The short term wealth effects of mergers and acquisitions in the Telecommunication Industry

The short term wealth effects of mergers and acquisitions

in the Telecommunication Industry

Evidence from the U.S. and European telecom market

The short term wealth effects of mergers and acquisitions

in the Telecommunication Industry

Evidence from the U.S. and European telecom market

Abstract

Table of contents 1. Introduction………....4 2. Literature………...6 2.1 Merger waves………..6 2.2 Telecom………...6 2.3 Valuation periods………8

2.4 Europe versus U.S………...9

2.5 Method of payment………...13

2.6 Type of acquisition………13

2.7 Control variable……….………...14

2.8 Hypothesis……….16

3. Data & Methodology………...17

3.1 Data………...17

3.2 Methodology abnormal returns……….……….. 21

3.3 Methodology control variables……….23

4. Results………...25

4.1 Event windows………..25

4.2 Acquiring versus target firms………....26

4.3 Valuation periods………..26

4.4 Europe versus U.S. ………...28

1. Introduction

The telecommunications industry in the U.S. and in Europe has gone through big developments in the last ten years. After the introduction of the Telecommunications Act in the U.S. in 1996, the European Union 1998 deregulation, and the Telecommunications Trade Pact of the World Trade Organization in 1998, there has been a large number of mergers, acquisitions, and partnerships. Beside these deregulations, the high merger activity was driven by globalization, expected synergies and rapid technological innovation (Warf, 2003). After the business services sector (i.e. banking, deregulated in 1994), the telecommunication industry was the largest industry in the fifth merger wave (Fuller et al, 2002).

The demand for telecommunication services and the positive effects of the deregulations and the technological innovation in the telecommunication sector were overestimated (Frieden, 2003). This overestimation of demand and technological innovation resulted in overcapacity and therefore in price pressure on the services of the telecommunication companies. The overpaid acquisitions of telecom companies in the late nineties decreased in value, which finally resulted in the market crash in 2000 (Frieden, 2003). Bouwman et al. (2003) show that shareholders realize positive abnormal returns at the announcement of mergers and acquisitions in high valuations periods, but negative abnormal returns in low valuation periods. However, the long-term value creating acquisitions are made in the low or neutral valuation periods.

The purpose of this paper is to analyze the wealth effects of shareholders of acquiring and target companies in U.S. and European domestic telecom mergers and acquisitions, before and after the market crash in 2000. This is interesting because the telecom industry had a large stake in the last merger wave, but the telecom companies were also the first companies that decreased in value after the market bubble in 2000 (Fuller et al., 2002). Moreover, the market reactions in the telecom industry are seldom studied specifically, since many companies were not publicly traded until the deregulation and privatization in the late nineties. The following research question will be used in this paper:

To which extent do market reactions on the announcement of domestic mergers and acquisitions in the telecom industry differ before and after the market crash in 2000 and what are the differences between U.S. and European market reactions?

To measure the market reactions, the movements of the stock prices of the acquirer, the target and the market have to be analyzed. The data consists of 55 U.S. mergers and acquisitions and 97 European mergers and acquisitions in the telecommunication industry over the period 1996 until 2005. The announcement effects are analyzed by applying the market model (Mackinlay 1997, Brown and Warner, 1985). Furthermore, the OLS regression method is used to analyze the influence of all explanatory variables simultaneously.

This paper presents evidence that shareholders of acquiring firms gain significantly higher short-term abnormal returns in neutral valuation periods (after the market crash) than in the high and low valuation periods. Shareholders of target firms gain significantly higher short-term abnormal returns in the neutral valuation period than in the high valuation period. Furthermore, this paper finds evidence that shareholders of acquiring and target firms realize significant higher abnormal returns in the U.S. than in Europe around the announcement date of mergers and acquisitions.

2. Literature

In section 2.1 a brief overview of the past five merger waves is given. In section 2.2 the question why merger waves occur is answered and literature on mergers in the telecommunication industry is analyzed. Furthermore, reasons behind the market crash in the telecommunication industry are analyzed. Section 2.3 gives an overview of studies on market reactions in different valuation periods and studies on differences between Europe and the U.S. are summarized in section 2.4. Studies on the method of payment of mergers and acquisitions are summarized in section 2.5 and studies on the type of acquisition are described in section 2.6. Section 2.7 summarizes studies that analyze the influence of market-to-book ratios on shareholder returns.

2.1 Merger waves

Past research (Sudarsanam, 2003) on M & A shows that there have been many mergers and acquisitions throughout the years. The United States (U.S.) has the longest history of takeover activities and several patterns, also called merger waves, can be observed throughout the history. The European merger waves show more or less the same pattern as in the U.S. The first merger wave (1880 – 1904) was caused by the industrial revolution which resulted in a horizontal consolidation. Since the creation of monopolies was arrested by the Clayton Act of 1914, a second merger wave (1919 – 1929) occurred which resulted in a move towards an oligopolistic structure in several industries (Sudarsanam, 2003). In order to face the global markets, conglomerates were created in the third merger wave (Peak Europe mid-60s, U.S.-end 60s). The fourth merger wave (1983-1989) was caused by technological progress and focus on new financial markets. Besides that, many divestures occurred during this merger wave. In the last completed merger wave (1993 – 2000) all historical records according to numbers and total values of mergers and acquisitions were broken. The emergence of new technologies, the internet and telecommunications boom and the further globalization were important drivers of this largest merger wave in history (Sudarsanam, 2003).

2.2 Telecom

The telecommunications industry was one of the most active industries on the field of M & A in the 1990s (Fama and French, 1997). What are the reasons behind this high merger activity and what are the main reasons behind the market crash in 2000? This section analyzes the telecommunication industry and provides an answer to this question.

Mitchell and Mulherin (1996) and Fuller et al. (2002) analyze the merger activities in the different industries and for the nineties they observe that the telecommunications industry has a large stake in the mergers merger of the nineties.

Table 2.1Top Five industries in the U.S. based on annual merger activity 1990 – 2000

Industry Number Acquirers % of total Number Targets % of total

Business Services 128 23,7 849 27,1

Telecommunications 64 11,9 330 10,5

Electronical Equipment 36 6,7 193 6,2

Healthcare 36 6,7 193 6,2

Wholesale 29 5,4 192 6,1

Note: Industry classifications are based on Fama and French (1997) Source: Fuller et al. (2002)

Fuller et al. (2002) analyze the successful mergers and acquisitions by industry in the U.S. over the period of 1990-2000. They observe that the telecommunications industry has a large stake in the nineties merger wave. According to Schleifer and Vishny (1988) the high merger activity within an industry is caused by industry shocks. Industry shocks are changes in the industry that have significant impact on the supply and demand in a certain industry. Mitchell and Mulherin (1996) studied the 1980s merger wave and show that deregulation, oil price shocks, foreign competition and financial innovations explain a significant part of the merger activities. Andrade et al. (2001) show that industry shocks create new investments opportunities for the industry and moreover they remove barriers that might have kept the industry dispersed.

Warf (2003) shows that deregulation in the telecom sector was an important driver of the high merger activity in the late nineties. The deregulation in the telecom sector started with the passage of the Telecommunications Act of 1996 in the U.S., the European Union 1998 deregulation, and the World Trade Organization Telecommunications 1998 Trade Pact. These changes in the law had the purpose to lower costs and to increase the service by increasing the competition among telecommunication firms. The increase in competition forced telecom companies to merge or acquire other telecom companies.

Furthermore, Warf (2003) shows that rapid technological innovation had impact on the mergers and acquisitions occurred. The innovations reduced the costs of traditional services and made many new services available at reasonable prices. Another reason is the digitalization of information in related sectors that created a huge demand of electronic networks provided by telecom companies. The consumer and the business telecommunications became closer to computers. This development created business opportunities, but resulted on the other side in a pressure on the prices in the traditional voice telephony. The innovations and the price pressure were two important factors that increased the merger activity.

The telecom industry has drastically changed because of the high merger activity, but what are the main reasons behind the market collapse in 2000? According to Frieden (2003) overestimation of market demand is one of the most important causes of the market collapse in 2000. In the late 1990s market experts predicted a growth of the internet market of 400 % a year, but only 100 % a year was realized1_{. As a result, huge excess capacity in the internet backbones and long distance-transmissions}

were created and once the demand did not match the supply, the telecom sector collapsed. In combination with the excess capacity the pressure on market prices of long distance-transmissions caused huge losses in this business unit. A third reason for the market collapse was the delayed effects of implementation of the Telecom Act of 1996. Implementation of the Act did not mean immediate and complete deregulation and therefore the delay had a negative impact on the companies in the telecommunications industry. The overpaid acquisitions of telecom companies in the late nineties decreased in value and this decrease in value finally resulted in the market crash in 2000 (Frieden, 2003).

2.3 Valuation periods

Do investors in telecom companies react different on mergers and acquisitions before (high valuation period) and after (low valuation period) the market collapse? Bouwman, Fuller and Nain (2003) analyze market reactions of acquires on M & A in low valuation periods (low stock prices), neutral valuation periods and high valuation periods (high stock prices). Bouwman et al. (2003) show shareholders of acquirers react positive on acquisitions made in high valuation periods but negative on acquisitions made in low valuation periods. These announcement effects fade away when stock returns are examined on a longer run. In the three years after the acquisition, shareholders of acquirers in low valuation periods earn significantly higher abnormal returns than shareholders of acquirers in high valuation periods. The causes of these results are possibly laid in managerial hubris and market irrationality. Managers are more optimistic about acquisitions in high valuation periods and they make therefore sub optimal decisions.

The abnormal returns for shareholders of these sub optimal acquirers are significantly lower in the long run than abnormal returns of acquirers in low valuation periods. This means that the market learns about the true quality of these sub optimal acquisition gradually.

According to Rhodes-Kropf and Viswanathan (2004) the synergy effects are overestimated during high valuation periods. This overestimation results in miss pricing of the target and therefore in overpaid acquisitions during high valuation periods. In low valuation periods investors are more critical about M & A and are also more negative about synergies. Shelton (2000) reports evidence that the abnormal returns of acquiring companies fall during neutral valuation periods and shareholders of acquiring companies gain significantly higher returns in high valuation periods. These results show that shareholders of acquiring companies react more positive on the announcement of M & A in high valuation periods.

2.4 Europe versus U.S.

Table 2.2 and table 2.3 present results about the abnormal return for shareholders of acquiring and target firms in Europe and the U.S. These tables show that shareholders of target companies gain significantly (Jensen and Ruback, 1983) and shareholders of acquiring companies sometimes have negative returns and sometimes zero or small positive returns around the announcement date (Fuller et al, 2002, Bradley, Desai & Kim, 1988, Goergen and Renneboog, 2004, Hulle et al., 1991).

Despite table 2.2 and table 2.3 show that results of European and U.S. studies do not differ much, some other researchers find differences in abnormal returns between European and U.S. M & A around the announcement date. Conn and Connell (1990) and Feils (1993) analyze acquisitions between U.S. and UK companies for the period of 1971-1980. They show that shareholders of U.S. target firms gain significantly higher abnormal returns than shareholders of UK target firms. De Long (2003) analyzes market reactions on the announcement of US domestic bank M & A and non-U.S. domestic bank M & A. De Long (2003) shows that shareholders of non-US acquirers earn significantly more than U.S. acquirers and shareholders of non-U.S. targets earn less than their U.S. counterparts. The relative market value of the target to the bidder is of significant negative influence on the cumulative abnormal returns of acquirers and targets. Similar to Feils (1993), De Long (2003) does not find other explaining factors for the differences between the non-US and U.S. announcement effects on M & A.

In common law countries a higher percentage of the total shares outstanding is openly traded on the stock markets than in civil law countries. La Porte et al. (1998) find evidence that countries with poor shareholder protection develop substitute systems to protect the shareholder. Concentrated ownership, mandatory dividends or legal reserve requirements are adoptions to poor legal protection of the shareholders. Furthermore, good accounting standards and shareholder protection measures are associated with a lower concentration of ownership, indicating that concentration is indeed a response to poor investor protection.

Table 2.2 Summary of the European event studies and their explaining factors

Note: The bidder or target is located in the countries mentioned in the third column.

Source: above mentioned studies

Author Period of

study (number of events)

Coun tries

Event window Bidder return % (days) Target return % (days) Explaining factors Goergen & Renneboog (2004) 1990 – 1997 (158) EU -60 to + 60 days 0,70 % (-1,0) 0,40 % (-40,0) 9 % (-1,0) 23 % (-40,0)

Higher abnormal returns for acquirers in case of: stock payments, acquisitions in high valuation periods, minor acquisitions and low market-to-book ratio of the target.

Higher abnormal returns for targets in case of: cash payments, 100 % acquisitions and high market-to-book ratio of the target.

Sudarsanam and Mahate (2003) 1983 – 1995 (519) UK - 1 to + 40 days -1,4 % (1,+1) -1,9 % (+2, +40)

- Value acquirers generate significant higher abnormal returns than glamour acquirers; - Glamour acquirers use significantly more equity compared to value acquirers. - 100 % acquisitions result in significant higher abnormal returns than mergers or minor acquisitions for shareholders of the target; Hulle, Vermaelen and de Wouters (1991) 1970 – 1985 (168) BE 1 month (-30, 0) Takeover: -1 % (-30, 0) Merger: -1,7 % (-30 ,0) Takeover: 5,1 % (-30, 0) Merger:

- 2,3 % (-30 ,0) - In case of mergers significantly more equity is used as the method of payment.

Franks and Harris (1989) 1955-1985 (1898) UK Bidding month (0), -4 months - +1 month

+1% (-30, 0) + 23 % (-30, 0) - Low relative size target/bidder results in higher abnormal returns for shareholders of targets;

- tender offers result in higher abnormal returns than mergers for shareholders of targets and bidders;

Table 2.3 Summary of U.S. event studies and their explaining factors

Author Period of study (nr. events) Countries Event window Bidder (days) Target (days) Explaining factors Jensen and Ruback (1983) 1958 – 1981 US 20 to 60 days 4 % (weighted average of 7 studies) 29 % (weighted average of 7 studies)

- Target gains of unsuccessful bids disappear by the time the failure becomes known. - Successful tender offers result in significant higher abnormal return than mergers for shareholders targets and acquirers. Bradley, Desai & Kim (1988) 1963 – 1984 (236) US 11 days (-5, + 5)

1,14 % (-5, +5) 28,1 % (-5, +5) - Multiple bidders increase abnormal returns for target shareholders and decrease the abnormal returns for bidders shareholders. - Late bidders (white nights) pay too much and have negative abnormal returns Bouwman

et al. (1998)

1979 – 2001 (1973)

US (-1 , +1) 0,54 % (-1, +1) - Higher abnormal returns (AR) for acquirers

during high valuation periods.

- Acquisitions during low valuation periods are value creating on the long-term. Fuller et al (2002) 1990 – 2000 (3135) US 5 days (-2, + 2) All 1,8 % (-2,+2) Public -1% (-2, +2) Private/subsidiary +2,3 % (-2,+2)

- Shareholders of acquiring firms gain significantly higher abnormal returns by acquiring a private target.

- larger relative size of target to the bidder results in more negative abnormal returns to shareholders of the acquirer.

- Acquirers of public targets that use cash as a method of payment generate significant higher abnormal returns than acquirers that use equity as a method of payment.

Travlos (1987) 1972 – 1981 (167) US 21 days (-10, + 10) Cash -0,1 % (-10,+10) Stock -1,6 % (-10,+10)

Shareholders of acquiring and target

companies realize higher abnormal returns in case of 100 % acquisitions and when cash is used as a method of payment.

Note: The bidder or target is located in the countries mentioned in the third column.

2.5 Method of payment

Researchers on M & A have discovered various factors that have significant impact on the cumulative abnormal return of acquiring and target companies. First, most cash-bids generate higher returns for targets as well as bidders than acquisitions where only stock is used. Goergen and Renneboog (2004) find that cash offers trigger significantly higher abnormal returns for targets than equity offers or combined offers. For acquirers the opposite results are found which implies that the choice to finance acquisitions with 100 percent equity does not suggest to shareholders of acquirers that the bidder’s equity is overvalued. Franks and Harris (1988) found significantly higher returns for targets and bidders when cash is used as the method of payment. Hulle et al. (1991) come up with a critical point on the method of payment in M & A. They state that tax considerations play an important role whether shares or cash is used as a method of payment. In some countries (Belgium) cash payments are taxable in case of mergers and therefore most of the time shares are used as a method of payment. If target owners receive new shares in return for their ownership stake, the shareholders delay their tax liability until the position in the bidder is liquidated. Therefore target shareholders might prefer stock payments in case of mergers and acquisitions. The signaling hypothesis states that if an acquisitions announcement will be paid by shares, this may signal to the market that the bidding management thinks that their shares are overvalued (Myers and Majluf, 1984). According to this hypothesis shareholders react negative on 100 percent equity acquisitions and therefore these acquisitions result in lower abnormal returns.

The results of U.S. studies are in line with the results that are found in Europe. Fuller et al. (2002) find significant negative returns for shareholders of firms that acquire public firms and use stock as a method of payment. With cash and combined offers they find insignificant positive returns for bidders. Travlos (1987) and Chang (1998) show also that equity offers result in more negative abnormal returns for shareholders of acquiring firms.

2.6 Type of acquisition

Goergen and Renneboog (2004) do not fully under scribe these results of Franks and Harris. Goergen and Renneboog (2004) find higher abnormal returns for targets in case of hostile bids compared to the abnormal returns for mergers. The opposite result is found for bidders and therefore Goergen and Renneboog conclude that shareholders of target firms are better of in case of mergers.

Jensen and Ruback (1983) analyze seven studies about successful and unsuccessful bidding and target firms around announcements of mergers and tender offers over the period of 1958 – 1981. The firms in the samples of all the studies are US firms. Jensen and Ruback (1983) show that targets of successful mergers and tender offers earn significantly positively abnormal returns around the announcement date of the event and also after completion of the dealing process. For the bidders Jensen and Ruback show that the abnormal returns for successful tender offers are positive, but for mergers they find zero abnormal returns. These results are in line with the European studies that tender offers gain significantly higher abnormal returns than mergers. Travlos (1987) came up with the same conclusion.

2.7 Control variable

Table: 2.4a Results of previous studies for acquirers

Authors Higher abnor-

mal return in high/low/neutral Higher in US or EU Cash /Stock/ mixed payments Higher abnor- mal return with minor or 100 % acq. MTBR to CAAR Deal value Goergen & Renneboog (2004)

High Stock Minor -/-

De Long (2003) EU Bouwman et al.(2000) High Cash Travlos (1987) Cash 100% Franks and Harris (1988) US Cash 100% Rau and Vermalen (1998) 100% -/-

Table: 2.4b Results of previous studies for targets

Cash/stock Minor/ Authors Higher return in

high/low/neutral

Higher in US

orEU Mixed pay. 100 % acq

MTBR to CAAR Deal Value Goergen & Renneboog (2004) Cash 100% + De Long (2003) US Bouwman et al.(2000) Travlos (1987) Cash 100% Franks and Harris (1988) Cash 100% Hulle et al. 100%

Note: The blue squares mean that no significant results were found for the independent variable or that the independent variable is not analyzed.

2.8 Hypothesis

This paper analyzes whether market reactions differ significantly before and after the market crash in the telecommunication industry in 2000. Furthermore, this paper analyzes whether U.S. market reactions on mergers and acquisitions in the telecommunications industry in the U.S. differ from the European market reactions on mergers and acquisitions in the same industry. Besides that, the impact of the method of payment and the type of acquisition on the abnormal returns are analyzed. F, the influence of the market-to-book ratio of the target and the deal value on the abnormal returns of shareholders are analyzed

Based on the analyzed literature the following hypotheses are tested:

- Shareholders of acquiring and target firms gain significantly higher abnormal returns (AR) at the announcement of M & A in high valuation periods than in low valuation periods;

- Shareholders of acquiring firms in the U.S. realize significantly lower AR than shareholders of European acquirers;

- Shareholders of US. targets realize significantly higher AR than shareholders of European targets at the announcement of M & A;

- M & A where cash is used as the method of payment result in significantly higher AR for shareholders of acquiring firms than full equity or mixed payments;

- M & A where cash is used as the method of payment result in significantly higher AR for shareholders of target firms than full equity or mixed payments;

- Shareholders of the acquiring and target firms generate significantly higher AR at the announcement of 100 percent acquisitions than at the announcement of minor acquisitions/mergers;

- The market-to-book ratio of the target firms is of positive significant influence on the AR of acquiring and target firms;

3. Data and Methodology

Within the research about mergers and acquisitions, the event study is the most used study to measure the wealth creation of mergers and acquisitions (Bruner, 2001). Within event studies a distinction can be made between short-term and long-term event studies. In short-term studies researchers analyze the abnormal return of a certain stock in the period surrounding the event. In long-term event studies a much larger time frame (after the event) is chosen to examine long-term stock performance after the event.

This paper uses the market return model (MacKinlay, 1997) to measure the abnormal returns for shareholders of acquiring and target firms around the announcement date of a merger or acquisition. The remaining of this section explains the data and methodology that are used for the analysis of the short-term wealth effects.

3.1 Data

Data on domestic mergers and acquisitions in the U.S. and in continental Europe and the UK are collected from the Zephyr database for the period of 1997 up until 2005. To be included in the sample, the following requirements must be satisfied:

1. The acquirer, target or vendor must be active in the telecommunications industry, defined by the Industry Classification Benchmark of Zephyr;

2. Acquirer must be a listed firm in the U.S. or in Europe; 3. The target must be a listed firm in the U.S. or in Europe;

4. The deal value is twenty five million dollars or more. Deal value is defined as total value paid by the acquirer for the percentage of shares (equity) acquired, fees and expenses excluded.

5. Status of the bid must be completed.

The sample consists of 134 domestic European and 97 domestic American mergers, 100 percent acquisitions and minority acquisitions. The adjusted prices (U.S.Dollar) of stocks of acquirers and targets are collected from the Thomson Datastream database. These prices are adjusted for splits and dividends. The Thomson Datastream database contains information about stock prices of international firms listed on the European as well as on U.S. stock exchanges. Furthermore, the database contains information about de-listed firms that have been acquired. If an acquiring firm bought a division of another firm, the stock prices of the vendor are collected from the database. The following events/firms are excluded from the sample:

1. Firms with stock prices that are not available in the Thomson Datastream database;

3. Firms with low trading volumes that result in small or no fluctuation of the share price in the estimation- and event window.

The final sample includes 97 domestic European events and 55 domestic U.S. events. Domestic mergers and acquisitions are chosen in order to make a proper analysis of the differences between Europe and the U.S.

Table 3.1 Sample characteristics valuation periods

Variable High valuation

(%) Low valuation (%) Neutral Valuation (%) Total Number of events 34 80 38 152 Minority acquisition 16 (47 %) 23 (29 %) 11 (29 %) 51 (34 %) 100 percent acquisition 18 (53 %) 55 (71 %) 27 (71 %) 101 (66 %) Cash payments 21 (62 %) 44 (55 %) 25 (66 %) 90 (59 %) Stock payments 4 (12%) 8 (10%) 2 (5 %) 14 (10 %) Mixed payments 6 (18 %) 18 (25 %) 6 (16 %) 30 (20 %) Unknown payments* 3 (8 %) 8 (10%) 5 (13%) 16 (11 %) Europe 29 (85 %) 43 (54 %) 25 (66 %) 97 (64 %) U.S. 5 (15%) 37 (46 %) 13 (34 %) 55 (36 %)

* Of some events the method of payment is unavailable in the Zephyr Database Source: own calculations

Table 3.1 and table 3.2 show the sample characteristics of the total sample, the differences between the valuation periods and the differences between Europe and the U.S.

Table 3.2 Sample characteristics of M & A in Europe and the U.S.

Variable Europe(percentage) U.S. (percentage) Total(percentage)

Number of events 97 55 152 Minority acquisition 31 (32 %) 20 (36 %) 51 (34 %) 100 percent acquisition 66 (68 %) 35 (64 %) 101 (66 %) Cash financing 57 (59 %) 33 (60 %) 90 (59 %) Stock payments 10 (10%) 5 (9 %) 15 (10 %) Mixed payments 16 (17 %) 14 (26 %) 30 (20 %) Unknown payments* 14 (14 %) 3 (5 %) 17 (11%)

* Of some events the method of payment is unavailable in the Zephyr Database Source: own calculations

In order to be able to test whether market reactions differ before and after the market crashes, the CAR are grouped into three categories. Similar to Bouwman et al. (2003) these categories are high-, low- and neutral valuation periods.

The valuation periods are classified by calculating the average P/E on the Europe or U.S. telecom price index during a certain year. A period is classified as a high valuation period when the yearly average is higher than the three-year average until that year. Figure 3.1 shows that the period from the start of the sample period until March 2000 can be classified as a high valuation period. A year is classified as a low valuation period if the P/E of that period is below the three-year average until that year. Figure 3.1 shows that this below three year average started in March 2000 and ended in January 2002. A year is classified as a neutral valuation period when the average P/E is lower than the average P/E of the high valuation period and higher than the average P/E of the low valuation period. Figure 3.1 shows that the neutral valuation period started in January 2002 and ended at the end of the sample period in 2005.

Figure 3.1 European and U.S. Telecom price indices between January 1997 and December 2005 U.S. 300 Telecom Price Index

0 200 400 600 800 1000 Jan 1997 Jan 2000 Jan 2002 Dec 2005 Time

European 300 Telecom Price Index

0 500 1000 1500 2000 2500 3000 Jan 1997 Jan 2000 Jan 2002 Dec 2005 Time

Similar to Fuller et al. (2002) and Martin (1996), the methods of payment are defined into three categories:

(1) Cash financing. Cash financing includes deals where cash and debt are used as a method of payment.

(2) Stock payments. Stock payments include deals where only shares are used as a method of payment.

(3) Mixed payments. Mixed payments compromises combinations of stocks, cash, and debt as a method of payment.

The methods of payments are collected from the Zephyr database. By including this variable it is possible to control the results for Europe and the U.S. for the method of payment.

Minority acquisitions. Minority acquisitions are defined as mergers and events where the acquirer buys less than 50 percent of the shares of the target company. The type of acquisition data are collected from the Zephyr database. This variable is included in my research, because past research shows that 100 percent acquisitions have greater impact on target and bidder returns (Travlos, 1987; Franks and Harris 1987).

100 percent acquisitions (tender offers). 100 percent acquisitions are defined as acquiring firms buying 100 percent of the shares or acquiring firms that already have a stake in the target and increase their stake above the level of 50 percent.

Market-to-book ratio. Market-to-book ratio is the ratio price to book in Zephyr database. The ratio is defined as the market value of all the shares divided by the total shareholders funds (book value equity plus minority interest) based on the year end book value before the announcement of the bid. This variable is included because the market-to-book ratio contains information about how the target is valued.

The FTSE World total return index is chosen as the benchmark index for calculating the abnormal returns for the events. This benchmark is composed of 2.700 Large/Mid Cap stocks from the FTSE Global Equity Index Series and it is value weighted. This index is chosen because it covers 90-95% of the investable market capitalisation in the world. The total return index calculates the performance of the 2.700 Large/Mid Cap stocks assuming that all dividends and distributions are reinvested. With this index a proper calculation of the abnormal returns can be made since the returns on the stocks are also adjusted for dividends. The benchmark index is collected from the Thomson Datastream database.

Besides the calculation of abnormal returns with the FTSE World total return index the abnormal return of the different stocks are calculated with the FTSE Telecom total return index. The telecom index might be biased because of the overvaluation of the telecom firms and because there are only telecom firms included in this benchmark. Table 3.1-3.4 of the appendix show that the results do not differ significantly with FTSE World total return index as benchmark.

Table 3. 3 Decomposition of independent variable deal value Variable Deal val

Europe Deal val U.S. Deal val. high Deal val. low Deal val. neutral Deal value (mln. Euros) Mean 5,796 3,869 9,085 5,106 1,516 5,099 Median 945 516 976 975 331 812,5 Maximum 204,730 71,762 204,730 71,762 10,181 204,730 Minimum 32 26 33 26 32 26 St. dev. 22,065 10,921 35,070 12,038 2,586 18,790 Nr. of observ. 97 55 34 80 38 152

Source: own calculations

The high deal value in the high valuation period can be explained by the large Vodofone Mannesmann deal of 205 billion Euros. The distribution of the deal value is highly skewed, since the average deal value is 5,099 billion and the median is only 812,5 million Euro.

Table 3. 4 Decomposition of independent variable market-to-book ratio Variable MTBR Europe MTBR U.S. MTBR high MTBR. low MTBR. neutral MTBR Mean 11,0 9,7 12,3 12,7 5,4 10,6 Median 3,8 2,9 5,2 3,7 2,8 3,6 Maximum 180,3 70,0 44,1 180,3 69,1 180,3 Minimum 0,4 0,2 1,1 0,4 0,2 0,2 St. dev. 23,6 16,5 14,6 26,9 12,0 21,5 Nr. of observ. 75 37 24 57 31 112

Source: own calculations

Table 3.4 shows the decomposition of the independent variable market-to-book ratio. Due to a data availability problem not all the market-to-book ratios of the target firms in the sample are available and therefore the number of observations is 112. Table 3.4 shows that the market-to-book ratio is much lower in the neutral valuation period than in the high and low valuation periods. A detailed analysis of the influence of the MTBR and deal value on the CARs of shareholders of acquiring and target firms is made in section 4.7.

3.2 Methodology Abnormal returns

The short-term wealth effects for acquirers and targets are measured by calculating the cumulative average abnormal return (CAAR). The CAAR is the average of the different cumulative abnormal returns (CAR) of the individual events in the sample. Similar to De Long (2003) the abnormal returns are calculated by applying the market model (MacKinlay, 1997):

(1) ARit = Ri – (αi + βi *

R

ARit = abnormal return for stock i at time t Ri = the total return on stock i at time t; and

Rm = the total return on the FTSE World index at time t

Abnormal returns are defined as the difference between the actual return and the expected return of an individual stock. The expected return is estimated by calculating the alpha and the beta of the market model, by using the daily stock- and benchmark return. The beta is calculated as follows:

(2) cov( ,₂ ) ( ) i m i m R R R

β

σ

.

Table 3.5 shows the descriptive statistics of the beta. The parameters of the market model are estimated over the estimation window which starts 285 days prior to the event date and ends 20 days prior to the event date.

Table 3.5 Descriptive statistics of the

β

Mean 0,68 0,51 Median 0,70 0,45 Maximum 2,06 1,78 Minimum -0,14 -0,42 St. dev. 0,04 0,04 Nr. of observ. 152 146

The abnormal returns are calculated for the estimation window and the event window. Table 4.1 shows that the average run up of the abnormal return of the target starts 18 days prior to the event date and the price run up ends one day after the event date. For the acquirer such a pattern is not observable. The abnormal returns for the target are significant on the three days around the announcement date. Because of these data properties the CAARs are analyzed for the event windows: (-1,+1), (-5, +5) and (-18, +1). The CAR is calculated with the following formula:

(3) CAR[-1,+1] = ∑(-1,+1) (Ri– ( αi + βi * Rm ))

In order to test whether the CAR for the acquiring and target companies differ significantly from zero, the t test is applied (equation 5). The t test procedure tests whether the mean of a single variable differs from a specified constant, in this research 0 %. It is not possible to use the z-score test, because the standard deviation of the population (all the telecom mergers occurred) is unknown due to the data availability problem. With the t-test this problem is solved since the t-test uses the variance of the sample2_{. The variances of the individual stock are calculated over the estimation window and}

similar to MacKinlay (1997) the one day sample variance is calculated as follows:

(4)

∑

⋅

=

)

t

,

(t

Var

N

Var

/

1

N = the number of events in the sample

In order to calculate the variances for the different event windows, the one day variance is multiplied by

n

The robustness of equation 4 is checked by calculating the variances of the individual stock over the event with the following equation:

(6)

∑

⋅

=

)

t

,

(t

Var

N

Var

/

1

The results of equation 4 are analyzed in section 4 and the robustness check of equation 6 is presented in table 1 of the appendix. In order to test whether the CAARs of the valuation periods, the deal types or method of payments differ significantly from each other, the following t-test is applied3_:

(7) t = 2 2 2 1 2 1 2 1 2 1

)

(

)

(

n

S

n

S

U

U

X

X

+

−

−

−

X1 = the CAAR of sub sample one X2 = the CAAR of the other sub sample

= the variance of sample one

= the variance of the other sample

2 1

S

S

U1 = the CAAR of population one U2 = the CAAR of the other population

With this t-test for independent populations it is possible to test whether CAAR (mean values) of independent samples differ significantly from each other. The test requires normality of the distribution of the dependent variable. Since the total sample of CAR of acquiring and target firms is split up in smaller groups (i.e. cash, stock, mixed payments), it is possible that the dependent variable is not normally distributed. However, Brown and Warner (1985) show in non-normality occurs, the above mentioned methodology is still valid.

3.3 Methodology control variables

The CARs of acquiring and target companies are regressed on the different independent variables in order to analyze the influence of these independent variables. The ordinary least squares (OLS) regression analysis is used to control the CARs for deal value, method of payment, market-to-book ratio and type of acquisition. The OLS is a model that is linear in mean and variance and requires the dependent and independent variables to be normally distributed.

The Jarque Bera (JB) test is applied to test whether the independent and dependent variables are normally disturbed. Because of outliers in the dependent and independent variables, the variables are not normally distributed. This non – normality problem is solved by deleting the observations that are larger than four times the standard deviation. After deleting the outliers, no observations are larger than four times the new standard deviation. The observations that are deleted from the sample do not exceed three percent of the total sample.

According to MacDonald and Robinson (1985) this is a proper solution as long as the truncation process does not change the results significantly. MacDonald and Robinson (1985) show that it does not matter whether 1 % or 5 % of the total sample is excluded as long as the explaining power remains the same. In this research this has been tested by deleting observations that are larger than four, six and eight times the standard deviation, but the results of the analysis did not change significantly. Furthermore, the power of the model remains the same and therefore this is a valid method on dealing with outliers.

The OLS model assumes that the variance of the errors is constant, also known as homoscedasticity. If the variance of the errors is not constant, this would be known as heteroscedasticity. Since cross sectional data are used, the White (1980) test is applied to check for heteroscedasticity. If heteroscedasticity occurs, the standard error estimates are modified to account for the heteroscedasticity following White (1980). The effect of this correction is that the hypothesis testing is more ‘conservative’, so that more evidence is required against the null hypothesis before it will be rejected4_.

A t-test measures if there is a significant relationship between the dependent and the independent variable. There is a significant relationship between the dependent and the independent variable when the outcome of the t-test for the independent variable is higher than 1.67 and lower than -1.67 in case of 10 degrees of freedom (two-side test).

In section 4.7 the impact for all explanatory factors is estimated simultaneously. Dummy variables are used in the regression analysis in order to test the influence of the method of payment, type of acquisition, geographical location and valuation periods. The dummy variables are used in the same way as deal value and market-to-book ratio of the target.

The model is:

CAR = α + β1 deal value + β2 MTBR + β3 cash + β4 Stock + β5 mixed + β6type of acquisition + β7geographical location + β8high valuation + β9Low valuation + ε

Since minority acquisitions and 100 percent acquisitions are highly related to each other, perfect multicollinearity occurs. In this case, it is not possible to calculate all of the coefficients in the model and therefore the dummy for minority acquisition is excluded from the OLS analysis. A 1 is assigned in case of a 100 % acquisition and a 0 otherwise. For geographical location a 1 is assigned for the U.S. and a 0 for Europe. For the valuation period a 1 is assigned in case of high or low valuation and a 0 for neutral valuation. The multicollinearity problem does not occur with the method of payment, because in 11 % of the cases the method of payment is unknown.

4. Results

In the next sub sections the CAARs of targets and bidders are analyzed. An analysis of the CAARs before and after the market crash is made in section 4.3 and in section 4.4 the CAARs for Europe and the U.S. are analyzed. The influence of the method of payment and type of acquisition are analyzed in section 4.5 and 4.6. All explanatory factors are regressed on the CARs for shareholders of acquiring and target simultaneously in section 4.7 in order to analyze their impact. The regression analysis is made for all event windows.

4.1 Event windows

In this section an analysis of the average abnormal returns for shareholders of acquiring and target firms is made. Table 4.1 shows that the average run up of the abnormal return of the target starts 18 days prior to the event date and the price run up ends one day after the event date. For the acquirer no significant pattern is observable. The abnormal returns for the target are significant on the three days around the announcement date. Because of these data properties the CAARs are analyzed for the event windows: (-1,+1), (-5, +5) and (-18, +1).

Table 4.1 Abnormal returns for acquirers and targets in the event window

Panel A Time AR CAR t-value Panel B AR CAR t-value

-20 0,002 0,002 0,871 -0,001 -0,001 -0,261 -19 0,004 0,007 1,621 0,002 0,001 0,473 -18 -0,005 0,002 -1,826 0,007 0,008 1,635 -17 0,002 0,004 0,779 0,003 0,011 0,650 -16 -0,001 0,003 -0,433 0,002 0,013 0,463 -15 -0,001 0,001 -0,511 0,005 0,018 1,043 -14 0,000 0,001 -0,049 -0,002 0,016 -0,346 -13 -0,003 -0,002 -1,258 0,000 0,016 0,095 -12 0,003 0,000 0,987 0,002 0,019 0,469 -11 0,000 0,001 0,119 0,005 0,024 1,184 -10 0,001 0,001 0,206 0,005 0,028 1,036 -9 0,004 0,005 1,415 0,000 0,028 0,043 -8 -0,002 0,003 -0,791 0,012 0,040 2,703 -7 -0,006 -0,003 -2,131 0,005 0,045 1,026 -6 0,000 -0,003 -0,151 -0,001 0,044 -0,311 -5 -0,003 -0,006 -1,025 0,003 0,047 0,724 -4 -0,001 -0,007 -0,406 0,003 0,050 0,728 -3 0,002 -0,005 0,747 0,005 0,054 1,031 -2 0,000 -0,005 -0,091 0,003 0,057 0,571 -1 -0,001 -0,006 -0,268 0,011 0,068 2,609 0 -0,002 -0,008 -0,788 0,022 0,090 4,896 1 -0,001 -0,008 -0,195 0,015 0,104 3,311 2 0,003 -0,006 0,974 -0,001 0,104 -0,128 3 -0,003 -0,009 -1,133 0,000 0,104 -0,086 4 0,000 -0,009 -0,175 -0,002 0,102 -0,394 5 0,002 -0,007 0,757 -0,001 0,101 -0,275

4.2 Acquiring versus target firms

The second column of table 4.1 shows the CAARs of the acquiring and target firms. Panel A shows that the effect of the announcement of M & A on the value of acquiring firms is small. The significant negative return for shareholders of acquirers in the largest event window seems that the bid was anticipated which might be due to insider trading (Goergen & Renneboog, 2004).

The second column of panel B shows that the shareholders of a target firm realize positive wealth effects when a merger or acquisition is announced. The different event windows show that all the positive CAARs are significant and increase for larger event windows. The CAARs obtained by this research are close to the results of Goergen & Renneboog (2004), Higson and Elliot (1998) and Franks and Harris (1989). Goergen and Renneboog find that targets gain on average 9 % (-1,0) and shareholders of acquirers gain on average 0,7 % (-1,0). Higson and Elliot, and Franks and Harris find significant target gains of respectively 21 % and 30 % in the event month (-20, +5). For acquirers they find insignificant returns around 0 % in the event month (-20, +5).

4.3 Valuation periods

In section 2 an overview is given of the telecom merger wave and furthermore an analysis of the causes of the market crash in 2000 has been made. Based on these findings and the research of Bouwman et al (2003) a segmentation towards high valuation, low valuation and neutral valuation is made. The CAARs in the different valuation periods are presented in table 4.1.

Table 4.1 CAARs in high, low and neutral valuation periods

Event Total High Low neutral High-Low High –neutral Low-neutral

window (t-value)

Panel A: Acquiring firms [-1, +1] -0,3 % (-1,0 ) 0,0 % (0,0) -1,0 %** (-2,0) 0,8 % (1,5) 1,0 %*** (7,4) -0,8 %*** (-5,5) -1,8 %*** (-17,2) [-5, +5] -0,4 % (-0,9) -0,5 % (-0,5) -1,5 %** (-2,0) 1,8 %** (2,4) 1,0 %*** (5,3) -2,3 %*** (-10,7) -3,3 %*** (-22,1) [-18, + 1 -1,5 %** (-2,5) -0,6 % (-0,5) -2,7%*** (-3,2) 0,2 % (0,3) 2,1 %*** (9,7) -0,8 %*** (-3,2) -2,9 %*** (-17,0) Events (% positive) 152 (49 %) 34 (51 %) 80(43 %) 38(58 %) Panel B: Target firms

[-1, +1] 4,8 %*** (8,3) 2,1 %*** (3,3) 4,8 %*** (7,6) 7,1 %*** (3,9) -2,7%*** (-19,9) -5,0%*** (-15,3) -2,3 %*** (-7,4) [-5, +5] 5,8 %*** (7,2) 3,4 %*** (3,9) 5,1 %*** (5,8) 9,4 %*** (3,8) -1,7%*** (-8,6) -6,0 %*** (-13,4) -4,3 %*** (-10,1) [-18, + 1] 10,5 %*** (10,7) 13,0 %*** (12,6) 10,2 *** (10,1) 8,9 %*** (3,1) 2,8 %*** (12,6) 4,1 %*** (7,8) 1,3 %** (2,7) Events 145 (61 %) 32 (61 %) 77(57 %) 36(68%) Average deal value 5,099 9,086 5,106 1,516 3,980 (0,25) 7,570** (2,1) 3,590** (2,1) Note: The average deal value is based on the total sample of the acquiring firms. The number of events in the sample of the target firms differs with the number of events of acquirers because of the data availability problem. ***, **, * significant at the 1 %, 5% and 10 % respectively

Panel A shows that shareholders of acquiring firms realize significantly lower abnormal returns in the low valuation period than in the high or neutral valuation period. Furthermore, panel A shows that shareholders of acquiring firms realize the highest CARs in the neutral valuation periods. 58 percent of the events in the neutral valuation period have a positive abnormal return compared to 43 percent in the low valuation period. In the low valuation period the decrease of the share prices started already 18 days prior to the announcement date. Therefore, the differences between the high and low valuation periods are the largest for the event window of [-18, + 1]. These results are in line with the hypothesis that investors of acquiring companies gain higher announcement wealth effects in high valuation periods. Furthermore, these results are in line with the findings of Bouwman et al. (2003) and Shelton (2000) who find also significant higher short-term wealth effects for acquirers in high valuation periods. However, Bouwman et al. (2003) do not find higher abnormal returns for acquirers in neutral valuation periods. Since the differences between high and low valuation periods are significant H0 can be rejected.

The differences between the valuation periods might be due to the fact that investors know that the better acquisitions are made in low or neutral valuation periods (Bouwman et al, 2003). Moreover, analysts expected the large acquisitions made up until the market crash to have great contributions to the value of the acquirers so this might also explain the differences between the high and low valuation period. Analysts predicted high market growth and therefore acquires could grow by making large acquisitions. According to Frieden (2003) the market growth was not as high as expected and therefore impairments have been made on the overpaid acquisitions in the high valuation periods. These statements are in line with the development of the average deal value during the different valuation periods. Table 4.2 shows that the average deal value is much higher in high valuation periods than in neutral valuation periods. This difference can be explained by the fact that acquisitions in the high valuation period decreased in value after the market crash and were divested in the neutral valuation period (Frieden, 2003).

Panel B shows that 68 percent of the M & A in the neutral valuation period result in positive abnormal returns, compared to 57 percent of the M & A in the low valuation period. Furthermore, shareholders of target firms realize significant higher abnormal returns in neutral valuation periods than in high and low valuation periods. These results are found for the two smallest event windows, but the opposite result is found for the largest event window. This opposite result is caused by the price run up in the high and low valuation period in the 18 days before the announcement date. Since

4.4 European versus U.S.

In this section a distinction is made between the wealth effects of domestic M & A in Europe and in the U.S. Table 4.3 shows the CAARs of acquiring and target firms in Europe and the U.S. From panel A can be concluded that shareholders of European acquiring firms realize wealth losses and shareholders of U.S. acquiring firms gain positive CAARs in all the three event windows. The differences between the European and U.S. CAARs are significant and therefore H0 can be rejected. These results are in contrast with De Long (2003) his analysis of the CAARs in the banking sector. De Long (2003) shows that Non-U.S. acquirers in the baking sector gain significantly higher short-term abnormal returns than U.S. acquirers.

Panel B of table 4.2 shows the CAARs of shareholders of firms that are acquired or firms that divested a business unit. The shareholders of these target firms in Europe as well as in the U.S. gain significant positive abnormal returns for all the three event windows. Similar to the CAARs of acquiring firms, the shareholders of target firms in the U.S. gain higher abnormal returns than shareholders of European target firms. These results are in line with De Long (2003), Conn and Connell (1990) and Feils (1993) and Danbolt (2002) find the same results, but they show that these differences are caused by deal characteristics instead of fundamental differences between Europe and the U.S. Table 4.2 shows the average deal value in Europe is higher than in the U.S.. In section 4,7 will be analyzed whether this has difference has significant impact.

Table 4.2 CAAR of acquiring companies in Europe and the U.S.

Event window

(t- value)

United States (t-value)

U.S. - Europe (t-value) Panel A: Acquiring firms

[-1, + 1] -0,5 % (-1,2) 0,0% (0,1) 0,5% *** (4,3) [-5, + 5] -1,0%* (-1,8) 0,5 % (0,6) 1,5 % *** (11,2) [-18, +1] -2,5 %*** (-3,9) 0,3 % (0,3) 2,8 % *** (17,9) Nr of events( % positive) 97 (48 %) 55 (47 %)

Panel B: Target firms

[-1, + 1] 3,0 %*** (7,0) 7,6 %*** (5,7) 4,6 %*** (24,6) [-5, + 5] 4,1 %*** (6,8) 8,4 %*** (4,5) 4,3 %*** (16,4) [-18, +1] 8,2%*** (11,8) 13,9 %*** (6,4) 5,7 %*** (18,8) Nr of events( % positive) 90 (58 %) 55 (65 %)

Average deal value

The differences between Europe and the U.S. can be explained by financial factors, but in section 2.4 more institutional factors are described. La Porta et al. (1998) show that shareholders are better protected in common law countries than in civil law countries. Since the U.K. is a common law country, a distinction between continental Europe and the U.K. is made. The results are presented in table 2 of the appendix. Table 2 shows that shareholders of U.K. acquirers gain positive CAARs and these returns are significantly higher than those of shareholders of acquirers in continental Europe. Shareholders of U.K. targets gain only higher CAARs than targets in continental Europe for the largest event window. Since shareholders in common law countries (U.K., U.S.) gain significantly higher CAARs than shareholders in continental Europe, the differences in common and civil law system might be an explaining factor for the CAARs. The U.S. and the U.K. have a higher dispersed ownership, a better protection of shareholders and creditors and a takeover regulation with a higher transparency (Goergen and Renneboog, 2004). Because of the higher transparency and better protection, shareholders in common law countries might react positive on the announcement of M & A.

4.5 Method of payment

In section 3.1 a distinction is made between all cash payments, stock payments and mixed payments. The influence of these methods of payments on the CAR is analyzed in more detail in this section. If managers of acquiring firms know that their shares are overvalued, they prefer to finance the acquisitions with their own equity. However, if the management of acquiring firms think that their shares are undervalued, they will prefer to pay the takeover with cash in order keep the benefit of future price risings with the existing shareholders (Goergen and Renneboog, 2004).

The CAARs of shareholders of acquiring and target firms of the different method of payments are shown in table 4.2. Panel A shows that shareholders of acquiring firms gain significantly higher returns when cash is used as a method of payment than in case of stock or mixed payments. These results are found for the two smallest event windows. The opposite result is found for cash and mixed payments in the largest event window.

Table 4.3 CAAR cash, stock and mixed payments

Cash Stock Mixed Cash-Stock Cash-mixed Stock-mixed (t-value) (t-value) Panel A: Acquiring firms [-1, +1] 0,3 % (0,6) 0,5% (0,4) -3,2%*** (-3,1) -0,2% (-0,6) 3,5%*** (17,9) 3,7 %*** (7,1) [-5, +5] -0,1% (-0,2) -2,5% (-1,3) -1,5% (-1,5) 2,4%*** (4,7) 1,4 %*** (7,3) -1,0%* (-1,8) [-18, + 1] -2,4 %*** (-3,5) -3,9 %* (-1,7) -0,9 % (-0,7) 1,5 %** (2,5) -1,5%*** (-6,7) -3,0%*** (-4,8) Events 90 15 30 Panel B: Target firms

[-1, +1] 4,6 %*** (5,2) 13,3 %*** (8,5) 2,9 % *** (3,3) -8,7%*** (-21,0) 1,7%*** (9,1) 10,4 %*** (24,1) [-5, +5] 6,1 %*** (5,0) 14,3 %*** (6,6) 2,6 % ** (2,2) -8,2%*** (-14,3) 3,5 %*** (13,6) 11,7 %*** (19,5) [-18+ 1] 9,4 %*** (6,6) 14,1 %*** (5,6) 12,6% *** (9,0) -4,7%*** (-7,0) -3,2%*** (-10,9) 1,5 %** (2,1) Events 84 15 30 Average deal value 1.847 4.316 15.808

***, **, * Significant at the 1 %, 5 % and 10 % level respectively Source: own calculations

Panel B of table 4.3 shows that shareholders of target firms realize significant wealth gains irrespective of the method of payment. From the differences analysis can be concluded that shareholders of target companies realize significantly higher abnormal returns when stocks are used as a method of payment than with events where cash or mixed payments are chosen as the method of payment. Therefore, the hypothesis which states that cash payments result in higher CARs for target shareholders can be rejected.

4.6 Type of acquisition

In this section the CAARs of the minor acquisitions and 100 percent acquisitions (tender offers) are analyzed in more detail. Panel A of table 4.4 shows that the differences in CAARs between 100 percent acquisitions and minor acquisitions for shareholders are only significant at the largest event window. This positive difference is line with the results of Hulle et al. (1991) and Franks in Harris. They state that the difference is caused by the method of payment. With 100 percent acquisitions more cash is used as the method of payment and in case of mergers or minority acquisitions more stock is used. Table 4.4 shows that this is also the case in this paper. 59 percent of the 100 percent acquisitions are financed by cash compared to 59 percent in case of mergers. Only two percent of the 100 percent acquisitions are financed by stocks compared to 15 percent with minority acquisitions/mergers.

Table 4.4 Deal type returns of target companies 100 % acquisition (t- value) Minor acquisition / mergers (t-value) 100 % - Minor (t-value) Panel B: Acquiring firms

[-1, +1] -0,1 % (-0,2) -0,8 % (-1,5) 0,7 % (1,2) [-5, +5] -0,4 % (-0,7) -0,4 % (-0,6) 0,0 % (0,4) [-18, +1] -0,5 % (-0,7) -3,5 %*** (-4,1) 3,0 %*** (3,2) Panel B: Target firms

[-1, +1] 6,4 %*** (8,2) 1,3 %* (1,8) 5,1 %*** (38,8) [-5, +5] 6,3 %*** (5,8) 4,7 %*** (4,7) 1,6 %*** (8,7) [-18, +1] 9,0 %*** (7,2) 13,6 %*** (11,8) - 4,6 %*** (-21,8) Cash 82 % 59 % Equity 2 % 15 % Mixed 16 % 26 %

***, * Significant at the 1 % and 10 % level respectively Source: own calculations

Panel B shows the wealth effects of shareholders of the target firms. The CAARs of both types of acquisitions are significant for all the three event windows. 100 percent acquisitions result in higher CAARs for target shareholders than minor acquisitions for the smallest event windows. These results are in line with the results of Hulle et al. (1991), Franks and Harris (1989) and Goergen and Renneboog (2004). However, shareholders of target firms realize higher abnormal return in case of minority acquisitions and mergers for the largest event window. This difference is caused by the price run up for minor acquisitions in the 18 days before the announcement date.

Because of this mergers are less unexpected and generate smaller CAARs for shareholders of the target around the announcement date. In the sample of this research acquirers already had a stake in the target company in case of 25 % of the minority acquisitions/mergers. Those toeholds might be an explanation for the lower abnormal returns for acquirers and targets for the smallest event windows.

4.7 Control variables

In this section the cumulative abnormal returns of acquiring and target firms are regressed on the following deal and firm specific characteristics:

1. Deal value of the event;

2. The method of payment (full cash payment, stock payment or mixed payments); 3. The market-to-book ratio of the target (MTBR);

4. The type of acquisition (1 = 100 percent acquisition, 0 = minority acquisition/merger); 5. Geographical location (1 = U.S., 0 = Europe);

6. Valuation period (1 = high, 1= low, 0 = neutral).

These variables are chosen because past research shows (section 2.7) that they are important value drivers of the CAR of acquiring and target firms. Due to the data restrictions it is only possible to analyze the influence of these variables.

Panel A shows the results of the independent variables tested on the CARs realized by shareholders of the acquiring firms. The deal value of the event has negative significant influence on the CARs of acquiring firms for all event windows so H0 can be rejected. Larger mergers or acquisitions have a greater impact on the strategy of the acquiring firm, which results in a greater impact on the share price of the acquirer. Franks and Harris (1989) find that the acquisition of larger targets result in lower abnormal returns to shareholders of acquiring firms. Franks and Harris measure the size of the target relative to the bidder, but due to data availability problem, this research focuses on the deal value. Shareholders of acquiring firms react more negatively on the acquisition of larger targets, because they might be anxious to overpay for the target (Goergen and Renneboog, 2004). These results are in line with panel A of table 4.1 that shows that the short-term wealth effects of acquiring firms in the high valuation period (average deal value ca 9 bln.) are much lower than in neutral valuation periods (average deal value ca. 1,5 bln).

According to Goergen and Renneboog (2004) the shareholders are anxious to overpay because the high market share price is based on future expectations that are not realized. The results of this paper show the opposite result for the smallest event window.

Shareholders of acquirers in the telecom industry prefer acquisitions of targets with high MTBRs because these firms have high growth potential.

The regressions also analyze whether the CARs are influenced by high, low or neutral valuation periods. This paper finds evidence that acquisitions made in the low valuation period (after the market crash) have significant negative influence on the CARs of shareholders of acquiring firms for the two smallest event windows. This negative influence is in line with the results presented in panel A of table 4.1. For the method of payment, type of acquisition and geographical location no significant results on the CARs of shareholders of acquiring firms are found.

Panel B shows that the deal value has also negative influence on the CAARs for shareholders of target firms so H0 can be rejected. Since the deal value is low in neutral valuation periods but also low in Europe, the deal value might explain the higher abnormal returns for shareholders of target firms in Europe and in the neutral valuation period.

Similar to the CARs of the acquirer, the MTBR has significant negative influence on the CARs of the target shareholders for the smallest event window. These results are in contrast with Goergen and Renneboog (2004). The negative influence of the target MTBR implies that the shareholders of these targets might prefer to stay independent because they have high growth opportunities. The results are only significant for the smallest event window.

Stock payments have significant positive influence on the abnormal returns of target firms for the two smallest event windows. These results are in line with the significant differences between cash and stock payments presented in section 4.4. Since in section 4.4 an analysis of the differences has been given, the influence of stock payments is not analyzed in more detail in this section. For the type of acquisition, geographical location and valuation period no significant influence on the CARs of shareholders of target firms are found.

The adjusted R-squared (R2) for the total model is presented in table 4.6. The R2 explains by which percentage the dependent variable is explained by the model (the independent variables). In contrast to R-squared, the adjusted R-squared takes the loss of degrees of freedom into account associated with adding extra variables5_{. In this research the value of R2 explains by how many percent}

the cumulative abnormal return is explained by the deal value, method of payment, market-to-book ratio of the target and type of acquisition. Table 4.6 shows that the explaining power of the model is only significant for the smallest event window. The model explains 6,0 % of the CAR of the acquirer and 10,9 percent of the CARs of the shareholders of the target firm. Goergen and Renneboog (2004) find a R2 of 15,2 % and 22,3 %.

These results are higher because Goergen and Renneboog used nineteen instead of nine variables in their model. Table 4 of the appendix shows an overview of the results of the OLS regression with only the significant variables included. This table shows that the explaining power of the model with only significant variables is higher. This means that adding non significant variables to the model has a decreasing effect on the explaining power of the total model.

Table 4.6 Determinants of short-term wealth effects for acquiring and target firms Acquring firms

Panel A

CAAR [-1, +1] (%) CAAR [-5, +5] (%) CAAR [-18, +1] (%) Value driver Coefficient t-value Coefficient t-value Coefficient t-value Deal value(bln €) -0,026 -2.76*** -0,037 -2,19** -0,060 -3,85*** Cash payment 1,59 1,36 -2,74 -1,21 -1,77 -0,44 Stock payment 2,68 1,17 -3,10 -1,04 -0,88 -0,19 Mixed payment 0,81 0,54 -0,45 -0,13 -1,17 -0,27 Target MTBR 0,033 2,37** 0,06 1,28 -0,009 -0,38 Type of acquisition 0,24 0,27 0,37 0,20 1,11 0,52 U.S. -0,23 -0,27 1,41 0,79 2,73 1,38 High valuation -0,31 -0,21 -1,76 -2,13 -3,66 -1,50 Low valuation -1,63 -1,68* -1,87 -1,83* -2,81 -1,32 Observations 106 109 107 R2 11,4 % 9,9 % 12,2 % Adjusted R2 6,0 % 1,7 % 4,0 % F-statistic 2,11* 1,20 1,50 Target firms Panel B

CAAR [-1, +1] (%) CAAR [-5, +5] (%) CAAR [-18, +1] (%) Value driver Coefficient t-value Coefficient t-value Coefficient t-value Deal value(mln €) -0,024 -3,49*** -0,010 -0,39 -0,057 -2,02* Cash payment -0,01 -0,01 4,07 0,94 1,27 0,27 Stock payment 6,17 1,78* 10,6 2,14 3,67 0,61 Mixed payment -0,20 -0,09 5,23 1,18 5,91 1,05 Target MTBR -0,17 -2,59*** 0,03 0,50 -0,09 -0,50 Type of acquisition 0,40 0,24 1,19 0,38 -1,94 -0,43 U.S. 0,41 0,25 2,94 1,02 3,56 0,83 High valuation -1,68 -0,75 -1,64 -0,53 3,54 0,75 Low valuation -1,94 -1,03 0,45 0,15 8,55 1,93 Observations 97 98 97 R2 19,3 % 7,4 % 1,7 % Adjusted R2 10,9 % 0,5 % 1,0 % F-statistic 2,30** 0,78 2,34

Note: The number of observations includes not the whole sample at the acquiring and target firms, because of the data availability problem of the target MTBR. The number of observations differs between the three event windows of the acquirer because outliers caused a normality problem and are therefore deleted from the sample. These outliers were more than for times the standard deviation and less than three percent of the total sample. ***, **, * significant at the 1 %, 5% and 10 % respectively