The Impact of Cross-Border Acquisitions on Acquirer Shareholder Value for Publicly Traded US Firms

(1)

The Impact of Cross-Border Acquisitions on

Acquirer Shareholder Value for Publicly

Traded US Firms

Jacob Hahn

University of Amsterdam

Bachelors of Business Administration

Specialization – Finance

Thesis Supervisor – Florencio Lopez de Silanes Molina

(2)

Statement of Originality

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

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

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

(3)

Abstract

M&A activity has continuously grown in value, particularly cross-border activity. It growing by 45% in value during 2018. While much research has been conducted on the motives of these acquisitions and conclusive findings exist for target firms, the existing research is

lacking concerning the acquirer. This paper uses an event study methodology to study the impact cross-border acquisitions have on US publicly listed acquiring firm shareholder value. The study found that insiders generate significant negative returns that do not diminish following the announcement of an international takeover. It was concluded that international acquisitions have an overall negative impact on acquiring firm shareholder value.

(4)

Page INTRODUCTION………..………...1 LITERATURE REVIEW...………...3 Overall Acquisition Research………...………..………..4

Isolating International Acquisitions…………..……….………..5

International Diversification….………...5

Macroeconomic Factors……….………..6

Value Building & Business Motives………8

Economies of Scale....………..8

Synergies……… ………...……….…….8

Local Politics & Greenfield ………...…………..………9

Managerial motives ……….………9

Managerial Entrenchment ……….………..………….9

Executive Compensation ………...………...…..10

Prestige and Career Growth ……….……….…………...10

(5)

Insider Trading ………...…………..……….……..11

Post-Announcement Drift …………..……….12

METHODS………..……….….…….13

Conceptual Model………..…..13

Sample collection………..…………...13

Determining of Daily Aggregated Abnormal Return………...……….………..16

Establishing the Event Windows to Test………...……..17

Lead-Up Windows………..……….17

Overall Windows……….18

Post-Announcement Drift Windows………...……….20

Testing the Hypotheses ………21

Additional Analysis………..……….…..22

RESULTS………...……….22

Daily Aggregated Abnormal Returns………..23

Plot of Abnormal Returns with Cumulative Abnormal Returns………..23

Results Against CRSP……….25

Lead-Up to Announcement……….……….25

Overall……….………25

(6)

Takeaway……….………27

Results with FF3 and FF Market Returns………27

Lead-Up To Announcement………...……….27

Overall……….………29

Following Announcement………..………..32

Additional Analysis……….………34

DISCUSSION……….……….35

Takeovers as an expensive proxy for outside board directors……….35

Empire Building………..………36

Synergies, Economies of Scale and Control………...……….36

Failure to properly determine premium…….………..36

Successful Value Chain & Market Control based takeovers……….…..37

Takeaway……….………38

Post-Announcement Returns………...……38

(7)

LIST OF TABLES

Table Page

1 Outline of Creation of Final Sample Set……….14

2 Daily Aggregated Abnormal Returns Against Fama French 3-Factor...23

3 Lead-up CAAR against CRSP………25

4 Overall CAAR Against CRSP………....26

5 Post-Announcement Drift Against CRSP………...27

6 Lead-Up CAAR against all Three Benchmarks……….28

7 Overall CAAR against all Three Benchmarks………...29

8 Post-Announcement Drift against all Three Benchmarks……….32

LIST OF FIGURES Figure Page 1 AAR vs. CAAR………...………...24

2 Lead-up Windows………..18

3 Overall Event Windows……….20

4 Post-Announcement Windows……… ……….….21

(8)

1 Introduction

Mergers and Acquisitions (M&A) is a rather broad and all-encompassing topic, it is in a related nature to the creation of strategic alliances, and joint ventures. All of these share the common basis that they involve the collection of two or more different firms coming together in some form with the often-intended goal to generate growth that would not be able to be achieved if pursued individually. In all of these cases, investors of the involved firms have a solid reason to monitor and keep track of the potential of one of these forms of cooperation due to the great complexity of such endeavors and the ability to offer both abnormal returns but also abnormal losses. This has indeed been confirmed to be the case through years of studies that highlight just how complicated it is to see if the overall activity that occurs, actually generates positive returns for its shareholders. Over time, research has begun to dive deeper into individual differences that can help indicate in which cases a merger may succeed and when it would result in potential destruction of value. Much research has shifted towards looking into the effect takeover attempts have on target firms. It has collectively started to see more results indicating general positive shareholder gains. However, the case for acquiring shareholders’ expectations have been proven to be largely inconclusive and slightly negative at best. Therefore, this paper makes a times series investigation into how the announcement of an acquisition impacts the value of acquiring shareholder’s stake.

With the rise of technology and globalization, cross-border transactions are becoming the dominant type of acquisition as companies seek to increase the rate of penetration into foreign markets, keep up with competitors and gain access to new technology. As such, the research will focus on US publicly listed firms that announce, and eventually complete the takeover of a foreign firm. Thus, the study aims to answer the question;

(9)

2

Overall do cross-border acquisitions generate abnormal returns for the acquiring firm’s shareholders?

When looking into past literature, the methods can be split into two general types; long term (looking years after) and short to medium term. The literature review will discuss how, when accounting for some studied inefficiencies, only a short to the medium-term horizon is needed due to the Efficient Market Hypothesis (EMH). Thus, the study operates on the

assumption that upon new information, the market in a short amount of time should fully reflect the adjusted value through changes in the share price. A study by Malkiel (2003) will serve as the foundation for this belief with a study by Beny and Seyhun (2012), as well as Bernard and Thomas (1989) combined with initial exploratory results providing the justification for including event windows that go beyond the day of the announcement.

Before getting into the analysis, previously studied reasoning for the conducting of acquisitions will be empirically discussed leveraging existing research on both the theoretical reasoning in terms of shareholder value and business reasons, as well as studies that highlight reasonings that focus on agency problems that lead to incentives to acquire firms even if it results in the potential destruction of shareholder value.

Interestingly, this study was started with the belief that with proven importance of technology in generating improved firm performance on an international level and with the existing acquiring firm research being inconclusive when taken as a whole, recent advancements in technology could have provided the opportunity to generate positive returns for shareholders. This led to the hypothesis;

(10)

3

It was believed that overall due to takeover premiums reflecting the value of synergies, acquiring firms at best could only receive small to moderate positive returns. Thus it was thought that a median above zero would be observed but the overall results would be skewed by agency problem-plagued transactions. Instead, the initial beliefs were proven wrong, and this study shows the continued struggle for firms to consistently generate positive returns for shareholders through the usage of acquisitions.

As such, the discussion, following the literature review and results, takes a deep dive into the potential agency problems at play, how to test for them in future studies and the next steps if future research confirms them to be at play. The discussion does also provide some thoughts, on how the returns of the firms can be spread as to still find many firms successfully taking over other firms, again discussing ways to test for this in the future.

The data was collected leveraging the advantages of Zephyr, the global M&A database, and the Center for Research in Securities Prices (CRSP). The list of completed acquisitions and their announcement date in which a firm with no prior ownership acquired one hundred percent of the target was obtained from Zephyr. Following this, the daily stock prices of the acquiring firms were obtained from the Center for Research of Security Prices (CRSP). The CRSP benchmark serves as the primary method of testing in this study, with Fama French 3 Factor Model and Market Returns serving as additional benchmarks to provide greater validity to the results.

Literature Review

For years, economists have conducted research to investigate how mergers and

acquisitions impact shareholders. As such it should be no surprise that across the globe M&A activity has existed for centuries, with many of today’s firms existing through a complex list of

(11)

4

mergers and acquisitions. First will be a review of performance-driven statistical studies. Next, a deep dive into studied motives for M&A activities. Finally, to establish the scientific reasoning for the event windows, a look at existing literature on the Efficient Market Hypothesis will be conducted.

Overall Acquisition Research

As the purpose of this research is to dive into the potential benefit to shareholders when their firm initiates an acquisition, the focus will be on discussing the existing conclusive framework on target firms before highlighting how the evidence is not as conclusive when looking at the acquirer. This has been highlighted in a study conducted by Cartwright and Schoenberg (2006) in which they empirically tried to rather than continue to generate new statistical data. In it, they attempt to decipher the existing reviews and find possible reasonings for the fact that target firms in almost all studies receive positive abnormal returns (Cartwright & Schoenberg, 2006). For example, a study by Michael Jensen (2003) concluded that on average takeover bids amounted to a thirty percent premium on target firms. Additionally, he concluded that the infamous golden parachutes protecting management from takeovers did not provide significant evidence to suggest any harm to the target firm’s shareholders. This aligns with Cartwright and Schoenberg’s (2006) belief that the abnormal returns are generated by the required price premium.

Concerning the acquirer, Cartwright and Schoenberg (2006) found that in most studies, the acquirer received slightly significant negative returns in, or more often, returns that were not significantly different than zero. They noticed that when diving into firm-level data that 35-40% generated positive returns after a two to three-year period. However, this was offset in overall

(12)

5

data analysis by the many that did not achieve positive returns. Thus, resulting in the numerous studies finding results at or close to zero (Cartwright & Schoenberg, 2006).

Isolating International Acquisitions

When looking specifically on an international basis, one of the most prominent studies is that by Datta and Puia (1995) who conducted a study on 112 acquiring firms between 1978 and 1990. They found no significant evidence for value creation. Moderate support for positive value creation was found when Meschi and Metais (2006) observed how French firms with different prior international acquisition experience performed after acquiring a US firm. Though they concluded the hypothesis could only be partially supported. The reason that this paper would like to relook at these past studies is in alignment with the fact that these two studies are much older and international cross-merger M&A activity is on the rise, becoming more prevalent, having increased by 48% in value during 2018 alone (Kengelbach et al., 2019).

International Diversification

Chang and Wang (2007) found that when firms diversify internationally they generated increased firm performance. Another study, by Douglas (2006), found that when Mexican firms diversify internationally, initially there is a negative impact on performance. Eventually, through continued expansion, the performance begins to stabilize and eventually becomes positive (Douglas, 2006).

Finally, a study by Chari, Devaraj and David (2007) to tried to figure out some of the mixed results with international diversification. It specifically looked at the impact of

information technology on firm performance during international expansion. By utilizing IT investment as a proxy for information exchange levels he found significantly positive results on

(13)

6

firm performance when there was a higher rate on IT investment (Chari, Devaraj, & David, 2007). What makes this paper interesting is the fact it used data from 1997, a time in which achieving global financial information required a much higher rate of investment while still yielding much slower and lower bandwidth data exchanges. This is confirmed by a look at some findings on the increase in average internet speed and a decrease in the cost of storage.

Backblaze a cloud storage company-maintained records of its bulk acquisition cost per gigabyte of storage from 2009 to 2017 in which 1GB cost about $0.11 in 2009 but by 2017 they could obtain 1GB of storage for less than $0.04 (Klein, 2017). Business Insider found that in 2009 the average US internet speed was about 5Mbps before increasing to about 100Mbps by 2019 (Villa-Boas, 2019). On top of this, as shown in a collection of publicly released iPhone sales by the Statista research team, iPhone sales only reached tens of millions per quarter in 2011 (Statista Research Department, 2020). Global smartphone sales combined in 2009 and 2010 did not reach the 2011 level of 494.4 million smartphones sold (O’Dea, 2020). This is important as Pitichat (2013) highlights how smartphone usage in the workplace has the benefit of increasing information exchange between employees. Therefore, strong evidence exists to suggest that companies’ access to information has changed and combined with Chari et al.’s (2007) findings raise the potential for firms to increase the ability to benefit from international acquisitions due to increased information sharing.

Macroeconomic Factors

A study on UK activity by Uddin and Boeteng (2011) observed an overlooked area of cross-border Mergers & Acquisitions (CBM&A) activity. Looking at macroeconomic factors they found that “GDP, exchange rate, interest rate and share prices have a significant impact on the level of outward UK CBM&As. On the other hand, GDP, money supply and share price have

(14)

7

a statistically significant impact on the UK CBM&As inflows” (Uddin & Boeteng, 2011, p. 547). This ties into studies which highlight the benefits of international diversification and information exchange.

There is a possibility that information exchange capabilities present a macro factor, and as Uddin and Boeteng (2011) pointed out, macro factors have a strong influence on CBM&A activity. The massive increase in global information exchange could have a positive influence on acquiring firm’s ability to generate positive returns in comparison to older data in which

information exchange was much slower and more costly. In other words, perhaps information technology must reach a certain level to facilitate more consistent positive returns for acquiring firms. The potential is that prior studies didn’t find positive returns, due to information exchange not existing in a significant enough magnitude. Target firms would not be affected as much due to the earlier stated findings that their positive returns can greatly be attributed to acquisition premiums, whereas acquiring firms must generate positive value to a substantial enough degree to overcome those premiums.

A further indication of a need to use recent data, in accordance with macroeconomic factors, stems from findings on the significant alterations in returns of M&A activity for acquiring and target firms during the 2008 Financial Crisis in comparison to non-recessionary times (Linssen, 2017). As per NBER who states that the crisis officially ended June 2009, the paper must look at acquisitions announced no sooner than June 2009 (US Business Cycle Expansions and Contractions, n.d.).

This paper combines the findings on the impact of the Financial crisis with that of the potential impact of information and general macroeconomic factors. In doing so, the paper concludes that acquisitions will need to have occurred starting no sooner than 2010 to avoid the

(15)

8

impact of the financial crisis and best capture potential benefits of increased information exchange. Thus, using the recent data, it is hypothesized that:

H1: Cross-Border Acquisitions by US firms generate positive cumulative abnormal returns for its shareholders.

Value Building & Business Motives

Economies of Scale. During times of economic growth, demand increases. This results in a need to generate ever-increasing output. As a result, M&A activity which targets the benefits of economies of scale are more prevalent during times of economic growth, and less prevalent during times of decreased demand, such as a recession. Through horizontal mergers and acquisitions, firms find themselves better able to meet customer supply demands, achieve competitive pricing through synergies and have the potential to be able to offer better products than their rivals (Lambrecht, 2004). Thus, if studying takeovers in times of economic growth, economies of scale benefits should help drive potential increased shareholder value.

Synergies. While in the 70s and 80s much M&A activity utilized leveraged buyouts and hostile takeovers, recent activity is much more cordial. In modern times synergies have become a very prominent determinant of M&A success - having the ability of the newly combined firm to offer similar products with substantially reduced costs leading to better returns than what could be achieved as two separate firms (Weber & Dholokia, 2000). Weber and Dholakia (2000) additionally highlight the need for identifying synergies as a critical factor in helping to determine potential merger success that translates into improved returns towards shareholders. Per the literature discussed on the premiums provided to target firms and inconclusive results

(16)

9

regarding shareholder returns for the acquirer, premiums are shown to reflect most, if not all of the synergistic value of a takeover (Damodaran, 2005). If an acquisition were to be valued only via synergies, the only way an acquiring firm would obtain shareholder value is if they managed a price worth less than the value of synergies. Inversely, a loss of shareholder value would symbolize an overestimation of synergies or a failure to fully realize them. This highlights, the importance of accurate identification synergies by acquiring firms as stated by Weber and Dholakia (2000).

Local Politics & Greenfield. Firms seeking expansion into a new country are likely to face different legal, economic, and political structures. A firm may choose to launch a greenfield venture. But as the study by Gaur & Kumar (2009) highlights, depending on the systems in place, a firm seeking to launch a greenfield may face significant hurdles that do not exist for current operating firms. Therefore, it may be an advantage for a firm to initiate an acquisition of an existing local firm to avoid potential hurdles (Gaur & Kumar, 2009). Another study compared cross border M&A with exporting and greenfield investments and found that the ability of M&A to be the most or least successful of the options was dependent on their mobile and immobile capabilities (Nocke & Yeaple, 2007).

Managerial Motives

Managerial Entrenchment. Primarily discussed by Jensen (1986), managerial

entrenchment revolves around the concept that managers manage to isolate themselves from the optimal level of control and accountability through various methods. Takeovers are one of these examples. Hughes, Lang, Mester, Moon, and Pagano, (2003) conducted a study on banks comparing the performance of entrenched vs non-entrenched management and looked at how different methods of growth and divestitures impacted overall performance. It was discovered that when the firm did not have entrenched management acquisitions, this resulted in increased

(17)

10

performance. The inverse was true for entrenched management. They discovered that acquired assets generate lower firm performance, but if entrenched management divested assets, increased performance was observed. This is especially true if management has become entrenched

without their substantial ownership to increase the cost of utilizing agency goods (Hughes et al., 2003). The continuance of entrenched managers to seek further acquisitions at the expense of firm value connects with the notion of prestige and executive compensation related to firm size.

Executive Compensation. Firm size was found to have a direct impact on executive compensation. The larger the firm on average, the larger the firm size, and the greater the executive compensation (McKnight, 1996). In direct relation to internationalization Sanders and Carpenter (1998) found that CEO’s of firms going through internationalization received

increased compensation as a result. Thus, top-level management would have motives to engage in takeovers for their total compensation independent of the reasons for shareholder value explored above.

Prestige and Career Growth. Managers have more than traditional financial metrics for maximizing their personal value. Two such metrics are career development and overall prestige. At a lower and middle-level prestige helps to generate career advancement (Monsen & Downs, 1965). The research highlights that even CEO’s have a desire for further prestige. However, it does not state how they would obtain it further, although in accordance with the other desires such as executive pay and entrenchment, perhaps there is some desire to takeover firms utilizing empire-building as a means to increase perceived prestige.

(18)

11 Efficient Market Hypothesis

The efficient market hypothesis had to be investigated, as this event study chose to observe returns on a relatively small scale. First off when EMH holds, it would be expected that the moment of the announcement becoming public that the stock price would immediately reflect any expected benefits for shareholders. This would generate the hypothesis:

H1: All abnormal returns would be observed on the day of the announcement.

If it did not, then that would be a sign of market inefficiencies (Damodaran, 2005). Malkiel (2003), provides an in-depth analysis and dispute to critics of the EMH. He explains how critics claim that predictable behavior and patterns can be found and exploited. Malkiel refutes this by showing how any patterns that do form, soon get exploited until the pattern no longer exists or can provide abnormal return per unit of risk. When Malkiel (2003) looked into mutual funds, he found the rate of losers to be much higher than winners when being compared to the S&P500, highlighting the challenge in beating the market through active trading. He concluded that while technology will help to uncover more patterns, it also will provide greater power to exploit, eliminating the pattern as a result. Finally, in regards to the beliefs that trends and market bubbles exist, he ascertains that in almost all cases these are only truly identified in an exploitable way after the fact and they don’t offer any ability to take advantage of future occurrences. Despite this, some market inefficiencies have been identified. The two being related to this study are insider trading and lag time in investor reaction.

Insider Trading. Beny and Seyhun (2012) studied the levels of insider trading occurring in the US between 1996 and 2011. They concluded the rate of a stock price run-up before an

(19)

12

announcement has increased. This implies a need to include returns before the announcement date to fully capture the impact of the acquisition on the acquiring firm value. This would generate the hypothesis:

H1: There are significant cumulative abnormal returns in the lead up to the announcement of a takeover

Post-Announcement Drift. Regarding the need to include time post-announcement Bernard and Thomas (1989) conducted a study on post-earnings announcement drift. They found there to be a significant drift both leading up to and following an earnings announcement. They concluded that a large portion of the drift occurs within five days following the earnings

announcement and depending on firm size, some drift can be noticed up to nine months after. They argue that by the sixty-day mark almost all valuable post earnings drift had concluded (Bernard & Thomas, 1989). The study highlights the need to take into account that while the market does react under the EMH, some inefficiencies require expanding the event window to include a longer time frame. This resulted in the following hypothesis:

H1: There are significant cumulative abnormal returns following the announcement of a takeover.

(20)

13 Methods

Conceptual Model

The purpose of the study is to investigate the impact that the announcement of an international acquisition has on shareholder value. Thus, an event study methodology has been used. The study operates following previously cited literature above, in which it is believed that the Efficient Market Hypothesis holds with some observable inefficiencies; insider trading leading up to the announcement and post-announcement drift.

The dependent variable is thus the mean cumulative average abnormal returns (CAR) of the sample and will be tested against the predicted returns in the event the announcement did not occur.

Previously cited research was determined to be both potentially outdated and inclusive. Combined with motive-based literature existing for both the creation and destruction of

shareholder value this study takes a two-sided approach.

Sample Collection

Table 1 provides a general overview of the pathway to the final sample size of 235. Below it further explanation is provided detailing the reasoning.

(21)

14 Table 1

Outline of Creation of Final Sample Set

Part of Process Reason for Reduction Sample Size Zephyr Extraction Max Initial Stake 0%

Final Stake 100%

575,083 Zephyr Extraction Acquirer Stock Exchange:

AMEX, NYSE or NASDAQ

41,070 Zephyr Extraction Deal Status:

Completion Confirmed

32,194

Zephyr Extraction Non-US Target Firm 9,027

Zephyr Extraction Time Period Of Completion 2011-Present

536 Zephyr Extraction Sub-Deal Type:

Acquisition

496 Ensuring OLS Prediction

Excludes effects of Recession

Removal of announcements before June 1st 2011

305 CRSP Zephyr Data Merge Failure to return CRSP

Results

296 Accounting For Multiple

Announcements on Same Day

Accounting for double announcements as one

290

Ensuring Unbiased Data Removal of Announcements on Non-Trading Days

235

The sample list of acquisitions was obtained utilizing the Zephyr database. Six were applied to ensure that the list contained only the acquisitions of interest. First, acquisitions were filtered to ensure that the acquiring firm held no initial stake in the firm and conducted a total acquisition of the target. Next, a reduction was made to include only acquiring firms that were listed on the three major US stock exchanges: AMEX, NYSE & Nasdaq. This was to ensure that acquiring firms were public and in accordance with the available stocks on the CRSP database. To isolate the effect of deals that occurred, a filter was placed requiring the deal to have been confirmed complete. Following literature regarding the impact of the Great Recession, the deals were required to have been completed no earlier than 2011. To complete the extraction from

(22)

15

Zephyr, a final filter was placed to ensure the deal was listed as an Acquisition. The reasoning for why this further reduced the sample size proved ambiguous. Upon further inspection of deals that were removed, it appeared that while the first filter would imply all results to exclusively consist of acquisitions, some mergers and joint ventures managed to stay in the data set depending on the method of executing such deals.

After retrieval, an inspection of the list was conducted. Due to some deals being completed in some cases years after the initial announcement, 90 acquisitions were removed. The result was announcements with the earliest date being March 1st_{, 2011. This was to further}

ensure that any residual effects of the Great Recession were minimized. The analysis was initially conducted with a 120-trading day prediction window. This was then adjusted to a

standard 252-trading day per year window to ensure greater predictive power and better highlight differences between the benchmarks. This expansion required an additional removal of 101 acquisitions. Thus, the earliest day, in the final analysis, required announcements to have occurred no earlier than June 1st 2011.

The next step required the retrieval of stock data from the CRSP database. 9 stocks were unable to retrieve some or all the required information. This resulted in a sample size of 305. Further inspection of the data revealed that six announcements days consisted of 2 acquisition announcements by the same firm. These duplicates were merged and thus treated as a single announcement bringing the sample size down to 290.

Finally, upon running the analysis in STATA, a further 55 announcements were

identified to have occurred on non-trading days. Due to the desire to test share price response on the day of and immediately following, it was decided to exclude these from the study. This generated the final sample size of 235 acquisitions.

(23)

16 Determining of Daily Aggregated Abnormal Return

The following steps were used for all three benchmarks, CRSP, Fama French 3-Factor Model (FF3) and Fama French Market (FF1) in the eventual test for cumulative aggregate abnormal returns. For each equation X, the addition of .1 has been added for FF3, .2 for FF1 and .3 for CRSP benchmarks.

The Fama French 3 Factor Model of daily aggregated abnormal returns were used in combination with previously studied literature to establish the event windows to be studied. The first step required the creation of an ordinary least squares (OLS) regression for each stock to be able to generate daily predicted returns if the announcement were not to occur. A standard 252-day trading year was used. This resulted in the estimation window T-31 to T-283. Equation (1.1), (1.2) and (1.3) shows the natural log regression to be able to generate the expected return on day t denoted 𝐸[𝑅𝑖𝑡] for FF3, FF1 and CRSP accordingly.

𝐸[𝑅𝑖𝑡] = 𝛼𝑖 + 𝛽𝑖(𝑚𝑘𝑡)𝑅𝑡(𝑚𝑘𝑡)+ 𝛽𝑖(ℎ𝑚𝑙)𝑅𝑡(ℎ𝑚𝑙) +𝛽𝑖(𝑠𝑚𝑏)𝑅𝑡(𝑠𝑚𝑏)

𝐸[𝑅𝑖𝑡] = 𝛼𝑖 + 𝛽𝑖(𝐹𝐹1)𝑅𝑡(𝐹𝐹1)

𝐸[𝑅_𝑖𝑡] = 𝛼_𝑖 + 𝛽_{𝑖(𝐶𝑅𝑆𝑃)}𝑅_{𝑡(𝐶𝑅𝑆𝑃)}

In Equation (2) for each benchmark b Daily abnormal returns 𝐴𝑅_𝑖𝑡 for each stock were generated by subtracting the predicted return 𝐸[𝑅_𝑖𝑡𝑏] from 𝑅_𝑖𝑡

𝐴𝑅_𝑖𝑡𝑏 = 𝑅_𝑖𝑡 − 𝐸[𝑅_𝑖𝑡𝑏]

Next in (3), the results were averaged to form the daily aggregated abnormal returns denoted 𝐴𝐴𝑅_𝑡𝑏 𝐴𝐴𝑅_𝑡𝑏 = 1 𝑁∑ 𝐴𝑅𝑖𝑡𝑏 𝑁 𝑖=1 (1.1) (1.2) (1.3) (2) (3)

(24)

17 Establishing the Event Windows to Test

Before proceeding to generate cumulative aggregated abnormal returns, the event windows needed to be established. To do so Table 2 was created and is shown in the results. It consisted of daily t-test for significant 𝐴𝐴𝑅_𝑡 against the FF3 𝐸[𝑅_𝑖𝑡]. Alongside this was the daily

AAR.

For additional insight, the cumulative aggregated abnormal returns (CAAR) against FF3 𝐸[𝑅𝑖𝑡] was generated from date T-30 to T+90. Equation (4) shows the calculation of CAAR for

each progressing day 𝑡2 holding 𝑡1 constant at -30.

𝐶𝐴𝐴𝑅_(𝑡₁_,𝑡₂₎ = 1

𝑁∑ 𝐶𝐴𝑅𝑖(𝑡1,𝑡2) 𝑁

𝑖=1

The resulting daily CAAR were plotted against the daily AAR to provide a visualization of the impact in Figure 1 which can be found in the results.

The event windows to test were then created by looking at Table 2 and Figure 1 in combination with the literature; the reasonings being discussed in the results. They can be broken down into three segmented types of sub-investigations: Lead-Up, Overall and Post-Announcement. Below the established event windows and the corresponding alternative hypotheses are elaborated. Finally, Figure 5 provides a timeline to visualize the overall event windows as well as the OLS prediction window in relation to the announcement date.

Lead-Up Windows. Three event windows were utilized to attempt to capture potential insider trading leading up to the announcement in accordance with Beny and Seyhun’s findings regarding the tendency to observe abnormal stock movement leading up to an announcement day. Three event windows were created to test against the null hypothesis:

(25)

18

HO: There are no significant cumulative abnormal returns in the lead up to the announcement of a takeover.

T-5 to T-1 H1: There are significant cumulative abnormal returns in the lead up to the announcement of a takeover.

Figure 2 provides a visualization of the lead-up event windows in relation to the announcement date.

Figure 2. Lead-Up Windows

Overall Windows. Overall Acquisition tests are used to highlight the main goal of the study in determining whether or not there is an overall effect on shareholder value with the null hypothesis:

(26)

19

H0: There is no significant overall effect on shareholder value upon the announcement of a takeover.

A test on the day of and seven expanded event windows were created. They can be found below with their corresponding alternative hypotheses. This is followed by Figure 3 which shows the windows visually in relation to the announcement day.

T=0 H1: On the day of the takeover announcement there are significant abnormal returns. T-1 to T+1 H1: There are significant overall cumulative abnormal returns upon the announcement of a takeover.

T-3 to T+3 H1: There are significant overall cumulative abnormal returns upon the announcement of a takeover.

(27)

20

Figure 3. Overall Event Windows

Post-Announcement Drift Windows. Per Bernard and Thomas’s (1989) findings on post earnings drift, five event windows were tested to potentially capture such drift. Below are the windows and corresponding alternative hypotheses followed by Figure 4 which shows them mapped out.

H0: There are no significant cumulative abnormal returns following the announcement of a takeover.

T=0 to T+4 H1: There are significant cumulative abnormal returns following the announcement of a takeover.

T=1 to T+30 H1: There are significant cumulative abnormal returns following the announcement of a takeover.

(28)

21

Figure 4. Post-Announcement Windows

Figure 5. Overview of Event Windows

Testing the Hypotheses

For each event window, a T-test was conducted to determine if significant CAAR was observed. This was done for each event window against all three benchmarks. This resulted in the creation

(29)

22

of Tables 3, 4, 5, 6, 7 and 8. In them the potential significance and magnitude of deviation for each event window against the null hypothesis.

Additional Analysis

Some additional analysis for discussion was also created. First, a histogram of the cumulative abnormal returns was created and confirmed that the data was normally distributed. Next, a box plot of the CAR was created over for the T-30 to T+90 event window to look for the distribution of the data and be able to see some insight about how the mean was being created; from a few poor acquisitions or a general negative trend.

Results

After first detailing the selection of the event windows, the results section was split into two parts, both following the same path but with two levels of detail. Part 1 one looks at the more traditional testing of Cumulative Abnormal returns by simply looking at the results

compared to the CRSP benchmark index. Part 2 will dive deeper into analysis by looking at how the results look when more benchmarks are used; Fama French 3 factor model and the daily Fama French market excess returns to provide greater validity to the results of part one.

For both parts: First, the build-up period was reviewed to see how potential insiders are reacting to the upcoming announcement. Next, an evaluation of the overall effect of the

announcement was observed by starting with a test on the day of the announcement before expanding the event window finishing at an overall view including the month preceding and the month following the announcement. Finally, an evaluation in accordance with literature will look at how the public reacts over time following the announcement to measure the public response and look for post-announcement drift.

(30)

23 Daily Aggregated Abnormal Returns

With the calculation of Table 2, shown below, the data highlighted that on average most days did not overall generate abnormal returns.

Table 2

Daily Aggregated Abnormal Returns Against Fama French 3-Factor

Plot of Abnormal Returns with Cumulative Abnormal Returns

Despite this, a negative trend was observed in the lead up to the announcement in which almost all days while not significant did present negative coefficients. This raised the notion that while not significant on any particular day (except T-4), there was reason to conduct a test on the significance of the cumulative abnormal returns leading up to the announcement. And thus led to

t Mean p Time p p t Mean p t Mean p

-30 -0.14% 0.2075 0 0.28% 0.1770 30 0.02% 0.8781 60 -0.20% 0.3264 -29 -0.05% 0.7529 1 0.25% 0.2132 31 0.11% 0.4677 61 -0.02% 0.9079 -28 -0.28% 0.1959 2 -0.13% 0.3205 32 -0.08% 0.5295 62 -0.18% 0.1750 -27 0.03% 0.8388 3 0.00% 0.9926 33 0.12% 0.4585 63 0.08% 0.5332 -26 0.14% 0.3812 4 0.10% 0.5504 34 -0.25% 0.0667 64 -0.11% 0.4681 -25 -0.14% 0.2607 5 -0.36% 0.0159 35 -0.12% 0.5203 65 0.05% 0.6574 -24 -0.06% 0.7560 6 -0.07% 0.5342 36 -0.09% 0.6009 66 0.16% 0.2924 -23 -0.13% 0.2719 7 -0.04% 0.7732 37 -0.11% 0.3948 67 0.26% 0.0666 -22 -0.16% 0.1821 8 -0.12% 0.3910 38 0.12% 0.4251 68 0.06% 0.6235 -21 -0.01% 0.9446 9 -0.15% 0.3011 39 -0.08% 0.5946 69 -0.07% 0.6191 -20 -0.14% 0.1759 10 0.21% 0.1176 40 -0.24% 0.1000 70 -0.01% 0.9339 -19 0.16% 0.2113 11 0.05% 0.7454 41 0.16% 0.4116 71 0.05% 0.6488 -18 -0.01% 0.9059 12 -0.11% 0.4091 42 -0.08% 0.4945 72 -0.24% 0.1589 -17 -0.09% 0.4950 13 -0.14% 0.3240 43 -0.04% 0.7215 73 -0.24% 0.0464 -16 -0.05% 0.6980 14 -0.04% 0.7546 44 0.00% 0.9722 74 -0.04% 0.7382 -15 0.01% 0.9532 15 -0.13% 0.3086 45 0.07% 0.5643 75 0.17% 0.2176 -14 0.02% 0.8499 16 0.08% 0.5471 46 0.04% 0.7742 76 0.03% 0.7640 -13 -0.09% 0.3733 17 0.02% 0.8924 47 -0.09% 0.4895 77 0.07% 0.6176 -12 -0.05% 0.7039 18 0.29% 0.0802 48 -0.01% 0.9281 78 0.08% 0.5500 -11 -0.01% 0.9719 19 0.18% 0.1810 49 -0.14% 0.2989 79 0.13% 0.4278 -10 -0.23% 0.0350 20 0.14% 0.2657 50 -0.10% 0.4778 80 -0.16% 0.3142 -9 0.02% 0.8427 21 0.00% 0.9911 51 0.14% 0.2925 81 -0.07% 0.6052 -8 -0.09% 0.3881 22 -0.27% 0.1185 52 -0.27% 0.0672 82 0.06% 0.6084 -7 0.04% 0.7799 23 -0.15% 0.3222 53 0.11% 0.4239 83 0.19% 0.1906 -6 -0.20% 0.0974 24 -0.17% 0.4008 54 -0.16% 0.1676 84 0.05% 0.6945 -5 0.15% 0.2333 25 -0.16% 0.3134 55 0.07% 0.6057 85 0.03% 0.8372 -4 -0.49% 0.0001 26 0.04% 0.7930 56 0.04% 0.8260 86 -0.03% 0.7805 -3 0.03% 0.8356 27 0.21% 0.3005 57 0.00% 0.9785 87 0.02% 0.8727 -2 -0.07% 0.5541 28 -0.05% 0.7045 58 -0.10% 0.4501 88 0.05% 0.7294 -1 -0.17% 0.1144 29 -0.14% 0.2731 59 -0.01% 0.8959 89 0.22% 0.1223 90 0.07% 0.5105

(31)

24

the establishment of the three lead up windows discussed in the methods. Figure 1 below shows this initial spike on announcement day and the days immediately following.

Figure 1. AAR vs. CAAR

The significance of announcement day abnormal returns, the following day as well as the general positive returns in the first five days of public awareness generated the belief that an event window to test the first five days was needed.

Despite consistently insignificant p-values and small coefficients thereafter, in accordance with the literature on post-announcement drift the additional two event windows; T0 to T+30, T0 to T+60 and T0 to T+90 were determined to be a useful inclusion none the less. These ensure that while seeming to not be existent from the daily abnormal returns, that any post-announcement drift was not overlooked.

To see the overall announcement impact, the event windows T-30 to T+30, T-15 to T-15, T-3 to T+3 & T-1 to T+1 were created. Finally, in the belief of EMH holding, the abnormal

(32)

25

return of the announcement day (T=0) was included, also representing the overall impact on shareholder value.

Results Against CRSP

Lead-Up to Announcement. As shown in Table 3, the findings suggest that there is indeed significant insider activity leading up to a takeover announcement and that activity is harmful to overall shareholder value. In all three event windows, at the 5% level of significance, negative cumulative abnormal returns are found.

Table 3

Lead-up CAAR against CRSP

Overall. As Table 4 shows, the overall impact it becomes clear that there is a lot of noise in the few days before and after the announcement. As the event window is expanded though it becomes very clear that value destruction occurs, peaking at over 4% when looking at the window -30 to 60. Mean -5 -1 -0.55% 0.0426** -14 -1 -1.08% 0.0278** -30 -1 -1.95% 0.0073*** CRSP Event Window p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1%

(33)

26 Table 4

Overall CAAR Against CRSP

After Announcement. Upon announcement day, shareholders a low level of significance respond positively on average to the announcement. As seen in Table 5 This positive reaction quickly diminishes though even resulting moderate significance when looking two months after. Shortly after though the negative response diminishes and no further drift is observed. This is in line with Bernard and Thomas’s (1989) findings that almost all post-announcement drift is observed in the first sixty days following an announcement.

Mean 0.34% 0.0999* -1 1 0.43% 0.1490 -3 3 0.22% 0.5882 -5 5 -0.44% 0.3704 -15 15 -1.61% 0.0342** -30 30 -2.60% 0.0226** -30 60 -4.07% 0.0051*** -30 90 -3.73% 0.0260** CRSP Event Window p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1% 0

(34)

27 Table 5

Post-Announcement Drift Against CRSP

Takeaway. Despite an initial positive spike from the public on announcement day, it becomes apparent that over time the public begins to share the views of the insiders in the lead up eventually to driving share value down. This is what leads to the significant overall

conclusion that international acquisitions by US firms lead to value destruction.

Results with FF3 and FF Market Returns

This section will further breakdown the results while also looking at difference across the three benchmarks.

Lead-Up To Announcement. When looking at the lead up to an announcement, as shown in Table 6 all three benchmarks provide nearly uniform results in both significance and magnitude of negative returns. The null hypothesis is rejected and the results across all three benchmarks at a 5% level of significance conclude that shareholders on average will obtain negative returns in the lead up to the announcement. The negative returns are apparent thirty

Mean 0 4 0.51% 0.2177 0 30 -0.66% 0.4108 0 60 -2.12% 0.0720* 0 90 -1.78% 0.1939 CRSP Event Window p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1%

(35)

28

trading days in advance peaking at around 2% negative returns and continue to exist at a diminishing rate up until the final few days before the announcement.

Table 6

Lead-Up CAAR against all Three Benchmarks

Regardless of the benchmark, the null hypothesis is rejected at the five percent level of significance. It is concluded that on average shareholders lose about 0.55% value in the final five trading days leading up to the announcement.

As the timeline is expanded some deviation begins to show across the benchmarks when looking closely at the p-values. Again though, at the five percent level of significance, all benchmarks will reject the null hypothesis. In the 14 trading days leading up to the

announcement, shareholders are expected on average to lose a little over 1% in share value.

Mean Mean Mean

-5 -1 -0.55% 0.0407** -0.56% 0.0379** -0.55% 0.0426** -14 -1 -1.16% 0.0156** -1.16% 0.0182** -1.08% 0.0278** -30 -1 -2.08% 0.0037*** -2.24% 0.0021*** -1.95% 0.0073*** CRSP Event Window p p p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1%

(36)

29

Despite expanding the horizon to thirty trading days, all three benchmarks continue to showcase the same results, especially regarding significance. All three will reject the null hypothesis at the one percent level of significance. While there is a .3% spread between the benchmarks, shareholders can expect around a 2% loss in shareholder value in the thirty trading days leading up to the announcement of an international takeover.

Overall

When looking at the overall returns, there is much noise and insignificance found in the days close to the announcement. As the window is expanded, as Table 7 shows, the results begin to reflect that of lead-up findings.

Table 7

Overall CAAR against all Three Benchmarks

0.28% 0.1770 0.33% 0.1070 0.34% 0.0999* -1 1 0.35% 0.2284 0.42% 0.1575 0.43% 0.1490 -3 3 0.17% 0.6706 0.21% 0.6149 0.22% 0.5882 -5 5 -0.45% 0.3550 -0.48% 0.3269 -0.44% 0.3704 -15 15 -1.58% 0.0303** -1.77% 0.0189** -1.61% 0.0342** -30 30 -2.48% 0.0275** -3.07% 0.0072*** -2.60% 0.0226** -30 60 -3.85% 0.0068*** -4.99% 0.0007**** -4.07% 0.0051*** -30 90 -3.22% 0.0477** -4.92% 0.0036*** -3.73% 0.0260**

Fama French 3-Factor Fama French Market CRSP

Event Window p p p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1% 0

(37)

30

T=0 H1: On the day of the takeover announcement there are significant abnormal returns.

When just using CRSP, to test for significance on announcement day, there was very mild support to reject the null hypothesis if willing to use a ten percent level of significance. Expanding the study to include the other two benchmarks, the results indicate a failure to reject the null hypothesis. It would not appear that there are any abnormal returns upon the day of an announced takeover.

When looking at the day before, the day of and the day after, the results indicate that there is no support for the alternative hypothesis. Therefore, regardless of the benchmark, the results failed to reject the null hypothesis. All three benchmarks provide similar magnitude and P-Values at this point.

As the window is expanded slightly, the P-Values rise as the returns decrease. Again all three benchmarks are similar in their findings and failed to reject the null hypothesis.

While the results still fail to reject the null hypothesis, it should be noted that the returns have decreased to the point of negative returns with lower p-values.

(38)

31

Now the results begin to reflect that of the lead-up findings. All three benchmarks reject the null hypothesis at the five percent level of significance. Interestingly the suggested negative returns are greater than that of the lead up for the similar event window. Around a 1.5% loss is observed over this horizon. Some spread between the Fama French Market against the other two benchmarks begins to become apparent.

When looking at the results of thirty trading days lead-up to and following, the results are very similar to the lead-up. All three benchmarks reject the null hypothesis at the five percent level of significance. They indicate an average shareholder loss of around 2.5 – 3%. It should be noted that the Fama French 3 Factor Model and CRSP benchmark maintain similar results. The Fama French Market though while in the same direction suggests more extreme findings.

As the window is expanded further, the negative returns become more extreme than that of just the lead-up. All three benchmarks reject the null hypothesis at the one percent level. Around -4% returns are found during this horizon. Again, the Fama French Market benchmark

(39)

32

provides much more extreme results. Now we also do observe some separation in results between the Fama French 3 Factor and CRSP benchmarks.

All the P-Values are significant at the five percent level, thus rejecting the null

hypothesis. In accordance with Bernard and Thomas, expanding the window beyond sixty days post doesn’t provide many changes in the results. The increase in P-values is in direct accordance with their findings as the earlier significant cumulative abnormal returns are muted by less significant returns in the additional thirty trading days.

Following Announcement. As shown in Table 8, the results following the

announcement are mixed. Following the findings against just the CRSP, very little significance is found to suggest a post-announcement drift.

Table 8.

Post-Announcement Drift against all Three Benchmarks

0 4 0.47% 0.2530 0.48% 0.2401 0.51% 0.2177 0 30 -0.40% 0.6097 -0.83% 0.2964 -0.66% 0.4108 0 60 -1.76% 0.1261 -2.75% 0.0215** -2.12% 0.0720* 0 90 -1.14% 0.3930 -2.67% 0.0538* -1.78% 0.1939 CRSP Event Window p p p * Significant at 10% ** Significant at 5% *** Significant at 1% **** Significant at 0.1%

(40)

33

T=0 to T+4 H1: There are significant post take over announcement drift.

This window was created because of insignificant but consistently positive returns when looking at Table 2 of daily abnormal returns. Despite that, no significant cumulative returns were found and thus the null hypothesis is rejected by all three benchmarks.

T0 to T+30 H1: There are significant post take over announcement drift.

This window was included to be in accordance with Bernard and Thomas’s (1989) conclusion that they observed post earnings drift. Again regardless of benchmark no support for the alternative hypothesis was found. The results fail to reject the null hypothesis as it would appear there is no post-announcement drift the thirty trading days following the announcement of an international takeover.

Bernard and Thomas (1989) concluded that if there is post-earnings drift almost all of it would be reflected after sixty days. Here the results are mixed but there is evidence to reject the null hypothesis partially. The Fama French 3 Factor model found no significance. The Fama French Market Returns found significance at the five percent level. Finally, the CRSP benchmark found significance at the ten percent level.

More so than focusing on the null hypothesis, these mixed findings highlight the potential significant difference in results that can occur when utilizing different benchmarks. Therefore, two conclusions are to be generated:

1. The null hypothesis is partially rejected. It would appear that in the long run there is some negative post-announcement drift.

(41)

34

2. Future studies using benchmarks should consider the usage of more than one to provide greater ability to generate conclusions when the results are only slightly significant.

Again, this window was created in accordance with Bernard and Thomas (1989). It was expected that there would be less significant results. The results due indicate that against all benchmarks the P-Value has risen. Other than for Fama French Market, returns no significance was found.

Again the results highlight the potential importance of using multiple benchmarks shown by the great variation in P-Values.

Additional Analysis

This section serves to provide some additional insight into the general findings and help to understand some potential reasoning behind the results found above, again as before as a part of the results, the interpretation will be kept strictly to data analysis leaving theoretical

interpretation for the discussion portion of this paper. Appendix 1 depicts a histogram

highlighting that despite the data being centered around a negative overall cumulative average, abnormal return over the window -30 to +30, returns appear to be randomly distributed.

Appendix 2 is a box plot which confirms what is seen in the histogram, as it highlights the fact that over half of the firms received negative cumulative abnormal returns over the horizon -30 to +30. It should be noted that over 25% firms were able to generate positive cumulative abnormal returns.

(42)

35 Discussion

In this section, the focus will be split into two segments to interpret the results following the existing literature. First, the negative overall shareholder value alongside the negative lead up returns will be discussed. Next, a look at the failure to observe post announcement drift will be looked at concluding with a discussion for why a substantial amount of the studied takeovers still resulted in positive returns and how futures studies can potentially identify the attributes that increase the potential for successful takeovers.

Takeovers as an expensive proxy for outside board directors

When evaluating managers and board desires for control, existing literature highlights the usage of takeovers as a way for CEO’s to gain further board control while on the inverse, be leveraged as a way for outsiders to regain control of the Board back from the CEO. Takeovers have the ability to proxy for the lack of outside director control in the board (Kini, Kracaw, & Mian, 1995). Unfortunately, Kini et al. (1995) do not study the returns generated overall of takeovers where the takeover was determined to take place with a board of insiders. When looking at the significant negative returns in the lead up to the announcement, there is a

possibility that this could be a price insiders place on the usage of a takeover to govern CEOs as opposed to if control was managed through a balanced board. This is one potential reason why the results show that some acquisitions generate substantial positive CAR, while the median is negative. A further study is needed to include what is analyzed in this paper with that of Kini et al.’s board composition and control to explore the validity of this idea. An extension of this paper could look at board composition to investigate how much of the successful and failed takeovers fit the Kini et al.’s concept of a takeover that acts as a proxy for outside directors. If many of the

(43)

36

unsuccessful takeovers were determined to be initiated for the purpose of control instead of direct shareholder value it could help lead to future segmentation into takeovers for control and takeovers for direct value creation.

Empire Building

In contrast, to the previous discussion on takeovers as a method of controlling managers. The negative returns are in accordance with the literature review’s findings that studies indicate managers initiate empire building and do receive greater compensation for managing a larger firm. Therefore, while some of the negative takeovers may reflect a control expense, others may reflect an acceptance of lost firm value if it can still lead to greater overall compensation for management.

As this study did not investigate executive pay changes following a takeover, a future study could investigate changes in executive pay following acquisitions, preferably in comparing international and domestic takeovers. This would allow researchers to first confirm existing studies on showing increased firm size leads to greater compensation, but also begin to look into potential differences in pay between international and domestic intensive firms.

Synergies, Economies of Scale and Control

Failure to properly determine the premium. This paper chooses to exclude

recessionary times to be able to exclude the likely negative ability for firms to merge and acquire for the benefit of economies of scale. Based on the overall negative results it does not seem that economies of scale and other synergies are achieved enough to offset the literature showing that firms often pay a significant premium to purchase a firm. This study is limited in this section as

(44)

37

it is unable to distinguish firms expected synergies that they expected through either economy of scale or other forms such as reduced administrative, marketing or logistical costs.

Successful Value Chain & Market Control based takeovers. Control can also be looked at as ways firm seek to gain better management of the value chain or market. By taking over a competing firm whether horizontally or vertically in the value chain, management is better able to utilize resources to develop competitive advantages and generate abnormal returns. The primary theory on this originates back to a theory by Henry G. Manne (1965) in which he investigates the notion of anti-trust laws. He explains how the very fact of their existence showcases the belief that if firms can gain too much control, whether vertically or horizontally, they can leverage it to extract excess returns, albeit at the expense of those outside the firm (Manne, 1965). We can see the desire to gain control during the deregulation of the US airline industry in 1978. There are benefits to consolidated control, as Alfred Kahn explains in his study regarding the US Deregulation of 1978, where he found that with the increased control and consolidation of airlines, consumers were able to gain access to reduced fair prices, increased route options and better service. Kahn (n.d.) also highlighted the concern that over time it has led to some monopolistic practices that in some areas of the industry re-established the harms of monopolies. Since deregulation, the United States has seen a plethora of mergers and

acquisitions resulting in a fraction of the number of major carriers.

Potential success can be seen in the box plot of overall cumulative abnormal returns (Appendix 2), in which over 25% of acquiring firms generated shareholder value. In combination with Kahn et al.’s (n.d.) study that showed the ability of mergers and acquisitions in the airline industry to result in larger firms that managed to grow and improve their offerings to consumers,

(45)

38

suggests the potential for firms to be successful in evaluating synergies and leveraging increased control.

Takeaway. The fact that over 25% of takeovers in the study generated positive CAR suggests that there are firms which can successfully achieve synergies. A future study could look into firms that have made their expected value of synergies public, and compare it with the takeover premiums to better understand if firms are indeed creating premiums based upon synergy value. Positive findings would then lead to the need to investigate reasoning for why so many mergers fail to properly identify or achieve the synergies. Following aforementioned prior studies, future researchers should be looking into factors that could lead to the misalignment in premiums paid and synergies; manager overconfidence, empire-building & executive

compensation.

Post-Announcement Returns

Despite an initial positive reaction from the public in the first few days following the announcement, in the long run, the post-announcement returns are insignificant to moderately negative. While a major significant long-term post-announcement drift in accordance with Bernard and Thomas (1989) was not observed, there are still some interesting findings that could be explored in further research. The results do not confirm nor deny the results of Bernard and Thomas, because of the inability to identify whether or not the lack of drift is merely because after the first day there is none, or if there is, in fact, a drift but in the form of a reversal. The fact that the public, excluding the initial few days, didn’t find any significant benefit overall to

international acquisitions, accepting the loss from the insiders before the announcement, suggests that future studies could investigate the opinions of shareholders after the announcement. Such a

(46)

39

study could be survey-based and orient itself around collecting shareholder’s opinion on the takeover on the day of, immediately after, and later in recurring segments to gain insight if shareholders, either do not notice the significant decline in value or in fact share insiders’ view that the takeover is harmful to shareholder value. In other words, there may be a drift, but unlike in a study of post-earnings drift, due to the complexities of a takeover, the drift exists in a non-linear form.

Conclusion

In response to the research question of this study, overall, an international takeover by US firms leads to negative shareholder value. The research is limited in its explanatory power due to its intention to serve as an exploratory study to help identify what future research should look at in regard to the topic of Mergers & Acquisition activity.

The key takeaway is that existing literature has confirmed the benefits of takeovers for target firms. With so much literature existing on the percentage of M&A related failures, and inconclusive findings regarding acquiring firm value, this study will need to be reinvestigated using either more metrics or a different sample set (European firms, small firms, etc.) to confirm the external validity of the results to a more global level.

The author believes that, while much activity is harmful to shareholders, the years of constant M&A activity with countless success stories, indicates that the issue is not M&A as a whole, but the inability to distinguish the factors of success beforehand. Such belief was confirmed in this study.

Thus, researchers should take away the fact that with so many theories on what can cause some acquisitions to succeed why so many others fail, the focus on future research should be

(47)

40

looking at ways to isolate factors of success due to a large amount of noise when trying to interpret takeovers at a macro level.

Of course, as the agency problems discussed in this paper point out, unsuccessful M&A is not likely to subside in the near future. But with research that helps identify the factors of success, managers and owners truly interested in achieving excess returns can hopefully better increase their success rate.

(48)

41 References

Beny, L. N., & Seyhun, H. N. (2012). Has illegal insider trading become more rampant in the United States? Empirical evidence from takeovers. Insider Trading Research Handbook, 50, 1-47. doi:10.4337/9780857931856.00021

Bernard, V., & Thomas, J. (1989). Post-Earnings-Announcement Drift: Delayed Price Response or Risk Premium? Journal of Accounting Research, 27, 1-36. doi:10.2307/2491062

Cartwright, S., & Schoenberg, R. (2006). Thirty Years of Mergers and Acquisitions Research: Recent Advances and Future Opportunities. British Journal of Management, 17, 1-5. doi:10.1111/j.1467-8551.2006.00475.x

Chang, S., & Wang, C. (2007). The effect of product diversification strategies on the relationship between international diversification and firm performance. Journal of World Business,

42(1), 61-79. doi:10.1016/j.jwb.2006.11.002

Chari, M., Devaraj, S., & David, P. (2007). International diversification and firm performance: Role of information technology investments. Journal of World Business, 42(2), 184-197.

doi:10.1016/j.jwb.2007.02.004

Damodaran, A. (2005). The Value of Synergy. SSRN Electronic Journal. doi:10.2139/ssrn.841486 Douglas, T. E. (2006). International diversification and firm performance in Mexican firms: A

curvilinear relationship? Journal of Business Research, 59(4), 501-507. doi:10.1016/j.jbusres.2005.08.008

Gaur, A. S., & Kumar, V. (2009). International Diversification, Business Group Affiliation and Firm Performance: Empirical Evidence from India. British Journal of Management, 20, 172-186. doi:10.1111/j.1467-8551.2007.00558.x

(49)

42

Hughes, J. P., Lang, W. W., Mester, L. J., Moon, C., & Pagano, M. S. (2003). Do bankers sacrifice value to build empires? Managerial incentives, industry consolidation, and financial

performance. Journal of Banking & Finance, 27, 417-447. doi:10.1016/s0378-4266(02)00385-0

Jensen, M. (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. The American Economic Review, 76(2), 323-329. Retrieved June 29, 2020, from

www.jstor.org/stable/1818789" www.jstor.org/stable/1818789

Jensen, M. C. (2003). The Free Cash Flow Theory of Takeovers: A Financial Perspective on Mergers and Acquisitions and the Economy. SSRN Electronic Journal, 102-143. doi:10.2139/ssrn.350422

Kahn, A. E. (n.d.). Airline Deregulation. Retrieved June 20, 2020, from https://www.econlib.org/library/Enc1/AirlineDeregulation.html

Kengelbach, J., Keienburg, G., Gell, J., Nielsen, J., Bader, M., Degen, D., & Sievers, S. (2019). 2019 Mergers & Acquisitions (M&A) Report: Downturns Are a Better Time for Deal Hunting. Retrieved from https://www.bcg.com/publications/2019/mergers-and-acquisitions-report-shows-downturns-are-a-better-time-for-deal-hunting.aspx

Kini, O., Kracaw, W., & Mian, S. (1995). Corporate takeovers, firm performance, and board composition. Journal of Corporate Finance, 1(3-4), 383-412. doi:10.1016/0929-1199(94)00011-i

Klein, A. (2017). Hard Drive Cost Per Gigabyte. Retrieved June 20, 2020, from https://www.backblaze.com/blog/hard-drive-cost-per-gigabyte/

Lambrecht, B. M. (2004). The timing and terms of mergers motivated by economies of scale.