Post crisis mergers and acquisitions in the United Kingdom : an event study approach

1

Bachelor thesis

Teun Grijzenhout

Semester 1, 2015-2016

10373721

Bachelor Economie & Bedrijfskunde

Dhr. F.S. Peters

BSc ECB

Universiteit van Amsterdam

Post Crisis Mergers and Acquisitions in the United Kingdom –

An event study approach

Abstract

This paper elaborates on the matter of mergers and acquisitions within the European Union, with the acquiring firm situated in the United Kingdom. The effects of merger and acquisition deals, in

particular the presence of abnormal returns, are examined for the acquiring firm’s shareholders in the period after the financial crisis. The investigated data comes from M&A announcements from 01-01-2011 until 01-01-2016 and is analyzed with a CAR analysis. In addition, an Ordinary Least Squares regression is executed to find out till what extend this abnormal returns are caused by the latest merger wave started at the beginning of 2014. The results are that acquiring shareholders in the United Kingdom experience a positive abnormal return of 1.586% from M&A announcements, significantly different from zero. This positive abnormal return is merely earnt on the event day and the two days thereafter. From the OLS regression it is found that the 2014 merger wave has a positive, but insignificant effect on the abnormal returns in this period. This result may be suffering from omitted variable bias, due to absence of significant explanatory or control variables.

2 Verklaring eigen werk

Hierbij verklaar ik, Teun Grijzenhout, dat ik deze scriptie zelf geschreven heb en dat ik de volledige verantwoordelijkheid op me neem voor de inhoud ervan.

Ik bevestig dat de tekst en het werk dat in deze scriptie gepresenteerd wordt origineel is en dat ik geen gebruik heb gemaakt van andere

bronnen dan die welke in de tekst en in de referenties worden genoemd. De Faculteit Economie en Bedrijfskunde is alleen verantwoordelijk voor de begeleiding tot het inleveren van de scriptie, niet voor de inhoud.

3

Table of contents

1. Introduction

1.1 Problem statement and research question

1.2 Delimitation

2. Literature review

2.1 Summary of M&A history

2.2 Motives for mergers and acquisitions

2.2.1 Neoclassical theories

2.2.2 Behavioral theories

2.3 Summary of M&A history

2.4 Market reactions to M&A announcements

3. Hypothesis

4. Methodology

5. Data

5.1 Data selection

5.2 Data adaptation

6. Results

7. Conclusion

8. References

4 1. Introduction

Nowadays, mergers and acquisitions are common in all kind of markets, for all kinds and sizes of companies. Due to globalization and deregulation today, both domestic and cross border takeovers are of frequent occurrence, representing enormous amounts of reallocated resources. In general, the purpose of this reallocation of resources is to create synergies for the merged companies, which lead to the creation of welfare and positive effects for the participating companies (Kleinert & Klodt, 2002).

According to Martynova and Renneboog (2006) and many others, mergers and acquisitions, hereinafter also referred to as M&A’s, come in waves and until now we have seen six examined in literature, namely the ones from the beginning of the 1900s, the whole 1920s, the end 1950s and 60s, the 1980s, the end of the 1990s and halfway the 2000s. They state that only in the most recent ones, levels of M&A activity in Europe could match the ones in the US. This was caused by the introduction of the Euro, technological inventions, the worldwide globalization process and deregulation and privatization. Rhodes-Kropf, Robinson, and Viswanathan (2004) found that in general, firms experience a higher return from M&A’s during merger waves. Since it is rumored by Codeiros (2014) among others that a seventh merger wave started with the start of 2014, the question rises whether the positive effect on stock returns is visible for this wave.

Next to that, extensive research has been done in the past to investigate whether M&A’s have positive or negative effects on the wealth of the participating shareholders. A study by Jensen (1988) revealed that the gains for selling firm shareholders worldwide in the period between 1977-1986 sum up to an amount of 346 billion dollars (measured in 1986 dollars). He estimates that the

revenues for the buying firm shareholders in the same period are positive, but that they are far lower and harder to estimate. Generally speaking, shareholders of the target firm experience positive short term results, while the returns for the acquiring firm’s shareholders vary and often suffer from share price underperformance in the short run.

1.1 Problem statement and research question

Although this study by Jensen provided clear insights in the effects of mergers and acquisitions at that time, time has gone by and two merger waves and a financial crisis later, the situation and conclusions today might be different than the ones in 1988. In the meantime, a lot of research has been done in this section and an overview of the results can be found below in the literature review. In order to keep the scope of this thesis in proportion, this paper focuses on the short term effects of U.K. firms acquiring both domestic firms as well as cross border firms within the Eurozone, after the

5 latest merger wave and the financial crisis. Therefore, as most important and therefore central question throughout this thesis stands the research question:

Do acquiring firm’s shareholders in theUnited Kingdomexperience abnormal return from mergers and acquisitions in the period after the crisis?

This abnormal return for the shareholders is measured based on movements in the stock prices of the companies. As a follow-up question, this thesis aims at finding what the drivers behind this abnormal return could be, in particular whether the latest merger wave of 2014 has a positive impact on the abnormal returns:

Does a merger or acquisition yield a higher abnormal return for acquiring U.K. firms if this event takes place during a merger wave?

The empirical tests used to answer the first question use the methodology of an event study, whereas an OLS regression on the factors which are possibly explaining the abnormal return is used to answer the second. The main question of interest in this regression is whether the M&A’s during the latest merger wave have a significant effect on the abnormal returns.

1.2 Delimitation

The focus is on the acquiring firm’s shareholders, because, according to Jensen (1988) and several other studies, the results for these group were less convincing than for the target firm’s

shareholders. Goergen and Renneboog (2004) found that when a UK firm is involved in an intra-European M&A transaction, the abnormal returns are almost twice as high as for M&A’s involving both target and bidder from the European continent. This can be explained by the fact that the UK has a more developed market for corporate control. This, and the fact that they found out that in the larger part of M&A deals in the European Union, the United Kingdom is involved, makes the U.K. the country of interest in this paper. Cross border transactions are included only within the European Union, because of the presence of the monetary union and standards for regulation that hold in the participating countries (European Union, 2016). The time frame for the investigation is set a year and a half after the end of the global financial crisis in June 2009 (according to the Bureau of Economic Research (2016)) until January 2015. It starts a year and a half after the end of the financial crisis in order to fully exclude it’s effects in the estimation window. The data is further delimited by following the selection criteria used by Moeller, Schlingemann and Stulz (2005), which include only deals with an initial stake less than 50%, a final stake of only 100% and a minimum deal value of 1 million Great Britain Pound. Besides the merger wave factor that is examined for the second question, control variables used in the OLS regression are retrieved through a screening of the existing literature.

6 2. Literature review

In this chapter of the thesis a concise overview is given, concerning papers and research in the M&A field done in the past. At first, a brief history of M&A’s is given in terms of the different M&A waves in the 20th and 21st century. This is followed by a theoretical background for the different reasons behind mergers and acquisitions, consisting of rational, irrational, internal and external factors. Thereafter the literature concerning reactions to takeover announcements is reviewed and at last an overview of the control variables used in the OLS regression is given.

2.1 A brief history of mergers and acquisitions in the 20th and 21st century

According to Martin Lipton (2006) and many other sources like the information website of KPMG (2016), mergers & acquisitions come in waves, and until now, six waves have been counted from the beginning of the 20th century. Below, all the waves are briefly addressed including their causes and characteristics.

The first wave lasted from 1893 until 1904 and ended due to a U.S. Supreme Court decision to make the antitrust laws applicable to horizontal mergers. These horizontal mergers creating the principal steel, telephone and oil giants of that time were typical for this first wave.

During the second wave (1919-1929), major automobile manufacturers such as Ford and FIAT came up. This wave was featured by a very significant increase in vertical integration and furthermore an increasing consolidation in the industries that were subject of the first wave. The second period ended due to the 1929 crash and following Great Depression.

The third period was one of great expansion and diversification, bringing in a US corporate management obsession with entering new markets. It endured from 1955 until 1970, when the conglomerates experienced a crash in their stock prices which caused this period to an end in the beginning of the 1970s.

The fourth wave endured officially from 1974 to 1989, but it was also referred to as the wave of the 1980s. It’s most remarkable feature was the introduction of the hostile takeover bid by Morgan Stanley on ESB in 1974, which inspired major investment banks to make hostile takeover bids as well on behalf of raiders. Next to that, cross border horizontal mergers entered the European market and the volume and size of LBO trades increased. The period ended with the collapse of the junk bond market and banks’ capital structures due to aggressive lending activities needed for funding. With the fifth wave (1993-2000), the period of the mega-deals began. This era was marked by appetite for larger economies of scale and the shaping of enormous companies and global sweep on

7 the strong idea that size of companies matters. The wave was driven by deregulation and

globalization and contained six of the ten largest M&A deals in history during the last three years (1998-2000). Although in beginning of the year 2000 the 165 billion dollar merger of Time Warner and AOL was settled, shortly thereafter the fifth wave ended due to the collapse of the internet stocks and financing problems of the telecom companies.

Only a small three years after the fifth wave ended, the sixth wave already began with further expansion of globalization and a boom in private equity, as shareholders had appetite for holding ownership of their companies with themselves. This wave lasted from 2003 until 2008, ending suddenly due to the 2008 great recession.

2.2 Motives for acquisition

In this section different ways of reasoning behind M&A’s are discussed. From the management’s point of view, motives for acquisitions must be split in two groups, namely internal and external motives. In this case, external motives for M&A’s are motives that are beyond the control of the management, such as globalization, (de)regulation or developments in technology. Internal motives, on the other hand, can be influenced by management. This internal motives, such as synergies or cost reduction, must be divided into two main categories, namely ‘neoclassical’ theories and

‘behavioral’ theories, as stated by Rhodes-Kropf, Robinson and Viswanathan (2004). They argue that in the distinction between these two categories, the concept of rationality is of importance.

2.2.1 Neoclassical theories

Rhodes-Kropf, Robinson and Viswanathan (2004) elaborate that under the neoclassical theory, it is assumed that the manager of the firm is fully rational. This means that the manager only participates in mergers and acquisitions if these result in an increase in wealth for shareholders of both the acquiring as the target firm. Following Berkovitch and Narayanan (1993), an example of this rationality is the synergy motive, which arises in the situation where the combination of the target and acquiring firm carries a greater value than the two firms apart of each other. They state that such a situation follows from possible economies of scope, economies of scale or increases in market power of the new firm. Furthermore, synergies can arise from cost reduction and improvement in resource reallocation, such as reallocation of capital or human resources. Berkovitch and Narayanan also name rationality in cases of external motives for M&A’s, in means of changes in regulation or technological developments. The rational manager makes his decisions taking these changes in consideration and still does what is in the firm’s best interest. The effective decision in times of

8 technological developments may be to arrange an M&A between the manager’s firm and a firm with a patent on this new feature.

2.2.2 Behavioral theories

According to Jensen and Meckling (1976), behavioral theories for acquisition can be split into two motives, being the agency motive where the manager of the firm is considered rational but not acting in the best interest of the shareholders, and the Hubris motive which considers the manager to be irrational. They state that the first situation arises when neither the manager himself or the other shareholders own a part of the shares, large enough to lead to incentives to act in the firms best interest or monitor the manager’s effort, respectively. In this situation, the agency problem arises due to managers getting more utility from acting in their own interest than in the interest of the shareholders. An example of such given by them is the case in which managers are reluctant to pay out excess cash, because it reduces their resource control. Instead they might arrange an acquisition with the excess cash, which does not maximize firm value. They continue with another agency example, which considers the one of empire building, which arises in case the manager’s compensation scheme is linked to total firm size. This leads to unnecessary acquisitions done to increase firm size, not firm value.

The second category of behavioral theories is the one of Hubris motives for acquisitions, which is described in a 1986 paper by Roll. This Hubris motivation considers managers as irrational, leading to M&A decisions made by mistakes of the manager, for example overpaying for the target firm in a takeover which decreases the possible synergies. Roll states that according to Hubris, the

management does try to maximize firm value, but does not have the capacity of doing so due to irrationality. This absence of good reasoning by the management results in lower or even negative proceeds for the acquiring firm’s shareholders.

2.3 Market reactions to the M&A announcement

In this section empirical papers on market reactions in the past are discussed. The writers of the existing empirical studies are clear in their conclusion about acquisitions creating value for the participating shareholders combined, with most of the revenues accrued to the target firm’s

shareholders, as found by Martynova and Renneboog (2006). There is a large amount of studies that found convincing evidence for positive effects for target shareholders in the UK and Europe, such as the studies by Franks and Harris (1989), Danbolt (2004) and Goergen and Renneboog (2004). These studies found target gain average announcement returns of 24% (1955-1985), 19% (1966-1991) and 13% (1990-2001). Schwert (1996) found evidence that for target firm shareholders, the share price

9 reaction commences already 42 days before the public announcement day. Next to that, from his study it appears that the price run-up is substantial and often exceeds the announcement effect itself. This implies that the announcements are anticipated with rumors, which are immediately reflected in the share price according to the efficient market hypothesis.

Opposed to the clear positive returns for target shareholders, the returns for the acquiring firm’s shareholders are often negligible. Empirical studies found results that acquiring shareholders observe abnormal returns immediately around the announcement day which are not significantly different from zero (Martynova & Renneboog, 2006). The severance between empirical evidence pointing in the direction of small positive and zero or small negative returns is about fifty fifty, and the share price run-up over the one-month period previous to the announcement is positive, but insignificant for acquiring shareholders (Schwert, 1996).

In the paper of Martynova and Renneboog (2006) is further shown that several factors influence the value of the acquiring firm. First, they notice in their paper that returns for bidders on the

announcement day are significantly lower in hostile than in friendly M&A’s. Secondly they state that when acquiring firm’s managers own a larger share of the equity, the share price reactions increase. This means that possible agency problems in the firm are discounted in the share price. At third, it is mentioned that cash offers generate higher returns for both target as acquiring firm, than all-equity offers do. Fourth, they continue naming that bidder firms with excess cash in general overpay for the target. At last, they conclude that the share price of bidders acquiring cross border firms is significantly lower than that of bidders acquiring domestic firms. This is due to the market

anticipating on regulation and cultural differences between countries leading to obstacles in managing the merged company.

2.4 Control Variables

In this section the control variables used in the regression are discussed. At first, a dummy is used for the difference in domestic and cross-border M&A’s. Goergen and Renneboog (2004) state that following the foreign direct investment theories, bidders in cross-border transactions pay higher premiums due to expected cross-border synergies, which lead to higher expected abnormal returns. Rose and Weisbach (2012) comment in their paper on determinants of cross-border M&A’s and mention the target firm value i.e. the deal value and the acquirer firm value prior to the deal.

Furthermore, they mention the influence of equality in industries of the target and acquirer firm. For this reason, the control variables ‘deal value’, ‘acquirer pre-deal market capitalization’ and a dummy for ‘industry equality’ are included in the regression as well.

10 3. Hypothesis

As shown in the literature, in particular in the last section, the expectation for the return for the acquiring firm’s shareholder is ambiguous, with about half the empirical evidence pointing in a negative direction and the other half pointing the other way to a positive or at least zero return. Because of this ambiguity, it is not possible to elaborate a prognosis on the direction of the result, but the expectation is to find the abnormal return to be significantly different from zero. The statistical null hypothesis will be the reversion, which makes the hypothesis below the alternative hypothesis H1:

In a merger and acquisition deal in the United Kingdom in the period after the financial crisis, shareholders of the acquiring firm do experience an abnormal return significantly different from zero.

In order to form a hypothesis in line with the second question concerning the influence of merger waves on the experienced abnormal return for acquirer firms, the literature is reviewed and among others, Rhodes-Kropf, Robinson, and Viswanathan (2004) state that abnormal returns are in general higher during merger waves. Therefore the expectation is that this holds for the 2014 merger wave as well, leading to the second hypothesis:

For merger and acquisition deals in the United Kingdom post-crisis, the merger wave started in 2014 has a positive effect on the abnormal return experienced by the acquiring firm’s shareholders.

4. Methodology

To come to the results needed to accept or reject the composed hypotheses, the event study method is used. This method was first developed by Fama, French, Jensen and Roll in their 1969 paper and is used as a standard method to measure the effects of events, such as M&A’s, in economics and finance since then (MacKinley, 1997).

The aim of this method is to compare the actual returns of a given firm around t = 0, the date on which the event takes place, with the returns considering that the event did not take place. Because of the efficient market hypothesis, that assumes full inclusion of all available information, the event date is the date of the announcement of the merger or acquisition instead of the date of the completion. According to MacKinley (1997), the period around the event, the event window, is set larger than 1 to facilitate the use of abnormal returns around the event day in the analysis. The estimation window, the period of trading days that is used to estimate the normal return of the

11 firm’s stock, is usually set to 200 trading days, which stop 40 trading days before the event to exclude the effect of rumors. For this purpose, my event window lasts from two days before until two days after the event (t = -2 until t = 2), and the estimation window lasts from 240 days until 40 days before the event (t = -240 until t = -40). The corresponding timeline can be found in below.

Estimation window Event Window

As mentioned above, the event study method composes the abnormal return (ARi) by comparing the

actual return over a given period (Ri) with the expected return given the conditioning information

(E(Ri|Xi)). This is reflected in the following formula:

𝐴𝑅_𝑖,𝑡= 𝑅_𝑖,𝑡− 𝐸(𝑅_𝑖,𝑡|𝑋_𝑡)

In general, there are two options for estimating the normal return in this equation, namely the constant mean return model and the market model (MacKinley, 1997). The difference between the two is that the first one assumes Xt to be a constant, whereas the latter assumes a stable linear

relation between the market and security return, making Xt the market return in this model. In this

paper the market model is used, because it yields more powerful results with a smaller variance under the assumptions of the asset returns to be normally distributed for a large number of observations and to be independently and identically draft (Brown & Warner, 1985). The market model equation is composed as follows:

𝑅𝑖,𝑡 = 𝛼𝑖+ 𝛽𝑖∗ 𝑅𝑚𝑘𝑡+ 𝜀𝑖,𝑡

With 𝐸(𝜀𝑖,𝑡) = 0, and 𝑉𝑎𝑟(𝜀𝑖,𝑡) = 𝜎𝜀² Combining these two formulas gives:

𝐴𝑅_𝑖,𝑡= 𝑅_𝑖,𝑡− 𝐸(𝑅𝑖,𝑡|𝑅𝑚𝑘𝑡) = 𝑅𝑖,𝑡− (𝛼𝑖+ 𝛽𝑖∗ 𝑅𝑚𝑘𝑡)

The alpha and beta in this model are estimated with an OLS regression on the basis of the daily returns in the estimation period, followed by a CAR analysis as described by MacKinley (1997) to make a conclusion regarding the abnormal return. This analysis starts with the aggregation of the Abnormal Returns. This aggregation is done through both the time as well as across the different

12 stocks, leading to the draft of inferences for the event of interest. First considering the aggregation through time, assume t1 and t2 as two dates within event window period, with t1<t2. The Cumulative

Abnormal Return CAR is then calculated by: 𝐶𝐴𝑅̂ (𝑡𝑖 1, 𝑡2) = ∑𝑡𝑡=𝑡2 1𝐴𝑅̂𝑖,𝑡. As the length of the estimation window increases, the variance asymptotically tends to 𝜎𝑖2(𝑡1, 𝑡2) = (𝑡2− 𝑡1+ 1)𝜎𝜀². The cumulative abnormal return under H0, no significant abnormal returns, is normally distributed,

with mean zero and variance 𝜎𝑖2(𝑡1, 𝑡2). Using this distribution, the null hypothesis can be tested and accepted or rejected. However, since only one event is taken into account, the test misses statistical power, and therefore aggregation along different M&A’s is added. For this aggregation the

assumption is used that the event windows of the securities do not overlap, i.e. the events and abnormal returns are independent of each other(MacKinley, 1997). The formula 𝐴𝑅𝑖,𝑡= 𝑅𝑖,𝑡− (𝛼𝑖+ 𝛽_𝑖∗ 𝑅_𝑚𝑘𝑡) is used for each event period, which leads to the aggregated abnormal returns for period t. The average of these abnormal returns can then be aggregated over the event window in the same way as for the time, resulting in the CAAR, the Cumulative Average Abnormal Return:

𝐶𝐴𝐴𝑅 = 𝐶𝐴𝑅̅̅̅̅̅̅(𝑡1, 𝑡2) =_𝑁1∑𝑖=1𝑁 𝐶𝐴𝑅̂ (𝑡1𝑖 , 𝑡2) and 𝑉𝑎𝑟(𝐶𝐴𝑅̅̅̅̅̅̅(𝑡1, 𝑡2)) = ∑𝑡𝑡=𝑡2 1𝑣𝑎𝑟(𝐴𝑅̅̅̅̅̅𝑡)

For the computation of the variance again the assumption that the events are independent of each other is used, so that covariance terms can be set equal to zero. The null hypothesis can now be tested using that the average CAR follows the normal distribution with mean zero and variance 𝑉𝑎𝑟(𝐶𝐴𝑅̅̅̅̅̅̅(𝑡₁, 𝑡₂)). This test is conducted by finding the t-values and p-values for the CAAR’s. Since the CAAR is equal to zero according to the null hypothesis, the test statistic is composed like this:

𝑇 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐 = 𝐶𝐴𝐴𝑅 − 0

𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑒𝑟𝑟𝑜𝑟; 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑒𝑟𝑟𝑜𝑟 =

𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 √𝑁

Following the normal distribution 𝑁(0, 𝑉𝑎𝑟(𝐶𝐴𝑅̅̅̅̅̅̅(𝑡1, 𝑡2))). Tests can be conducted at the rejection level of α = 1%, α = 5% and α = 10%. In order to test the second hypothesis, an OLS regression is performed with the statistical software program Stata13. The coefficients found with this analysis are tested in the same way, with the T statistic constructed as shown above.

5. Data

In this chapter the steps taken in the process of the data selection are described. The selection criteria as described in section 1.2 delimitation are used in order to select the data that match the purpose of this thesis. To find the data that match the selection criteria, the process starts with a screening for M&A deals in the database Zephyr. Zephyr is a database by Bureau van Dijk that contains data on worldwide deals of various type running from the beginning of the year 2000, with

13 information of over 550,000 mergers and acquisitions. Since Zephyr does not contain financial data on companies, the ISIN number of the companies is extracted and afterwards further analyzed using the database of DataStream. Next to that, because all the firms examined in this paper are situated in the United Kingdom, the Rmkt is given by the levels of the Financial Times Stock Exchange 100 index, a

share index of the 100 London Stock Exchange listed companies with the highest market capitalization. This is also referred to as the benchmark index.

5.1 Data selection

In Zephyr the following selection criteria are used, leading to a result of 717 M&A deals for acquiring U.K. firms (see figure 2):

Deal type: Acquisition, Merger; Time period: announced on and after 01/01/2011 and up to and including 01/01/2015; Country: United Kingdom (GB) (Acquiror); World Regions: European Union (Target); Listed/Unlisted/Delisted companies: listed acquirer; Percentage of stake: Percentage of initial stake (max: 0%); Percentage of final stake (min: 100%); Deal value (mil GBP): min=1 (including estimates).

The criterion for listed acquirer companies is included since unlisted companies are excluded from DataStream. After that, the ISIN numbers of the acquirer companies are added and the deals with an acquirer without this number, with a country code other than GB (Great-Britain) and acquirer firms for which the pre-deal market capitalization is not announced, are deleted (83 deals deleted). Next, the request is extracted from Zephyr to an excel file, including the announcement dates, ISIN numbers for the acquiring companies, company country codes, pre-deal market capitalization of the acquirer, deal value and Zephus industry codes for both acquirer and target . The DataStream option P#T for datatype is added, so that there is no data in excel after a stock stops trading. Excluding this option makes DataStream repeat the last known value for all the missing values, which biases the outcome.

5.2 Data adaptation

After the request to DataStream has been made with the event dates and the company ISIN numbers, the table is processed resulting in stock prices and market indices for the estimation and event window. After removing events for which stock prices are not announced, 842 M&A events remain for further examination. For these events, both stock as well as daily market returns are calculated and converted into the mean adjusted returns, i.e. stock returns in the event window reduced by the average return in the estimation period. Eventually the market model as described in

14 the methodology section is estimated and combined with the actual stock returns during the event window, the abnormal returns for these five days can be estimated for each of the 842 events. In order to perform the second test, dummies are added in excel for an event to take place during the latest merger wave (from the start of 2014), for an M&A to be domestic or cross-border and for the takeover to take place within an industry. In this case the merger wave dummy is equal to 1 for events after 1-1-2014 and 0 otherwise, the dummy for domestic/border is equal to 1 for cross-border M&A’s and 0 otherwise and the industry dummy equals 1 for takeovers within the same industry and 0 otherwise. Furthermore the data for deal value and for pre-deal acquirer market capitalization are added. Both the deal value as well as the pre-deal acquirer market capitalization values are transformed into their logarithm forms. This is done because differences in these values are that large that percentage differences must be used for the interpretation (Stock & Watson, 2012).Events for which the Zephus industry codes for target or acquirer firms are not announced, are removed, leaving 837 events (5 removed). Together with the CARs calculated in the first test, the data is imported and tested in Stata13. The results can be found in chapter 6.

6. Results

In this chapter the results of the CAR analysis and OLS regression are presented and used to test the hypotheses. The tests are conducted as described in the methodology section and the results are interpreted with respect to the empirical evidence found in the literature review. For this tests, the hypotheses composed in section 3 are reversed, so that the null hypothesis says ‘no results’. The statistical hypotheses are as follows:

H0: No abnormal returns  CAAR = 0 H1: Abnormal returns  CAAR ≠ 0

H0: No influence of merger waves  β1 = 0 H1: Influence of merger waves  β1 > 0

The first alternative hypothesis is set ‘unequal to zero’, covering the scenarios of positive and

negative abnormal return in the first test. For the second test, the alternative hypothesis is set ‘larger than zero’, because according to the literature it is expected to find a positive effect for the ‘during a wave’ factor. In this hypothesis, β1 is the coefficient corresponding with the slope of the ‘during a wave’ factor. The alternative hypothesis states that this slope is positive. Including the control variables, the OLS equation is as follows:

𝐶𝐴𝑅(−2,2) = β₀+ β₁∗ Wave + β₂∗ CB + β₃∗ ln(Deal) + β₄∗ ln(MktCap) + β₅∗ Industry In this equation, the betas are coefficients for the constant, for the event to take place during the merger wave started in 2014 (Wave), the dummy for cross-border mergers (CB), the logarithm of the deal value (ln(Deal)), the logarithm of the pre-deal market capitalization of the acquirer (ln(MktCap)

15 and the dummy for equality in industries (Industry), respectively. The results for the first test can be found in the table below:

ACQUIRER CAAR

From event

day AAR Std. dev Std. error t-value p-value

-2 -0.006% 0.0242 0.00083 -0.073 0.941 -1 0.028% 0.0212 0.00073 0.376 0.706 0 0.751%*** 0.0569 0.00196 3.824*** 0.000 1 0.605%*** 0.0491 0.00169 3.567*** 0.000 2 0.209%** 0.0262 0.00090 2.312** 0.021 CAAR(-2,2) 1.586%*** 0.0897 0.00309 5.126*** 0.000 statistical significance at *10%, **5%, ***1%

From this table it can be concluded that the Cumulative Average Abnormal Return for acquiring firms in the U.K. is equal to 1.586% for the five day period around the announcement of a merger or acquisition. The corresponding t-value of 5.126 is significant at the 1% significance level, which means that this result is very significant as can be seen as well from the p-value which is smaller than 0.001. Furthermore, the conclusion can be draft that acquiring shareholders experience a negative but insignificant return two days before the announcement and a negligible positive return one day in advance. However, the results of 0.751% on the event day, 0.605% on the day after and 0.209% two days after the event are significant at the 1%, 1% and 5% level respectively. From this

insignificant results before the event day and significant results from the event day onwards it can be concluded that rumors and other factors influencing the outcome are successfully excluded. The overall conclusion is that the null hypothesis is rejected and that the alternative hypothesis, which is the main hypothesis of this thesis, is accepted.

In order to test the second hypothesis, an Ordinary Least Squares regression is conducted in Stata. The CAR for every event is set as the dependent variable and regressed on the dummy for the event to take place during a wave and the control variables. The option ‘robust’ is used in Stata, because of assumed heteroscedasticity of the error term. The result of this linear regression is as follows:

16

Dependent variable: Acquiring firm CAR

Ordinary Least Squares

(1) (2) (3) (4) (5) (6) Constant 0.01390 (0.001)** 0.01593 (0.000)*** 0.01028 (0.607) 0.03766 (0.305) 0.02084 (0.011) 0.03077 (0.404) During wave 0.00301 (0.629) 0.00098 (0.871) Cross-border -0.00421 (0.519) -0.00316 (0.608) Log deal value 0.00052 (0.786) 0.00273 (0.241) Log pre-deal mkt. Cap. (acquirer) -0.00178 (0.517) -0.00288 (0.381) Industry -0.00710 (0.419) -0.00731 (0.408) Obs 837 837 837 837 837 837 R-Squared 0.0003 0.0003 0.0001 0.0018 0.0011 0.0047 statistical significance at *10%, **5%, ***1%

The numbers in brackets are the p-values corresponding with the coefficients. From this table it can be seen that if the control variables are included, the ‘during a wave’ dummy has a positive effect of 0.00098. The interpretation of this number that if a M&A event took place during the 2014 merger wave, the Cumulative Abnormal Return is 0.098%-point higher than for M&A events in the sample before 2014. However, from the corresponding p-value it can be seen that this effect is not

significant. In case the control variables are excluded, it is estimated that an event during the latest wave has an 0.3%-point higher abnormal return than events in the period before. This estimation is insignificant as well. For this reason, the second null hypothesis is not rejected and therefore we can not conclude that the second hypothesis composed in chapter 3 is true. Additionally, the summary statistics for the variables included in the OLS regression can be found in the table below:

Summary Statistics

Variable Obs. Mean Std. Dev. Min. Max.

CAR 837 0.015 0.0895 -0.586 1.103

During Wave 837 0.481 0.5000 0 1

Cross-border 837 0.140 0.3470 0 1

Deal Value 837 189,645 2,335,988 1,193.84 6.44e+07 Pre-Deal Mkt. Cap. 837 2,139,740 7,762,989 290.89 1.32e+08

Industry 837 0.774 0.4183 0 1

From this table, it can be seen that from the total observations of 837 used for the second test, 403 events took place during the latest merger wave, 117 were cross-border M&A’s versus 720 domestic and 648 of the events took place between two firms equal in operating industries. Means, standard deviations, minimums and maximums can be distracted from the table.

17 7. Conclusion

From the results presented in section 6, it can be concluded that the first hypothesis is confirmed, but the second is not. After conducting the CAR analysis on the selected data, a positive CAAR value of 1.586% came out, which is significant at the 1% rejection level. Furthermore it is shown that the AAR values for the two days before the announcement were insignificant, which suggests that there were no or few rumor effects. The event day, the day after and the day thereafter had positive AARs, significant at the 1%, 1% and 5% rejection level respectively, which endorsed the expectation that abnormal returns measured during the event day are caused by the event itself. This leads to rejection of the null hypothesis and acceptance of the alternative hypothesis and hypothesis draft in chapter 3. Regarding the second the second test, the conclusion can be draft that if an event takes place during the merger wave started in 2014, the cumulative abnormal return experienced from this event is 0.098%-point higher than from events before this period. However, this effect is

insignificant at the 10% rejection level and therefore the null hypothesis is not rejected. It follows that the hypothesis composed in chapter 3 can not be confirmed.

A suggestion for further research or improvements of this thesis is to broaden the scope of the events in terms of time. This may help to improve the significance of the results. The same holds for including only M&A’s with a target in the European Union, which can be extended to inclusion of the rest of the world. This increased number of observations could improve the value of the results, in particular those of the second test. Next to that, the quality of this test could be improved by including more control variables in the OLS regression. The small R-squared of the regression qualifies the factors as weak in their ability to estimate the CARs. This may be due to omitted variables, causing omitted variable bias. Variables that could be included in further research are for example dummies for hostile takeovers and method of payment.

18 8. References

Berkovitch, Elazar and Narayanan, M.P. (1993) Motives for Takeovers: An Empirical Investigation. The

Journal of Financial and Quantitative Analysis. Vol. 28, No. 3, Retrieved

http://www.jstor.org/stable/2331418?seq=1#page_scan_tab_contents

Brown, Stephen J. and Warner, Jerold B. (1985) Using Daily Stock Returns, The Case of Event Studies.

Journal of Financial Economics. No. 14, pp. 3-31 Retrieved

http://www.simon.rochester.edu/fac/warner/Jerry%20Papers/JFE-March%2085.pdf Cordeiro, Marcos (2014) The seventh merger wave. Camaya Partners.

Danbolt, J. (2004) Target company cross-border effects in acquisitions into the UK. European Financial Management 10, 83-108

European Union (2016) Website, visited 11-1-2016

http://europa.eu/about-eu/basic-information/money/euro/index_nl.htm

Franks, J. and Harris, R. (1989) Shareholder wealth effects of corporate takeovers: the U.K. experience 1955-1985. Journal of Financial Economics 29, 81-96

Goergen, M. and Renneboog, L. (2004) Shareholder wealth effects of European Domestic and Cross Border Takeover Bids. European Financial Management 10. 9-45

Jensen, Michael C. (1988) Takeovers: Their Causes and Consequences. Journal of Economic

Perspectives. Vol. 2, No. 1, Retrieved http://ssrn.com/abstract=173455

Jensen, Michael C., and Meckling, William H. (1976) Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics. Vol. 3 pp. 305-360. Retrieved http://www.sciencedirect.com/science/article/pii/0304405X7690026X

Kleinert, Jörn and Klodt, Henning (2002) Causes and Consequences of Merger Waves. Kiel Working

Paper No. 1092, Retrieved www.ifw-kiel.de KPMG (2016) Website, visited 14-1-2016

http://www.kpmg.com/za/en/issuesandinsights/articlespublications/transactions-restructuring/pages/seventh-wave-of-ma.aspx

Lipston, Martin (2006) Merger Waves in the 19th, 20th and 21st centuries. Retrieved http://cornerstone-business.com/MergerWavesTorontoLipton.pdf

MacKinlay, A. Craig (1997) Event Studies in Economics and Finance. Journal of Economic Literature.

Vol. 35, No. 1, pp 13-39 Retrieved http://www.jstor.org/stable/2729691

Martynova, Marina and Renneboog, Luc. (2006) Mergers and Acquisitions in Europe. Advances in

Corporate Finance and Asset Pricing. Retrieved via www.ssrn.com

Moeller, Sara B., Schlingemann, Frederik P. and Stulz, René M. (2005) Wealth Destruction on a Massive Scale? A Study of Acquiring-Firm Returns in the Recent Merger Wave. The Journal of

19 Finance. Vol. LX, No. 2, pp. 757-782 Retrieved

http://fisher.osu.edu/fin/faculty/stulz/publishedpapers/Wealth%20Destruction_JF.pdf National Bureau of Economic Research (2016) Website, visited 12-1-2016

http://www.nber.org/cycles.html

Rhodes-Kropf, Matthew, Robinson, David T. and Viswanathan, S. (2004) Valuation Waves and Merger Activity: The Empirical Evidence. Journal of Financial Economics. Retrieved

http://poseidon01.ssrn.com/

Roll, Richard (1986) The Hubris Hypothesis of Corporate Takeovers. The Journal of Business. Vol. 59, No. 2, part 1, pp 197-216. Retrieved

http://pendientedemigracion.ucm.es/info/jmas/doctor/roll.pdf

Rose C. Liao, Isil Erel and Weisbach, Michael S. (2012) Determinants of Cross-Border Mergers and Acquisitions. Journal of Finance. Vol. LXVII, No. 3 Retrieved

http://u.osu.edu/weisbach.2/files/2015/01/ELWJFfinalpdf-16sy82h.pdf

Schwert, G.W. (1996) Markup Pricing in Mergers and Acquisitions. Journal of Financial Economics 41. 153-162

Stock, James H. and Watson, Mark M. (2012) Introduction to Econometrics. Pearson Education Limited. Chapter 8.2, p. 307