total, 21 variables are included in the regression to represent the determinants individually analyzed in Section 3.4 as well as the omitted determinants time, public status of the target, and acquirer continent. In the following, the respective parameter values are specified in detail.

CAARi = a0 + β1 ⋅ Cross _ Border + β 2 ⋅ Cross _ Continental + β 3 ⋅ Diversifying + β 4 ⋅ TransactionValue + β 5 ⋅ 33%MediumTrans. + β 6 ⋅ 33%SmallestTrans. (3.10) + β 7 ⋅ Time2 + β 8 ⋅ Time3 + β 9 ⋅ Time4 + β10 ⋅ Time5 + β11 ⋅ Pr ivate + β12 ⋅ OtherTar

+ β13 ⋅ Engine + β14 ⋅ Drive + β15 ⋅ Electrical + β16 ⋅ Interior + β17 ⋅ Tires + β18 ⋅ Europe + β19 ⋅ Asia + β 20 ⋅ Multi _ Bidder + β 21 ⋅ Bidder _ Champions

17

Since the abnormal returns previously determined by the FF3F-model represent averages of weighted takeover portfolios, the regression in Section 3.5.2 will focus on BHARs as the main indicator of long-term abnormal performance.

74

3 Study 1: Determinants of Capital Market Performance

Cross_Border and Cross_Continental – The results presented in Section 3.4.2.1

provide a first indication that national transactions might have a positive valuation effect on short-term abnormal returns. Following the argument of Mentz and Schiereck (2008), realizing synergies in cross-border and cross-continental transactions appears harder than within country borders. However, since the theory of foreign direct investments (FDI) claims that imperfections in capital, factor, and product markets give multinational firms a competitive edge over local firms, cross-border acquisitions can also be regarded to create more wealth than domestic takeovers (Kang (1993)). Furthermore, transcontinental transactions could stem from the motive to enter new markets or from reasons related to favorable industry-specific prospects (Bley and Madura (2003)). As previous results remain ambiguous, a significant relation underlying this variable would not only clarify the impact but could also provide further information on the motive underlying the move to acquire targets across country borders. Both effects are included in the regression using the dummy variables “Cross_Border” and “Cross_Continental.” Product Scope – As predicted by Hypothesis 3.3, diversifying transactions that

do not involve overlapping product groups are expected to have a negative effect on short-term announcement gains in the automotive supply industry. If the acquirer and the target company have no overlap in produced product segments, the dummy variable "diversifying" assumes the value "1"; otherwise, it remains at "0". Transaction Value – Following the argument of Durand and Vargas (2003),

company size can be regarded as an indicator of a certain level of productive efficiency, accumulated over time. The larger the targets, the higher their implied efficiency and the larger the economies of scale potentially realized through the transactions. To test for the implication of this assumption, the transaction value is included as a variable. In addition, two dummy variables are added to reflect the respective size clusters of the acquisition. If the acquisition belongs to the smallest 33 per cent of all transactions, the dummy variable "33% smallest transactions" assumes the value "1", likewise for the medium 33 per cent of all transactions.

3.5 Robustness Cheques and Cross-Sectional Regressions

75

Time Period – Previous research finds significant differences between abnormal

returns accumulated in different time periods of the merger wave in the automotive supply industry. To test for the significance of the supposed relation, dummy variables for time periods 2 to 5 are included representing six-year periods starting in 1987. If the transaction takes place between 1980 and 1986 ("Time1"), all four dummy variables are "0" ("Time2:" 1987-1992, "Time3:" 1993-1998, "Time4:" 1999-2004, "Time5:" 20042007). Public Status of Target – "Private" represents all targets that are privately held,

"Other Targets" contain all joint ventures and separately sold subdivisions of larger groups. Product Group – Dummy variables comprise the product groups “Drive,” “Engine,” “Electrical,” “Interior,” and "Tires." The acquirer operates in the exterior/chassis

group if all dummy variables are “0.” Acquirer Location – Preceding findings from the automotive supply industry

also provide an indication for a significant difference between the effects M&A have on European, Asian, and American acquirers (Mentz (2006)). In order to test the robustness of this finding, two dummy variables have been included to indicate the acquirer continents "Asia" and "Europe". In addition, dummy variables are included to reflect the acquisition experience of the respective acquirer. If the same acquirer completed at least five transactions between 1980 and 2007, the dummy variable "Bidder_Champion" assumes the value "1". If the acquirer was active more than once, the variable "Multi_Bidder" is set to "1". Consequently, the regression attempts to explain abnormal returns to the total acquirer portfolio using 21 variables. Table 3.23 provides an overview of four regression models regressing the CAARs in the event-windows [-1,0], [-1,1], [-5,5], and [-20,20] against the complete list of variables described above. Although all four models exhibit low or no overall signifi-

76

3 Study 1: Determinants of Capital Market Performance

cance, two significant factor-effects consistently persist: On the one hand, medium and small transactions gain significantly less abnormal returns than large transactions (Hypothesis 3.4). The previously determined differentiation between medium and small transactions only holds true for the largest event-window: Here, medium transactions lose a highly significant 8.2% (1% significant) compared to the large transactions, while small transactions lose 5.5% (10% significant) in direct comparison. On the other hand, transactions between the years 2004 and 2007 represent the second consistent factor effect as they consistently yield higher returns than transactions in earlier time periods. As previously assumed, this pattern could potentially indicate another upcoming merger wave. However, these models do not yield findings sufficiently robust to draw significant conclusions: While two event-windows yield overall insignificant models inherent in low F-values, the [-1,1] and [-20,20] event-windows are only significant at the 10% level. In connection with the low explanatory power as indicated by the provided adjusted R2 values, their overall quality is limited. To arrive at more reliant findings regarding the determinants of short-term announcement effects within the industry, a stepwise regression of the acquirer returns in the smallest event-window [-1;0] is performed as presented in Table 3.24, where the variable with the lowest t-statistic is deleted from one model to the other. The final model, regression 18, represents a highly significant model at the 1%-level (F=4.350), and its Durbin-Watson statistic of 1.806 nearly excludes the possibility of autocorrelation between residuals.

3.5 Robustness Cheques and Cross-Sectional Regressions

77

Table 3.23: Regression of Short-term CAARs to Acquirers Estimated coefficients and t-statistics are determined using a multivariate regression of the CAARs to the acquirers of automotive supply companiesa against a number of explanatory variables, including deal geography, takeover value, time period of the takeover, public status of the target, product group the acquirer operates in, acquirer continent, and the acquisition experience of the acquirer.b

Variable

CAAR [-1,0] Coeftficient value

CAAR [-1,1] Coeftficient value

CAAR [-5,5] Coeftficient value

CAAR [-20,20] Coeftficient value

(Constant)

0.027

1.515

0.042 *

1.683

0.054

1.325

0.097

1.496

CROSS_BORDER

0.008

0.728

0.011

0.684

-0.011

-0.419

-0.013

-0.310 -0.338

CROSS_CONTINENTAL

-0.006

-0.528

-0.015

-0.998

-0.017

-0.705

-0.013

DIVERSIFYING

0.001

0.118

-0.005

-0.509

-0.003

-0.194

-0.036

-1.466

TRANSACTION_ VALUE

0.000

-0.244

0.000

-1.122

0.000

-0.819

0.000

-1.533

33%_MEDIUM_TRANS.

-0.015 *

-1.724

-0.035 *** -2.998

-0.044 **

-2.267

-0.082 ***

-2.677

33%_SMALLEST_TRANS.

-0.020 **

-2.190

-0.034 *** -2.716

-0.044 **

-2.161

-0.055 *

-1.702

1987-1992

0.001

0.040

-0.008

-0.321

-0.005

-0.121

-0.027

-0.422

1993-1998

0.012

0.845

0.013

0.664

0.021

0.631

0.032

0.624

1999-2004

0.012

0.815

0.012

0.558

0.004

0.128

0.013

0.245

2004-2007

0.039 **

2.065

0.057 **

2.161

0.111 **

2.565

0.172 **

2.506

PRIVATE_TARGET

-0.011

-1.118

0.009

0.678

0.007

0.304

-0.004

OTHER_TARGET

0.004

0.479

0.017

1.511

0.018

0.971

0.022

0.783

ENGINE

0.005

0.564

0.004

0.348

-0.001

-0.035

-0.010

-0.336

DRIVE

-0.002

-0.203

-0.010

-0.923

-0.009

-0.501

-0.041

-1.454

ELECTRICAL

-0.016 **

-2.155

-0.020 *

-1.923

-0.017

-1.011

-0.016

-0.609

INTERIOR/TEXTILE TIRES

-0.110

0.001

0.171

0.000

0.017

-0.010

-0.509

-0.001

-0.042

-0.002

-0.147

-0.005

-0.371

-0.035

-1.459

-0.054

-1.437

EUROPE

-0.005

-0.623

0.004

0.383

0.021

1.145

0.034

1.146

ASIA

-0.015

-1.192

-0.008

-0.469

-0.026

-0.927

-0.036

-0.799

MULTI_BIDDER

-0.018 **

-2.311

-0.014

-1.330

0.002

0.092

0.001

0.040

BIDDER_CHAMPION

-0.005

-0.520

-0.006

-0.458

-0.013

-0.550

-0.035

-0.960

R-squared

0.123

0.144

0.141

0.142

Adjusted R-squared

0.023

0.046

0.043

0.044

Durbin-Watson statisticc

1.769

1.476

1.905

1.802

F-statistic

1.231

1.473 *

1.439

1.450 *

Probability (F-stat)

0.229

0.091

0.105

0.100

a

CAARs are derived for a sample of 206 takeover transactions in the automotive supply industry involving publicly-listed acquirers using the market model approach in an event-study. For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively. b

78

3 Study 1: Determinants of Capital Market Performance

Table 3.24: Step-wise Regression of Short-term CAARs to Acquirers Estimated coefficients and t-statistics are determined using a multivariate regression of the CAARs to acquirers of automotive supply companies in the two-day event-window [-1,0]a against a number of explanatory variables.b The variable with the lowest significance is excluded from one regression model to the next regression model.

1

2

CoefVariable (Constant) CROSS_BORDER CROSS_CONTINENTAL

0.027 0.008 -0.006

t

ficient

1.515

0.028 *

0.728

0.008

-0.528 -0.006

4

Coeft

ficient

5

Coeft

ficient

Coeft

ficient

t

1.913

0.028 ** 2.044

0.028 ** 2.092

0.028 ** 2.190

0.749

0.008

0.008

0.008

0.718

-0.525 -0.005

-0.502

-0.529 -0.006

0.745

-0.527 -0.006

0.739

DIVERSIFYING

0.001

0.118

0.001

0.118

TRANSACTION_VALUE

0.000

-0.244

0.000

-0.246

33%_MEDIUM_TRANS.

-0.015 *

33%_SMALLEST_TRANS.

-0.020 ** -2.190 -0.020 ** -2.225 -0.020 ** -2.228 -0.020 ** -2.229 -0.020 ** -2.242

-1.724 -0.015 *

0.000

-1.737 -0.015 *

-0.235

0.000

-1.741 -0.015 *

-0.268

0.000

-1.739 -0.014 *

-0.253 -1.731

1987-1992

0.001

0.040

1993-1998

0.012

0.845

0.012

1.224

0.012

1.227

0.012

1.269

0.012

1.266

1999-2004

0.012

0.815

0.012

1.173

0.012

1.182

0.012

1.244

0.012

1.238

2004-2007

0.039 ** 2.065

PRIVATE TARGET

-0.011

0.039 ** 2.493

0.039 ** 2.498

0.039 ** 2.561

0.039 ** 2.563

-1.118 -0.011

-1.131 -0.011

-1.129 -0.011

-1.138 -0.011

OTHER TARGET

0.004

0.479

0.004

0.479

0.004

0.488

0.004

0.486

0.004

0.493

ENGINE

0.005

0.564

0.005

0.565

0.005

0.571

0.005

0.615

0.005

0.592

-0.213 -0.002

-0.262

-0.002

ELECTRICAL

-0.016 ** -2.155 -0.016 ** -2.161 -0.016 ** -2.211 -0.016 ** -2.253 -0.016 ** -2.266 0.001

-0.203 -0.002

0.171

0.002

-0.200 -0.002

0.179

-0.246 -0.002

-1.147

DRIVE

INTERIOR/TEXTILE

a

ficient

3

Coef-

0.001

0.165

TIRES

-0.002

-0.147 -0.002

-0.145 -0.002

-0.163

0.002

0.210

EUROPE

-0.005

-0.623 -0.005

-0.624 -0.005

-0.620 -0.005

-0.652 -0.005

-0.629

ASIA

-0.015

-1.192 -0.015

-1.196 -0.015

-1.216 -0.015

-1.241 -0.015

-1.292

MULTI_BIDDER

-0.018 ** -2.311 -0.018 ** -2.323 -0.018 ** -2.326 -0.018 ** -2.390 -0.018 ** -2.390

BIDDER_CHAMPION

-0.005

-0.520 -0.005

-0.521 -0.005

-0.541 -0.005

-0.542 -0.005

Adjusted R-squared

0.023

0.028

0.034

0.039

0.044

F-statistic

1.231

1.300

1.375

1.457

1.548 *

Probability (F-stat)

0.229

0.183

0.144

0.110

0.082

Durbin-Watson statisticc

1.769

1.769

1.769

1.764

1.768

-0.522

CAARs are derived for a sample of 206 takeover transactions in the automotive supply industry involving publicly-listed acquirers using the market model approach in an event-study. b For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively.

3.5 Robustness Cheques and Cross-Sectional Regressions

79

Table 3.24: Step-wise Regression of Short-term CAARs to Acquirers (2/4) Estimated coefficients and t-statistics are determined using a multivariate regression of the CAARs to acquirers of automotive supply companies in the two-day event-window [-1,0]a against a number of explanatory variables.b The variable with the lowest significance is excluded from one regression model to the next regression model.

6

7

CoefVariable (Constant) CROSS_BORDER

ficient

8

Coeft

ficient

9

Coeft

ficient

10

Coeft

ficient

0.028 ** 2.186 0.027 ** 2.176 0.029 ** 2.518 0.029 **

Coeft

ficient

t

2.504 0.031 *** 2.693

0.008

0.731 0.008

0.752 0.009

0.817 0.009

0.880 0.009

0.895

-0.005

-0.507 -0.005

-0.516 -0.006

-0.529 -0.006

-0.597 -0.006

-0.597

33%_MEDIUM_TRANS.

-0.014 *

-1.738 -0.014 *

-1.769 -0.013 *

-1.715 -0.013 *

-1.711 -0.013 *

-1.721

33%_SMALLEST_TRANS.

-0.019 ** -2.270 -0.019 ** -2.318 -0.018 ** -2.269 -0.018 ** -2.253 -0.018 ** -2.252

CROSS_CONTINENTAL DIVERSIFYING TRANSACTION_VALUE

1987-1992 1993-1998

0.012

1.262 0.012

1.248 0.012

1.242 0.012

1.235 0.011

1999-2004

0.012

1.235 0.012

1.224 0.012

1.224 0.012

1.222 0.011

1.153

2004-2007

0.038 ** 2.564 0.038 ** 2.563 0.037 ** 2.539 0.037 **

2.547 0.037 **

2.540

-1.690 -0.014 *

-1.709 -0.014 *

-1.746

0.518 0.004

0.492

PRIVATE TARGET OTHER TARGET ENGINE

-0.011

-1.150 -0.011

-1.153 -0.014

0.004

0.506 0.004

0.499

0.597 0.004

0.546 0.004

0.005

1.204

DRIVE

-0.002

-0.259

ELECTRICAL

-0.016 ** -2.266 -0.016 ** -2.377 -0.017 ** -2.424 -0.016 ** -2.393 -0.016 ** -2.353

INTERIOR/TEXTILE TIRES

a

EUROPE

-0.005

-0.652 -0.005

-0.642 0.008

-0.698 -0.005

-0.673 -0.005

-0.659

ASIA

-0.015

-1.290 -0.015

-1.282 0.012

-1.346 -0.015

-1.308 -0.016

-1.354

MULTI_BIDDER

-0.018 ** -2.414 -0.018 ** -2.435 0.007 ** -2.460 -0.019 *** -2.699 -0.019 *** -2.730

BIDDER_CHAMPION

-0.006

-0.571 -0.005

-0.531 0.009

Adjusted R-squared

0.048

0.053

0.057

-0.485 0.060

F-statistic

1.649 *

1.764 **

1.879 **

2.014 **

2.170 **

Probability (F-stat)

0.060

0.042

0.031

0.022

0.015

Durbin-Watson statisticc

1.771

1.776

1.787

1.787

1.790

0.064

CAARs are derived for a sample of 206 takeover transactions in the automotive supply industry involving publicly-listed acquirers using the market model approach in an event-study. b For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively.

80

3 Study 1: Determinants of Capital Market Performance

Table 3.24: Step-wise Regression of Short-term CAARs to Acquirers (3/4) Estimated coefficients and t-statistics are determined using a multivariate regression of the CAARs to acquirers of automotive supply companies in the two-day event-window [-1,0]a against a number of explanatory variables.b The variable with the lowest significance is excluded from one regression model to the next regression model.

11

12

CoefVariable

ficient

13

Coeft

ficient

t

(Constant)

0.030 *** 2.680

0.029 *** 2.631

CROSS_BORDER

0.005

0.003

0.682

14

Coefficient

15

Coeft

0.030 *** 2.746

ficient

Coeft

0.038 *** 4.264

ficient

t

0.039 *** 4.727

0.533

CROSS_CONTINENTAL DIVERSIFYING TRANSACTION_VALUE 33%_MEDIUM_TRANS.

-0.013 *

33%_SMALLEST_TRANS.

-0.018 ** -2.199 -0.017 ** -2.184 -0.017 ** -2.130 -0.016 ** -2.085 -0.016 ** -2.044

-1.664 -0.013 *

-1.653 -0.013 *

-1.657 -0.013 *

-1.761 -0.013 *

-1.764

1987-1992 1993-1998

0.011

1.161

0.011

1.206

0.011

1.205

1999-2004

0.011

1.155

0.011

1.155

0.011

1.180

2.542

0.037 **

2.527

0.036 **

2004-2007 PRIVATE TARGET

0.037 ** -0.014 *

-1.755 -0.014 *

-1.752 -0.013 *

2.489

0.004

0.028 **

-1.695 -0.014 *

0.534

2.189

0.027 **

-1.743 -0.013 *

2.128 -1.702

OTHER TARGET ENGINE DRIVE ELECTRICAL

-0.016 ** -2.334 -0.015 ** -2.292 -0.015 ** -2.278 -0.015 ** -2.246 -0.015 ** -2.239

INTERIOR/TEXTILE TIRES EUROPE

-0.004

-0.536

ASIA

-0.017

-1.433 -0.015

MULTI_BIDDER

-0.019 *** -2.765 -0.019 *** -2.737 -0.019 *** -2.692 -0.019 *** -2.688 -0.019 *** -2.719

-1.343 -0.015

-1.329 -0.013

-1.214 -0.015

-1.351

BIDDER_CHAMPION

a

Adjusted R-squared

0.067

0.071

0.074

0.072

F-statistic

2.342 **

2.557 ***

2.820 ***

2.993 ***

0.075 3.392 ***

Probability (F-stat)

0.010

0.006

0.004

0.003

0.002

Durbin-Watson statisticc

1.789

1.787

1.791

1.778

1.772

CAARs are derived for a sample of 206 takeover transactions in the automotive supply industry involving publicly-listed acquirers using the market model approach in an event-study. b For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively.

3.5 Robustness Cheques and Cross-Sectional Regressions

81

Table 3.24: Step-wise Regression of Short-term CAARs to Acquirers (4/4) Estimated coefficients and t-statistics are determined using a multivariate regression of the CAARs to acquirers of automotive supply companies in the two-day event-window [-1,0]a against a number of explanatory variables.b The variable with the lowest significance is excluded from one regression model to the next regression model.

16

17

CoefVariable (Constant)

ficient

18

Coeft

0.037 *** 4.564

ficient

Coeft

0.035 *** 4.336

ficient

t

0.038 *** 4.901

CROSS_BORDER CROSS_CONTINENTAL DIVERSIFYING TRANSACTION_VALUE 33%_MEDIUM_TRANS.

-0.014 *

33%_SMALLEST_TRANS.

-0.017 ** -2.189 -0.018 ** -2.302 -0.019 ** -2.420

-1.851 -0.015 ** -2.025 -0.016 ** -2.057

1987-1992 1993-1998 1999-2004 2004-2007 PRIVATE TARGET

0.021 * -0.012

1.791

0.016

1.382

-1.519

OTHER TARGET ENGINE DRIVE ELECTRICAL

-0.014 ** -2.171 -0.014 ** -2.136 -0.015 ** -2.254

INTERIOR/TEXTILE TIRES EUROPE ASIA MULTI_BIDDER

-0.018 ** -2.568 -0.016 ** -2.397 -0.018 *** -2.696

BIDDER_CHAMPION

a

Adjusted R-squared

0.072

0.066

F-statistic

3.637 ***

3.878 ***

0.061 4.350 ***

Probability (F-stat)

0.002

0.002

0.002

Durbin-Watson statisticc

1.797

1.828

1.806

CAARs are derived for a sample of 206 takeover transactions in the automotive supply industry involving publicly-listed acquirers using the market model approach in an event-study. b For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively.

82

3 Study 1: Determinants of Capital Market Performance

The final multivariate regression of short-term CAARs as presented in Table 3.24 provides several insights to advance and confirm prior findings from the univariate analyses. First of all, the intercept representing the overall announcement effect to acquirers in this industry is highly significant at +3.8%. This reconfirms the short-term assumption of Hypothesis 3.1 and the findings presented in section 3.4.1. Once again, it points out the exceptional return position of the automotive supply industry vis-à-vis other industries. The second finding relates to the impact of transaction size. Both the medium and the smaller size clusters have a negative impact on the overall announcement effect of -1.6% and -1.9% respectively. However, these returns only partly concur with the empirical findings presented in section 3.4.2.3. While larger transactions seem to yield more positive announcement returns as presented before, a differentiation between medium and small transactions is not confirmed on the basis of this regression. Nevertheless, the short-term prediction of Hypothesis 3.4 holds true within the presented regression model. In addition, the regression confirms the negative outlying position of the electrical product sector in the observed two-day event-window (H3.5). With regard to other industries, acquirers with operations in this field realize a significant 1.5% less upon the announcement of a takeover. It could therefore be assumed that the particular industry characteristics previously attributed to the tire product sector actually apply to the electrical components suppliers. However, as Section 3.4.3.1 reveals that larger eventwindows revert the effect, the evidence remains inconclusive. A last finding from this regression can be drawn from the impact of acquisition experience (H3.6). Consistent with the findings presented in Section 3.4.3.2, bidders that previously completed a takeover within the observed time period realize a significantly lower announcement return by 1.8%. This finding is consistent with the assumption that capital markets value potentially unrealized synergies from previous takeovers negatively by applying a discount. However, the regression model does neither confirm the hypotheses concerning geographical expansion (H3.2) and product diversification (H3.3), nor the additional impact of timing, public status of the target and acquirer continent.

3.5 Robustness Cheques and Cross-Sectional Regressions

83

3.5.2 Regression of Long-term BHARs

After analyzing the short-term announcement effects in the previous section, this section shifts focus to the determinants of the acquirers' long-term capital market performance. Therefore, the same regression procedure as applied in Section 3.5.1 is now applied to the BHARs determined in Section 3.4.1. The underlying model again follows Equation (3.10) with the exception of the variable "Time5." This variable is omitted from the BHAR- regression because the BHARs represent a three-year holding period with an implicit latest start in 2004. Consequently, the regression attempts to explain abnormal returns to acquirers using 20 variables. Table 3.25 provides an overview of the complete regression models of 12-, 24-, and 36-month BHARs against all 20 variables. The 24- and 36-month models are both highly significant at the 1%-level, explanatory power is remarkably high with adjusted R2 ranging between 10% and 20%, potential autocorrelation problems can be neglected based on the high Durbin-Watson-Statistics. The overall abnormal long-term performance as represented by the constant yields a positive 13.7% for the 12-month BHARmodel and becomes increasingly negative for the longer models. Over 36 months, the constant yields an insignificant underperformance of -57.7%. Although being less significant, this finding corresponds to the underperformance determined in the univariate analysis above and to the prediction of Hypothesis 3.1. In addition, the two models involving longer time periods also indicate a first set of significant determinants of long-time performance. First of all, cross-border transactions are determined to have a significant negative impact on the long-term performance of acquirers and, thereby, confirm the assumption of the internationalization hypothesis (H3.2). Over the 36-month holding period, cross-border transactions decrease the overall performance by a highly significant 81%. The cross-continental variable yields a significant positive influence of about 60% and appears to positively influence the performance. However, the cross-continental variable has to be interpreted in connection with a cross-border effect as it cannot appear on its own. So while the overall cross-

84

3 Study 1: Determinants of Capital Market Performance

border effect confirms earlier findings from Section 3.4.2 and from prior research (Conn et al. (2005)), which finds post-merger integration to be more challenging in more complex international settings, the cross-continental variable seems to offset this effect slightly although leaving an overall negative impact on performance. This could potentially be the result of two effects: Firstly, following the argument of Bley and Madura (2003) as well as Mentz and Schiereck (2008), the motive to enter new markets as well as favorable industry-specific prospects in form of a global presence in the automotive triad (North America, Europe, Asia) could favorably influence long-term performance, i.e. decrease the overall loss from engaging in a crossborder transaction. Secondly, as cross-continental deals require significantly larger resources, the negative cross-border effect might in this case be distorted by a positive size effect observed for very large transactions. Nonetheless, it can be concluded that cross-border transactions negatively influence long-term performance while advantages of a cross-continental presence or transaction size appear to offset the negative effect slightly. The ultimate motive cannot be determined as part of this study. The second main finding relates to the diversification hypothesis (H3.3). Over the first two holding periods, diversifying transactions without a product overlap significantly underperform focusing transactions by 18 and 28% respectively. Over the longest event-window, there is still a negative but insignificant influence. In this case, these findings advance prior results presented in Section 3.4.2.2. Diversification seems to have a negative overall effect on post-acquisition performance as the realization of synergy and efficiency potentials is more difficult across different product groups. The contradicting results from the FF3F-model only hold true for an equal-weighted approach that can be regarded as inferior to a value-weighted approach representing the true investor return. With regard to transaction size (H3.4), the findings are not fully consistent with the univariate results presented above. In the regression model, there is a positive effect of small-transactions on long-term performance while the returns to acquirers in me-

3.5 Robustness Cheques and Cross-Sectional Regressions

85

dium transactions drop from positive to negative across the three holding horizons (all insignificant). It appears that prior assumptions seem to hold: Large transactions yield less significant negative long-term returns while small transactions even point to positive returns for the acquirer. However, inferences are limited due to lack of statistical significance. However, the findings concerning the influence of different products sectors on long-term performance of the acquirer are again more consistent (H3.5). Compared to the 'Exterior' product group, suppliers in the 'Drive Components' and 'Electrical Components' component sectors appear to significantly underperform by 32% and 44% respectively over a three-year holding period. Returns to electrical component suppliers are insignificantly more negative than returns to exterior component suppliers, tire supplier appear to even generate more positive returns than any other product group. This might be an indication for the special position of these product groups in the proximity of other industries outside automotive suppliers. Bidding experience, though, has a positive influence on long-term performance, the experience hypothesis (H3.6) seems to hold as more experienced acquirers outperform single acquirers in integrating the target and realizing synergy potentials. In addition, the findings in Table 3.25 also indicate a positive influence of European and Asian transactions on long-term performance, North-American acquirers seem to underperform.

86

3 Study 1: Determinants of Capital Market Performance

Table 3.25: Regression of Long-term BHARs to Acquirers Estimated coefficients and t-statistics are determined using a multivariate regression of the BHARs to the acquirers of automotive supply companiesa against a number of explanatory variables, including deal geography, takeover value, time period of the takeover, public status of the target, product group the acquirer operates in, acquirer continent, and the acquisition experience of the acquirer.b

Variable (Constant) CROSS_BORDER CROSS_CONTINENTAL DIVERSIFYING

24-month BHARs Coeft-value ficient

36-month BHARs Coeft-value ficient

0.137

0.527

-0.436

-1.208

-0.577

-1.230

-0.157

-1.064

-0.474 **

-2.313

-0.810 ***

-3.043

0.138

1.005

1.297

0.574 **

-0.184 **

-2.114

0.248 -0.282 **

-2.335

-0.226

2.314 -1.440

TRANSACTION_VALUE

0.000

0.340

0.000

0.058

0.000

-0.735

33%_MEDIUM_TRANS.

0.053

0.466

0.005

0.030

-0.022

-0.107

33%_SMALLEST_TRANS. 1987-1992

0.087

0.695

0.355 **

2.046

0.314

1.390

-0.050

-0.205

0.414

1.233

0.678

1.554

1993-1998

0.065

0.308

0.345

1.179

0.539

1.418

1999-2004

-0.234

-1.075

-0.052

-0.171

0.206

0.524

PRIVATE_TARGET

-0.066

-0.492

0.090

0.480

0.126

0.516

0.091

0.889

0.110

0.772

-0.056

-0.304

ENGINE

-0.006

-0.053

DRIVE

-0.132

-1.338

OTHER_TARGET

0.220 -0.287 **

1.530

0.141

0.753

-2.100

-0.318 *

-1.789

ELECTRICAL

-0.045

-0.485

-0.077

-0.597

-0.086

-0.517

INTERIOR/TEXTILE

-0.181

-1.607

-0.425 ***

-2.710

-0.436 **

-2.137

TIRES

-0.105

-0.784

0.014

0.075

0.112

0.460

0.080

0.735

0.385 **

2.556

0.478 **

2.443

-0.071

-0.427

0.405 *

1.740

0.524 *

1.733

0.075

0.741

0.484 ***

3.426

0.463 **

-0.024

-0.205

EUROPE ASIA MULTI_BIDDER BIDDER_CHAMPION

a

12-month BHARs Coeft-value ficient

-0.049

-0.301

-0.015

R-squared

0.137

0.289

0.225

Adjusted R-squared

0.017

0.189

0.117

Durbin-Watson statisticc

1.756

1.806

1.930

F-statistic

1.140

2.901 ***

2.076 ***

Probability (F-stat)

0.317

0.000

0.007

2.524 -0.073

BHARs are derived for a sample of 164 takeover transactions in the automotive supply industry between 1980 and 2004. b For a detailed description of the variables and the underlying equation see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively.

3.5 Robustness Cheques and Cross-Sectional Regressions

87

Table 3.26: Step-wise Regression of 36-month BHARs to Acquirers Estimated coefficients and t-statistics are determined using a multivariate regression of the 36-month BHARs to the acquirers of automotive supply companiesa against a number of explanatory variables, including deal geography, takeover value, time period of the takeover, public status of the target, product group the acquirer operates in, acquirer continent, and the acquisition experience of the acquirer.b

1

2 Coefficient

3 Ceofficient

4 Coefficient

Variable

Coefficient

(Constant)

-0.577

CROSS_BORDER

-0.810 *** -3.043 -0.808 *** -3.061 -0.812 *** -3.111 -0.824 *** -3.204

CROSS_CONTINENTAL DIVERSIFYING

0.574 **

t

-1.230 -0.579 2.314

0.572 **

-0.226

-1.440 -0.225

TRANSACTION_VALUE

0.000

33%_MEDIUM_TRANS.

-0.022

t

-1.241 -0.594 2.329

0.574 **

t

-1.340 -0.639 2.350

0.576 **

t -1.525 2.367

-1.446 -0.226

-1.469 -0.225

-1.463

-0.735 0.000

-0.740 0.000

-0.764 0.000

-0.713

-0.107 -0.022

-0.109

33%_SMALLEST_TRANS.

0.314

1.390

0.314

1.397

0.329 *

1.851

0.327 *

1.846

1987-1992

0.678

1.554

0.680

1.566

0.685

1.592

0.693

1.618

1993-1998

0.539

1.418

0.539

1.423

0.542

1.440

0.543

1.446

1999-2004

0.206

0.524

0.206

0.526

0.210

0.540

0.212

0.547

PRIVATE TARGET

0.126

0.516

0.125

0.515

0.123

0.511

0.168

0.855

-0.312 -0.058

-0.319

OTHER TARGET ENGINE DRIVE

-0.056 0.141

-0.304 -0.057 0.753

0.139

-0.318 *

-1.789 -0.315 *

ELECTRICAL

-0.086

-0.517 -0.085

INTERIOR/TEXTILE

-0.436 ** -2.137 -0.437 **

0.752

0.140

-1.823 -0.318 *

0.756

0.147

-1.870 -0.322 *

0.807 -1.901

-0.515 -0.084

-0.512 -0.078

-0.478

-2.165 -0.441 **

-2.233 -0.445 **

-2.263 0.456

TIRES

0.112

0.460

0.112

0.464

0.110

0.459

0.109

EUROPE

0.478 **

2.443

0.480 **

2.483

0.483 **

2.535

0.493 *** 2.634

ASIA

0.524 *

1.733

0.526 *

1.758

0.524 *

1.761

0.530 *

MULTI_BIDDER

0.463 **

2.524

0.460 **

2.615

0.462 *** 2.673

0.464 *** 2.690

BIDDER_CHAMPION

-0.015

-0.073

Adjusted R-squared

0.117

0.123

0.129

0.134

F-statistic

2.076 ***

2.201 ***

2.338 ***

2.485 ***

Probability (F-stat)

0.007

0.005

0.003

0.002

Durbin-Watson statisticc

1.930

1.929

1.928

1.929

a

1.791

BHARs are derived for a sample of 164 takeover transactions in the automotive supply industry between 1980 and 2004. For a detailed description of the variables and the underlying equation, see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively. b

88

3 Study 1: Determinants of Capital Market Performance

Table 3.26: Step-wise Regression of 36-month BHARs to Acquirers (2/3) Estimated coefficients and t-statistics are determined using a multivariate regression of the BHARs to the acquirers of automotive supply companiesa against a number of explanatory variables, including deal geography, takeover value, time period of the takeover, public status of the target, product group the acquirer operates in, acquirer continent, and the acquisition experience of the acquirer.b

5

6 Coefficient

7 Ceofficient

8 Coefficient

Variable

Coefficient

(Constant)

-0.615

CROSS_BORDER

-0.817 *** -3.191 -0.801 *** -3.155 -0.796 *** -3.143 -0.794 *** -3.141

CROSS_CONTINENTAL DIVERSIFYING TRANSACTION_VALUE

0.566 **

t

-1.483 -0.426 * 2.342

0.568 **

t

-1.699 -0.467 * 2.353

0.557 **

t

-1.944 -0.456 * 2.322

0.558 **

t -1.910 2.330

-0.240

-1.605 -0.232

-1.562 -0.222

-1.507 -0.223

-1.519

0.000

-0.679 0.000

-0.632 0.000

-0.640 0.000

-0.614

33%_MEDIUM_TRANS. 33%_SMALLEST_TRANS.

0.327 *

1.850

0.331 *

1.881

0.349 **

2.018

0.346 **

2.006

1987-1992

0.726 *

1.726

0.528 **

2.202

0.535 **

2.237

0.547 **

2.298

0.365 **

2.232

0.369 **

2.269

0.381 **

2.362

0.773

0.153

0.792

0.151

0.782

0.643

0.109

1993-1998

0.557

1.493

1999-2004

0.221

0.573

PRIVATE TARGET

0.164

0.838

0.150

0.133

0.743

0.113

OTHER TARGET ENGINE DRIVE

-0.349 ** -2.210 -0.347 **

-2.200 -0.360 **

ELECTRICAL

-0.096

-0.584

INTERIOR/TEXTILE

-0.466 ** -2.443 -0.460 **

-0.608 -0.092

-2.421 -0.448 **

0.624 -2.316 -0.340 **

-2.240

-2.376 -0.465 **

-2.496

TIRES EUROPE

0.508 *** 2.764

0.511 *** 2.786

0.524 *** 2.884

0.526 *** 2.903

ASIA

0.543 *

0.557 *

0.575 **

0.559 *

MULTI_BIDDER

0.477 *** 2.810

0.466 *** 2.770

0.457 *** 2.736

0.463 *** 2.778

Adjusted R-squared

0.139

0.143

0.146

0.150

F-statistic

2.641 ***

2.808 ***

2.998 ***

3.212 ***

Probability (F-stat)

0.001

0.001

0.000

0.000

Durbin-Watson statisticc

1.942

1.934

1.918

1.909

1.845

1.906

1.981

1.939

BIDDER_CHAMPION

a

BHARs are derived for a sample of 164 takeover transactions in the automotive supply industry between 1980 and 2004. For a detailed description of the variables and the underlying equation, see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively. b

3.5 Robustness Cheques and Cross-Sectional Regressions

89

Table 3.26: Step-wise Regression of 36-month BHARs to Acquirers (3/3) Estimated coefficients and t-statistics are determined using a multivariate regression of the BHARs to the acquirers of automotive supply companiesa against a number of explanatory variables, including deal geography, takeover value, time period of the takeover, public status of the target, product group the acquirer operates in, acquirer continent, and the acquisition experience of the acquirer.b

9

10 Coefficient

11 Ceofficient

Variable

Coefficient

(Constant)

-0.486 ** -2.085 -0.440 *

CROSS_BORDER

-0.787 *** -3.124 -0.755 *** -3.035 -0.745 *** -2.983

CROSS_CONTINENTAL DIVERSIFYING

0.548 ** -0.217

t

2.297

0.538 **

-1.483 -0.222

t

t

-1.942 -0.576 *** -2.759 2.261

0.534 **

2.237

-1.518

TRANSACTION_VALUE 33%_MEDIUM_TRANS. 33%_SMALLEST_TRANS.

0.375 **

2.265

0.370 **

2.240

0.368 **

2.219

1987-1992

0.552 **

2.327

0.554 **

2.335

0.571 **

2.403

1993-1998

0.378 **

2.350

0.400 **

2.521

0.405 **

2.542

0.161

0.842

-0.342 ** -2.257 -0.354 **

-2.350 -0.304 **

-2.061

-0.463 ** -2.491 -0.463 **

-2.497 -0.425 **

-2.304

1999-2004 PRIVATE TARGET OTHER TARGET ENGINE DRIVE ELECTRICAL INTERIOR/TEXTILE TIRES EUROPE

0.529 *** 2.923

0.511 *** 2.848

0.502 *** 2.788

ASIA

0.559 *

0.533 *

0.561 *

1.957

MULTI_BIDDER

0.453 *** 2.737

0.420 *** 2.615

0.416 **

2.580

Adjusted R-squared

0.153

0.155

0.148

F-statistic

3.462 ***

3.720 ***

3.829 ***

Probability (F-stat)

0.000

0.000

0.000

Durbin-Watson statisticc

1.912

1.927

1.930

1.944

1.866

BIDDER_CHAMPION

a

BHARs are derived for a sample of 164 takeover transactions in the automotive supply industry between 1980 and 2004. For a detailed description of the variables and the underlying equation, see Chapter 3.5. c The Durbin-Watson statistics are estimated to test for autocorrelation of the residuals. Conclusions on this test are challenged against a visual test for normality and equal variance. ***, **, * denote statistical significance at the 1%, 5%, and 10% level respectively. b

90

3 Study 1: Determinants of Capital Market Performance

Although the overall models exhibit high statistical quality, the results are again challenged by a step-wise regression similar to the approach previously applied to shortterm CAARs. Table 3.26 contains the results and reveals that regression model 11 is able to increase overall significance to an F-value of 3.829. In addition to the determinants presented above, this model also reveals a highly significant overall negative long-term performance in form of the intercept at -57.6% (1% significant) and a positive impact of the earlier time periods from 1987 to 1998. The later could potentially be the result of the overall consolidation effort: As the merger wave in the automotive supply industry turned toward its peak in 1998, easy to realize takeover targets increasingly disappear from the market. Consequently, long-term performance decreases over the time periods.

3.6

Conclusion

The objective of this study was to measure and analyze the short- and long-term performance of acquirers in the automotive supply industry. For this purpose, a sample of 230 horizontal takeover transactions involving automotive suppliers between 1981 and 2007 was identified and analyzed for the short- and long-term capital market performance of acquirers. In contrast to the majority of prior research on post-acquisition capital market performance, this study focuses on a single industry, the automotive supply industry, and applies a combination of the short-term market model and the two leading long-term methodologies: the matching-firm approach in event-time and the Fama-French-3-Factor-model in calendar-time. The derived results provide new insights into the perceived long-term success of transactions in the automotive supply industry and its corresponding evaluation through capital markets. Firstly, the results indicate that acquirers in the automotive supply industry are not able to sustain their positive announcement returns in the long-run. Although this study confirms the outstanding positive short-term returns to acquirers as identified by prior research (Mentz and Schiereck (2008)), both long-term performance models indi-

3.6 Conclusion

91

cate a consistent value destruction of 16% to 20% over a three-year period following an M&A announcement. This negative finding is consistent with, but yet more negative than the negative returns previously determined by Loughran and Vijh (1997) or Mitchell and Stafford (2000) in cross-industry studies. Therefore, it appears that the above-average synergy potentials perceived by capital markets in the short-term perspective cannot be sufficiently realized by the acquirers within the three years following a transaction. Secondly, this study finds that geographic expansion and transaction size consistently influence both short- and long-term capital market performances. On the one hand, the derived return patterns provide evidence for a significantly negative influence of international diversification. Consistent with the findings of Conn et al. (2005) and Aw and Chatterjee (2004), calendar- and event-time approaches consistently yielded a negative impact of internationally diversifying transactions on long-term value prospects. The more challenging post-merger integration of targets in cross-border transactions negatively impacts the post-merger performance of automotive suppliers. At the same time, the short-term announcement returns are significantly higher for transactions with a national focus. On the other hand, this study also determines a consistent positive effect of larger transactions on short- and long-term performance given a higher probability to experience economies of scale, as argued by Ferris and Park (2002). Thirdly, this study also reveals that bidding experience of the acquirer has a different effect on short-term announcement returns than on long-term performance: While short-term returns are negatively influenced for multi-acquirers, their long-term performance is significantly better than the returns to single acquirers. Consistent with the results of Antoniou and Zhao (2004), it can therefore be assumed that multi-bidders carry experience in integrating takeover targets, which enables them to consistently outperform their inexperienced peer group in the long-run. The short-term announcement effects are stronger for inexperienced acquirers representing some initial synergies as well as the element of surprise.

92

3 Study 1: Determinants of Capital Market Performance

The influence of product diversification remains inconclusive and could not ultimately be determined in this study. However, a corresponding acquisition strategy to maximize long-term capital market performance comprises the integration of smaller, national targets to gain experience as a first step. Afterwards, the gained experience could be leveraged in a cross-continental deal of significant size in order to gain access to the global synergy structure of this industry.

4

Study 2: Does Operating Performance Meet Market Expectations?

4.1

Introduction

In the past, empirical research has applied two approaches to assess the performance of mergers and acquisitions (M&A): one evaluating stock returns (event studies) and one evaluating accounting profits (accounting studies). Although both approaches have been frequently applied over the last decades, only event studies have converged towards a consistent conclusion about the value creation potential of M&A: While the transactions seem to initially create value for the combined entities (Bruner (2002)), slightly negative announcement returns for acquirers are found to turn into consistent and significant value losses in long-term event studies.18 By contrast, accounting studies yield a more ambiguous view on the overall value creation potential as conclusions in different studies vary from significant performance losses to significant gains.19 However, since accounting studies evaluate actual past performance of acquirers, understanding the determinants behind long-term operating performance is key to the longterm success of companies engaging in M&A activity. The automotive supply industry has been facing increased merger activity since the 1990s. The pressure to produce better equipped and less expensive automobiles created a growing trend towards specialization and internationalization of the industry. While some product ranges are now dominated by very few players, acquisitions also led to geographical expansion of suppliers across borders and across continents (Sadler (1999)). In the light of these conditions, previous research on M&A announcement returns identifies the industry as an outlier: Acquiring companies are able to realize significant positive short-term returns as an expression of the global synergy and efficiency 18

19

See, for example, Agrawal et al. (1992), who find a significant value loss of about 10% to acquiring companies over a five-year period following merger completion. Over various studies, performance changes observed for combined entities range from -2% (Ravenscraft and Scherer (1989)) to +2.8% (Healy et al. (1992)) from before to after a transaction.

94

4 Study 2: Does Operating Performance Meet Market Expectations?

potential underlying the transactions (Mentz and Schiereck (2008)). However, Chapter 3 also shows that these acquirers are not able to sustain the positive returns beyond a short-term perspective. In the long-run, they fail to realize synergies and experience significant value losses. Given the ambiguity in previous accounting studies, further research is needed to challenge these capital market returns against the long-term accounting performance within the industry. While the long-term underperformance in share returns potentially serves as an indicator for the operating performance among acquirers in the industry, research still leads an on-going discussion about how the results of event and accounting studies correlate and whether they could be treated as substitutes. While Healy et al. (1992) find a positive relationship between the two approaches, Fridolfsson and Stennek (2005) develop an economic model to show that both approaches have to be used as complements. More recently, Betzer and Metzger (2007) find no statistical relationship between the two. For the automotive supply industry, this study attempts to determine a correlation between a three-year performance change measure focusing on the postmerger period as an accounting indicator and 36-month Buy-and-Hold Abnormal Returns (BHARs). Consequently, this study determines whether acquirers in this industry are consistent long-term underperformers or able to surprise by fulfilling the initial market expectations about superior synergy potentials. The objective of this study is twofold: First of all, this study determines the long-term abnormal accounting performance based on a sample of 93 mergers and acquisitions in the automotive supply industry between 1981 and 2003. For this purpose, the change in several accounting performance indicators over the three years after a transaction is tracked and measured against the same indicator of non-merging matching firms from the same industry. In addition, the observed long-term performance patterns are analyzed to detect and categorize potentially underlying determinant variables. To interpret the findings, a regression analysis is used to determine correlations and to test for statistical significance. Secondly, this study examines the relationship between the

4.2 Literature Review and Hypotheses

95

observed abnormal changes in accounting performance and previously observed 36month BHARs on the same 93 transactions in order to assess the degree to which acquirers deliver market expectations in their operating performance. Overall, this study finds significant evidence that acquirers in the automotive supply industry underperform their non-merging rivals between -4% and -12%, depending on the observed performance indicator. As this result concurs with previously determined negative BHARs and is likewise influenced by similar determinants, this study also determines a strong positive correlation between capital market returns and the observed accounting performance. As a result, it is assumed that an analysis of firm profitability cannot only serve as a substitute for short-term announcement returns but also for long-term event studies based on Buy-and-Hold Abnormal Returns. The remainder of this study is organized as follows: Section 4.2 provides a brief overview of the relevant literature and outlines the hypotheses concerning the long-term accounting performance and its correlation with event-study results. Section 4.3 presents the sample selection procedure as well as the employed methodology. The following Section 4.4 contains the empirical results including the overall long-term accounting performance and its respective determinants. In addition, Section 4.4 also complements the results with a cross-sectional analysis on the underlying value drivers. Section 4.5 presents the results of the correlation analysis between event- and accounting-study results before Section 4.6 summarizes the findings and concludes.

4.2

Literature Review and Hypotheses

4.2.1 Related Literature

In contrast to the large number of existing event studies, studies comparing the financial performance of merging firms before and after M&A transactions remain fewer in number and yield mixed results. On the one hand, some studies support the argument that M&A lead to a significant reduction in the merging firm's profitability.

96

4 Study 2: Does Operating Performance Meet Market Expectations?

For example, Ravenscraft and Scherer (1989) analyze the pre- and post-merger performance of 471 US corporations between 1957 and 1977 and conclude that, on average, merger activity decreases the ratio of operating income to total assets by -2.75%. The authors also analyze the impact of different accounting methods and attribute the majority of the observed effect to deals completed under the purchasing accounting method. Dickerson, Gibson, and Tsakalotos (1997) support this negative view by analyzing 613 UK transactions between 1948 and 1977: under a simple model, becoming an acquirer reduces the rate of return on net assets by -2.4%; including determinants into a more detailed model yields significant value performance losses of approximately -1% (1% significance level). On the other hand, empirical evidence for a significant performance increase through M&A also exists. Healy et al. (1992) analyze accounting data from annual reports of the 50 largest transactions between US public firms between 1979 and 1984. Applying an industry-adjusted operating cash flow (CF) to total asset measure, they conclude that the merging firms outperform their industry peers by a significant +2.8% (1% significance level). This performance increase is particularly strong for firms with overlapping businesses. Ramaswamy and Waegelein (2003) find a significant increase of +17.2% (1%-level) by applying the approach of Healy et al. (1992) to a sample of 162 transactions in the later time period of 1975 to 1990. This performance increase is positively influenced by dissimilar industries as well as mergers that happened before 1983. After 1983, the observed positive accounting effect diminishes to an insignificant 4.8%. Yet more recently, Betzer and Metzger (2007) apply the same procedure to a sample of 160 US transactions between 1981 and 2000 and find an insignificant performance increase of 0.21% compared to industry peers and 0.51% compared to nonindustry peers. In addition, studies analyzing the accounting performance in certain industries or regions yield mixed results as well. For example, Cornett and Tehranian (1992) analyze 30 large bank mergers between 1982 and 1987 and thereby extend the methodol-

4.2 Literature Review and Hypotheses

97

ogy of Healy et al. (1992) to the regulated banking industry. Overall, they conclude that merging banks outperform their peers by a significant 1.2%. Cheng and Leung (2004) find no significant improvement of accounting-based performance in 36 mergers in Hong-Kong between 1984 and 1996. Liu, Chen, and Pai (2007) analyze the performance of 60 Taiwanese telecommunications companies between 1993 and 2003: They conclude that M&A activity reduces overall technical efficiency, a measure derived by weighting 20 financial indicators in a data envelopment analysis, by approximately 3%. Overall, it becomes apparent that previous accounting studies, both general and in specific industries or regions, yield mixed results. There appears to be no consistent performance impact of M&A on the merging firms. Besides analyzing the general impact of M&A activity on accounting performance, previous literature also discusses the correlation of event- and accounting-study results. To do so, these studies usually compare the long-term operating performance as measured by an accounting indicator to the short-term announcement returns to test whether stockholder expectations expressed in short-term returns hold true in the longrun. However, given the diversity in the results described above, findings concerning the relationship of the two approaches are also contradicting. Main contributions include Healy et al. (1992) who find a strong positive relationship between operating performance and the short-term returns between announcement and delisting of the target. Consequently, they argue that both approaches can be used as substitutes. Fidolfsson and Stennek (2005), by contrast, develop a model arguing that both approaches have to be used as complements. If a merger confers strong negative externalities, i.e. companies outside a deal are worse off than the participating parties, share prices of the merging firms might increase even though profitability decreases. In such a situation, a company entering into an M&A situation has a strong incentive to pre-empt its competitors from becoming a successful acquirer. As a result, its profitability decreases while capital markets credit the fact that it did not become an outsider exposed to externalities. Betzer and Metzger (2007) compare their long-term accounting results

98

4 Study 2: Does Operating Performance Meet Market Expectations?

to the 41-day announcement returns to the combined entities and find no significant statistical relationship. Therefore, they support previous literature that both approaches have to be used as complements. This study compares the abnormal changes in accounting performance primarily to the corresponding Buy-and-Hold Abnormal Returns (BHARs) over a 36-month horizon. Studies focusing on BHARs are more numerous and have been consistently pointing towards a significant value loss in terms of share return: Loughran and Vijh (1997) use a sample of 947 US transactions between 1970 and 1989 and find negative Buyand-Hold Abnormal Returns of -6.5% within a five-year time period following the transaction. Mitchell and Stafford (2000) calculate BHARs for a sample of 2,068 US transactions between 1961 and 1993 and report insignificant negative returns of -1% (equal-weighted) and significant abnormal returns of -3.8% (value-weighted) over a three-year period. Similar results are presented by Black et al. (2001), Sinha (2004), Gregory and Matatko (2004).20 Given the consistency of these findings, it is assumed that a regression based on BHARs will yield a more consistent view on the substitutability of the long-term event- and accounting-study approaches. 4.2.2 Hypotheses 4.2.2.1 The Overall Effect

While previous research provides mixed results concerning the effect of M&A activity on accounting performance, some clear indications for the potential effect in the automotive supply industry do exist. First of all, the share price returns within the industry have been extensively analyzed: Mentz and Schiereck (2008) identify the industry as an outlier with significant positive announcement returns to acquirers of approximately

20

For the automotive supply industry, Chapter 3 determines a significant long-term value loss: Based on a sample of 164 transactions, 36-month BHARs of -16.68% (equal-weighted) and -17.38% (valueweighted) are derived for the acquirers. However, as the original sample is reduced to yield those with available accounting data, a detailed description of the results is omitted at this point. An update of the findings with a corresponding sample is contained in the following sections.

4.2 Literature Review and Hypotheses

99

+2%. In the long-run, acquirers are not able to sustain this position but lose a significant -17% compared to their peers (Chapter 3). Following the argument of Fridolfsson and Stennek (2005), the extraordinary positive short-term announcement returns could be an indication for the existence of strong externalities. Consequently, the acquirers in this industry would be entering unprofitable deals to prevent becoming an outsider themselves. Their long-term profitability should therefore decrease to reflect these moves. However, in the light of the observed negative long-term BHARs for acquirers, the interpretation of this theoretical model is limited: The existence of strong externalities would also have negatively influenced long-term share returns of the peers. A second indicator lies in previous accounting studies with industry focus. Liu et al. (2007) provide evidence for a negative effect of M&A on the performance of acquirers in the telecommunications industry. They argue that the observed effect is due to a lack of efficiency increase through the transactions, the pre-merger performance gap between target and acquirer does not suffice to offset the premium paid. For the automotive supply industry, the strong trends towards specialization and globalization (Sadler (1999)) in connection with the merger wave of the 1990s could be an indication that transactions have been completed without realizing the full synergy potential originally underlying the transaction. Eventually, profitability decreases with the realization of additional acquisitions over time. Lastly, since long-term share returns frequently react to published information, the observed underperformance in long-term share returns (BHARs) can be assumed to be more directly linked to accounting performance than the short-term announcement returns applied in previous studies. It is therefore hypothesized that the accounting performance also reveals a significant underperformance as compared to the non-merging industry peers. H4.1: Acquirers in the automotives supply industry experience a long-term abnormal decrease in financial performance following M&A activity.

100

4 Study 2: Does Operating Performance Meet Market Expectations?

4.2.2.2 The Impact of Transaction and Acquirer Characteristics

Previous research on long-term BHARs in the automotive supply industry shows that the significant underperformance of acquirers is mainly driven by crossborder transactions, diversification in product lines, inexperienced bidders and insufficient transaction size (Chapter 3). However, a comparably comprehensive list of determinants for the accounting performance could not be obtained. Only the diversification in businesses has been previously discussed: Dividing their sample in mergers of low, medium and high business overlap, Healy et al. (1992) support the positive influence of focusing mergers on the post-merger performance of acquirers. Similarly, Megginson, Morgan, and Nail (2004) show that focusing mergers result in marginal improvements in long-term performance while focus-decreasing mergers realize a significant decline of operating cash flows three years after the transaction. Therefore, it is assumed that operational performance in the automotive supply industry is likewise positively influenced by non-diversifying transactions. Given the continuous inclusion of new information in share returns, it is assumed that the other determinants influence accounting results in a similar way as they influence long-term share price returns. H4.2: The financial performance of acquirers in the automotive supply industry will be positively influenced by regional focus, product focus, acquisition experience, and transaction size. 4.2.2.3 The Correlation of Accounting- and Event-Study Results

The diverging opinions on the correlation of accounting- and event-study results might be the result of the specific returns analyzed: Most of the studies compare the short-term announcement returns over a few days with the long-term performance changes over five years.21 Thereby, the authors are able to evaluate whether the longterm performance is already correctly anticipated and incorporated at the merger an-

21

See, for example, Healy et al. (1992) and Betzer and Metzger (2007).

4.3 Data and Methodology

101

nouncement. However, this approach does not answer the question whether both approaches can be used as substitutes in determining long-term success of M&A. Therefore, this study extends prior research by also comparing long-term performance directly to long-term share price returns. In addition, the accounting analysis focuses completely on published accounting data. A revaluation of total assets to market value of assets, as for example performed by Healy et al. (1992), is intentionally omitted. Given the argument that share prices should over time reflect the financial performance of a firm, it is therefore assumed to find a strong positive correlation between the longterm event- and accounting-study results in the automotive supply industry. H4.3: Accounting- and event-study results positively correlate in the automotive supply industry.

4.3

Data and Methodology

4.3.1 Identifying Merging Companies

The sample of mergers and acquisitions for the empirical study is drawn from the SDC/Thomson One Banker database and the Bloomberg M&A database. It includes all takeover events announced between January 1st, 1981, and September 1st, 2007. The total number of M&A transactions is reduced to yield only those transactions meeting the following criteria.22

At the time of the takeover announcement, the target and acquirer company both possess active operations in the automotive supply industry.

The acquirer intends to purchase 50% or more of the outstanding shares or of the private equity, for publicly traded and privately held targets respectively.

22

In addition, all targets and acquirers were double-checked by a press research using the Factiva database to ensure that all transactions are horizontal and that announcement dates are correct as provided by the databases. All non-horizontal deals as well as deals involving financial investment companies have been excluded.

102

4 Study 2: Does Operating Performance Meet Market Expectations?

The total transaction value accumulates to at least USD 50 million.

The acquiring company is located in one of the following geographic regions: Europe, North and South America, and Asia. The described selection criteria result in a takeover sample of 230 events in the

automotive supply industry. After eliminating private acquirers and acquirers with insufficient data availability, 164 transactions remain which are challenged against their available accounting data. For each acquirer, accounting information including 'CF', 'EBIT', 'EBITDA', 'Sales', and 'Total Assets' is obtained from Standard&Poors (S&P), which consolidates information from the Compustat database (for US firms), Global Vantage and Xpressfeed databases (for other global firms) and manually collected data from the companies' annual reports. The sample is then reduced by another 43 transactions with incomplete data over the four-year time period from the merger year (t = 0) to the third year (t = 3) following the transaction. An additional 17 transactions yield negative asset and/or sales values in year 0 or year 3 and are therefore considered to be inconsistent for the purpose of this study. For 11 transactions, the matching firms carry performance growth rates larger than 300% over the three-year time period following the transaction. It is assumed that this performance growth is rather the result of restructuring than of organic growth. As organic growth is one of the conditions for serving as a non-merging matching firm and since the matches are kept consistent with the matches determined in Chapter 3, these 11 transactions are also dropped from the sample. These described selection criteria result in a final takeover sample of 93 events to be used in this accounting study. Table 4.1 provides an overview of the frequency distribution over time and reveals a strong concentration of events between the years 1995 and 2000. The average transaction value ranges from USD 94.1 million to USD 1,100.0 million, whereas the later is mainly driven by one larger acquisition by Eaton Corporation in 1993. Overall, it becomes apparent that both the frequency distribution as well as the regional affilia-

4.3 Data and Methodology

103

tion of the sample is not significantly altered through the sample reduction. The majority of the transactions still take place in the 1995 to 2000 time period; 69% of the deals are initiated through an American acquirer (as opposed to 60% in the original sample). Therefore, it is assumed that the reduced sample is still representative of the M&A activity in the automotive supply industry. In order to determine long-term BHARs for this reduced sample, monthly adjusted stock prices as well as monthly market values and yearly market-to-book ratios for all acquirers are downloaded from the Thomson Datastream database.23 Table 4.1: Overview of the Transaction Sample – Descriptive Statistics

Year

Transactions

(%)

Average Transaction Value (USD million)

Acquirer Region Number of Transactions Americas Europe Asia

1986 1988 1989 1990 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

1 3 4 1 1 1 2 6 10 15 10 20 8 3 4 4

1.1 3.2 4.3 1.1 1.1 1.1 2.2 6.5 10.8 16.1 10.8 21.5 8.6 3.2 4.3 4.3

200.0 453.1 269.6 239.8 355.0 1100.0 166.2 209.3 312.4 366.3 309.2 421.9 353.7 94.1 266.0 286.9

1 1 2 4 9 13 6 11 4 2 3 1

Total

93

100.0

337.6

64

1 3 3

1 1

2 1 2 4 6 3

3 1 1

1 1

2

20

9

4.3.2 Portfolio of Matching Firms

Both approaches, determining abnormal accounting performance as well as long-term BHARs, require a set of comparable, non-merging firms from the same in23

To reflect the influence of dividend payments as well as share issuances or repurchases on return data, the adjusted stock prices denoted by data type “RI” were selected.

104

4 Study 2: Does Operating Performance Meet Market Expectations?

dustry. Since there is no broad global index available for the automotive supply industry which comforts the purpose of this study, this study determines a portfolio of matching firms from the constituents' lists of broader country-specific indices. A first step in constructing the portfolio of matching firms includes screening the constituents' lists underlying the country-specific industry indices for companies in the automotive supply industry. Downloading all available constituents' lists of indices in the automotive parts industry as supplied and constructed by the Thomson Datastream database results in 42 lists of international automotive suppliers.24 In a second step, all 42 constituent's lists are aggregated and compared to the respective lists from prior time periods. In a third step, the initial set of 109 companies derived from the Datastream database is crossreferenced to a list of the top 100 OEMs in the automotive supply sector as published on a yearly basis by the Automotive News (Automotive News (2003)). Another 21 publicly traded automotive supply companies are added to yield a total number of 130 matching firms. The final portfolio contains 49 Asian, 46 North American, and 35 European publicly traded companies from the automotive supply industry. For each firm, monthly stock returns, monthly market values, and yearly market-to-book ratios are downloaded from the Thomson Datastream database; accounting information is collected in the same way as described for the acquirers. The matches between acquirers and non-merging industry peers are kept consistent with the ones applied in Chapter 3 and follow the character-based matching procedure as proposed by Lyon et al. (1999):

For each public acquiring firm from the takeover sample, the relevant size (market value) and market-to-book ratio are obtained. Market values are determined as the last quote in the last June preceding the transaction, market-

24

Constituents' Lists are denominated by the industry-specific MNEMONIC code “LAUPRT” plus a two-digit country-specific suffix. For example, a list of US automotive suppliers constituting the automotive parts industrial index in the US can be derived using the code “LAUPRTUS.”

4.3 Data and Methodology

105

to-book ratios are derived for the last completed fiscal year before the transaction.25

Likewise, for each year, market values at the end of June and market-to-book ratios are determined for the full list of matching firms. For each acquisition, the total list of potential matches is then reduced to yield only those firms for which both descriptive data fields as well as monthly return data are available.

In a second step, the list of potential matches is reduced to those companies with a market value in the range of 70% to 130% of the acquirer's market value. This procedure ensures an overall comparability between acquirer and its match.

Finally, from the list of companies with a market value between 70% and 130% of the acquirer's market value, the one with the smallest absolute difference in market-to-book ratio is selected as the control-firm for the analysis.

4.3.3 Econometric Strategy

To measure the impact on operating performance, the majority of the more recent studies employ the same or a derivation of the operating cash flow return measure as originally proposed by Healy et al. (1992).26 Their approach scales the operating cash flow of a firm by its market value of assets in order to circumvent potential biases from different accounting standards and from inter-temporal comparisons. This study employs a similar, yet stricter performance indicator. Operating cash flows (CF) are measured for each acquirer and rival as net operating profit less adjusted taxes (NOPLAT) plus depreciation and goodwill expenses, while NOPLAT is defined as sales minus cost of goods sold (COGS), selling, general, and administrative expenses (SG&A), depreciation and goodwill expenses, and adjusted taxes. As described by Equation (4.1), the

25 26

All information is downloaded from the Thomson Datastream database. See, for example, Ramaswamy and Waegelein (2003) and Betzer and Metzger (2007).

106

4 Study 2: Does Operating Performance Meet Market Expectations?

operating cash flows are then scaled by the book value of assets (BVA) in order to arrive at a comparable performance measure across firms and time.27

(4.1)

Pt * =

CFt * BVAt *

; for * ȯ (Acquirers, Rivals)

While this performance indicator ensures a stricter separation of accounting and market information, it is at the same time exposed to a potential manipulation through different accounting standards (Betzer and Metzger (2007)). Therefore, a second performance indicator using 'total sales' as a denominator is employed. As opposed to book value of assets, this later indicator is assumed to be less sensitive to different accounting standards and deal structures. In order to generate a comprehensive view on the accounting performance changes and to separate the effect of depreciation and amortization on performance, this study also employs EBIT and EBITDA as two additional nominators for the performance. EBITDA is defined as sales minus COGS and SG&A; EBIT is derived by further deducting depreciation and amortization from EBITDA. Based on Equation (4.1), a total of six performance indicators (CF/BVA; CF/Sales; EBIT/BVA; EBIT/Sales; EBITDA/BVA; EBITDA/Sales) are measured and tracked for each acquirer and non-merging matching firm from the transaction year (t = 0) to the third year following the transaction (t = 3). In order to determine the abnormal accounting performance of the involved acquirers, this study focuses on their post-merger performance development over the three years following the transactions. Consequently, the original methodology of Healy et al. (1992) is slightly adjusted in two ways: Firstly, the approach applied in this study mimics the BHAR-approach as proposed by Barber and Lyon (1997) and Lyon et al. (1999). Therefore, instead of constructing a separate benchmark portfolio, the original matches 27

The consistency of the applied Compustat data items with the information from other data sources as provided by Standard&Poors is manually probed and therefore assumed to be valid in the course of this study.

4.3 Data and Methodology

107

as derived under the matching procedure in Chapter 3 apply as a benchmark.28 Secondly, this study concentrates more strictly on the post-merger performance development than the pre- and post-merger comparison applied in Healy et al. (1992). Switzer (1996) proposes a similar analytic focus, when she uses the first available published accounting information after the transaction as a starting point for the analysis while refraining from including pre-merger information. s +T

(4.2)

s +T

APi ,T = ∏ (1 + g x ,i ,t ) −∏ (1 + g x ,control ,t ) t =s

; with:

t =s

APi ,T = Abnormal performance change of company i over T years g x ,i ,t = Yearly performance change in x of company i in year t g x ,control ,t = Yearly performance change in x of the control-firm in year t

The abnormal performance change of an acquirer i (APi) is derived as the difference between the change of a performance indicator x of the acquirer i over T years and the same indicator change of the control-firm (see Equation (4.2)). This measure is determined for a period of three years following the takeover announcement, with t = 0 being the year of the announcement. The average abnormal performance for the total sample and for the respective subsamples is calculated as an equal-weighted portfolio average. Statistical significance is tested using standard t-statistic. A Wilcoxon-SignedRank-Test serves as a test for statistical significance of differences in subsample means. In addition, a second approach based on a linear regression model is applied to generate additional evidence for the existence of an effect of M&A on accounting performance. This model regresses the median peer-group adjusted performance indicator of firm i after the merger (IAPpost,i) on the same measure in the transaction year (IAP0,i) as described by Equation (4.3). The intercept Į is defined as the measure of abnormal 28

Becker-Blease, Goldberg, and Kaen (2008) follow a similar reasoning and apply a control sample based on size and operating performance in their analysis of post-merger operating performance.

108

4 Study 2: Does Operating Performance Meet Market Expectations?

peer-adjusted post-merger performance change; the slope coefficient ȕ describes the correlation between the accounting performance in year 0 and the median after-merger performance. This approach corresponds to a regression approach performed by Betzer and Metzger (2007), while again excluding pre-merger performance from the analysis.

IAPpost ,i = α + β * IAP0,i + ε i

(4.3)

The corresponding Buy-and-Hold Abnormal Returns for determining a potential correlation between both approaches are derived by using the standard BHAR-approach as proposed by Lyon et al. (1999): Replacing the indicator change g in Equation (4.2) with the respective monthly rate of return r yields that abnormal returns (BHARs) are calculated as the difference between the Buy-and-Hold Return (BHR) of an investor in the acquiring company and the BHR of an investor in the control-firm. BHARs are determined for a period of 36 months following the takeover announcement. The average BHARs are calculated as an equal-weighted average and as a value-weighted average where the respective market value at the end of the preceding June serves as the weight. Statistical significance is tested using standard t-statistic. Since rebalancing and new listing biases do not exist within the control-firm approach and since Barber and Lyon (1997) also show that skewness can be neglected in this context, no further adjustments to test-statistics are necessary. The applied regression models to test for the correlation between the two approaches are described in Section 4.5.

4.4

Empirical Results

4.4.1 The Overall Effect The first aspect to be analyzed in the course of this study is the overall effect of mergers and acquisitions on the post-merger performance of acquirers. Table 4.2 provides an overview of the average performance indicators of acquirers and non-merging

4.4 Empirical Results

109

control-firms over the three years following a transaction. At first sight, the comparison of means carries a number of remarkable insights: Firstly, the average 'CF/BVA'-ratios of about 9% are consistently lower than the ones derived by Healy et al. (1992) or Betzer and Metzger (2007), who both arrive at a median measure of above 20%. This difference can partly be attributed to the application of asset book values rather than market values as well as to the inherent focus on a single, more resource-driven industry. In general, it is therefore assumed that the derived indicators are valid and applicable.29 Table 4.2: Average Performance Following M&A Transactions Average Post-Merger Performance

Relative Year

CF/

CF/

EBIT/

EBIT/

EBITDA/

EBITDA/

BVA

Sales

BVA

Sales

BVA

Sales

Acquirers 0

10.65%

9.27%

9.50%

8.20%

14.04%

12.16%

+1

10.61%

9.05%

9.17%

7.71%

14.05%

11.93%

+2

10.37%

8.88%

8.76%

7.46%

13.58%

11.59%

+3

9.83%

8.58%

7.87%

6.75%

12.66%

10.99%

MDN[+1,+3]

9.95%

8.87%

8.37%

6.75%

13.08%

11.24%

0

9.84%

9.29%

9.08%

7.86%

13.43%

12.32%

+1

9.80%

9.38%

8.81%

7.87%

13.28%

12.48%

+2

9.69%

8.61%

8.23%

6.65%

12.90%

11.24%

+3

9.42%

8.76%

7.67%

6.62%

12.31%

11.29%

Control-Firms

MDN[+1,+3] 9.40% 8.45% 6.60% 6.62% 12.28% 10.95% This table shows the average performance of acquiring companies and non-merging control-firms over three years following mergers and acquisitions in the automotive supply industry. Performance is measured using six different performance indicators. All transactions between 1980 and 2004 are included for which the selection criteria described in Section 4.3.1 apply (n=93).

Secondly, the average level of performance is higher for acquirers than for nonmerging rivals across the majority of indicators and years. This observation seems to

29

Other studies focusing on single industries also determine different performance levels than provided by cross-industry studies: For example, Cornett and Tehranian (1992) determine pre-tax cash-flow returns on assets for the banking industry of 2 to 3%.

110

4 Study 2: Does Operating Performance Meet Market Expectations?

contradict Hypothesis 4.1 as it indicates that acquirers outperform their control-firms. However, averages aggregated across deals and time make the observation biased and invalid. Finally, it becomes apparent that all six performance indicators are consistently declining for both acquirers and non-merging rivals over the three years following a transaction. While a decline in the performance of acquirers would by itself be in line with the proposed argument under Hypothesis 4.1, its consistency across the industry more generally points towards an industry trend influencing performance. The existence of such a trend, in turn, represents the importance to control for industry-wide effects in order to determine the incremental effect of M&A transactions (Cornett and Tehranian (1992)). To address potential industry effects, this study follows preceding literature and employs a peer-adjusted performance measure to arrive at a reliable indication of the abnormal performance of acquirers. Table 4.3 presents the growth rates of different performance indicators for acquirers and control-firms over the three years following a transaction. For acquirers, all six performance indicators significantly decrease (means and medians) over the observed time period. While this decrease confirms the descriptive pattern observed in Table 4.2, the realized growth rates for the non-merging rival companies are less consistent. In this case, the mean growth rate for the CF-returns becomes positive, while statistical significance generally suffers from diverging means and medians. Obviously, the decrease in performance averages does not necessarily translate into a performance down-trend of non-merging rivals. Some rival companies appear to outperform the industry with strong performance growth rates. Since the growth rates are derived by following a character-based matching procedure as described in Section 4.3, they overcome the averaging and the industry factor biases inherent in Table 4.2, making a comparison across deals and time possible. For all six performance indicators, acquirers in this industry underperform their matching firms up to -11.98% (CF/BVA-growth). Statistical significance in either test is given for

4.4 Empirical Results

111

the CF/BVA-, CF/Sales-, and EBITDA/Sales-indicators.30 As argued before, the later denominator appears to be least sensitive to different accounting methods and policies and is therefore the most reliable indicator. Based on this indicator, the observed patterns promote the assumption that the performance development of acquirers is significantly worse than the development of non-merging rivals. Overall, Hypothesis 4.1 is supported. Table 4.3: Abnormal Performance Changes in the Automotive Supply Industry Performance Change Over Three Years After the Transaction Acquirers Mean Performance Indicator

Median

t-value

CF/BVA

-4.32%

-1.66 **

-7.16% CF/Sales

-4.73%

-2.21 **

-10.26% -2.57 ***

-11.80% -3.54 ***

-6.78%

-7.28% -9.15%

1.28

-11.98% -1.87 0.03 **

5.78%

1.09

-5.35% -0.98

-8.23% -1.83 **

-0.93% -0.20 -5.05%

-3.54 ***

-3.53% -1.10 -4.10%

t

p

Wilcox. Sign.-R. Z

p

-1.19

0.23

-1.75

0.08 *

-0.87

0.39

-1.04

0.30

-1.19

0.24

-1.86

0.06 *

-3.75%

-8.33% -2.61 ***

-10.16% EBITDA/Sales

Median

-13.01%

-9.40% EBITDA/BVA

7.66%

t-Test

t-value

-0.42%

-13.61% EBIT/Sales

Median

Difference Mean

-4.66%

-5.93% EBIT/BVA

Control-Firms Mean

-10.52% -1.98 0.02 ** -5.23% -4.90% -0.82 0.21 -1.05% -3.57% -0.75 0.23 -8.23% -5.85% -1.16 0.12 -3.20% -3.75% -1.13 0.13 -4.71%

This table shows the change in different performance indicators of acquiring companies and non-merging control-firms over the 3 years following transactions in the automotive supply industry. Means across transactions are derived as simple averages. The difference displayed describes the abnormal performance change of acquirers and is derived under a control-firm matching approach as proposed by Lyon et al. (1999). All changes relate to a three-year period starting in the year of the merger. All acquirers between 1980 and 2004 were included for which the selection criteria as described in Section 4.3.1 apply. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test as well as a Wilcoxon Signed-Rank Test (Wilcox. Sign.-R.). 30

All differences described in Table 4.3 represent simple averages across the 93 transactions. Appendix 4 provides an overview of the respective value-weighted averages with the transaction volume serving as the weighting factor. Since the overall pattern including significance levels does not change significantly, the presented results are assumed to be robust to outliers. For the remainder of this study, the simple averages will serve as the basis for further analysis.

112

4 Study 2: Does Operating Performance Meet Market Expectations?

Table 4.4: Regression of Median Post-Merger Abnormal Performance Regression Results Performance Indicator CF/BVA

a

t-value

b

-0.12%

-0.54

0.793

t-value 20.55 ***

Statistical Quality F = 422.19 *** R2adj. =

CF/Sales

0.05%

0.13

0.476

9.41 ***

F = 88.47 *** R2adj. =

EBIT/BVA

-0.18%

-0.56

0.771

15.82 ***

0.23%

0.69

0.560

9.76 ***

-0.12%

-0.36

0.787

17.23 ***

0.09%

0.22

0.602

11.95 ***

0.51

F = 296.82 *** R2adj. =

EBITDA/Sales

0.73

F = 95.29 *** R2adj. =

EBITDA/BVA

0.49

F = 250.28 *** R2adj. =

EBIT/Sales

0.82

0.76

F = 142.80 ***

R2adj. = 0.61 This table shows the results of the models regressing the median annual peer-group adjusted performance indicator of firm i after the merger (IAPpost,i) on the same measure in the transaction year (IAP0,i). The intercept 'a' is defined as the measure of abnormal post-merger performance change; the slope coefficient 'b' describes the correlation between the combined performance in year 0 and the median after-merger performance. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test for individual coefficients and the F-Test for the overall model respectively.

The negative long-term performance is in line with literature promoting a negative impact of M&A on firm profitability, including Ravenscraft and Scherer (1989) and Dickerson et al. (1997). It advances these studies by applying a different methodology of tracking post-merger performance changes against the performance of matching firms and by focusing on the automotive supply industry as a whole. In addition, this finding seems to confirm earlier findings based on share price returns where acquirers underperform their matches by a significant 17% over three years (Chapter 3). Alternatively, Table 4.4 presents the results of an additional regression model of the median peer-adjusted post-merger performance, a derivation of the frequently-used approach in more recent studies such as Betzer and Metzger (2007). Overall, it becomes apparent from the coefficients b that there is a strong positive correlation between the peer-

4.4 Empirical Results

113

adjusted performances in the merger year and the median peer-adjusted post-merger performances. Therefore, there is no reason to assume that pre-merger outperformers as approximated by the combined performances in the merger year lose their position over the three years following the transaction and become underperformers. Consequently, a significant post-merger abnormal performance impact through M&A, as inherent in the intercept a, cannot be obtained by applying this methodology.31 4.4.2 Determinants of Profitability After the previous section has shown that CF-returns of acquirers in the automotive supply industry seem to underperform those of non-merging rivals, this section focuses on analyzing the determinants of this effect. One of the determinants analyzed in previous research is the effect of diversifying versus focusing transactions: Both, Healy et al. (1992) and Megginson et al. (2004), find a positive relationship between focusing mergers and post-merger performance of acquirers. Likewise, prior research in the automotive supply industry determines a positive influence of non-diversifying transactions on long-term post-merger share returns: diversifying transactions lead to a significant value loss of about 25% over three years (5% significant), whereas focusing transaction yielded an insignificant loss of 10% (Section 3.4.2.2). It appears that focusing transactions yield a greater chance of realizing synergies. In order to define diversifying and focusing transactions, one common approach is the comparison of four-digit SIC codes.32 However, the applicability of SIC codes for the purpose of this study is limited. On the one hand, SIC codes generally depend on the data source and may change over time (Kahle and Walkling (1996)). On the other hand,

31

32

Betzer and Metzger (2007) find an insignificant increase of 0.21% (0.51%) compared to industry (non-industry) matches and conclude that there is no abnormal effect observable based on this methodological approach. Given the significant findings from the approach presented in Table 4.3, it is therefore assumed that the applicability of the later method is limited. For the remainder of this study, the discussion will focus on the results presented in Table 4.3. For example, Comment and Jarrell (1995) use industry classifications via four-digit SIC codes to identify firms with similar focus of operations within the same industry; Eckbo (1992) uses SIC codes to describe horizontal and vertical mergers within a given sample.

114

4 Study 2: Does Operating Performance Meet Market Expectations?

SIC codes are not available in sufficient detail to distinguish between product groups within the automotive supply industry. Therefore, this study clusters all acquiring and matching companies according to their operational focus into six product groups: 'Exterior Body Work and Chassis', 'Engine and Combustion-Related Components', 'Driving Components', 'Electrical Parts', 'Interior Parts', and 'Tires.'33 Companies operating in more than one of these product sectors are attributed to each of the first three sectors stated either by the Automotive News in their business description of the company or the respective company home pages. All transactions in which acquirer and target are part of the same cluster are treated as "within product group/intra-industry," all others represent "diversifying transactions." Table 4.5 provides an overview of the performance changes for the resulting subsamples "intra-industry" and "diversifying". It becomes apparent that especially diversifying transactions decrease acquirer performance significantly, namely -8.05%. Focusing on intra-industry transactions results in an insignificant and smaller decrease of -2.12%. The same pattern holds true for the abnormal performance as described by the growth differences. Again, only diversifying transactions yield a significant underperformance of -13.37% for acquirers compared to their non-merging matches. The observed patterns can serve as an indication for a negative effect of diversification on profitability of the acquiring firms; the integration of targets from other product sectors appears to require significant resources that, in turn, decrease profitability above industry-average. This result is consistent with previous findings on the diversification effect (Healy et al. (1992), Megginson et al. (2004)) and complements prior research from the industry: Over a long-term perspective, M&A activity outside the own product sector does not only carry the risk of failing to realize inherent synergies but also requires significant resources for integrating the target. While the resource requirement becomes 33

The product groups are developed on the basis of the business descriptions applied by the Automotive News (2003). For a detailed description of the products groups, see Section 3.4.2.2.

4.4 Empirical Results

115

apparent in a decreased profitability as shown here, the increased likelihood of failure has been previously determined as a significant underperformance in stock returns (Chapter 3.4.2.2). A similar argument applies to the effect of internationalization on profitability. Again, cross-border transactions rather tend to be associated with significantly higher financial and non-financial resources than national transactions. Therefore, the internationalization effect on profitability is analyzed next. Table 4.5: Abnormal Performance Change – Differences by Product Scope Performance Change in CF/Sales – Differentiated by Product Scope Acquirers Mean

Control-Firms Mean

Difference Mean

t-Test

Wilcox. Sign.-R.

Sample

Median

t-value

Median

t-value

Median

t

Z

p

Intra-Industry

-2.12%

-0.86

6.15%

0.92

-8.27%

-1.24

0.11

-1.41

0.16

-2.19 **

5.32%

0.62

-13.37%

-1.55

0.06 *

-0.95

0.34

(n=52)

-3.33% -8.05%

Diversifying (n=41)

-8.19%

5.31%

-8.51%

p

-5.45%

-5.10%

This table shows the change in CF/Sales of acquiring companies and non-merging control-firms over the three years following transactions in the automotive supply industry, differentiated by product scope of the transaction. Means across transactions are derived as simple averages. The difference displayed describes the abnormal performance change of acquirers and is derived under a control-firm matching approach as proposed by Lyon et al. (1999). All changes relate to a three-year period starting in the year of the merger. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test as well as a Wilcoxon Signed-Rank Test (Wilcox. Sign.-R.).

Table 4.6 documents performance changes within the three subsamples of national, cross-border and cross-continental transactions. As with product diversification, a similar pattern holds: The observed performance decrease is stronger for international than for national transactions. National transactions reduce profitability by -4.04% while cross-border deals result in a -5.28% decrease. Cross-continental transactions yield a highly significant performance reduction of -6.68%. Significant abnormal performance changes can only be obtained for cross-border and cross-continental transactions with -13.12% and -16.78% respectively. The underperformance of crosscontinental transactions is consistently significant under both tests applied. As a result, cross-border transactions significantly reduce profitability by requiring significant re-

116

4 Study 2: Does Operating Performance Meet Market Expectations?

sources and by posing a larger challenge to realize transaction synergies across countryborders. Hypothesis 4.2 seems to hold for its diversification and internationalization assumptions. Table 4.6: Abnormal Performance Change – Differences by Geographical Scope Performance Change in CF/Sales – Differentiated by Regional Scope Acquirers Mean Median

Sample Cross-Continental (n=41)

-6.68%

(n=52)

(n=41)

t-value

Median t-value

-0.64% -1.68 **

7.84%

-1.44 *

3.17% -0.96%

Median -16.78%

t-Test t

p

Wilcox. Sign.-R. Z

p

-2.08 0.02 **

-1.98

0.05 **

-1.96 0.03 **

-1.79

0.07 *

-0.84 0.20

-0.60

0.55

-6.53% 1.18

-13.12%

0.37

-7.21%

0.04%

-6.62%

Difference Mean

-2.63 *** 10.10% 1.23

-5.90% -4.04%

National

Mean

-5.87% -5.28%

Cross-Border

Control-Firms

-6.12%

-4.90%

This table shows the change in CF/Sales of acquiring companies and non-merging control-firms over the three years following transactions in the automotive supply industry, differentiated by regional scope of the deal. Means across transactions are derived as simple averages. The difference displayed describes the abnormal performance change of acquirers and is derived under a control-firm matching approach as proposed by Lyon et al. (1999). All changes relate to a three-year period starting in the year of the merger. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test as well as a Wilcoxon Signed-Rank Test.

In order to challenge these findings and test for a potential influence of other determinants, a cross-sectional regression is performed on the abnormal peer-adjusted performance growth (APi) based on the CF/Sales-indicator. Besides product diversification (intra_industry) and internationalization (cross_border; cross_continental), the model as described by Equation (4.4) also includes dummy variables testing for the effect of transaction size, transaction timing, public status of the target, acquirer continent, and acquisition experience of the acquirer. These additional effects and the corresponding variables are specified below.

4.4 Empirical Results

117

APi = a0 + β1 ⋅ cross _ border + β 2 ⋅ cross _ continental + β 3 ⋅ intra _ industry (4.4)

+ β 4 ⋅ trans _ value _ usd + β 5 ⋅ time93_98 + β 6 ⋅ time99_04 + β 7 ⋅ private _ target + β 8 ⋅ other _ target + β 9 ⋅ europe _ acq + β10 ⋅ asia _ acq + β11 ⋅ multi _ bidder + β12 ⋅ bidder _ champ Transaction Value – A highly valued target can be regarded as an indicator for

productive efficiency (Durand and Vargas (2003)). If the target company is highly productive, realizing additional synergies through the transaction becomes more difficult. In general, more resources are needed to integrate a larger target, which in turn would decrease the abnormal performance. Following the argument of resource requirements, larger transactions should negatively influence operating performance. Time Period – Previous research finds that abnormal performance of acquirers depends on the respective time periods in which the transaction takes place (Ramaswamy and Waegelein (2003)). To test for the significance of this supposed relation, dummy variables for the time periods between 1993 to 1998 and 1999 to 2004 are included into the model.34 If both time variables assume the value “0,” the takeover took place before 1993. Public Status of Target – Dummy variables are also included to test for a significant relationship between the public status of the target and the abnormal operating performance of acquirers. "Private_target" represents all targets that are privately held, "other_target" contains all other non-public targets including joint ventures and separately-sold subdivisions of larger groups. Acquirer Location – In order to test for the effect of acquirer location on abnormal performance, two dummy variables are included to differentiate between the acquirer continents "Asia" and "Europe." If the acquirer is located in North America, the two variables take a value of “0.”

34

As presented in Table 4.1, 1998 represents the peak year of M&A activity within the industry.

118

4 Study 2: Does Operating Performance Meet Market Expectations?

Additional dummy variables reflect the acquisition experience of the respective acquirer. If the same acquirer completes at least five transactions within the total sample (1980-2004), the dummy variable "bidder_champ" assumes the value "1". If the acquirer is active more than once, the variable "multi-bidder" assumes "1". For single acquirers, both variables remain at "0." Consequently, the regression attempts to explain abnormal performance using 12 variables. Table 4.7 provides an overview of the stepwise regression model starting with the full set of variables described above. With each step between models 1 and 10, the least significant variable is dropped until the most significant overall model remains. However, the overall statistical quality of all models including the final model 10 remains fairly poor. The F-statistic of 1.79 does not suffice to generate a significant model at the 10% significance-level; explanatory power (adj. R2) of 3% limits the overall interpretation of the results. Therefore, the derived results can only serve as an indication; reliable inferences about the determining relationship cannot be obtained on the basis of this model. Nevertheless, model 10 yields a significant negative intercept of -17%, which is in line with the overall results derived above. In addition, the insignificant negative effect of cross-border transactions also promotes the univariate results presented above. A resulting positive effect for European acquirers and transactions involving other/subsidiary targets potentially point toward additional effects influencing abnormal performance. However, given the poor statistical quality of the model, its interpretation is not overvalued.

4.4 Empirical Results

119

Table 4.7: Stepwise Regression of Abnormal Performance Changes Estimated coefficients and t-statistics are determined using a multivariate regression of the abnormal changes in CF/Sales of acquirers in the automotive supply industry over a three-year perioda against a number of explanatory variables.b The variable with the lowest significance is excluded from one model to the other.

Regression Model Coefficients 1

2 t

Coef.

3 t

Coef.

4 t

Coef.

5 t

Coef.

6

Variables

Coef.

t

Coef.

t

Constant

-0.10 -0.43 -0.10 -0.43 -0.10 -0.47 -0.09 -0.43 -0.12 -0.64 -0.11 -0.57

cross_border

0.08 0.39 0.08 0.40 0.08 0.42 0.09

cross_continent

-0.20 -1.04 -0.20 -1.06 -0.20 -1.08 -0.20 -1.08 -0.20 -1.09 -0.14 -1.20

0.45 0.08 0.43

intra_industry

0.07 0.61 0.07 0.61 0.07 0.63 0.07

0.63 0.07 0.64 0.07 0.62

trans_value_usd

0.00 -0.05 0.00 -0.05

time 93-98

-0.17 -0.84 -0.17 -0.84 -0.17 -0.85 -0.16 -0.83 -0.15 0.64 -0.16 -0.86

time 99-04

-0.10 -0.49 -0.10 -0.50 -0.10 -0.51 -0.09 -0.48 -0.09 -0.81 -0.09 -0.48

private_target

0.04 0.21 0.04 0.21 0.04 0.23

other_target

0.17 1.15 0.16 1.16 0.17 1.24 0.15

europe_acq

0.18 1.08 0.18 1.11 0.18 1.12 0.17

1.10 0.17 1.32 0.21 1.53

asia_acq

0.08 0.38 0.08 0.39 0.08 0.40 0.08

0.39 0.09 1.11 0.10 0.49

multi_bidder

-0.03 -0.23 -0.03 -0.24 -0.03 -0.27 -0.04 -0.33

1.34 0.15 -0.46 0.15 1.37

bidder_champ

0.00 -0.02

Adjusted R2

-0.06

-0.04

-0.03

-0.02

-0.01

F

0.59

0.65

0.72

0.81

0.90

1.01

Significance

0.85

0.78

0.70

0.61

0.52

0.43

0.00

DWS 2.05 2.05 2.05 2.05 2.05 2.04 The abnormal performance changes are derived for a sample of 93 transactions in the automotive supply industry involving publicly listed acquirers measured against their non-merging peers. The matching procedure follows a character-based matching approach as proposed by Barber and Lyon (1997). b For a detailed description of the variables, see Section 4.4.2. ***, **,* denote statistical significance at the 1%, 5%, and 10% level respectively.

a

120

4 Study 2: Does Operating Performance Meet Market Expectations?

Table 4.7: Stepwise Regression of Abnormal Performance Changes (2/2) Estimated coefficients and t-statistics are determined using a multivariate regression of the abnormal changes in CF/Sales to acquirers in the automotive supply industry over a three-year perioda against a number of explanatory variables.b The variable with the lowest significance is excluded from one model to the other.

Regression Model Coefficients 7

8 t

Coef.

9 t

Coef.

10 t

Coef.

11

Variables

Coef.

Constant

-0.17 -1.38 -0.17 -1.33 -0.12 -1.13 -0.17 -1.87 *

-0.14 -1.39

-0.15 -1.31 -0.13 -1.24 -0.13 -1.23 -0.15 -1.38

-0.28 -1.71 *

cross_border cross_continent intra_industry

t

Coef.

t.

0.22 1.36

0.07 0.67

0.08 0.70

0.07 0.62

trans_value_usd time 93-98

-0.09 -0.79 -0.10 -0.97 -0.10 -0.97

time 99-04 private_target other_target

0.15 1.36

0.15 1.40

0.15 1.39

0.16 1.51

europe_acq

0.20 1.49

0.19 1.43

0.17 1.36

0.19 1.45

asia_acq

0.09 0.47

multi_bidder bidder_champ Adjusted R2

0.01

0.02

0.02

0.03

0.00

F

1.16

1.36

1.58

1.79

1.06

Significance

0.34

0.25

0.19

0.15

0.37

DWS 2.03 2.03 2.05 2.02 2.05 a The abnormal performance changes are derived for a sample of 93 transactions in the automotive supply industry involving publicly listed acquirers measured against their non-merging peers. The matching procedure follows a character-based matching approach as proposed by Barber and Lyon (1997). b For a detailed description of the variables, see Section 4.4.2. ***, **,* denote statistical significance at the 1%, 5%, and 10% level respectively.

4.5 The Correlation of Accounting- and Event-Study Results

121

Besides the stepwise regression, model 11 focuses on the determinants analyzed in the univariate analyses above, namely diversification and internationalization. However, this model also lacks explanatory power and statistical significance. Nevertheless, it yields a significant negative impact of cross-continental transactions on abnormal performance of -28%. Since this effect is a combined effect of the cross-border and cross-continental variables, the true negative impact accumulates to approximately -6%, which, in turn, is consistent with the results presented in Table 4.6 above. Staying within the own product sector (intra-industry) yields an additional positive effect of +7%. Overall, it can be concluded that there is a significant and consistent influence of internationalization on abnormal post-merger performance of acquirers. In addition, the results also point to a likewise negative effect of diversification although it is not as significant. For both effects, Hypothesis 4.2 seems to hold. However, none of the other determinants including transaction size, experience, continents, and public status of the target turn out to be significant.

4.5

The Correlation of Accounting- and Event-Study Results

In order to assess the potential correlation between the results of the preceding performance analyses and long-term event-study results, corresponding 36-month BHARs are derived for the 93 transactions underlying the accounting analyses of this study. Table 4.8 reveals that the resulting returns remain fairly similar to the ones previously derived on the basis of 206 transactions in the industry (see Chapter 3.4): Both, equal- and value-weighted returns decrease over time and become statistically significant over a 36-month horizon. Equally-weighted, the 93 acquirers in this sample underperform the share returns of their non-merging matches by a significant -14.24%, as opposed to the -16.68% observed for the full sample. However, although the overall pattern looks similar, it becomes apparent that the negative effect is less strong and less significant for the selected 93 transactions and especially regarding the value-weighted returns. Therefore, it is assumed that the se-

122

4 Study 2: Does Operating Performance Meet Market Expectations?

lected 93 transactions contain a number of positively outperforming transactions of significant size. Compared to the original sample, these transactions are over-represented and therefore decrease the observable negative effect. Consequently, it could be assumed that the observed accounting results in Section 4.4.1 are likewise influenced and lead to the non-significant results in EBIT- and EBITDA-returns.35 While this potential limitation is again considered in the overall discussion of the accounting results, it does not affect the relationship between the two approaches so that the derived BHARs are subsequently used as a starting point for the regression analysis. Table 4.8: BHARs to Acquirers Applied in the Accounting Study Total Sample BHARs (n=93) Equal-weighted Event-window

BHAR

t-value

12 months

0.84%

0.15

24 months

-5.46%

36 months

-14.24%

Value-weighted p-value

BHAR

t-value

p-value

0.44

3.72%

0.66

0.26

-0.65

0.26

1.91%

0.23

0.41

-1.32

0.09 *

-7.73%

-0.71

0.24

This table shows the average Buy-and-Hold Abnormal Returns to acquiring companies in mergers and acquisitions in the automotive supply industry. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table differentiates between a 12-, 24-, and 36-month holding period. All 93 acquisitions between 1980 and 2004 are included which are applied in the overall sample of this study. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test.

Table 4.9 provides an overview of the six resulting OLS regression models each regressing the 36-month BHARs on one of the six indicators of abnormal performance change as the independent variable. For three of the four EBIT-related performance indicators, a strong positive relationship between the two approaches is obtained. The EBIT/Sales indicator positively relates to the 36-month BHARs with a factor of 0.972 (at the 1% significance level). Therefore, Hypothesis 4.3 seems to hold for the EBIT35

Appendix 4 confirms that especially EBIT- and EBITDA-returns are positively influenced by larger transactions. The CF-returns, on the contrary, are negatively influenced under a value-weighted average. The subsequent discussion will distinguish between the three performance nominators.

4.5 The Correlation of Accounting- and Event-Study Results

123

related indicators as well as the EBITDA/Sales performance indicator. All three correlate significantly with the long-term event-study results. In general, these results are in line with the findings of Healy et al. (1992) who conclude that there is a positive relationship between event- and accounting-study results and that both approaches can be used as substitutes. However, as this preceding work focuses on the comparison with short-term event studies, Table 4.9 advances previous findings by extending the substitution characteristic to long-term event studies. Table 4.9: Regression of 36-month BHARs on Abnormal Performance Changes Regression Results Independent Variable CF/BVA

a

t-value

b

-15.56%

-1.41

-0.110

t-value -0.62

Statistical Quality F = 0.39 R2adj. = -0.01 DWS = 1.75

CF/Sales

-13.19%

-1.19

0.100

0.47

F = 0.22 R2adj. = -0.01 DWS = 1.72

EBIT/BVA

-12.20%

-1.15

0.415

2.24 **

F = 5.04 ** R2adj. = 0.04 DWS = 1.70

EBIT/Sales

-10.77%

-1.09

0.972

4.53 ***

F = 20.55 *** R2adj. = 0.18 DWS = 1.83

EBITDA/BVA

-14.05%

-1.28

0.033

0.14

F = 0.02 R2adj. = -0.01 DWS = 1.72

EBITDA/Sales

-11.33%

-1.06

0.775

2.32 **

F = 5.39 ** R2adj. = 0.05

DWS = 1.76 This table shows the results of the models regressing the 36-month long-term abnormal share price return (BHARs) of firm i on the abnormal performance growth of the same firm. The slope coefficient b captures the correlation between the two approaches. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test for individual coefficients and the F-Test for the overall model respectively.

124

4 Study 2: Does Operating Performance Meet Market Expectations?

The results derived for CF-returns do not correlate positively to the returns derived under the BHAR-approach. The long-term capital market performance seems to be unaffected by the significant underperformance in CF-returns as derived in Section 4.4.1. One potential reason for this pattern lies in the evaluation behavior of capital markets. If investors follow earnings measures more closely in order to make their investment decisions, a decrease in CF performance will not influence capital market performance as strongly as a decrease in EBIT-returns. A second potential explanation lies in the negative cash flow effects inherent in M&A activity. If mergers are considered cash- and resource- intense, CF-returns will decrease independently from a potential increase or decrease in profitability (=EBIT). If this was true, the effect on CF-returns should be consistently negative for all instances in other measures such as positive and negative short-term announcement returns as well as positive and negative long-term BHARs. In order to test for this hypothetical relationship, the following model as described by Equation (4.5) incorporates additional variables including the short-term announcement returns (CAARs) to acquirers and the multiplicative combination of the abnormal operating performance change with both short- and long-term capital market returns. In the following, these variables are described in detail.

(4.5) BHARi, 36months = a 0 + b ⋅ PI i + c ⋅ PI i * BHARi + d ⋅ CAARi ; −5to 5 + e ⋅ PI i * CAARi ; −5to 5

PIi – As in the previous regression, PI describes the abnormal performance change over three years following a transaction. The respective performance indicator applied as an independent variable varies from one model to the other. If the positive relationship determined above holds, this variable is expected to assume values of similar size and significance.

4.5 The Correlation of Accounting- and Event-Study Results

125

PIi * BHARi – This multiplicative combination of the performance indicator changes and long-term BHARs provides an indication of how well the observed accounting and capital market performance correspond. Overall, a negative value for c is expected. If both measures, BHARs and PI, are negative, a negative coefficient increases the negative effect on BHARs additionally punishing the overall negative development. If both measures are positive, the overall positive BHARs are weakened potentially signaling that capital markets had expected the development. If the two measures diverge in direction, though, the negative coefficient would actually decrease a negative BHAR and increase a positive BHAR representing either additional information not represented in operating results or a "surprise premium" for unexpected positive operational development. CAARi,-5to5 – The majority of preceding literature tests how well the long-term accounting performance is already represented in the short-term announcement returns to acquirers. To include the effect of short-term results in this study, short-term announcement returns are estimated using the event-study methodology in connection with the standard market-model as described, for example, by Brown and Warner (1985). Local indices are determined for each country represented in the transaction sample and serve as the return benchmark. The market models are estimated by using OLS regression over a 200-trading-day period starting at trading day t = -250 relative to the earliest announcement date of the M&A event.36 If capital markets correctly anticipate the long-term performance development, the variable will assume a significant positive value. If the argument of Fridolfsson and Stennek (2005) holds true, the relationship will be negative as an indication of the externalities inherent in the market. PIi * CAARi,-5to5 – In addition, the multiplicative variable combining the observed short-term announcement effect with the long-term profitability development is included into the model. In contrast to the PI*BHAR-variable, the multiplicative con36

Appendix 5 presents an overview of the derived short-term CAARs for the 93 acquirers in the sample. The derived values confirm the extraordinary return position of acquirers in the industry as previously determined by Mentz and Schiereck (2008).

126

4 Study 2: Does Operating Performance Meet Market Expectations?

nection with short-term returns is assumed to carry a positive premium for correct anticipation of long-term share price returns. If short-term CAARs and long-term performance changes move in the same direction, the anticipation premium increases positive BHARs and decreases negative BHARs. If the two diverge, the effect is the opposite. The expected coefficient is significantly positive. Table 4.10 presents the resulting six regression models each including all four variables described above. At first sight, it becomes apparent that only those models using returns on sales yield highly significant results whereas the ones focusing on returns on BVA remain insignificant. This observation again favors the previously-held assumption that sales represent the more reliable denominator in terms of measuring the impact of M&A on acquirer performance. Assets are subject to a variety of different accounting standards and provide an opportunity for designing balance sheet policy to meet the acquirers' needs. Hence, no statistically significant relationship can be determined on asset-returns while sales models are highly significant. While the relationship between long-term 36-month BHARs and accounting performance changes as presented by the coefficient b remains consistent with the results presented in Table 4.9, the additional coefficients determined carry a number of additional insights. Firstly, the negative variable derived as coefficient c might be an indication for a "surprise premium": if long-term BHARs and observable performance development in published accounting information diverge (and therefore have different signs), the resulting positive impact will increase positive BHARs and decrease negative BHARs. While positive operating returns improve a negative capital market performance, positive BHARs concurring with accounting underperformance are most likely the result of additional information inherent in the market. If the positive BHARs persist through the underperformance in accounting data, the positive capital market returns are amplified in magnitude. If, however, both measures correspond in direction, no surprise premium affects the long-term BHARs; both positive and negative BHARs experience an additional negative effect.

4.5 The Correlation of Accounting- and Event-Study Results

127

Table 4.10: Relationship between Event- and Accounting-Study Results Regression Results Performance

Constant

Indicator

a

CF/BVA.

-0.185

Stat. Quality CF/Sales

EBIT/BVA Stat. Quality EBIT/Sales Stat. Quality

-1.59

F = 1.06 -0.086

Stat. Quality

t-value

F=

-0.83

-1.11

F=

1.62

-0.032

-0.29

EBITDA/BVA -0.177

t-value

R2adj. =

0.00

DWS = 1.76

0.268

1.30

-1.254

0.457

2.17 **

R2adj. = 0.03 0.838

0.49

F=

0.37

R2adj. =

-0.03

EBITDA/Sales -0.045

-0.41

0.643

F=

1.90 *

5.62 *** R2adj. = 0.17

-1.55

-5.66 ***

CAAR

PI*CAAR

d

t-value

e

1.179

0.39

-0.199

t-value 0.92

0.139

0.11

4.975

1.75 *

1.043

0.78

-1.022 -0.57

-0.651

-0.49

2.773

1.421

1.03

-1.280 -0.66

0.225

0.16

3.037

DWS = 1.78 -0.107

-0.63

DWS = 1.72

3.60 *** -0.444

7.55 *** R2adj. = 0.22 0.137

Stat. Quality

-0.377

t-value

0.98

-1.51

Stat. Quality

PI*BHAR c

-0.006

8.41 *** R2adj. = 0.24

-0.129

F=

PI b

-2.80 ***

1.07

DWS = 1.74 0.060

0.21

DWS = -0.03 -1.180

-4.01 ***

0.91

DWS = 1.80

This table shows the results of the models regressing the 36-month long-term abnormal share price return (BHARs) of firm i on the abnormal performance growth of the same firm (PI), the product of the performance growth and the BHARs (PI*BHAR), the short-term announcement returns over 11 days surrounding the announcement date (CAARs), and the product of the later and the observed BHARs. The slope coefficients b, c, d, and e capture the correlations between the different approaches. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test for individual coefficients and the F-Test for the overall model.

Secondly, the impact of CF-returns on BHARs still remains insignificant at 0.268. However, the derived "surprise premium" also seems to hold for this indicator with a highly significant coefficient of -1.254 (1% significant). CF-returns appear to have at least an indirect connection to capital market performance. If the changes in CFs are moving in the opposite direction than the BHARs, capital markets react and award the surprise premium. Therefore, the assumption that this performance indicator yields negative returns independently from its share price returns can be rejected. Nevertheless, a direct link between the measures does not exist so that a substitutability of the long-term event- and accounting-study approaches, as predicted in Hypothesis 4.3, must be limited to EBIT-/EBITDA-measures.

128

4 Study 2: Does Operating Performance Meet Market Expectations?

Thirdly, the product of the two variables PI and CAARs positively influences long-term BHARs across all the three returns on sales and therefore indicates a similar, yet opposite premium-return as previously observed between BHARs and PI: If shortterm announcement returns correctly anticipate operational performance, capital markets appreciate this anticipation by granting an "anticipation premium". If the two measures diverge and acquirers do not develop as predicted by capital markets, BHARs decrease significantly expressing the disappointment of not realizing advanced synergy gains inherent in positive CAARs to acquirers. However, the statistical significance of the anticipation effect is lower than the significance of the surprise premium. The only indicator yielding a significant impact is CF/Sales, for which the coefficient amounts to a significant 4.975 (10% significant). The other coefficients stay insignificant. The short-term announcement returns to acquirers by themselves do not significantly influence long-term BHARs. However, long-term capital market returns continuously incorporate new information and upcoming changes in a company's field of activities. Therefore, it does not seem surprising that long-term BHARs and short-term announcement returns do not significantly correlate. In addition, the majority of previous research on the correlation of the two approaches tests whether long-term operating performance is correctly anticipated in short-term returns. To advance this discussion, Appendix 6 contains an additional regression of accounting performance changes on the short-term announcement returns to acquirers. In line with the argument of Healy et al. (1992), a highly significant positive coefficient for EBIT/Total Sales and EBITDA/BVA-returns can be obtained. Overall, it is concluded that the findings in this section correspond to previous findings indicating that accounting- and event-study returns can be treated as substitutes. However, this study advances previous literature in a number of ways: Firstly, the substitutability does not only apply to the comparison between short-term event-study and accounting-study returns, but also to long-term event-study and accounting-study returns. Secondly, the degree of substitutability depends on the accounting performance

4.6 Conclusion

129

indicator chosen. While capital markets closely follow EBIT- and EBITDA-returns, CFreturns do not correlate with long-term BHARs. In addition, 'sales' turns out to be the preferable denominator of post-merger performance due to the exposed position of assets to other effects. And thirdly, there is significant evidence for a "surprise premium" rewarding a divergence in BHARs and accounting-study returns as well as for an "anticipation premium" rewarding a correctly anticipated long-term performance by shortterm announcement returns.

4.6

Conclusion

The first objective of this study was to measure and analyze the long-term accounting performance of acquirers in the automotive supply industry. For this purpose, a sample of 93 horizontal takeover transactions involving automotive suppliers between 1986 and 2003 was identified and analyzed for their long-term peer-adjusted performance change. Overall, acquirers in this industry seem to consistently underperform their non-merging matching firms up to -12%, depending on the observed performance indicator. The abnormal change in operating cash flow over sales is a significant -10.52% lower than the change in the same indicator of the non-merging control-firms. As these results concur with a previously determined significant underperformance in long-term capital market returns, they are in line with capital market's continuous assessment of the acquirers' performance. However, as they stand in clear contrast to significant positive announcement returns observed in the industry, this average analysis supports the assumption that acquirers are unable to realize the outstanding synergies perceived by capital markets. Nevertheless, the significant underperformance in accounting returns is in line with previous research promoting a negative effect of M&A activity on firm profitability including Ravenscraft and Scherer (1989) and Dickerson et al. (1997). In addition, this study provides some indications that the underperformance in firm profitability is influenced by similar determinants as long-term

130

4 Study 2: Does Operating Performance Meet Market Expectations?

capital market performance. Due to larger resource requirements, diversifying and cross-border transactions negatively influence firm profitability. The second objective of this study was to determine a potential correlation between the observed profitability changes and event-study results. In line with previous findings of Healy et al. (1992), this study finds a strong positive correlation between short-term announcement returns and the observed accounting performance. While contradicting the results of the overall average perspective described above, the correlation results support the assumption that acquirers tend to fulfill capital market expectations as expressed in short-term announcement returns. While the overall average remains significantly negative, some acquirers seem to be able to realize a significant positive performance development in line with positive announcement returns previously granted by capital markets. In addition, this study's correlation analysis extends the substitutability argument to long-term event-study results. Significant evidence indicates that in the automotive supply industry, an analysis of firm profitability cannot only serve as a substitute for short-term announcement returns but also for long-term event studies on Buy-and-Hold Abnormal Returns. Furthermore, this study provides a number of additional insights into the correlating relation: Firstly, the substitutability of the two approaches depends on the chosen performance indicator. For EBIT- and EBITDA-returns, the relations hold true and both approaches can be used as substitutes. However, CF measures do not correspond to event-study results and must be individually analyzed in future studies. Secondly, 'sales' are proven to be the more accurate denominator of post-merger performance. Models focusing on returns on assets appear to be influenced by other effects than M&A transactions including balance sheet policy and choices in asset realization. And thirdly, significant evidence for the existence of a surprise and an anticipation premium exists. A divergence of long-term BHARs and firm profitability development is positively rewarded by long-term capital market returns; and a divergence between short-term

4.6 Conclusion

131

CAARs and firm profitability is followed by a consistent decrease in long-term capital market returns as an expression of investor disappointment with synergy realization. This study serves as a comprehensive overview of the long-term effect of mergers and acquisitions on firm profitability in the automotive supply industry. It contributes to the methodological discussion on the correlation between accounting and event studies with significant evidence for substitutability between the two approaches. However, the results presented are influenced by the sample selection requiring data availability across the time period under analysis. As pointed out before, these 93 transactions are affected by a few outperformers of significant transaction size. This could partly explain the less significant underperformance observed in EBIT-/ EBITDAreturns. In order to provide further evidence for a consistent performance effect and the exceptional role of cash positions such as operating cash flow, further research could challenge the observed findings against a case study from the industry and compare realized gains with intended transaction motives.

5

Study 3: How a Good Bidder Becomes a Good Target – The Case of Continental AG Acquiring Siemens VDO

5.1

Introduction

On July 25, 2007, after almost half a year of weighing strategic and financial alternatives, the technology conglomerate Siemens AG agreed to sell its automotive supply division ("Siemens VDO Automotive") to the German-based automotive supplier Continental AG. At a value of EUR 11.4 billion (USD 15.7 billion), this acquisition represented the largest in Continental's corporate history as well as in the automotive supply industry up to that date. The combined firm ranked among the five largest automotive suppliers in the world, within reach of the few industry leaders. From the perspective of its management board, engaging in this significant transaction carried an opportunity not only to realize synergy potentials, but, more importantly, to meet industry trends by expanding Continental's market position, by increasing its innovative abilities and by balancing its product portfolio across multiple product segments (Continental (2007a)). At the time of its announcement, the transaction appeared to be a reasonable strategy to prepare Continental for future market challenges. However, one year after the deal announcement, the risk of unsuccessfully integrating the target and, hence, finally overpaying for the acquisition remained noticeable for the investor: Over the twelve months following the acquisition, Continental's share price dropped from EUR 108.5 to EUR 73.0 per share. Although Continental claimed to have paid a fair price for VDO, it realized a loss in market value of approximately EUR 4 billion within this one-year time frame. Furthermore, Continental eventually became a target itself and was taken over by the German Schaeffler Group in August 2008. In the light of the originally positive public deal appraisal, this significant value loss in connection with the subsequent takeover raises the question whether Continental was in fact a 'bad bidder' predetermined to become a 'good target' (Mitchell and Lehn (1990)).

134

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

This question calls for a comprehensive assessment of the overall merger success and its underlying determinants. Whether the takeover of Siemens VDO can be regarded as a success also depends on the underlying industry trends. Unlike other producing industries, the automotive supply industry provides a particularly challenging market environment to supply companies. On the one hand, the pressure to produce better equipped and less expensive automobiles creates a growing trend towards specialization and internationalization among suppliers (Sadler (1999)). Following their original equipment manufacturers (OEMs), many suppliers relocate their production facilities abroad to meet 'localcontent'-requirements and circumvent customs (Abrenica (1998)). On the other hand, increasing prices for raw materials represent an additional burden for the profit situation of automotive suppliers. As a result, many suppliers suffer from significant profit reductions of up to 50% in less than 2 years (Fitzgerald (2002)). Under these conditions, investors seem to value M&A as a valuable response strategy: Acquirers are able to realize significant positive short-term returns as an expression of the global synergy and efficiency potential underlying the transactions (Mentz and Schiereck (2008)). However, the preceding chapters also show that suppliers are generally not able to sustain these announcement returns beyond a short-term perspective. In the long-run, they fail to realize synergies and experience significant value losses from a capital market perspective (Chapter 3) as well as from an operating performance perspective across various performance indicators (Chapter 4). Only transactions involving national targets, non-diversifying product segments, significant transaction size, and bidding experience of the acquirer partially mitigate the negative returns, yielding these characteristics as potential success factors. Since the Continental/Siemens VDO transaction combines all of these attributes, Continental appears to be well prepared for long-term success in the industry. Consequently, the VDO transaction is particularly relevant for further analysis in the form of a case study.

5.2 Literature Review and Industry Overview

135

The following case study explores the motivation and background behind Continental's takeover of Siemens VDO and advances previous literature by an additional perspective on which M&A strategies might lead to long-lasting positive returns in the automotive supply industry. By overcoming the methodological shortcomings of broader empirical studies, this case study serves as an additional opportunity to validate key success factors for future takeovers. Therefore, the objective of the following case study is twofold: firstly, the motivation behind the transaction is compared to the realized capital market returns and accounting performance as well as to Continental's relative position in the market. As a result, this study presents a comprehensive assessment of the overall transaction success. Secondly, general key success factors as developed by preceding literature are validated and complemented with additional insights from the case study at hand. The remainder of this study is organized as follows: Section 5.2 provides a brief overview of the relevant literature and outlines the predominant market conditions in the automotive supply industry at the time of the observed transaction. Section 5.3 introduces the two transaction partners, the transaction motivation, and a detailed description of the course of events around the acquisition. The following Section 5.4 contains the empirical assessment of the merger success including the short- and long-term capital market reactions to the deal announcement. In addition, Section 5.4 also provides an analysis of Continental's accounting performance over the 12 months following the deal and compares the results to the preceding capital market assessment. Section 5.5 discusses the findings and concludes.

5.2

Literature Review and Industry Overview

5.2.1 Related Literature The value creation potential of mergers and acquisitions has been extensively discussed in empirical finance literature. However, merely a few empirical event and

136

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

accounting studies focused on the specific characteristics of the automotive supply industry. Mentz and Schiereck (2008) were the first to conduct an event study on 201 M&A transactions in the automotive supply industry between 1981 and 2004. Over an 11-day event-window surrounding the announcement date, they determine a significant positive abnormal announcement return to acquiring companies of +1.6%. As the majority of broader event studies determines negative announcement returns to acquirers, this outstanding positive return characteristic identifies the industry as an outlier.37 The authors argue this finding to be the result of industry-specific synergy potentials perceived by capital markets; in the short-term, mergers and acquisitions are perceived as a feasible measure to realize synergy- and efficiency-gains. However, Chapter 3 shows that acquirers in the automotive supply industry are not able to sustain this outstanding return position beyond a short-term horizon. Over a 36-month period following the deal announcements, the long-term event study presented in Chapter 3 reports significantly negative BHARs of approximately -17% for a sample of 164 transactions. While these negative returns are generally in line with returns derived in broader long-term event studies, 38 Chapter 3 also determines a number of characteristics that significantly reduce or mitigate the derived underperformance: national transactions, large transaction volumes, focusing transactions, and acquisition experience of the acquirer. As the Continental/Siemens VDO transaction carries all of these characteristics, their applicability as general key success factors are validated in the course of this case study. In addition to the capital market perspective of event studies, the post-merger operating performance of acquirers in the automotive supply industry has also been analyzed in a broad accounting study: Chapter 4 finds a significant underperformance of

37

38

Loughran and Vijh (1997) and Bruner (2002) each provide a comprehensive overview of existing event studies across various industries. While the later concludes that abnormal returns to acquiring companies are essentially zero, Loughran and Vijh find overall negative or, at most, insignificantly positive acquirer returns. The majority of more recent long-term event studies point to negative abnormal returns to acquirers (see, for example, Black et al. (2001), Sinha (2004), Gregory and Matatko (2004)).

5.2 Literature Review and Industry Overview

137

acquiring automotive suppliers in 93 transactions when compared to their non-merging peers. Comparing this underperformance to the long-term capital market returns, the derived analyses also indicate a significant correlation between the two measures for success pointing to the conclusion that event- and accounting-study returns can be used as substitutes in the automotive supply industry. The underperformance in accounting data is likewise reduced for national and non-diversifying transactions. While event and accounting studies are generally accepted as well-proven empirical approaches to measure value creation through M&A transactions, both approaches carry methodological shortcomings potentially limiting the interpretation of their results. Event studies are generally forward-looking and build on the efficient market hypothesis: it is assumed that share prices represent the present value of expected future cash flows to shareholders (Bruner (2002)). However, given the large amount of available information, some information may not be correctly incorporated into the current share price (Eberhart et al. (2004)). In addition, event-study results are sensitive to the event and estimation periods chosen; while abnormal returns are usually mitigated in larger event-windows, larger estimation periods are likely contaminated by other confounding events (Rhoades (1994)). Accounting studies, however, examine reported financial results and are therefore primarily backward-looking. While published data carries credibility to the reader, a general comparability of published data across different years and reporting standards does not always apply (Bruner (2002)). In the light of these methodological shortcomings, case studies have developed into a commonly-used methodological approach complementing event and accounting studies. They facilitate a more detailed analysis of outstanding transaction phenomena that usually exceed the scope of large-sample event or accounting studies (Kaplan et al. (1997)). This case study attempts to find new insights from the case of the Continental transaction as well as to validate key success factor previously developed in largesample event and accounting studies.

138

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

While the value creation of M&A in the automotive supply industry has been studied in large-sample studies, a case study involving an automotive supplier does not yet exist. However, there are a number of substantial case studies offering insights and potential research foundations for the purpose of this study. For example, Bruner (1999) examines the 1993 merger attempt of AB Volvo with Renault in France. Shortly after the deal announcement, a sharp drop in Volvo's market value of 22 per cent forced the management board to withdraw from its original plans. Apparently, investors doubted the supposed merger synergies and feared the planned transfer of business control to France. In 1991, AT&T acquired NCR Corporation and thereby decreased its shareholder wealth by more than USD 4 billion. In this case, Lys and Vincent (1995) offer a mismatch in management objectives and shareholder wealth, management hubris, and a psychological phenomenon to move against better knowledge as explanations for the underperformance. Calomiris and Karceski (1998) analyze nine bank transactions in the 1990s and find that some transactions are driven by pure efficiency gains while others are motivated by revenue synergies from cross-selling and market re-positioning. While analyzing the banks' capital market performance, the authors find two cases in which the transactions cause negative share price developments; these banks were previously traded at a premium while being rumored to become takeover targets themselves. As Continental AG also became a takeover target in 2008, its performance has to be evaluated in the light of a potentially existing target premium and anticipation effects. The structure of this study generally follows the framework provided by Calomiris and Karceski (1998): Based on a detailed description of the transaction partners and their industry environment, the motivation and the applied M&A-strategy are derived. Afterwards, the performance of the acquirer is evaluated based on capital market and accounting data as opposed to a peer benchmark. Other substantial case studies exploring the value creation through M&A follow a similar approach and include, for example, the takeover of Cameron Iron Works by Cooper and of Florida Tile by Pen-

5.2 Literature Review and Industry Overview

139

mark (Kaplan et al. (1997)), the acquisition of Federated by Campeau (Kaplan (1989)), and the takeover of Conoco by Dupont (Ruback (1982)).39 5.2.2 Overview of the Automotive Supply Industry By 2007, consolidation efforts among automobile producers were well advanced and created global revenue opportunities as well as an omnipresent global competition. Most obviously, this advanced consolidation becomes apparent in the total number of existing OEMs: Mentz and Schiereck (2008) are able to identify 23 OEMs that are still active or were active between the years 1981 and 2004. By contrast, the number of automotive suppliers is estimated significantly higher. In 2000, the CEO of Dana Corporation estimated the global number of tier-1 suppliers at approximately 9,000 (Morcott (2000)).40 Consequently, the level of consolidation is far lower for automotive suppliers than for their customers. Nonetheless, consolidation efforts have significantly increased among suppliers over the last two decades, especially in Europe, Northern America and Japan/South Korea. It is estimated, for example, that the number of direct suppliers in Europe dropped from 10,000 in the early 1970s to 3,000 in 1995 and to about 500 in the year 2000 (Sadler (1999)). Figure 5.1 emphasizes this consolidation activity and shows how a strong merger wave affects the automotive supply industry during the 1990s. Between 1991 and 1999, significant M&A transactions (with a transaction value of more than USD 50 million) steadily increased both in total number as well as in inflation-adjusted transaction volume. After the year 2000, the extent to which these transactions influence the industry appears to weaken in direct comparison to the 1990s. However, a closer look at the average transaction values reveals that they peaked three times over the last 30 39 40

For an overview of the derived conclusions, see Bruner (2002). The concept of distinguishing between different supplier tiers relates back to the pyramid-shaped supply chains developed mainly in the Japanese automotive industry. First-tier suppliers deliver subassembled units (e.g. complete seats or transmissions) to OEMs. Lower-tier suppliers deliver components for the sub-assembled units to first-tier suppliers. Consequently, manufacturers are able to reduce their direct relations to a few first-tier suppliers while the suppliers coordinate themselves down the supply chain (Von Corswant et al. (2003), Fujimoto (2001)).

140

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

years: once during an early consolidation wave in the 1980s, once during the merger wave of the 1990s and once just recently with the USD 15.7 billion transaction of Continental AG. Consequently, it is likewise reasonable to assume that, instead of just one merger wave in the 1990s, consolidation in the automotive supply industry follows a continuing activity pattern with the 2007 Continental acquisition of VDO potentially representing the starting point of the next significant consolidation wave. Figure 5.1: Transaction Volume in the Automotive Supply Industry 25

35 30

20

15

20 15

10

Count

USD billion

25

10 5

Transaction Value (inflation adj.)

2007

2005

2003

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

0

1981

5 0

Number of transactions

Source: Thomson Datastream, own calculations

A number of key industry trends and conditions promote this on-going consolidation among suppliers. The remainder of this section outlines the most relevant trends grouped by their relative source of pressure and aims to create an understanding for the industry setting at the time of the VDO takeover. The first group of trends relates to the relationship between the automotive suppliers and their customers, the automotive OEMs. With the globalization of the automobile industry, the OEMs are increasingly interested in sourcing supplies from the same supplier on a world-wide basis (Sadler (1999)). In order to meet their customer's demands regarding just-in-time delivery as

5.2 Literature Review and Industry Overview

141

well as local regulative requirements including customs and in-country quotas, the pressure on automotive suppliers to create a costly global presence is growing. Wherever financially affordable, automotive suppliers are therefore following the internationalization of their customers either by geographical expansion or cross-border acquisitions (Abrenica (1998), Sadler (1999)). In addition, automotive producers increasingly seek to outsource parts of their production facilities and to purchase full systems of components from their suppliers rather than individual parts (Sadler (1999)). While this development enables OEMs to reduce their coordination efforts to a few relationships with first-tier suppliers, it likewise increases the coordination efforts imposed on the supplier itself. As first-tier suppliers also outsource a growing part of their activities, first-tier suppliers are increasingly required not only to manage their customer demands up-stream, but also a growing number of lower-tier suppliers down-stream. At the same time, automotive producers are increasingly transferring product development tasks. While the producers' share of total product development resources averaged at around 70% in 1988, it dropped to approximately 60% ten years later (Von Corswant and Fredriksson (2002)). As a result, suppliers acquire valuable production expertise and product development capabilities in order to be capable of delivering innovative products at high frequency. The closer a supplier is situated to the automobile producer in the supply chain, the higher its actual product development activity becomes (Von Corswant et al. (2003)). The second group of main industry trends relates to the increasing competition among automotive suppliers: As producers try to source complete systems from a limited number of first-tier suppliers, supply companies develop an increasing tendency towards specialization on particular products or segments. As of today, some leading firms have even become inseparably connected with particular systems or technologies (Sadler (1999)). One relevant example for this instance comprises the speedometers of Siemens VDO, whose success story started with their prominent placement into the Volkswagen Beetle in 1939. Where this specialization strategy is successful, the level of

142

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

competition is fairly low, due to a limited number of rivaling companies (for example, only a few players provide very complex products such as Xenon headlights). Where specialization is not able to create a competitive edge, though, the particularly high number of suppliers creates a very strong competition and rivalry. This is the case, for example, in the leather and tire industries (Aktas et al. (2004)). A last set of pressuring industry trends stems from the increasing prices for raw materials. Although the slowing down US economy and the weak US Dollar offset some of the price effects on suppliers, especially for European and Asian companies, the majority of raw materials reached record prices in 2007. Over the year 2007, for example, the average price of crude oil rose by 11%, the price of processed metals increased by 9% and of natural rubber by 10%. The majority of other raw materials including copper, steel, and nickel experienced similar price increases with different volatilities (Continental (2007b)). In connection with the increasing pressure from customers and competitors, these costs for raw materials put a strong strain on the profit situation of automotive suppliers. As a result, many suppliers realized losses or significant profit reductions over the first years of this century. Between 2000 and 2002, for example, many suppliers suffered from significant profit reductions of up to 50% (Fitzgerald (2002)). Given these challenging industry conditions, mergers and acquisition appear to be a valuable answer for automotive suppliers to offset some of the problems described.

5.3

Case Study Background

5.3.1 The Transaction Partners 5.3.1.1 Continental AG With an original focus on soft rubber products, rubberized fabrics, and solid tires for carriages and bicycles, Continental-Caoutchouc- and Gutta-Percha Compagnie

5.3 Case Study Background

143

was founded in Hanover, Germany in 1871.41 While being closely related to the vehicle industries, Continental's initial growth was largely driven by innovations in the field of rubber tires. In 1892, for example, Continental was the first German producer of pneumatic tires for bicycles. Nine years later, the first Daimler-produced car won the NiceSalon-Nice car race on Continental's automobile pneumatic tires without tread pattern. The list of significant tire innovations continued through the first half of the 19th century and comprised the world's first automobile tire with patterned tread (1904), a detachable rim for sedans (1908), the cord tire (1931), the tubeless tire for commercial vehicles (1943), and M+S tires for winter driving (1952). Nevertheless, innovations in other product segments also contributed to Continental's early success: The first airplane to cross the English Channel in 1909 was covered with Continental's Aeroplan material. In the 1950s, the company invented air springs for trucks and busses. In the 1970s, Continental started to combine its innovation-driven organic growth with an exploration of the increasing global market for mergers and acquisitions. At first, Continental focused on internationalizing its purely German tire business. By acquiring the European tire activities of American Uniroyal Inc. in 1979 and the Austrian-based Semperit in 1985, Continental made its first move into the European tire market. Two years later, the company entered the North American market by acquiring US-based General Tire. The following decade, Continental shifted its focus from internationalizing its tire business to product diversification. In 1998, Continental managed to become a leading brake and chassis specialist by acquiring the Automotive Brake & Chassis division of an American competitor. By acquiring Phoenix AG in 2004, Continental was able to complete the reorganization of its industrial product lines under the brand 'ContiTech' and thereby create an internationally competitive position in the hoses and conveyor belts market. With the acquisitions of Temic Microelectronic GmbH in 2001 and Motorola's automotive electronics business in July 2006, Continental did not

41

The following description of Continental's corporate history is based on Continental (2007c).

144

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

only extend once again its product portfolio into vehicle electronics, but also underlined its frequently-proven acquisition capabilities. At the time of the Siemens VDO takeover, Continental was considered Europe's second largest automotive supplier. With revenues of EUR 14.9 billion, it was able to generate an EBIT-margin of 10.8% in 2006, namely EUR 1.6 billion in earnings. Being active in 37 countries, the company operated more than 100 production and research facilities. A total of 85,224 employees worked in one of the four business divisions Automotive Systems, Passenger and Light Truck Tires, Commercial Vehicle Tires and ContiTech. In terms of revenue contribution, the Automotive Systems and Passenger Tire divisions represented the strongest divisions with 40% and 32% contribution respectively. Commercial Tires was the smallest unit with 10% revenue contribution; the remaining 18% originated from ContiTech products. While the earnings contribution generally followed the same pattern, the Commercial Vehicle division represented a positive outlier generating the largest profit with EUR 651 million, or 41% of Continental's total EBIT (Continental (2006)). By the year 2007, each of the four business divisions held at least one significant market leadership position within its product range: The Automotive Systems division was considered the number one worldwide provider for foundation brakes and number two for electronic brakes. With the acquisition of the electronics business of Motorola, the division recently gained operative access to the interior electronics and telematics product segment, in which it believed to find additional growth opportunities. In terms of sales, the Passenger Tires Division was considered the leading European supplier for original equipment tires while the Commercial Vehicle Tire division ranked at number three among European truck tire suppliers. ContiTech, the world's largest manufacturer of rubber and plastics technologies outside the tire industry, held the world leadership position in the production of conveyor belts, hoses and air springs for rail vehicles (Continental (2006)).

5.3 Case Study Background

145

Supported by the records of its divisions, Continental realized an outstanding financial performance in its fiscal year 2006. For the fifth time in a row, it was able to grow sales and earnings on a year to year basis: compared to 2005, sales increased by 5.8%, earnings by 6.3%. The resulting cash flows enabled Continental to propose a 100% increase in dividends per share in the following shareholders' meeting. Continental's balance sheet benefited from its performance as the company was able to limit leverage to a gearing ratio of 25% (Continental (2006)).42 However, a number of challenges confronting the company at the time also influenced the outlook on Continental's future performance. First of all, a deleveraged balance sheet as exhibited by Continental in 2006 also carried the disadvantage of increasing the company's attractiveness for financial investors. In summer 2006, a first private equity firm assessed a potential public takeover bid but withdrew at an early stage (Continental (2006)). Secondly, Continental largely depends on raw materials such as natural rubber and oil. With raw material prices rising, Continental suffered a strong negative earnings impact of EUR -316 million between 2005 and 2006. At the same time, automotive manufacturers expected their suppliers to decrease prices by three to five per cent annually (Continental (2006)). With 61% of its sales originating from OEMs, Continental was therefore facing significant reductions of its profit margin over the coming years while the pressure from both sides of the supply chain continued. In addition, the 2006 business model of Continental was still strongly focused on European markets. All business units realized more than 50% of their sales within Europe; the Passenger Tire division even sold 72% of their products within its home continent. As the whole industry becomes increasingly global, decreasing the regional focus could determine future success as a whole. And lastly, as Continental has acquired a number of significant targets over the last years, it is still heavily engaged in integration efforts of different acquired targets, especially its 2006 acquisition of the Motorola electronics division.

42

The gearing ratio is defined as net indebtness divided by total equity.

146

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

In response to these challenges, the management board focused its 2007 strategy on a number of key objectives. Besides on-going research activities and fostering innovations, Continental also actively sought acquisition targets to continue growing at a constant rate. As a side effect, a potential acquisition would also enable the company to rebalance its financing structure and to protectively increase debt levels. To reduce dependencies on single markets and customers, Continental tried to expand its product range by finding complementary businesses around their present segments. In addition, the company continuously increased its production at low-cost sites and explored alternative raw material sources in order to ease the pressure on its profit margin (Continental (2006)). 5.3.1.2 Siemens VDO Automotive Siemens VDO Automotive originally emerged from a merger between Mannesmann VDO and Siemens AG's automotive branch ('Siemens Automotive') in 2001. Although both predecessor companies had developed into comparably diversified producers of automotive electronics by 2001, their roots and development paths over the preceding century were fairly distinct.43 The foundation of VDO's initial success lies in the invention of a speed-measuring device for automobiles, the tachometer. Based on an original patent filed in 1902, mass production of these speed measuring devices flourished in the 1920s when the OSA Apparate GmbH in Frankfurt, Germany took over their production and distribution. After merging with the tachometer operations of Deutsche Tachometerwerke GmbH in 1928, the company founded VDO Tachometer AG and began to spread the brand "VDO" across Germany and the world. The famous Volkswagen Beetle, for example, included VDO's tachometers as of 1939. Supported partly by the large sales volumes of this particular car, annual production of VDO tachometers reached 4 million by the 1960s. Around the same time, VDO commenced tachometer production in Australia. 43

The following description of the corporate history of Siemens VDO Automotive is based on Continental (2007d).

5.3 Case Study Background

147

Over the following decades, VDO diversified its tachometer business into the automotive electronics sector. With a number of significant innovations including the electronic cruise control, the first quartz clock within a cockpit, and the central information system, VDO did not only manage to continue growing but also to establish itself as an industry standard in a number of product segments. In 1991, Mannesmann AG took over the majority shareholding of the so far independent VDO Adolf Schindling AG and acquired complete ownership three years later. The automotive business of Siemens AG, however, gained its first experiences in the field as an automotive producer rather than a supplier. In 1905, the SiemensSchuckert-Werke manufactured one of the world's first electric cars called 'Victoria.' Three years later, production of an award-wining gasoline-driven automobile called 'Protos' started. A total of 25,000 automobiles of this type were produced until 1927. Afterwards, Siemens realized that success in the automotive business at the time was driven by mechanical systems rather than electronic advancements. Consequently, it temporarily withdrew from the industry in pursuit of its core competence: electrical engineering. With the merger of Siemens-Schuckert-Werke, Siemens & Halske and SiemensReiniger-Werke into Siemens AG in 1966, the company resumed its automotive business and started supplying electrical harnesses to automotive producers in the 1960s. Starting from that point, Siemens followed a similar path as VDO and grew through innovations and diversification. Main advancements in the 1960s and 1970s included a central locking system, a childproof safety lock, and automotive semi-conductors. Later on, onboard computers, ignition systems and alarm systems further increased the range of supplied products. Given its growing importance and revenue contribution, Siemens founded an independent business unit for their automotive businesses in 1989 ('Siemens Automotive'). As both companies, Mannesmann VDO and Siemens Automotive, had independently developed into diversified producers of automotive electronics, the

148

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

merger of Mannesmann VDO and Siemens Automotive in 2001 created a global market leader in the field of automotive electronics and mechatronics. In 2006, Siemens VDO Automotive represented one of eleven operative business units held by Siemens AG.44 With EUR 10.0 billion in revenues, the division accounted for approximately 11.5% of Siemens' total revenues and realized a profit margin of 6.7%, namely EUR 669 million (Siemens (2006)). A total of 53,000 employees worked in 14 different product divisions ranging from gasoline and diesel systems (powertrain), over infotainment and radio navigation (interior electronics) to safety electronics and electric motor drives (safety and chassis). Especially in the field of navigation, radio, and surround sound, VDO created a significant development edge in future trend technologies. Examples of first successes in emerging technologies include the HD Radio, a satellite radio, and rear-sear entertainment systems (Continental (2007e)). The high degree of innovativeness in connection with its broad product range enabled Siemens VDO to grow revenues and earnings in the past. Compared to 2005, the company increased revenues by 4.2% in 2006; earnings grew by 6.2% over the same period. While this development was generally in line with the development of the industry and its competitors (see Continental above), it still decreased in comparison with 2005 growth numbers and fell behind the overall performance of Siemens AG: Overall, Siemens was able to increase revenues by 15.7% and earnings by 10.4% in 2006. This relative underperformance could potentially be one reason why Siemens decided to pursue a new strategy "Fit4More" in 2005: following this program, opportunities for taking the automotive division public were assessed in order to re-focus on its core competencies (Siemens (2007)). At the same time, while being active in the same industry as Continental, Siemens VDO was also exposed to similar threats. Consequently, the management strategically focused its resources on an increase in R&D expenditures and introduced a cost reduction program. It divested a number of joint ventures in the US

44

Other business units of Siemens AG at the time included, for example, Power Transmission (PT), Power Generation (PG), Medical Solutions (Med), and Industrial Solutions and Services (I&S).

5.3 Case Study Background

149

and Europe in order to focus resources. Going forward, staying innovative and continuously producing new products was to remain its main strategic focus (Siemens (2006)). 5.3.2 Transaction Motives In order to determine the overall success of the Continental/Siemens VDO takeover, one main objective of this study is to compare the original motivation behind the analyzed transaction with the realized capital market returns and the operating performance gains or losses. In general, value-creating motives are associated with either synergy or efficiency gains (Piloff and Santomero (1998)). While efficiency gains can be realized by a stand-alone entity without engaging in M&A activity, synergy gains can exclusively be realized through a combining transaction. Therefore, the following analysis of the transaction motives behind the VDO takeover focuses on the synergy motives, both on the revenue as well as on the cost side. In addition, since the takeover at hand was initiated and mainly driven by the acquirer Continental and since the amount of publicly available background information on Siemens VDO as an integrated division of Siemens AG is limited, the following section focuses on the buy-side motivation for buying Siemens' automotive business.45 Immediately after the deal announcement on July 25th, Continental launched an integration program entitled "winning the future – together." In the eyes of the acquirer, this slogan described the intended flexibility, creativity, and performance orientation for successfully integrating VDO into the Continental organization (Continental (2007b)). However, the slogan also appears to point out a strong interest in fostering former VDO strengths. For the acquirer, the target did not only yield the potential for cost synergies but, likewise, a number of strong revenue synergies. Consequently, Continental communicated its intention to "add" these strengths to its existing organization as a partner 45

In addition, Kaplan et al. (1997) discuss a potential sampling bias in case studies and find that mainly acquirers in successful transactions are willing to generously provide information on case studies in the form of interviews. In order to prevent this bias, this study refrains from conducting interviews but focuses on publicly available sources of information. It is assumed that especially testified regulatory publications carry the required degree of credibility.

150

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

rather than simply integrating them. With regard to the underlying revenue synergies, four main value drivers can be distinguished. Complementary Product Portfolios – In 2007, Continental identified three main technology trends which were to determine its future product sector priorities: an increasing strictness in emission regulation (mainly CO2-related), an accelerated flow of information within and between automobiles, and the convergence of active and passive safety systems towards integrated safety concepts (Continental (2007b)). Consequently, the company derived the three sectors Powertrain, Infotainment & Telematics, as well as Brake, Chassis & Safety Systems as its future business priorities. Before the VDO takeover, however, Continental carried a significant product range in only one of these three sectors, namely the brakes and safety product sector; in the other two, it obtained a mere presence while offering a few isolated products. The takeover of Siemens VDO enabled Continental to reposition itself as a market leader in all three of its future key segments as the product portfolios of both companies, with just a few exceptions, can be regarded as highly complementary. In the Powertrain segment, for example, the deal complemented Continental's engine management and transmission control systems with a full array of VDO pumps, sensors and injection systems. With the addition, Continental was able to comprehensively address the problem of reducing emissions towards its customers as it afterwards supplied the majority of emission-related components. In Infotainment, the transaction carried a similar effect. VDO contributed the complete range of radio, navigation, and sound systems which Continental needed to reach a higher connectivity with the control electronics it supplied within the car. Even in Continental's well-occupied safety arena, the takeover of VDO enabled the acquirer to gain access to further technologies such as 24 GHz radar sensors (Continental (2007e)). Altogether, the combined product portfolio after the takeover allowed Continental to further exploit the growing demand of its existing customers for pre-assembled systems while adhering to their efforts of reducing the number of supplier relations. At the same time, Continental could use its newly cre-

5.3 Case Study Background

151

ated market position to generate innovations and, thereby, to create industry standards which potentially provide access to further business with OEMs (Continental (2007b)). Innovation Capabilities – Although being closely related to the synergies from a complementary product portfolio, synergies from innovation capabilities positively affected Continental in two ways. On the one hand, combining R&D facilities increased the level of innovativeness. As all engineers after the deal had access to a larger variety of different technologies in the three targeted product sectors, Continental aimed to increase the probability and frequency of developing more complex, standard-setting products while at the same time fostering economies of scope. On the other hand, the combined R&D department also decreased the time to market of its innovations through increased development capacities. As automotive producers were increasingly outsourcing development activities to the suppliers, the ability to create standard technology platforms and offer development capacities to its customers became a key success factor for Continental. "Acquiring" a large number of automotive engineers gave Continental the engineering capacity to approach additional customers (Continental (2007b)). Global Presence – By globally combining the research and production facilities of both companies, Continental also aimed at creating a stronger and more significant global presence towards its customers. This becomes apparent in the fact that postmerger the combined firm ranked among the largest five suppliers worldwide whereas the individual companies pre-merger ranked just in or below the global top ten in terms of sales to OEMs. This newly acquired size enabled Continental to counterbalance the strong US-American and Japanese competitors (Continental (2007e)). In addition, Continental also aimed to reduce the dependencies of its revenues on the European OEM markets. Through acquiring VDO, it attempted to emphasize its regional footprint in North America and Asia, potentially reaching additional local customers and creating new revenue potential (Continental (2007b)). Customer Access – Besides supplying a larger product variety to its existing customers, the takeover also provided Continental with an opportunity to gain access to

152

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

new customers previously served by Siemens VDO. Given the limited number of existing automotive producers, gaining access to additional customers can significantly impact revenues. One main customer of VDO's electronic components was Hyundai Motor Corporation to which Continental could widen its braking systems and tire businesses (DowJonesNewswires (2007a)). In addition to the described revenue synergies, the takeover of Siemens VDO also conveyed significant cost synergies for the acquiring Continental. On the day of the deal announcement (July 25), Continental already presented synergy estimates amounting to a minimum of EUR 170 million per year as of 2010 (Continental (2007e)). A month later, Continental's management announced that these expected annual synergies carried upside potential (DowJonesNewswires (2007b)); and by February 2008, estimates of the yearly net synergy potential reached a sum between EUR 300 and 350 million (DowJonesNewswires (2008)). Although significant parts of the product portfolios were complementary, some overlap between the two portfolios did exist and was estimated between 20 and 25 per cent of total revenues (Handelsblatt (2007)). Continental intended to realize these synergies from a number of different sources. Firstly, by eliminating its redundant production capacities, the acquirer aimed to contribute significantly to the cost synergies without reducing revenues with its existing and newly acquired customers. Additional plant closures and staff reductions would further increase savings in personnel costs. Secondly, Continental also intended to consolidate its purchasing, administrative and research departments. After founding three new automotive divisions in its focus areas powertrain, infotainment, and safety, Continental had the opportunity to establish cross-divisional, centralized administrative functions representing additional contributions toward the cost synergy estimates above. While all revenue and cost synergies presented above can be regarded as generally value-adding, the takeover of VDO also followed a number of additional strategic considerations with potentially ambiguous value effects. Among these considerations was the application of the VDO transaction as a general takeover defense. At the time of

5.3 Case Study Background

153

the takeover, a potential investor interested in taking advantage of Continental's deleveraged balance sheet structure had already approached Continental. With a significant takeover such as VDO, Continental was able to leverage its capital structure and decrease the risk of becoming a takeover target itself. Another potential transaction motive was the acquisition of engineering talent. Before the deal, Continental had been short of engineers, merely employing approximately 7,000 employees with engineering degrees. By acquiring VDO, Continental was able to triple its engineering staff by adding an additional 12,000 automotive engineers from VDO (Wortham (2007)). And lastly, Continental's takeover motive could also stem from the on-going competition with its largest national and global rival Robert Bosch. After the takeover of VDO, Continental not only converged towards Bosch in terms of total size, but is now also facing off its main rival in Bosch's core business of advanced fuel-injections systems (Reuters (2007a)). 5.3.3

The Acquisition Event When Continental first announced its interest in Siemens VDO in January 2007,

the news took Siemens' management board by surprise and provoked not more than a conservative acknowledgement among Siemens executives (Reuters (2007b)). One reason for this reaction potentially lies in a misalignment of Continental's offer with the original plans Siemens management had for its automotive business: As early as 2005, Siemens had already decided to refocus its business portfolio on its core competencies and, therefore, to take its automotive business 'Siemens VDO' public (Siemens (2007)). In early 2007, these floatation plans were restated more precisely in a way that Siemens wanted to float 25% of its VDO business unit but keep a majority stake in the long-term (Reuters (2007b)). However, when Continental entered the market for Siemens' automotive business in the end of January, its offer started a six-month public bidding war that attracted a growing number of interested bidders and eventually led Siemens to withdraw from its original floatation plans. Table 5.1 shows how the resulting takeover occurred in two major phases: the first one comprising a six-month bidding war until the

154

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

final management decision on July 25th, and the second one focusing on preliminary integration measures preceding the final regulatory approval on December 5th. Table 5.1: Milestones in the Continental/Siemens VDO Transaction Date

Milestone

January 25

Continental publicly announces interest in acquiring Siemens VDO

February 22

Continental underlines its primary interest in industrial, not financial leadership

March 6

TRW Automotive joins bidding war for VDO

April 12

Siemens postpones the scheduled spin-off from May 1 to June 1

May 16

Siemens calls first Continental offer no competitive alternative to scheduled IPO Siemens mandates three investment banks to ready floatation Valeo (France) expresses interest in acquiring VDO

May 23

Continental submits first indicative bid for Siemens VDO (EUR 10 bn, estimated) TRW Automotive, KKR and Permira submit competitive bids

May 25

Siemens rejects indicative offer and continues IPO preparations

June 1

Siemens confirms IPO for September 30

June 13

Continental continues talks with Siemens

June 22

Continental explores alternative acquisitions

July 4

Due Diligence starts; further proceedings to be decided on July 25

July 25

Siemens Supervisory board agrees to sell VDO to Continental for EUR 11.4 bn

August 5

Continental expects annual growth between 6 and 7% after the VDO takeover

August 24

Continental expects annual synergies from the deal to exceed EUR 170 million

October 30

Continental issues new shares worth EUR 1.48 billion

November 29

EU Commission approves the acquisition

December 5

Continental concludes the takeover transaction of Siemens VDO

Source: Factiva Press Database, own illustration

Initiated by the first public announcement of Continental' interest in VDO on January 25th, the bidding war of phase 1 started uneventful: Given the large overlap in product segments between the two companies, Continental identified its business opportunity and announced interest in acquiring a stake of the soon-to-be spun-off Siemens VDO. On February 22nd, it underlined its potentially legitimate objectives by offering to

5.3 Case Study Background

155

accept industrial leadership over a majority stake in the equity (DowJonesNewswires (2007c)). However, the obvious synergy potentials between the two companies in connection with the positive track record of Siemens VDO soon attracted industry competitors and private equity firms into the competition. On March 6th, it became public that TRW Automotive, a US-based competitor controlled by its majority-shareholder Blackstone, entered the competition setting the stage for an increasingly more eventful bidding war. Over the following three months, a number of other interested players were mentioned in the press including competitors like French Valeo (May 16th) and private equity firms KKR and Permira (May 23rd). At the same time, Siemens gave these bidders the run-around. Although it publicly delayed the planned IPO repeatedly, it also publicly refused an initial offer and an indicative bid of Continental as being noncompetitive to the IPO strategy. By June 2007, after Siemens had repeatedly confirmed to adhere to its IPO plans, Continental indeed left communication channels with Siemens open but also assessed five other potential acquisition targets; public indications for a withdrawal from the deal arose. On July 4th, however, Siemens gave in and opened books for all remaining bidders which were still interested in buying its division. The only serious bidders remaining at that time were Continental and TRW Automotive, which both submitted competitive offers of more than EUR 10 billion by the official deadline three weeks later. On July 25th, with strong public support of German politicians and economists, the supervisory board of Siemens agreed to sell Siemens VDO to Continental for EUR 11.4 billion. As part of the deal, Continental received a tax credit of approximately EUR 1 billion. To finance the remaining amount, Continental planned on offering additional shares worth EUR 1.5 billion and taking on the larger part as debt. After both sides had approved the deal on July 25th, Continental did not lose time waiting for the official regulatory approval but quickly focused on integration issues and operative challenges. In phase 2 of the takeover, the acquirer used the time to quantify synergy and efficiency potentials. On July 29th, Continental confirmed that it

156

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

did not accept binding job guarantees as part of the takeover deal. In August, it announced that the expected annual synergies of EUR 170 million were conservative with further upside potential. At the same time, it set its growth aspirations between 6 and 7% in the mid-term indicating that further takeovers are possible. In November, target profit margins of 10 to 12% followed and set the future development path of VDO. While increasing growth through synergies on the revenue side, Continental consequently aimed for realizing efficiency gains through plant closure and staff reductions. On December 5th, after the European Antitrust Commission had approved the deal, Continental concluded the so far largest takeover in the automotive supply industry.

5.4

Acquisition Performance

5.4.1 The Capital Market Perspective In order to evaluate the success of this significantly sized transaction, this study follows the example of preceding case study research and analyzes both share price information and available published accounting data (Calomiris and Karceski (1998)). The following section focuses on the reaction of capital markets to the VDO transaction by analyzing the short- and long-term impact on Continental's share prices and corresponding returns. Therefore, the share price developments of Continental as the acquirer and Siemens as the seller are validated against their abnormal returns in various time periods. In order to determine abnormal returns to shareholders, a combination of different event-study methodologies is applied including the standard market model over short-term event-windows and the Buy-and-Hold Abnormal Return methodology using control-firms over a long-term horizon. As a result, this section provides a comprehensive overview of the absolute and relative value impact for the shareholders involved in the Continental/Siemens VDO transaction. As outlined in the previous section, the original takeover announcement took place on July 25th, 2007. Consequently, this date applies as the main reference point for

5.4 Acquisition Performance

157

the short- and long-term capital market analyses. However, Continental had already announced an interest in acquiring VDO about six months earlier, namely on January 25th. Therefore, it is assumed that the abnormal capital market reaction was partly included into the share price by July, in anticipation of the upcoming takeover event. To capture this anticipation and to provide a comprehensive overview of the short-term value creation, short-term abnormal returns are also derived for the dates of Continental's first notion of intent (January 25th) and the following regulatory approval of the deal (December 5th). For the long-term analyses, the focus lies on determining the continuous post-takeover return impact rather than short-term announcement reactions. Since potential anticipation effects become marginal with increasing time frames, the long-term perspective focuses on the announcement date in July as its reference point. Short-term abnormal returns are assessed using the event-study methodology in connection with the standard market model as described by Brown and Warner (1985). As the return of the market portfolio within the model generally refers to a market index associated with the given securities over time, the German DAX 30 index applies as the corresponding market portfolio (Coutts et al. (1994)). The market models are estimated by using OLS regression over a 200-trading-day period starting at trading day t = -250 relative to the first public mention of the transaction, namely Continental's first announcement of its intention to acquire VDO on January 25th. On the basis of the estimated market model parameters, abnormal returns for both Continental AG as well as Siemens AG are derived for the three event dates described. The longest event-window is 41 days: t = [-20,+20] days, t = 0 being the respective event date with regard to the transaction.46 From the share price development depicted in Figure 5.2, it becomes apparent that especially Continental's initial announcement of its acquisition intent induces the capital market reaction pattern normally observed for takeover events: Upon Continen46

Since Siemens VDO is not publicly traded at the time of the takeover, an assessment of the value creation potential for the combined entity cannot be obtained. However, the short-term event study also determines the abnormal returns to Siemens AG as the seller involved in this transaction.

158

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

tal's press release on January 25th, its share price drops by -1.8% to EUR 91.77 while the share price of Siemens AG as the seller of Siemens VDO significantly increases by 5.8% to EUR 82.44. This capital market reaction may reflect the uncertainty about the outcome of the bid. A potential competitive bidding war is a positive perspective for the target shareholders but includes the risk of overpaying for Continental. Figure 5.2: Share Price Development of Continental AG and Siemens AG 120

Euro

110 100 100 90 90 80 80

Continental AG

Dec-07

Nov-07

Oct-07

Sep-07

Aug-07

Jul-07

Jun-07

May-07

Apr-07

Mar-07

Feb-07

Jan-07

70 70

Siemens AG

However, this effect turns into the opposite upon the announcement of the deal six months later. On July 25th, Continental gains an additional 1.6% in share price while Siemens AG loses 6.2%. For Continental, this additional positive impact could be the expression of positively perceived synergies outweighing the now known takeover premium; the share price of Siemens decreases to a more moderate (but still increased) level after speculations about deal details continuously increased its share price over the preceding 6 months. As a result, the original pattern is in fact reversed, but in total both participating parties seem to gain value over the 6 months of deal negotiations. Upon completion of the deal, both acquirer and seller again react favorably and gain approximately 2% in market value. The short-term abnormal returns as presented by Table 5.2 confirm this first positive assessment of the transaction. Although Continental realizes a negative abnormal return of -1.27% upon the day of its first notion of intent, this negative reaction de-

5.4 Acquisition Performance

159

creases in the longer event periods around the first event date: Over the 41 days surrounding January 25th, Continental realizes a positive abnormal return of 6.26%, over the twenty days after the announcement a positive 2.18%. This positive reaction stands in clear contrast to preceding findings on abnormal returns to acquiring companies which conclude that abnormal returns to acquirers are essentially zero (Bruner (2002)) or negative (Loughran and Vijh (1997)). However, the returns are in line with positive returns previously determined in the automotive supply industry: Positive announcement returns to acquirers represent the capital market's perception of extraordinary synergy potentials in the industry (Mentz and Schiereck (2008)). Table 5.2: Abnormal Announcement Returns to Continental AG and Siemens AG Cumulative Abnormal Returns (CARs), in % EventWindow

January 25 Conti. Siemens

July 25 Conti. Siemens

[-20,20] [-20,10] [-10,10] [-5,5] [-1,1] [-1,0] [0] [0,1] [0,5] [0,10] [0,20]

6.26 4.16 4.05 2.65 -1.92 -2.63 -1.27 -0.56 0.73 0.08 2.18

-0.73 2.52 -0.32 6.93 2.43 3.70 3.58 2.31 3.14 -4.61 -7.86

13.33 10.72 6.77 8.98 7.59 6.84 6.29 7.04 9.89 6.64 9.26

-7.24 -4.79 -6.76 -6.26 -7.54 -4.16 -4.76 -8.14 -7.52 -7.62 -10.07

December 5 Conti. Siemens

4.37 5.52 4.47 2.74 4.20 0.35 0.92 4.77 5.36 5.77 4.62

27.45 28.26 16.05 5.36 5.35 3.46 1.37 3.26 1.93 8.44 7.63

This table shows the cumulative abnormal returns (CARs) to Continental AG and Siemens AG around three different event dates with respect to the Continental/Siemens VDO transaction. On January 25th, Continental announced its interest in acquiring a stake in Siemens VDO. On July 25th, both parties agreed on the deal terms and officially announced the takeover. Regulatory approval completed the transaction on December 5th. Market models are estimated by using OLS regression over a 200-trading-day period starting at trading day t = -250 relative to the first public mention of the transaction on January 25.

The positive reaction of Continental's stock returns is again confirmed for the other two event dates: On July 25th, Continental realizes abnormal returns of 3.58%;

160

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

upon deal completion in December, Continental's shareholders experience an additional positive abnormal return of 0.92%. Although some negative abnormal returns are also existent for Continental especially on the day of the first mentioning in January and the days after the deal announcement on July 25th, the abnormal returns in general remain positive and point toward a positive perception of the deal by capital markets as described above. At the three dates investigated here, investors appear to primarily perceive the inherent deal synergies as clearly outweighing the downside from overpaying for the acquisition. For Siemens AG, positive abnormal returns are mainly realized around the January and December announcement. It appears as if the negative returns in July represent a corrective measure decreasing share price to a regular level after a period of price increasing speculations about competitive bidders. In order to assess the long-term capital market performance of Continental, Figure 5.3 provides an overview of Continental's long-term share price development after July 25th, 2007. The graph contains daily return information for Continental AG, the German DAX 30 index and the Dow Jones Euro Auto & Parts index; all three time series are indexed to 100 on the day of the deal announcement. From Figure 5.3 it becomes apparent that Continental's total return index consistently develops worse than the comparable regional and industry indices. The figure also provides evidence that the appearance of the new takeover offer by the Schaeffler Group seems to temporarily offset the negative development on July 13th, 2008. For the purpose of this study, the derived return development is again challenged against abnormal returns in comparison to a peer-group benchmark. Therefore, long-term abnormal returns are determined as BHARs using a character-based matching approach. This matching procedure follows the approach proposed by Lyon et al. (1999) and determines the matching firm as the firm with a market value between 70% and 130% of Continental's market value and the smallest absolute difference in market-to-book ratios. The population of potential matches originates from the 130 companies comprised in the automotive supplier portfolio applied in Chapters 3 and 4. Based on the

5.4 Acquisition Performance

161

described matching procedure, US-based Johnson Controls represents the closest match for Continental AG. In order to extend the results from this one-on-one comparison, this study also applies three additional peer groups that serve as benchmarks in the following analysis. Peer group 1 comprises all remaining competitor companies in the portfolio with a market value between 70% and 130% of Continental (namely Michelin, Bridgestone, Asahi Glass, and Toyota Industries Corporation). For peer group 2, the market value criterion is extended to all companies between 50% and 150% of Continental's market value at the end of June 2007. This extension adds Alcan, Eaton, PPG Industries, and Aisin Seiki to the original peer group 1. Peer Group 3 is manually constructed, based on a matching product portfolio between the peers and Continental. This peer group is the largest and comprises Denso Corp., Michelin, Bridgestone, Goodyear, TRW Automotive, BorgWarner, Nokian Renkaat, Cooper Tire, and Autoliv. Figure 5.3: Long-term Relative Share Price Development of Continental AG 140 120 100 80 60 40 20

Continental AG

DAX 30

Jul-08

Jun-08

May-08

Apr-08

Mar-08

Feb-08

Jan-08

Dec-07

Nov-07

Oct-07

Sep-07

Aug-07

Jul-07

0

DowJones TM EURO Auto&Parts

The single matching firm and the different peer groups represent the benchmarks in determining long-term abnormal returns to Continental, which are derived as the difference between the Buy-and-Hold Return (BHR) of an investor in Continental and the BHR of an investor in the control-firm or an equally-weighted portfolio of control-firms. Table 5.3 presents the resulting Buy-and-Hold Returns for Continental AG and Johnson Controls.

162

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

Over a six-month holding period, an investor in Continental AG realizes a loss of -16.23%. In connection with the positive return to an investor in Johnson Controls, this loss corresponds to an abnormal underperformance of -20.79% (BHAR) over the six months following the VDO takeover. While being fairly significant in magnitude, the direction of this value effect is in line with previous findings of long-term postacquisition performance: For example, Loughran and Vijh (1997) find negative BHARs of -6.5% within a five-year time period following a transaction. The underperformance also corresponds to the preceding findings in the automotive supply industry. Chapter 3 provides evidence that an insignificant underperformance of -2.79% over a one-year period increases to a significant -16.68% over three years. Table 5.3: Buy-and-Hold Abnormal Returns to Continental AG BHARs (control-firm approach), in % BHR Acquirer Continental AG Time frame 6 months (months -1 to 5) -16.23 12 months (months - 1 to 11) -10.48

BHR Peer Johnson Controls 4.56 -9.96

BHAR -20.79 -0.52

This table shows the Buy-and-Hold Abnormal Returns to Continental AG following the Siemens VDO takeover. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999), the table differentiates between 6- and 12-month holding periods.

However, this pattern changes in the 12-month period. As presented in Table 5.3, Continental is able to decrease its negative returns to -10.48% in the longer time period. At the same time, the performance of the matching firm is significantly worse and falls from a positive return to -9.96%. As a result, the abnormal underperformance of Continental decreases to -0.52% and almost becomes non-existent. Consequently, it becomes apparent that Continental is not only able to improve its self-standing capital market performance but also gains ground in relation to its matching peer. Table 5.4 confirms that this finding remains robust against a change in peer group compilation. Across all three determined peer groups, Continental is able to reduce its abnormal underperformance from a clear negative BHAR to almost non-exiting negative abnormal returns. In two of the cases, Continental even realizes positive abnormal returns under-

5.4 Acquisition Performance

163

lining once again its extraordinary positive return position. Table 5.4 also contains the results of an extended market model estimate as applied in the short-term analysis. Based on market model estimates derived in the short-term analysis, expected daily returns for the full year are compared against the actual performance of the Continental share. The results are in line with the determined BHARs and yield a positive abnormal return of 6.05% (12 months). Table 5.4: Abnormal Returns to Continental AG versus Different Peer Groups Abnormal Returns, in % Approach BHAR (Single Peer) BHAR (Peer Group 1) BHAR (Peer Group 2) BHAR (Peer Group 3) Market Model

6 months (months -1 to 5) -20.79 -3.80 -11.38 -7.07 -13.70

12 months (months -1 to 11) -0.52 6.65 0.80 -1.62 6.05

This table shows abnormal returns to Continental AG following the Siemens VDO takeover subject to different methodologies and peer groups. Abnormal returns are derived using a control-firm matching approach as proposed by Lyon et al. (1999) and the standard market model as described by Brown and Warner (1985). The table differentiates between 6- and 12-month holding periods.

Therefore, it can be concluded that at first Continental does suffer a significant underperformance as a direct reaction to the takeover. However, it is also able to quickly recover from this underperformance and offsets the negative capital market effect within a year. Over a 12-month horizon, Continental realizes an insignificant underperformance of at most -1.6%; depending on the peer group benchmark, the underperformance even disappears and turns into a positive outperformance. The extraordinary synergy potential perceived in the short-term announcement returns appears to be realized in the long-run. After takeover costs and premium are paid, capital markets perceive Continental to be able of quickly recovering and realizing the originally anticipated synergies. Given the positive short-term announcement returns and the positive trend in long-term abnormal returns, capital markets perceive the VDO takeover as a

164

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

strong success and put Continental in a positive position compared to its non-merging peers. The following section focuses on the long-term development of published accounting information and provides an additional opportunity to support this positive capital market assessment. In addition, it is also determined in how far the positive stock returns in 2008 are already influenced by rumors about an upcoming takeover bid by the Schaeffler group. 5.4.2 Performance Analysis While the preceding abnormal return analysis focuses on the perspective and expectations of capital markets about the takeover transaction, a performance analysis of published accounting data allows for a more detailed analysis of the realized synergies and the corresponding value creation. With its inherent focus on the past, it is not exposed to the subjective assessment of investors constantly incorporating future expectations into the stock price. Still, the quality of the derived findings depends on the quality of available published accounting information. In the Continental/Siemens VDO transaction, the consolidation of the external financial reporting did not take place until after the regulatory approval of the acquisition on December 5th, 2007. Consequently, the 2007 annual report of Continental to date represents the only annual report including the financial performance of both companies. Appendix 7 provides an overview of Continental's past annual balance sheets and income statements. It shows how the takeover of VDO significantly increases Continental' total assets as well as non-current liabilities. However, the full consolidation is only evident in balance sheet items, the income statement only includes one month of combined performance. Consequently, an analysis of annual data sources remains inconclusive in assessing the long-term post-acquisition performance of Continental. For the remainder of this section, the analysis of accounting information focuses on information published in quarterly reports of Continental AG. Unlike the annual statements, an analysis of the quarterly reports yields four reporting instances after the

5.4 Acquisition Performance

165

original deal announcement in July 2007, three of which include the consolidated Continental AG after taking over Siemens VDO. Table 5.5 provides an overview of the main balance sheet and income statement items and outlines the significant size effect of the analyzed merger. In the fourth quarter of 2007, total assets of Continental more than doubled from EUR 12.0 billion to EUR 27.3 billion. In order to finance the EUR 11.4 billion acquisition, Continental issued additional shares worth EUR 1.48 billion in the third quarter of 2007 and increased its long-term liabilities to finance the remainder. As a result, it is able to significantly increase the leverage on its balance sheet. The physical transfer of the transaction price becomes apparent in the extraordinary free cash flow of the fourth quarter that accumulates to EUR -10.7 billion. Table 5.5: Quarterly Balance Sheet/Income Statement Items of Continental AG Balance Sheet and Income Statement Items (in EUR million) Q3/06 Total Assets

Q4/06

Q1/07

Q2/07

Q3/07

Q4/07

Q1/08

Q2/08

11,375.7 10,853.0 11,750.8 11,927.7 12,091.2 27,737.6 26,910.2 27,077.5

Equity

4,320.1 4,709.9 4,970.8 4,965.4 5,191.6

Non-current liabilities

2,735.7 2,156.8 2,076.6 2,040.1 2,381.3 11,668.3 11,664.9 11,665.7

Sales

3,714.4 3,941.7 3,964.8 4,049.1 3,906.6

Cost of Goods Sold

-2,821.0 -2,919.7 -2,964.1 -3,023.4 -2,925.5

6,856.1

4,698.9

6,912.3 7,019.6

6,639.4 6,614.6

-3,682.6 -5,252.6 -5,203.7

R&D Expenses

-185.8

-174.5

-185.2

-204.0

-187.8

-257.8

-415.2

-424.4

Selling, Admin.&Other Expenses

-314.7

-372.8

-389.0

-352.1

-372.7

-424.7

-535.3

-551.0

572.4

696.1

613.6

650.0

605.5

621.5

884.0

890.8

-178.3

-210.3

-176.8

-175.3

-179.4

-283.3

-427.3

-435.1

EBIT

394.1

485.8

436.8

474.7

426.1

338.2

456.7

455.7

Net Income

240.8

326.7

277.3

308.2

258.2

206.2

179.9

206.3

Operating Cash Flow

159.8

699.0

66.6

279.3

364.8

1,202.9

19.1

617.1

Free Cash Flow

-659.4

361.0

-119.8

104.3

77.1 -10,687.2

-316.7

469.5

# Employees

84,561

85,224

87,284

89,082

EBITDA Depreciation/Amortization

89,375

151,654 153,587 149,113

Source: Quarterly Reports

A similar size effect relates to the income sheet items. Quarterly sales increase from an average of EUR 4 billion to more than EUR 6.6 billion in the first quarter of

166

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

2008. Likewise, the cost items increase as a result of combining the two entities. Besides this simple combination effect, however, the effect of realized synergies and cost efficiencies remains rudimental at this early stage after the takeover while yielding mixed results. While quarterly revenues decrease slightly from the first to the second quarter 2008 despite the targeted revenue synergies, Continental is able to reduce COGS disproportionately by almost EUR 50 million. Selling and R&D expenses, though, increase and dilute a potential effect on earnings. In addition, significant increases in depreciation and interest charges due to the merger nearly offset the entire size effect on the earnings items; EBIT and Net Income remain almost constant or lose compared to pre-takeover levels. However, Continental gives an indication of the general trend as it manages to slightly increase EBITDA and Net Income from Q1 and Q2 2008. Table 5.6 provides an overview of the resulting performance indicators and generally supports the previous findings concerning a size and a preliminary synergy effect. Upon consolidation of both entities, all profit-oriented performance indicators decrease as a result of increasing cost items and of the significant interest and depreciation expense associated with the transaction. The revenue-oriented indicators also decrease as they suffer from a disproportionate increase in total assets and number of employees. However, the preliminary synergies effect as evident in the indicator changes from the first to the second quarter of 2008 appears to be more positively represented by the performance indicators. Except for the EBIT/Total Assets indicator, Continental is already able to grow all other profit-oriented indicators in 2008. The cost-income-ratio decreases slightly driven by savings in the costs of goods sold. In accordance with the original deal motivation, expenses for research and development activities slightly increase. The cost synergies in administrative functions do not become effective to date, SG&A expenses increase from 8.06% to 8.33%. Although the preceding analysis provides some evidence for the existence of a positive post-takeover performance trend, it cannot determine whether this trend originates from synergies related to the takeover or from general market conditions affecting

5.4 Acquisition Performance

167

the complete automotive supply industry. Therefore, this study also analyzes a peergroup adjusted set of the same performance indicators provided in Table 5.6. For this purpose, quarterly results of all peer companies in peer group 3 are collected from their quarterly reports and aggregated into a simple average performance. The performance indicators of Continental are then reduced by this performance to yield the abnormal performance attributable to the VDO takeover. Table 5.7 indicates how Continental consistently outperforms the peer-group benchmarks over the years preceding the VDO takeover. Profit- and revenue-oriented indicators outperform the industry by several percentage points. Table 5.6: Unadjusted Quarterly Performance Indicators of Continental AG Q3/06

Q4/06

Q1/07

Q2/07

Q3/07

Q4/07

Q1/08

Q2/08

Profit-oriented Performance Indicators (in %) EBITDA/Total Assets EBITDA/Sales EBIT/Total Assets

5.03

6.41

5.22

5.45

5.01

2.24

3.28

3.29

15.41

17.66

15.48

16.05

15.50

13.23

13.31

13.47

3.46

4.48

3.72

3.98

3.52

1.22

1.70

1.68

10.61

12.32

11.02

11.72

10.91

7.20

6.88

6.89

Net Income/Total Assets

2.12

3.01

2.36

2.58

2.14

0.74

0.67

0.76

Net Income/Sales

6.48

8.29

6.99

7.61

6.61

4.39

2.71

3.12

CF/Total Assets

1.40

6.44

0.57

2.34

3.02

4.34

0.07

2.28

EBIT/Sales

CF/Sales Cost Income Ratio -COGS/Sales

4.30

17.73

1.68

6.90

9.34

25.60

0.29

9.33

89.42

87.96

89.24

88.40

89.23

92.90

93.43

93.42

75.95

74.07

74.76

74.67

74.89

78.37

79.11

78.67

-R&D Expenses/Sales

5.00

4.43

4.67

5.04

4.81

5.49

6.25

6.42

-Selling Expenses/Sales

8.47

9.46

9.81

8.70

9.54

9.04

8.06

8.33

Revenue-oriented Performance Indicators Sales/Total Assets (in %)

32.65

36.32

33.74

33.95

32.31

16.94

24.67

24.43

Sales/Employee (in EUR)

43,926

46,251

45,424

45,454

43,710

30,984

43,229

44,360

Balance-Sheet Structure (in %) Equity/Total Assets

37.98

43.40

42.30

41.63

42.94

24.72

25.69

25.92

Debt/Equity-Ratio

163.32

130.43

136.40

140.22

132.90

304.57

289.31

285.74

Source: Quarterly Reports, own calculations

168

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

Table 5.7: Abnormal Quarterly Performance Indicators of Continental AG Q3/06

Q4/06

Q1/07

Q2/07

Q3/07

Q4/07

Q1/08

Q2/08

Profit-oriented Performance Indicators (in %) EBITDA/Total Assets

2.44

3.50

2.28

2.06

1.40

-1.88

-0.06

-0.01

EBITDA/Sales

4.67

7.81

4.15

3.20

1.65

-0.84

0.81

1.08

EBIT/Total Assets

2.14

2.88

2.08

1.85

1.18

-1.57

-0.30

-0.28

EBIT/Sales

4.79

7.35

4.50

3.45

1.78

-2.29

-0.80

-0.74

Net Income/Total Assets

1.32

2.01

1.51

1.25

0.44

-1.02

-0.67

-0.44

Net Income/Sales

2.79

5.34

3.45

2.29

0.01

-1.57

-2.55

-1.67

Operating CF/Total Assets

1.20

-0.14

0.94

-0.23

2.09

-3.87

0.14

2.00

Operating CF/Sales

4.74

-4.36

4.32

-1.96

7.38

-1.71

1.51

8.89

Cost Income Ratio

-4.29

-5.17

-3.54

-3.15

-2.84

2.46

1.19

1.10

-3.98

-5.69

-4.16

-2.80

-4.20

1.03

0.49

0.29

-COGS/Sales -R&D Expenses/Sales -Selling Expenses/Sales

0.64

0.28

0.52

1.03

0.70

1.87

2.30

2.64

-4.09

-2.60

-2.88

-4.03

-2.36

-2.91

-4.42

-4.31

Revenue-oriented Performance Indicators Sales/Total Assets (in %)

6.21

8.84

6.27

6.01

5.06

-12.46

-3.54

-4.10

Sales/Employee (in EUR)

-3,060

-6,195

-2,368

-7,315

-4,230

-24,946

-4,827

-7,909

Balance-Sheet Structure (in %) Equity/Total Assets Debt/Equity-Ratio

2.97

9.24

7.43

7.15

6.58

-13.65

-15.02

-13.05

-34.14

-54.95

-43.30

-43.48

-38.75

143.43

144.05

129.71

Source: Quarterly Reports, own calculations

Upon the consolidation with VDO, this outperformance turns into an underperformance representing the incremental efforts of the takeover such as implementation costs and additional interest expenses. However, over the following three quarters, Continental continuously develops all indicators towards benchmark levels. EBITDA/Sales already beat the benchmark by more than one percentage point in the second quarter of 2008. It becomes apparent that Continental is able to regain relative performance quickly and could potentially outperform its peer group in the following quarters. Although the intended cost synergies of at least EUR 170 million per year cannot be traced and evaluated at this early stage, the relative performance trend supports the assumptions that Continental will manage to realize the cost synergies by 2010 as presented in

5.4 Acquisition Performance

169

its original press release. From the perspective of an investor, the overall performance development is therefore positive; there are convincing indications that the takeover was successful and remains successful in the future. Figure 5.4: Revenue Split of Continental AG across Divisions Total Revenues 100%

80% ContiTech Commercial Vehicle Tires

60%

Passenger and Light Truck Tires Interior 40%

Powertrain Chassis and Safety

20%

0% Q1/07

Q2/07

Q3/07

Q4/07

Q1/08

Q2/08

Source: Quarterly Reports

While these observations hold true for the majority of cost- and efficiencydriven synergies, a conclusion about the realization of revenue synergies turns out to be less obvious. Although Section 5.3 provides an overview of the different revenue synergies underlying the transaction, quarterly revenues seem to decrease after the takeover became effective. However, this observation is derived from the very limited data sources of two post-consolidation quarterly reports. As Continental also plans on discontinuing a few competitive products from the VDO product portfolio, these efforts could potentially already account for the observed revenue decrease. In addition, many of the described revenue synergies require significant time spans to become effective. Especially the synergies in research and development most likely become effective after years of research work. Therefore, a comprehensive judgment on the realization of revenue synergies as part of the VDO takeover could not be obtained. However, Continental manages to achieve some of the structural revenue changes outlined in Section 5.3. From Figure 5.4, it becomes apparent how Continental is able to shift its product

170

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

focus from its original tire business into the newly-defined growth sectors interior, powertrain, and chassis/safety. The dependency of Continental on its tire business is reduced from 40% of revenues to approximately 20% after the takeover.

5.5

Discussion and Conclusion

Regarding the post-acquisition performance of acquirers, Mitchell and Lehn (1990) made a significant contribution and asked whether bad bidders subsequently become good targets. Based on a transaction sample from the 1980s, they find that firms which subsequently become takeover targets made acquisitions significantly reducing their equity value. However, firms that do not become takeover targets were able to substantially increase their equity value. Not more than a year after the VDO-takeover, the German-based Schaeffler Group started a hostile takeover bid for Continental AG, which ultimately led Continental's management to accept a bid of EUR 75 per share in August 2008. Given that Continental became a target within months of its VDO takeover, the original contribution of Mitchell and Lehn (1990) raises two questions: Firstly, was Continental a bad bidder? And secondly, if it wasn't, why did it become a good target within such a short time period? With regard to the first question, this case study presents strong evidence that Continental is in fact not a "bad bidder" entering into a value-reducing acquisition. Unlike the results of Mitchell and Lehn (1990), Continental realizes a positive announcement return of +3.58% on the day of the deal announcement and an additional +0.92% on the day of the regulatory deal approval. It becomes apparent that capital markets seem to appreciate the underlying synergy potentials imminent in the industry and particularly in this transaction. The results are in line with previous findings in the automotive supply industry (Mentz and Schiereck (2008)). The only outlier in this pattern is the negative announcement return determined on the day of Continental's first statement of interest in January 2007. However, these negative returns disappear over

5.5 Discussion and Conclusion

171

longer event-windows when the acquisition offer becomes more concrete. Overall, a positive short-term assessment of the value impact prevails. The assessment of Continental's long-term performance supports the positive outcome of the short-term announcement returns. Although the derived BHARs indicate a severe loss of up to -20.79% over a six-month holding period, these losses vanish over a 12-month horizon and turn to an abnormal outperformance when measured against three out of five benchmarks. In this regard, Continental does suffer an initial decrease in long-term share price performance as predicted by preceding research (Chapter 3), but it is quickly able to recover to an outstanding positive position in comparison to its peers. The accounting performance follows a similar pattern: Although all indicators almost consistently fall from above-benchmark levels to below-market levels, a clear development path towards reaching the benchmarks in the near future can be determined. Overall, it can be concluded that Continental does not permanently reduce equity value and, therefore, cannot be considered a "bad bidder." The question remains why Continental nonetheless became a good target. As part of this discussion, two lines of thought are offered potentially contributing to Continental's future takeover. One line relates to the issue of timing. While Continental focused on quickly realizing synergies, the few months between consolidation of VDO and the takeover bid by Schaeffler did not suffice to benefit from the first synergies and thereby recover from integration charges. Although Continental was on the best way to reach market levels again, Schaeffler's hostile bid arrived as Continental was still below or exactly at market levels both in terms of share price and accounting performance. While Schaeffler could not have afforded to acquire Continental during its performance peak before July 2007, chances exist that substantially higher acquisition efforts would have been required to acquire Continental after it had regained its strengths at some point in 2009. A second line of thought relates to the relative position of Continental in its long-term performance. While measured against peer benchmarks, the capital market

172

5 Study 3: How a Good Bidder Becomes a Good Target – The Case

returns and accounting performance of Continental are both regaining position. After 12 months, the underperformance in capital market returns has almost vanished. However, these relative results do not change the fact that both the industry as a whole as well as Continental by itself lost significant equity values. As stated before, Continental shareholders lost approximately EUR 4 billion in equity value within a year. In this situation, privately held Schaeffler took advantage of Continental's unfavorable, absolute value position. With regard to the objectives of this case study, it is generally concluded that Continental AG was a "good bidder" in acquiring Siemens VDO. However, it is also assumed that the "bad bidder, good target" discussion remains newsworthy.

6

Conclusion This thesis determines and analyzes the long-term success of mergers and acqui-

sitions in the international automotive supply industry. A detailed and thorough analysis of this topic becomes particularly interesting for two main reasons, namely the relevance of the automotive supply industry as a research object and the general complications involving the measurement of long-term M&A success. With regard to the earlier, the competitive pressures within individual product segments, the growing demands of OEMs eager to produce higher quality at lower costs, and the increasing raw material prices have created a uniquely competitive situation in which M&A became a common strategic response for many suppliers. As a result, this industry has been facing significant consolidation activity continuously decreasing the number of active players over the last twenty years. At the same time, previous research determines the industry to be an outlier: Unlike other industries, acquirers in the automotive supply industry are able to generate significant positive announcement returns (Mentz and Schiereck (2008)). As capital markets appear to initially appreciate M&A as a valuable response strategy, this thesis addresses the question whether these suppliers are able to sustain their outstanding positive returns in the long-run. The on-going consolidation among suppliers underlines the academic and hands-on need to understand the overall effect as well as its underlying determinants. The second reason contributing to the overall relevance of the topic relates to several complications involving the measurement of long-term M&A success. Unlike the well-proven event-study methodology frequently applied in assessing short-term announcement returns, a number of different research approaches exist to determine long-term transaction success: The long-term performance of acquirers can be determined based on different data sets (share returns vs. accounting information), different statistical methodologies (calendar-time vs. event-time) and different research archetypes (event studies vs. case studies). As a result, a comprehensive assessment of the long-term value creation potential through M&A must consider a number of different

174

6 Conclusion

research approaches and build a consistent judgement across different results. Consequently, this thesis follows three different objectives: (1) determining the overall longterm success of acquirers in the automotive supply industry by answering the question whether these acquirers are able to sustain their extraordinary positive return position beyond a short-term perspective, (2) analyzing a comprehensive set of determinants for the long-term post-merger performance enabling senior management to derive a valuemaximizing M&A strategy, and (3) challenging the derived performance against results from a most comprehensive set of available research approaches ensuring a valid and consistent assessment of the long-term value creation potential. Addressing the first objective, the first empirical study of this thesis analyzes the long-term post-merger capital market performance of acquirers in 230 horizontal takeover transactions involving automotive suppliers between 1981 and 2007. After validating the outstanding positive announcement returns described in preceding literature, a combination of the matching-firm approach in event-time and the Fama-French-3Factor-model in calendar-time reveals that acquirers in the automotive supply industry are in fact not able to sustain positive returns in the long-run. On the contrary, both long-term performance models indicate a consistent value destruction of 16 to 20% over a three-year period following a takeover announcement. While being consistent with the majority of academic literature on the long-term value creation through M&A, this result is yet more negative pointing to the assumption that the above-average synergy potentials initially perceived by capital markets cannot be sufficiently realized by the acquirers in the long-run. Although entering promising M&A transactions at first, the comparison to different return benchmarks reveals that acquirers in the automotive supply industry rather realize significant value losses in the years following the M&A transactions. With regard to the second overall objective, this first study also analyzes the impact of different return determinants. The derived results provide evidence for a consistent negative impact of geographic expansion and of lacking acquisition experience on

6 Conclusion

175

long-term capital market performance. One explanation for the negative impact of internationalization lies in the operational challenges and required resources in order to complete a transaction across country borders or even across continents. Consequently, the more challenging and resource-intense post-merger integration of cross-border transactions negatively impacts post-merger performance of automotive suppliers. Likewise, a lack of acquisition experience decreases post-merger capital market returns. In this case, it can be assumed that multi-bidders carry experience in integrating takeover targets, which enables them to consistently outperform their inexperienced peer group in the long-run. The size of an M&A transaction, however, appears to have significant positive effect on long-term post-merger performance given a higher probability to experience economies of scale. While these insights advance previous literature on performance determinants, they likewise enable top management to identify a corresponding value-maximizing acquisition strategy: a first step of such a strategy comprises the integration of smaller, national targets to gain M&A experience. Afterwards, acquirers should leverage their gained experience in a large cross-continental deal gaining access to the global synergy structure of this industry. The second empirical study addresses similar objectives as the first, but rather focuses on analyzing the acquirers' operating performance as provided in their accounting statements than on their capital market performance. By introducing an additional research approach based on alternative performance data, it challenges the significant underperformance previously determined for acquirers in the industry and aims at creating a more consistent view in accordance with the third objective of the study. For this purpose, a subsample of the original transaction sample, namely 93 horizontal takeover transactions involving automotive suppliers, is analyzed for its long-term peer-adjusted operating performance change. In general, acquirers in this industry also underperform their non-merging matching firms in terms of operating performance changes. Over the three years following a transaction, the operating performance of acquirers develops between -4% and -12% worse than the performance of non-merging peers, depending on the observed performance indicator. The peer-adjusted abnormal change in opera-

176

6 Conclusion

tional cash flow over sales is a significant -10.52% (5% level) lower than the change in the same indicator of non-merging rivals. Consequently, as the abnormal operating performance concurs with the losses determined in capital market returns, it is also in line with capital market's continuous assessment of the acquirers' performance and with the assumption that acquirers were unable to realize the outstanding synergies initially perceived by capital markets. In addition, this second study provides evidence that similar determinants impact both operating performance as well as long-term capital market performance. Due to larger resource requirements, diversifying and cross-border transactions also negatively influence firm profitability. In the light of the third objective, a consistent view on the longterm value creation through M&A within the industry emerges. To further probe consistency, the second study also determines the correlation between the observed profitability changes and the event-study results determined in the first study. The derived regression models indicate that an analysis of firm profitability in the automotive supply industry can serve as a substitute not only for short-term announcement returns but also for long-term event studies based on Buy-and-Hold Abnormal Returns. While the statistical quality of these relations seems to depend on the applied performance indicator, the models also reveal the existence of potential surprise- and anticipation premiums. Longterm capital market returns positively reward a divergence of long-term BHARs and firm profitability development; and a divergence between short-term CAARs and firm profitability is followed by a consistent decrease in long-term capital market returns as an expression of investor disappointment with synergy realization. In general, the first two studies point towards a significant and consistent longterm underperformance of acquirers in the automotive supply industry. The third empirical study included in this thesis explicitly turns toward a positive outlier by introducing the case of Continental AG acquiring Siemens VDO. Besides its significant value, the fact that Continental became a takeover target itself within a one-year time span significantly contributes to the relevance of this transaction. Addressing the popular

6 Conclusion

177

discussion on how bad bidders become good targets (Mitchell and Lehn (1990)), this study finds that Continental is in fact not a bad bidder. Upon the announcement of the takeover, Continental realizes a positive announcement return of +3.58% and an additional +0.92% on the day of the regulatory deal approval. It becomes apparent that capital markets seem to appreciate the underlying synergy potentials imminent in the industry and particularly in this transaction. The assessment of Continental's long-term performance supports the positive outcome of the short-term announcement returns. Although BHARs indicate a severe loss of up to -20.79% over a six-month period, these losses vanish over a 12-month horizon and turn into an abnormal outperformance when measured against three out of five benchmarks. In this regard, Continental does suffer an initial decrease in long-term share price performance as predicted by the first two studies, but it is quickly able to recover to an outstanding positive position in comparison to its peers. The accounting performance follows a similar pattern: Although all indicators almost consistently fall from above-benchmark levels to below market-levels, a clear development path towards reaching the benchmarks in the near future is determined. Overall, it becomes apparent that Continental does not permanently reduce equity value and, therefore, cannot be considered a "bad bidder." On the contrary, this case study provides a valuable example for suppliers intending to engage in M&A activity. It offers an exemplary path for a consistent, target-oriented post-merger management process that achieves to recover from the largest deal in the industry's history within a one-year time frame. This positive example might also hold the key to the on-going consolidation observable among automotive suppliers today. As both general studies consistently point towards an overall long-term underperformance of acquirers within the industry, the legitimate question arises why M&A are still occurring and potentially increasing among automotive suppliers. One potential answer lies in the transaction success of Continental, which proves that there is a way to achieve long-term success and, eventually, increase shareholder value through M&A. Continental rigorously identified and

178

6 Conclusion

pursued future growth markets by choosing a target to deliver complementary technical strengths and product competencies. After the transaction was legally completed, Continental immediately focused on integrating the target and realizing revenue and cost synergies. The strategic fit in connection with a stringent and immediate realization of synergies turn out to be the specific key success factors in this case. In addition, Continental's acquisition also follows the positive determinants developed in the preceding event and accounting studies. Continental is an experienced acquirer that engaged in a large national deal yielding a complementary product scope. Consequently, following the proposed acquisition strategy in connection with the key success factors from the Continental case potentially enables acquirers to achieve long-term post-merger success despite the overall negative picture yielded by the general event and accounting studies. A second reason for the continuing consolidation potentially lies in the risk of becoming a transaction outsider. Mentz and Schiereck (2008) show how non-merging rivals, especially within the same country as an acquiring automotive supplier, suffer from short-term value losses. As the merging companies gain market strength, their increased negotiation power and efficiency exerts negative value effects on the nonmerging rivals. As a result, non-merging rivals might risk entering unprofitable deals in order to prevent themselves from becoming a transaction outsider. With merger waves increasingly affecting the industry, many players are consequently entering a valuedestroying spiral: With every transaction, the pressure to enter the next transaction increases. At the same time, the time to properly integrate targets decreases and acquirers are facing significant long-term losses as determined in this thesis. However, as the focus of this study does not include a self-contained analysis of the long-term performance of non-merging rivals, further research could focus on determining the reasons for automotive suppliers to engage or refrain from M&A activity including an assessment of the "transaction outsider hypothesis" as described above. In this regard, further research could determine strategic alternatives to M&A, which potentially likewise promote long-term value creation within the industry.

6 Conclusion

179

In addition, this thesis focuses on analyzing a resource-intense production industry exposed to various competitive pressures. Further research could expand the presented findings by validating them against the specific situation in other production and service industries. In this way, the applicability of the derived acquisition strategy and key success factors could be generalized across industries or, alternatively, be tied to the specific situation of the automotive supply industry. With regard to the timing of the presented research, the chosen transaction sample comprises the early transactions of the 1980s, the large M&A wave of the 1990s, as well as the potential start of a new transaction wave inherent in the takeover of Siemens VDO in 2007. As the long-term assessment of M&A success also carries a continuous perspective, expanding the observed time horizon beyond 2007 and analyzing value changes after the next potential merger wave might also carry additional insights into continuous long-term success of acquirers in the automotive supply industry. While the inability to realize synergy potentials currently drives the consistent underperformance of acquirers, an increasingly consolidated industry setting might carry additional positive effects such as increased negotiation power, economies of scale and scope, and global customer access. Further research focusing on later time periods might determine that these effects partly or in total offset the negative value impact presented above. The case study on Continental presented a very positive case example on how an automotive acquirer might succeed in the long-term. Additional research could diversify the presented key success factors by analyzing other case studies of different size, product focus and geographical scope. Thereby, further research could differentiate the derived key success factors and tailor them to various economic situations suppliers are facing. With regard to the presented accounting analysis, the focus of this thesis lies on providing a most comprehensive view on value creation by applying various available methodologies. Further research could carry out a more detailed assessment of the longterm profitability effect. By differentiating the impact on different cost items, future

180

6 Conclusion

studies might follow the objective to determine the source for a long-term underperformance such as a lack of R&D expenditure, costly raw material sourcing, or managerial overconfidence in granting salaries. This thesis provides a comprehensive assessment of the long-term success of M&A in the automotive supply industry. Across a number of different research approaches, it determines a consistently negative picture of the acquirers' long-term performance, which seem to be unable to live up to the positive expectations carried by capital markets. However, this thesis also provides evidence that there is a way to escape this negative destiny: following a mature acquisition strategy in connection with a rigorous post-merger realization of synergies. As a result, further consolidation within the automotive supply industry does not solely provide a threat for their capital market and operating performance, but also represents a strong opportunity for automotive suppliers to realize consistent and long-term success through mergers and acquisitions.

Appendix Appendix 1: CAARs to Target Companies .................................................................. 182 Appendix 2: CAARs to Combined Entities.................................................................. 183 Appendix 3: Daily Average Abnormal Returns to Acquirers and Targets .................. 184 Appendix 4: Value-weighted Abnormal Performance Changes .................................. 185 Appendix 5: CAARs to Acquirers Applied in the Accounting Study.......................... 186 Appendix 6: Regression of Abnormal Performance Changes on 11-day CAARs ....... 187 Appendix 7: Annual Balance Sheet/Income Statement Items of Continental AG ....... 188

182

Appendix

Appendix 1: CAARs to Target Companies Targets (n=111) Event-

t-Test p-value

z-Test

Window

CAAR

t-value

z-value

[-20,20]

15.24%

4.99

<0.01 ***

7.09

[-20,10]

15.79%

5.39

<0.01 ***

[-10,10]

15.02%

5.85

<0.01 ***

[-5,5]

13.64%

6.03

[-1,1]

11.12%

5.87

p-value

Gen. Sign Test P

z-Value

<0.01 ***

61%

3.44

<0.01 ***

p-value

8.06

<0.01 ***

68%

4.77

<0.01 ***

8.99

<0.01 ***

66%

4.39

<0.01 ***

<0.01 ***

10.54

<0.01 ***

67%

4.58

<0.01 ***

<0.01 ***

12.48

<0.01 ***

68%

4.96

<0.01 ***

[-1,0]

9.03%

5.44

<0.01 ***

11.84

<0.01 ***

70%

5.34

<0.01 ***

[0]

7.59%

5.25

<0.01 ***

13.10

<0.01 ***

71%

5.54

<0.01 ***

[0,1]

9.68%

5.65

<0.01 ***

12.97

<0.01 ***

67%

4.58

<0.01 ***

[0,5]

10.20%

5.48

<0.01 ***

10.19

<0.01 ***

65%

4.20

<0.01 ***

[0,10]

10.51%

5.33

<0.01 ***

8.61

<0.01 ***

60%

3.25

<0.01 ***

[0,20] 9.96% 4.78 <0.01 *** 6.76 <0.01 *** 59% 3.06 <0.01 *** This table shows the cumulative average abnormal returns (CAAR) to target companies in mergers and acquisitions in the automotive supply industry. It contains all public targets for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test, the cross-sectional test as proposed by Boehmer et al. (1991) (z-Test), and the Generalized Sign Test as proposed by Cowan et al. (1990). P stands for the percentage of transactions in the sample that result in a positive abnormal return.

Appendix

183

Appendix 2: CAARs to Combined Entities Combined Entities (n=101) Event-

t-Test

z-Test

Gen. Sign Test

Window

CAAR

t-value

p-value

z-value

p-value

P

z-Value

[-20,20]

4.50%

3.82

<0.01 ***

3.58

<0.01 ***

62%

3.31

<0.01 ***

p-value

[-20,10]

4.45%

4.06

<0.01 ***

3.90

<0.01 ***

68%

4.51

<0.01 ***

[-10,10]

3.59%

4.13

<0.01 ***

3.94

<0.01 ***

59%

2.71

<0.01 ***

[-5,5]

3.71%

4.57

<0.01 ***

4.87

<0.01 ***

64%

3.71

<0.01 ***

[-1,1]

2.35%

4.41

<0.01 ***

4.96

<0.01 ***

67%

4.31

<0.01 ***

[-1,0]

2.20%

5.08

<0.01 ***

5.71

<0.01 ***

64%

3.71

<0.01 ***

[0]

2.12%

5.19

<0.01 ***

7.01

<0.01 ***

68%

4.51

<0.01 ***

[0,1]

2.27%

4.55

<0.01 ***

5.62

<0.01 ***

65%

3.91

<0.01 ***

[0,5]

2.66%

4.54

<0.01 ***

4.92

<0.01 ***

70%

4.91

<0.01 ***

[0,10]

2.50%

3.71

<0.01 ***

3.82

<0.01 ***

64%

3.71

<0.01 ***

[0,20] 2.54% 2.98 <0.01 *** 2.95 <0.01 *** 54% 1.71 0.04 ** This table shows the cumulative average abnormal returns (CAAR) to the combined entities in mergers and acquisitions in the automotive supply industry. It contains all potential public-public combinations for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using a standard t-Test, the cross-sectional test as proposed by Boehmer et al. (1991) (z-Test), and the Generalized Sign Test as proposed by Cowan et al. (1990). P stands for the percentage of transactions in the sample that result in a positive abnormal return.

184

Appendix

Appendix 3: Daily Average Abnormal Returns to Acquirers and Targets Acquiring Companies (n=206) Target Companies (n=111) AAR z-value p-value AAR z-value p-value Event Day -20 -0.09% 0.08 0.47 0.32% 0.37 0.35 -19 0.22% 1.99 0.02 ** -0.01% 0.97 0.17 -18 -0.12% -0.70 0.76 0.60% 1.89 0.03 ** -17 0.20% 1.72 0.04 ** 0.12% 0.18 0.43 -16 -0.08% -0.72 0.77 -0.32% -0.67 0.75 -15 -0.09% -0.59 0.72 -0.30% -0.89 0.81 -14 0.01% 0.19 0.42 0.07% 0.14 0.45 -13 0.05% -0.05 0.52 -0.07% -0.61 0.73 -12 -0.07% 0.05 0.48 0.03% 0.22 0.41 -11 0.16% 1.24 0.11 0.33% 0.81 0.21 -10 -0.03% -0.22 0.59 0.19% 0.20 0.42 -9 -0.14% -1.07 0.86 -0.09% -0.75 0.77 -8 0.16% 1.03 0.15 0.22% 0.36 0.36 -7 0.09% 0.70 0.24 0.12% 0.84 0.20 -6 0.21% 0.60 0.27 0.63% 2.38 0.01 *** -5 0.40% 1.12 0.13 0.49% 1.73 0.04 ** -4 -0.18% -0.94 0.83 0.08% 0.80 0.21 -3 0.02% -0.10 0.54 0.74% 2.57 0.01 *** -2 -0.01% -0.39 0.65 0.70% 2.51 0.01 *** -1 0.03% -0.26 0.60 1.43% 3.08 0.00 *** 0 0.94% 4.81 0.00 *** 7.59% 13.40 0.00 *** 1 0.78% 3.78 0.00 *** 2.09% 4.44 0.00 *** 2 0.29% 1.09 0.14 0.21% 1.66 0.05 ** 3 0.16% 1.56 0.06 * -0.28% -1.19 0.88 4 -0.22% -1.05 0.85 0.39% 1.30 0.10 * 5 0.08% 1.26 0.10 0.19% 0.79 0.21 6 -0.24% -1.37 0.92 0.09% 0.74 0.23 7 -0.01% -0.48 0.68 -0.02% -0.39 0.65 8 -0.19% -1.17 0.88 0.06% 1.11 0.13 9 -0.13% -0.96 0.83 -0.25% -1.36 0.91 10 -0.21% -1.24 0.89 0.44% -0.11 0.54 11 0.03% -0.07 0.53 -0.10% -0.41 0.66 12 -0.13% -1.08 0.86 -0.31% -1.10 0.86 13 0.13% 0.78 0.22 0.09% 0.09 0.46 14 0.02% 0.13 0.45 -0.16% -0.94 0.83 15 0.05% 0.58 0.28 0.12% 0.64 0.26 16 0.12% 1.35 0.09 * -0.25% -0.49 0.69 17 0.21% 1.48 0.07 * -0.21% -0.71 0.76 18 -0.01% 0.21 0.42 0.18% 0.27 0.39 19 -0.18% -1.10 0.86 -0.05% 0.46 0.32 20 -0.04% -0.86 0.81 0.15% 0.73 0.23 This table shows the daily average abnormal returns to acquirer and target companies in mergers and acquisitions in the automotive supply industry. It contains all public acquirers and targets for which trading data is available between 250 before and 20 days after the transaction. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively. The statistical significance is tested using the cross-sectional test as proposed by Boehmer et al. (1991) (z-Test).

Appendix

185

Appendix 4: Value-weighted Abnormal Performance Changes Performance Change Delta of Acquirers and Rivals over 3 years after the transaction - value-weighted (n=93) Performance Mean Median Indicator (PI)

CF/BVA CF/Sales EBIT/BVA EBIT/Sales EBITDA/BVA EBITDA/Sales

-16.32% -14.06% 0.48% 2.77% -3.57% -0.84%

-1.22% -3.00% -0.49% -3.75% -1.37% -2.17%

t-Test

t -1.53 -1.55 -0.07 -0.42 -0.58 -0.14

p 0.06 * 0.06 * 0.47 0.34 0.28 0.44

This table shows the abnormal change in different performance indicators of acquiring companies over the 3 years following transactions in the automotive supply industry. Means across transactions are derived as value-weighted averages. The difference is derived under a control-firm matching approach as proposed by Lyon et al. (1999). All changes relate to a 3-year period starting in the year of the merger. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and has been tested using a standard t-Test as well as a Wilcoxon Signed-Rank Test.

186

Appendix

Appendix 5: CAARs to Acquirers Applied in the Accounting Study CAARs to Acquirers (n=93) EventWindow

t-Test CAAR

t-value

p-value

[-20,20] [-5,5] [-1,1] [-1,0] [0] [0,1]

2.10% 2.73% 1.71% 0.96% 0.82% 1.56%

1.47 3.26 3.39 2.36 2.34 3.33

0.07 * <0.01 *** <0.01 *** <0.01 *** <0.01 *** <0.01 ***

This table shows the cumulative average abnormal returns (CAAR) to acquiring companies in mergers and acquisitions in the automotive supply industry. It contains all 93 transactions which are applied in the accounting study. Statistical significance is tested using a standard t-Test and denoted by *, **, and *** at the 10%, 5%, and 1%-level respectively.

Appendix

187

Appendix 6: Regression of Abnormal Performance Changes on 11-day CAARs Regression Results Dependent Variable Operating CF/BVA

a

t-value

b

-12.53%

-1.84 *

0.203

t-value 0.25

Statistical Quality F = 0.06 R2adj. = -0.01 DWS = 1.66

Operating CF/Sales

-13.36%

-2.41 **

1.043

1.59

F = 2.54 R2adj. = 0.02 DWS = 1.99

EBIT/BVA

-7.25%

-1.15

0.860

1.34

F = 1.34 R2adj. = 0.00 DWS = 1.98

EBIT/Sales

-8.55%

-1.78 *

1.824

3.22 ***

F = 10.36 *** R2adj. = 0.09 DWS = 1.93

EBITDA/BVA

-5.73%

-1.07

-0.044

-0.07

F = 0.00 R2adj. = -0.01 DWS = 1.82

EBITDA/Sales

-6.75%

-2.00 **

1.101

2.77 ***

F = 7.68 *** R2adj. = 0.07

DWS = 1.98 This table shows the results of six models regressing the abnormal peer-group adjusted performance growth of firm i on the short-term announcement returns (CAARs) over the 11 days surrounding the announcement date. The slope coefficient b captures the correlation between the two approaches. Statistical significance at the 10%, 5%, and 1%-level is denoted by *, **, and *** respectively, and is tested using a standard t-Test for individual coefficients and the F-Test for the overall model respectively.

188

Appendix

Appendix 7: Annual Balance Sheet/Income Statement Items of Continental AG Balance Sheet and Income Statement Items (in EUR billion) 2003

2004

2005

2006

2007

CAGR

CAGR

(US-GAAP)

(IFRS)

(IFRS)

(IFRS)

(IFRS)

03-06

03-07

Total Assets

8.30

9.68

10.55

10.85

Equity

1.98

2.84

3.80

n/a

3.35

2.79

Sales

11.53

12.60

Cost of Goods Sold

-8.82

R&D Expenses

-0.50

Selling, Admin. & Other Expenses

Non-current liabilities

EBITDA

27.74

9.3%

35.2%

4.71

6.86

33.5%

36.4%

2.16

11.67

n/a

n/a

13.84

14.89

16.62

8.9%

9.6%

-9.45

-10.33

-11.21

-12.60

8.3%

9.3%

-0.52

-0.59

-0.68

-0.83

10.8%

13.5%

-1.38

-1.48

-1.43

-1.43

-1.54

1.2%

2.8%

1.46

1.83

2.25

2.30

2.49

16.4%

14.3%

Depreciation/ Amortization

0.60

0.67

0.74

0.70

0.81

5.3%

7.8%

EBIT

0.86

1.16

1.51

1.60

1.68

23.0%

18.2%

Net Income

0.31

0.74

0.95

1.00

1.05

47.8%

35.7%

Operating Cash Flow

1.23

1.53

1.54

0.82

1.91

-12.6%

11.6%

Free Cash Flow

0.61

0.68

0.66

-0.64

-10.63

< -100%

< -100%

68,829 80,586 79,849 85,224 151,654

7.4%

21.8%

# Employees (end of year) Source: Annual Reports

References Abrenica, J. (1998). The Asian Automotive Industry: Assessing the Roles of State and Market in the Age of Global Competition. Asian Pacific Economic Literature, 12(1), 12-26. Agrawal, A., & Jaffe, J. F. (2000). The Post-Merger Performance Puzzle. In G. Cooper & A. Gregory (Eds.), Advances in Mergers and Acquisitions (pp. 7-41). Stamford, CT: JAI Press. Agrawal, A., Jaffe, J. F., & Mandelker, G. N. (1992). The Post-Merger Performance of Acquiring Firms: A Re-examination of an Anomaly. Journal of Finance, 47, 1605-1621. Akdo÷u, E. (2009). Gaining a Competitive Edge Through Acquisitions: Evidence from the Telecommunications Industry. Journal of Corporate Finance, 15, 99-112. Aktas, N., de Bodt, E., & Derbaix, A. (2004). Horizontal, Downstream and Upstream Effects of Merger & Acquisition Operations in the Car Industry. Banque et Marché, Mars-Avril, 40-49. Andrade, G., Mitchell, M., & Stafford, E. (2001). New Evidence and Perspectives on Mergers. Journal of Economic Perspectives, 15(2), 103-120. Andrade, G., & Stafford, E. (2004). Investigating the Economic Role of Mergers. Journal of Corporate Finance, 10, 1-36. André, P., Kooli, M., & L'Her, J.-F. (2004). The Long-Run Performance of Mergers and Acquisitions: Evidence from the Canadian Stock Market. Financial Management, 33(4), 27-43. Antoniou, A., & Zhao, H. (2004). Long-run Post Takeover Stock Return: The Impact of Overlapping Return, Takeover Premium, and Method of Payment. Working Paper (January), Centre for Empirical Research in Finance, University of Durham, Durham. Automotive News (2003). Top 100 OEM Parts Suppliers – Ranked on 2002 global OEM automotive parts sales. Retrieved October 23, 2003, from http://www.autonews.com

190

References

Aw, M. S. B., & Chatterjee, R. A. (2004). The Performance of UK Firms Acquiring Large Cross-border and Domestic Takeover Targets. Applied Financial Economics, 14, 337-349. Baradwaj, B. G., Dubofsky, D. A., & Fraser, D. R. (1992). Bidder Returns in Interstate and Intrastate Bank Acquisitions. Journal of Financial Services Research, 5, 261-273. Barber, B. M., & Lyon, J. D. (1996). Detecting Abnormal Operating Performance: The Empirical Power and Specification of Test Statistics. Journal of Financial Economics, 41, 359-399. Barber, B. M., & Lyon, J. D. (1997). Detecting Long-run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics. Journal of Financial Economics, 43, 341-372. Barnes, P. (1984). The Effect of a Merger on the Share Price of the Attacker, Revisited. Accounting and Business Research, 15, 45-49. Becker-Blease, J. R., Goldberg, L. G., & Kaen, F. R. (2008). Mergers and Acquisitions as a Response to the Deregulation of the Electric Power Industry: Value Creation or Value Destruction? Journal of Regulatory Economics, 33, 21-53. Beitel, P., Schiereck, D., & Wahrenburg, M. (2004). Explaining M&A Success in European Banks. European Financial Management, 10, 109-139. Berry, S. K. (2000). Excess Returns in Electric Utility Mergers during Transition to Competition. Journal of Regulatory Economics, 18, 175-188. Betzer, A., & Metzger, D. (2007). Event Studies and Accounting Studies in Merger Analysis – Are They Really Substitutes? Working Paper (January 3), University of Bonn, Bonn, Germany. Black, F. (1972). Capital Market Equilibrium with Restricted Borrowing. Journal of Business, 45, 444-454. Black, E., Carnes, T., & Jandik, T. (2001). The Long-term Success of Cross-border Mergers and Acquisitions. Working Paper (May), University of Arkansas, Fayetteville, AR. Bley, J., & Madura, J. (2003). Intra-Industry and Inter-Country Effects of European Mergers. Journal of Economics and Finance, 27, 373-395.

References

191

Boehmer, E., Musumeci, J., & Poulsen, A. (1991). Event-Study Methodology under Conditions of Event-induced Variance. Journal of Financial Economics, 30, 253-272. Bouwman, C., Fuller, K., & Nain, A. (2009). Market Valuation and Acquisition Quality: Empirical Evidence. Review of Financial Studies, 22, 633-679. Bradley, M., Desai, A., & Kim, E. H. (1988). Synergistic Gains from Corporate Acquisitions and their Division between the Stockholders of Target and Acquiring Firms. Journal of Financial Economics, 21, 3-40. Bradley, M., & Jarrell, G. A. (1988). Comment. In J. C. Coffee, L. Lowenstein, & S. Rose-Ackerman (Eds.), Knights, Raiders and Targets: The Impact of the Hostile Takeover (pp. 253-259). New York, NY: Oxford University Press. Brown, S., & Warner, J. B. (1980). Measuring Security Price Performance. Journal of Financial Economics, 8, 205-258. Brown, S., & Warner, J. B. (1985). Using Daily Stock Returns: The Case of Event Studies. Journal of Financial Economics, 14, 3-31. Bruner, R. F. (1999). An Analysis of Value Destruction and Recovery in the Alliance and Proposed Merger of Volvo and Renault. Journal of Financial Economics, 51, 125–166. Bruner, R. F. (2002). Does M&A Pay? A Survey of Evidence for the Decision-Maker. Journal of Applied Finance, 12(1), 48-68. Calomiris, C., & Karceski, J. (1998). Is the Bank Merger Wave of the 1990s Efficient? Lessons Learned from Nine Case Studies. Washington, DC: American Enterprise Institute Press. Carr, C. (1993). Global, National, and Resource-Based Strategies: An Examination of Strategic Choice and Performance in the Vehicle Components Industry. Strategic Management Journal, 14, 551-567. Cheng, L. T. W., & Leung, T. Y. (2004). A Comparative Analysis of the MarketBased and Accounting-Based Performance of Diversifying and NonDiversifying Acquisitions in Hong Kong. International Business Review, 13, 763-789. Comment, R., & Jarrell, G. A. (1995). Corporate Focus and Stock Returns. Journal of Financial Economics, 37, 67-87.

192

References

Conn, R., Cosh, A., Guest, P. M., & Hughes, A. (2005). The Impact on UK Acquirers of Domestic, Cross-border, Public and Private Acquisitions. Journal of Business Finance & Accounting, 32, 815-870. Continental (2006). Annual Report 2006. Hanover: Author. Continental (2007a). Continental AG and Siemens VDO Automotive AG join together to form automotive supplier at forefront of global market. Press Release (July 25), Continental AG, Hanover. Continental (2007b). Annual Report 2007. Hanover: Author. Continental (2007c). Continental: 136 Years of Safe Mobility. Retrieved August 1, 2008, from Continental Web site: http://www.conti-online.com/genertor/www/ com/en/continental/portal/themes/press_services/acq/download/day1_continenta l_history_en.doc Continental (2007d). History of Siemens VDO Automotive AG. Retrieved August 01, 2008, from Continental Web site: http://www.conti-online.com/generator/www/ com/en/continental/portal/themes/press_services/acq/download/day1_vdo_histor y_en.doc Continental (2007e). "winning the future – together." Corporate Presentation, Continental AG, Hanover. Presented at the Press Conference on July 25. Copeland, T. E., & Weston, J. F. (1988). Financial Theory and Corporate Policy. Reading, MA: Addison-Wesley. Cornett, M. M., & Tehranian, H. (1992). Changes in Corporate Performance Associated with Bank Acquisitions. Journal of Financial Economics, 31, 211-234. Coutts, J. A., Mills, T. C., & Roberts, J. (1994). The Market Model and the Event Study Method: A Synthesis of the Econometric Criticisms. International Review of Financial Analysis, 3, 149-171. Cowan, A., Nandkumar, N., & Singh, A. (1990). Stock Returns before and after Calls of Convertible Bonds. Journal of Financial and Quantitative Analysis, 25, 549555. Damodaran, A. (2004). Valuing Acquisitions and Takeovers. In A. Damodaran (Ed.), Corporate Finance: Theory and Practice (2nd ed., pp. 834-878). New York, NY: John Wiley & Sons Inc.

References

193

Dewenter, K. (1995). Does the Market React Differently to Domestic and Foreign Takeover Announcements? - Evidence from the U.S. Chemical and Retail Industries. Journal of Financial Economics, 37, 85-101. Dickerson, A., Gibson, H., & Tsakalotos, E. (1997). The Impact of Acquisitions on Company Performance: Evidence from a Large Panel of UK Firms. Oxford Economic Papers, 49, 344-361. DowJonesNewswires (2007a). Continental AG's CFO Upbeat on Synergies From VDO Deal. October 31. DowJonesNewswires (2007b). Continental AG Close To Takeover in China - Newspaper. August 24. DowJonesNewswires (2007c). Continental AG Would Accept Minority Stake in VDO Deal – CEO. February 22. DowJonesNewswires (2008). Continental AG: Synergies Through VDO Up To EUR 350M in 2010. February 21. Durand, R., & Vargas, V. (2003). Ownership, Organization, and Private Firms’ Efficient Use of Resources. Strategic Management Journal, 24, 667-675. Eberhart, A., Maxwell, W., & Siddique, A. (2004). An Examination of Long-term Abnormal Stock Returns and Operating Performance Following R&D Increases. Journal of Finance, 59, 623-649. Eckbo, B. E. (1992). Mergers and the Value of Antitrust Deterrence. Journal of Finance, 47, 1005-1029. Eisenhardt, K. (1989). Building Theories from Case Study Research. Academy of Management Review, 14, 532-550. Fama, E., & French, K. (1993). Common Risk Factors in the Returns on Stocks and Bonds. Journal of Financial Economics, 33, 3-56. Ferris, S. P., & Park, K. (2002). How Different is the Long-Run Performance of Mergers in the Telecommunications Industry? Advances in Financial Economics, 7, 127-144. Fitzgerald, C. (2002). Are Supplier Profits Right Around the Corner? Automotive Industries, 182, 26-28.

194

References

Franks, J., & Harris, R. (1989). Shareholder Wealth Effects of Corporate Takeovers: The U.K. Experience 1955-1985. Journal of Financial Economics, 23, 225-249. Franks, J., Harris, R., & Titman, S. (1991). The Postmerger Share-Price Performance of Acquiring Firms. Journal of Financial Economics, 29, 81-96. Fridolfsson, S.-O., & Stennek, J. (2005). Why Mergers Reduce Profits and Raise Share Prices? A Theory of Preemptive Mergers. Journal of the European Economic Association, 3, 1083-1104. Fujimoto, T. (2001). The Japanese Automobile Parts Supplier System: the Triplet of Effective Inter-firm Routines. International Journal of Automotive Technology and Management, 1, 1-34. Fuller, K., Netter, J., & Stegemoller, M. (2002). What Do Returns to Acquiring Firms Tell Us? Evidence from Firms that Make Many Acquisitions. Journal of Finance, 57, 1763-1793. Goergen, M., & Renneboog, L. (2004). Shareholder Wealth Effects of European Domestic and Cross-Border Takeover Bids. European Financial Management, 10, 9-45. Gregory, A. (1997). An Examination of the Long Run Performance of UK Acquiring Firms. Journal of Business Finance & Accounting, 24, 971-1002. Gregory, A., & Matatko, J. (2004). Biases in Estimating Long Run Abnormal Returns and Conditional Measures of Performance: The Evidence on Takeovers Reexamined. Working Paper (April), School of Business & Economics, University of Exeter, Exeter. Haleblian, J., & Finkelstein, S. (1999). The Influence of Organizational Acquisition Experience on Acquisition Performance: A Behavioral Learning Perspective. Administrative Science Quarterly, 44, 29-56. Handelsblatt (2007). German Conti Accelerates VDO Integration, December 5. Healy, P. M., Palepu, K. G., & Ruback, R. S. (1992). Does Corporate Performance Improve After Mergers? Journal of Financial Economics, 31, 135-175. Jensen, M. C., & Ruback, R. S. (1983). The Market for Corporate Control – the Scientific Evidence. Journal of Financial Economics, 11, 5–50.

References

195

Kahle, K. M., & Walkling, R. A. (1996). The Impact of Industry Classifications on Financial Research. Journal of Financial and Quantitative Analysis, 31, 309335. Kamath, R. R., & Liker, J. K. (1994). A Second Look at Japanese Product Development. Harvard Business Review, 72(6), 154-170. Kang, J.-K. (1993). The International Market for Corporate Control – Mergers and Acquisitions of U.S. Firms by Japanese Firms. Journal of Financial Economics, 34, 345-371. Kaplan, S. (1989). Campeau’s Acquisition of Federated: Value Destruction or Value Added. Journal of Financial Economics, 25, 191–212. Kaplan, S., Mitchell, M. L., & Wruck, K. (1997). A Clinical Exploration of Value Creation and Destruction in Acquisitions. Working Paper (July), Graduate School of Business, University of Chicago, Chicago, OH. Kennedy, V., & Limmack, R. (1996). Takeover Activity, CEO Turnover, and the Market for Corporate Control. Journal of Business Finance & Accounting, 23, 267-285. Kiymaz, H., & Mukherjee, T. (2000). The Impact of Country Diversification on Wealth Effects in Cross-Border Mergers. The Financial Review, 35(2), 37-58. Laabs, J.-P., & Schiereck, D. (2009). The Long-term Success of M&A in the Automotive Supply Industry: Determinants of Capital Market Performance. Journal of Economics and Finance, 33, forthcoming. Liu, H.-H., Chen, T.-Y., & Pai, L-Y. (2007). The Influence of Merger and Acquisition Activities on Corporate Performance in the Taiwanese Telecommunications Industry. The Service Industries Journal, 27, 1041-1051. Loderer, C., & Martin, K. (1992). Postacquisition Performance of Acquiring Firms. Financial Management, 21(3), 69-79. Loughran, T., & Vijh, A. (1997). Do Long-Term Shareholders Benefit from Corporate Acquisitions? Journal of Finance, 52, 1765-1790. Lyon, J. D., Barber, B. M., & Tsai, C.-L. (1999). Improved Methods for Test of Long-run Abnormal Stock Returns. Journal of Finance, 54, 165-201.

196

References

Lys, T., & Vincent, L. (1995). An Analysis of Value Destruction in AT&T’s Acquisition of NCR. Journal of Financial Economics, 39, 353–378. Malatesta, P. H. (1983). The Wealth Effect of Merger Activity and the Objective Functions of Merging Firms. Journal of Financial Economics, 11, 155-181. Mandelker, G. (1974). Risk and Return: The Case of Merging Firms. Journal of Financial Economics, 1, 303-335. McIvor, R. T., Humphreys, P. K., & McAleer, W. E. (1998). European Car Makers and Their Suppliers: Changes at the Interface. European Business Review, 98, 87-99. Megginson, W. L., Morgan, A., & Nail, L. (2004). The Determinants of Positive Long-term Performance in Strategic Mergers: Corporate Focus and Cash. Journal of Banking & Finance, 28, 523-552. Mentz, M. (2006). Mergers & Acquisitions in der Automobilzuliefererindustrie – Wertschöpfungspotentiale durch internationale Positionierung. Wiesbaden: Deutscher Universitäts-Verlag. Mentz, M., & Schiereck, D. (2008). Cross-border Mergers and the Cross-border Effect: The Case of the Automotive Supply Industry. Review of Managerial Science, 2, 199-218. Mercer, G. (1995). Modular Supply in the 1990s – the Keys to Success. Europe's Automotive Components Business, Second Quarter, 112-135. Mikkelson, W. & Partch, M. (1988). Withdrawn Security Offerings. Journal of Financial and Quantitative Analysis, 23, 119-133. Mitchell, M. L., & Lehn, K. (1990). Do Bad Bidders Become Good Targets? The Journal of Political Economy, 98, 372-398. Mitchell, M. L., & Stafford, E. (2000). Managerial Decisions and Long-term Stock Price Performance. Journal of Business, 73, 287-329. Morcott, S. (2000). Today's Automotive Supply Industry – Evolution or Revolution? Retrieved December 21, 2004, from: http://www.mapi.net/html/abstract.cfm?PubID =88 Mulherin, J., & Boone, A. L. (2000). Comparing Acquisitions and Divestitures. Journal of Corporate Finance, 6, 117-139.

References

197

Piloff, S. J., & Santomero, A. M. (1998). The Value Effects of Bank Mergers and Acquisitions. In Y. Amihud & G. Miller (Eds.), Bank Mergers & Acquisitions (pp. 59-78). New York, NY: Kluwer Academic Publishers. Ramaswamy, K. P., & Waegelein, J. F. (2003). Firm Financial Performance Following Mergers. Review of Quantitative Finance & Accounting, 20, 115-126. Rau, P., & Vermaelen, T. (1998). Glamour, Value, and the Post-Acquisition Performance of Acquiring Firms. Journal of Financial Economics, 49, 223-253. Ravenscraft, D., & Scherer, F. M. (1989). The Profitability of Mergers. International Journal of Industrial Organization, 7, 101-116. Reuters (2007a). Continental AG joins industry elite with VDO buy. July 25. Reuters (2007b). Continental AG says interested in Siemens VDO unit. January 25. Rhoades, S. (1994). A Summary of Merger Performance Studies in Banking, 19801993, and an Assessment of the "Operating Performance" and "Event Study" Methodologies. Working Paper (July), Board of Governors of the Federal Reserve System, Washington, DC. Ritter, J. (1991). The Long-run Performance of Initial Public Offerings. Journal of Finance, 46, 197-216. Ruback, R. (1982). The Conoco Takeover and Stockholder Returns. Sloan Management Review, 23, 13–33. Sadler, D. (1999). Internationalization and Specialization in the European Automotive Components Sector: Implications for the Hollowing-out Thesis. Regional Studies, 33, 109-119. Siemens (2006). Annual Report 2006. Munich: Author. Siemens (2007). Annual Report 2007. Munich: Author. Siems, T. F. (1996). Bank Mergers and Shareholder Wealth: Evidence from 1995's Megamerger Deals. Financial Industry Studies of the Federal Reserve Bank of Dallas, August, 1-12. Sinha, M. (2004). The Long-run Performance of Mergers and Acquisitions: Risk or Mispricing. Working Paper (November), Cornell University, Ithaca, NY.

198

References

Switzer, J. A. (1996). Evidence on Real Gains in Corporate Acquisitions. Journal of Economics and Business, 48, 443-460. Von Corswant, F., & Fredriksson, P. (2002). Sourcing Trends in the Car Industry: A Survey of Car Manufacturers' and Suppliers' Strategies and Relations. International Journal of Operations and Production Management, 2, 741-758. Von Corswant, F., Wynstra, J. Y. F., & Wetzels, M. (2003). In Chains? Automotive Suppliers and Their Product Development Activities. Research in Management, ERIM Report Series (April), Erasmus University, Rotterdam. Retrieved June 16, 2008, from SSRN Web site: http://ssrn.com/abstract=423652 Weston, J. F., Chung, K. S., & Siu, J. A. (1990). Takeovers, Restructuring, and Corporate Governance. Upper Saddle River, NJ: Prentice Hall. Wortham, A. (2007). Continental: No plans to sell parts of Siemens VDO. Automotive News, 82(6275), 44.

The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry

Mergers and Acquisitions in the Machinery Industry

Mergers and acquisitions basics

Mergers and acquisitions

Mergers and Acquisitions

Acquisitions and Mergers

Advances in Mergers and Acquisitions, Volume 4 (Advances in Mergers and Acquisitions) (Advances in Mergers and Acquisitions)

Mergers and Acquisitions

Advances in Mergers and Acquisitions, Volume 4 (Advances in Mergers and Acquisitions) (Advances in Mergers and Acquisitions)

The Management of Mergers and Acquisitions

Mergers & Acquisitions

Mergers, acquisitions, and corporate restructurings

Copper in the Automotive Industry

Advances in Mergers and Acquisitions, Volume 5

Mergers & Acquisitions

Plastics in the Automotive Industry

Mergers and Acquisitions (Express Exec)

Mergers, Acquisitions and Corporate Restructuring

German Mergers & Acquisitions in the USA

Advances in Mergers and Acquisitions, Volume 7 (Studies in Mergers and Acquisitions)

Advances in Mergers and Acquisitions, Volume 3 (Advances in Mergers and Acquisitions)

Intellectual Property Assets in Mergers and Acquisitions

The Social Side of Mergers and Acquisitions: Cooperation relationships after mergers and acquisitions

Mergers & Acquisitions

Advances in Mergers and Acquisitions, Volume 2 (Advances in Mergers and Acquisitions)

Value Creation in Mergers, Acquisitions, and Alliances

Mergers, Acquisitions and Corporate Restructuring

Mergers & Acquisitions

Advances in Mergers and Acquisitions, Volume 6 (Advances in Mergers and Acquisitions)

Advances in Mergers and Acquisitions, Volume 8

Intellectual Property Assets in Mergers and Acquisitions

The Long-Term Success of Mergers and Acquisitions in the International Automotive Supply Industry

Mergers and Acquisitions in the Machinery Industry

Mergers and acquisitions basics

Mergers and acquisitions

Mergers and Acquisitions

Acquisitions and Mergers

Advances in Mergers and Acquisitions, Volume 4 (Advances in Mergers and Acquisitions) (Advances in Mergers and Acquisitions)

Mergers and Acquisitions

Advances in Mergers and Acquisitions, Volume 4 (Advances in Mergers and Acquisitions) (Advances in Mergers and Acquisitions)

The Management of Mergers and Acquisitions

Mergers & Acquisitions

Mergers, acquisitions, and corporate restructurings

Copper in the Automotive Industry

Advances in Mergers and Acquisitions, Volume 5

Mergers & Acquisitions

Plastics in the Automotive Industry

Mergers and Acquisitions (Express Exec)

Mergers, Acquisitions and Corporate Restructuring

German Mergers & Acquisitions in the USA

Advances in Mergers and Acquisitions, Volume 7 (Studies in Mergers and Acquisitions)

Advances in Mergers and Acquisitions, Volume 3 (Advances in Mergers and Acquisitions)

Intellectual Property Assets in Mergers and Acquisitions

The Social Side of Mergers and Acquisitions: Cooperation relationships after mergers and acquisitions

Mergers & Acquisitions

Advances in Mergers and Acquisitions, Volume 2 (Advances in Mergers and Acquisitions)

Value Creation in Mergers, Acquisitions, and Alliances

Mergers, Acquisitions and Corporate Restructuring

Mergers & Acquisitions

Advances in Mergers and Acquisitions, Volume 6 (Advances in Mergers and Acquisitions)

Advances in Mergers and Acquisitions, Volume 8

Intellectual Property Assets in Mergers and Acquisitions

Recommend Documents