Docstoc

Share Repurchases in Germany an Empirical Investigation

Document Sample
Share Repurchases in Germany an Empirical Investigation Powered By Docstoc
					SHARE REPURCHASES IN GERMANY – AN EMPIRICAL
              INVESTIGATION



                    Diploma Thesis


                       Submitted to

                  Professor Dr. Schmidt
                         Chair for
            International Banking and Finance,
              Faculty of Applied Economics
           Johann Wolfgang Goethe-University
                      Frankfurt/Main




                              by
           cand. rer. oec. Alexandre Zdantchouk
                  Ginnheimer Landstr. 42
                 60487 Frankfurt am Main
                   Phone: 0175-58-48-903
          e-mail: alexandre.zdantchouk@lazard.com
              Business Administration (BWL)
              Matriculation number: 1774490




                                                    1
TABLE OF CONTENTS


TABLE OF CONTENTS................................................................................................. 1
INDEX OF FIGURES......................................................................................................4
INDEX OF ABBREVIATIONS ......................................................................................5
1. INTRODUCTION .......................................................................................................6
   1.1. MOTIVATION........................................................................................................ 6
   1.2. OUTLINE............................................................................................................... 7
   1.3. BUY-BACK DEFINITIONS .................................................................................... 8
   1.4. MAIN RESULTS OF THE EMPIRICAL STUDY ...................................................... 8
2. REVIEW OF PRIOR RESEARCH ON SHARE BUY-BACKS ................................ 10
   2.1. SIGNALLING/UNDERVALUATION HYPOTHESIS ............................................ 11
   2.2. TAKEOVER DEFENCE HYPOTHESIS ................................................................ 16
   2.3. OTHER HYPOTHESES ........................................................................................ 17
      2.3.1. Free Cash Flow hypothesis ..................................................................................................... 17
      2.3.2. Taxation Hypothesis .............................................................................................................. 18
      2.3.3. Leverage Hypothesis ............................................................................................................... 19
      2.3.4. Minority Shareholder/Bondholder Expropriation Hypothesis ................................................. 19
   2.4. IMPLICATIONS FOR OWN RESEARCH AND CONCLUDING REMARKS........... 19
3. SHARE BUY-BACKS – SITUATION IN GERMANY ............................................. 21
   3.1. SHARE BUY-BACK IN GERMANY – A HISTORICAL DEVELOPMENT............. 21
   3.2. CURRENT SITUATION IN GERMANY – A LEGAL PERSPECTIVE .................... 21
   3.3. ANALYSIS OF THE BUY-BACK ACTIVITY IN GERMANY ................................. 23
   3.4. RATIONALE FOR SHARE BUY-BACKS IN GERMANY ...................................... 28
   3.5. FURTHER SPECIFICS OF SHARE BUY-BACKS IN GERMANY........................... 30
      3.5.1. Buy-Back Methods................................................................................................................. 30
      3.5.2. Completion Rates ................................................................................................................... 30
      3.5.3. Pricing of Share Buy-Backs .................................................................................................... 30
   3.6. CONCLUDING REMARKS ................................................................................... 31
4. INVESTIGATION METHODOLOGY .................................................................... 32
   4.1. EXCURSE: EVENT STUDY MODELS – BACKGROUND THEORY .................... 32
   4.2. MARKET MODEL AND CALCULATION METHODOLOGY............................... 34
   4.3. SIGNIFICANCE TESTS ........................................................................................ 35
   4.4. SAMPLE DESIGN ................................................................................................ 36
   4.5. INVESTIGATION STRUCTURE............................................................................ 36
5. EMPIRICAL INVESTIGATION OF BUY-BACK ANNOUNCEMENTS ............. 38
   5.1. ANALYSIS OF ABNORMAL RETURNS FOR THE COMPLETE SAMPLE OF SHARE
        BUY-BACKS ........................................................................................................ 38
   5.2. DETAILED INVESTIGATION – UNIVARIATE ANALYSES ................................ 40
                                                                                                                                                  2
      5.2.1. Analysis of Abnormal Returns by Motives as Stated by Companies ........................................ 41
      5.2.2. Analysis of Abnormal Returns by Market-to-Book Value Ratio ........................................... 43
      5.2.3. Analysis of Abnormal Returns by Company Size ................................................................... 45
      5.2.4. Analysis of Abnormal Returns by Industrial Classification of Share Buy-Backs ..................... 51
      5.2.5. Analysis of Abnormal Returns by Exchange Listing Sector .................................................... 53
      5.2.6. Analysis Abnormal Returns by Cumulative Returns before Announcement ............................. 55
      5.2.7. Analysis of Abnormal Returns by the Percentage of Outstanding Shares Repurchased ............. 56
      5.2.8. Analysis of Abnormal Returns by Financial Position of Company .......................................... 57
      5.2.9. Analysis of Abnormal Returns by Shareholder Structures ....................................................... 59
   5.3. DETAILED INVESTIGATION – A CROSS-SECTIONAL MULTIVARIATE
        ANALYSIS ........................................................................................................... 64
   5.4. EXCURSIVE: ANALYSIS OF INDUSTRY-WIDE INFORMATION EFFECTS OF
        SHARE BUY-BACK ANNOUNCEMENTS ........................................................... 70
   5.5. CONCLUDING REMARKS ................................................................................... 72
6. EMPIRICAL INVESTIGATION OF BUY-BACK PLAN ANNOUNCEMENTS
AND AGM AUTHORISATIONS OF BUY-BACKS ..................................................... 73
   6.1. ANALYSIS OF PRICE EFFECTS OF BUY-BACK PLAN ANNOUNCEMENTS ..... 73
   6.2. ANALYSIS OF PRICE EFFECTS OF AGM AUTHORISATIONS .......................... 75
   6.3. CONCLUDING REMARKS ................................................................................... 82
7. RESULTS AND CONCLUSIONS............................................................................. 84
8. APPENDICES ............................................................................................................ 88
   APPENDIX I: COMPARISON OF DIFFERENT EVENT STUDY METHODOLOGIES 88
   APPENDIX II: SHARE-BUY BACKS IN GERMANY – THE MAIN SAMPLE.............. 93
   APPENDIX III: DECILE ANALYSIS OF ANNOUNCEMENTS RETURNS BY
      COMPANY SIZE.................................................................................................. 94
REFERENCES............................................................................................................... 96
WORD-OF-HONOUR-STATEMENT (IN GERMAN) ............................................ 100




                                                                                                                              3
INDEX OF FIGURES

FIGURE 1: TOP 10 SHARE BUY-BACKS IN GERMANY ........................................................................ 24
FIGURE 2: ANALYSIS OF REPURCHASING ACTIVITY BY INDUSTRY ................................................ 25
FIGURE 3: DEVELOPMENT OF REPURCHASING ACTIVITY IN GERMANY ...................................... 26
FIGURE 4: RATIONALE FOR SHARE BUY-BACKS IN GERMANY ........................................................ 29
FIGURE 5: TIME SEQUENCE OF A SHARE BUY-BACK PROCESS IN GERMANY ................................ 37
FIGURE 6: ANALYSIS OF ABNORMAL RETURNS FOR THE COMPLETE SAMPLE OF SHARE BUY-
    BACKS ............................................................................................................................................. 38
FIGURE 7: ANALYSIS OF ABNORMAL RETURNS BY MOTIVES AS STATED BY COMPANIES ............. 42
FIGURE 8: ANALYSIS OF ABNORMAL RETURNS BY MTB RATIO ...................................................... 44
FIGURE 9: REGRESSION ANALYSIS OF ABNORMAL RETURNS BY MTB RATIO ............................... 45
FIGURE 10: DISTRIBUTION OF THE COMPANY SAMPLE BY SIZE...................................................... 46
FIGURE 11: ANALYSIS OF ABNORMAL RETURNS BY COMPANY SIZE ............................................... 47
FIGURE 12: DECILE SUB-SAMPLES ....................................................................................................... 48
FIGURE 13: DECILE ANALYSIS OF ANNOUNCEMENT RETURNS BY COMPANY SIZE ...................... 49
FIGURE 14: DISTRIBUTION OF THE TOTAL SAMPLE ACCORDING TO INDUSTRIAL
    CLASSIFICATION OF COMPANIES................................................................................................. 51
FIGURE 15: ANALYSIS OF ABNORMAL RETURNS BY INDUSTRIAL CLASSIFICATION OF SHARE
    BUYBACKS ...................................................................................................................................... 52
FIGURE 16: ANALYSIS OF ABNORMAL RETURNS BY EXCHANGE LISTING SECTOR........................ 54
FIGURE 17: REGRESSION ANALYSIS OF ABNORMAL RETURNS PRIOR TO ANNOUNCEMENTS OF
    SHARE BUY-BACKS ........................................................................................................................ 56
FIGURE 18: ANALYSIS OF ABNORMAL RETURNS BY FINANCIAL POSITION OF COMPANIES ...... 58
FIGURE 19: ANALYSIS OF ABNORMAL RETURNS BY MANAGEMENT HOLDINGS ........................... 60
FIGURE 20: ANALYSIS OF ABNORMAL RETURNS BY SHAREHOLDER STRUCTURES ........................ 62
FIGURE 21: FIRST-STEP-ANALYSIS OF VARIABLES ............................................................................. 67
FIGURE 22: MULTIVARIATE REGRESSION ANALYSIS......................................................................... 68
FIGURE 23: ANALYSIS OF INDUSTRY-WIDE INFORMATION EFFECTS OF SHARE BUY-BACK
    ANNOUNCEMENTS ........................................................................................................................ 71
FIGURE 24: ANALYSIS OF BUY-BACK PLAN ANNOUNCEMENTS....................................................... 74
FIGURE 25: ANALYSIS OF PRICE EFFECTS OF AGM AUTHORISATIONS ......................................... 77
FIGURE 26: ANALYSIS OF NEGATIVE PRICE REACTIONS AFTER AGM APPROVALS ..................... 79
FIGURE 27: ANALYSIS OF NON-REPURCHASING COMPANIES .......................................................... 81
FIGURE 28: RELATIVE DEVELOPMENT OF CDAX INDUSTRIAL INDEXES .................................... 89
FIGURE 29: COMPARISON OF EVENT STUDY METHODOLOGY ........................................................ 90
FIGURE 30: COMPARATIVE GAINS FROM USING A MORE COMPREHENSIVE MODEL .................... 92
FIGURE 31: DECILE ANALYSIS OF ANNOUNCEMENT RETURNS BY THE COMPANY SIZE.............. 94




                                                                                                                                                        4
Index of Abbreviations



INDEX OF ABBREVIATIONS

AG            –      Aktiengesellschaft (German)
AGM           –      Annual General Meeting
AktG          –      Aktiengesetz (German)
AR            –      Abnormal Returns
BAW           –      Bundesaufsichtsamt für den Wertpapierhandel
CAR           –      Cumulated Abnormal Return(s)
CDAX          –      Composite DAX
CFO           –      Chief Financial Officer
CMR           –      Constant Mean Return
DAI           –      Deutsches Aktieninstitut (German)
DC            –      Dispersed Control (in this context, the shareholders ability to control
                     the company)
EV            –      Enterprise Value
FCF           –      Free Cash Flow
DJ            –      Dow Jones
HC            –      High Control (in this context, the shareholders ability to control the
                     company)
IT            –      Information Technology
KonTraG       –      Gesetz zur Kontrolle und Transparenz im
                     Unternehmensbereich (German) (English: Control and Transparency
                     in Business Act)
NM            –      Neuer Markt (German)
NPV           –      Net Present Value
M&A           –      Mergers & Acquisitions
MC            –      Medium Control (in this context, the shareholders ability to control
                     the company)
MTB           –      Market-to- Book Value Ratio
OMR           –      Open Market Repurchase
SEC           –      Securities and Exchange Commission
TPR           –      Transferable Put Rights
Vol.          –      Volume



                                                                                               5
1. Introduction


1. INTRODUCTION

1.1. MOTIVATION
Since the "Control and Transparency in Business Act" (KonTraG) came into effect on 1st May
1998, German firms have been permitted to repurchase their own shares. The German
business community broadly welcomed this liberalisation step1. Until now, more than 160
German corporations have engaged in share buy-backs. The popularity of this financial
instrument is expected to increase even more in the future against the background of a
changing capital market environment and corporate governance cultures in Germany.
However, still different views exist on the usefulness of share buy-backs.

The companies can execute share buy-backs to reduce their cost of capital and more flexibly
manage their capital structures. Share buy-backs can be also implemented to signal
management‟s views with respect to company‟ valuation or cash flow expectations. They may
also be a useful instrument for tax efficient pay-outs to shareholders. Finally, repurchase of
own stock can be used as payment in corporate acquisitions or stock option plans.

However, as many opponents would argue, the buy-back activity may simply be motivated by a
management‟s inability to identify profitable investment opportunities. Share buy-backs can be
mis-used for „entrenchment‟ purposes – by influencing the ownership structure of the firm or
fending off takeover attempts. This could damage the effectiveness of corporate control over
the management and further increase information asymmetries between the company and the
capital market – and, as a result, destroy shareholder value2.

Much empirical research has been done so far on the subject of share buy-backs. The research
in this field has been able to provide insights into how capital markets react to share buy-back
announcements. However, the vast majority of this research has been dedicated to the US
capital market where this subject has been investigated in much detail. Primarily, that is why
the US research has been the benchmark for buy-back research in other countries. But, as we
will see, the subject of buy-backs is largely influenced by the specific factors in particular
country environments.

In this respect, Germany, where share buy-backs were legalised only four years ago, provides a
unique opportunity for any empirical researcher. This, in particular, against the background of
the fact that almost no significant empirical research on German share buy-backs has been

1   See [DAI, 1999], p. 1.
2   See [Pellens/Schremper, 2000], p.2.


                                                                                              6
1. Introduction

done so far3. The questions as to how the German capital market perceives share buy-backs
and what the major explanations for its (abnormal) reaction are, remain broadly unanswered4.

My investigation will attempt to reduce the gap in the empirical research on the subject of
share buy-backs in Germany. As time passes, evidence from Germany would become more
and more valuable as it would provide new insights into the problem, in particular from a
scientific perspective. In addition, the results of the analyses below might have great potential
for practical implementation as they might help repurchasing firms to gauge their possible
gains and avoid potential risks associated with a buy-back program.

1.2. OUTLINE
In my diploma thesis, the primary focus is made on the empirical investigation of share buy-
backs in Germany. Moreover, the analyses done are primarily focused on the testing of the
validity of signalling/undervaluation hypothesis for German repurchasing firms.

Chapter two provides a general review of the major results of prior empirical studies on share
buy-backs. The chapter also presents the major hypotheses which have been advanced as
explanation for abnormal market reaction caused by share buy-back announcements. Chapter
three describes the current situation on share buy-backs in Germany – particularly from
historical and legal perspectives. In addition, also in this chapter, various qualitative analyses of
the share buy-back activity in Germany are provided. Chapter four presents the methodological
framework for the ensuing empirical investigation and provides a general structural outline for
the analyses to follow. Also in this chapter, different event study methodologies are analysed
and the question is answered on what methodology is best applicable for the purposes of this
investigation. In Chapter five, the major results of the empirical investigation of share buy-back
announcements in Germany are presented5. Afterwards, various detailed analyses are
performed, examining the announcement effects of share buy-backs by means of univariate
and multivariate analysis. As a logical addition to this chapter, Chapter six examines the other
two major buy-back events – announcements of the buy-back plans and authorisations by the
Annual General Meetings (AGMs). Finally, Chapter seven summarises the findings of the paper
and outlines some ideas for further buy-back research in Germany.



3 Probably except for some earlier research contributions like [Pellens/Schremper, 2000] and [Pellens/Schremper,
1999].
4 See [Pellens/Schremper, 2000].
5 Here, I would like to draw your attention that there are three major events which are associated with a buy-back

program in Germany. They are (in the time sequence): (1) buy-back plan announcements, (2) AGM approvals (3)
buy-back announcements. All three events are investigated in my paper.

                                                                                                                7
1. Introduction


1.3. BUY-BACK DEFINITIONS
The „share buy-back‟ or „equity repurchase‟ refers to a situation whereby a company
repurchases a proportion of its own outstanding shares in the open market or through a
repurchase tender offer. There are four main methods of repurchasing shares6. Each has
strengths and weaknesses relative to specific corporate finance objectives. Here, a brief review
of each technique follows:7

   Open-market buy-backs – this is the most widely used technique by far. The company buys
    shares in the open-market (on the Stock Exchange) from stockholders. Normally in this
    case, the market participants are not aware that they are selling shares to the company.
    There is no specific expiration date for open-market buy-backs, though their timing is
    limited by the AGM resolution. This type of buy-back may take place over several months
    or even years.

   Fixed-price tender offer – the company offers to buy a specified amount of shares at a given
    price, typically at a premium to the market. A period (usually three-four months upon the
    announcement) is set during which the shares could be sold to the company.

   Dutch-auction tender-offer – this type of buy-back is very similar to a book-building procedure.
    Its difference to the fixed-price tender offer is that the company states a price range within
    which shareholders may tender their shares. The company then selects the price to
    purchase the minimum number of shares it wishes to acquire. All shareholders tendering at
    or below the selected price receive the selected price for their shares.

   Targeted buy-backs – the company acquires its own shares in a block from a particular
    shareholder by direct negotiation. In this case, problems of equal treatment of all
    shareholders may arise.

1.4. MAIN RESULTS OF THE EMPIRICAL STUDY
In my paper, I investigate abnormal share price behaviour associated with German share buy-
backs. The study is done by means of event study methodology (market model). Presumably,
all German buy-back announcements for the time period between May/1998 and May/2002
have been examined.

6 As a separate class of buy-back transactions, the use of buy-backs via transferable put rights (TPR) may be
considered. Under a TPR, the company issues put options to the shareholders in proportion to the number of
shares they own. Each TPR gives the shareholder the right to sell one share at a fixed price during a specified
period. The shareholders who wish to hold their shares can sell their TPRs to the open market. More on that, see
See [Kale et al, 1989].
7 See [Lamba/Ramsay, 2000], p. 4.


                                                                                                              8
1. Introduction

Based on my empirical study, I came to the conclusion that German corporations are mainly
driven by the signalling/undervaluation hypothesis. Management mainly uses the buy-back
instrument to signal positive information about the firm‟s future cash generation prospects.
For the complete sample of buy-back announcements (in total 189), I found evidence which
indicates that there is a highly significant market reaction on the announcement day
(AR[0]=+4.74%) and days around it (e.g. CAR[-1;+1]=+6.00%, CAR[-1;-5]=+6.66%).
Moreover, the announcement event is generally preceded by a one-two month period of
underperformance, which portends to significant timing skills of the repurchasing firms and
also supports the signalling hypothesis.

A more detailed analysis of factors influencing the magnitude of abnormal returns leads to the
conclusion that the firm‟s size, market-to-book ratio, exchange listing sector as well as some
specific factors of the firm‟s governance (in particular, shareholder structure), are the most
important factors influencing the perception of share buy-backs by the German capital market.

I found evidence that the other two buy-back events – i.e. announcements of buy-back plans
and AGMs authorisations – have also significant impacts on the share prices. Moreover, in my
analysis, I detected a new phenomenon which stands in close connection with AGM buy-back
authorisation and equals in significant share price underperformance upon the AGM
authorisation of share buy-back. I have provided an in-depth analysis of this phenomenon,
however, some questions remain unanswered.




                                                                                            9
2. Review of Prior Research on Share Buy-Backs


2. REVIEW OF PRIOR RESEARCH ON SHARE BUY-BACKS

Over the past decades, share buy-backs have gained an ever increasing interest from practical
and scientific perspectives. In its practical implementation, share buy-backs appear to gain an
ever increasing importance as an effective financial instrument of a company. On the other
hand, from a research perspective, the investigation of share buy-backs may provide further
valuable insights for better understanding of capital markets.

The main purpose of this chapter is to review the results of empirical research on share buy-
backs in the US and world-wide. The analysis of this research will provide further implications
for my empirical investigation that follow in the next chapters. Moreover, it will create the
basis for the evaluation of the results of my investigation.

A great deal of prior empirical research has been engaged in the examining of the reasons why
companies buy back their shares. The main rationale for share buy-backs is summarised
below8:

      Information signalling – as documented by the majority of empirical research, a share buy-
       back usually conveys favourable information about the company to the capital market, as
       managers would rather buy back shares when the company is undervalued.

      Takeover defence mechanism – a share buy-back may be used as a defence mechanism. Since a
       share buy-back reduces the liquidity on the market, increases the leverage of the company
       and, as shown empirically, props up the share price, it may significantly reduce the
       possibility of a hostile takeover.

      Cash flow pay-outs – this rationale, which goes back to Jensen (1986), implies that when a
       company generates substantial free cash flows it faces problems of their efficient
       investments on behalf of shareholders because of a lack of profitable investment
       opportunities. Share buy-backs in this case, as well as dividends, are an efficient means for
       returning funds to shareholders.

      Favourable taxation – the favourable tax treatment of share buy-backs in several countries,
       such as e.g. the US and Germany, provides a further explanation for the use of share buy-
       backs in these countries – in addition to or in place of dividends.

      Wealth transfer – the wealth is being transferred from the creditors to the shareholders.
       While increasing the leverage, a share buy-back reduces the assets of the company at the


8   See [Lamba/Ramsay, 2000], p. 3.

                                                                                                 10
2. Review of Prior Research on Share Buy-Backs

    same time and therefore the value of the claims of the creditors. The wealth transfer can
    also occur between the controlling and minority shareholders.

   Adjustment of capital structure – share buy-backs often increase the financial leverage of a
    company. The buy-back instrument allows the company to flexibly structure the
    proportion between the equity and the debt capital. However, in many cases a problem of
    creditor protection may arise. Moreover, share buy-backs might appear an effective means
    of introducing changes in ownership structure of the company9.

   Earnings per share – while reducing equity base, share buy-backs increase earnings per share.
    However, this might appear quite a short term oriented means of communication with the
    capital market („window dressing‟). A buy-back needs financing which, at the same time,
    reduces the asset base and the size of the company and, ceteris paribus, its ability to generate
    earnings in the future.

   Other reasons – for example, using repurchased shares for implementation of employee
    incentive schemes such as employee share schemes and stock option plans.

Based on this rationale, in the empirical research, various hypotheses have been advanced to
explain the stock price behaviour surrounding share buy-back announcements. Among them
are (1) Signalling/undervaluation hypothesis, (2) Take-over defence hypothesis, (3) Free cash
flow hypothesis, (4) Taxation hypothesis, (5) Leverage hypothesis and (6) Expropriation
hypotheses. These hypotheses are most popular ones of those mentioned in the empirical
research on buy-backs. However, by no means, this should be an exhaustive list of all existing
hypotheses. Furthermore, these alternative hypotheses mentioned above are not mutually
exclusive10.

In the following sections, the major results of empirical studies on the main hypotheses will be
presented in more detail.

2.1. SIGNALLING/UNDERVALUATION HYPOTHESIS
A considerable set of buy-back literature has been engaged in the testing of the information
signalling hypothesis (which is also known as the asymmetric information or undervaluation




9  The remaining shareholders may increase their relative share position in the ownership structure of the
company; see [Moreira/Procianoy, 2001], p. 11.
10 See [Vermaelen, 1981], p. 142.


                                                                                                       11
2. Review of Prior Research on Share Buy-Backs

hypothesis). In the empirical research, this hypothesis found the strongest support11, and it
seems to offer the best explanation of the market reaction of share buy-backs so far12.

Research studies demonstrated that (1) buy-backs mainly result in permanent share price
increases of the companies, (2) are followed by their higher profitability in the future and (3)
result in financial analysts revising upwards their earnings forecasts for those companies13.

Management uses the announcement of a share buy-back to signal positive information about
the firm‟s future prospects. Here, the assumption is made that the management (or controlling
shareholders) has advanced information about the true value of the company than other
market participants. Since managers only have incentives to reveal positive information, such
buy-backs are expected to yield positive stock price responses if the signal is credible14.

One of the most prominent studies of signalling hypothesis of share buy-backs was performed
by Vermaelen (1981) for the US buy-backs. He found an average premium offered to
shareholders as part of the tender offer buy-back was +23%. According to Vermaelen, this
reaction (of which he found +13% to be permanent and persistent in the future) could
primarily be attributed to an information signalling effect, indicating an undervaluation of the
company stock prior to the tender offer. Thus, he found evidence that the management
undertakes a buy-back to convince investors that the shares of the company are undervalued.

Furthermore, Vermaelen found that the magnitude of this premium was positively related to
the amount of the shares repurchased by the company, as well as to the shareholding owned
by the management. This evidence was also consistent with the signalling hypothesis, as these
two factors should be positively related to the investors‟ perception of the strength of the
managers‟ conviction that the shares are traded under the fair value15. Their true conviction
was indeed confirmed by abnormal high performance of the companies during the five years
upon the share buy-back.

A wide range of other findings largely supported the signalling hypothesis. Those are, for
example, Dann (1981), Vermaelen (1984), Netter and Mitchell (1989), Bartov (1991),
Comment and Jarrell (1991) etc. According to Dann‟s findings, share buy-backs (tender offers)
led to positive share returns of approx. +15% and which did not return to their pre-buy-back


11 See [Lamba/Ramsay, 2000], p. 7.
12 See [Moreira/Procianoy, 2001], p. 2.
13 See [Lamba/Ramsay, 2000], p. 13.
14 However, this signal must not always indicate an undervaluation of the company. It may be just that the

company has no profitable projects for which to use the funds and therefore undertakes a buy-back as a means of
returning the funds to shareholders. On this, see [Lamba/Ramsay, 2000], p. 8.
15 See [Vermaelen, 1984], p. 163.


                                                                                                            12
2. Review of Prior Research on Share Buy-Backs

levels (i.e. permanent returns). Thus, these results „are consistent with the hypothesis that
tender-offer buy-back announcements constitute a revelation by management of favourable
information about the value of the company‟s future prospects‟16

In the framework of the undervaluation/signalling hypothesis, Ikenberry et al (1995) became
one of the most significant contributions regarding the abnormal price behaviour surrounding
US open market share repurchases (OMRs). Ikenberry et al (1995) examined some 1,239
OMRs between 1980 and 1990. The authors found an average positive abnormal return CAR
[-2;+2] of approx. 3.5% resulting from OMR announcements. Ikenberry et al showed, inter alia,
that the announcement effects were positively correlated with the percentage size of the
repurchase program (higher commitment by the management means a stronger signal sent to
the market) and negatively correlated with the firm size decile (firm size was assumed to be a
proxy for information asymmetries). Moreover, the authors found that the market under-reacts
to OMRs which results in a buy-and-hold abnormal performance of over +12% over 4 years
following the OMRs announcement. Thus, combined with the announcement effect, total
undervaluation amounted to about +15%. In addition, the authors examined the sample using
a market-to-book (MTB) ratios of the repurchasing firms and found a strong negative
correlation between the MTB ratio and the abnormal performance following the buy-back
announcement. The authors suggested that firms with a lower MTB ratios are, generally, more
undervalued.

Considering the signalling/undervaluation hypothesis, the results of the study of Wansley et al
(1989) are particular interesting as the authors carried out a direct evaluation of reasons behind
the share buy-backs in the US. Wansley et al examined some 140 questionnaires of CFO‟s of
US companies which undertook share buy-backs. The authors asked the CFO‟s to comment
upon the possible reasons for why the companies had undertaken share buy-backs. One of the
major significant explanations was that the buy-back was undertaken to convey management
confidence in their company to the market: „Managers agreed with the information signalling
hypothesis of share buy-backs, both as a reason of share buy-backs and as an important
component of buy-back premiums‟17.

Hertzel and Jain (1991) found strong support of the information signalling hypothesis
observing revisions of analysis‟s forecasts upon share buy-back announcements. The authors
document the fact that financial analysts revise their earnings forecasts upwards18.


16 See [Dann, 1981], p. 136.
17 See [Wansley et al, 1989], p. 97.
18 See Hertzel/Jain, 1991], p. 254.


                                                                                               13
2. Review of Prior Research on Share Buy-Backs

Moreover, as several empirical researches documented, buy-backs convey information to
shareholders not only in regard to management‟s expectations on future company earnings,
but also to the level of risk associated with the company. According to Bartov (1991), a share
buy-back may convey information about a decline in the company risk. This might happen in
two ways. Firstly, the decline in risk may represent a reduction in the volatility of the
company‟s future operating cash flows. Secondly, the higher future earnings (that are not fully
paid out) could imply a lower debt-to-equity ratio which in turn implies a lower financial risk19.

In the context of information signalling hypothesis, the following question deserves particular
consideration: If management has confidential information about the prospects of the
company to convey to shareholders, why not make a public announcement (such as a profit
forecast) to them rather than undertake an expensive buy-back?20. In the empirical literature,
several reasons explaining this fact can be found, e.g.: (1) the risk of personal liability of
managers for the announcements they make21 and (2) credibility of the signal. The
announcement of a share buy-back is an effective signal because it is “… backed by hard and
cold cash” and “alternative communications may lack the credibility that comes from „saying it
with cash‟ ”22. Moreover, signals conveyed via a buy-back announcement can conduct positive
information to the capital market without releasing sensitive details to competitors.

In addition, a number of further interesting issues can be found in the empirical literature in
regard to the signalling hypothesis:

    As was stated above, a significant negative relationship between the company size and the
     magnitude of abnormal returns has been documented. So what are the reasons for that the
     smaller the company undertaking the buy-back, the larger, on average, the returns received
     by shareholders? Lakonishok and Vermaelen (1990) outlined several reasons explaining the
     phenomenon:

             Small companies are much less likely to enjoy a deeper analytical and press
              coverage than large companies. Thus, the uncertainty about the quality of a smaller
              company is higher. That is why every (positive) signal from these companies
              renders a higher market reaction.



19 See [Bartov, 1991], p. 277.
20 See [Lamba/Ramsay, 2000], p. 9.
21 For example, in the US, it is the Corporations Law which imposes the risk of personal liability on those who

make a representation with respect to a future matter, e.g. profit forecasts, if they do not have reasonable grounds
for making the representation.
22 See [Asquith/Mullis, 1986], p. 28.


                                                                                                                 14
2. Review of Prior Research on Share Buy-Backs

            Small companies have fewer institutional investors than larger companies23. But it is
             typically institutional investors who have greater resources and higher effectiveness
             in obtaining, analysing and communicating corporate information. Moreover,
             institutions also have greater incentives to monitor the activities of the companies
             they invest in than smaller individuals who just enjoy „free-riding” benefits. In
             contrast, the proportion of individual investors and, in particular, owner-managers
             is in tendency higher at smaller companies24.

    Furthermore, one of the major findings in the framework of the signalling hypothesis was
     the fact that different types of share buy-backs - open-market, tender-offer (fixed-price and
     Dutch auction25) and targeted buy-backs - convey signals of different strengths.

            Vermaelen (1981) found that tender-offers resulted in permanent gains to
             shareholders of +13% while the open-market buy-backs – of only +2%. According
             to Vermaelen (1981), open-market transactions convey less information to the
             market about the true value of the company as small (and normally unspecified)
             quantities of shares are bought over a longer period. So the market can not
             precisely estimate the validity of the signal sent. In contrast, in a tender-offer the
             quantity of shares to be repurchased, the expiration of the offer and the single price
             are explicitly defined.

            Dutch auctions were also found to convey a weaker (and less credible) signal to
             shareholders than traditional fixed-price tender-offer buy-backs. Comment and
             Jarrell (1991) found that Dutch-auctions resulted in an average positive return to
             shareholders of +7.7% vs. +11.9% for fixed-price buy-backs. As the authors
             argued: “…Because a Dutch auction allows managers to guarantee a relatively low
             (marginal) offer price, it follows that Dutch auctions should provide a less-credible
             signal than would an otherwise equivalent fixed-price offer”26 and “In a Dutch
             auction, however, the outsiders have an active role in establishing the terms of
             trade by choosing their tendering prices. To the extent that it is the outsiders‟
             prices that are discovered in a Dutch auction, the latter seems a curious vehicle for
             the signalling of inside information.”27

23 Pugh and Jahera (1990) provided further evidence for US buy-backs that the stock market reaction to the buy-
back announcements is less for companies with larger market caps or widely held by institutional investors.
24 As we saw above, the higher the ownership of the management, the higher the market reaction of share buy-

back announcements.
25 Here it is worth noting the previous Chapter for recalling the major features of a Dutch auction.
26 See [Comment/Jarrell, 1991], p. 1243.
27 See [Comment/Jarrell, 1991], p. 1247.


                                                                                                            15
2. Review of Prior Research on Share Buy-Backs

The signalling hypothesis was also largely confirmed by non-US research, such as Li and
McNally (1999), Ikenberry et al (2000) - Canada, Pellens and Schremper (1999) - Germany,
Zhang (2000) - Japan, Lamba and Ramsay (2000), Otchere and Ross (2000) - Australia,
Moreira and Procianoy (2001) - Brazil, Ginglinger and L‟Her (2002) – France, etc. At this stage
it is very important to stress the significant influence of local capital market environment of
the patters of latter‟s perception of buy-backs. This could be seen on the following example:

The evidence from US that tender offers result in higher abnormal returns– though plausible
and convincing – was not confirmed for Japan28. Zhang (2000) found that announcement
period abnormal returns in Japan were much higher for open-market buy-backs (+7.6%) than
for fixed-price tender-offers (+5.4%). As Zhang (2000) argues, the less positive reaction to
tender-offer buy-backs in Japan is due to the fact that tender offer prices offered in Japan were
not much different from the market price at the announcement (i.e. no premium was offered).
In contrast, in the US the companies were offered significant premiums to the market (up to
+20%)29. On the other hand, the more positive reaction to open-market announcements in
Japan is possibly due to the fact that Japanese firms are more likely to complete their
announced target percentage when using open market transactions than US firms30. Thus, the
evidence from Japan suggests that fixed-price tender offers are just more credible signalling
devices in the US than in Japan and not the tender offer method per se31.

2.2. TAKEOVER DEFENCE HYPOTHESIS
A number of studies have tested the market reaction of share buy-backs undertaken as part of
a hostile takeover defence. Surprisingly, a great deal of empirical research such as e.g. Dann
and DeAngelo (1988), Davidson and Garrison (1989) and Denis (1990) indicate negative
reactions to the buy-back announcements, although, in general, supporting the evidence that
the capital market reacted positively to buy-back announcements not made as a part of a
takeover defence.

This evidence primarily supported the hypothesis of ‘management entrenchment’ which suggests that,
when managers undertake buy-backs in defensive situations, they mainly act in their own
interests at the expense of shareholders. By setting up defensive tactics, managers seek to
„entrench‟ in their positions and prevent changes in the corporate control. In this respect, a


28 See [Zhang, 2000], p. 1.
29 See [Comment/Jarrell, 1991], p. 1245.
30 See [Zhang, 2000], p. 12. As to [Stephens/Weisbach, 1998], the completion rates are in US are 54% after one

year, 69% after two years and 74% after three years upon buy-back announcements.
31 See [Zhang, 2000], p. 3.


                                                                                                           16
2. Review of Prior Research on Share Buy-Backs

share buy-back can be considered as a defensive mechanism increasing the company‟s leverage,
reducing the number of shares available on the market or reducing the cash, i.e. making the
company not that available (or attractive) as a takeover target.

In contrast to the hypothesis of management entrenchment, the shareholders’ interest hypothesis
tries to interpreted defensive tactics as favourable to shareholders, especially when these tactics
allow the managers to focus on the investment decisions which enhance the shareholder value
in the long-term perspective. Thus, a negative reaction to buy-backs in a hostile take-over
situation is, logically, an interplay of the entrenchment and the shareholders‟ interest
hypotheses. When the former prevails, the buy-back results in a overall share-price decline.

Furthermore, a number of researchers have studied the share price reactions of targeted share
buy-backs, i.e. a re-acquisition of a block of shares from one shareholder who has acquired a
substantial shareholding in the company and is threatening a hostile takeover. The results of
these studies are mixed32. In theory, both effects could have impacts here – the management
entrenchment and the shareholders‟ interest factors. Several US studies, such as e.g. Bradley
and Wakeman (1983) indicate that these buy-backs were associated with a decline in the share
prices. These results were largely supported by more recent studies, such as those of Klein and
Rosenfeld (1988) and Mikkelson and Ruback (1991). These studies also documented a negative
reaction following targeted buy-backs. However, the authors also found evidence that the
combined effect of the initial acquisitions of the stakes and their buy-backs later resulted in
positive returns of up to +7 to 12%.

Thus, although believed to be overall positive, the reaction of share buy-back announcements
might differ depending on the situation in which the share buy-back is made. Negative
reactions of share buy-back announcements are broadly observed when the company is under
threat of a hostile takeover.

2.3. OTHER HYPOTHESES
2.3.1. Free Cash Flow hypothesis

The Free Cash Flow hypothesis is primarily attributed to Jensen (1986) 33. This hypothesis
focuses on the agency conflict between managers and shareholders. According to Jensen
(1986), managers have financial and non-financial incentives to invest the free cash flows to
new projects and even to those with a negative NPV34. Thus, since the management also takes

32 See [Lamba/Ramsay, 2000], p. 6.
33 See e.g. [Lakonishok/Vermaelen, 1990].
34 See [Moreira/Procianoy, 2001], p. 8.


                                                                                                17
2. Review of Prior Research on Share Buy-Backs

into account the maximisation of its own convenience, phenomena such as fringe benefits
consumption, „entrenchment‟ or „empire building‟ emerge.

Jensen‟s hypothesis predicts that companies engaging in negative NPV projects will receive an
unfavourable reaction from the market when new investments and acquisitions are
announced35. In those companies, however, an increase of dividends or a share buy-back will
reduce the amount of free cash available for management disposal. Thus, the possible agency
conflicts between managers and shareholders could be reduced. The pay-out of a free cash
flow, which would otherwise be wasted, leads to an increase in the stock prices of those
firms36.

Several researches have carried out empirical tests of FCF hypothesis, such as e.g. Shoven and
Simon (1987), Bagwell and Shoven (1988) or Evans et al (2000). Most of these studies found
evidence to support Jensen‟s point of view.


2.3.2. Taxation Hypothesis

As we have seen, share buy-backs are often used as a means of FCF pay-outs. In this respect,
in several countries, such as e.g. the US or Germany, share buy-backs enjoy a more favourable
tax treatment than dividends37. This could provide a further explanation for (positive) price
reaction surrounding share buy-back announcements.

As stated for the US, the dividend taxes were higher than the taxes on capital gains and
American managers were expected to take advantage of buy-back programs to reduce the total
amount of taxation to their shareholders38. However, such a situation is not always the case
(e.g. in Brazil or Japan) where the taxes on dividends have been sometimes lower than taxes on
capital gains. As long as capital gains are taxed and dividends are not, stock buy-backs increase
prices and the total taxation imposed on shareholders. Expectedly, managers would have no
incentives to use buy-backs to improve shareholder wealth, instead they would pay out special
dividends39. Nevertheless, share buy-backs were launched even in those countries and, in
general, led to positive announcement effects.



35 See [Moreira/Procianoy, 2001], p. 8.
36 See [Moreira/Procianoy, 2001], p. 8.
37 At least at the level of shareholders, the dividends are subject to capital gains taxes whereas gains from

increased share prices are not taxed if the shares are sold after a required holding period (non-speculative gains).
38 See [Moreira/Procianoy, 2001], p. 3.
39 See [Moreira/Procianoy, 2001], p. 3. A combination of a special dividend which would also authorise the

shareholders to purchase stock in the market is also possible. This would bring about a net impact both on prices
and on the reduction of total taxation.

                                                                                                                 18
2. Review of Prior Research on Share Buy-Backs

Furthermore, according to Vermaelen (1981), it still remains unclear whether stock price
reactions to share buy-backs reflect favourable taxation effects or not. This view is supported
by Copeland and Weston (1992) who interviewed some 609 companies in the US which had
conducted at least one share buy-back. None of the 29 reasons classified by the authors feature
a tax shield explanation, neither for the stockholders nor for the companies.


2.3.3. Leverage Hypothesis

The repurchase price may be financed via a debt issue. The possible positive effect of such
share buy-backs could arise from the tax-subsidy connected with the deductibility of interest
payments which is passed on to the shareholders. Originally proposed by Modigliani and Miller
(1963), this hypothesis has been further expanded to share buy-backs by Vermaelen (1981).
According to his results, no sufficient evidence has been found which would allow the
acceptance of the leverage hypothesis as the predominant one.


2.3.4. Minority Shareholder/Bondholder Expropriation Hypothesis

Inter alia, the repurchase strategy can be used to transfer wealth from minority shareholders to
a present controlling ownership group („squeeze‟ of minority interests). Bondholders (or
generally creditors) might also be expropriated when the company lowers its assets base or
takes on more debt to finance the repurchase programs. The plausibility of this hypothesis is,
however, weakened by the existence of laws in many states which restrict buy-backs only to
those made out of surplus or earned surplus accounts. Moreover, most bond covenants put
limits on repurchases in the same way as they put restrictions on dividend payments40.

There is not much evidence on the bondholder expropriation hypothesis available in the past
empirical research. In a study by Wansley and Fayez (1986), the authors found that buy-backs
transfer wealth from creditors to shareholders. However, Dann (1981) as well as Vermaelen
(1981) found no evidence that creditors suffer as a result of share buy-backs.

2.4. IMPLICATIONS FOR OWN RESEARCH AND CONCLUDING REMARKS
A great deal of research has been done on US companies so far. Therefore, there is not much
innovative potential for any further research in this area in the US. Contrary to the US,
however, there is only scarce research done on non-US countries. So there is still a significant
potential for testing the major buy-back hypotheses, whether they hold good in these countries


40   See Vermaelen, 1981], p. 141.

                                                                                             19
2. Review of Prior Research on Share Buy-Backs

or not. In this respect, the countries where buy-back bans or restrictions were implemented
but have been recently lifted, are of particular interest for empirical research. From this
perspective, Germany represents one of those unique countries to investigate.

In prior research, various hypotheses have been advanced to explain abnormal share price
behaviour surrounding announcements of share buy-backs. As far as we can compare them,
the signalling/undervaluation hypothesis appears to stand out as the most fruitful when it
comes to explaining stock price increases surrounding share buy-back announcements41. At
this point, I would argue that my investigation can only be conclusive with respect to the
investigation of the predominant effects behind stock repurchases among a sample of
repurchasing          companies42.       Thus,   particularly   for   this   reason,   I   take   the
signalling/undervaluation hypothesis as the core of my investigation. In the analyses that
follow, I will focus on testing the validity of the signalling/undervaluation hypothesis in the
German capital market environment.

In addition, based on prior buy-back research, the first implications can be drawn for the
empirical testing of the undervaluation hypothesis already at this stage:

      The signalling hypothesis can also be tested by directly examining the buy-back motives as
       stated by the companies themselves. These results belongs to the most valuable as the
       hypothesis is tested without any proxy references;

      Stronger market response to larger share repurchases (i.e. higher percentages of the
       outstanding shares bought back) is consistent with the signalling hypothesis;

      Tender offers result in a significantly higher abnormal price reaction than OMRs. So
       tender offers convey a more favourable signal to the market participants;

      Companies with low market-to-book ratios that engage in share repurchasing show higher
       abnormal returns as they tend to be more undervalued by the market. Thus, this inverse
       relationship between MTB and abnormal returns is also consistent with signalling
       hypothesis;

      If firm size is a proxy for information asymmetries, inverse relationship between size and
       abnormal returns is also consistent with signalling hypothesis. In the framework of
       information asymmetries further proxies can be introduces such as market listing sector or
       industry affiliation.


41   See [Moreira/Procianoy, 2001], p. 11.
42   See Vermaelen, 1981], p. 141.

                                                                                                  20
3. Share Buy-Backs – Situation in Germany


3. SHARE BUY-BACKS – SITUATION IN GERMANY

3.1. SHARE BUY-BACK IN GERMANY – A HISTORICAL DEVELOPMENT
The regulations of share buy-backs in Germany have experienced a long and alternating
history. The absolute ban of share buy-backs had been in effect until the end of the 19th
century. This ban was largely lifted in 1897. Hence, through the world-wide crisis of 1929-
1931, German companies actively used the weakness of the stock markets and bought back
their shares. However, when companies did so, they sometimes came into significant liquidity
problems and bankruptcies were the result43. The emergency regulation of 19th September
1931, introduced a new general prohibition of share buy-backs. So this ban continued in the
subsequent editions of the German „Aktiengesetz‟ of 1937 and 1965.

The major rationale behind the buy-back prohibition during these times was the idea of
creditors and minorities protection from the hazardous changes of shareholder structures. The
creditors should be secured from the potential risks of liability capital reduction which, as a
rule, resulted in an increased risk exposure of creditors. On the other hand, the ban should
facilitate the equal treatment of all shareholders and prevent privileging of selected ones at the
expense of minority shareholders and creditors (as would be the case in a targeted buy-back44).

Thus, until 1st May 1998, in accordance with § 57, No. 1, AktG, share buy-backs in Germany
had been basically prohibited and were only allowed in a limited number of exceptional cases,
for example, when the company intended to implement the repurchased shares for employee
incentive programs.

3.2. CURRENT SITUATION IN GERMANY – A LEGAL PERSPECTIVE
On 1st May 1998, the German „Control and Transparency in Business Act‟ (KonTraG) came
into effect allowing German firms to repurchase their own shares. The new regulation aimed at
increasing firms' financial flexibility and thus took a further step towards adapting the German
capital market to international standards45. This step of liberalisation was largely welcomed by
the business community in Germany46.

According to the KonTraG (in conjunction with § 71 Abs. 1 Nr. 8 AktG), the AGM of the
company can authorise the Management Board to repurchase company‟s shares. This


43 See [Rosen, 1996], p. 435.
44 Or when e.g. only common stock is repurchased and preferred is not, as featured in the Volkswagen case.
45 See [DAI, 1999], p. 1.
46 See [DAI, 1999], p. 1.


                                                                                                             21
3. Share Buy-Backs – Situation in Germany

authorisation can be given for a time period of max. 18 months, during which the company
can buy back the authorised amount of share capital. The authorised amount may not exceed
10% of the share capital, however. The authorisation given by the AGM must also include the
highest and lowest repurchase price at which the shares may be bought back by the company.

Thus, although largely liberalised, the following restrictions can still be observed in the newest
regulation:

    The company may not hold more than 10% of its share capital as a treasury stock (§ 71
     Abs. 2 Satz 1 AktG). This restriction is introduced in order to avoid possible liquidity
     shortages of the company47. It also prevents high-scaled moves of capital structure of the
     company (leverage) and insofar facilitates the protection of minority shareholders and
     creditors.

    Moreover, German companies are not allowed to repurchase their shares for the purposes
     of trading with them. This regulation (§ 71 Abs. 1 Nr. 8 AktG) prevents speculative
     (insider) trading in company‟s securities by the management.

    Inter alia, § 71 Abs. 1 Nr. 8 AktG also preserves the principal of equal treatment of all
     shareholders saying that a buy-back offer has to be directed equally to all stockholders of
     the company.

    The company has to provide the required transparency on the subject of the share buy-
     backs. Firstly, it has to inform the Bundesaufsichtsamt für den Wertpapierhandel (BAW)48
     about its authorisation to buy back shares given to the company by the AGM. Secondly, as
     a rule, the Supervisory Board of the company has to support the share buy-back in a
     separate approval (immediately before the buy-back itself). Thirdly, the company has to
     announce the imminent share buy-back in an ad-hoc announcement. In addition to that, at
     the next AGM, the Management Board has to report to the shareholders about the reasons
     for the buy-back performed, the amount of shares repurchased, as well as the repurchase
     prices. This information must also be published in the Annual Report of the company.

    In the balance sheet of the company, the repurchased shares must be included on the
     assets side according to the lowest value principle (§ 253 Abs. 3 HGB). On the passive




47 If held as a treasury stock, i.e. the share capital is not reduced, the treasury stock is being considered as a cash
position of the balance sheet. However, in higher amounts, such a position could appear illiquid.
48 Now – The Bundesanstalt für Finanzdienstleistungsaufsicht. The latter took over the responsibilities of BAW,

beginning 1st May 2002.

                                                                                                                   22
3. Share Buy-Backs – Situation in Germany

     side, the same amount should be balanced in as a provision. Similar to a pay-out lock49, this
     provision primarily carries the creditor protection function.

3.3. ANALYSIS OF THE BUY-BACK ACTIVITY IN GERMANY
Since the ban lift for share buy-backs in Germany in May 1998, some 160 companies50 have
engaged in buy-back of their own shares. They represent more than 20% of all German public
corporations listed on the German Stock Exchanges51. During four years between May 1998
and May 2002 these 160 companies announced some 193 share buy-backs. In the following
section, a more detailed analysis of the repurchasing activity in Germany is being presented.

Already during the AGM season of 1998, a large number of German corporations obtained an
authorisation to buy-back their shares, in order to possess additional option for their financial
flexibility. However, the actual buy-backs did not materialise, as the question of tax treatment
of buy-backs remained unclear. That is why the first wave of share buy-backs in Germany did
not break off until the tax situation was cleared by the German Finance Ministry at the
beginning of 1999.




49 Translation of: Ausschüttungssperre.
50 All the figures shown here are sourced from my own investigation.
51 The total number of 790 corporations is taken as basis here, all companies-constitutes of the CDAX, as of 17 th

May 2002.

                                                                                                               23
3. Share Buy-Backs – Situation in Germany


On FIGURE 1 below, the Top 10 German share buy-backs are presented. As we can see, the
values of the underlying transactions are considerable and easily comparable with the largest
German M&A transactions of the recent past52:

     FIGURE 1: TOP 10 SHARE BUY-BACKS IN GERMANY (MAY/1998 – MAY/2002, €BN)
       BUY-BACK VALUE MAX. AUTHORISED BY AGM53,54                         BUY-BACK VALUE AS REALLY PAID BY COMPANY FOR
                                                                                     THE SHARES ACQUIRED53




           Siemens                                                 9.2          Allianz                                      5.8


           Allianz                                       5.8                      E.On                                 4.5


              E.On                                 4.5                            RWE                        2.8

       Muenchener
                                                 4.1                        Volkswagen                     2.3
          Re

              RWE                      2.8                                        Bayer              1.4


     Deutsche Bank                   2.4                                          BASF               1.3


             Bayer                   2.3                                       Siemens               1.2


             BASF                    2.3                                       Hoechst              1.1


       Volkswagen              1.7                                        Deutsche Bank        0.5


     Commerz-bank             1.5                                              Schering       0.2


                     0         2             4           6     8     10                   0            2           4         6     8        10



 Source: own investigation, Annual Reports, IR Interviews




52 Such as for example Siemens, Bosch/Atecs Mannesman (€10bn), Bayer/Aventis Cropscience (€7bn), Imperial
Tobacco/Reemtsma (€5bn), Interbrew/Brauerei Beck (€2bn) etc.
53 Based on the average share price for five days before the buy-back announcement.
54 On the left side, only a potential value of the buy-back transactions is shown as the figures are based on the

amount of shares to which the companies are authorised. The actual value paid for the shares bought back is
shown on the right side of the figure.

                                                                                                                                       24
       3. Share Buy-Backs – Situation in Germany


       In FIGURE 2 below, more detailed results of analysis of repurchasing activity of German
       companies are presented:

  FIGURE 2: ANALYSIS OF REPURCHASING ACTIVITY BY INDUSTRY
    RELATIVE ACTIVITY THROUGHOUT INDUSTRIES, %                                                                           PARTICIPATING ACTIVITY WITHIN INDUSTRIES, %


                                                                                                                                                                                                                    Total number of
                                                                                                                                                                                                                    constituents in the
                                                                                                                                                                                                                    industry

                                                                                                   22%
            Software                                                                   17%                            Chemicals                                                                             38%         13
                                                                     10%
        Technology                                                         12%                                       Automobile                                                                 30%                     27
                                                                9%
  Financial Services                                          8%                                                         Utilities                                                             29%                      14
                                                       6%
             Pharma                                      7%                                                                Retail                                                         27%                           44
                                                    6%
            Industrial                                6%                                                                Software                                                         26%                            133
                                             4%
         Machinery                                     6%                                                                 Banks                                                      26%                                23
                                                        6%
               Media                                   6%                                                      Financial Services                                             22%                                       64
                                                             8%
                Retail                              6%                                                         Consumer Cyclical                                             21%                                        33
                                                     6%
      Construction                                  5%                                                              Construction                                             21%                                        43
                                               4%
 Consumer Cyclical                            4%                                                                          Media                                              21%                                        48
                                                  5%
       Automobile                         3%                                                                           Industrial                                      18%                                              49
                               1%
                Food                     3%                                                                              Pharma                                        18%                                              55
                                             4%
               Banks                    3%                                                                           Technology                                   16%                                                   97
                              1%
             Telecom                    3%                                                                             Insurance                                 16%                                                    19

                                   2%
            Insurance               2%                                                                                Resources                            13%                                                          15

                              1%
         Transport                  2%                                                                                Machinery                           13%                                                           48

                               1%
         Resources                                                   Weight of the industry (number of                     Food                     8%                                                                  25
                                 2%

                                     3%                              companies) in the sample
             Utilities                                                                                                 Transport               6%                                                                       17
                                   2%
                                                                     Weight of the industry (number of
                                         3%
         Chemicals                 2%
                                                                     companies) in the CDAX                             Telecom           4%                                                                            23


                         0%                    5%                 10%            15%         20%         25%                         0%   5%        10%   15%          20%         25%     30%        35%   40%   45%



Source: own investigation


       FIGURE 2 (on the left) shows a comparative repurchasing activity throughout different CDAX
       industries55. For these purposes, the relative weight of the respective industry in the buy-back
       sample (based on the number of repurchasing companies) is being compared with the relative
       weight of the industry in the CDAX (also based on the number of companies). As we can see,
       the Software, Retail, Automotive and Chemicals industries experienced relatively higher
       repurchasing activity56 with Technology, Machinery, Food & Beverages, Telecom and
       Transport & Logistics lagging somewhat behind57. However, the overall picture is well-
       balanced, featuring no particularly dominating industry according to its level of the buy-back
       activity.

       55 For these purposes, CDAX industry sub-indexes (numbers of constitutes) have been used.
       56 E.g. 22% of the sample firms were Software companies whereas the software industry accounted for „only‟ 17%
       of all publicly traded companies in Germany.
       57 These results are to be treated carefully, as the weights are based on the number of companies and not on their

       market caps, for example.

                                                                                                                                                                                                                        25
3. Share Buy-Backs – Situation in Germany

FIGURE 2 (on the right) shows the participating activity in the respective industries, i.e. the
proportions (in %) of industry companies that already launched a share buy-back program58.
Here, we also can see that, without extreme deviations, the majority of industries tends to
group on the proportions in the range of 15-25%. However, the Chemical industry appears to
have the highest level of company participation with 38% of the companies having
implemented share buy-backs. The Food, Transport and Telecommunication industries have
the lowest participations levels ranging between 4% and 8%. The financial services industries
(Banks, Financial Services and Insurance) tend to group in the middle of the range.

FIGURE 3 shows the behaviour of share buy-back announcements in the time of the
investigation period between May/1998 and May/2002:

             FIGURE 3: DEVELOPMENT OF REPURCHASING ACTIVITY IN GERMANY
             MAY/1998 – MAY/2002

                                 15
                                                                         1998:                                                                                                         1999:                                                                                                                              2000:                                                                                                                          2001:                                                                      2002:
                                                                         4 buy-backs                                                                                                   51 buy-backs                                                                                                                       67 buy-backs                                                                                                                   63 buy-backs                                                               7 buy-backs

                                                                                                                                                                                                                                                                                                                                                                                                                                                                       12
     Buy-back number per month




                                 10
                                                                                                                                                                                                                                                                                                                               9

                                                                                                                                                                                                                             8                                8                                                                           8                                          8                                                                                             8         8

                                                                                                                                                                                                        7                                          7                                7                                                                                     7                    7

                                                                                                                                                                                                                                                                                                                                                                6                                                            6                                6

                                                                                                                                                                                                                   5                    5                                                                                                                                                                                               5                                                              5
                                   5
                                                                                                                                                                                                                                                                                                                     4                                                                                                                             4                                                             4

                                                                                                                                     3                                                       3                                                                           3                               3                                           3                                                                                                                                                                     3                                        3

                                                                                                                           2                                2                                                                                                                                 2                                                                                                          2                                                                                                                                               2

                                                                    1                                          1                                1                      1          1                                                                                                                                                                                                                                                                                                                                                  1         1

                                              0          0                     0         0           0                                                                                                                                                                                                                                                                                                             0
                                   0
                                                                                                   Oct-98




                                                                                                                                               Feb-99




                                                                                                                                                                                                                                       Oct-99




                                                                                                                                                                                                                                                                                  Feb-00




                                                                                                                                                                                                                                                                                                                                                                         Oct-00




                                                                                                                                                                                                                                                                                                                                                                                                                 Feb-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     Oct-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                             Feb-02
                                            May-98
                                                       Jun-98
                                                                  Jul-98


                                                                                        Sep-98


                                                                                                              Nov-98


                                                                                                                                    Jan-99




                                                                                                                                                                                May-99
                                                                                                                                                                                           Jun-99
                                                                                                                                                                                                      Jul-99


                                                                                                                                                                                                                            Sep-99


                                                                                                                                                                                                                                                  Nov-99


                                                                                                                                                                                                                                                                        Jan-00




                                                                                                                                                                                                                                                                                                                   May-00
                                                                                                                                                                                                                                                                                                                              Jun-00
                                                                                                                                                                     Apr-99




                                                                                                                                                                                                                                                                                                        Apr-00




                                                                                                                                                                                                                                                                                                                                         Jul-00


                                                                                                                                                                                                                                                                                                                                                              Sep-00


                                                                                                                                                                                                                                                                                                                                                                                   Nov-00


                                                                                                                                                                                                                                                                                                                                                                                                       Jan-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                            Jun-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               Nov-01
                                                                                                                                                                                                                                                                                                                                                                                                                                       Apr-01
                                                                                                                                                                                                                                                                                                                                                                                                                                                  May-01


                                                                                                                                                                                                                                                                                                                                                                                                                                                                       Jul-01


                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           Sep-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   Jan-02




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Apr-02
                                                                                                                                                          Mar-99




                                                                                                                                                                                                                                                                                             Mar-00




                                                                                                                                                                                                                                                                                                                                                                                                                            Mar-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        Mar-02
                                                                                                                         Dec-98




                                                                                                                                                                                                                                                             Dec-99




                                                                                                                                                                                                                                                                                                                                                                                             Dec-00




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         Dec-01
                                                                             Aug-98




                                                                                                                                                                                                                 Aug-99




                                                                                                                                                                                                                                                                                                                                                    Aug-00




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Aug-01
                                                                                                 Oct-98




                                                                                                                                             Feb-99




                                                                                                                                                                                                                                     Oct-99




                                                                                                                                                                                                                                                                                 Feb-00




                                                                                                                                                                                                                                                                                                                                                                       Oct-00




                                                                                                                                                                                                                                                                                                                                                                                                                Feb-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Oct-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            Feb-02
                                                                                                                                                        Mar-99




                                                                                                                                                                                                                                                                                           Mar-00




                                                                                                                                                                                                                                                                                                                                                                                                                          Mar-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      Mar-02
                                                                                                                       Dec-98




                                                                                                                                                                                                                                                           Dec-99




                                                                                                                                                                                                                                                                                                                                                                                            Dec-00




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                        Dec-01
                                                     Jun-98




                                                                                      Sep-98


                                                                                                            Nov-98




                                                                                                                                                                                         Jun-99




                                                                                                                                                                                                                          Sep-99


                                                                                                                                                                                                                                                Nov-99




                                                                                                                                                                                                                                                                                                                            Jun-00




                                                                                                                                                                                                                                                                                                                                                             Sep-00


                                                                                                                                                                                                                                                                                                                                                                                  Nov-00




                                                                                                                                                                                                                                                                                                                                                                                                                                                           Jun-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          Sep-01


                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Nov-01
                                 Apr-98
                                          May-98


                                                                Jul-98




                                                                                                                                  Jan-99



                                                                                                                                                                   Apr-99
                                                                                                                                                                              May-99


                                                                                                                                                                                                    Jul-99




                                                                                                                                                                                                                                                                      Jan-00




                                                                                                                                                                                                                                                                                                      Apr-00
                                                                                                                                                                                                                                                                                                                 May-00


                                                                                                                                                                                                                                                                                                                                       Jul-00




                                                                                                                                                                                                                                                                                                                                                                                                      Jan-01



                                                                                                                                                                                                                                                                                                                                                                                                                                     Apr-01
                                                                                                                                                                                                                                                                                                                                                                                                                                                May-01


                                                                                                                                                                                                                                                                                                                                                                                                                                                                     Jul-01




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                  Jan-02



                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                 Apr-02
                                                                           Aug-98




                                                                                                                                                                                                               Aug-99




                                                                                                                                                                                                                                                                                                                                                  Aug-00




                                                                                                                                                                                                                                                                                                                                                                                                                                                                                Aug-01




 Source: own investigation


FIGURE 3 (the upper part) shows the development of cumulated monthly announcements in
time whereas the spectrum-like part below presents the non-cumulated daily density of share



58 Here again, the CDAX Industry indexes have been used in order to define the industry itself and the number of
its constitutes.

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           26
3. Share Buy-Backs – Situation in Germany

buy-back announcements as the time passed. Thus, a set a set of very interesting results can be
derived from this analysis:

    During the first calendar year following the ban lift on share buy-backs, only a very weak
     buy-back activity could be observed. At a first glance, this is quite an astonishing result –
     after the ban lift in May 1998, a big wave of share buy-backs had been expected. These
     expectations were mainly based on the fact that there was a large overhang of companies
     willing to buy-back their own shares at that time. However, due to the unclear tax
     treatment of the capital gains derived from share repurchases, the major buy-backs were
     postponed at that time. On the other hand, this blank period could also be explained by
     the fact that there is always a certain time lag until the buy-back can be started as the
     option to repurchase shares has to be authorised by the AGM59. It is worthwhile to say that
     the buy-back restrictions were first lifted in May, i.e. in the middle of the AGM season.

    Contemplating the density spectrum of the share buy-backs (the lower chart), we can
     clearly distinct four major time clusters – they are highlighted with the figured parentheses
     on the chart. Understandably, these periods of higher buy-back density correspond with
     the higher numbers of share buy-backs on a monthly basis (the upper chart).

    As a result, at least four major cyclical waves can be identified within the last four years –
     the cycles that also feature well observed seasonal ups and downs (!). The frequency of
     share buy-backs rises from the middle of the year steadily towards the year end (which is
     particularly well distinguishable for 1999 and 2000) to nearly break off in the beginning of
     the next year. Apparently, a conscious timing of the share buy-backs by the management of
     the companies can explain this cyclical behaviour of the share buy-back activity in
     Germany.

    I would suggest that one reason for this phenomenon could possibly be the management‟s
     efforts to improve the share-price towards the financial year end and to show possibly
     higher valuation ratios60. On the other hand, the cycles could possibly be created by the
     seasonality of AGMs which predominantly take place in April-May of each year. If
     approved by the AGM, the buy-back programs would normally start in the second half of
     the year. In contrast, however, managers would rather not start a share buy-back program
     short before the AGM (i.e. in the first half of the year) in order to prevent having to report
     additionally about it at the AGM.

59 It should be mentioned at this stage, however, that many of the authorisations have been granted before the law
came into effect.
60 Under assumption that the financial year of the most companies ends 31 st December.


                                                                                                               27
3. Share Buy-Backs – Situation in Germany

    There is an overall rising trend in share buy-back numbers throughout the past four years.
     However, a major downturn of buy-back activity can be observed in the recent past – i.e.
     from the middle of 2001 to May 2002. This might be possibly explained by an overall
     downturn in economic activity in Germany, as well as worldwide61. On the other hand, the
     treatment of repurchased shares in the framework of the new takeover code which is being
     introduced is also still unclear62.

    Thus, for example, the cyclical behaviour of share buy-backs must be taken into
     consideration before making any forecasts of share buy-back activity. In the case of 2002,
     the higher number of share buy-backs is probably still to come.

3.4. RATIONALE FOR SHARE BUY-BACKS IN GERMANY
In Germany, managers are not legally required to disclose their motives for share repurchases.
However, in contrast to many other countries, such as the US, it is general practice in
Germany to reveal the motives of the share buy-backs in the share buy-back announcements63.

In FIGURE 4 below, the rationale for share buy-backs in Germany is presented. The data for
1999 is sourced from a study carried out by DAI (Deutsches Aktieninstitut) in June 199964.
Thus, the availability of these data provided a nice opportunity of comparing those results with
the rationale as given by the repurchasing companies in their buy-back ad hoc announcements
during the period May/1999 – May/200265:




61 The question of how the share buy-back activity is related to the economic activity cycles could probably yield
some very interesting results when investigated in more depth.
62 This was, for example, why Siemens postponed its recent share buy-back up to next year. As Siemens argues,

there is still a degree of uncertainty whether several types of share buy-backs are to fall under the consideration by
the new take-over code. For example, in the case where Siemens is tendering Infineon‟s shares for its own shares,
this share buy-back is probably subject to regulation by the take-over code.
63 Apparently in line with the legal requirement to provide transparency with respect to share buy-backs.
64 This study included 67 companies possessing the right to buy-back shares. The companies were questioned

directly for their reasons to buy back shares.
65 The data was primarily sourced from the buy-back announcements of the repurchasing firms. Unfortunately,

the majority of the firms have given more than one reason for why they are buying back their own stock. In these
cases, every reason was counted equally in the final results. The data was prepared appropriately, in order to
provide better comparability between the results.

                                                                                                                   28
3. Share Buy-Backs – Situation in Germany


     FIGURE 4: RATIONALE FOR SHARE BUY-BACKS IN GERMANY

                         Up to June/1999                                                   Up to May/2002
       Employees              Others                   Acquisition         Cancellation/         Others     Acquisition
       Participation          7%                       Currency            Pay-Outs to           2%         Currency
       Programmes                                      36%                 Shareholders                     43%
       4%                                                                  9%

       Optimisation of                                                    Employees
       Shareholer                                                         Participation
       Structure                                                          Programs
       13%                                                                12%




      Cancellation/
      Pay-Outs to
      Shareholders
      18%                                                                 Undervaluation
                                                                          34%
                                         Adjustment of Capital
                                         Structure
                                         22%

 Source: DAI, Deutsches Aktieninstitut                               Source: own investigation


     As we can see, the most popular reason for buying back own shares, both in 1999 and
      2002, was to use the repurchased stock as currency for future acquisitions. However, at this
      stage, this reason seems to be rather a pseudo-motive, since no obvious advantages could
      be drawn from paying in shares instead of cash in acquisitions66.

     Undervaluation of the stock seems to be the second important rationale for share buy-
      backs67.

     Moreover, as we can see, the role of treasury stock for its implementation in employees
      incentivisation programs has significantly increased over time.




66 However, in acquisitions where shares are used as payment currency, companies would rather pay the entire
price in shares if they intended to anyway. Thus, transaction costs could be saved if the entire price is paid in „one
currency‟. Furthermore, a slight indication to an advantage of the payment in shares instead of cash could be the
fact that according to German GAAP (HGB), the acquired assets can be balanced in under continuation of their
book values (i.e. hidden reserves are not realised) if the at least 90% of the purchasing price for the assets is paid
in shares. However, this argument is ever diminishing in Germany as more and more companies switch to
accounting to standards other than HGB.
67 Understandably the data of the two charts are only limited comparable. „Adjustment of capital structure‟ (DAI,

1999) should probably correspond to „Undervaluation‟ – the DAI study investigated only companies which simply
possessed an AGM authorisation to buy-back shares (i.e. they must not have bought back their own shares).
Thus, at this point in time, it does not make sense to announce an undervaluation reason.

                                                                                                                          29
3. Share Buy-Backs – Situation in Germany


3.5. FURTHER SPECIFICS OF SHARE BUY-BACKS IN GERMANY

3.5.1. Buy-Back Methods

All possible types of share buy-back transactions are allowed in Germany68. However, the most
preferred method of share buy-backs in Germany is the open market transaction on the Stock
Exchange. From the total of 193 buy-backs announced in Germany only 4 could be identified
as being non-open market. Those were AGIV AG (4-Apr-00, fixed-price tender offer on
shareholders), Friedrich Grohe AG (7-Oct-99, fixed-price tender offer on minorities holders
of preferred stock), Koegel Fahrzeuge AG (7-Dec-98, tender offer on common shareholders)
and Krones AG (18-Jan-99, Dutch auction tender-offer)69.


3.5.2. Completion Rates

The absolute majority (i.e. more than 70%) of companies has been authorised for the
repurchase of the maximum possible amount of 10% of share capital. However, the companies
did not buy back 10% at once but gradually, in several steps, sometimes over months or even
years. Furthermore, German companies are not obliged to buy back the entire amount of
shares which is authorised by the AGMs. And even the targeted percentage of shares to buy
back which is given in the ad-hoc announcement is not legally binding.

During the time period in which the buy-back option is granted to the company, the firm may
use its treasury shares as payment in acquisitions, in employee compensation schemes or even
cancel the shares reducing its share capital. Therefore, it is very difficult to talk about the
completion rates of share buy-backs by German companies70,71.


3.5.3. Pricing of Share Buy-Backs

In general, the Stock Exchange price is broadly chosen to be the major criterion for the price
at which the company would buy back its shares. As usual, the buy-back authorisation given by

68 Although a targeted buy-back contradict the general principle of equal treatment of all company shareholders
(stated e.g. in the German Aktiengesetzt and the new Takeover Code), it might take place in certain circumstances
such as e.g. buy-back of company‟s own shares held by one single shareholder, i.e. if there are no free-float or
other shareholders.
69 At this stage, it would be interesting to further investigate the question why predominantly open market

transactions are used by German firms. Possibly, the 10% limit as imposed by the legislation can largely account
for this fact. Moreover, as German firms have usually strong banking relationships (Hausbanken), they would
more often use brokerage services of the banks.
70 At least by means of analysing of treasury stock figures available in the Annual Reports of the companies.
71 In the US, announced share buy-back programs are not firm commitments either. Stephens and Weisbach

(1998) estimate that US firms repurchase about 54% of their announced target within a year of the initial
announcement and 74% within three years.

                                                                                                              30
3. Share Buy-Backs – Situation in Germany

the AGM would include the price range (levels above and below the average price) in which
the shares may be bought back72.

By setting up the price range, the company attempts to avoid pushing up its own share price if
the company enters the market directly and attempts to treat all shareholders equally as
required by the law.

3.6. CONCLUDING REMARKS
The pace of the share buy-back activity in Germany has slowed down up to the end of 2001 -
the beginning of 2002. As it seems, this slow-down could not alone be explained by a cyclical
down-turn, as an additional negative factor apparently influenced the trend during the second
half of the year 2001. In my view, that might possibly be the influence caused by the overall
economic recession which led to falling numbers of share buy-backs announced during this
period.

Over the past four years, share buy-backs have gained increasing popularity among German
firms. This instrument has been used by some 160 companies predominantly for the purposes
of financing acquisitions and share price management, as well as a substitute for cash flow pay-
outs or implementation in employees incentivisation programs. As the persistence of these
needs does not seem to be on decrease now and in the future, the popularity of share buy-
backs should not diminish either. According to the DAI, there is still a substantial and ever
increasing backlog of AGM authorisations for share buy-backs73.

It is worthwhile to say at this stage, however, that the buy-back activity is very sensitive to the
capital market environment since every buy-back is simply an option and there is no obligation
to exercise it. Thus, with regards to the future buy-back activity in Germany, much will depend
on any possible changes in the taxation treatment of share buy-backs, on their treatment within
the framework of the new takeover code, as well as on the ability of German companies to
generate excess cash flows to pay out to their shareholders.




72 E.g. an excerpt from an ad-hoc announcement: “The equivalent of the repurchased stock must be within a
price range of +10/-15% of the actual share price. The relevant share price is defined as the average closing price
at the Frankfurt Stock Exchange (FSE) during the last five (5) trading days prior to the repurchase stock or before
the public announcement of the buying offer.”
73 This fact is also confirmed by my own investigation, as the German companies usually tended to extend their

authorisations to buy back shares at the following AGMs.

                                                                                                                31
4. Investigation Methodology


4. INVESTIGATION METHODOLOGY

In this chapter, the methodology of the ensuing investigation will be described and a general
structural framework for the following analyses will be provided.

In my empirical investigation, I used event study methodology to analyse the abnormal returns
behaviour around the announcement of share buy-backs. Using financial market data, an event
study measures the impact of a specific event on the value of the firm (share price). The
usefulness of such a study comes from the fact that given rationality in the market place, the
effects of an event will be reflected immediately in security prices74. The event study has a very
broad range of application fields in accounting and finance research, such as M&A, earnings or
dividends announcements, issues of debt or equity etc. In this respect, share buy-backs also
represent a classical case for examination by an event study - particularly due to its clear
distinguishable event day (the day of share buy-back announcement) and a significant
information content which presumably exists behind a share buy-back announcement and
should be therefore reflected in share prices.

4.1. EXCURSE: EVENT STUDY MODELS – BACKGROUND THEORY
Measuring of the normal performance with a subsequent calculation of the abnormal
returns)represents the core of every event study methodology. According to McKinlay (1997),
the following event study models can be differentiated 75:

1) Constant Mean Return Model

                                                        Rit = μi + ξit

                                        E(ξit)=0                     var(ξit)=σ2(ξit)

         Rit is the period-t return on security i and ξit is the disturbance term for security i.

         As we can see from the formula above, the main advantage of the constant mean
         return model is its simplicity. In addition to this, as stated by several researchers76, the
         model often yields results similar to those of more sophisticated models as the variance
         of the abnormal returns is frequently not reduced much within a more sophisticated
         model. Another pro in favour of this model is its applicability to the situations when
         only limited statistical data is available.


74 See [MacKinlay, 1997], p. 13.
75 See [MacKinlay, 1997], p. 17. Only statistical models are presented here.
76 Such as [Brown and Warner, 1980] and [Brown and Warner, 1985].


                                                                                                    32
4. Investigation Methodology

2) Market Model

         This is the standard event study methodology of Brown and Warner (1985). The stock
         return process assumedly follows the simple market model specification:

                                                  Rit = αi + βiRmt + εit

                                       E(εit)=0                    var(εit)=σ2(εit)

         Rit and Rmt are the period-t return on security i and the market portfolio, respectively,
         and εit is the zero mean the disturbance term. αi, βi, σ2(εit) are the parameters of the
         market model which are calculated within the estimation window.

         Thus, the market model is a statistical model which relates the return of any given
         security to a return of the market portfolio. This represents a potential improvement
         over the constant mean return model77 - the variance of the abnormal returns is
         reduced as the proportion of the return that is related to variation in the market return
         is removed.

         As can be seen, the market model is not much more complex than the CMR model.
         Thus, it provides a significant quality improvement at very limited costs78. This is
         probably why there is almost no research literature which would claim superiority of
         the constant mean return model over the market model.

         However, as it now seems, the major weakness of the market model is the definition of
         the market portfolio itself. Since the ideal market index does not exist and even the
         broadest defined market indexes do not perfectly seize all the traded assets79, only their
         substitutes can be used – such as e.g. S&P 500, FTSE 100 or DJ Industrials (for
         German empirical investigations, the CDAX (Composite DAX) became the most
         frequently used benchmark as it has 790 constituent firms – all German publicly traded
         companies80).

3) Other Statistical Models

         This group includes a number of other statistical models which are predominantly
         represented by multifactor models. In order to further reduce the variance of abnormal


77 See [MacKinlay, 1997], p. 18.
78 See [Brown and Warner, 1985], p. 4.
79 For example, in the CAPM model, the definition of the market portfolio includes all available capital assets,

even including such assets as human capital. Thus the market portfolio defined in this way is quite a theoretical
construction.
80 Figure as of June 2002.


                                                                                                              33
4. Investigation Methodology

         returns by explaining more of the variation in the normal returns, new factors are
         introduced into the model81. Those are for example industry groups or company size82.

The quality of an event study model could be measured by different criteria such as e.g. the
residual variance of the abnormal returns, model‟s applicability to the given data limitations, or
model simplicity. Moreover, in a framework of the data available, a model should provide the
lowest possible variance (ideally a zero-variance) of the abnormal returns in the control period.
I.e. as much as possible distillate the returns attributed to the event only. So for example, the
implementation of multifactor models might yield additional gains for the quality of the final
results. However, in most cases, this improvement is achieved at very high costs83.

In my particular case, I was neither constrained by the data availability nor by any other
limitations such as time or lack of the challenge to implement a more comprehensive model.
Thus, at that stage, the implementation of a multifactor market model in the framework of my
investigation seemed challenging and exciting at the same time.

Therefore, I carried out a comprehensive comparative analysis as on the possible gains which
can be achieved through the implementation of a multifactor model based on the CDAX
industrial sub-indexes. I presented the results of this analysis are presented in Appendix I84. In
this analysis, I came to a conclusion, however, that the marginal gains from the implementation
of a multifactor model could barely make up for the increased complicity of the analysis. Thus,
I have decided to chose the standard market model for the event study analyses that follow in
my investigation.

4.2. MARKET MODEL AND CALCULATION METHODOLOGY
I used the standard market model methodology to evaluate the abnormal reaction to
announcements of share buy-backs. The daily stock returns were computed on the basis of the
company daily stock prices sourced from the Datastream capital market database. As an
appropriate benchmark for the market model, the Composite DAX (CDAX) has been used.

81 See [MacKinlay, 1997], p. 18.
82 Thus, the normal returns of the company would be modelled by means of not only one factor such as a market
index in the market model, but also by means of other applicable factors. For these purposes, peer groups are
created for the sample firms under investigation and the firm‟s returns are linked to the returns of the peer
groups.
83 See [MacKinlay, 1997], p. 18.
84 The implementation of a multifactor model based on industrial sub-indexes for the purposes of my empirical

investigation has also induced a very lively discussion with my fellow students and research colleges, in
particularly, with Dr. Ralf Elsas (Chair of Corporate Finance, Frankfurt University). From this perspective, my
empirical study provided a great opportunity to look deeply into these problematics. Although not the subject of
this paper, the inclusion of such an analysis seemed important to me particularly on behalf of my research
colleagues.

                                                                                                             34
4. Investigation Methodology

CDAX is the most broadly defined market index available in Germany as it includes some 790
German publicly traded companies.

The market model parameters were estimated over the days [-270;–60] relative to the
announcement day [0]. Such a long time lag (60 days) between the estimation window end and
the event day can be justified since the possibility of information leakage regarding the planned
share buy-back announcement might have been high85. On the other hand, this long pre-
announcement lag could be very helpful for the detection of any possible underperformance
prior to the buy-back announcements.

In main analyses of my investigation, the abnormal returns on the announcement day (AR [0])
and on the following day (AR [+1]) have been examined. Moreover, several time windows
around the day [0] have been investigated as well - usually the time windows [-1;+1] and [-
1;+5].

In most analyses, the 0-hypothesis has been tested, i.e. that the average abnormal returns (AR)
or cumulative abnormal returns (CAR) over the investigation period are not significantly
different from zero. In order to test the statistical significance of the obtained results a t-test
has been performed86.

4.3. SIGNIFICANCE TESTS
To test the 0-hypothesis, i.e. that the abnormal returns on the event day [0] (or the day i
around the event day) are not statistically different from 0, I computed the t-statistic:

                                               t = ARi/σ(ARi)

where σ(ARi) is the sample standard deviation of the portfolio returns for the estimation
window87.

The t-statistic can also be computed for cumulative abnormal returns:

                                          t = CARii+n/σ(CARii+n),


85 Typically, German firms have more stable shareholder structures, with prevailing larger controlling stakes,
much higher cross-holding density, interlocking directorships and the more important role of banking institutions
in the corporate management (the Germany AG). That could possibly increase the probability of information
leakage (or insider trading).
86 The major methodological patterns used here are taken from [Lamba/Ramsay, 2000], an investigation of buy-

back announcements in Australia.
87 Vermaelen, (1981) uses the combination of two periods [-60;-10] and [+10;+60] as control period for the

calculation of σ(ARi). In order to calculate σ(ARi), I used the estimation window [-270;-60] – that secures the
higher overall consistency of the calculations as the major model parameters are estimated for this period.
Furthermore, as we will see below, the sample companies feature overall underperforming trends, which is why
using the control period of [-60;-10] and [+10;+60] would not be appropriate enough.

                                                                                                              35
4. Investigation Methodology

                                            σ(CARii+n)=√n • σ(ARi)

where CARii+n is the cumulative abnormal returns for the time period from t until t+n. σ(ARi)
is the sample standard deviation of the sample abnormal returns computed as before88.

4.4. SAMPLE DESIGN
My analysis focused on all types of share buy-backs announced by German public companies
during the time period between May/1998 and May/2002. The initial buy-back sample
(approx. 70 buy-back announcements) was obtained from the M&A database of SDC (Security
Data Company). This sample was then enlarged with new companies found in Reuters and
Bloomberg news databases, as well as in the ad-hoc announcements database of the Deutsche-
Börse AG89. This research resulted in a sample of 160 companies which carried out some 193
buy-backs it total90.

To be included in the Main Sample91, the following restrictions were implemented: (1) there are
no other confounding events reported in the two days before and after the announcement date
of the buy-back92, (2) daily returns for at least 60 days before the event and the event period
are available. These two criteria resulted in a final sample of 189 share buy-backs.

4.5. INVESTIGATION STRUCTURE
The subject of my empirical investigation is the abnormal share price behaviour in association
with share buy-backs in Germany. Moreover, I mainly focused on the testing of the validity of
undervaluation/signalling hypothesis for the sample of German repurchasing firms.

The German regulation environment provides a unique opportunity to examine the effects of
at least three major consequent buy-back events: (1) buy-back announcements, (2)


88 See [Vermaelen, 1981], p. 148.
89 It is understood, that only publicly announced buy-back transactions have been included in the final sample.
The comprehensiveness of the research carried out, as well as the search in all the available sources of public
information, let me conclude that there is a very high probability that all the buy-backs have been registered in the
final sample, possibly leaving some space for some 2 or 3 unregistered transactions.
90 During the investigation period, a number of companies have made more than one buy-back announcement.

Therefore, the number of buy-backs in the sample exceeded the number of companies.
91 The definition „Main Sample‟ is used here, in order to distinguish if from a number of other samples created

during the preparation of this paper.
92 Only those confounding events have been considered relevant which stood in no obvious connection with the

share buy-back itself. Examples of such events are board changes announcements, windfall gains or losses etc. In
contrast, it is very difficult to estimate the possible confounding specifics of other information with which share
buy-backs are more closely connected. Those are for example financial forecasts, share splits, threats of take-
overs, issuing of strategic guidelines, etc. In such cases, the share-buy backs would be derivative signals for these
good (or bad) news. I.e. a share buy-back cannot be qualified as good or bad news per se, as it is simply a signal of
the underlying developments within the company. Thus, they cannot be „separated‟ from the underlying news.

                                                                                                                 36
4. Investigation Methodology

announcements of (or news on) share buy-back plans and (3) buy-back authorisations by
AGMs. This time sequence of events is outlined in FIGURE 5 below:


     FIGURE 5: TIME SEQUENCE OF A SHARE BUY-BACK PROCESS IN GERMANY




     Buy-Back Plans                            Authorisation by               Share Buy-Back
     (Seeking Approval)                        Annual General
                                               Meeting                        (Ad-hoc Announcement)




                               Chapter 6                                            Chapter 5

 Source: own investigation




In the sequence of the buy-back events, the buy-back announcement is the last and the most
important one93. The buy-back announcement is the core event of the share buy-back
programs in Germany as it marks up the time point when the company really intends to buy
back its own shares paying „real and cold cash‟ for them. A detailed investigation of share buy-
back announcements is provided in Chapter 5. And separately, in Chapter 6, analyses of share
price reactions to the announcement of buy-back plans and buy-back authorisations by the
AGMs is presented.




93This is also documented by the fact that the investigation of buy-back announcements has been the primary
core of the vast majority of the prior empirical research.

                                                                                                        37
5. Empirical Investigation of Buy-Back Announcements


5. EMPIRICAL INVESTIGATION OF BUY-BACK ANNOUNCEMENTS
In this chapter, a detailed empirical investigation of share buy-back announcements in
Germany is provided. Firstly, I carry out an analysis of the abnormal share price behaviour
surrounding share buy-back announcements for the complete sample of the companies and
draw first implications for the validity of the signalling hypothesis. Afterwards, I examine the
price reaction to share buy-back announcements and test the signalling hypothesis by means of
univariate analyses of different parameters which could influence the signalling effect of share
buy-back announcements. Finally, I expand my analyses to a number of additional influencing
variables and carry out a cross-sectional multivariate regression analysis for the entire set of
these variables.

5.1. ANALYSIS OF ABNORMAL RETURNS FOR THE COMPLETE SAMPLE OF SHARE
BUY-BACKS
FIGURE 6 presents the results for the daily average abnormal returns and cumulative abnormal
returns for the full sample of 189 share buy-backs announced between May/1998 – May/2002:

   FIGURE 6: ANALYSIS OF ABNORMAL RETURNS FOR THE COMPLETE SAMPLE OF SHARE BUY-BACKS

           CAR, %                                                                                                        TOTAL SAMPLE
                                                                                                                         TOTAL SAMPLE
          5%                                                                                                             N=189
                                                                                                                         N=189
                                                                                                                         AR [0]
                                                                                                                          AR [0]             =
                                                                                                                                             =    +4.74%
                                                                                                                                                   +4.74%       (( t=
                                                                                                                                                                    t=   +16.37 ))
                                                                                                                                                                          +16.37
                                                                                                                         AR [+1]
                                                                                                                          AR [+1]            =
                                                                                                                                             =    +1.32%
                                                                                                                                                   +1.32%       (( t=
                                                                                                                                                                    t=   +4.57 ))
                                                                                                                                                                          +4.57
                                                                                                                         CAR [-1;+1]
                                                                                                                          CAR [-1;+1]        =
                                                                                                                                             =    +6.00%
                                                                                                                                                   +6.00%       (( t=
                                                                                                                                                                    t=   +11.96 ))
                                                                                                                                                                          +11.96
                                                                                                                         CAR [-1;+5]
                                                                                                                          CAR [-1;+5]        =
                                                                                                                                             =    +6.66%
                                                                                                                                                   +6.66%       (( t=
                                                                                                                                                                    t=   +8.69 ))
                                                                                                                                                                          +8.69
                                                                                                                         CAR [-30;-2]
                                                                                                                          CAR [-30;-2]       =
                                                                                                                                             =    -9.14%
                                                                                                                                                   -9.14%       (( t=
                                                                                                                                                                    t=   -5.86 ))
                                                                                                                                                                          -5.86
                                                                                                                         CAR [-60;-2]
                                                                                                                          CAR [-60;-2]       =
                                                                                                                                             =    -11.52%
                                                                                                                                                   -11.52%      (( t=
                                                                                                                                                                    t=   -5.18 ))
                                                                                                                                                                          -5.18
          0%
                                                                                             -8
                                                                                                  -4
               -60
                     -56
                           -52
                                 -48
                                       -44
                                             -40
                                                   -36
                                                         -32
                                                               -28
                                                                     -24
                                                                           -20
                                                                                 -16
                                                                                       -12




                                                                                                       0
                                                                                                           4
                                                                                                               8
                                                                                                                   12
                                                                                                                        16
                                                                                                                             20
                                                                                                                                  24
                                                                                                                                       28
                                                                                                                                            32
                                                                                                                                                 36
                                                                                                                                                      40
                                                                                                                                                           44
                                                                                                                                                                48
                                                                                                                                                                     52
                                                                                                                                                                           56
                                                                                                                                                                                 60
                                                                                                                                                                          days




         -5%




        -10%




        -15%



 Source: own investigation




                                                                                                                                                                                      38
5. Empirical Investigation of Buy-Back Announcements

    The announcements of share buy-backs have significant positive effects on share prices of
     the announcing companies. The share price improvement amounts to AR [0] = +4.74%
     immediately upon the buy-back announcement (significant at the 99% confidence level and
     better) and on the day following the announcement day, AR [+1] =+1.32% (significant at
     the 99% confidence level and better).

    The cumulative abnormal return CAR [-1;+1] equalled +6.00% and CAR [-1;+5] reached
     the highest level with +6.66%, both figures are significant at the 99% confidence level and
     better. Thus, the results are not being driven by a few large positive abnormal returns.

    The sample companies feature an overall significant underperforming trend over the time
     period [-60;+60] - understandably except for the event window.

    The origins of this overall negative trend are difficult to explain at this stage. They will be
     analysed further below in more detail. But intuitively, this negative tendency might be
     partially accounted for by the overall stock market downturn over the past year94.

    Starting from the day [-31], an increased underperformance can be observed which persists
     until the announcement day [0]. Cumulated over the time period [-30;-2] this
     underperformance amounts to CAR [-30;-2]95 =-9.14% (significant at the 99% confidence
     level and better). The underperformance has been largely set off by the share buy-back
     announcement. Interestingly, however, this just helped the price to return to the trend level
     prior to the increased underperformance period (see the red line). Afterwards, the
     underperforming trend continued through the period [+6;+60].

    The „come back‟ to the previous trend level remains persistent 96. This fact provides strong
     evidence that the sample companies were „mis-priced‟ by the capital markets. Thus, the
     buy-back announcement is considered as signal to the capital market which returns the
     company value to its „correct‟ level.

    Thus, the evidence from the analysis of the complete sample provides a strong support of
     the signalling/undervaluation hypothesis.




94 The existence of a persistent (negative) trend cannot be eliminated through the implementation of the market
model. The market model solely re-models the normal trend of the share price through the event day (as if no
announcement were made) and it is not the purpose of the market model to „take-out‟ the overall trend from the
price development of the company.
95 For the purposes of better comparability with the analyses that will follow, the time period [-30;-2] is used

instead.
96 This could probably be more visible if there were no negative trends during [-60;+60].


                                                                                                             39
5. Empirical Investigation of Buy-Back Announcements


5.2. DETAILED INVESTIGATION – UNIVARIATE ANALYSES
In this section, I will examine the relation between abnormal returns surrounding share buy-
back announcements and different parameter-factors which could possibly influence the
magnitude of the abnormal returns. Those factors, which are analysed here, are (1) Motives for
share buy-backs as stated by the company, (2) Market-to-book value ratio, (3) Company size,
(4) Industrial affiliation, (5) Exchange listing sector, (6) Cumulative returns prior to a buy-back
announcement, (7) Percentage of outstanding shares repurchased, (8) Financial position of the
company, (9) Company manager-ownerships and (10) Company shareholder control
structures.

In the buy-back research literature, these factors have been most frequently used for a detailed
investigation of the abnormal price behaviour surrounding share buy-backs. As stated by many
researchers (such as Vermaelen (1981), Zhang (2000), Lamba and Ramsay (2000)), these
parameters could possibly influence the signal sent to the capital market by a share buy-back
announcement so they provide a nice opportunity to test the signalling hypothesis in more
detail (see p. 19) 97.

The following investigation is primarily carried out this by means of a univariate analysis, i.e. I
investigated the influence of just one factor in one particular case. Such an analysis is valuable
in the respect that it helps to distinguish, step by step, the major firm‟s characteristics
influencing the share price behaviour surrounding buy-back announcements. Furthermore,
investigation of just one factor can help abstract from many other confounding firm‟s
characteristics.

However, the major disadvantage of such a method is that it does not take into consideration
the possibility of strong correlation between the tested firms characteristics, such as e.g. firm
size and exchange listing sector98. Thus, the univariate analyses will be supplemented (and
further expanded) by the multivariate analysis which follows subsequently in section 5.3.




97 The choice of the variables above is largely based on a comprehensive analysis of the prior empirical buy-back
research. The MTB ratio as well as cumulative returns prior to buy-back announcement are taken as proxies for a possible
undervaluation of the firm; company size, industrial affiliation, exchange listing sector, financial position of the company are
possible proxies for information asymmetries existing between the company and the capital market; motives for
share buy-backs as stated by the company, percentage of outstanding shares repurchased, company manager-ownership and company
shareholder control structures could possibly express the credibility of the signal sent to the capital market via
announcement of a share buy-backs.
98 It is observed that the NM companies are on average smaller than the average of CDAX companies.


                                                                                                                            40
5. Empirical Investigation of Buy-Back Announcements

5.2.1. Analysis of Abnormal Returns by Motives as Stated by Companies

In the US, managers are not legally required to disclose their motives for share repurchases 99
and in most cases they do not explicitly reveal them when they announce stock repurchases.
Any attempt to directly test the signalling hypothesis in the US is therefore blurred by the lack
of disclosure of the motive100. In contrast to the US, it is general practice in Germany to reveal
the motives of the share buy-backs in the share buy-back announcements, although there is no
explicit legal obligation to do so. Therefore, the German share buy-back environment provides
a unique data set to test the undervaluation hypothesis.

The motives for share repurchases in Germany have been sourced from the buy-back ad-hoc
announcements made by German companies during the time period between May/1998 and
May/2002. This resulted in a data set containing some 148 data entries (not all the companies
revealed the motives for their buy-back). The vast majority of the sample companies have
given several motives in one announcement, on average 2-3, and even up to 5 motives. So I
constructed a dummy variable for testing the undervaluation hypothesis. This variable equalled
1 if at least one of the stated motives was undervaluation101 and 0 otherwise.

I analysed the data set by means of linear regression. The dummy variable (independent
variable) was regressed against several measures of abnormal returns as outlined in FIGURE 7
below:




99 See e.g. [Evans et al, 2000].
100 See [Otchere/Ross, 2000], p. 3.
101 The undervaluation was in selected cases expressed indirectly, like „good company prospects‟ or share price

stabilisation. However, in most of the cases it was clearly stated that the company is buying its own shares for
reasons of unfair undervaluing by the market.

                                                                                                             41
5. Empirical Investigation of Buy-Back Announcements


  FIGURE 7: ANALYSIS OF ABNORMAL RETURNS BY MOTIVES AS STATED BY COMPANIES

    Dependent Variables
           AR [0]                                    Coefficients        Standard Error   t-Statistics        P-Value
                                 Intercept                  0.034                 0.011           3.14          0.002
                                 x Variable                 0.037                 0.015           2.40   **     0.018


           CAR [-1;+1]                               Coefficients        Standard Error   t-Statistics        P-Value
                                 Intercept                  0.036                 0.013           2.64          0.009
                                 x Variable                 0.058                 0.019           3.06 ***      0.003


           CAR [-1;+5]                               Coefficients        Standard Error   t-Statistics        P-Value
                                 Intercept                  0.044                 0.019           2.34          0.021
                                 x Variable                 0.064                 0.026           2.45   **     0.015


           CAR [+1;+30]                              Coefficients        Standard Error   t-Statistics        P-Value
                                 Intercept                 -0.011                 0.028          -0.40          0.688
                                 x Variable                 0.015                 0.039           0.40    *     0.692


           CAR [+1;+60]                              Coefficients        Standard Error   t-Statistics        P-Value
                                 Intercept                 -0.034                 0.036          -0.95          0.342
                                 x Variable                 0.053                 0.050           1.06    *     0.291


                            *** - Significant at the 99% confidence level and better
                             ** - Significant at the 98% confidence level
                              * - Not statistically significant

Source: own investigation


     As we can see, the abnormal returns surrounding the day of a buy-back announcement
      were highly dependent on the motives stated by the repurchasing companies. The relation
      with AR [0] and CAR [-1;+5] was statistically significant at the 98% confidence level and
      with CAR [-1;+1] even on the level better than 99%.

     Thus, the capital market reacted more positively to those share buy-backs where
      undervaluation was mentioned among the stated motives.

     Over the periods [+1;+30] and [+1;+60], no significant over-performance of the
      undervalued companies has been found. However, a far more important result of the
      regression with CAR [+1;+30] and CAR [+1;+60] is that no under-performance of the
      undervalued companies has been found for these periods either.

     That means that the higher reaction on [0]-[-1;+5] was not a „one day effect‟ and as it
      seems the capital market has put real confidence in the motives which were stated in the
      share buy-back announcements. This evidence also supports the hypothesis that the
      repurchasing companies were undervalued.

                                                                                                                        42
5. Empirical Investigation of Buy-Back Announcements

5.2.2. Analysis of Abnormal Returns by Market-to-Book Value Ratio

According to Ikenberry et al (1995), repurchasing firms with lower MTB (Market-to-Book)
ratios are more likely to be undervalued102. Thus, testing the relationship between MTB and
announcement period returns would provide a unique opportunity to test the
signalling/undervaluation hypothesis.

In order to test the validity of this hypothesis for the German sample by means of MTB, I
formed a high and a low-valued portfolio of companies (based on the MTB ratio). MTB ratios
were calculated by dividing the market cap of the company (average for five days prior to the
buy-back announcement) by the book values of equity which were sourced from the
Bloomberg database103. Afterwards, the average MTB ratio for the entire sample was calculated
(MTBavg = 2.05). Thus, the „undervalued‟ portfolio included firms with MTB<2.05 (109
announcements, MTB ranging from 0.17 to 2.03), and the „overvalued‟ portfolio included
firms with MTB>2.05 (80 announcements, MTB ranging from 2.06 to 36.92). The
comparative analysis of the two sub-samples is presented in FIGURE 8 below:




102 See [Ikenberry et al, 1995], p. 21. Here, the MTB ratio is possibly taken as the most universal valuation
multiple. Since the sample companies feature strong differences according to e.g. their sizes or industries, other
multiples such as EBIT, EBITDA, sales, P/E would be less applicable.
103 For every data set entry from the Main Sample (189). The next available figures prior to announcement have

been used.

                                                                                                               43
5. Empirical Investigation of Buy-Back Announcements


    FIGURE 8: ANALYSIS OF ABNORMAL RETURNS BY MTB RATIO

     CAR, %                                                                                                       TOTAL SAM PLE
                                                                                                                  TOTAL SAM PLE
                                                                                                                  N =189
                                                                                                                  N =189
      5%                                                                                                          AR [0]
                                                                                                                  AR [0]               =
                                                                                                                                       =    +4.74%
                                                                                                                                             +4.74%                   ((   t=
                                                                                                                                                                            t=   +16.37 ))
                                                                                                                                                                                  +16.37
                                                                                                                  AR [+1]
                                                                                                                  AR [+1]              =
                                                                                                                                       =    +1.32%
                                                                                                                                             +1.32%                   ((   t=
                                                                                                                                                                            t=   +4.57 ))
                                                                                                                                                                                  +4.57
                                                                                                                  CAR [-1;+1]
                                                                                                                  CAR [-1;+1]          =
                                                                                                                                       =    +6.00%
                                                                                                                                             +6.00%                   ((   t=
                                                                                                                                                                            t=   +11.96 ))
                                                                                                                                                                                  +11.96
                                                                                                                  CAR [-1;+5]
                                                                                                                  CAR [-1;+5]          =
                                                                                                                                       =    +6.66%
                                                                                                                                             +6.66%                   ((   t=
                                                                                                                                                                            t=   +8.69 ))
                                                                                                                                                                                  +8.69
                                                                                                                  CAR [-30;-2]
                                                                                                                  CAR [-30;-2]         =
                                                                                                                                       =    -9.14%
                                                                                                                                             -9.14%                   ((   t=
                                                                                                                                                                            t=   -5.86 ))
                                                                                                                                                                                  -5.86
                                                                                                                  CAR [-60;-2]
                                                                                                                  CAR [-60;-2]         =
                                                                                                                                       =    -11.52%
                                                                                                                                             -11.52%                  ((   t=
                                                                                                                                                                            t=   -5.18 ))
                                                                                                                                                                                  -5.18
      0%




                                                                                                -5
            -60

                  -55

                        -50

                              -45

                                        -40

                                                  -35

                                                        -30

                                                                  -25

                                                                           -20

                                                                                    -15

                                                                                          -10



                                                                                                     0

                                                                                                         5

                                                                                                             10

                                                                                                                  15

                                                                                                                       20

                                                                                                                              25

                                                                                                                                      30

                                                                                                                                            35

                                                                                                                                                    40

                                                                                                                                                              45

                                                                                                                                                                           50

                                                                                                                                                                                  55

                                                                                                                                                                                       60
                                                                                                                       'UNDERVALUED' SAMPLE
      -5%                                                                                                              N=109
                                                                                                                       AR [0]          =   +5.79%   (   t=   +14.88    )
                                                                                                                       AR [+1]         =   +1.30%   (   t=   +3.35     )
                                                                                                                       CAR [-1;+1]     =   +7.11%   (   t=   +10.54    )
                         'OVERVALUED' SAMPLE                                                                           CAR [-1;+5]     =   +8.03%   (   t=   +7.79     )
                         N=80                                                                                          CAR [-30;-2]    =   -7.96%   (   t=   -3.80     )
                                                                                                                       CAR [-60;-2]    =   -9.95%   (   t=   -3.33     )
     -10%                AR [0]               =   +3.31%      (   t=    +8.35   )
                         AR [+1]              =   +1.35%      (   t=    +3.41   )
                         CAR [-1;+1]          =   +4.48%      (   t=    +6.54   )
                         CAR [-1;+5]          =   +4.80%      (   t=    +4.58   )
                         CAR [-30;-2]         =   -10.74%     (   t=    -5.03   )
                         CAR [-60;-2]         =   -13.65%     (   t=    -4.49   )

     -15%

                                    Undervalued Companies
                                    (MTB<2.05)
                                    Overvalued Companies
                                    (MTB>2.05)
     -20%



 Source: own investigation


    As we can see, both sub-samples experienced statistically significant abnormal returns
     surrounding the buy-back announcement. The „undervalued‟ sample featured on average
     higher abnormal returns with AR [0] = 5.79%, CAR [-1;+1] = 7.11% and CAR [-1;+5] =
     8.03% - in contrast to the „overvalued‟ sample with AR [0] = 3.31%, CAR [-1;+1] = 4.48%
     and CAR [-1;+5] = 4.80%.

    The differences in market reaction between the two samples were tested by means of t-
     statistic and have been found statistically significant at AR [0], CAR [-1;+1] which were
     different at the confidence levels of 95%, 90% respectively.

    Thus, this evidence supports the hypothesis that the undervalued companies (i.e. those
     with lower MTB ratios) have a more positive signalling effect by the share buy-back
     announcement, i.e. via share buy-backs, they send to the capital market a stronger signal
     that the real motive of their share repurchases is undervaluation. Thus, this implication
     strongly supports the undervaluation hypothesis.

In addition I carried out a regression analysis in which the relation between the magnitude of
the MTB ratio and the abnormal returns in various event windows (i.e. AR [0], CAR [-1;+1],

                                                                                                                                                                                             44
5. Empirical Investigation of Buy-Back Announcements

CAR [-1;+5], as well as CAR [-60;-2] and CAR [-30;-2]) was examined. The results of this
analysis is shown in FIGURE 9 below:

    FIGURE 9: REGRESSION ANALYSIS OF ABNORMAL RETURNS BY MTB RATIO

    Dependent Variables
            AR [0]                            Coefficients        Standard Error             t-Statistics        P-Value
                             Intercept               0.058                 0.008                     7.18          0.000
                             x Variable             -0.004                 0.002                    -2.17 **       0.031


            CAR [-1;+1]                       Coefficients        Standard Error             t-Statistics        P-Value
                             Intercept               0.076                 0.010                     7.67          0.000
                             x Variable             -0.006                 0.002                    -2.66 ****     0.008


            CAR [-1;+5]                       Coefficients        Standard Error             t-Statistics        P-Value
                             Intercept               0.086                 0.014                     6.30          0.000
                             x Variable             -0.007                 0.003                    -2.38 ***      0.018


            CAR [-30;-2]                      Coefficients        Standard Error             t-Statistics        P-Value
                             Intercept              -0.107                 0.021                    -5.03          0.000
                             x Variable              0.003                 0.004                     0.67 *        0.506


            CAR [-60;-2]                      Coefficients        Standard Error             t-Statistics        P-Value
                             Intercept              -0.124                 0.030                    -4.14          0.000
                             x Variable              0.003                 0.006                     0.48 *        0.631

                             ****         - Significant at the 99% confidence level and better
                             ***          - Significant at the 98% confidence level
                             **           - Significant at the 95% confidence level
                             *            - Not statistically significant

 Source: own investigation


    As we can see there is a highly statistically significant negative relation between the MTB
     ratios and the proxies for abnormal returns such as AR [0], CAR [-1;+1], CAR [-1;+5], at
     levels of confidence 95%, 99% and 98% respectively.

    Thus, both analyses carried out above provided strong evidence consistent with the
     signalling/undervaluation hypothesis: the firms with lower MTB ratios, i.e. those firms
     which are rather undervalued, have in tendency experienced higher abnormal returns
     surrounding the announcement.


5.2.3. Analysis of Abnormal Returns by Company Size

Smaller firms receive less attention from analysts and are therefore more likely to be mis-priced
by the capital market. As e.g. Ikenberry et al (1995) argue, the announcement period returns


                                                                                                                           45
5. Empirical Investigation of Buy-Back Announcements

are related to company size decile. The authors suggested that company size can be a proxy for
information asymmetries104. Therefore, the market should react more favourably to the buy-
back announcements made by smaller firms.

First, I analysed my sample according to the size of the companies in it and found it very
skewed to the right – which can be seen in FIGURE 10 below. The ten largest companies in the
sample account for some 78% of the total enterprise value of all companies in the sample.
Thus, as we can see, the skewed distribution of the enterprise values of the companies in the
sample can be better described by a parabolic function of a kind y=xn.


      FIGURE 10: DISTRIBUTION OF THE COMPANY SAMPLE BY SIZE
 Distribution of the firms with market value of under €500m. The negative enterprise                             Distribution of the firms with market value of more
 values result when cash positions exceed the capitalisation of the equity.                                      than €500m. The lowest firm size on this chart
                                                                                                                 corresponds to the highest firm size on the left chart!

      500

                                                                                                                   90,000
      400

                                                                                                                   70,000
      300

                                                                                                                   50,000
      200

                                                                                                                   30,000
      100

                                                                                                                   10,000
        0
             1   6   11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106 111 116 121 126 131 136 141




                                                                                                                             1
                                                                                                                                 5
                                                                                                                                     9
                                                                                                                                         13
                                                                                                                                              17
                                                                                                                                                   21
                                                                                                                                                        25
                                                                                                                                                             29
                                                                                                                                                                  33
                                                                                                                                                                       37
                                                                                                                                                                            41
                                                                                                                                                                                 45
                                                                                                                   -10,000
      -100


 Source: own investigation

Thereafter, I divided the whole sample of 189 companies into 2 equal groups: large and small
firms according to their enterprise values prior to the announcement day105,106.

The comparative analysis of the two sub-samples is presented in FIGURE 11 below:




104 See [Ikenberry et al, 1995], p. 201
105 For my investigation, the enterprise value seems to be the most universal proxy for the company size, as it
includes both components of funding capital – equity and loans, and cash is excluded from the value as a non
value creating component. Other proxies like total sales, number of employees, total assets or book capitalisation
would not be as good, given the industrial inhomogeneousness of the total sample.
106 Different from the major empirical studies which used equity market cap, I took the enterprise value as a

measure of company size. Besides the equity capitalisation, enterprise value also includes net debt and minority
interests. Net debt has been calculated by subtracting cash and cash equivalents from the total interest bearing
debt. The data has been sourced from the next available financials reports released before the announcement day.
The equity value was calculated as an average over the days [-6;-2]. In my view, the inclusion of net debt, as well
as minority interests (both are not reflected in the equity value) provides a better proxy for „the true‟ company
size.

                                                                                                                                                                        46
5. Empirical Investigation of Buy-Back Announcements


      FIGURE 11: ANALYSIS OF ABNORMAL RETURNS BY COMPANY SIZE

        CAR, %                  "LARGE" FIRMS
                                 "LARGE" FIRMS                                                                           TOTAL SAMPLE
                                                                                                                         TOTAL SAMPLE
                                N=95
                                 N=95                                                                                    N=189
                                                                                                                         N=189
        5%                      AR [0]
                                AR [0]           =
                                                 =   +2.40%
                                                      +2.40%   ((   t=
                                                                     t=   +6.86
                                                                           +6.86   ))                                    AR [0]
                                                                                                                         AR [0]                =
                                                                                                                                               =   +4.74%
                                                                                                                                                    +4.74%        ( ( t=
                                                                                                                                                                       t=   +16.37 ) )
                                                                                                                                                                             +16.37
                                AR [+1]
                                AR [+1]          =
                                                 =   +0.98%
                                                      +0.98%   ((   t=
                                                                     t=   +2.81
                                                                           +2.81   ))                                    AR [+1]               =   +1.32%         ( ( t=    +4.57 ) )
                                                                                                                         AR [+1]               =    +1.32%             t=    +4.57
                                CAR [-1;+1]
                                CAR [-1;+1]      =
                                                 =   +3.67%
                                                      +3.67%   ((   t=
                                                                     t=   +6.06
                                                                           +6.06   ))                                    CAR [-1;+1]           =   +6.00%         ( ( t=    +11.96 ) )
                                CAR [-1;+5]      =   +4.14%    ((   t=    +4.48    ))
                                                                                                                         CAR [-1;+1]           =    +6.00%             t=    +11.96
                                CAR [-1;+5]      =    +4.14%         t=    +4.48                                         CAR [-1;+5]
                                                                                                                         CAR [-1;+5]           =
                                                                                                                                               =   +6.66%
                                                                                                                                                    +6.66%        ( ( t=
                                                                                                                                                                       t=   +8.69 ) )
                                                                                                                                                                             +8.69
                                CAR [-30;-2]
                                CAR [-30;-2]     =
                                                 =   -6.80%
                                                      -6.80%   ((   t=
                                                                     t=   -3.61
                                                                           -3.61   ))
                                                                                                                         CAR [-30;-2]
                                                                                                                         CAR [-30;-2]          =
                                                                                                                                               =   -9.14%
                                                                                                                                                    -9.14%        ( ( t=
                                                                                                                                                                       t=   -5.86 ) )
                                                                                                                                                                             -5.86
                                CAR [-60;-2]
                                CAR [-60;-2]     =
                                                 =   -8.00%
                                                      -8.00%   ((   t=
                                                                     t=   -2.98
                                                                           -2.98   ))
                                                                                                                         CAR [-60;-2]
                                                                                                                         CAR [-60;-2]          =
                                                                                                                                               =   -11.52%
                                                                                                                                                    -11.52%       ( ( t=
                                                                                                                                                                       t=   -5.18 ) )
                                                                                                                                                                             -5.18
        0%
             -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10                                -5   0    5    10       15   20      25     30        35    40       45   50        55     60
                                                                                                                                                                              days


       -5%




      -10%
                                                                                                 "SMALL" FIRMS
                                                                                                  "SMALL" FIRMS
                                                                                                 N=94
                                                                                                  N=94
                       Large                                                                     AR [0]
                                                                                                 AR [0]              =
                                                                                                                     =   +7.11%
                                                                                                                          +7.11%    ( ( t=
                                                                                                                                         t=   +15.36
                                                                                                                                               +15.36   ))
                                                                                                 AR [+1]
                                                                                                 AR [+1]             =
                                                                                                                     =   +1.67%
                                                                                                                          +1.67%    ( ( t=
                                                                                                                                         t=   +3.61
                                                                                                                                               +3.61    ))
                       Small                                                                     CAR [-1;+1]
                                                                                                 CAR [-1;+1]         =
                                                                                                                     =   +8.35%
                                                                                                                          +8.35%    ( ( t=
                                                                                                                                         t=   +10.42
                                                                                                                                               +10.42   ))
                                                                                                 CAR [-1;+5]
                                                                                                 CAR [-1;+5]         =
                                                                                                                     =   +9.20%
                                                                                                                          +9.20%    ( ( t=
                                                                                                                                         t=   +7.52
                                                                                                                                               +7.52    ))
      -15%                                                                                       CAR [-30;-2]        =   -11.50%    ( ( t=    -4.61     ))
                                                                                                 CAR [-30;-2]        =    -11.50%        t=    -4.61
                                                                                                 CAR [-60;-2]
                                                                                                 CAR [-60;-2]        =
                                                                                                                     =   -15.08%
                                                                                                                          -15.08%   ( ( t=
                                                                                                                                         t=   -4.24
                                                                                                                                               -4.24    ))




      -20%



    Source: own investigation


     Small firms had higher price run-ups on the day [0] and higher cumulative returns around it
      - with AR[0] = +7.11%, CAR [-1;+1] = +8.35% and CAR [-1;+5] = +9.20%. All the
      figures are significant at the 99% confidence level and better.

     Larger companies feature on average significantly lower abnormal returns surrounding the
      announcement day – with AR [0] = +2.40%, CAR [-1;+1] = +3.67%, CAR [-1;+5] =
      +4.14%. All these figures are significant at the 99% confidence level and better.

     The difference between the sub-samples appear to be also highly statistically significant at
      the 99% confidence level and better.

     These results strongly support the signalling hypothesis – the stronger the information
      asymmetries between the company and the capital markets, the stronger is the positive
      signal to the capital market that the company is undervalued.

     Moreover, smaller companies experienced a higher cumulated underperformance prior to
      the share buy-back announcement: CAR [-30;-2] =-11.50%, CAR [-60;-2] =-15.08%
      compared with the sub-sample of large companies CAR [-30;-2] =-6.80%, CAR [-60;-2] =-
      8.00%. However, the differences between the two sub-samples featured only weak

                                                                                                                                                                                         47
5. Empirical Investigation of Buy-Back Announcements

       statistical significance at levels not higher than 84% and 86% for CAR [-30;-2] and CAR [-
       60;-2] respectively.

      This higher underperformance of smaller companies prior to share buy-backs supports the
       thesis that the smaller companies are more likely to be incorrectly undervalued by the
       capital market. The share buy-back represents in this case a positive correction of the share
       price that permanently returns to a higher level.

In addition to the analysis of announcements effects by size provided above, I further divided
the whole sample into ten equal groups107 (à 19 companies in the first nine groups and 18 in
the tenth) in order to more deeply investigate the influence of company size decile on the
abnormal announcement returns. Thus, the following sub-samples have been created:



               FIGURE 12: DECILE SUB-SAMPLES

                                           No. of         ENTERPRISE VALUE
                     Decile              Companies        From          To
                        1                            19         -39           19
                        2                            19          19           36
                        3                            19          37           54
                        4                            19          55           80
                        5                            19          85          133
                        6                            19         133          236
                        7                            19         236          371
                        8                            19         382          771
                        9                            19         865         4192
                       10                            18        4586        96827


             Source: own investigation




The results of the decile group analysis are visualised in FIGURE 13 below:




107   Like suggested in [Vermaelen, 1981], p. 163.

                                                                                                 48
5. Empirical Investigation of Buy-Back Announcements


      FIGURE 13: DECILE ANALYSIS OF ANNOUNCEMENT RETURNS BY COMPANY SIZE

      CAR, %
      30%
                                      DECILES                 AR [0]                    CAR [-1;+1]           CAR [-1;+5]
                                                  1                 14.80%                     15.75%              18.69%
                                                                      11.01 ***                   6.76 ***            5.68 ***
                                                 2                   8.29%                      9.68%               8.68%
                                                                       7.21 ***                   4.86 ***            3.08 ***
      25%                                        3                   5.16%                      5.82%               7.63%
                                                                       5.22 ***                   3.40 ***            3.15 ***
                                                 4                   4.38%                       5.14%              5.41%                AR [0]
                                                                       4.12 ***                    2.79 ***           2.07 **
                                                 5                   2.21%                      5.37%               5.31%                CAR [-1;+1]
                                                                       2.64 ***                   3.70 ***            2.58 ***
      20%                                                                                                                                CAR [-1;+5]
                                                 6                   4.88%                      5.60%               5.10%
                                                                       4.82 ***                   3.19 ***            2.06 **
                                                 7                   2.98%                      4.59%               5.60%
                                                                       3.33 ***                   2.96 ***            2.56 ***
                                                 8                   3.00%                      6.09%               8.80%
      15%                                                              4.14 ***                   4.85 ***            4.96 ***
                                                 9                   0.75%                      0.69%              -0.65%
                                                                       1.09 *                     0.58 *              -0.39 *
                                                10                   0.75%                       1.01%              1.65%
                                                                       1.39 *                      1.08 *             1.25 *
                                     t-statistics in italics
      10%                                               *** - significant at levels 99% and better
                                                        ** - significant at levels 95% and better
                                                        * - not statistically significant



       5%



       0%
               1             2   3               4                        5                          6              7            8   9           10
                                                                                DECILES


      -5%


 Source: own investigation


      Over the ten groups of companies, there is a clear negative interrelation between the size
       of the company (decile number) and abnormal returns AR [0], CAR [-1;+1] and CAR [-
       1;+5]. This relation is statistically highly significant – tested both over the ten groups with
       aggregated returns of the sub-sample (t=-4.66) and over the whole sample of companies
       (t=-4.68)108.

      The relation is even non-linear, and rather hyperbolic, i.e. better described by a formula
       like y=1/xn. This is particularly interesting in the light of the parabolic distribution of the
       enterprise values of the sample companies (see FIGURE 10). Thus, a non-linear regression
       would probably yield some interesting results in this case109.

      Almost all abnormal returns AR (0), CAR [-1;+1] and CAR [-1;+5] for the deciles 1-8 are
       statistically significant at the 99% confidence level and better. However, no statistically
       significant returns were found for the deciles 9 and 10, i.e. the largest companies of the
       sample.



108 The t-values given here refer to CAR [-1;+1]. The significance of AR [0] and CAR [-1;+5] was approximately
at the same confidence level.
109 Not carried out in this paper.


                                                                                                                                                       49
5. Empirical Investigation of Buy-Back Announcements

     Thus, a significant negative relation between the firm size (in this case decile number) and
      the abnormal price return is confirmed by this analysis.



FIGURE 31, (which is separately presented in Appendix III)110 visualises the developments of
cumulative abnormal returns of the ten decile sub-samples. This analysis also yielded some
very interesting results:

     The smallest companies of the first decile had the highest announcement returns with AR
      [0] = +14.80%, CAR [-1;+1] = +15.75% and CAR [-1;+5] = +18.69%. Thus, this was the
      only decile which experienced a turn-around in trend upon the share buy-back
      announcement.

     Thus, this group of companies had been „truly‟ undervalued and had been left without
      attention by the capital markets111. In these cases, the share buy-backs have been a very
      effective means of drawing the attention of investors to the stock and managing the share
      price of the company.

     The majority of the decile groups (i.e. deciles 3,4,5,6,7 and 9) experienced underperforming
      trends over the observation period of [-60;+60].

     The inclusion of these deciles in the total sample is apparently the reason why the total
      sample underperformed (this problem was discussed on the page 38).

     The decile 8 over-performed through the period of [-60;+60] which breaks the ranks of the
      overall underperforming picture (particularly when compared with deciles 7 and 9).
      However, this positive trend in the decile was largely accounted for by strong run-ups in
      prices of Entrium Direct Bankers AG and Senator Film AG during the time period [-60;-
      10] of some 89% and 72% respectively112.

     No significant underperformance can be observed for the decile 10. This decile accounts,
      however, for some 90% of the enterprise value of the total sample of 189 companies.

     Thus, knowing this, the overall underperformance of the total sample (this problem was
      discussed on the page 38) can be explained rather by the prevailing underperforming
      trends in prices for smaller companies in general, compared with the „blue chips‟.



110 The figure was placed in appendix mainly for reasons of its large size.
111 We can also say the same about the companies of the decile 2, however, the positive CARs upon the buy-back
announcement are not significant.
112 An analysis of this sub-sample excluding these two firms turned around the trend to underperformance.


                                                                                                           50
5. Empirical Investigation of Buy-Back Announcements

     The largest companies in the sample (i.e. deciles 9 and 10) experienced no significant price
      run-ups upon the share buy-back announcements (in line with the results presented in
      FIGURE 13 above).


5.2.4. Analysis of Abnormal Returns by Industrial Classification of Share Buy-Backs

Analyses based on industry classification of the sample companies advocate the idea that the
price reaction to a share buy-back announcement could be a subject to the industry affiliation
of the companies. The industry groups are more homogeneous according to their
characteristics such as average company size, shareholder structures, operating data and
financial ratios. The following analysis examines whether the market‟s reaction to German
companies announcing share buy-backs is related to their industrial affiliation.

I divided the total sample into the major three categories: Manufacturing, Financial Services
and Non-Financial Services113,114. This distribution is visualised in FIGURE 14 below:

             FIGURE 14: DISTRIBUTION OF THE TOTAL SAMPLE ACCORDING TO INDUSTRIAL
             CLASSIFICATION OF COMPANIES


                             Financial Services
                             13%
                             (25 companies)

                                                                            Manufacturing
                                                                            61%
                             Non-Financial                                  (114 Companies)
                             Services
                             26%
                             (50 Companies)




           Source: own investigation


This is however, a very rough classification. However, a more deep classification would
probably further liquefy the sizes of the sub-samples. The results of the analysis are visualised
in FIGURE 15 below:

113 Such a rough classification is applicable here, see [Lamba/Ramsay, 2000], p.19. A more detailed split could also
be imaginable. However, the sub-samples would be far to small.
114 The sample companies have been classified in the following way: Manufacturing: included, inter alia, industrial

products, chemicals, IT/software (non-services), utilities, food products, building construction, building materials,
and printing; Non-Financial Services: included shipping, transportation, trucking, logistics, retail/wholesale,
media/publishing, telecommunication, various business services (including IT business services) and consulting;
Financial Services: included banking, investment banking, mortgage banking, real estate, brokerage services,
insurance and insurance broking. In the case of a conglomerate, the prevailing sector was decisive for the final
classification of the company.

                                                                                                                 51
5. Empirical Investigation of Buy-Back Announcements


      FIGURE 15: ANALYSIS OF ABNORMAL RETURNS BY INDUSTRIAL CLASSIFICATION OF SHARE BUYBACKS

      CAR, %                                                                                                                                   TOTAL SAMPLE
                                                                                                                                                TOTAL SAMPLE
                                                                                                                                               N=189
      10%                                                                                                                                       N=189
                                                                                                                                               AR [0]
                                                                                                                                                AR [0]                 ==    +4.74%
                                                                                                                                                                              +4.74%            ( ( t=
                                                                                                                                                                                                     t=   +16.37 ) )
                                                                                                                                                                                                           +16.37
                          FINANCIAL SERVICES
                           FINANCIAL SERVICES                                                                                                  AR [+1]
                                                                                                                                                AR [+1]                ==    +1.32%
                                                                                                                                                                              +1.32%            ( ( t=
                                                                                                                                                                                                     t=   +4.57 ) )
                                                                                                                                                                                                           +4.57
                          N=25
                           N=25                                                                                                                CAR [-1;+1]
                                                                                                                                                CAR [-1;+1]            ==    +6.00%
                                                                                                                                                                              +6.00%            ( ( t=
                                                                                                                                                                                                     t=   +11.96 ) )
                                                                                                                                                                                                           +11.96
                          AR [0]                = +4.19%      ((   t= +6.03    ))                                                              CAR [-1;+5]
                                                                                                                                                CAR [-1;+5]            ==    +6.66%
                                                                                                                                                                              +6.66%            ( ( t=
                                                                                                                                                                                                     t=   +8.69 ) )
                                                                                                                                                                                                           +8.69
                           AR [0]                = +4.19%           t= +6.03                                                                   CAR [-30;-2]            =     -9.14%             ( ( t=    -5.86 ) )
                          AR [+1]               = +1.46%      ((   t= +2.10    ))                                                               CAR [-30;-2]            =     -9.14%                 t=    -5.86
                           AR [+1]               = +1.46%           t= +2.10                                                                   CAR [-60;-2]            =     -11.52%            ( ( t=    -5.18 ) )
                          CAR [-1;+1]
                           CAR [-1;+1]          = +5.17%
                                                 = +5.17%     ((   t= +4.30
                                                                    t= +4.30   ))                                                               CAR [-60;-2]            =     -11.52%                t=    -5.18
       5%                 CAR [-1;+5]           = +6.03%      ((   t= +3.28    ))
                           CAR [-1;+5]           = +6.03%           t= +3.28
                          CAR [-30;-2]
                           CAR [-30;-2]         = -9.29%
                                                 = -9.29%     ((   t= -2.48
                                                                    t= -2.48   ))
                          CAR [-60;-2]
                           CAR [-60;-2]         = -7.44%
                                                 = -7.44%     ((   t= -1.39
                                                                    t= -1.39   ))                                                 MANUFACTURING
                                                                                                                                   MANUFACTURING
                                                                                                                                  N=114
                                                                                                                                   N=114
                                                                                                                                  AR [0]
                                                                                                                                   AR [0]             ==    +5.34%
                                                                                                                                                             +5.34%     ((   t=
                                                                                                                                                                              t=   +14.30 ) )
                                                                                                                                                                                    +14.30
                                                                                                                                  AR [+1]
                                                                                                                                   AR [+1]            ==    +1.44%
                                                                                                                                                             +1.44%     ((   t=
                                                                                                                                                                              t=   +3.85 ) )
                                                                                                                                                                                    +3.85
                                                                                                                                  CAR [-1;+1]
                                                                                                                                   CAR [-1;+1]        ==    +7.32%
                                                                                                                                                             +7.32%     ((   t=
                                                                                                                                                                              t=   +11.32 ) )
                                                                                                                                                                                    +11.32
       0%                                                                                                                         CAR [-1;+5]         =     +7.87%      ((   t=    +7.97 ) )
                                                                                                                                   CAR [-1;+5]         =     +7.87%           t=    +7.97




                                                                                                -5
             -60

                   -55

                         -50

                                -45

                                          -40

                                                  -35

                                                        -30

                                                                    -25

                                                                           -20

                                                                                    -15

                                                                                          -10




                                                                                                     0

                                                                                                            5

                                                                                                                   10

                                                                                                                         15

                                                                                                                                  20

                                                                                                                                         25

                                                                                                                                                   30

                                                                                                                                                             35

                                                                                                                                                                       40

                                                                                                                                                                               45

                                                                                                                                                                                          50

                                                                                                                                                                                                    55

                                                                                                                                                                                                               60
                                                                                                                                  CAR [-30;-2]
                                                                                                                                   CAR [-30;-2]       ==    -9.76%
                                                                                                                                                             -9.76%     ((   t=
                                                                                                                                                                              t=   -4.85 ) )
                                                                                                                                                                                    -4.85
                                                                                                                                  CAR [-60;-2]
                                                                                                                                   CAR [-60;-2]       ==    -11.01%
                                                                                                                                                             -11.01%    ((   t=
                                                                                                                                                                              t=   -3.84 ) )
                                                                                                                                                                                    -3.84                 days


      -5%




      -10%

                               Financial
                               Services
                                                                                                         NON-FINANCIAL SERVICES
                               Non-Financial Services                                                     NON-FINANCIAL SERVICES
                                                                                                         N=50
                                                                                                          N=50
      -15%
                                                                                                         AR [0]
                                                                                                          AR [0]         ==   +4.23%
                                                                                                                               +4.23%    ( ( t= +6.54 ) )
                                                                                                                                              t= +6.54
                               Manufacturing                                                             AR [+1]
                                                                                                          AR [+1]        ==   +0.82%
                                                                                                                               +0.82%    ( ( t= +1.26 ) )
                                                                                                                                              t= +1.26
                                                                                                         CAR [-1;+1]
                                                                                                          CAR [-1;+1]    ==   +3.78%
                                                                                                                               +3.78%    ( ( t= +3.37 ) )
                                                                                                                                              t= +3.37
                                                                                                         CAR [-1;+5]
                                                                                                          CAR [-1;+5]    ==   +4.89%
                                                                                                                               +4.89%    ( ( t= +2.86 ) )
                                                                                                                                              t= +2.86
                                                                                                         CAR [-30;-2]
                                                                                                          CAR [-30;-2]   ==   -6.41%
                                                                                                                               -6.41%    ( ( t= -1.84 ) )
                                                                                                                                              t= -1.84
                                                                                                         CAR [-60;-2]
                                                                                                          CAR [-60;-2]   ==   -12.75%
                                                                                                                               -12.75%   ( ( t= -2.56 ) )
                                                                                                                                              t= -2.56
      -20%



 Source: own investigation


      Manufacturing companies feature the highest reaction of AR[0]=+5.34% on the
       announcement day and a CAR [-1;+1]=+7.32%. The abnormal returns in the Financial
       Services industry, as well as Non-Financial Services are below the average, both in terms of
       AR(0) and CAR [-1;+5].

      The Non-Financial Services industry reveals the lowest announcement results with
       AR[0]=+4.23%, CAR [-1;+1]=+3.78% and CAR [-1;+5]=+4.89%. However, the group
       experiences further positive price run-ups after the day [+5]. So for example, the CARs [-
       1;+7] reaches +5.57%.

      I tested the statistical significance of the differences between the samples by means of t-
       statistics115. Only for the proxy CAR [-1;+1], evidence has been found only for the
       manufacturing companies that their price reaction is a significantly different from other
       companies (here, a positive industry impact at the significance level of 95% has been
       found) and for the non-financial services companies that their price reaction is a


115In every case the bilateral differences between the industry sample under investigation (dummy equalling 1)
and the sample containing the other two industries (dummy variable equalling 0) has been tested for significance.

                                                                                                                                                                                                                       52
5. Empirical Investigation of Buy-Back Announcements

    significantly different from other companies (here, a negative industry impact at the
    significance level of 90% has been found). However, when tested for other proxies, i.e.
    AR[0] and CAR [-1;+5], no differences between industries with respect to their reaction to
    share buy-back announcements has been detected.

   Thus, I would rather suggest that there is no strong evidence that the factor of industry
    affiliation conveys any additional information which could significantly influence the
    market perception of share buy-back announcements.


5.2.5. Analysis of Abnormal Returns by Exchange Listing Sector

I split the whole sample of 189 firms into the Neuer Markt (NM) (71 firms) and Non-NM (118
firms). The abnormal returns were calculated for the two samples separately using the same
market model.

As one possibly would expect, the positive reaction on share buy-back announcements made
by the NM companies should be higher, primarily due to the fact that higher information
asymmetries between these companies and the capital market exists:

   NM companies have on average, a higher risk class – due to the simplified admission
    standards for the listing on the Neuer Markt trading segment. Moreover, the models of
    NM companies are attached with more business risk, as they are rather innovation driven
    start-up businesses.

   NM companies are, on average, of smaller sizes. As we have seen above and as
    documented by many other empirical researchers, smaller companies are less covered by
    analysts and are less known to the market.

   Furthermore, NM companies have higher concentrated levels of management control and
    management shareholdings. These factors also prop up the abnormal returns, as the signals
    made by these companies might be considered more credible.




                                                                                           53
5. Empirical Investigation of Buy-Back Announcements


FIGURE 16 below visualises the differences between the two sub-samples of companies:

    FIGURE 16: ANALYSIS OF ABNORMAL RETURNS BY EXCHANGE LISTING SECTOR
    (NON-NM VS. NM COMPANIES)

     CAR, %                                                                                                      TOTAL SAMPLE
                                                                                                                 TOTAL SAMPLE
                                                                                                                 N=189
     5%                                                                                                          N=189
                                                                                                                 AR [0]
                                                                                                                  AR [0]              =
                                                                                                                                      =   +4.74%
                                                                                                                                           +4.74%     ( ( t=
                                                                                                                                                           t=   +16.37 ) )
                                                                                                                                                                 +16.37
                                                                                                                 AR [+1]
                                                                                                                  AR [+1]             =
                                                                                                                                      =   +1.32%
                                                                                                                                           +1.32%     ( ( t=
                                                                                                                                                           t=   +4.57 ) )
                                                                                                                                                                 +4.57
                                                                                                                 CAR [-1;+1]
                                                                                                                  CAR [-1;+1]         =
                                                                                                                                      =   +6.00%
                                                                                                                                           +6.00%     ( ( t=
                                                                                                                                                           t=   +11.96 ) )
                                                                                                                                                                 +11.96
                                                                                                                 CAR [-1;+5]
                                                                                                                  CAR [-1;+5]         =
                                                                                                                                      =   +6.66%
                                                                                                                                           +6.66%     ( ( t=
                                                                                                                                                           t=   +8.69 ) )
                                                                                                                                                                 +8.69
                                                                                                                 CAR [-30;-2]
                                                                                                                  CAR [-30;-2]        =
                                                                                                                                      =   -9.14%
                                                                                                                                           -9.14%     ( ( t=
                                                                                                                                                           t=   -5.86 ) )
                                                                                                                                                                 -5.86
                                                                                                                 CAR [-60;-2]
                                                                                                                  CAR [-60;-2]        =
                                                                                                                                      =   -11.52%
                                                                                                                                           -11.52%    ( ( t=
                                                                                                                                                           t=   -5.18 ) )
                                                                                                                                                                 -5.18

     0%




                                                                             -5
           -60

                 -55

                       -50

                             -45

                                   -40

                                         -35

                                               -30

                                                     -25

                                                           -20

                                                                 -15

                                                                       -10



                                                                                  0

                                                                                      5

                                                                                          10

                                                                                                15

                                                                                                     20

                                                                                                                25

                                                                                                                        30

                                                                                                                               35

                                                                                                                                      40

                                                                                                                                              45

                                                                                                                                                     50

                                                                                                                                                            55

                                                                                                                                                                      60
                                                                                                                                                                days


                                                                                      NON-NM COMPANIES
                                                                                       NON-NM COMPANIES
                                                                                      N=118
                                                                                       N=118
     -5%                                                                              AR [0]        = +3.23%         ( t= +12.41 )
                                                                                       AR [0]        = +3.23%         ( t= +12.41 )
                                                                                      AR [+1]       = +1.72%         ( t= +6.61 )
                                                                                       AR [+1]       = +1.72%         ( t= +6.61 )
                                                                                      CAR [-1;+1]   = +4.92%         ( t= +10.91 )
                                                                                       CAR [-1;+1]   = +4.92%         ( t= +10.91 )
                                                                                      CAR [-1;+5]   = +5.90%         ( t= +8.56 )
                                                                                       CAR [-1;+5]   = +5.90%         ( t= +8.56 )
                                                                                      CAR [-30;-2]  = -7.68%         ( t= -5.48 )
                                                                                       CAR [-30;-2]  = -7.68%         ( t= -5.48 )
                                                                                      CAR [-60;-2]  = -8.18%         ( t= -4.09 )
                                                                                       CAR [-60;-2]  = -8.18%         ( t= -4.09 )


    -10%


                                                                                      NM COMPANIES
                                                                                       NM COMPANIES
                                                                                      N=71
                                                                                       N=71
                                                                                      AR [0]        =    +7.25%      ( t= +11.13 )
                                                                                       AR [0]        =     +7.25%     ( t= +11.13 )
                                                                                      AR [+1]       =    +0.67%      ( t= +1.02 )
                                                                                       AR [+1]       =     +0.67%     ( t= +1.02 )
                                                                                      CAR [-1;+1]   =    +7.79%      ( t= +6.90 )
    -15%                                                                               CAR [-1;+1]
                                                                                      CAR [-1;+5]   =
                                                                                                     =     +7.79%
                                                                                                         +7.93%
                                                                                                                      ( t= +6.90 )
                                                                                                                     ( t= +4.60 )
                                                                                       CAR [-1;+5]   =     +7.93%     ( t= +4.60 )
                                                                                      CAR [-30;-2]  =    -11.56%     ( t= -3.29 )
                                                                                       CAR [-30;-2]  =     -11.56%    ( t= -3.29 )
                                                                                      CAR [-60;-2]  =    -17.07%     ( t= -3.41 )
                                                                                       CAR [-60;-2]  =     -17.07%    ( t= -3.41 )
                              Non-NM Companies
                              NM-Company

    -20%



 Source: own investigation


    Absolutely in line with our expectations, the stock market reaction of buy-back
     announcements by NM-companies is more than twice as high as that of Non-NM
     companies with AR[0]=+7.25% vs. AR[0]=+3.23%. The CARs [-1;+1] are +7.29% and
     +4.29% for NM and non-NM companies, respectively. The CARs [-1;+5] are +7.93% and
     +5.90% for NM and non-NM companies, respectively.

    This evidence also supports the signalling hypothesis, as the companies with higher
     information asymmetries send a more powerful signal to the market when they announce
     share buy-backs.

    Both samples experience a significant underperformance prior to the share buy-back
     announcement. However, both groups feature somewhat different trends during the time
     period [-60;-2]. For the Non-NM sample, the underperformance sets in on the day [-34]
     and there is no underperforming trend to be seen prior to this point in time. In contrast,


                                                                                                                                                                             54
5. Empirical Investigation of Buy-Back Announcements

      the NM sample underperforms almost on the whole length of [-60;-6] with the
      underperformance somewhat increasing after the day [-28]116.

     Starting from day [-6] the underperforming trend of NM companies turns around which
      indicates the penetration of positive information to the market and probably insider
      trading.

     After the share buy-back, both company samples continue their development from a
      higher level which persists through the observation period [+5;+60], however,
      accompanied through a significant negative underperformance. There are no significant
      difference in the slopes between NM and non-NM companies.


5.2.6. Analysis Abnormal Returns by Cumulative Returns before Announcement

By examining the cumulative abnormal returns prior to the announcement day we can gauge
whether the management reacts to the prior share performance when it decides when and
under what market conditions the share buy-back will be announced. According to the
signalling hypothesis, those buy-back announcements would induce a higher price reaction
which have experienced a higher underperformance prior to the announcement117, (i.e. the fact
of the company undervaluation would be in this case more credible).

Moreover, if signalling/undervaluation is one of the key motivating factors behind the buy-
back decision, the management is more likely to send a stronger signal to the market when a
firm‟s share price has experienced significant negative returns before the announcement. Thus
higher repurchase targets, higher premiums and longer repurchase periods are announced in
such cases118.

I examined the interrelation between the cumulative abnormal returns before the buy-back
announcements and price reaction on the announcement day [0] and around it. The method of
linear regression has been used. In total, I carried out 6 regression analyses: investigating in
pairs interrelations between CAR [-60;-2] and CAR [-30;-2] vs. AR [0], CAR [-1;1] and CAR [-
1;+5]. The results of these analyses are shown in FIGURE 17 below:




116 Returning back to the problem of the underperformance of the entire sample (p.38) – as we can see above, the
underperformance of the whole sample during [-60, 0] can primarily be explained by the effect of the NM
companies. The „normal‟ companies do not experience underperformance prior to [-34].
117 See [Zhang, 2000], p. 15.
118 See [Zhang, 2000], p. 14.


                                                                                                             55
5. Empirical Investigation of Buy-Back Announcements


  FIGURE 17: REGRESSION ANALYSIS OF ABNORMAL RETURNS PRIOR TO ANNOUNCEMENTS OF SHARE
  BUY-BACKS

                                              I n d             e    p     e    n     d       e   n   t      v     a r i a         b   l   e   s
                                         CAR [-30;-2]                                                              CAR [-60;-2]
 Dependent variables:
        AR [0]                            Coefficients       t-Statistics           P-Value                         Coefficients   t-Statistics      P-Value
                            Intercept            0.046               6.51 **           0.00           Intercept            0.046           6.68 **      0.00
                            x Variable          -0.015              -0.52 *            0.61           x Variable          -0.012          -0.62 *       0.54


        CAR [-1;+1]                       Coefficients       t-Statistics           P-Value                         Coefficients   t-Statistics      P-Value
                            Intercept            0.057               6.53 **           0.00           Intercept            0.057           6.78 **      0.00
                            x Variable          -0.033              -0.96 *            0.34           x Variable          -0.022          -0.89 *       0.38


        CAR [-1;+5]                       Coefficients       t-Statistics           P-Value                         Coefficients   t-Statistics      P-Value
                            Intercept            0.061               5.12 **           0.00           Intercept            0.067           5.75 **      0.00
                            x Variable          -0.054              -1.14 *            0.26           x Variable           0.007           0.21 *       0.84

                                    ** - Significant at the 99% confidence level and better
                                     * - Not statistically significant

Source: own investigation


     In contrast to the implications of the signalling/undervaluation theory, no statistically
      significant relation could be detected for any pair of regression. As we can see in this figure
      above, there is but insignificant negative119 relation between the cumulative abnormal
      returns prior to the announcement and the magnitude of the abnormal returns surrounding
      the announcement day.


5.2.7. Analysis of Abnormal Returns by the Percentage of Outstanding Shares
Repurchased

Vermaelen (1981), as well as Comment and Jarrell (1991) found that the magnitude of the
premium offered to shareholders was positively related to the percentage of outstanding shares
repurchased120. The authors suggested that the repurchase target indicates the quality of
management information about the undervaluation of the company.

I partitioned the whole sample into high-percentage and low-percentage portfolios, according
to the number of shares the company intended to repurchase as a percentage of shares
outstanding121. However, this is the German specific that the company cannot buy back more
than 10% of its share capital. Moreover, German corporations do not have to commit
themselves to (and announce) a specific percentage of shares they are going to repurchase.




119 Negative due to the „negativeness‟ of the CARs prior to announcement.
120 See [Vermaelen, 1981], p. 155 and [Comment/Jarrell, 1991], p. 1245.
121 The technology is taken from [Zhang, 2000], p. 9.


                                                                                                                                                               56
5. Empirical Investigation of Buy-Back Announcements

Therefore, this sample was largely based on the percentages to which the companies were
authorised by the AGMs.

The high-percentage sub-sample included all buy-backs where the company was authorised to
repurchase the maximum allowed amount of 10% of its share capital. Thus, the sample
included some 133 firms. The low-percentage sub-sample included all repurchases where the
company was authorised to buy-back less than 10% of its share capital. This sample consisted
of 54 firms. However, the results did not deliver any statistically significant differences in
reaction between the two samples. As it seems, this fact can primarily be explained by the
following reasons:

     German companies are allowed to buy-back max. 10% of their shares. Thus, this amount
      might be not large enough so that any difference in reaction became perceivable.
      Moreover, the periods during which companies are allowed to buy back shares are
      sometimes quite long. This indeed may affect the management decisions on how much
      shares they in fact want to buy-back. Thus, on the day of announcement, this information
      is not being reflected in the share price with a significant degree.

     The 10% to which the companies are authorised are by no means a firm commitment to
      repurchase exactly this amount of shares – but simply a frame within the scope of which
      the company may move. In this light, an authorisation (or announcement) to buy-back say
      5% of the shares might even seem (from the capital market perspective) to be more a firm
      commitment than 10% (!). Thus, an intention to buy-back a smaller percentage could be
      perceived even more positively by the capital market.


5.2.8. Analysis of Abnormal Returns by Financial Position of Company

In this section, I tested a hypothesis that the abnormal return of repurchasing firms should be
higher, the higher amount of net cash122 they have. Although a part of the cash flow
hypothesis, this argumentation would not contradict the signalling hypothesis: the buy-back
signal will be perceived more positively since the release of free cash to stockholders would
discipline the management and eliminate the possibility that it invests the funds sub-
optimally123.



122 Net debt = Interesting Bearing Debt – Cash and Cash Equivalents. A negative net debt is equivalent to a net
cash position. The data was mainly sourced from the Bloomberg database, as well as from the annual reports of
the repurchasing companies. The reporting dates of financial years most adjacent to the buy-back before and after
it have been used.
123 See [Evans et al, 2000], p. 4.


                                                                                                              57
5. Empirical Investigation of Buy-Back Announcements

I calculated net debt124 positions for the data set entries from the Main Sample, before and
after the announcement of their share buy-back. At a first glance, the total sample featured a
very high concentration (i.e. 53%) of cash positive companies which was quite unusual. As a
proxy for cash amounts available to the company, I used a proportion of net debt to the
company market cap. Afterwards I divided the entire sample (177 announcements) entries into
10 equal decile groups à 17-18. The results of the regression analysis are shown in FIGURE 18
below:


  FIGURE 18: ANALYSIS OF ABNORMAL RETURNS BY FINANCIAL POSITION OF COMPANIES

 Dependent Variables:
         AR [0]                                 Coefficients    Standard Error t-Statistics          P-Value
                               Intercept                  0.031            0.015            2.10               0.037
                               x Variable                 0.004            0.002            1.49 *             0.138


         CAR [-1;+1]                            Coefficients    Standard Error t-Statistics          P-Value
                               Intercept                  0.048            0.018            2.65               0.009
                               x Variable                 0.003            0.003            0.96 *             0.339


         CAR [-1;+5]                            Coefficients    Standard Error t-Statistics          P-Value
                               Intercept                  0.065            0.025            2.57               0.011
                               x Variable                 0.001            0.004            0.19 *             0.849


                            * - Not statistically significant

Source: own investigation




     As we can see there is a positive relation between the (decile) relative net debt position of
      the company and the abnormal price returns surrounding the announcement day.
      However, this relation is very weak with the highest confidence level of approx. 86% as
      applied for the AR [0] regression. Thus, the 0-hypothesis could not be denied with any
      significant level of confidence125.




124 I used net debt as a substitute for cash flows, primarily due to the reasons of the data availability and different
cash flow definitions used by the sample companies.
125 The results of this analysis could probably be further improved by investigating cash flows instead of net debt

or by further narrowing the reporting dates (from „yearlies‟ to „quarterlies‟) at which net debt positions have been
estimated.

                                                                                                                       58
5. Empirical Investigation of Buy-Back Announcements

5.2.9. Analysis of Abnormal Returns by Shareholder Structures

As documented by a number of studies, e.g. Vermaelen (1981), the magnitude of the premium
offered to shareholders was positively related to the fraction of the company‟s shares owned
by managers. Understandably, the signal carried by a share buy-back is stronger in such a case,
as it expresses the undervaluation beliefs of owner-managers who pay their own cash for it.
Moreover, researchers such as Zhang (2000) and Ginglinger and L‟Her (2002) examined the
influence of shareholder structures in more detail extending their analyses to multivariate tests
with several input variables describing shareholder structures in more depth (see below, Cross-
Sectional Multivariate Analysis)

In the ensuing analysis, I investigated the relationship between the shareholder structure of the
company and the magnitude of abnormal returns upon a share buy-back announcement. The
general patterns of the investigation technology I used have been mainly taken from Zhang
(2000) and Ginglinger and L‟Her (2002).

Analysis of abnormal returns by management shareholdings

Firstly, I investigated the influence of the management holding on the magnitude of the
abnormal returns on the announcement day - AR [0] and around it - CAR [-1;+1], CAR [-
1;+5], by means of linear regression. Expectedly, the companies with a higher percentage of
the shares owned by managers would, in tendency, produce a more positive reaction to share
buy-back announcements.

The data on the management holding at the time of share buy-backs was sourced from
different sources such as Bloomberg and Hoppenstedt, as well as Annual Reports of the
repurchasing companies. The results of this analysis are presented in FIGURE 19 below (the
table simply contains the slope-coefficients and the t-statistics in italics underneath):




                                                                                              59
5. Empirical Investigation of Buy-Back Announcements


  FIGURE 19: ANALYSIS OF ABNORMAL RETURNS BY MANAGEMENT HOLDINGS

                                                  I n d e p e n d e n t              v a r i a b l e s
                                                                                      DUMMY FOR
                                                  MANAGEMENT                        MANAGEMENT
                     Dependent variables:            HOLDING                            HOLDING

                     AR [0]                                      0.016                              0.023
         SECTION 1


                                                                  0.52 *                             1.94 **
                     AR [-1;+1]                                 0.022                               0.025
                                                                 0.62 *                              1.64 **
                     AR [-1;+5]                                 -0.005                              0.028
                                                                 -0.09 *                             1.31 *


                     ABS AR [0]                                 0.045                               0.036
                                                                 1.83 **                             3.50 ****
         SECTION 2




                     ABS AR [-1;+1]                             0.054                               0.039
                                                                 1.85 **                             3.23 ****
                     ABS AR [-1;+5]                              0.051                              0.042
                                                                  1.29 *                             2.52 **

                                      ****   - significant at the 99% confidence level and better
                                       ***   - significant at the 95% confidence level
                                        **   - significant at the 90% confidence level
                                         *   - not statisically significant

Source: own investigation


     In Section 1, I tried to find any significant relation by performing a regression analysis
      between the magnitude of the management holding and the proxies for abnormal returns
      AR [0], CAR [-1;+1] and CAR [-1;+5]. However, contrary to my expectations, no
      statistically significant positive price relation has been found.

     Then I replaced the absolute magnitude of the management holdings by a dummy variable
      equalling 1 when management held any reportable amounts of share capital at the time of
      share buy-backs and 0 otherwise. The results show that there is indeed a statistically
      significant positive relation between the variables, at least for AR [0] and CAR [-1;+1] at
      the confidence level of 90%.

     Accidentally, I noticed that rather the absolute (!) magnitudes of abnormal returns on the
      announcement days depend on the amount of the management holdings. I then carried out
      an additional regression analysis which was applied to the absolute magnitudes of abnormal
      returns (see Section 2 of FIGURE 19 above). I found statistically significant relations, at
      least for ABS AR [0] and ABS CAR [-1;+1] at the confidence levels of 90%. The statistical

                                                                                                                 60
5. Empirical Investigation of Buy-Back Announcements

      significance of this relation even increased to 99% and better (see Section 2, on the right
      side) upon the replacement of the management holding variable with the dummy variable
      for management holdings as applied above.

     Thus, I came to a conclusion that the management holding is primarily decisive for the
      strengthening of the buy-back signal sent to the capital market. However, a management
      holding might be the reason for strengthening of both a more positive and a more negative
      reaction. Which reaction – depends on the capital market perception of the role of the
      management in the company. If the market participants think the company is buying back
      shares for a „good reason‟ (e.g. it is undervalued), the existence of owner-managers might
      increase the credibility of the signal sent. However, if the market participants (e.g.
      minorities) are concerned about possible threats brought about by a share buy-back
      (further squeeze out of minority interests, buy-back in a take-over defence situation,
      further management „entrenchment‟, etc.) the existence of management holdings would
      further increase the negative signal126.

     Furthermore, the fact that the significance of my analysis increased by switching to the
      dummy variable, led me to a conclusion that not the magnitude of the management
      shareholding but simply the existence of it is the most decisive factor for the final
      perception of the buy-back by the capital market. Thus, even small stakes held by the
      management of the company could significantly influence the credibility of the signals sent
      via a buy-back to the capital markets.



Analysis by level of shareholder control

In addition, I carried out an analysis of the influence of insider control concentration on the
abnormal returns upon share buy-back announcements. The data on companies‟ shareholder
structures was put together from different sources such as Bloomberg, the home-pages of the
German Stock Exchange, Primark Extel Reports (Thomson Financial), ad-hoc announcements
of the repurchasing companies, Annual Reports, as well as Hoppenstedt Stock Guides (1998-
2002). For the purposes of my investigation the next adjacent available shareholder structure
of the sample companies have been used127. I assumed them to be the best approximation for
the actual shareholder structures at the time of the buy-back announcement.


126 Those are my own thoughts on a possible explanation of this phenomenon. I have not found any second
opinion or a reference in a literature source on this problem.
127 In some cases, the current shareholder structures of the companies have been used. Searching for the exact

shareholder structures on the announcement day could be a subject of a whole research paper, as the costs of

                                                                                                           61
5. Empirical Investigation of Buy-Back Announcements

The complete sample has been divided into three groups according to the level of corporate
control128 over the company by the capital market. Dispersed control (DC) resulted when the
company had no shareholders with a voting stake higher than 25%. Medium control (MC)
required a stake of at least 25% of voting rights (but not exceeding 50%). The Highly
controlled (HC) companies had to have a shareholder with more than 50% of voting rights 129.
The results of my investigation are shown in FIGURE 20 below:


  FIGURE 20: ANALYSIS OF ABNORMAL RETURNS BY SHAREHOLDER STRUCTURES

     CAR, %                                             HIGH CONTROL
                                                        HIGH CONTROL
                                                                                                                                           TOTAL SAMPLE
                                                                                                                                            TOTAL SAMPLE
                                                        N=71                                                                               N=189
     5%                                                 N=71                                                                                N=189
                                                        AR [0]              =    +4.53%     ((   t=    +9.30    ))                         AR [0]
                                                                                                                                            AR [0]                    ==   +4.74%
                                                                                                                                                                            +4.74%     ( ( t=
                                                                                                                                                                                            t=   +16.37 ) )
                                                                                                                                                                                                  +16.37
                                                         AR [0]              =    +4.53%          t=    +9.30
                                                        AR [+1]             =    +2.42%     ((   t=    +4.96    ))                         AR [+1]
                                                                                                                                            AR [+1]                   ==   +1.32%
                                                                                                                                                                            +1.32%     ( ( t=
                                                                                                                                                                                            t=   +4.57 ) )
                                                                                                                                                                                                  +4.57
                                                         AR [+1]             =    +2.42%          t=    +4.96
                                                        CAR [-1;+1]         =    +7.49%     ((   t=    +8.87    ))                         CAR [-1;+1]
                                                                                                                                            CAR [-1;+1]               ==   +6.00%
                                                                                                                                                                            +6.00%     ( ( t=
                                                                                                                                                                                            t=   +11.96 ) )
                                                                                                                                                                                                  +11.96
                                                         CAR [-1;+1]         =    +7.49%          t=    +8.87
                                                        CAR [-1;+5]         =    +10.10%    ((   t=    +7.83    ))                         CAR [-1;+5]
                                                                                                                                            CAR [-1;+5]               ==   +6.66%
                                                                                                                                                                            +6.66%     ( ( t=
                                                                                                                                                                                            t=   +8.69 ) )
                                                                                                                                                                                                  +8.69
                                                         CAR [-1;+5]         =    +10.10%         t=    +7.83
                                                        CAR [-30;-2]        =    -7.93%     ((   t=    -3.02    ))                         CAR [-30;-2]
                                                                                                                                            CAR [-30;-2]              ==   -9.14%
                                                                                                                                                                            -9.14%     ( ( t=
                                                                                                                                                                                            t=   -5.86 ) )
                                                                                                                                                                                                  -5.86
                                                         CAR [-30;-2]        =    -7.93%          t=    -3.02
                                                        CAR [-60;-2]        =    -10.13%    ((   t=    -2.71    ))                         CAR [-60;-2]
                                                                                                                                            CAR [-60;-2]              ==   -11.52%
                                                                                                                                                                            -11.52%    ( ( t=
                                                                                                                                                                                            t=   -5.18 ) )
                                                                                                                                                                                                  -5.18
                                                         CAR [-60;-2]        =    -10.13%         t=    -2.71
     0%

                                                                                            -5
           -60

                 -55

                       -50

                              -45

                                      -40

                                             -35

                                                        -30

                                                                -25

                                                                      -20

                                                                            -15

                                                                                    -10



                                                                                                       0

                                                                                                                5

                                                                                                                     10

                                                                                                                           15

                                                                                                                                 20

                                                                                                                                           25

                                                                                                                                                   30

                                                                                                                                                             35

                                                                                                                                                                       40

                                                                                                                                                                               45

                                                                                                                                                                                      50

                                                                                                                                                                                            55

                                                                                                                                                                                                      60
                                                                                                                                                                                                 days



     -5%
             DISPERSED CONTROL
             N=57
             AR [0]            =    +4.24%    (    t=   +8.60   )
             AR [+1]           =    +1.06%    (    t=   +2.15   )
             CAR [-1;+1]       =    +4.91%    (    t=   +5.75   )
    -10%     CAR [-1;+5]       =    +4.40%    (    t=   +3.38   )
             CAR [-30;-2]      =    -5.27%    (    t=   -1.99   )
             CAR [-60;-2]      =    -9.25%    (    t=   -2.44   )




                                                                                                           MEDIUM CONTROL
                                                                                                           MEDIUM CONTROL
    -15%                                                                                                   N=61
                                                                                                           N=61
                            Dispersed Control                                                              AR [0]
                                                                                                            AR [0]         ==   +5.45%
                                                                                                                                 +5.45%      ((   t=
                                                                                                                                                   t=   +9.67
                                                                                                                                                         +9.67   ))
                                                                                                           AR [+1]
                                                                                                            AR [+1]        ==   +0.30%
                                                                                                                                 +0.30%      ((   t=
                                                                                                                                                   t=   +0.53
                                                                                                                                                         +0.53   ))
                            Medium Control                                                                 CAR [-1;+1]
                                                                                                            CAR [-1;+1]    ==   +5.28%
                                                                                                                                 +5.28%      ((   t=
                                                                                                                                                   t=   +5.41
                                                                                                                                                         +5.41   ))
                                                                                                           CAR [-1;+5]     =    +4.78%       ((   t=    +3.20    ))
                            High Control                                                                    CAR [-1;+5]
                                                                                                           CAR [-30;-2]    =
                                                                                                                            =    +4.78%
                                                                                                                                -14.16%      ((
                                                                                                                                                   t=
                                                                                                                                                  t=
                                                                                                                                                         +3.20
                                                                                                                                                        -4.66    ))
                                                                                                            CAR [-30;-2]    =    -14.16%           t=    -4.66
                                                                                                           CAR [-60;-2]
                                                                                                            CAR [-60;-2]   ==   -15.25%
                                                                                                                                 -15.25%     ((   t=
                                                                                                                                                   t=   -3.52
                                                                                                                                                         -3.52   ))

    -20%



Source: own investigation

There are some very interesting results which can be derived from the analysis above:

     The HC companies have generated the highest abnormal returns upon the share buy-back
      announcements. Although this group did not produce the highest first day returns AR[0]

time required would be tremendous. But in fact, even that would bring only limited advantages for the model
applied here. It should be mentioned at this stage, that German firms have no obligation to report the shareholder
structures on the announcement day – in contrast to their French counterparts (see Ginglinger and L‟Her (2002),
p.7, which describe their shareholder structures in the registration statements at the time of adoption of the share
buy-back plans.
128 Or, even better expressed, „controllability‟.
129 However, in each case, the decision on the group affiliation was made rather individually. For example, I

considered the stakes held by the management members as one stake or the company belonged to the „High
Control‟ group even with the largest stake of slightly under 50% but held by a financial institution.

                                                                                                                                                                                                              62
5. Empirical Investigation of Buy-Back Announcements

    =+4.53%, the cumulated abnormal returns CAR [-1;+5] amounted to +10.13% (!). The
    DC and MC companies had average abnormal returns CAR[0]DC = +5.45% and CAR[0]MC
    =+4.24% respectively. For them, the cumulative abnormal returns around the
    announcement day CAR [-1;+1] and CAR [-1;+5] have not been much different from the
    return on the announcement day [0].

   In contrast to its peers, the HC group of companies has succeeded to return its share price
    to the levels of [-60;-35].

   These results above largely support the signalling/undervaluation hypothesis, as they
    provide evidence that the signal sent to the capital market by a „better‟ controlled firm is
    significantly stronger than those sent by other companies.

   Interestingly, approximately 50% of CARs [-1;+5] of the HC group have been generated
    after the announcement day (!). The CAR [+1;+5] amounts to +5.03% and is statistically
    significant at the 99% confidence level and better. This implies the possibility of receiving
    abnormal profits investing in the stocks during the period [+1;+5] upon the
    announcement – i.e. during the time when the buy-back information has already become
    public (!). Moreover, in a situation of a limited liquidity on the market (i.e. when the free
    float is in the range 5%-10% ) the minority shareholders are not willing to tender their
    shares to the company in anticipation of a possible squeeze out for a much higher price.

   Although this looks like a new phenomenon which should certainly be investigated in
    more detail, I would nevertheless suggest that this deferred price reaction could have been
    caused by limited liquidity which is available for trading on the market. The HC company
    group, i.e. those companies whose shares are concentrated in the hands of a few
    shareholders, would rather tend to have a more limited free float. Thus, this deferred
    reaction appears to be simply a technical problem which by no means can be used for
    achieving significant speculative profits.

   Also a very interesting point which is worth mentioning here is the highest cumulative
    underperformance CAR [-60;-2] of the MC group equalling -15.25%, whereas the HC and
    DC groups experienced an underperformance which was not significantly different from
    that of the total sample. This underperforming trend can be seen clearly on FIGURE 20
    above, where the MC curve starts from the „all-highs‟ around the day [-40] down to the „all-
    lows‟ just before the announcement.




                                                                                               63
5. Empirical Investigation of Buy-Back Announcements

      An intuitive explanation for this phenomenon could probably be the combined impact of
       two negative factors. On the one hand, a limited free float of the MC companies130 leads to
       limited liquidity available on the market and, as a result, lower interest by investors. On the
       other hand, the situation is further worsened by the absence of a majority shareholder131
       and, as a result, a higher probability of ineffective control. Thus, this „double cave‟ leads to
       an increased „mis-pricing‟ of these companies prior to the buy-back announcement.

All the results above should be mainly viewed in the light of the fact that there are possibly
several confounding influencing factors which could somewhat blur the results. For example,
the company size can produce much influence in the HC group since the proportion of
individual investors and, in particular, owner-managers is in tendency higher at smaller
companies.

The following section will examine the reaction patterns of share buy-backs via a cross-
sectional multivariate regression analysis. In this analysis, the issue of shareholder structure will
be, inter alia, expanded to a number of further variables - proxies for the shareholding structure
of repurchasing companies (see below in detail).

5.3. DETAILED INVESTIGATION – A CROSS-SECTIONAL MULTIVARIATE ANALYSIS
The analyses above illustrated features of share-price behaviour surrounding buy-back events
for the full sample and several sub-samples created according to firms‟ characteristics and
terms of buy-backs. However, some of these characteristics might have been correlated. So for
example, characteristics such as market listing segment, company size, industrial affiliation and
even some features of shareholder structures may well be inter-correlated. Thus, in order to
gauge the key determinants of the buy-back announcement returns, a multi-variable regression
analysis is being performed below.

In the ensuing analysis, I test the 0-hypothesis that the regression coefficients to the
independent variables of the multivariate regression equal zero. Thus I investigated the
following independent variables:

      MTB – the Market-to-Book variable was included in this analysis as it already provided
       good results in the univariate analyses above.

      SIZE – Here I tested two variables Size Decile and Size Dummy. Size Decile equals the
       firm size deciles numbers (which have been used in the analysis above), in the range from 1


130   Understandably, the free float of MC companies is significantly lower than that of DC.
131   In the MC group, no shareholder hold more than 50% of voting rights.

                                                                                                    64
5. Empirical Investigation of Buy-Back Announcements

      to 10; Size Dummy equals 1 for „Large‟ firms and 0 for „Small‟ firms, according to the
      definition given on page 43.

     LISTING is a dummy variable for the exchange listing sector equalling 1 for NM and 0 for
      non-NM companies. In the univariate analysis carried out above, this variable has been
      highly significant. However, LISTING seems to be a conglomerate variable comprising
      other „sub-characteristics‟ such as (smaller) size and (higher) ownership concentration (see
      p. 53). Thus, the ensuing multivariate analysis will show the „distillate‟ significance of this
      variable, i.e. stripped of its size and ownership concentration influencing components.

     CAR [-30;-2] and CAR [-60, -2] are the cumulative abnormal returns before the
      announcement, CAR [-30;-2] and CAR [-60, -2] respectively, in %. The possible influence
      of the prior returns on the share buy-back reaction has been explained on page 55;

     CONTROL – Here I tested two variables CONTROL-1 and CONTROL-2. CONTROL-
      1 is a dummy variable which equals 1 if the company is of HC type and 0 if the company
      was of the type MC or D. CONTROL-2 also shows the degree of corporate control by
      shareholders, whereas it equals 3 for HC-firms, 2 for MC and 1 for D (see p. 62);

     WDIFF = W*DIFF where W=VRS2/VRS1 is the weight of the second shareholding‟s
      voting rights relative to the first shareholder (VRS2 is the percentage of voting rights held
      by second largest shareholder; VRS1 is the percentage of the voting rights held by the first
      largest shareholder). DIFF is a dummy variable that equals 1 if the second shareholder is of
      a different type than the first and 0 otherwise132. The higher the influence of the second
      shareholder, the more favourable would be the expected reaction. The presence of a strong
      second shareholder, inter alia, disciplined the controlling ones133.

     TAKEOVER – a dummy variable, that equals 1 if the company is in a potential position as
      a target of a (hostile) takeover and 0 otherwise. Expectedly, a firm which is likely to be a
      target of a (hostile) takeover elicits a less favourable price reaction when the repurchase
      program is announced134;


132 For example, different are family shareholders and management shareholders, management shareholders and
institutional (banks) shareholders. However, the most important feature in this difference is the pursuit of
different interests by these shareholders and, therefore, existing incentives to control each other.
133 E.g. compare the case of mg technologies and Otto Happel, or Warren Buffett and Vendex KBB. The fact that

the presence of the second largest shareholder raises firm value and profitability is documented by e.g. Boehmer
(2000) and Lehman and Weigand (2000).
134 Compare [Denis, 1990], p. 1433-1434. In my particular case, every sample company has been considered

individually. However, as a major criterion of a defence situation I considered a „physical‟ possibility of a potential
acquirer to obtain majority control in the company via a hostile offer plus the possibility that the repurchase of
max. 10% of the capital by the company might significantly turn around the situation.

                                                                                                                    65
5. Empirical Investigation of Buy-Back Announcements

     FOREIGN – a dummy variable that equals 1 if foreign institutional investors are identified
      among the firm‟s shareholders. We can assume that these investors contribute to the
      adoption of international standards of corporate governance and are more vigilant towards
      the introduction of mechanisms that run counter to their interest135.

     MANAGEMENT – a dummy variable that equals 1 for firms with significant
      management shareholdings and 0 otherwise. The possible influence of management
      shareholdings is described on p. 60;

     FAMILY - a dummy variable that equals 1 for family controlled firms and 0 otherwise. In
      family controlled firms, there are large differences between cash-flow rights and control
      rights, thus, enhancing the power of the controlling family. As a result, minority
      shareholders are probably rather poorly protected because of the absence of reference
      shareholders that tend to discipline the controlling shareholders136. Accordingly, these type
      of firms would expectedly yield a less favourable price reaction to a share buy-back. On the
      other hand, a family may consolidate control over the firm which would positively impact
      the performance of the firm.

     Other possible variables such as (1) dummy for motives as stated by the company, (2)
      percentage of the outstanding shares repurchased, (3) relative amount of net debt have not
      been included into the analysis because of the missing data for many of the data set
      entries137. The Industry-variable has also not been included in the multivariate analysis,
      mainly for reasons of its conglomerate character, i.e. it is rather a compound of other
      factors like size, MTB, financial position etc.

I investigated three major models relating to the investigated variables with the three major
proxies of abnormal returns surrounding the announcement day AR [0], CAR [-1;+1] and
CAR [-1;+5]. In the first step, I carried out several univariate regressions in order to spot
variables to exclude from the final analysis. The results of the univariate regressions are shown
in FIGURE 21 below (the t-statistics showing the significance of the slope-coefficients from the
regression of the variables to the respective proxies of abnormal returns are featured in italics).
Thus, at the first stage I dropped the following variables SIZE DUMMY, CAR [-60;-2],



135 See [Ginglinger/L‟Her 2002], p. 16.
136 See [Ginglinger/L‟Her 2002], p. 16-17.
137 An attempt to use the data available has led to more than halving the final sample. Moreover, the variables (2)

and (3) have been found barely significant in the previous analysis. The variable (1) is per se not a company specific
factor but a directly announced reason for a share buy-back. Therefore, testing it along with the other variables
would not make sense.

                                                                                                                   66
5. Empirical Investigation of Buy-Back Announcements

CONTROL-2 and WDIFF138 and left those counterparts which better correlated with the
abnormal returns.


  FIGURE 21: FIRST-STEP-ANALYSIS OF VARIABLES

                                                D     e p e n d e n t       v                      a r   i   a b l e s
         Independent variables                       AR [0]      AR [-1;+1]                                  AR [-1;+5]
              MTB                                      -2.17 ***                        -2.66 *****              -2.38 ****
              SIZE DECILE                              -5.45 *****                      -4.36 *****              -3.37 *****
              SIZE DUMMY                               -3.76 *****                      -2.99 *****              -2.33 ****
              LISTING                                   3.07 *****                       1.75 **                 0.90 *
              CAR [-30;-2]                             -0.52 *                          -0.96 *                  -1.14 *
              CAR [-60;-2]                             -0.62 *                          -0.89 *                  0.21 *
              CONTROL-1                                -0.25 *                           1.45 *                  2.44 ****
              CONTROL-2                                 0.14 *                           1.36 *                  2.20 ***
              DIFF                                      1.45 *                           0.71 *                  0.33 *
              WDIFF                                     0.70 *                           0.15 *                  0.31 *
              TAKEOVER                                 -0.22 *                           0.04 *                  -0.48 *
              FOREIGN                                   0.09 *                          -0.17 *                  1.23 *
              MANAGEMENT                                1.94 **                          1.64 *                  1.31 *
              FAMILY                                    1.01 *                           1.13 *                  1.03 *

                                 *****   - Significant at the 99% confidence level and better
                                  ****   - Significant at the 98% confidence level
                                   ***   - Significant at the 95% confidence level
                                    **   - Significant at the 90% confidence level
                                     *   - Not statistically significant

Source: own investigation




Thus, I pursued my analysis with the following 10 variables: MTB, SIZE DECILE, LISTING,
CAR [-30;-2], CONTROL-1, DIFF, TAKEOVER, FOREIGN, MANAGEMENT and
FAMILY. Upon the conclusion of the analysis for every model (i.e. AR [0], CAR [-1;+1] and
CAR [-1;+5], those variables that appeared to be significant139 were tested further in a reduced
model (on the right side of every column)140. The results of the multivariate regression are
presented in FIGURE 22 below:


138 Apparently, the quality of this variable has not improved by multiplying it with the weight of the second
shareholder. Thus, as it seems, the nominal existence of a second disciplining shareholder is more decisive than its
voting weight.
139 I took the level of t=1.5 because in the ensuing analysis with a reduced number of variables, the level of

significance might rise.
140 This reduced-variables analysis is necessary because of the multi-colinearity problem which might have existed

between those variables which remained in the second step and those variables which have been excluded. As we

                                                                                                                               67
      5. Empirical Investigation of Buy-Back Announcements



  FIGURE 22: MULTIVARIATE REGRESSION ANALYSIS
                                                                     D       e   p        e   n   d    e      n     t      v a       r   i    a     b   l   e   s
       Independent variables                            AR [0]                                                    CAR [-1;+1]                                        CAR [-1;+5]
           INTERCEPT                        0.110                     0.102                           0.112                 0.128                           0.105                  0.119
                                            5.56         ****         5.80           ****             4.45          ****    7.76             ****           3.00         ****      4.77     ****
           MTB                             -0.003                    -0.003                           -0.005                -0.004                          -0.006                 -0.005
                                            -1.62        *            -1.73          **                -2.14        ***      -1.82           **              -1.90       **         -1.88   **
           SIZE DECILE                     -0.010                     -0.010                          -0.009                -0.010                          -0.009                 -0.010
                                            -4.12        ****          -4.06         ****              -2.95        ****     -3.86           ****            -2.07       ***        -2.70   ****
           LISTING                         0.032                      0.034                           0.023                      -                          0.014
                                           2.11          ***          2.30           ***              1.20          *            -                          0.52         *
           CAR [-30;-2]                    -0.006                        -                            -0.023                     -                          -0.044
                                            -0.23        *               -                             -0.68        *            -                           -0.93       *
           CONTROL-1                       -0.015                        -                            0.017                      -                          0.052                  0.052
                                            -1.04        *               -                            0.92          *            -                          2.06         ***       2.36     ***
           DIFF                            0.010                         -                            0.004                      -                          -0.001
                                           0.71          *               -                            0.23          *            -                           -0.05       *
           TAKE-OVER                       -0.009                        -                            -0.006                     -                          -0.021
                                            -0.73        *               -                             -0.39        *            -                           -0.95       *
           FOREIGN                         -0.002                        -                            -0.007                     -                          0.034
                                            -0.10        *               -                             -0.32        *            -                          1.19
           MANAGEMENT                      -0.024                     -0.021                          -0.014                     -                          0.003        *
                                            -1.55        *             -1.44         *                 -0.70        *            -                          0.12         *
           FAMILY                          0.031                      0.022                           0.019                      -                          0.005
                                           1.66          **           1.31           *                0.79          *            -                          0.16         *

                               ****   - Significant at the 99% confidence level and better
                                ***   - Significant at the 95% confidence level
                                 **   - Significant at the 90% confidence level
                                  *   - Not statistically significant


Source: own investigation




      The multivariate analysis above led me to the following conclusions:

           MTB (Market-to-Book value ratio) and firm size appear to be the most significant factors
            influencing a signal sent to the capital market via a buy-back announcement. These two
            variables „work‟ well for all three proxies of abnormal returns on and around the day of the
            announcement, as they were statistically significant for AR [0], CAR [-1;+1] and CAR [-
            1;+5].

           Moreover the CONTROL-1 variable features a significant positive relation only with the
            third proxy CAR [-1;+5]. As we have seen in the univariate analysis (see p. 62), the price
            reaction to share buy-backs of the HC group is not fully reflected in the share price on the
            day [0] and even for the period [-1;+1] and the market needs approx. next four-five days to
            fully saturate the information content of a share buy-back. This can be seen in the table as
            the regression coefficient of the CONTROL-1 variable changes its sign from negative to


      can see, the significance of the remaining variables did not change very much. Thus, there is apparently no severe
      multi-colinearity problem between these variable and the variables excluded in the second step have been in fact
      not influencing.

                                                                                                                                                                                                  68
5. Empirical Investigation of Buy-Back Announcements

       positive as time passes. The variable becomes first significant at the 95% confidence level
       for CAR [-1;+5].

      The LISTING variable was found significant only for the proxy AR [0]. Interestingly,
       however, the significance of this variable decreased when tested to CAR [-1;+1] and CAR
       [-1;+5] – exactly the period when CONTROL-1 gained its significance. So in my view (and
       here is worth mentioning the discussions on p.53 and p.64), this fact could point out to the
       suggestion that LISTING variable probably „disappears‟ on behalf of CONTROL-1. This,
       in turn, would infer that the higher price reaction from buy-backs by NM-companies could
       rather be explained by a higher proportion of HC-firms (High Control, see p. 62) on the
       Neuer Markt listing sector141.

      The evidence provided above leads to a conclusion that, ceteris paribus, the share buy-back
       announcement of the firms with a smaller size, lower MTB ratio and (contingently) those
       listed on the Neuer Markt are perceived by the market as more credible signal of
       undervaluation.

      The expansion of the investigation to a number of new variables did not provide any new
       insights into the influence on the signal sent to the market by a share buy-back
       announcement. Interestingly, however, although not statistically significant, the other
       variables such as CAR [-30;-2], DIFF, TAKEOVER, and FOREIGN (at least for the
       proxy CAR [-1;+5]) have featured „correct‟ sign of the relation with abnormal returns, i.e.
       in line with our expectations.

      The MANAGEMENT variable has not featured any sign of increasing significance (at
       least for the proxies AR [0], CAR [-1;+1] and CAR [-1;+5]). However, these proxies are
       apparently not „the right ones‟ but ABS AR [0], ABS CAR [-1;+1] and ABS CAR [-1;+5]
       (as shown in the univariate analysis).

      Interestingly and contrary to our expectations, family controlled firms apparently
       experienced on average more positive abnormal returns. However, this evidence was very
       weak as the FAMILY-variable lost its significance in the second step of the multivariate
       analysis (for proxy AR[0]).




141   However, in my view this point needs to be investigated more deeply

                                                                                                69
5. Empirical Investigation of Buy-Back Announcements


5.4. EXCURSIVE: ANALYSIS OF INDUSTRY-WIDE INFORMATION EFFECTS OF
SHARE BUY-BACK ANNOUNCEMENTS
In this section, I examine the information transfer effects of share buy-back announcements142,
i.e. whether share buy-back announcements also convey industry-wide information and, hence,
the announcements have valuation effects of industry counterparts. The answer to this
question is particular interesting in the context of the information signalling hypothesis – i.e. if
buy-back announcements are motivated by manager‟s desire to communicate information
about the future prospects of the firm, then it is possible that the undervaluation may be an
industry wide phenomenon (contagion effects)143. On the other hand, the repurchasing firms might
have gained market share of industry value, thus, rivals will be negatively affected (competitive
effects)144. An interplay of these two effects can possibly result in a positive or negative industry-
wide reaction.

In order to investigate the industry wide information effects of share buy-back
announcements, I created a „mirror-sample‟ in addition to the Main Sample. Instead of the
share-price data series of the companies this sample contained historical data series of the
corresponding CDAX sub-indexes for every company in the sample.

The event model technology remained the same in its major features. In addition, however, in
order to eliminate the influence if the repurchasing company on the sub-index returns, I
undertook a major correction of the daily returns of the sub-indexes by means of subtracting
the daily return of the repurchasing company multiplied by the weight of the company in the
sub-index from the daily return of the sub-index:

                                         Rind,icleared = Rind,i – w*Rfirm,i

Where Rind,icleared – the day [i]-return of the corresponding CDAX sub-index, cleared from the
influence of the repurchasing company (as the companies assumably experience abnormal
price returns around the announcement day). Rind,i and Rfirm,i are the daily price returns of the
corresponding CDAX sub-index and the company respectively; w is the weight of the
company in the corresponding CDAX sub-index145. The weightings of the sub-indexes have

142 The patterns of the investigation technology used here have been largely taken from the research paper of
[Otchere/Ross, 2000].
143 See [Otchere/Ross, 2000], p. 3.
144 See [Otchere/Ross, 2000], p. 3.
145 The trick described above has been invented in order to evade the major difficulties of the investigation

methodology proposed by Otchere and Ross (2000). For every repurchasing firm from the main sample (73
firms), the authors created a sub-sample of rival firms according to the ASX three-digit industry classification
scheme. Thus, the final sample counted in total 3,200 rival firms. In my opinion, this methodology has two major
advantages: (1) the influence of the repurchasing firms is totally eliminated from the final results and (2) the

                                                                                                             70
5. Empirical Investigation of Buy-Back Announcements

been sourced from the Deutsche Börse AG, such as they were on the days of buy-back
announcements of the repurchasing companies.

The results of my analysis are visualised in FIGURE 23 below:
  FIGURE 23: ANALYSIS OF INDUSTRY-WIDE INFORMATION EFFECTS OF SHARE BUY-BACK
  ANNOUNCEMENTS

    CAR, %                                                                     INDUSTRY-WIDE REACTION
    1%                                                                         N=188
                                                                               AR [0]               =   +0.02%    (    t=   +0.22   )
                                                                               AR [+1]              =   -0.05%    (    t=   -0.49   )
                                                                               CAR [-1;+1]          =   -0.04%    (    t=   -0.19   )
                                                                               CAR [-1;+5]          =   -0.04%    (    t=   -0.14   )
                                                                               CAR [0;+20]          =   +0.25%    (    t=   +0.49   )
                                                                               CAR [+10;+20]        =   +0.57%    (    t=   +0.88   )
                                                                               CAR [-30;-2]         =   -0.53%    (    t=   -0.88   )
                                                                               CAR [-60;-2]         =   -0.45%    (    t=   -0.52   )

    0%
          -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10   -5   0   5   10   15   20   25   30    35       40   45       50    55      60
                                                                                                                             days




    -1%




    -2%



Source: own investigation


     As we can see, there are no statistically significant abnormal returns of the industrial sub-
      indexes on the announcement day [0] and around it. Thus, since the 0-hypotheis could not
      be denied, we can state that (on average and for the whole sample of the companies) the
      undervaluation signals of the repurchasing companies are not being transferred to the
      company rivals. Thus, the undervaluation cannot be considered as an industry-wide
      phenomenon.



methodology allows to flexibly chose the best comparable competitors for the repurchasing firm. In contrast, a
CDAX industry sub-index consists of a fixed number of companies which are not always comparable (e.g. Henkel
and Metro in the CDAX Retail sub-index). Furthermore, the influence of the repurchasing company cannot be
fully eliminated since the weight of the company in the sub-index varies every day. However, it is assumed to
remain constant in my model. But on the other hand, it would not have been possible to implement this
methodology considering the amounts of time and data it requires. In any case, it would also be very interesting to
see whether the methodology as applied by Otchere and Ross (2000) could bring about different results.

                                                                                                                                         71
5. Empirical Investigation of Buy-Back Announcements

   However, this evidence needs to be taken with „a pinch of salt‟ as it refers to the entire
    sample. For example, the missing reaction might be accounted for by the parallel influence
    of both contagion and competitive effects. Thus, a deeper investigation is necessary to
    provide further insights into the problem.

5.5. CONCLUDING REMARKS
In Germany, announcements of share buy-backs cause a significant positive price reaction
surrounding the announcement day. This result strongly supports the signalling/
undervaluation hypothesis. On average, the sample companies were in fact „mis-priced‟ by the
capital markets and share buy-backs appeared to be an effective means of returning their share
prices to their previous levels.

As shown by means of univariate analyses, the strength of the buy-back signal was significantly
dependent on such factors as (undervaluation) motives as stated by the company, market-to-
book ratio, company size, exchange listing sector, as well as manager-holdings and shareholder
control    structure    of    the   company.     This    evidence    strongly    supports    the
signalling/undervaluation hypothesis. However, by means of a cross-sectional multivariate
analysis, only firm size, market-to-book ratio, shareholder control structure and (contingently)
exchange listing have been found to induce significant influence on the abnormal returns
surrounding the announcement of share buy-backs.

In contrast to my expectations, however, no evidence has been found that the signal sent to
the market via a share buy-back stands in any significant relation to cumulative returns prior to
the buy-back, industrial affiliation, percentage of outstanding shares repurchased as well as the
financial position of the company. Moreover, no significant industry-wide effects caused by
share buy-backs have been detected.




                                                                                              72
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs


6. EMPIRICAL INVESTIGATION OF BUY-BACK PLAN
ANNOUNCEMENTS AND AGM AUTHORISATIONS OF BUY-BACKS

As was already discussed above, every share buy-back is preceded by an authorisation by the
AGM. Moreover, as was observed in many cases, the companies announced their plans to buy
back shares even before the authorisations by the AGMs happened. In this Chapter, these two
buy-back events are examined in more detail. This kind of investigation would be particularly
interesting as it would greatly complement the results of the analyses of share buy-back
announcements provided in the previous chapter:

     Firstly, German buy-backs provide a unique opportunity for a researcher due to the legal
      requirements existing in this country that determine this three-events-framework as
      scheduled above. For instance, for a US sample, the first two events (i.e. plans and AGMs)
      could hardly be distinguished.

     Secondly, in order to estimate the overall effects of share buy-backs, all abnormal price
      reactions, i.e. induced by all the three events, should be considered as in combination.

     Thirdly, it is also interesting to see whether buy-back plan announcements and AGM
      authorisations alone would provide effective tools for the purposes of share price
      management – without having to buy back stock in fact.

6.1. ANALYSIS OF PRICE EFFECTS OF BUY-BACK PLAN ANNOUNCEMENTS
The announcements of the share buy-back plans considered by a company are not subject to
an ad-hoc release. The company simply sometimes makes public its intention to repurchase its
own stock or, more usually, its plans to seek shareholders approval for a buy-back program at
the next AGM. Thus, during the collection of the sample containing buy-back plans, the first
news found in the major article databases (such as Reuters Business Briefing, Bloomberg and
Factiva) has been considered as price relevant and included into the final sample 146. However,
in many cases, spotting the date was quite a difficult task, which is why much of the
information leakage can be expected in the final analysis.




146In this case, the task of seeking the dates was somewhat facilitated by the fact that the companies used to
announce their buy-back plans in quite a standard wording like “For the first time, the company plans to seek
shareholder approval for a share buy-back program at the next AGM … “.

                                                                                                           73
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

The final sample of share buy-back plans contained in total some 95 buy-back plan
announcements147,148. The results of the analysis of the share price reaction on a share buy-back
are shown in FIGURE 24 below:

      FIGURE 24: ANALYSIS OF BUY-BACK PLAN ANNOUNCEMENTS

      CAR, %
      10%
                 BUY-BACK PLANS
                 BUY-BACK PLANS
                 N=95
                 N=95
                 AR [0]
                 AR [0]           =
                                  =   +1.54%
                                      +1.54%   ((   t=
                                                     t=   +4.49
                                                          +4.49   ))
                 AR [+1]
                 AR [+1]          =
                                  =   +0.05%
                                      +0.05%   ((   t=
                                                     t=   +0.14
                                                          +0.14   ))
       8%
                 CAR [-1;+1]
                 CAR [-1;+1]      =
                                  =   +2.37%
                                      +2.37%   ((   t=
                                                     t=   +3.99
                                                          +3.99   ))
                 CAR [-1;+5]
                 CAR [-1;+5]      =
                                  =   +2.42%
                                      +2.42%   ((   t=
                                                     t=   +2.67
                                                          +2.67   ))
                 CAR [-14;-1]
                 CAR [-14;-1]     =
                                  =   +2.78%
                                      +2.78%   ((   t=
                                                     t=   +2.17
                                                          +2.17   ))
                 CAR [-60;+60]
                 CAR [-60;+60]    =
                                  =   +5.82%
                                      +5.82%   ((   t=
                                                     t=   +1.54
                                                          +1.54   ))

       6%




       4%




       2%




                                                                                                                             days
       0%
            -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10            -5   0   5   10   15   20   25   30   35   40   45   50   55     60



 Source: own investigation


      News on share buy-back plans produced positive abnormal returns: AR [0]=+1.54% and
       CAR [-1;+1] = +2.37%, which are highly significant at the 99% confidence level and
       better. The new price level after the event day obviously persists over the remaining period



147Primarily, I searched for buy-back plan announcements of the companies of the Main Sample (by „backward
induction‟). However, the number of such companies which have ever planned a share buy-back (but have not
bought back any shares yet) is much larger. So I completed my sample with some new companies of the kind that
I found accidentally. It would be also interesting (but also very difficult) to try to find all the buy-back plan
announcements. However, for my purposes, the sample of 95 constituents appears large enough.
148The sample was not cleared from confounding events on the event day and around it, like was done for the
Main Sample – mainly for technical reasons, as it was already difficult to spot the right date itself. But also by the
fact that almost every piece of news on share buy-back plans was a part of an „broader‟ announcement of overall
positive company news such as higher profit expectations, strategy changes etc., i.e. the announcements of plans
to buy back shares have been used by management to commit to the truthfulness of this information. Thus, these
pieces of news could not be separated from each other.

                                                                                                                                         74
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

    of observation of [+1;+60], i.e. the reaction on the plans contains value relevant
    information and is apparently not a one-day-effect.

   A positive trend already sets in after the day [-14], resulting in a positive run-up of CAR [-
    14;-1] = +2.78%, which is statistically significant at the 99% confidence level. This trend
    can be apparently explained by some positive company news with which the buy-back
    plans were announced (see footnote 148). Moreover, there were probably much rumours
    and information leakages prior to the day [0] as spotted by me.

   There is an overall positive trend prevailing throughout the entire observation period with
    CAR [-60;+60] = +5.82%. It is, however, significant only at the level of 88%.

   The evidence above supports the information signalling hypothesis and leads to the
    conclusion that even the news on a possible share buy-back can result in positive price run-
    ups and can be actively implemented by companies for the management of their share
    prices. However, the overall impact of buy-back plans is much lower than that induced by
    buy-back announcements. In fact, at the time of the buy-back plans, the company provides
    no firm commitment to buy back shares.

   The significant positive price reaction following the announcements of buy-back plans can
    be mis-used by the company‟s insiders for insider trading. This is why the buy-back plans
    should be subject to ad-hoc announcements.

6.2. ANALYSIS OF PRICE EFFECTS OF AGM AUTHORISATIONS
This analysis of the price effects of AGM authorisations of share buy-back programs delivered
number of very interesting results which are presented below. Logically, as we would expect,
an authorisation of a share buy-back should not result in somewhat significant price reactions,
mainly for the following reasons:

   The authorisation by the AGM gives the company an additional option to buy back its
    own shares. However, it still remains simply an option without concrete commitments and
    terms to exercise. In Germany, there are much more companies which possess an AGM
    authorisation but which have not bought back their shares yet. Thus, in light of this
    argumentation, the authorisation should expectedly induce, if at all, a slightly positive
    impact on the share price .




                                                                                               75
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

     The possibility that the AGM would not give the green light for the buy-back program is
      relatively low149. As soon as concrete buy-back plans are outlined by the management, the
      authorisation by the AGM becomes simply a formal procedure. That is why, expectedly,
      only the share buy-back plans and the share buy-backs themselves – and not the AGM
      authorisations – should carry any price-relevant information.

I created an additional sample which included 181 authorisation events granted by AGMs of
some 157 companies, during a period between May/1997 – Sep/2001150. The data was taken
from the same sources as before. Here again, the sample was created on the basis of the Main
Sample. The main implication from it is that the sample primarily contained the companies
which later in fact decided to buy back their shares. That does not mean, however, that this
sample includes all the German AGM authorisations, i.e. it did not cover the companies which
have not bought back their stocks yet.

It is worth mentioning at this stage that, like in the case above, no confounding events (or
news) from the AGMs were removed from the final sample. It appeared quite an impossible
task, due to the fact that the approval of a share buy-back program has never been the only
topic during an AGM. However, here I would argue that in a large company sample like this
one, all other confounding news (good and bad) would counter-balance each other.

The results of the analysis are visualised in FIGURE 25 below:




149 The statement is largely based on my own experience gained during the preparation of this investigation.
However, this question could be investigated more deeply.
150 As we see, some authorisations have been already given by AGMs conditional to then largely expected ban lift

on share buy-backs.

                                                                                                             76
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

      FIGURE 25: ANALYSIS OF PRICE EFFECTS OF AGM AUTHORISATIONS

        CAR, %
        5%                                                                    TOTAL SAMPLE
                                                                               TOTAL SAMPLE
                                                                              N=181
                                                                               N=181
                                                                              AR [0]
                                                                               AR [0]                    ==   +0.005%
                                                                                                               +0.005%    ((   t=
                                                                                                                                t=   +0.02
                                                                                                                                      +0.02   ))
                                                                              AR [+1]
                                                                               AR [+1]                   ==   -1.47%
                                                                                                               -1.47%     ((   t=
                                                                                                                                t=   -4.94
                                                                                                                                      -4.94   ))
                                                                              CAR [-1;+1]
                                                                               CAR [-1;+1]               ==   -1.19%
                                                                                                               -1.19%     ((   t=
                                                                                                                                t=   -2.32
                                                                                                                                      -2.32   ))
                                                                              CAR [-1;+5]
                                                                               CAR [-1;+5]               ==   -2.49%
                                                                                                               -2.49%     ((   t=
                                                                                                                                t=   -3.17
                                                                                                                                      -3.17   ))
                                                                              CAR [-1;+20]
                                                                               CAR [-1;+20]              ==   -6.65%
                                                                                                               -6.65%     ((   t=
                                                                                                                                t=   -4.78
                                                                                                                                      -4.78   ))
                                                                              CAR [-45;-1]
                                                                               CAR [-45;-1]              ==   +2.80%
                                                                                                               +2.80%     ((   t=
                                                                                                                                t=   +1.41
                                                                                                                                      +1.41   ))
                                                                              CAR [-45;-25]
                                                                               CAR [-45;-25]             ==   +3.06%
                                                                                                               +3.06%     ((   t=
                                                                                                                                t=   +2.30
                                                                                                                                      +2.30   ))


        0%
              -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10   -5   0   5   10     15    20       25   30    35     40      45     50      55         60
                                                                                                                                        days




        -5%




       -10%



 Source: own investigation


      Although not statistically significant on the event day [0] (which was probably caused by
       the fact that an AGM lasts the whole day and the first results are not publicly available
       until the next trading day), there is a highly significant negative price reaction of –1.47% on
       the next day (significant at the 99% confidence level and better) (!).

      The following days further resulted in falling prices with CAR [-1;+1] = -1.19% (significant
       at the 98% confidence level), CAR [-1;+5] = -2.49% and CAR [-1;+20] = -6.65% (both
       significant at the 99% confidence level and better). Thus, in contrast to my expectations,
       the evidence suggests that the AGM approvals of share buy-backs resulted in highly
       significant underperforming trends of the underlying companies. This underperformance
       persisted on average during the next 20 trading days following the authorisation event.
       However, at this stage, it is difficult to draw any implications from these results.

      Prior to the AGM approval, a positive price run-up could be observed, starting at the day
       [-45]151: CAR [-45;-1] = +3.09% (significant at near 90%). This run-up is particularly strong

151In fact at the day [-43], however here and below, uniform time periods have been used in order to provide
better comparability with analyses that follow.

                                                                                                                                                   77
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

      between days [-45] and [-25] with CAR [-45;-25]=+3.06% (significant at the 98%
      confidence level). These positive CARs might possibly be accounted for by a higher
      concentration of the announcements of buy-back plans (or possibly other positive news) in
      periods 1-2 months prior to the AGMs themselves - indeed, the agenda of the AGM
      becomes public at least one month before the AGM itself when the AGM invitation is
      being sent round to the shareholders. Moreover, and it seems also very plausible, as the
      majority of the AGMs are concentrated in the „AGM season‟ (April – May), the releases of
      the first quarter results precede the AGMs by exactly 1-2 months. If positive, they are
      often accompanied by buy-back plan announcements and, therefore, could induce positive
      price run-ups within this time period152.

These phenomenon-like results stood in full conflict both with my expectations and with the
market efficiency hypothesis – assuming that the information about the buy-back plans
(although contingent to the AGM approval) is largely on the market and should be reflected in
the share prices.

Below, I have outlined some possible explanations for this phenomenon:

     Firstly, the data quality – as was mentioned above, the confounding news have not been
      removed (and also technically cannot be removed) from the final sample. A far more
      important point of an AGM is certainly how the annual profits shall be distributed, i.e. the
      issue dividends should not be underestimated in this analysis.

     Secondly, the negative reaction described above does not have to be attributed purely to
      the share buy-back approval by the AGM. Otherwise, it might be a set of other (negative)
      information with which a share buy-back might be closely linked (or indicate to).

     Thirdly, the results above (and in particular the relatively long underperforming trend
      following AGM approval) might largely be explained by the sample composition. That
      means, I have here only examined the companies which in fact did buy back their own
      shares later. However, knowing that there is a significant underperformance prior to the
      buy-back announcements – could my results not be just an ex-ante observation of this
      same underperformance, but from the earlier point in time? Thus, this negative price
      impact would be caused, not by the AGM approvals, but by the sample selection design.

In order to investigate the last point, I divided my sample into two major sub-samples:


152Even taken without buy-back plans, after a series of positive quarterly results, the AGM is more likely to
approve the buy-back program.

                                                                                                          78
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

      The data set entries with a time lag between the AGM authorisation and the buy-back
       announcement of less than 100 trading days. In total – 69 AGM approvals.

      The data set entries with a time lag between the AGM authorisation and the buy-back
       announcement of more than 100 trading days. In total – 92 AGM approvals.

The main idea behind this separation is that, expectedly, the first sample of companies would
‟react‟153 more negatively and faster to the authorisations by AGMs than the second sample –
i.e. the underperformance period begins, on average, for first sample companies sooner and
for second sample companies later. The results are visualised in FIGURE 26 below:

      FIGURE 26: ANALYSIS OF NEGATIVE PRICE REACTIONS AFTER AGM APPROVALS

      CAR, %
                                                                                                             >100 DAYS
                                                                                                              >100 DAYS
      10%                                                                                                    N=92
                                                                                                              N=92
                         >100                                                                                AR [0]            =    +0.147%    ((   t= +0.39    ))
                                                                                                              AR [0]            =    +0.147%         t= +0.39
                         Total Sample                                                                        AR [+1]
                                                                                                              AR [+1]          ==   -1.24%
                                                                                                                                     -1.24%    ((   t= -3.29
                                                                                                                                                     t= -3.29   ))
                                                                                                             CAR [-1;+1]       =    -0.86%     ((   t= -1.31    ))
                         <100                                                                                 CAR [-1;+1]
                                                                                                             CAR [-1;+5]       =
                                                                                                                                =    -0.86%
                                                                                                                                    -1.84%     ((
                                                                                                                                                     t= -1.31
                                                                                                                                                    t= -1.84    ))
                                                                                                              CAR [-1;+5]       =    -1.84%          t= -1.84
                                                                                                             CAR [-1;+20]
                                                                                                              CAR [-1;+20]     ==   -4.03%
                                                                                                                                     -4.03%    ((   t= -2.27
                                                                                                                                                     t= -2.27   ))
       5%                                                                                                    CAR [-45;-1]      =    +4.58%     ((   t= +1.81    ))
                                                                                                              CAR [-45;-1]      =    +4.58%          t= +1.81
                                                                                                             CAR [-45;-25]
                                                                                                              CAR [-45;-25]    ==   +5.67%
                                                                                                                                     +5.67%    ((   t= +3.36
                                                                                                                                                     t= +3.36   ))




       0%
                                                                                      -5
             -60

                   -55

                         -50

                                -45

                                      -40

                                            -35

                                                  -30

                                                         -25

                                                               -20

                                                                      -15

                                                                             -10



                                                                                           0

                                                                                               5

                                                                                                   10

                                                                                                        15

                                                                                                              20

                                                                                                                    25

                                                                                                                          30

                                                                                                                               35

                                                                                                                                     40

                                                                                                                                          45

                                                                                                                                                50

                                                                                                                                                        55

                                                                                                                                                                60
                                                                                                                                                        days



       -5%
                               <100 DAYS
                                <100 DAYS
                               N=69
                                N=69
                               AR [0]
                                AR [0]            ==    +0.008%
                                                         +0.008%     ((   t= +0.02
                                                                           t= +0.02   ))
                               AR [+1]
                                AR [+1]           ==    -2.10%
                                                         -2.10%      ((   t= -4.15
                                                                           t= -4.15   ))
                               CAR [-1;+1]
                                CAR [-1;+1]       ==    -1.78%
                                                         -1.78%      ((   t= -2.03
                                                                           t= -2.03   ))
      -10%                     CAR [-1;+5]        =     -3.26%       ((   t= -2.43    ))
                                CAR [-1;+5]        =     -3.26%            t= -2.43
                               CAR [-1;+20]
                                CAR [-1;+20]      ==    -10.02%
                                                         -10.02%     ((   t= -4.21
                                                                           t= -4.21   ))
                               CAR [-45;-1]
                                CAR [-45;-1]      ==    +0.82%
                                                         +0.82%      ((   t= +0.24
                                                                           t= +0.24   ))
                               CAR [-45;-25]
                                CAR [-45;-25]     ==    -0.33%
                                                         -0.33%      ((   t= -0.14
                                                                           t= -0.14   ))



      -15%



 Source: own investigation


      As we see, the two curves feature quite different tendencies – in line with our expectations.
       The first sub-sample (<100 days until buy-back) experiences a much stronger
       underperformance after the AGM approval compared with the second sub-sample (>100


153 Here, the wording „react‟ is, however, not the most appropriate in terms of the hypothesis tested here. Per se,
this would represent not just a reaction, but a simply retrospectively observed development of the share prices –
with reaction probably not caused by the AGM itself.

                                                                                                                                                                     79
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

      days until buy-back): AR1 [+1] = -2.10% vs. AR2 [+1] = -1.24% and CAR1 [-1;+20] = -
      10.02% vs. CAR2 [-1;+20] = -4.03%.

     In addition, I carried out a significance test of differences in reactions to the two
      samples154. The differences in AR [0] between the samples are not statistically significant.
      The abnormal returns AR [+1] are slightly different at the confidence level of 80%; the
      returns CAR [-1;+20] are different with significance of 96%, the returns CAR [-45;-25] are
      different with a confidence of 91%.

     Thus, our „ex-post‟ – hypothesis outlined above can but, rather vaguely, be confirmed.

In an attempt to improve the quality of my evidence, I also created an additional sample of
companies which, although having authorised their share buy-back programs, have not
repurchased their shares yet. The sample contained 30 authorisations by companies of that
kind155. As we would expect, the sample companies should not show any significant price
reaction on the AGM day. In FIGURE 27 below, I present the price development of this sub-
sample („non-repurchasing‟ companies) in comparison to the previous sub-samples:




154 A linear regression method (y=bx + ε) was used with x=0 for the companies in the first sub-sample (buy-backs
within 100 days after the AGM) and x=1 for the companies in the second sub-sample (buy-backs after 100 days
after the AGM). The significance of the b co-efficient is mentioned here.
155 Surely, these 30 authorisations do not cover all the companies which have ever authorised a share buy-back

program. Those were only the companies found accidentally during the creation of the Main Sample.
Understandably, for a more comprehensive investigation a much broader company sample is needed.

                                                                                                             80
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

      FIGURE 27: ANALYSIS OF NON-REPURCHASING COMPANIES

      CAR, %
                                                                                                      NON-REPURCHASING SAMPLE
                                                                                                       NON-REPURCHASING SAMPLE
      20%                                                                                             N=30
                                                                                                       N=30
                                                                                                      AR [0]
                                                                                                       AR [0]          ==   +0.117%
                                                                                                                             +0.117%   ((   t= +0.21
                                                                                                                                             t= +0.21   ))
                                                                                                      AR [+1]
                                                                                                       AR [+1]         ==   -1.20%
                                                                                                                             -1.20%    ((   t= -2.16
                                                                                                                                             t= -2.16   ))
       15%                                                                                            CAR [-1;+1]
                                                                                                       CAR [-1;+1]     ==   -0.27%
                                                                                                                             -0.27%    ((   t= -0.29
                                                                                                                                             t= -0.29   ))
                                                                                                      CAR [-1;+5]
                                                                                                       CAR [-1;+5]     ==   -2.47%
                                                                                                                             -2.47%    ((   t= -1.69
                                                                                                                                             t= -1.69   ))
                                                                                                      CAR [-1;+20]
                                                                                                       CAR [-1;+20]    ==   -7.88%
                                                                                                                             -7.88%    ((   t= -3.04
                                                                                                                                             t= -3.04   ))
                                                                                                      CAR [-45;-1]
                                                                                                       CAR [-45;-1]    ==   +12.60%
                                                                                                                             +12.60%   ((   t= +3.40
                                                                                                                                             t= +3.40   ))
                                                                                                      CAR [-45;-25]    =    +16.95%    ((   t= +6.85    ))
       10%                                                                                             CAR [-45;-25]    =    +16.95%         t= +6.85




       5%



       0%




                                                                               -5
             -60

                   -55

                         -50

                               -45

                                     -40

                                           -35

                                                 -30

                                                       -25

                                                             -20

                                                                   -15

                                                                         -10



                                                                                    0

                                                                                        5

                                                                                            10

                                                                                                 15

                                                                                                       20

                                                                                                            25

                                                                                                                  30

                                                                                                                       35

                                                                                                                            40

                                                                                                                                 45

                                                                                                                                        50

                                                                                                                                               55

                                                                                                                                                        60
                                                                                                                                                 days

       -5%



      -10%

                          Non-Buying
      -15%                >100
                          <100

      -20%



 Source: own investigation


      Similar to the peer samples analysed before, „non-repurchasing‟ companies showed no
       significant reaction on the event day [0]. On the next day, the AR [+1] equalled -1.20%
       which was statistically significant at the 95% confidence level156.

      The sample also experienced a steep decline following the event resulting in a CAR [-
       1;+20] of -7.88% which was highly statistically significant (at the 99% confidence level and
       better). However, this negative trend did not start at the day [0] but consistently persisted
       throughout the time window [-25;+30]. Thus, although significant, the event (AGM
       approval) can only conditionally be attributed to the founding event [0].

      The positive reaction CAR [-45;-25] further increased compared to the peer samples and
       reached +16.95% (significant at the 99% confidence level and better). The most
       astonishing fact was the steep jump on the day [-44] by +4%. By deeper analysis, however,



156Here, I draw your attention to the fact that the event day was not the trigger of a negative reaction although in
the framework of the overall analysis statistically significant, the event does not seem to be as much significant
during a shorter period of time, say [-25;+25] where the negative trend persisted.

                                                                                                                                                             81
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

       this effect was mainly caused by a 98% increase in Hornblower-Fisher AG157. But even
       after removal of the firm from the sample, the price reaction after [-45] remained highly
       positive and significant.

6.3. CONCLUDING REMARKS
      Announcements of (or news on) buy-back plans cause a significant positive price reaction
       to the buy-back announcement. The new price level persists at least during the following
       two months. Thus, the buy-back plan announcements could already be used for the
       purposes of share price management. However, it would be an interesting point to
       investigate how usable it is to implement this instrument on a stand alone basis, i.e.
       without „paying money‟ in a real buy-back.

      The news-releases of buy-back plans should be a subject of an ad-hoc announcement in
       order to prevent possible insider trading.

      There is a significant negative price reaction on the day of the buy-back approval by the
       AGM with a subsequent period of significant underperformance. This behaviour can
       largely be explained by the sample composition – as it mainly consists of the companies
       which in fact did buy back their shares later (ex-ante observation). The reaction of those
       companies which bought back their shares sooner was more negative and steeper.

      Assuming the quality of the sample is sufficient, the sample composition hypothesis does
       not explain the highly significant reaction on the day after the approval ([+1]), however.
       Apart from its statistical significance on a stand-alone basis, the price decrease on the day
       [+1] was, in all cases, statistically different from any price decreases around this date.
       Apparently, this negative reaction should be investigated in light of other events
       confounding an AGM, particularly dividend payments.

      I also would suggest to investigate the following issue in more detail. Possibly assuming
       that the explanation power of the „sample composition‟ hypothesis which I tested above is
       not sufficient enough (the negative trend could be observed for the „later-repurchasing‟
       and even for the „non-repurchasing‟ companies), could the idea of somewhat negative
       effects of AGM buy-back approvals be advanced? In light of this, might not the approval
       of the AGM possibly be considered as a negative signal of the anticipated
       underperformance period in the future? I.e. the management (inside shareholders) need an
       instrument of share price improvement for possible worse times to come. In contrast, the

157   Thus, this largely abnormal increase was not set off by quite a small sample size.

                                                                                                 82
6. Empirical Investigation of Buy-Back Plan Announcements and AGM Authorisations of
Buy-Backs

    companies without an buy-back option are in a position of „burned bridges‟, they do not
    need the option as the underperformance period is not anticipated?

   Also from this perspective, the question deserves a more detailed analysis as to whether the
    capital market, based on the currently available information about a company considering a
    share buy-back, can anticipate the possible fate of the now exercisable option to buy back
    shares – whether the company exercises its right or not?

   The last two points outlined above are quite vague and should not be considered as more
    than just a hint to a possible hypothesis. There is no empirical evidence or second opinion
    to be found in literature on this matter. So I would suggest that a much more
    comprehensive empirical investigation is needed at this stage.




                                                                                             83
7. Results and Conclusions

7. RESULTS AND CONCLUSIONS
Since the ban lift on share buy-backs on 1st May 1998, a large number of German firms158 have
started implementing share buy-back programs. More than 160 of them have already
repurchased their own shares. The abandonment of buy-back restrictions in Germany has been
broadly welcomed by the German business community as a further step towards the
liberalisation of the German capital market. However, different views still exist on the
usefulness of share buy-backs and much criticism is still being expressed towards share-backs
whether or not they represent an effective financial instrument in the tools portfolio of a
company.

An overall analysis of the share buy-back activity in Germany indicates a considerable cyclicity
in the number of share buy-back announcements. This fact apparently points out to a
conscious timing of share buy-backs by German corporations during the financial year.
Moreover, German share buy-back activity appears to be largely dependent on the changing
regulation environment, in particular, with respect to taxation aspects.

In this paper, I focused on examining the share price impact caused by announcements of
stock buy-backs in Germany. In particular, I directed my investigation to the testing of the
signalling/undervaluation hypothesis. In prior empirical research, this hypothesis was found to
be the most powerful for explaining abnormal share price behaviour surrounding share buy-
backs.

The empirical results indicate that the announcements of share buy-backs in Germany are
associated with a positive price reaction, confirming evidence obtained in other countries. For
example, similar to findings from studies using US data, there was a statistically significant
abnormal return of 4.74% upon the announcement of buy-backs in Germany.

Furthermore, it becomes apparent that repurchasing firms on average have suffered a
significant stock price decline prior to the buy-back announcements. The negative
performance particularly prevailed during the last one-two months prior to the buy-back
announcements.

Overall, the evidence from German repurchases is consistent with the signalling/
undervaluation hypothesis which was advocated by the majority of empirical researchers
worldwide. First of all, the positive abnormal returns obtained for the days immediately
following the announcement of buy-backs strongly support this hypothesis.


158   This number is not mentioned here upon my concerns of its correctness.

                                                                                             84
7. Results and Conclusions

Germany provides a unique buy-back research environment, primarily due to the fact that the
undervaluation hypothesis can be tested directly as the vast majority of the German
repurchasing firms reveal their motives for share repurchases. My analysis indicated that
abnormal returns at the time of announcements were more positive if the company stated the
undervaluation rationale among the reasons mentioned in the buy-back announcement.

Absolutely in line with the prediction of the undervaluation hypothesis, further analyses
provided evidence that abnormal returns were negatively related to firm size, as well as market-
to-book ratio. Furthermore, the abnormal returns at the time of announcements were more
positive if the company was traded on the Neuer Markt exchange listing sector.

However, I found no strong evidence that industry affiliation conveys any additional
information which could significantly influence the market perception of share buy-back
announcements.

My results also stress the importance of the ownership structure in determining the market
price reaction to the announcement of a buy-back program in Germany. In particular, the buy-
back announcement plays a positive role for controlled firms. Thus, the firms with the largest
shareholder, controlling more than 50% of the voting rights of the company, induced the
abnormal returns of up to CAR [-1;+5] = +10.13% through the buy-back announcement.
Moreover, more than a half of this reaction was deterred and was first reflected in the share
prices during the days [+1;+5]. In my view, this phenomenon could be possibly explained by a
limited liquidity available on the market for this group of companies.

Contrary to my expectations, however, the role of the management shareholdings, appeared to
be rather ambiguous in Germany – the price reaction to share buy-backs seemed, at least at a
first glance, to be not sensitive to the existence (and even the amount) of management
shareholdings in the repurchasing companies. However, a highly significant relation has been
found between the management shareholding dummy and the absolute magnitude of abnormal
returns. So I came to a conclusion that by far more important, management shareholdings
were for increasing the amplitude of the signal – positive or negative – sent to the capital
market. If the market anticipates „good‟ reasons behind share buy-backs, a management
shareholding would strengthen the credibility of the signal. Otherwise, if e.g. the management
intends to further „entrench‟ in the company, a management shareholding will be considered as
a negative signal. In addition, I found evidence that the existence of any significant
management holding is by far more important for the strength of the signalling effect than the
amount of this holding.



                                                                                             85
7. Results and Conclusions

Moreover, in contrast to the expectations, no evidence has been found that the buy-back signal
stands in any significant relation to cumulative returns prior to buy-back announcement,
percentage of outstanding shares repurchased as well as the financial position of the company.
Moreover, no significant industry-wide effects caused by share buy-backs have been detected.
However, industry-wide effects of share buy-backs could further be investigated in a more
detailed analyis which possibly could yield some interesting results.
By means of a cross-sectional multivariate analysis, only firm size, market-to-book ratio, level
of shareholder control and (contingently) the exchange listing dummy have been found to
induce significant influence on the abnormal returns surrounding the announcement of buy-
backs. Non of the additional variables describing firm‟s governance structure (such as family
control, existence of a second largest controlling shareholder, a foreign investor among
company‟s owners or potential threats of a hostile takeover) have been found to significantly
influence the signalling effect of share buy-backs.
Furthermore, thanks to specific regulations and disclosure requirements in Germany, some
further interesting issues have been investigated in this paper, such as reaction to share buy-
back plan announcements and market reaction to buy-back authorisations by AGMs. The
evidence suggests that there is a significant share price increase upon the first news on existing
share buy-back plans.
An analysis of buy-back approvals by AGMs provided evidence that there is a significant share
price decrease on the day following the AGM session. This evidence led me to a conclusion
that the buy-back approvals by AGMs may not be considered out of the framework of other
confounding events surrounding AGM sessions – in particular, dividend announcements.
Moreover, upon the AGM session, the share price trend features a significant
underperformance during approximately the next 20 trading days. I found evidence that, at
least partially, this phenomenon could be explained by the sample composition, as the sample
under investigation predominantly consisted of companies which have in fact repurchased
their shares later. A deeper analysis of this problem did not, however, absolutely confirm this
suggestion. Thus, the AGM-phenomenon still needs to be investigated in more detail.

There is still a large potential for empirical buy-back research in Germany. Many issues have
been neither covered by the contemporary empirical German research nor are featured in the
framework of this paper. In my view, an in-depth analysis of AGM buy-back approvals, as well
as the examination of the role of the management shareholding in the framework of share
buy-back announcements could possibly provide some valuable insights into the problem. It
would also be interesting to test German buy-backs upon the validity of other possible buy-

                                                                                               86
7. Results and Conclusions

back hypotheses such as take-over defence, free-cash flow, taxation or creditors expropriation
hypotheses.

As time passes, evidence from Germany would become more and more valuable as it would
provide new insights into the buy-back problematic – both from a scientific, practical and
regulative perspective.




                                                                                           87
8. Appendices


8. APPENDICES
APPENDIX I: COMPARISON OF DIFFERENT EVENT STUDY METHODOLOGIES
My empirical study provides a great opportunity to look more deeply into the problematics of
event study technologies. In the following section, I provide a comparative analysis of three
different types of event study models: (1) Constant mean returns model, (2) Market model and
(3) Multifactor model based on CDAX Industrial sub-indexes.

In my view, the following rationale might be posed why not use a standard market model:

1. Employing the CDAX for the purposes of this buy-back investigation seems not
       convincing. CDAX includes only publicly traded companies. However, the index cannot
       pretend to be the market portfolio.

2. Intuitively, due to high heterogeneity of the sample companies (regarding their size or
       industry affiliations), the CDAX would not correctly reduce the external variance of
       abnormal returns, as probably an industry index would do.

3. Returning to MacKinlay (1999): although “….generally the gains from employing
       multifactor models for event studies are limited (as the marginal explanatory power of
       additional factors is small), the variance reduction would typically be greatest in cases
       where the sample firms have a common characteristic such as e.g. one industry…”159.

4. The observations of the past 2-3 years feature large differences in the development of the
       German capital market across its industrial segments (see FIGURE 28 below). Thus, the
       implementation of a homogeneous benchmark like the CDAX could highly distort the
       results. As we see in the figure below, the CDAX sub-indexes have developed in quite an
       inhomogeneous modus – with respect to both their trends and volatility.




159   See [MacKinlay, 1997], p. 18.

                                                                                             88
8. Appendices


      FIGURE 28: RELATIVE DEVELOPMENT OF CDAX INDUSTRIAL INDEXES

       in %
       450
                                                                                             CDAX COMPOSITE
                                                                                             Automobile
       400                                                                                   Banks
                                                                                             Basic Resources
       350                                                                                   Chemicals
                                                                                             Construction
                                                                                             Consumer Cyclical
       300
                                                                                             Financial Services
                                                                                             Food & Bevereges
       250                                                                                   Industrial
                                                                                             Insurance
       200                                                                                   Machinery
                                                                                             Media
                                                                                             Pharma & Healthcare
       150
                                                                                             Retail
                                                                                             Software
       100                                                                                   Technology
                                                                                             Telecom
        50                                                                                   Transport & Logistics
                                                                                             Utilities

         0
             Oct-98




             Feb-99
             Mar-99




             Oct-99




             Feb-00
             Mar-00




             Oct-00




             Feb-01
             Mar-01




             Oct-01




             Feb-02
             Mar-02
             Sep-98



             Dec-98




             Sep-99



             Dec-99




             Sep-00



             Dec-00




             Sep-01



             Dec-01
             May-98
              Jun-98
               Jul-98




             Nov-98




             May-99
              Jun-99
               Jul-99




             Nov-99




             May-00
              Jun-00
               Jul-00




             Nov-00




             May-01
              Jun-01
               Jul-01




             Nov-01
              Jan-99


             Apr-99




              Jan-00



             Apr-00




              Jan-01


             Apr-01




              Jan-02


             Apr-02
             Aug-98




             Aug-99




             Aug-00




             Aug-01
 Source: Datastream




Thus, a multi-factor model could be implemented based on the industrial indexes for every
company in the sample according to the index affiliation of the company. I.e. the market-
model formula changes from

                                                 Rit = αi + βiRmt + εit

into

                                Rit = αi + βiRind,t + εit, (for every industry separately)

                                      E(εit)=0                     var(εit)=σ2(εit)

Where Rit and Rind,t are the period-t returns on security i and the industry portfolio,
respectively, and εit is the zero mean the disturbance term. αi, βi, σ2(εit) are the parameters of
the market model160.

Thus, the abnormal returns for the event window can be similarly calculated:

                                 AR=Rit - αi + βiRind,t (for every industry separately)




160   These parameters are calculated on the basis of the respective industry sub-index.

                                                                                                                     89
8. Appendices


Furthermore, the constant mean return (CMR) model was also included into the comparative
analysis:

                                                                                           Rit = μi + ξit

                                                              E(ξit)=0                                            var(ξit)=σ2(ξit)

Rit is the period-t return on security i and ξit is the disturbance term for security i.

The mean returns have been estimated for the same estimation window, using the geometrical
mean:

                                                                        μi =(P-60/P-270)^(1/211)-1161

In FIGURE 29 below, the Main Sample results of the buy-back announcements, calculated on
the basis of the three methodologies, are presented:

      FIGURE 29: COMPARISON OF EVENT STUDY METHODOLOGY

       10%
                                                                                                                                                 MULTIFACTOR MODEL
                                                                                                                                                  MULTIFACTOR MODEL
                                                                                                                                                 N=189
                                                                                                                                                  N=189
                                                                                     MARKET MODEL
                                                                                      MARKET MODEL                                               AR [0]           =    +4.64%      ((   t= +15.97    ))
                                                                                     N=189                                                        AR [0]           =    +4.64%           t= +15.97
                                                                                      N=189                                                      AR [+1]          =    +1.40%      ((   t= +4.80     ))
                                                                                                                                                  AR [+1]          =    +1.40%           t= +4.80
                   CONSTANT MEAN RETURNS MODEL                                       AR [0]               =    +4.74%     ((   t= +16.37    ))   CAR [-1;+1]
                                                                                                                                                  CAR [-1;+1]     ==   +5.98%
                                                                                                                                                                        +5.98%     ((   t= +11.88
                                                                                                                                                                                         t= +11.88   ))
                    CONSTANT MEAN RETURNS MODEL                                       AR [0]               =    +4.74%          t= +16.37
                   N=189                                                             AR [+1]              =    +1.32%     ((   t= +4.57     ))   CAR [-1;+5]
                                                                                                                                                  CAR [-1;+5]     ==   +6.43%
                                                                                                                                                                        +6.43%     ((   t= +8.36
                                                                                                                                                                                         t= +8.36    ))
                    N=189                                                             AR [+1]              =    +1.32%          t= +4.57
        5%         AR [0]              =    +4.72%           ((   t= +14.99    ))
                                                                                     CAR [-1;+1]
                                                                                      CAR [-1;+1]         ==   +6.00%
                                                                                                                +6.00%    ((   t= +11.96
                                                                                                                                t= +11.96   ))   CAR [-30;-2]
                                                                                                                                                  CAR [-30;-2]    ==   -9.35%
                                                                                                                                                                        -9.35%     ((   t= -5.97
                                                                                                                                                                                         t= -5.97    ))
                    AR [0]              =    +4.72%                t= +14.99         CAR [-1;+5]          =    +6.66%     ((   t= +8.69     ))   CAR [-60;-2]
                                                                                                                                                  CAR [-60;-2]    ==   -12.23%
                                                                                                                                                                        -12.23%    ((   t= -5.48
                                                                                                                                                                                         t= -5.48    ))
                   AR [+1]             =    +1.42%           ((   t= +4.50     ))     CAR [-1;+5]          =    +6.66%          t= +8.69
                    AR [+1]             =    +1.42%                t= +4.50          CAR [-30;-2]         =    -9.12%     ((   t= -5.85     ))
                   CAR [-1;+1]         =    +6.19%           ((   t= +11.33    ))     CAR [-30;-2]         =    -9.12%          t= -5.85
                    CAR [-1;+1]         =    +6.19%                t= +11.33         CAR [-60;-2]         =    -11.50%    ((   t= -5.17     ))
                   CAR [-1;+5]         =    +7.29%           ((   t= +8.74     ))     CAR [-60;-2]         =    -11.50%         t= -5.17
                    CAR [-1;+5]         =    +7.29%                t= +8.74
                   CAR [-30;-2]
                    CAR [-30;-2]       ==   -8.08%
                                             -8.08%          ((   t= -4.76
                                                                   t= -4.76    ))
                   CAR [-60;-2]
                    CAR [-60;-2]       ==   -8.29%
                                             -8.29%          ((   t= -3.42
                                                                   t= -3.42    ))

        0%
                                                                                                -8
                                                                                                     -4
             -60
                   -56
                         -52
                               -48
                                     -44
                                           -40
                                                 -36
                                                       -32
                                                              -28
                                                                     -24
                                                                           -20
                                                                                    -16
                                                                                          -12




                                                                                                          0
                                                                                                                4
                                                                                                                     8
                                                                                                                               12
                                                                                                                                    16
                                                                                                                                         20
                                                                                                                                                 24
                                                                                                                                                      28
                                                                                                                                                           32
                                                                                                                                                                 36
                                                                                                                                                                       40
                                                                                                                                                                            44
                                                                                                                                                                                  48
                                                                                                                                                                                        52
                                                                                                                                                                                             56
                                                                                                                                                                                                     60
       -5%




      -10%                 Constant Mean Returns
                           Market Model
                           Multi Factor Model



      -15%


 Source: own investigation




161   211 is the number of days between the day [-270] and the day[-60] plus one.

                                                                                                                                                                                                          90
8. Appendices


Thus, the following results can be derived from the comparison above:

      The calculation according to the CMR-model results in tendency in somewhat higher
       returns particularly for the time windows around the announcement day (i.e. [-1;+1]; [-
       1;+5]). This can obviously be explained by the fact that the CMR-model does not take into
       account the market movements and measures changes in share price behaviour directly
       without reference to the market. Thus, even if we have a company sample which on
       average underperforms during the entire investigation period (our case), this
       underperformance might not be registered by the CMR-method.

      In general, however, all three curves feature very similar trends over the investigation
       period and, particularly, in the key event windows. The underperforming period can be
       clearly distinguished in every model as it is starting up around the days [-31]-[-32]. The
       CAR [-30;-2] differ insignificantly throughout the methods.

One of the major parameters of the quality of an event study model is the latter‟s ability to
reduce the variance (or standard deviation) of the abnormal returns162. In FIGURE 30 below, I
compared the qualities of the three models according to their abilities of reduction of the
standard deviation of the abnormal returns. Different definitions of standard deviation are
used (according to the control period selected, both including and excluding the period close
to the event [-10;+10]). In FIGURE 30 below the gains of using the new model (in comparison
to the model to the left) are shown:




162   See [MacKinlay, 1997], p. 17-18.

                                                                                              91
8. Appendices


      FIGURE 30: COMPARATIVE GAINS FROM USING A MORE COMPREHENSIVE MODEL

                                             Constant Mean
                                             Returns Modell            Market Model            Multi Factor Model
                        Control Period
        STDEV*          [-270;-60]                            0.315%                  0.290%                  0.291%
                        [-270;+60]                            0.417%                  0.402%                  0.399%
                        [-120;+60]                            0.473%                  0.468%                  0.459%
                        [-60;+60]                             0.546%                  0.543%                  0.534%
                        [-30;+30]                             0.713%                  0.713%                  0.701%

        excluding [-10;+10]
        STDEV            [-270;+60]                           0.316%                  0.292%                  0.290%
                         [-120;+60]                           0.290%                  0.272%                  0.261%
                         [-60;+60]                            0.296%                  0.277%                  0.268%
                         [-30;+30]                            0.281%                  0.249%                  0.226%

        STANDARD DEVIATION REDUCTION (Relative to the Previous Model)

        STDEV           [-270;-60]                                 -                  0.025%                 -0.001%
                        [-270;+60]                                 -                  0.014%                  0.003%
                        [-120;+60]                                 -                  0.005%                  0.009%
                        [-60;+60]                                  -                  0.003%                  0.009%
                        [-30;+30]                                  -                  0.000%                  0.012%

        excluding [-10;+10]
        STDEV            [-270;+60]                                -                  0.024%                  0.002%
                         [-120;+60]                                -                  0.018%                  0.011%
                         [-60;+60]                                 -                  0.019%                  0.009%
                         [-30;+30]                                 -                  0.032%                  0.023%


                                         *   Standard deviation

 Source: own investigation


As we can see, the model quality is improving – since, in major cases, the standard deviation of
abnormal returns is being reduced. However, the results above show that the gains from
employing a multifactor model are very limited - there is but only marginal variance reduction
of abnormal returns in comparison with the market model163 (particularly for the time period
of [-270;-60], the estimation window which was used in my empirical investigation). Thus, the
classic market model has been used for the major analyses of my study, because it best
combines such quality factors as precision and simplicity.




163   These results are fully in consensus with [MacKinlay, 1997], p. 18.

                                                                                                                       92
8. Appendices


APPENDIX II: SHARE-BUY BACKS IN GERMANY – THE MAIN SAMPLE
(MAY/1998 – MAY/2002):

             No        Date Announced                         Company              CAR [-1;+1]      No         Date Announced                        Company            CAR [-1;+1]
                   1             1-Jul-98    AGIV                                            2.2%         96          28-Sep-00     Trius                                        34.3%
                   2         11-Nov-98       Schering                                        2.4%         97          10-Oct-00     SW Umwelttechnik                             -1.7%
                   3           7-Dec-98      Koegel Fahrzeugwerke                          21.6%          98          10-Oct-00     Teles                                        11.2%
                   4         29-Dec-98       BHF Bank                                       -1.2%         99          11-Oct-00     Gold-Zack                                     0.1%
                   5           12-Jan-99     Ti Tabbert-Industrie                          30.2%         100          13-Oct-00     AHAG Wertpapierhandelshaus                    7.6%
                   6           13-Jan-99     BASF                                            3.5%        101          20-Oct-00     Sartorius                                    -2.9%
                   7           18-Jan-99     Krones                                        15.2%         102          24-Oct-00     Wella                                         5.4%
                   8          10-Feb-99      Eurobike                                       -3.8%        103          27-Oct-00     Berliner Effektengesellschaft                 5.9%
                   9           9-Mar-99      Bayer                                           7.1%        104          1-Nov-00      Jean Pascale                                 28.4%
                  10         22-Mar-99       Commerzbank                                     2.6%        105          3-Nov-00      Bien-Haus                                     5.9%
                  11          23-Apr-99      Pfeiffer Vacuum                                -6.2%        106          7-Nov-00      Renk                                         13.0%
                  12         17-May-99       Hoechst                                         5.2%        107         10-Nov-00      Birkert & Fleckenstein                       11.7%
                  13          25-Jun-99      Jungheinrich                                    7.2%        108         13-Nov-00      Maxdata                                       4.8%
                  14          28-Jun-99      Lobster Technology Holding                    19.1%         109         16-Nov-00      Gebhard & Schuster                           23.0%
                  15          30-Jun-99      Pongs & Zahn                                   -2.5%        110         17-Nov-00      Easy Software                                -8.1%
                  16           15-Jul-99     TDS Informationstechnologie                     8.3%        111         23-Nov-00      Swing! Enternainment Media                   -6.2%
                  17           21-Jul-99     A S Creation Tapeten                            7.8%        112           4-Dec-00     InnoTec TSS                                   8.2%
                  18           27-Jul-99     SoftM Software und Beratung                     7.8%        113         11-Dec-00      LPKF Laser & Electronics                      6.7%
                  19           28-Jul-99     Augusta Technologie                             1.6%        114         11-Dec-00      TDS Informationstechnologie                  -2.6%
                  20           28-Jul-99     BHW Holding                                     4.7%        115         15-Dec-00      Adolf Ahlers                                  0.0%
                  21           28-Jul-99     Moebel Walther                                  0.1%        116         18-Dec-00      MAN                                          -1.4%
                  22           30-Jul-99     Rinol                                          -4.2%        117         19-Dec-00      Lintec Computer                              -2.7%
                  23           2-Aug-99      Senator Film                                   -2.2%        118         27-Dec-00      Value Management & Reseach                  -14.6%
                  24           6-Aug-99      Adolf Ahlers                                    4.8%        119            4-Jan-01    Edscha                                        8.9%
                  25         13-Aug-99       Turbon                                         -1.8%        120           25-Jan-01    Travel24.com                                 20.7%
                  26         24-Aug-99       Jenoptik                                        2.3%        121           1-Mar-01     Curanum Bonifatius                            1.6%
                  27         26-Aug-99       Bijou Brigitte Modische Accessories            -0.4%        122           2-Mar-01     Mühl Product & Service                        0.5%
                  28            3-Sep-99     Lobster Technology Holding                      3.7%        123         21-Mar-01      Analytik Jena                                -3.7%
                  29            3-Sep-99     SER Systems                                   17.3%         124         21-Mar-01      BKN International                            -0.3%
                  30            6-Sep-99     Gerry Weber International                      -1.2%        125         22-Mar-01      RTL Group                                    -0.9%
                  31            6-Sep-99     infor business solution                       10.1%         126         28-Mar-01      Sartorius                                    32.5%
                  32            7-Sep-99     Hochtief                                        3.5%        127           1-Apr-01     Allianz                                      -2.4%
                  33            8-Sep-99     Berliner Effektengesellschaft                   3.8%        128           1-Apr-01     Muenchener Rueckversicherungs                 0.7%
                  34            9-Sep-99     K+S                                            -5.8%        129           4-Apr-01     Fritz Nols Global Equity                      4.8%
                  35          13-Sep-99      Uzin Utz                                        3.6%        130           5-Apr-01     German Brokers                                4.5%
                  36            1-Oct-99     Concord Effekten                               -1.0%        131          12-Apr-01     Net AG                                       27.8%
                  37            5-Oct-99     Muehlbauer Holding                              3.9%        132           7-May-01     SER Systems                                  27.6%
                  38          14-Oct-99      BID Boerseninformationsdienst                   1.2%        133         17-May-01      d+s online                                   18.0%
                  39         10-Nov-99       Infomatec Integrated Info Sys                  -9.7%        134         17-May-01      Ludwig Beck                                   1.5%
                  40         19-Nov-99       MG Technologies                                 2.5%        135         30-May-01      Web.de                                       -2.1%
                  41         22-Nov-99       Graphisoft                                    36.4%         136          13-Jun-01     I-D Media                                     5.0%
                  42         22-Nov-99       K&M Moebel                                      0.3%        137          13-Jun-01     Lindner Holding                               0.0%
                  43         25-Nov-99       RWE                                             1.0%        138          20-Jun-01     ComRoad                                       0.2%
                  44         29-Nov-99       Creaton                                         5.2%        139          27-Jun-01     AdLINK Internet Media                         4.0%
                  45         30-Nov-99       Mania Technologie                             12.3%         140          28-Jun-01     Feedback                                     10.3%
                  46           1-Dec-99      H.I.S sportswear                                5.1%        141          29-Jun-01     OnVista                                       7.4%
                  47           1-Dec-99      Walter                                        13.7%         142             3-Jul-01   InnoTec TSS                                  -3.4%
                  48           3-Dec-99      Turbon                                          7.7%        143             6-Jul-01   BUS Berzelius Umwelt-Service                  2.0%
                  49         10-Dec-99       Stuttgarter Hofbräu                             8.9%        144           11-Jul-01    OHB Teledata                                  7.8%
                  50         15-Dec-99       Buderus                                        -0.1%        145           11-Jul-01    Wella                                         1.6%
                  51         19-Dec-99       Bertelsmann                                    -0.2%        146           16-Jul-01    HIT Intl. Trading                             6.3%
                  52         20-Dec-99       Gontard & Metallbank                            9.0%        147           20-Jul-01    Gebhard & Schuster                           -5.2%
                  53         22-Dec-99       Mühl Product & Service                         -3.1%        148           20-Jul-01    RWE                                          -1.7%
                  54           24-Jan-00     Krones                                        12.7%         149           24-Jul-01    Cancom IT Systeme                             8.7%
                  55           25-Jan-00     Cenit Systemhaus                                8.3%        150           24-Jul-01    i:FAO                                        -1.2%
                  56            4-Feb-00     BERU                                            5.3%        151           24-Jul-01    Lintec Computer                              11.9%
                  57          18-Feb-00      Salzgitter                                      0.0%        152           26-Jul-01    Brainpool TV                                 19.0%
                  58          23-Feb-00      Douglas Holding                                -2.8%        153           30-Jul-01    Euwax Broker                                 -1.1%
                  59          23-Feb-00      Heidelberger Zement                            -1.4%        154           8-Aug-01     K+S                                          -0.7%
                  60          25-Feb-00      Villeroy & Boch                               20.0%         155           8-Aug-01     Turbon                                        7.7%
                  61          29-Feb-00      CeWe Color Holding                              6.0%        156           9-Aug-01     CPU Softwarehouse                            32.3%
                  62           8-Mar-00      BASF                                           -1.3%        157         17-Aug-01      Atoss Software                               30.5%
                  63         20-Mar-00       Bijou Brigitte Modische Accessories             3.1%        158         21-Aug-01      CyBio                                         3.0%
                  64           3-Apr-00      Leifheit                                        2.5%        159         22-Aug-01      schlott sebaldus                              6.4%
                  65           4-Apr-00      AGIV                                            5.2%        160         27-Aug-01      Das Werk                                     -9.1%
                  66          18-Apr-00      Deutsche Bank                                  -6.8%        161         29-Aug-01      Kling Jelko Wertpapierhandelsbank            55.2%
                  67           2-May-00      Siemens                                         0.1%        162            4-Sep-01    Beta Systems Software                        35.1%
                  68           9-May-00      Ceotronics                                      3.3%        163            5-Sep-01    iXOS Software                                15.2%
                  69         11-May-00       Celanese                                        0.5%        164          11-Sep-01     IBS AG                                       26.0%
                  70         29-May-00       Solarworld                                     -5.9%        165          14-Sep-01     Plenum                                       14.9%
                  71            6-Jun-00     SER Systems                                   16.0%         166          21-Sep-01     PC-Ware Information Technology                0.7%
                  72             7-Jun-00    AVA Allgemeine Handelsges.                    10.6%         167           25-Sep-01    Muehlbauer Holding                          27.6%
                  73             8-Jun-00    Hawesko Holding                               -1.5%         168           27-Sep-01    Eckert & Ziegler                            15.7%
                  74           16-Jun-00     Escada                                        19.4%         169           28-Sep-01    SAP                                         -5.2%
                  75           26-Jun-00     K+S                                            3.7%         170             4-Oct-01   Ad Pepper                                   22.7%
                  76           28-Jun-00     Baader Wertpapierhandelsbank                   0.7%         171             9-Oct-01   Deutsche Pfandbrief Bank                    -8.8%
                  77           29-Jun-00     AdCapital                                     23.3%         172           22-Oct-01    DIS Deutscher Industrie Service             -1.6%
                  78           29-Jun-00     Curtis 1000 Europe                             6.1%         173           24-Oct-01    Buderus                                      1.3%
                  79           29-Jun-00     M-Tech Technologie                             7.4%         174           24-Oct-01    DEAG Deutsche Entertainment                 -4.2%
                  80              5-Jul-00   Schuler Pressen                                3.7%         175          15-Nov-01     Plambeck Neue Energien                      -0.7%
                  81              7-Jul-00   Artnet.com                                   -36.2%         176          17-Nov-01     SER Systems                                  0.2%
                  82            11-Jul-00    Centrotec Hochleistungskunststoffe             5.6%         177          30-Nov-01     Mologen Holding                              0.8%
                  83            13-Jul-00    Ludwig Beck                                    6.5%         178          30-Nov-01     USU Openshop                                -2.0%
                  84            14-Jul-00    GESCO Industrie Holding                       12.7%         179            3-Dec-01    Net AG                                      29.8%
                  85            19-Jul-00    B.I.S.                                        -8.4%         180            5-Dec-01    MWG Biotech                                 -7.3%
                  86            19-Jul-00    Continental                                    5.6%         181            6-Dec-01    Beate Uhse                                   4.7%
                  87            25-Jul-00    BUS Berzelius Umwelt-Service                  14.4%         182            21-Jan-02   iXOS Software                               14.8%
                  88            3-Aug-00     Tria Software                                  5.1%         183           13-Feb-02    PC-SPEZIALIST Franchise                     -0.9%
                  89          17-Aug-00      Abacho                                        20.7%         184          18-Mar-02     Girindus                                    21.8%
                  90          18-Aug-00      Odeon Film                                     2.2%         185          27-Mar-02     Entrium Direct Bankers                      13.0%
                  91           12-Sep-00     Gedys Internet Products                        6.4%         186           22-Apr-02    Thiel Logistik                             -18.3%
                  92           13-Sep-00     Pro DV Software                                8.7%         187           23-Apr-02    Eckert & Ziegler                             8.4%
                  93           15-Sep-00     Volkswen                                       7.5%         188           30-Apr-02    BayWa                                       11.0%
                  94           22-Sep-00     E.On                                           6.0%         189          13-May-02     Renk                                        -4.3%
                  95           26-Sep-00     Phoenix                                        6.6%



 Source: own investigation



                                                                                                                                                                                         93
8. Appendices


APPENDIX III: DECILE ANALYSIS OF ANNOUNCEMENTS RETURNS BY COMPANY
SIZE

      FIGURE 31: DECILE ANALYSIS OF ANNOUNCEMENT RETURNS BY THE COMPANY SIZE
 EV164 [-39;19]                                                                                         EV [19;36]

                                                    Decile 1                                                                                                        Decile 2
       25%                                                                                                25%

       20%            AR [0]             CAR [-1;+1]           CAR [-1;+5]                                20%               AR [0]                     CAR [-1;+1]             CAR [-1;+5]
                          14.80%              15.75%                 18.69%                                                      8.29%                        9.68%                  8.68%
       15%                                                                                                15%                      7.21 ***                     4.86 ***               3.08 ***
                            11.01 ***            6.76 ***               5.68 ***
       10%                                                                                                10%

        5%                                                                                                 5%

        0%                                                                                                 0%




                                                                                                                                                                     -5
                                                                                                                                                                            0
                                                                                                                                                                            5
                                                                                                                                                                           10
                                                                                                                                                                                      15
                                                                                                                                                                                      20
                                                                                                                                                                                      25
                                                                                                                                                                                      30
                                                                                                                                                                                                       35
                                                                                                                                                                                                       40
                                                                                                                                                                                                       45
                                                                                                                                                                                                                      50
                                                                                                                                                                                                                      55
                                                                                                                                                                                                                                60
                                                                                                                -60
                                                                                                                      -55
                                                                                                                             -50
                                                                                                                             -45
                                                                                                                             -40
                                                                                                                                           -35
                                                                                                                                                 -30
                                                                                                                                                 -25
                                                                                                                                                 -20
                                                                                                                                                              -15
                                                                                                                                                                    -10
                                               -5
                                                             0
                                                             5
                                                            10
                                                                     15
                                                                     20
                                                                     25
                                                                     30
                                                                                   35
                                                                                   40
                                                                                   45
                                                                                             50
                                                                                                  55
                                                                                                  60
              -60
              -55
                      -50
                      -45
                      -40
                                 -35
                                 -30
                                 -25
                                 -20
                                              -15
                                              -10




       -5%                                                                                                -5%

       -10%                                                                                              -10%

                                                                                                         -15%
       -15%
                                                                                                         -20%
      -20%
                                                                                                         -25%
      -25%




 EV [37;54]                                                                                             EV [55;80]

                                                    Decile 3                                                                                                    Decile 4
       25%                                                                                               25%

       20%           AR [0]             CAR [-1;+1]            CAR [-1;+5]                               20%           AR [0]                          CAR [-1;+1]             CAR [-1;+5]
                          5.16%               5.82%                  7.63%                                                     4.38%                         5.14%                   5.41%
       15%                  5.22 ***            3.40 ***               3.15 ***                          15%                     4.12 ***                      2.79 ***                2.07 **

       10%                                                                                               10%

       5%                                                                                                 5%

       0%                                                                                                 0%




                                                                                                                                                                     -5
                                                                                                                                                                           0
                                                                                                                                                                                5
                                                                                                                                                                               10
                                                                                                                                                                                      15
                                                                                                                                                                                             20
                                                                                                                                                                                             25
                                                                                                                                                                                                  30
                                                                                                                                                                                                       35
                                                                                                                                                                                                            40
                                                                                                                                                                                                                 45
                                                                                                                                                                                                                      50
                                                                                                                                                                                                                           55
                                                                                                                                                                                                                                60
                                                                                                                -60
                                                                                                                      -55
                                                                                                                            -50
                                                                                                                            -45
                                                                                                                                     -40
                                                                                                                                           -35
                                                                                                                                                 -30
                                                                                                                                                       -25
                                                                                                                                                       -20
                                                                                                                                                              -15
                                                                                                                                                                    -10
                                                          -5
                                                           0
                                                               5
                                                                   10
                                                                   15
                                                                         20
                                                                              25
                                                                              30
                                                                                   35
                                                                                        40
                                                                                        45
                                                                                             50
                                                                                             55
                                                                                                   60
              -60
                    -55
                    -50
                          -45
                          -40
                          -35
                                       -30
                                       -25
                                       -20
                                              -15
                                                    -10




       -5%                                                                                                -5%

      -10%                                                                                               -10%

      -15%                                                                                               -15%

      -20%                                                                                               -20%

      -25%                                                                                               -25%




 Source: own investigation




  Enterprise Value. EV = Equity Value (Market Cap) + Minority Interests + Interest Bearing Debt – Cash and
164

Cash Equivalents.

                                                                                                                                                                                                             94
8. Appendices



   FIGURE 31 (CONT’D): DECILE ANALYSIS OF ANNOUNCEMENT RETURNS BY THE COMPANY SIZE
 EV [85;133] (in €m)                                                                                            EV [133;236] (in €m)

                                                     Decile 5                                                                                                              Decile 6
    25%                                                                                                          25%

    20%          AR [0]                     CAR [-1;+1]             CAR [-1;+5]                                  20%            AR [0]                             CAR [-1;+1]             CAR [-1;+5]
     15%                2.21%                     5.37%                      5.31%                                                         4.88%                          5.60%                    5.10%
                                                                                                                 15%
                          2.64 ***                  3.70 ***                   2.58 ***                                                      4.82 ***                       3.19 ***                 2.06 **
     10%                                                                                                         10%
     5%
                                                                                                                  5%
     0%
                                                                                                                  0%


                                                               -5
                                                                0
                                                                    5
                                                                        10
                                                                        15
                                                                                 20
                                                                                      25
                                                                                      30
                                                                                           35
                                                                                                40
                                                                                                45
                                                                                                      50
                                                                                                      55
                                                                                                           60
           -60
                 -55
                 -50
                        -45
                        -40
                        -35
                                      -30
                                      -25
                                      -20
                                                   -15
                                                         -10
     -5%




                                                                                                                                                                                 -5
                                                                                                                                                                                       0
                                                                                                                                                                                       5
                                                                                                                                                                                              10
                                                                                                                                                                                                   15
                                                                                                                                                                                                        20
                                                                                                                                                                                                        25
                                                                                                                                                                                                               30
                                                                                                                                                                                                                    35
                                                                                                                                                                                                                         40
                                                                                                                                                                                                                         45
                                                                                                                                                                                                                                   50
                                                                                                                                                                                                                                   55
                                                                                                                                                                                                                                             60
                                                                                                                        -60
                                                                                                                              -55
                                                                                                                              -50
                                                                                                                                          -45
                                                                                                                                          -40
                                                                                                                                                      -35
                                                                                                                                                            -30
                                                                                                                                                                  -25
                                                                                                                                                                  -20
                                                                                                                                                                         -15
                                                                                                                                                                         -10
                                                                                                                  -5%
    -10%
                                                                                                                 -10%
    -15%
                                                                                                                 -15%
    -20%
    -25%                                                                                                         -20%

    -30%                                                                                                         -25%




 EV [236;371]                                                                                                   EV [382;771]

                                                         Decile 7                                                                                                              Decile 8
     25%                                                                                                         25%           AR [0]                             CAR [-1;+1]          CAR [-1;+5]
                                                                                                                                          3.00%                         6.09%                   8.80%
     20%          AR [0]                    CAR [-1;+1]             CAR [-1;+5]                                  20%                        4.14 ***                      4.85 ***                4.96 ***
                        2.98%                     4.59%                      5.60%
     15%                  3.33 ***                  2.96 ***                   2.56 ***                          15%

     10%                                                                                                         10%

      5%                                                                                                          5%

      0%                                                                                                          0%




                                                                                                                                                                                 -5
                                                                                                                                                                                       0
                                                                                                                                                                                            5
                                                                                                                        -60
                                                                                                                              -55
                                                                                                                                    -50
                                                                                                                                    -45
                                                                                                                                                -40
                                                                                                                                                      -35
                                                                                                                                                            -30
                                                                                                                                                                  -25
                                                                                                                                                                  -20
                                                                                                                                                                         -15
                                                                                                                                                                                -10




                                                                                                                                                                                           10
                                                                                                                                                                                                   15
                                                                                                                                                                                                        20
                                                                                                                                                                                                        25
                                                                                                                                                                                                               30
                                                                                                                                                                                                                    35
                                                                                                                                                                                                                         40
                                                                                                                                                                                                                              45
                                                                                                                                                                                                                                   50
                                                                                                                                                                                                                                        55
                                                                                                                                                                                                                                             60
                                                               -5
                                                                    0
                                                                    5
           -60
                 -55
                       -50
                       -45
                             -40
                             -35
                                      -30
                                            -25
                                            -20
                                                   -15
                                                   -10




                                                                        10
                                                                             15
                                                                             20
                                                                             25
                                                                                          30
                                                                                          35
                                                                                          40
                                                                                                 45
                                                                                                      50
                                                                                                      55
                                                                                                           60




     -5%                                                                                                          -5%

    -10%                                                                                                         -10%

    -15%                                                                                                         -15%

    -20%                                                                                                         -20%

    -25%                                                                                                         -25%




 EV [865;4,192]                                                                                                 EV [4,586;96,827]

                                                         Decile 9                                                                                                          Decile 10
     25%                                                                                                         25%

                   AR [0]                    CAR [-1;+1]             CAR [-1;+5]                                 20%            AR [0]                             CAR [-1;+1]             CAR [-1;+5]
     15%                   0.75%                   0.69%                     -0.65%                                                        0.75%                          1.01%                    1.65%
                                                                                                                 15%                         1.39 *                         1.08 *                   1.25 *
                             1.09 *                  0.58 *                     -0.39 *
                                                                                                                 10%
      5%
                                                                                                                  5%
                                                                0
                                                                    5
           -60
                 -55
                 -50
                        -45
                        -40
                        -35
                                      -30
                                      -25
                                      -20
                                                   -15
                                                         -10
                                                               -5



                                                                        10
                                                                        15
                                                                                 20
                                                                                      25
                                                                                      30
                                                                                           35
                                                                                                40
                                                                                                45
                                                                                                      50
                                                                                                      55
                                                                                                           60




     -5%                                                                                                          0%
                                                                                                                                                                                       0
                                                                                                                                                                                            5
                                                                                                                        -60
                                                                                                                              -55
                                                                                                                                    -50
                                                                                                                                          -45
                                                                                                                                                -40
                                                                                                                                                -35
                                                                                                                                                            -30
                                                                                                                                                            -25
                                                                                                                                                                      -20
                                                                                                                                                                      -15
                                                                                                                                                                                -10
                                                                                                                                                                                 -5



                                                                                                                                                                                           10
                                                                                                                                                                                                   15
                                                                                                                                                                                                   20
                                                                                                                                                                                                              25
                                                                                                                                                                                                              30
                                                                                                                                                                                                                    35
                                                                                                                                                                                                                         40
                                                                                                                                                                                                                              45
                                                                                                                                                                                                                              50
                                                                                                                                                                                                                                        55
                                                                                                                  -5%                                                                                                                        60
    -15%
                                                                                                                 -10%

                                                                                                                 -15%
    -25%
                                                                                                                 -20%

    -35%                                                                                                         -25%




 Source: own investigation




                                                                                                                                                                                                                          95
References


REFERENCES

[Asquith/Mullis, 1986]   Asquith, P./Mullis, D. Signalling with Dividends, Stock Repurchases and
                         Equity Issues; Financial Management, Vol. 15, 1986, p. 27-44
[Bartov, 1991]           Bartov, E.: Open-Market Repurchases as Signals for Earnings and Risk
                         Changes; Journal of Accounting and Economics, Vol. 14, 1991, p. 275-294
[Bhattacharya/Dittmar,   Bhattacharya, U./Dittmar, A.: Costless Versus Costly Signalling: Theory and
2001]                    Evidence from Share Repurchases; Working Paper, Indiana University, 2001, p.
                         1-34
[Boehmer, 2000]          Boehmer E.: Business Groups, Bank Control and Large Shareholders: an
                         Analysis of German Takeovers; Journal of Financial Intermediation, Vol. 9,
                         2000, p. 117-148
[Brown/Warner, 1980]     Brown, S./Warner, J.: Measuring Security Price Performance; Journal of
                         Financial Economics, Vol. 8, 1980, p. 205-258
[Brown/Warner, 1985]     Brown, S./Warner, J.: Using Stock Returns: The Case of Event Studies;
                         Journal of Financial Economics, Vol. 14, 1985, p. 3-31
[Bradley/Wakeman,        Bradley, M./Wakeman, L.: The Wealth Effects of Targeted Share
1983]                    Repurchases; Journal of Financial Economics, Vol. 11, 1983, p. 301-328
[Chan et al, 2001]       Chan, K./Ikenberry, D./Lee, I.: Do Firms Knowingly Repurchase Stock for
                         Good Reason?; Working Paper, National Taiwan University, 2001, p. 1-30
[Comment/Jarrell,        Comment, R./Jarrell, G.: The Relative Signalling Power of Dutch-Auction
1991]                    and Fixed-Power Self-Tender Offers and Open-Market Share Repurchases;
                         Journal of Finance, Vol. 46, 1991, p. 1243 - 1271
[Cook/Krigman, 2001]     Cook, D./Krigman, C.: An Analysis of SEC Guidelines for Executing Open
                         Market Repurchases; Working Paper, University of Arizona, 2001, p. 1-38
[Cook et al, 2000]       Cook, D./Krigman, L./Leach, C.: On the Timing and Execution of Open
                         Market Repurchases; Working Paper, University of Arizona, 2000, p. 1-27
[Copeland/Weston,        Copeland, T./Weston, J.: Financial Theory and Corporate Policy; 3. ed.
1992]                    Edison Wesley, Chapters 19 and 20, 1992
[DAI, 1999]              DAI (Deutsches Aktieninstitut e.V.): Der Erwerb eigener Aktien in
                         Deutschland; Deutsches Aktieninstitut e.V. Frankfurt am Main, 1999, p. 1-18
[Dann, 1981]             Dann, L.: Common Stock Repurchases: an Analysis or Return to
                         Bondholders and Stockholders; Journal of Financial Economics, Vol. 9, 1981, p.
                         113-138
[Dann/DeAngelo,          Dann, L./DeAngelo, H.: Corporate Financial Policy and Corporate Control:
1988]                    a Study of Defensive Adjustments in Asset and Ownership Structure; Journal
                         of Financial Economics, Vol. 20, 1988, p. 87-127
[Davidson/Garrison,      Davidson, W./Garrison, S.: The Stock Market Reaction to Significant
1989]                    Tender Offer Repurchases of Stock: Size and Purpose Perspective; Financial
                         Review, Vol. 24, 1989, p. 93-125
[Denis, 1990]            Denis, D.: Defensive Changes in Corporate Payout Policy: Share
                         Repurchases and Special Dividends; Journal of Finance, Vol. 45, 1990, p.
                         1433-1456
[Ding, 1996]             Ding, D.: The Effects of Common Stock Repurchases on Shareholder
                         Wealth; Working Paper, Nanyang Technological University, 1996, p. 1-15
[Evans et al, 2000]      Evans, J./Evans, R/Gentry, J.: The Decision to Repurchase Shares: A Cash
                         Flow Story; Working Paper, Curtin University of Technology, 2000, p. 1-28

                                                                                                     96
References


[Fried, 2001]             Fried, J.: Open Market Repurchases: Signalling or Managerial
                          Opportunism?; Theoretical Inquiries in Law, Vol. 2, 2001, p. 865-894
[Gelb, 2000]              Gelb, C.: Payout Composition and Investors‟ Reaction to Dividend and
                          Stock Repurchase Announcements; Working Paper, Seton Hall University,
                          2000, p. 1-16
[Ginglinger/L‟Her         Ginglinger, E./L‟Her, J.: Ownership Structure and Open Market Stock
2002]                     Repurchases in France; Working Paper, University of Paris, 2002, p. 1-24
[Green/Hollifield,        Green, R./Hollifield, B.: The Personal-Tax Advantages of Equity; Working
1999]                     Paper, Carnegie Mellon University, 1999, p. 1-40
[Grullon/Michaely,        Grullon, G./Michaely, R.: The Information Content of Share Repurchase
2000]                     Programs; Working Paper, Rice University, 2000, p. 1-33
[Guay/Harford, 2000]      Guay, C./Harford, J. : The Cash Flow Permanence and Information
                          Content of Dividend Increases vs. Repurchases; Journal of Financial Economics,
                          Vol. 57, 2000, p. 385-416
[Hillebrandt/             Hillebrandt, F./Schremper, R.: Analyse des Gleichbehandlungsgrundsatzes
Schremper, 2001]          beim Rückkauf von Vorzugsaktien; Betriebs-Berater (BB), Vol. 56, 2001, p.
                          533 - 538
[Hirtle, 2001]            Hirtle, B.: How Do Stock Repurchases Affect Bank Holding Company
                          Performance?; Working Paper, Federal Reserve Bank of New York, 2001, p. 1-47
[Hertzel et al, 2000]     Hertzel, M./Naveen, L.: Stock Repurchase Tender Offer Announcements
                          and Revaluation of Prior Accounting Information; Working Paper, Arizona
                          State University, 2000, p. 1-32
[Hertzel/Jain, 1991]      Hertzel, M./Jain, P.: Earnings and Risk Changes Around Stock Repurchase
                          Tender Offers; Journal of Accounting and Economics, Vol. 14, 1991, p. 253-274
[Ikenberry et al, 2000]   Ikenberry, D./Lakonishok, J./Vermaelen, T.: Stock Repurchases in Canada:
                          Performance and Strategic Trading; Journal of Finance, Vol. 55, 2000, p. 2373-
                          2397
[Ikenberry et al, 1995]   Ikenberry, D./Lakonishok, J./Vermaelen, T.: Market Underreaction to
                          Open-Market Share Repurchases; Journal of Financial Economics, Vol. 39,
                          1995, p. 181 – 208
[Ikenberry/Vermaelen,     Ikenberry, D./Vermaelen, T.: The Option to Repurchase Stock; Financial
1996]                     Management, Vol. 25, 1996, p. 9 -24
[Jagannathan et al,       Jagannathan, M./Stephens, C./Weisbach, M.: Financial Flexibility and the
2000]                     Choice Between Dividends and Stock Repurchases; Journal of Financial
                          Economics, Vol. 57, 2000, p. 355-384
[Jagannathan/             Jagannathan, M./Stephens, C.: Motives for Multiple Open-Market
Stephens, 2001]           Repurchase Programs; Working Paper, University of Missouri-Columbia, 2001, p.
                          1-29
[Jensen, 1986]            Jensen, M.: Agency Costs of Free Cash Flow, Corporate Finance and
                          Takeovers; American Economic Review, Vol. 76, 1986, p. 323-329
[Kahle, 2002]             Kahle, K.: When a Buy-Back Isn‟t a Buy-Back: Open Market Repurchases
                          and Employee Options; Journal of Financial Economics, Vol. 63, 2002, p. 235-
                          261
[Kale et al, 1989]        Kale, J./Noe T./Gay G.: Share Repurchase through Transferable Put
                          Rights: Theory and Case Study; Journal of Financial Economics, Vol. 76, 1989,
                          p. 141-160
[Klein/Rosenfeld,         Klein, A./Rosenfeld, J.: The Impact of Targeted Share Repurchases on the
1988]                     Wealth of Non-Participating Shareholders; Journal of Financial Research, Vol.

                                                                                                      97
References

                          11, 1988, p. 89-122
[Lamba/Ramsay, 2000]      Lamba, A./Ramsay, I.: Share Buy-Backs: an Empirical Investigation;
                          Working Paper, The University of Melbourne, 2000, p. 1-34
[Lamba, 2000]             Lamba, A.: Share Buy-Backs in Highly Regulated and Less Regulated
                          Market Environment; Working Paper, The University of Melbourne, 2000, p. 1-36
[Lakonishok/              Lakonishok, J./Vermaelen, T.: Anomalous Price Behaviour Around
Vermaelen, 1990]          Repurchase Tender Offers; Journal of Finance, Vol. 45, 1990, p. 455-477
[Lehman/Weigand,          Lehman, E. and Weigand, J.: Does the Governed Corporation Perform
2000]                     Better? Governance Structures and Corporate Performance in Germany;
                          European Finance Review, Vol. 4, 2000, p 157-195
[Li/McNally, 2001]        Li, K./McNally, W.: The Decision to Repurchase, Announcement Returns
                          and Insider Holdings: a Conditional Event Study ; Working Paper, University of
                          British Columbia, 2001, p. 1-26
[Li/McNally (2), 1999]    Li, K./McNally, W.: Open Market Versus Tender Offer Share Repurchases:
                          a Conditional Event Study; Working Paper, University of British Columbia, 1999,
                          p. 1-32
[Li/McNally (3), 1999]    Li, K./McNally, W.: Information Signalling or Agency Conflicts: What
                          Explains Canadian Open Market Share Repurchases?; Working Paper,
                          University of British Columbia, 1999, p. 1-28
[Liljeblom/Pasternack,    Liljeblom, E./Pasternack, D.: Share Repurchases, Dividends and Options;
2002]                     Empirical Evidence from Finland; Working Paper, Swedish School of
                          Economics and Business Administration, 2002, p. 1-20
[McNally, 1999]           McNally, W.: Open Market Stock Repurchase Signalling; School of Business
                          and Economics, Wilfrid Laurier University, 1999, p. 1-37
[MacKinlay, 1997]         MacKinlay, C.: Event Studies in Economics and Finance; Journal of Economic
                          Literature, Vol. 35, 1997, p. 13-39
[Mikkelson/Ruback,        Mikkelson, W./Ruback, R.: Targeted Repurchases and Common Stock
1991]                     Returns; Rand Journal of Economics, Vol. 22, 1991, p. 544-561
[Moreira/Procianoy,       Moreira, L./Procianoy, J.: Open Market Stock Repurchases at Sao Paulo
2001]                     Stock Exchange – Bovespa; Working Paper, Universidade Federal do Rio Grande
                          do Sul, 2001, p. 1-33
[Netter/Mitchell, 1989]   Netter, J./Mitchell, M.: Stock - Repurchase Announcements and Insider
                          Transactions after the October 1987 Stock Market Crash; Financial
                          Management, Vol. 18, 1989, p. 84-95
[Oded, 2001]              Oded, J.: Why Do Firms Announce Open Market Repurchase Programs?;
                          Working Paper, Boston University, 2001, p. 1-40
[Otchere/Ross, 2000]      Otchere, I./Ross, M.: Do Share Buy-Back Announcements Convey Firm
                          Specific or Industry-Wide Information? A Test of the Undervaluation
                          Hypothesis; Working Paper, University of Melbourne, 2000, p. 1-33
[Pellens/Schremper,       Pellens, B./Schremper, R.: Theorie und Praxis des Aktienrückkaufs in den
2000]                     USA und in Deutschland; Betriebswirtschaftliche Forschung und Praxis (BFuP),
                          Vo. 52, Heft 2, 2000, p. 132-156
[Pellens/Schremper,       Pellens, B./Schremper, R.: Aktienrückkauf - reine Modewelle?; Börsen-Zeitung
1999]                     11/06/1999, No. 109, p. 15
[Pellens/Schremper,       Pellens, B./Schremper, R.: Aktienrückkauf in Deutschland - Eine
1999]                     theoretische und empirische Untersuchung der Erwerbsmotive; Working
                          Paper No. 76 of Institute for Corporate Management and Research, Ruhr-University
                          Bocum, 1999

                                                                                                       98
References


[Persons, 1995]         Persons, J.: Heterogeneous Shareholders and Signaling with Share
                        Repurchases; Journal of Corporate Finance, Vol. 3, 1995, p. 221-249
[Persons, 1994]         Persons, J.: Signalling and Takeover Deterrence with Stock Repurchases:
                        Dutch Auctions versus Fixed Price Tender Offer; Journal of Finance, Vol. 49,
                        1994, p. 1373-1402
[Pugh/Jahera, 1990]     Pugh, W./Jahera, J.: Stock Repurchases and Excess Returns: an Empirical
                        Examination; Financial Review, 1990, Vol. 25, p. 127-142
[Rosen/Helm, 1996]      Rosen, R./Helm, L.: Der Erwerb Eigener Aktien durch die Gesellschaft; Die
                        Aktien-Gesellschaft, Vol. 10, 1996, p. 434-442
[Sharpe et al, 1995]    Sharpe, W./Gordon J./Bailey J.: Investments; Englewood Cliff, Fifth Edition,
                        NJ, 1995
[Shoven/Simon, 1987]    Shoven, J./Simon L.: Share Repurchase and Acquisitions: an Analysis of
                        Which Firms Participate; Working Paper No. 2243, NBER Working Paper
                        Series, 1987, p. 1-38
[Stephens/Weisbach,     Stephens, C./Weisbach M.: Actual Share Reacquisitions in Open-Market
1998]                   Repurchase Programs; Journal of Finance, Vol. 53, 1998, p. 313-333
[Vermaelen, 1984]       Vermaelen, T.: Repurchase Tender Offers, Signalling and Management
                        Incentives; Journal of Financial and Quantitative Analysis, Vol. 19, 1984, p. 163-
                        181
[Vermaelen, 1981]       Vermaelen, T.: Common Stock Repurchases and Market Signalling: an
                        Empirical Study; Journal of Financial Economics, Vol. 9, 1981, p. 139-183
[Volkart, 2001]         Volkart, R.: Aktienrückkäufe und Eigenkapitalherabsetzungen –
                        Wertkonsequenzen aus Analytischer Sicht; Working Paper, University of Zurich,
                        2001, p. 1-24
[Wansley et al, 1989]   Wansley, J./Lane W./Sarkar S.: Management‟s View on Share Repurchases
                        and Tender Offer Premiums; Financial Management Vol. 18, 1989, p. 97-110
[Wansley/Fayez, 1986]   Wansley, J./Fayez E.: Stock Repurchases and Securityholder Returns: a Case
                        Study of Teledyne; Journal of Financial Research, Vol. 9, 1986, p. 179-191
[Weisbenner, 2000]      Weisbenner, S.: Corporate Share Repurchases in the 1990s: What Role Do
                        Stock Options Play?; Working Paper, Federal Reserve Board, 2000, p. 1-42
[Zhang, 2000]           Zhang, H.: Share Repurchases in Japan: Market Reaction and Actual
                        Implementation; Working Paper, Chinese University of Hong Kong; 2000, p. 1-32




                                                                                                       99
Word-of-Honour-Statement (in German)



WORD-OF-HONOUR-STATEMENT (IN GERMAN)


      „Ich versichere hiermit ehrenwörtlich, daß ich die vorliegende Arbeit selbständig und
      nur unter Benutzung der angegebenen Literatur und Hilfsmittel angefertigt habe.
      Wörtlich übernommene Sätze und Satzteile sind als Zitat belegt, andere Anlehnungen
      hinsichtlich Aussage und Umfang unter Quellenangabe kenntlich gemacht. Die Arbeit
      hat in gleicher oder ähnlicher Form noch keiner Prüfungsbehörde vorgelegen.“




             Ort und Datum                                             Unterschrift




                                                                                       100

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:18
posted:9/16/2011
language:English
pages:100