The influence of M&A on acquiring firm performance

1

The influence of M&A on acquiring firm performance

‘Focus on the U.S. High-Tech industry’

Abstract

This paper investigates the M&A performance of acquiring U.S. firms in the high-tech industry using event study methodology. I provide robust evidence that acquirers of high-tech targets experience significant positive announcement returns. Next to this, I investigate the relation between the financial crisis and high-tech M&A performance and find weak evidence that the financial crisis has a positive influence on the announcement returns. Finally, I investigate the long-term high-tech M&A performance and find negative long-term abnormal returns over the one-year period. The negative long-term abnormal returns indicate that investors are overly optimistic about the expected benefits of high-tech M&As.

JEL classifications: G30; G32; G34; G39; G01; G02

Keywords: Mergers & Acquisitions; stock price performance; high-tech industry; financial crisis

Author: Jelle Assink

Studentnr: s1892746

2 ‘’U.S. dealmaking at record year-to-date high’’

Reuters, May 29, 2015 ‘’A zeal for deals: corporate takeovers are booming once again’’

The Economist, April 18, 2015

The U.S. M&A market is growing at a strong pace. Within this growing M&A market the high-tech industry clearly stands out by growing 62% in deal value over 2014, almost doubling the average market growth rate of 33% (KPMG, 2015).

While the acquiring firm performance of mergers and acquisitions (‘M&As’) has been extensively researched (see, e.g., Jensen and Ruback, 1983; Moeller et al., 2005), previous research focusing on M&A performance in the high-tech industry is severely lacking (Benou et al., 2008). This lack of research is remarkable considering the distinctive high-growth and high-risk nature of the high-tech industry and the questions this raises on acquiring firm performance. On the one hand the attractive growth prospects of high-tech targets give strong potential for value creation leading to higher announcement returns to the acquiring firm. On the other hand, the uncertainty associated with many high-tech firms holds that these prospects may never be realized, which makes investors skeptical on the potential benefits leading to lower announcement returns (Kohers and Kohers, 2000).

This study investigates the acquiring firm performance of high-tech M&As. I use event study methodology to calculate announcement returns to the acquiring firm. In order to do this, I use a unique sample of 735 U.S. high-tech M&As in the 2003-2012 period.

Next to analyzing announcement returns, I investigate the influence of different firm, deal, and environmental characteristics on high-tech M&A performance using a multiple regression model. Although it is well-known that M&A activity is lower during a period of economic downturn, no clarity exists on the acquiring shareholder wealth effects of M&As during a financial crisis (Harford, 2005). Therefore, among the different characteristics, the main focus is on the influence of the financial crisis on acquiring firm performance.

Finally, this paper examines long-term stock returns to see to what extent the announcement stock price reaction on high-tech M&As is an unbiased forecast of future long-run firm performance. In summary, I determine the main research question of this paper as follows:

3 The distinct high-growth, high-risk nature of the high-tech industry combined with a lack of prior research makes that this study fills an important gap in the current M&A literature. The results of this study give insight in high-tech M&A performance, which can be of great value to key-decision makers of high-tech firms when shaping their corporate growth strategy. Furthermore, this study makes a unique contribution by covering a timeframe not earlier investigated in this context. Finally, it adds to the existing literature by being the first to investigate the influence of the financial crisis on M&A performance in the high-tech industry.

In line with previous studies, the results provide robust evidence of significant positive announcement returns to acquiring firms following U.S. high-tech M&As. The findings confirm the influence of the distinct industry nature and indicate that investors are optimistic about the growth prospects and associated value creation potential of high-tech M&As. Next to this, I find weak evidence for a positive influence of the financial crisis on high-tech M&A performance. Finally, the results show significantly negative one-year abnormal returns for acquiring high-tech firms indicating that investors’ perceptions on high-tech M&A are overly optimistic and positive announcement returns are not an unbiased forecast of long-term firm performance.

The remainder of this paper is organized as follows. Section 2 gives an overview of the relevant M&A literature. Based on the literature, the hypotheses are formed. Section 3 describes the data and methodology used in this study. Subsequently, section 4 will discuss and interpret the empirical results. Finally, section 5 will conclude and provide limitations and suggestions for future research.

2. Literature review

4 2.1 M&A motives

M&A performance can be greatly influenced by its underlying motivations. Therefore, this section discusses the most important motives.

The creation of synergies is widely agreed to be the key principle and most important driver of M&As (see, e.g., Seth, 1990; Gaughan, 2010). Synergies are defined as the concept in which the value of two firms combined is greater than the sum of the value of the two firms individually (Gaughan, 2010). Synergies can broadly be divided into operational, financial and managerial synergies and, among others, can be achieved through economies of scale, elimination of duplicate cost factors and taxation benefits. In the high-tech sector synergies are especially realized by sharing the costs and outcomes of R&D.

Next to synergies, M&As can be motivated by monopoly theory (Eckbo, 1983). The monopoly theory states that firms undertake M&As in order to increase market power. Expansion of market power is desirable because it results in increased bargaining power, which in turn positively influences firm profits. The $18.7 billion acquisition of Compaq Corporation by Hewlett-Packard in 2001 provides a good example of a high-tech deal in which monopoly theory played an important role. However, expanding market power through M&As is limited by antitrust regulations.

Additionally, Roll (1986) points at the influence of managerial hubris in M&A decision making. The hubris theory states that the management of acquiring firms in general is overly optimistic on their ability to critically evaluate potential target firms. The over optimism of the management of the acquiring firm typically results in overpriced takeovers and consequently negatively affects the acquiring firm’s shareholder wealth (Trautwein, 1990).

Next to this, agency problems play an important role in M&A decisions (Trautwein, 1990; Jensen, 1986; Amihud, Lev, and Travlos, 1990). Agency problems arise when the interests between managers and firm stakeholders are misaligned. A clear application of agency problems is reflected in the empire building theory (Trautwein, 1990). The empire building theory states that managers primarily initiate M&As to lift their own status and wealth instead of those of the firm’s shareholders. M&As based on both the managerial hubris and the empire building theory are generally value decreasing for the shareholders of the acquiring firm.

5 gained access to Motorola’s extensive patent portfolio and manufacturing technology. Based on this, Link (1988) concludes that the importance of M&As is greater for firms in the high-tech industry compared to firms where competition is non- or less technology based.

2.2 Short-term M&A performance

Previous studies find consistent empirical evidence of significant positive returns to the shareholders of the target firm (Schwert, 1996; Goergen & Renneboog, 2004). However, it remains a matter of ongoing debate whether the shareholders of the acquiring firm lose or benefit from M&As.

Jensen and Ruback (1983) conduct a survey in which they compare the findings of several empirical studies on M&A performance between 1960 and 1970. They conclude that the shareholders of the acquiring firm ‘don’t lose’ from M&As. Bruner (2004) however finds significant negative announcement returns (-1.02%) to the acquiring firm using a sample of 138 U.S. firms between 1990 and 1999. Bruner’s results are supported by Fuller, Netter, and Stegemoller (2002), who also find negative announcement returns to the acquiring firm. In contrast to the negative results, Moeller and Schlingemann (2004) find zero and positive announcement returns to the shareholders of the acquiring firm. Campa and Hernando (2004) confirm the mixed evidence by comparing the results from previous studies and showing that ten studies report negative abnormal returns and seven studies report zero or positive abnormal returns.

Very few studies focus on M&A performance in the high-tech industry. As mentioned in the introduction, this is remarkable considering the industry’s distinct nature. High-tech firms possess desirable growth opportunities and are characterized by relatively high-growth rates (Kohers and Kohers, 2000, 2001). Because of this, they are expected to deliver higher returns to acquiring shareholders compared to targets from other industries. The attractive growth opportunities and associated returns lead to optimism among investors which in turn positively influences high-tech M&A announcement returns (Kohers and Kohers, 2000; Benou et al. 2008; Raggozino, 2006).

6 The high-growth and high-risk nature of the high-tech industry raises questions on the expected announcement returns for shareholders of the acquiring firm. Do investors focus on the strong growth potential of high-tech firms, most likely leading to positive announcement returns? Or do they mainly see the higher degree of risk and uncertainty of the high-tech industry, which would negative influence the announcement returns?

Kohers and Kohers (2000) empirically analyze the performance of high-tech M&As. They investigate announcement returns to the acquiring firm using a sample of 1,634 M&As between 1987 and 1996. In support of their hypothesis that the market tends to assess high-tech M&As favorably, they find significant positive announcement returns of 1.26% over the two-day event window.

In addition to Kohers and Kohers (2000), other empirical studies on the high-tech industry come from Benou et al. (2008) and Ragozzino (2006). Benou et al. (2008) study the wealth effects to sellers and buyers in response to high-tech divestures. They use a sample of 2040 divestitures between 1980 and 2001 and find significant positive announcement returns for acquiring firm shareholders of 2.32% over the three-day event window. Next to this, they show that high-tech deals during the tech-bubble experience a more favorable market reaction than deals before the tech-bubble. The results confirm that stock price reactions in response to high-tech asset divestitures are significantly different from asset divestitures in other industries.

Ragozzino (2006) investigates firm valuation effects of high-tech M&As and focuses on comparing new venture acquiring firms with established acquiring firms. The results show that new ventures experience lower announcement returns than the established firms. Ragozzino (2006) also empirically tests the acquiring firm performance of the total sample and finds statistically significant positive announcement returns of 2.23% over the three-day event window.

In contrast to the ambiguous results of studies focusing on acquiring firm performance in other industries, the limited amount of studies focusing on the high-tech industry consistently find positive announcement returns to the shareholders of the acquiring firm. This indicates that in general, investors seem to be more optimistic about the growth prospects of high-tech firms than concerned about the unique risk profile of high-tech firms.

7 Investor optimism on high-tech firms increased strongly while investor skepticism on the uncertainty and risk involved decreased (Gao and Sudarsanam, 2003). The over optimism on the high-tech industry during the tech-bubble most likely had a positive influence on announcement returns to acquiring high-tech firms.

The tech-bubble burst in 2000 after which the Nasdaq Composite Index went down by 41% in one year. Consequently, investor optimism on high-tech firms decreased accordingly. During this study’s timeframe, the Nasdaq Composite Index did not greatly outperform the S&P 500 and investor optimism on the high-tech industry is therefore assumed to be considerably lower than during the tech-bubble. Following the theory this would lead to lower announcement returns relative to the announcement returns found in previous studies.

To summarize, the existing literature finds conflicting results on the announcement returns for acquiring firms. The tech industry distinguishes itself from other industries by its growth, high-risk nature. Previous studies find evidence that the desirable growth opportunities of the industry stimulate investor optimism and lead to significant positive announcement returns to acquiring firm shareholders. Although it is assumed that this study’s time period inhibits less ‘high-tech optimism’, I follow the theory and empirical evidence from previous high-tech studies and still expect positive announcement returns to acquiring high-tech firms. However, based on the mixed results of previous non-industry specific M&A performance studies, I formulate hypothesis 1 as follows:

Hypothesis 1: high-tech M&As generate announcement returns to the acquiring firm that are significantly different from zero

2.3 Financial crises and M&A performance

The M&A literature identifies several firm and deal characteristics which influence M&A announcement returns. However less research focuses on the influence of environmental factors such as a financial crisis. Therefore, this section investigates the influence of different economic conditions (financial crisis versus non-financial crisis) on announcement returns to the acquiring firm.

8 constrained during a financial crisis. Financial constraints in turn make it more difficult for firms to do large investments such as M&As, resulting in lower M&A activity (Harford, 2005).

However, according to Malmendier and Tate (2005), being financially constrained as an acquirer, although leading to lower M&A activity, has a positive influence on acquiring firm performance. Supported by Roll (1986), they argue that top corporate decisions makers are in general too optimistic about the outcome of their decisions and persistently overestimate their own skills relative to those of others. Financial constraints force managers to evaluate their investment decisions more critically which reduces the overconfidence problem and results in less so-called ‘bad M&As’ (value decreasing M&As). Based on their results, Malmendier and Tate (2005) conclude that: ‘’in the absence of financial constraints, overconfidence can explain a significant portion of acquiring shareholder value lost in merger deals.’’

According to a study by Berger and Bouwman (2009) on the banking industry, the weaker financial position of banks during a crisis gives opportunity to financially strong banks to acquire their weaker brothers at distressed prices. Acquiring at distressed prices in turn has a positive effect on announcement returns. In line with this, Beltratti and Paladino (2011) state that during a period of financial crisis acquiring banks can achieve discounts because targets are forced to sell their assets, primarily because they have to strengthen capital and liquidity indicators. Both papers specifically focus on the banking industry, which inhibits unique characteristics for instance in terms of industry specific capital and liquidity requirements. However, although no empirical evidence exists, the general logic in the theory of financial crises leading to more targets available at discount prices due to financial distress potentially holds for other industries such as the high-tech industry as well.

Finally, announcing a takeover in a period of financial crisis can be seen as a sign of strength of the acquiring firm. A sign of firm strength can give rise to optimism among investors which in turn positively influences acquiring firm shareholder returns.

Based on the theory discussed a positive influence of the financial crisis on high-tech M&A performance can be expected. However, no previous studies have investigated the effects of a financial crisis on M&A performance in the high-tech industry. I use the following hypothesis to test the influence of a financial crisis on high-tech M&A performance:

Hypothesis 2: high-tech M&As generate significantly different announcement returns during a crisis compared to during non-crisis period

9 Next to the potential influence of the financial crisis, the literature identifies several other determinants of M&A performance. In order to control for these determinants influencing the impact of the financial crisis, I included the most well-known M&A performance determinants as control variables in the multiple regression model.

Following Kohers and Kohers (2000) and Moeller, Schlingemann, and Stulz (2004) the relative size of the target firm has a significant influence on acquiring firm performance. The main reason for this is that large targets, relative to their acquirers, offer greater synergy potential than smaller targets. Especially sizeable synergies such as economies of scale are easier achieved when acquiring a relatively large target. Martynova et al. (2006) however warn the acquirer by pointing at integration difficulties accompanied with acquiring a relatively large target.

According to Conn et al. (2005), the ownership status of the target might be of influence on acquiring firm performance. Private firms are expected to be younger and smaller and are generally in an earlier growth stage and more resource restricted than public firms. The marginal benefits of getting access to the resources of the acquiring firm are therefore expected to be higher for a private target than for a public target. In addition, the higher ownership complexity of public firms often leads to a relatively high transaction value for a public target as more stakeholders have to be satisfied. Therefore acquiring a private target is expected to generate higher abnormal returns compared to acquiring a public target.

Additionally, Moeller and Schlingemann (2005) studied the influence of domestic versus cross-border M&As. They find that acquiring firms experience lower abnormal returns in case the target is located in a foreign country. In support of this, Goergen and Renneboog (2004) argue that cross-border M&As more easily lead to complications in the post-merger process due to regulatory and cultural differences between the target and acquiring firm.

Another characteristic of influence is identified by Seth (1990). He argues that industry relatedness between the acquirer and target increases the synergy potential and therefore positively influences announcement returns. As an example, operational costs of highly industry related firms can be minimized more easily due to similar practices and technologies.

10 and Stulz (1996) shows that firms with valuable investment opportunities experience higher abnormal returns when they issue equity. Therefore stock financed deals should positively influence the stock price reaction as it can be seen as a sign of ample profitable investments opportunities. Kohers and Kohers (2000) empirically test the relation between the method of payment and announcement returns. They did not find a significant relation, indicating that the market’s perception of stock versus cash financed M&As does not differ notably.

Lastly, I add the acquiring firm’s market-to-book ratio and debt-to-equity ratio for control purposes. According to Rau and Vermaelen (1998), acquirers with a low market-to-book ratio, the so-called value firms, are expected to outperform acquirers with a high market-to-book ratio, the so-called glamour firms. Furthermore, acquiring firms with higher a debt-to-equity ratio are expected to earn higher announcement returns because highly leveraged firms are more closely monitored, which makes it less likely that they will participate in ‘bad M&As’.

2.5 Long-term M&A performance

In addition to investigating the announcement returns of acquiring high-tech firms, I investigate the long-term abnormal returns to see whether the initial stock price reaction is a reliable forecast of future long-term performance.

Compared to short-term abnormal returns, a much smaller body of research is devoted to examine long-term abnormal returns (Martynova and Renneboog, 2008). The reason for this is that there are severe methodological issues involved in measuring the long-term post M&A performance. The main problem lies in the difficulty of isolating the acquisition effect on long-term performance from other events or factors influencing the firm’s long-term performance.

An often used approach for measuring long-term performance is the buy-and-hold abnormal return approach described by Barber and Lyon (1997). The BHARs show the average multi-year return from investing in firms that complete an acquisition, and selling the firm’s stock at the end of a pre-specified holding period. The returns obtained are then compared to applying the same investment strategy to a benchmark that did not experience an acquisition. In this way, the BHAR approach is said to most accurately measure the investor experience.

11 calendar-time approach because in this approach every period is equally weighted and thus the time-varying misevaluation behavior of managers is ignored.

The literature fails to agree on a robust research design for long-term M&A performance. Therefore, previous studies on long-term M&A performance used different methodologies leading to mixed results. On the one hand, Agrawal et al. (1992) find that shareholders of the acquiring experience a significant loss in the five year post deal period. On the other hand, Loughran and Vijh (1997) find significant positive abnormal returns in the years after the deal.

Only two studies examine the long-term returns of high-tech acquiring firms. These studies come from Kohers and Kohers (2001) and Gao and Sudarsanam (2003). The results of Kohers and Kohers (2001) show that while the market reacts positively to high-tech M&As in the short-term, they significantly underperform compared to their benchmark in the long-term. Their findings suggest that the market exhibits over optimism on the expected benefits of high-tech M&As. In addition to Kohers and Kohers (2001), Gao and Sudarsanam (2003) also find evidence of high-tech acquiring firms underperforming their non-acquiring industry peers in the long-term.

Studies by Kohers and Kohers (2001) and Gao and Sudarsanam (2003) focusing on the high-tech industry shows negative long-term abnormal returns to the acquiring firm. However, previous, non-industry specific studies find mixed results. I formulate the following hypothesis to test the influence of high-tech M&As on long-term firm performance:

Hypothesis 3: High-tech M&As generate long-term abnormal returns to the acquiring firm that significantly differ from zero

3. Data and Methodology

Section 3 presents the research design of the empirical study. First the data collection will be described, followed by the methodology.

12 The data used in this study covers M&A transactions from January 2003 until December 2012. The timeframe includes M&As from two very distinct economic periods. The first period (2003-2007) is a non-crisis period while the second period (2008-2012) is characterized by the global financial non-crisis.

The transaction data is collected from the Thomson ONE database. The Thomson ONE database provides integrated access to several other financial databases1 _{and is used as an industry standard}

around the world (Pagan, Chu, 2009). The transaction data includes: the announcement date, completion date, transaction value, target nation, percentage of shares acquired, target ownership status, acquirer primary industry, target primary industry and the percentage of cash and/or stock used for payment.

In order to calculate the abnormal returns of the acquiring firms, the daily stock returns of the acquiring firms and the daily benchmark returns are needed. Both data should at least be available for a period covering the estimation and event window. The daily stock price returns of the acquiring firms as well as the daily benchmark returns are obtained from Thomson Reuters Datastream.

As a benchmark the Nasdaq Composite Index is used. The Nasdaq is a broad tech-orientated index containing over 3000 firms. Due to its tech nature, the Nasdaq index is expected to most accurately estimate the abnormal returns of this study’s high-tech sample and therefore is the relevant benchmark to use. Other financial information such as the total assets, debt-to-equity ratio and market-to book-ratio of the acquiring firms are also obtained from Thomson Reuters Datastream.

In order to be eligible for the sample, seven filters are applied to the transaction data. First of all the acquirer as well as the target has to be active in the high-technology industry. In line with Kohers and Kohers (2000), this study largely follows the TechAmerica industry classification of the high-tech industry to determine if a firm is considered to be high-tech. TechAmerica, formerly known as AEA (American Electronics Association) is the leading U.S. association for the technology industry. Following the OECD 2011 high-tech industry classification, I exclude the communications services industry. Appendix A provides a list of the different high-tech sectors included in this study’s sample.

A second filter is applied to the ownership status of the acquiring firm. Only public acquirers are included because daily stock data is necessary to calculate the abnormal returns. Third, the acquiring firm has to be located in the U.S. and fourth, the transactions need to have disclosed announcement and completion dates. Furthermore, all transactions with an undisclosed deal value or a deal value lower than 50 million USD are deleted. Additionally, all transactions involving an acquisition of less than 50% of the shares are deleted from the sample. Finally, following Loughran and Vijh (1997), all overlapping events

13 are removed from the sample. If a firm undertakes more than one acquisition, and the estimation and/or event window of the different M&As overlap, the latter transaction is excluded from the sample. After cleaning the data for missing values a final sample of 735 high-tech M&As is left.

Table 1 provides an overview of the final sample. It shows the per year number and value of the included M&A. Average deal value does not significantly differ between the non-crisis (369.4 million USD) and crisis period (385.9 million USD). In line with theory from Harford (2005), the sample shows a decrease in M&A activity after the start of the financial crisis in 2008. The non-crisis period on average contains 88 deals per year adding up to a total of 442 deals, while the crisis period on average contains 57 deals per year adding up to total of 293 deals. However, due to the minimum deal value filter, the difference in merger activity should be interpreted with caution as positive selection bias could be present.

3.2 Methodology Table 1

Overview of the number and value of M&A deals in each year during the sample period This table presents the number of mergers and acquisitions (M&As) during each year of the sample period. Next to this it provides the year's average and total value of the deals in million U.S. dollars.

Year Number of deals

Average value of deals ($m)

14 Section 3.2 discusses the methodology used to perform the event study and regression analyses. Calculating the short-term effects of M&As requires a different methodology from calculating the long-term effects. Based on this, this section is divided in section 3.2.1. and section 3.2.2.

3.2.1. Short-term abnormal returns

Following Brown and Warner (1985) I use standard event study methodology to assess the impact of high-tech M&As on acquiring firm performance. An event study measures the effect of an event on the value of the specific firm by assessing to which degree the stock price performance around the event has been abnormal (Brown and Warner, 1980). The abnormal stock return is estimated by subtracting the expected/normal (henceforth, normal) stock return from the actual stock return during the event window.

The semi-strong efficient market hypothesis (EMH) forms the basis of event study methodology. The EMH assumes that the stock price reflects the present discounted value of all future returns and adjusts immediately to newly available public information (Fama, 1970). Therefore, according to the EMH, announcement returns provide a good measure of M&A performance.

In order to perform an event study I first have to define the estimation window, the event window and the event date. The event date in this study is the given by the announcement date of the acquisition. The event window is defined by the period surrounding the event over which the abnormal returns of the

firm involved in the event are measured. Different event windows can be used to analyze announcement returns. Most studies prefer to keep the event window as short as possible because the smaller the window the smaller the chance that events other than the acquisition have influence on the stock price (Mackinlay, 1997). On the other hand, Andrade et al. (2001) show that the first news leak regarding a deal often precedes the official announcement by a week, which could already lead to event related abnormal returns a week before the announcement. In order to account for the issues addressed by Mackinlay (1997) and Andrade et al. (2001), this study uses five different event windows: (-1,1), (-2,2), (-3,3), (-5,5), (-10,10). Furthermore, the multiple event windows also serve as a test of robustness.

Finally, the estimation window is the period used to estimate the parameters needed to determine the normal returns in the event window. The estimation window should not contain any dates that are also used in the event window and should be sufficiently long to generate a good estimate of the normal returns (Mackinlay, 1997). The estimation window in this study spans from 150 days before the announcement date until 11 days before the announcement, covering 140 days in total.

Figure 1 provides an overview of the time sequence of the event study. The event date is defined as

τ

15 Figure 1. Time sequence of an event study.

To appraise the impact of an event a measure of abnormal return is needed. Different models can be used to estimate abnormal returns (see, e.g., constant mean return model, market model, factor model). Brown and Warner (1985) use simulation procedures to investigate the reliability of the different event study methodologies and conclude that the simple one-factor market model procedure is well-specified. Therefore, as in most previous M&A performance studies (see, e.g., Moeller and Schlingemann, 2005; Kohers and Kohers, 2000), I use the market model approach to estimate the abnormal returns. The market model relates the stock price return of a given security to the stock price return of the market index. It assumes a constant and linear relation between the market index return and the firm’s individual return. For any firm i at time t in the estimation window the market model is:

t i t m i i t i R R_, 







in which Ri,tand Rm,t are the daily returns over the estimation window and the market index respectively. The zero mean standard error is reflected by





ˆ



ˆ

parameters. After the market model parameters are estimated, the abnormal returns over the event window can be measured by subtracting the normal returns from the actual return in the event window. The abnormal return of firm i at time t is described as follows:

) ˆ ˆ ( _, , ,t it i i mt i R R AR  





(

( Event window ]

Time

0

𝑇1 𝑇2

16 where 𝑅𝑖,𝑡 is the actual return and (





defined as the expected return of the firm without conditioning on the event taking place. The normal returns are based on the actual returns of the market index and the parameters



ˆ



ˆ

𝐶𝐴𝑅𝑖(𝜏1,𝜏2) = ∑ 𝐴𝑅i,t

𝑇

𝑡=1 , (3)

in which 𝐶𝐴𝑅𝑖(𝜏1,𝜏2) is defined as the cumulative abnormal return of firm i over the event window.

The cumulative average abnormal return, as described in Eq. (4), is an average of the CARs of the entire sample: 𝐶𝐴𝐴𝑅(𝜏1,𝜏2) = 1 𝑁∑ 𝐶𝐴𝑅𝑖 𝑁 𝑖=1 . (4)

After the abnormal returns are calculated, I apply the traditional method of testing if the cumulative average abnormal returns are significantly different from zero, following Mackinlay (1997) and Brown and Warner (1985). To increase robustness, I add a standardized cross-sectional test-statistic as opposed by Boehmer et al. (1991). The test-statistic following the traditional method is calculated as follows:

T-statistic = _{𝜎𝐶𝐴𝐴𝑅(𝜏}𝐶𝐴𝐴𝑅(𝜏1,𝜏2)

1,𝜏2) , (5)

where 𝐶𝐴𝐴𝑅(𝜏1,𝜏2)represents the cumulative event window average abnormal return, and

𝜎𝐶𝐴𝐴𝑅(𝜏1,𝜏2) represents the standard deviation of the cumulative average abnormal return. The

𝜎𝐶𝐴𝐴𝑅(𝑡1,𝑡2) is calculated as follows: 𝜎𝐶𝐴𝐴𝑅(𝜏1,𝜏2) = √ 1 𝑁2∑ 𝜎2(𝐶𝐴𝑅𝑖) 𝑁 𝑖=1 , (6)

17 𝜎2(𝐶𝐴𝑅𝑖) = (𝜏1− 𝜏2+ 1) 𝜎2(𝐴𝑅𝑖,𝑡), (7)

in which (𝜏1− 𝜏2+ 1) represents the number of days in the event window and 𝜎2(𝐴𝑅𝑖,𝑡) represents

the conditional variance of the abnormal returns. Following Mackinlay (1997), the estimation window used in this study is large enough to assume that the additional variance due to the sampling error in



ˆ



ˆ

2_(𝐴𝑅

𝑖,𝑡) is assumed to be equal to the variance of the error terms

from Eq. (1).

In addition to the traditional test-statistic, I add a standardized cross-sectional test-statistic developed by Boehmer et al. (1991). The main improvement of the standardized cross-sectional test-statistic over the traditional test-statistic is that it accounts for event-induced increases in the variance of stock returns. Event-induced variance occurs when the variance of the event window exceeds the variance of the estimation period. Boehmer et al. (1991) prove that even if an event causes only a minor increase in variance the null hypothesis of zero cumulative average abnormal returns is rejected more often than actually true. Therefore, adding the standardized cross-sectional test statistic will improve robustness of the event study results.

In order to calculate the standardized cross-sectional test-statistic, the abnormal returns are first standardized by dividing the abnormal returns by the estimation-period standard deviation. The cumulative standardized abnormal returns in the event window are then calculated following:

𝑆𝐶𝐴𝑅𝑖 (𝜏1,𝜏2)= ∑𝑁𝑖=1𝑆𝐴𝑅𝑖,, (8)

after which the cumulative standardized average abnormal returns are calculated following:

𝑆𝐶𝐴𝐴𝑅(𝜏1,𝜏2) = 1

𝑁∑ 𝑆𝐶𝐴𝑅𝑖(𝜏1,𝜏2)

𝑁

𝑖=1 . (9)

The cumulative standardized average abnormal return is then estimated from the cross-section of event-window standardized average abnormal returns following:

𝜎𝐶𝑆𝐴𝐴𝑅(𝜏1,𝜏2) = √ 1

𝑁(𝑁−1)∑ (𝑆𝐶𝐴𝑅𝑖(𝜏1,𝜏2) −

𝑁

18 After which the standardized cross-sectional test-statistic can be calculated as follows:

𝑇𝐵𝑜𝑒ℎ𝑚𝑒𝑟 =

𝑆𝐶𝐴𝐴𝑅(𝜏1,𝜏2)

𝜎𝑆𝐶𝐴𝐴𝑅(𝜏1,𝜏2) . (11)

Finally, I split the total sample in a crisis (2008-2012) and non-crisis sample (2003-2007) and apply the same significance testing methodologies as with the total sample. To test if the CARs of the crisis and non-crisis sample are significantly different from each other, I apply a t-test for equality of means.

After the abnormal returns of the three samples are calculated and tested for significance, a multiple regression analysis is performed. In this regression model, the CARs serve as the dependent variable. Next to this a set of firm, deal and environmental characteristics, as described in section 2.4, are added to the model as control variables. All variables used in the regression model are described in more detail in table 2. The regression analysis is performed for the total sample, crisis sample and non-crisis sample. The regression model is described as follows:

𝐶𝐴𝑅𝑖 = 𝛼𝑖+ 𝛽1𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒𝑆𝑖𝑧𝑒𝑖+ 𝛽2𝐶𝑟𝑜𝑠𝑠𝐵𝑜𝑟𝑑𝑒𝑟𝑖+ 𝛽3𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦𝑅𝑒𝑙𝑎𝑡𝑒𝑑𝑛𝑒𝑠𝑠𝑖+ (12)

𝛽4𝐶𝑟𝑖𝑠𝑖𝑠𝑖+ 𝛽5𝑂𝑤𝑛𝑒𝑟𝑠ℎ𝑖𝑝𝑖+ 𝛽6𝑃𝑎𝑦𝑚𝑒𝑛𝑡𝑀𝑒𝑡ℎ𝑜𝑑𝑖+ 𝛽7𝐷/𝐸𝑖+ 𝛽8𝑀/𝐵𝑖+ 𝜀𝑖

Table 2

Description of the regression variables and data sources

This table presents the variables used in the multiple regression analysis. It provides the variable name, type, definition and data source

Variable Type Definition Source

CAR(-1,1) Numerical Cumulative abnormal return over the three-day event window

Datastream CAR(-2,2) Numerical Cumulative abnormal return over the

five-day event window

Datastream CAR(-3,3) Numerical Cumulative abnormal return over the

seven-day event window

Datastream CAR(-5,5) Numerical Cumulative abnormal return over the

eleven-day event window

Datastream CAR(-10,10) Numerical Cumulative abnormal return over the

21-day event window

Datastream Crisis Dummy Coded 1 for M&A between 2008-2013; 0

for 2003-2007

19 Relative size Numerical Relative size of the target firm. Derived

by: Relative size = transaction value / (transaction value + market value acquirer at day t=-11)

Datastream

Ownership Dummy Coded 1 if the target is privately owned; 0 if otherwise

SDC Method of payment Dummy Coded 1 if the transaction is cash

financed; 0 if otherwise

SDC Industry relatedness Dummy Coded1 if the 4-digit SIC code of the

acquirer corresponds with the target; 0 if otherwise

SDC

Cross-border Dummy Coded 1 if the target is located outside U.S.; 0 if otherwise

SDC D/E Numerical Debt to equity ratio of the acquiring firm.

Derived by: D/E =(long term debt + short term debt and current portion of long term debt) / common equity )

Datastream

M/B Numerical Market to book ratio of the acquiring firm. Derived by: M/B = (Market value / common equity)

Datastream

3.2.1 Long-term abnormal returns

Following Barber and Lyon (1997) I use the buy-and-hold abnormal return (BHAR) approach to investigate the long-term post high-tech M&A performance. The BHARs are calculated for the one-year, two-year and three-year period after completion of the deal and are defined as the difference between the realized buy-and-hold return and a benchmark buy-and-hold return.

As a benchmark I use the Nasdaq Composite market index. According to Barber and Lyon (1997), next to using a market index as benchmark, matched control firms and matched reference portfolios based on comparable market to book- and size ratios should be used to prevent for biased test statistics2_{. Biases}

particularly may arise from rebalancing of the benchmark index, new listings and skewness of multiyear abnormal returns.

A benchmark should represent the return in case no event takes place. Therefore, firms that did undertake an acquisition in the three-year event period are not eligible to serve as a control firm or reference portfolio firm. More strictly, even when the control firm is involved in a takeover in the three years before the three-year event window, the long-term stock price effect of this event still mixes with

20 the effect of the relevant event and therefore the firm should not be selected. Applying this rule to this study’s sample makes it difficult to find truly comparable control firms. The Nasdaq Composite comprises over 3000 stocks. Therefore, single firm corporate events within the event window period have a negligible influence on the returns of the complete index. For this reason I solely use the Nasdaq Composite Index as a benchmark. However, acknowledging the potential influence of the biases addressed by Barber and Lyon (1997), future research is needed to improve robustness of the long-term M&A performance results. Unfortunately, this is beyond the scope of this research.

To ensure stock return data availability for the one-year, two-year, and three-year post event period I delete all M&As taking place in 2012. After this, the sample covers 672 M&As in the 2003-2011 period. Finally, following Loughran and Vijh (1997), all overlapping events are removed from the sample which leaves a final long-term sample of 317 M&As. Following Barber and Lyon (1997, 1999), A T-year BHAR for event firm i starting from the date of completion of the deal is defined as:

𝐵𝐻𝐴𝑅𝑖,(𝑡, 𝑇) = ∏𝑇𝑡=1(1 + 𝑅𝑖,𝑡)− ∏𝑇𝑡=1(1 + 𝑅(𝐵,𝑡)) , (13)

where 𝑅𝑖,𝑡 represents the firm’s stock return and 𝑅(𝐵,𝑡) represents the Nasdaq Composite index. The

sample average buy-and-hold abnormal returns is calculated as follows:

𝐵𝐻𝐴𝐴𝑅(𝑡, 𝑇) = 1

𝑁∑ (𝐵𝐻𝐴𝑅𝑖(𝑡, 𝑇)) 𝑁

𝑖=1 . (14)

The cross-sectional test statistic, used for testing if the BHAARs are significantly different from zero, is calculated as follows:

T-statistic = _{𝜎𝐵𝐻𝐴𝐴𝑅(𝑡,𝑇)}𝐵𝐻𝐴𝐴𝑅(𝑡,𝑇) , (15)

in which 𝜎𝐵𝐻𝐴𝐴𝑅(𝑡,𝑇) is the standard deviation of the buy-and-hold average abnormal returns, estimated

from the cross-section of buy-and-hold abnormal returns following:

𝜎𝐵𝐻𝐴𝐴𝑅(𝑡, 𝑇) = √ 1

𝑁−1∑ (𝐵𝐻𝐴𝑅𝑖(𝑡, 𝑇) − 𝐵𝐻𝐴𝐴𝑅(𝑡, 𝑇)) 2 𝑁

21 4. Results

In this section the results of the different analyses are presented. First the results of the short-term event study are discussed. After this, the descriptive statistics and results of the multiple regression analysis are discussed, followed by the long-term event study results.

4.1 Short-term event study results

The results of the short-term event study are presented in table 3. The table shows the CARs and the accompanied significance statistics for the total sample, the crisis sample and the non-crisis sample.

Table 3

Cumulative average abnormal announcement returns for the total sample, crisis sample and non-crisis sample

This table includes the CAARs for the total sample, crisis sample and non-crisis sample over the three-, five-, seven-, eleven-, and 21-day event window. The CAARs are measured using the market model. The difference between crisis and non-crisis abnormal returns is analyzed using t-tests for equality of means. P-value (B) is the p-value obtained by following the Boehmer et al. (1991) method. *** p<0.01, ** p<0.05, * P<0.1.

Variables Total Sample Crisis sample (1)

22 CAAR(-5,5) 0.0068 0.0093 0.0051 -0.0042 p-value 0.0180** 0.5708 0.1978 0.5579 p-value (B) 0.0145** 0.1204 0.0593* CAAR(-10,10) 0.0041 0.0105 -0.0002 -0.0106 p-value 0.3011 0.8888 0.9708 0.2237 p-value (B) 0.1909 0.1783 0.5698 N 735 293 442

The second column represents the results for the total sample of high-tech M&As. Consistent with the findings of Kohers and Kohers (2000) the results show positive announcement returns. The positive announcement returns earned over the (-1,1), (-2,2), (-3,3) windows are significant at the 1% level, the returns over the (-5,5) event window are significant at the 5% level and the (-10,10) event window returns do not significantly differ from zero. The significant positive announcement returns can be explained by the attractive growth opportunities of the high-tech industry, which stimulated investor optimism on high-tech M&As (Kohers and Kohers, 2000). The finding indicates that investors are optimistic about the growth prospects and associated value creation potential of high-tech M&As.

The positive announcement returns vary among the different event windows between 0.68% and 0.92%, which is significantly lower than announcement returns found in previous studies. The lower abnormal returns could be explained by the lower high-tech optimism among investors during this study’s timeframe compared to studies using a timeframe covering the tech-bubble in the 1990’s (Kohers and Kohers, 2000). Based on the results I reject the first null hypothesis and conclude that firms acquiring high-tech targets generate significant positive announcement returns.

The third and the fourth column of table 3 represent the CAARs of the crisis and non-crisis sample. Both samples earn a positive significant CAAR of 1.39% and 0.61% respectively over the three-day event window. The significance of the CAARs decreases as the length of the event window increases. The decreasing effect could be explained by the lower influence of event related news leaks when pre-event days in the event window increase (Andrade et al. ,2001).

23 results of the multiple regression analysis, which shows if the financial crisis does significantly influence M&A performance after controlling for a set of firm and deal characteristics.

As described in section 3.2, the multiple event windows do not only provide information on different short-term timeframes, they also serve as a robustness check. As can be seen in table 3, the CARs are consistently significant for the three-, five- and seven-day event window, which increases reliability of the results. To further improve robustness, I also test the CAARs for being significantly different from zero by applying the standardized cross-sectional test-statistic developed by Boehmer et al. (1991). The results of standardized cross-sectional significance test are given by ‘p-value (B)’ in table 3. The p-values obtained by the traditional method and the method developed by Boehmer et al. (1991) do not differ notably, which indicates that the announcement returns are robust against event-induced variance.

4.2 Descriptive statistics

Table 4 presents the descriptive statistics of the variables used in the regression model for the total sample, the crisis sample and the non-crisis sample. Panel A shows the total sample of 735 M&As, panel B shows the crisis sample of 293 M&As and panel C shows the non-crisis sample of 442 M&As. As discussed in section 4.1, the CARs in the crisis sample are higher than in the non-crisis period, which holds for all five event windows. The relative size of the target for the total sample on average is 0.1369. During the crisis the relative size of the target is significantly lower (0.1175) than during the non-crisis period (0.1500). For the other variables, the differences between the variables of the two samples are considered too be small.

Table 4

Descriptive statistics for the total sample, the crisis sample and the non-crisis sample

Panel A represents the total sample and includes 735 M&As in the 2003-2012 period. Panel B includes the M&As from the total sample announced in the 2008-2012 period. The non-crisis sample includes the M&As from the total sample announced in the 2002-2007 period. Relative size is defined as the transaction value divided by the transaction plus the total assets of the acquiring firm. The D/E variable represents the debt-to-equity ratio of the acquirer and the M/B variable represents the market-to-book ratio of the acquirer. The method of payment dummy is coded 1 if the transaction is completely cash financed. The industry relatedness dummy is coded 1 if the 4-digit SIC code matches between bidder and target. The ownership dummy is coded 1 if the target is privately owned. The crisis dummy is coded 1 if the transaction takes place in the 2008-2012 period. The cross-border dummy is coded 1 if the target is located outside the United States.

Panel A: Total sample

Variable N Min Max Mean Standard deviation

CAR(-1,1) 735 -0.2836 0.5037 0.0092 0.0694

CAR(-2,2) 735 -0.2893 0.5470 0.0091 0.0732

24 CAR(-5,5) 735 -0.4246 0.5882 0.0068 0.0944 CAR(-10,10) 735 -0.4669 0.5498 0.0041 0.1159 Relative size 735 0.0010 0.9720 0.1369 0.1372 D/E 735 -63.1389 72.9182 0.3679 3.8149 M/B 735 -99.3100 224.0800 3.2484 10.1813 Method of payment 735 0 1 0.5361 0.4990 Industry relatedness 735 0 1 0.4857 0.5001 Ownership 735 0 1 0.5864 0.4928 Crisis 735 0 1 0.3986 0.4900 Cross-border 735 0 1 0.1973 0.3982

Panel B: Crisis sample

Variable N Min Max Mean Standard deviation

CAR(-1,1) 293 -0.2350 0.5037 0.0139 0.0676 CAR(-2,2) 293 -0.2368 0.5470 0.0121 0.0755 CAR(-3,3) 293 -0.4413 0.5262 0.0132 0.0834 CAR(-5,5) 293 -0.4246 0.5882 0.0093 0.0931 CAR(-10,10) 293 -0.3242 0.5282 0.0105 0.1109 Relative size 293 0.0010 0.7770 0.1175 0.1191 D/E 293 -17.9220 5.7178 0.3067 1.5255 M/B 293 -63.2100 32.4100 2.7203 6.2968 Method of payment 293 0 1 0.5495 0.4984 Industry relatedness 293 0 1 0.5085 0.5008 Ownership 293 0 1 0.6177 0.4868 Cross-border 293 0 1 0.2287 0.4207

Panel C: Non-crisis sample

Variable N Min Max Mean Standard deviation

CAR(-1,1) 442 -0.2840 0.4050 0.0060 0.0700

25 CAR(-3,3) 442 -0.2500 0.3100 0.0040 0.0770 CAR(-5,5) 442 -0.3030 0.5320 0.0050 0.0950 CAR(-10,10) 442 -0.4670 0.5500 0.0000 0.1190 Relative size 442 0.0010 0.9720 0.1500 0.1470 D/E 442 -63.1390 72.9180 0.4080 4.7620 M/B 442 -99.3100 224.0800 3.5990 12.0820 Method of payment 442 0 1 0.5270 0.5000 Industry relatedness 442 0 1 0.4710 0.5000 Ownership 442 0 1 0.5660 0.4960 Cross-border 442 0 1 0.1760 0.3820 4.3 Regression results

26 Table 5

Cross-sectional regression analysis of cumulative abnormal returns for the total sample, crisis sample and non-crisis sample

This table provides the estimates of the parameters of the multiple regression analysis on the total sample, crisis sample and non-crisis sample. *** p<0.01, ** p<0.05, * P<0.1.

Panel A: total sample

CAR(-1,1) CAR(-2,2) CAR(-3,3) CAR(-5,5) CAR(-10,10)

Variable B P-value B P-value B P-value B P-value B P-value

27 Panel B: crisis sample

CAR(-1,1) CAR(-2,2) CAR(-3,3) CAR(-5,5) CAR(-10,10)

Variable B P-value B P-value B P-value B P-value B

P-value Constant -0.001 0.941 -0.003 0.849 0.001 0.925 -0.004 0.794 0.011 0.592 Relative size 0.103 0.004*** 0.105 0.008*** 0.059 0.180 0.036 0.461 0.060 0.303 D/E 0.006 0.275 0.011 0.062* 0.010 0.129 0.007 0.355 0.007 0.463 M/B -0.001 0.343 -0.003 0.068* -0.002 0.136 -0.002 0.272 -0.002 0.355 Method of payment -0.004 0.675 -0.002 0.831 0.006 0.599 0.009 0.428 0.001 0.957 Industry relatedness 0.013 0.098* 0.011 0.210 0.014 0.167 0.026 0.020** 0.016 0.220 Ownership 0.002 0.794 0.008 0.417 0.001 0.898 -0.003 0.778 -0.010 0.503 Cross-border -0.008 0.408 -0.014 0.172 -0.011 0.334 -0.016 0.228 -0.029 0.063 N 293 293 293 293 293 R-Squared 0.058 0.059 0.027 0.034 0.029

Panel C: non-crisis sample

CAR(-1,1) CAR(-2,2) CAR(-3,3) CAR(-5,5) CAR(-10,10)

Variable B P-value B P-value B P-value B P-value B

28 For the total sample the crisis dummy has a significant positive sign of 0.010. This means that if an acquisition takes place during the financial crisis it leads to a 1.0% higher CAR compared to when it takes place in the non-crisis period. The positive relation is only significant at the 10% level for the three- and seven-day event window. Therefore the evidence for the positive influence of the financial crisis on short-term acquiring firm performance is considered to be weak.

The findings on the influence of the financial crisis are in line with the theory by Malmendier and Tate (2005), which states that being financially constrained leads to decreased optimism among acquiring managers and therefore reduces the number of ‘bad M&As’. The results are also supported by theory from Berger and Bouwman (2009) and Beltratti and Palladino (2011). In their study on the banking industry they argue that a financial crisis ultimately leads to more bargains in the markets which in turn have a positive effect on the abnormal returns for the acquiring firm. In order to improve robustness of the relation between the financial crisis and acquiring high-tech firm performance, future research using a more extensive dataset is needed.

The control variables used in the regression model mostly have signs consistent with the literature. Relative size of the transaction has a significant positive influence on acquiring firm performance across all event windows. The positive relation supports the results of previous studies claiming that large target firms are able to provide greater synergies (Kohers and Kohers, 2000).

In line with the literature, the private ownership status of the target has a positive effect on the announcement returns (Moeller et al. 2004, Kohers and Kohers 2000). The finding suggests that the market has higher expectations from a young, private firm because its growth is more stimulated by getting access to the acquirers resources compared to a more established, public firm.

The method of payment variable shows a slightly positive sign across all event windows. The relation is significant for the seven- and eleven-day event window at the 5% level. The slightly positive relation can be explained by the signaling hypothesis which states that stock payments are a sign of overvalued stocks (Rau and Vermaelen, 1998).

29 Cross-border M&As do not have a significant influence on the CARs, which contradicts the negative relation found by Moeller and Schlingemann (2005). A reason for this could be that technology is less bounded by borders and thus can be more easily integrated into the acquiring firm compared to more culturally restricted assets. Furthermore the acquirer’s debt-to-equity ratio and market-to-book ratio do not significantly influence the dependent variable.

Next to the total sample, the multiple regression analysis is also performed on the crisis and non-crisis sample. In line with the results for the total sample and the existing literature, the relative size of the target has a highly significant and positive influence on the CARs for both the crisis and the non-crisis sample.

The private ownership status of the target firm in the non-crisis sample has a positive influence on the CARs in all event windows. The relation between private ownership and the CARs in the crisis sample is not significant. The method of payment variable shows a positive sign for the non-crisis sample, which means investors prefer cash financed deals. In line with the results for the total sample, the relation is only significant for the seven- and eleven-day event window. In the crisis sample, the method of payment is not of significant influence on the CARs.

In contrast to the results for the total sample, the industry relatedness variable significantly affects the CARs in the crisis as well as the non-crisis sample. For the non-crisis sample the relation appears to be negative which is remarkable as previous studies all point at a positive relationship (Healy et al., 1992; Kohers and Kohers, 2000). The crisis sample shows a significant positive relation between industry relatedness and firm performance for the three-, and eleven-day event window. The positive relation consistent with the theory by Hotchkiss et al. (1998), which argues that firms that acquire other financially distressed firms tend to know very well the true value of their target, and thus are better able to acquire at a bargain price.

Cross-border M&As in the non-crisis sample earn a 1.7% significant higher announcement return than domestic M&As. The positive cross-border result is remarkable because it contradicts with the results of Goergen and Renneboog (2004). They argue that cross-border M&A negatively influence abnormal returns because they more easily lead to complications in the post-merger process due to higher regulatory and cultural differences between the target- and acquiring firm. Knowing the rapid technological development in the past decades, the contradicting results could potentially stem from the different timeframe used.

30 financial crisis dummy, has a VIF value of 1.029. All other variables have VIF values below seven. Therefore, the regression results are not influenced by multicollinearity.

4.4 Long-term abnormal returns

The results of the short-term study show that acquiring firm shareholders respond favorably to high-tech M&A announcements. The long-term study examines if the high expectations about the future performance are actually justified. Are the positive announcement returns an unbiased forecast for the long-term performance of tech M&A? Or are investors overly optimistic on the prospects of the high-tech industry and will long-term stock price returns be negative?

The results of long-term study are presented in table 9 containing both a sample with overlapping events and a sample excluding the overlapping events. As described in the methodology section, the sample with only non-overlapping events is considered to deliver the most reliable results (Loughran and Vijh, 1997). Therefore the conclusions will be based on the results from the sample that excludes overlapping events.

In line with Kohers and Kohers (2001) and Gao and Sundersanam (2004), the long-term BHARs for the non-overlapping sample are negative for the one-, two- and three-year period. The results are only significant for the one-year BHARs at the 5% level. The one-year result shows a negative BHAR of 5.4% while the two- and three-year BHAR are -4.3% and -4.1% respectively.

The BHARs are also calculated for the crisis- and non-crisis sample. Consistent with the total sample results, significant negative returns of 5.8% are found for the oneyear period for the crisis sample and -5.5% for the non-crisis sample. Next to this, the three-year BHAR of the crisis sample shows a significant negative abnormal return of 13.9%.

31 5. Conclusions, limitations and future research

5.1 Conclusions

This paper examines the acquiring shareholder wealth effects of 735 U.S. high-tech M&As between 2003 and 2012. The cumulative abnormal returns around the announcement date are measured by using the market model. I expect that high-tech M&As generate positive announcement returns to the acquiring firm. Additionally, I argue that M&As during a financial crisis generate higher announcement returns then M&As during a non-crisis period. Finally, long-term acquiring shareholder wealth effects are investigated to see if the initial announcement stock price reaction is an unbiased forecast of future long-term performance.

The results show that high-tech M&As generate significant positive announcement returns to the acquiring shareholder. The positive short-term results are in line with the literature and can be explained by the distinct high-growth character of the high-tech industry. Investors tend to assess the high-tech

Table 6

Buy-and-hold average abnormal returns for the total sample, crisis sample and non-crisis sample

The first column shows the one-year, two-year, and three-year buy-and-hold average abnormal returns for the total sample the crisis sample and the non-crisis sample. For each sample the BHAARs are calculated for a sample including overlapping events and a sample excluding overlapping events. *** p<0.01, ** p<0.05, * P<0.1.

Including overlapping events Excluding overlapping events

Variables N Mean p-value N Mean P-value

32 industry favorably relative to other industries because it contains attractive growth opportunities with great upside potential, translated into a positive stock price reaction after a high-tech takeover.

The positive announcement returns are lower compared to findings from previous studies. An explanation for this can be found in the different timeframes used in previous studies (see, e.g., Kohers and Kohers, 2000; Benou et al., 2008; Ragozinno, 2006). Previous studies use M&A samples from the 1990s which is a period characterized by relatively high investor optimism on high-tech firms (Kohers and Kohers, 2000). The relatively high investor optimism positively stimulates announcement returns and could therefore be explanatory to the lower announcement returns in this study.

I find weak evidence for the positive influence of a financial crisis on high-tech M&A performance. After controlling for several firm, and deal characteristics the results show a at 10% level significant positive influence on the acquiring firm performance. The results can be explained by the lower overconfidence problem due to more financial constraints during a financial crisis as outlined by Malmendier and Tate (2005). The weak evidence asks for future research to increase robustness and provide more details on the relation between a financial crisis and M&A performance.

In contrast to the positive announcement returns, the long-term event study shows significantly negative buy-and-hold abnormal returns for the one-year period after completion of the deal. The results for the two- and three year period after completion of the deal are negative as well but not statistically significant. Given the initial positive reaction to a high-tech M&A announcement, the results indicate that investors tend to be overly optimistic about the growth prospects of high-tech firms and underestimate the associated unique industry risk.

5.2 Limitations and suggestions for future research

The following limitations have to be considered when interpreting the results and conclusions of this study. First of all a larger sample size, especially after splitting up the data in a crisis and non-crisis sample and after excluding overlapping events, would improve robustness of the results.

33 Third, as discussed in section 2.5, measuring long-term M&A performance is subject to methodological issues. In this study I use a market index as a benchmark in the long-term event study. Stronger evidence could be obtained by adding a matched control firm and a matched reference portfolio approach. Furthermore, to improve robustness, Fama’s (1998) calendar-time approach could be added to the long-term study as better deals with the bad model problem. Finally, Kohers and Kohers (2001), point at the influence of agency problems on term high-tech M&A performance. The determinants of long-term M&A performance provide a rich area for future research.

6. Reference list

Amihud, Y., Lev, B., Travlos, N., 1990. Corporate control and the choice of investment financing: the case of corporate acquisitions. Journal of Finance 45(2), 603-616.

Andrade, G., Mitchell, M., Stafford, E., 2001. New evidence and perspectives on mergers. The Journal of Economic Perspectives 15, 103–120.

Asquith, P., Bruner, R., Mullins, D., 1990. Merger returns and the form of financing. Working Paper. Barber, B., Lyon, J., 1997. Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of Financial Economics 43, 341-372.

Barber, B., Lyon, J., 1999. Improved methods for tests of long-run abnormal stock returns. Journal of Finance 54(1), 165-201.

Beltratti, A., Paladino, G., 2013. Is M&A different during a crisis? Evidence from the European banking sector. Journal of Banking & Finance 37(12), 5394-5405.

Benou, G., Madura, J., Ngo, T., 2008. Wealth creation from high-tech divestitures. The Quarterly Review of Economics and Finance 48, 505-519.

Berger, A., Bouwman, C., 2009. Bank liquidity creation. The Review of Financial Studies 22(9), 3779-3837. Boehmer, E., Masumeci, J., Poulsen, A., 1991. Event-study methodology under conditions of event-induced variance. Journal of Financial Economics 30(2), 253-272.

34 Brown, S., Warner, J., 1985. Using daily stock returns: the case of event studies. Journal of Financial Economics 14(1), 3-31.

Bruner, R., 2004. Where M&A pays and where it strays: a survey of the research. Journal of Applied Corporate Finance 16(4), 63-76.

Campa, J., Hernando I., 2004. Shareholder value creation in European M&A. European Financial Management 10(1), 47-81.

Cartwright, S., Schoenberg, R., 2006. Thirty years of mergers and acquisitions research: recent advances and future opportunities. British Journal of Management 17, 1-15.

Chakrabarti, A., Hausschildt, J., Süverkrüp, C., 1994. Does it pay to acquire technological firms? R&D Management 24(1), 47-56.

Conn, R., Cosh, A., Guest, P., Hughes, A., 2005. The impact of UK acquirers of domestic, cross-border, public and private acquisitions. Journal of Business Finance & Accounting 32(5-6), 815-870.

Dutta, S., Jog, V., 2009. The long-term performance of acquiring firms: A re-examination of an anomaly. Journal of Banking & Finance 33(8), 1400-1412.

Eckbo, B., 1983. Horizontal mergers, collusion, and stockholder wealth. Journal of Financial Economics 11, 241-273.

Fama, E., 1970. Efficient capital markets: a review of theory and empirical work. Journal of Finance 25(2), 383-417.

Fama, E., 1998. Market efficiency, long-term returns, and behavioral finance. Journal of Financial Economics 49(3), 283-306.

Fuller, K., Netter, J., Stegemoller, M., 2002. What do returns to acquiring firms tell us? Evidence from firms that make many acquisitions. Journal of Finance 57(4), 1763-1793.

Gao, L., Sudarsanam, P., 2003. Value creation in UK high technology acquisitions. EFMA 2004 Basel Meetings Paper.

Goergen, M., Renneboog, L., 2004. Shareholder wealth effects of European domestic and cross-border takeover bids. European Financial Management 10(1), 9-45.

35 Harford, J., 2005. What drives merger waves? Journal of Financial Economics 77, 529-560.

Healy, P., Palepu, K., Ruback, S., 1992. Does corporate performance improve after mergers? Journal of Financial Economics 31, 135-175.

Hotchkiss, E., Mooradian, R., 1998. Acquisitions as means of restructuring firms in chapter 11. Journal of financial intermediation 7, 240-262

Jensen, M. and Ruback, R., 1983. The market for corporate control. Journal of Financial Economics 11, 5-50.

Jensen, M. 1986. Agency cost of free cash flow, Corporate Finance, and Takeovers. American Economic Review 76(2).

Jung, K., Kim, Y., Stulz, R., 1996. Timing, investment opportunities, managerial discretion, and the security decision. Journal of Financial Economics 42(2), 159-186.

Kennedy, P. 1992. A guide to econometrics. Blackwell publishers, Oxford.

Kohers, N., Kohers T., 2000. The value creation potential of high-tech mergers. Financial Analysts Journal 53, 40-50.

Kohers, N., T. Kohers, T., 2001. Takeovers of technology firms: expectation versus reality. Financial Management 30(3) 35-54.

KPMG, 2015. 2015 M&A outlook survey report. Retrieved from: http://www.kpmgsurvey-ma.com/pdf/298400_NSS_M&A_OutlookSurvey2015v14_WEB.pdf

Link, A., 1988. Acquisitions as sources of technological innovation, Mergers and Acquisitions 23, 36-39. Loughran, T., Vijh, A., 1997. Do long-term shareholders benefit from corporate acquisitions? Journal of Finance 52(5), 1765-1790.

Loughran, T., Ritter, J., 2000. Uniformly least powerful tests of market efficiency. Journal of Financial Economics 55(3), 361-389.

MacKinlay, A., 1997. Event studies in economics and finance, Journal of Economic Literature, 35(1), 13-39.