Long-run post-merger stock price performance : do mergers increase the stock’s performance?

(1)

Faculty of Economics and Business

Long-run post-merger stock price performance

Do mergers increase the stock’s performance?

29 June, 2016

Abstract:

This paper measures the long-run post-merger performance of the acquiring firm based on its stock prices. Performance is measured as the cumulative average abnormal return. Additionally, the results found are compared with the pre-merger cumulative average abnormal return. The tests are run using a sample of 26 takeovers over the time period of 2005 to 2010. The sample solely includes publicly listed firms from Germany. Thereby, this paper intentionally uses an unused sample. This is assumed to be needed because previous literature primarily focuses on the NYSE. The main findings of this paper are as follows: the total period cumulative average abnormal returns are negative the first four years, but equal out in the fifth year. Furthermore, the sub-periods cumulative average abnormal returns increase yearly. The examination of the pre- and post-merger cumulative average abnormal returns shows that the firm worsens its abnormal performance compared to its pre-merger self. This finding may however be due to measurement errors. Future research is needed to confirm the findings.

Bachelor Thesis Economics and Business Specialization: Economics and Finance

Name: Stijn Smit

Student Number: 10559515 Thesis Supervisor: Dr. Ilko Naaborg

(2)

2

Statement of Originality

This document is written by Stijn Smit who declares to take full responsibility for the contents of this document.

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

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

(3)

3

Street Journal, 2015). A similar claim is made in Europe, where in the Netherlands the number of mergers in the period 2011 to 2015 has consistently been double the number of mergers in the years 2007-2010 (Centraal Bureau voor de Statistiek, 2016). However, for Europe as a whole this phenomenon is less extreme. In Europe the number of takeovers does not set a new high, as shown in figure 1.1, but the trend line shows that there is an upward trend in Europe. This trend may be due to the currently low interest rates, as set by the central bank. With discussions about further lowering the interest rates to a negative amount, another increase in the number of mergers may occur. This large number of mergers requires an in debt knowledge about its potential for success.

Figure 1.1

The number of takeover transactions in Europe.

[Data source: www.imaa-institute.org]

The success of USA-based mergers has been thoroughly researched with papers like Agrawal, Jaffe, & Mandelker (1992), Dodd (1980) and Healy, Palepu, & Ruback (1992). However, surprisingly few studies focus on Europe. Therefore, this paper will examine merger performance in Europe, and more specifically in Germany. Germany is often described as the engine of Europe and is therefore interesting to examine.

0 5000 10000 15000 20000 1980 1985 1990 1995 2000 2005 2010 2015 2020 N um be r of T ake ove rs

(5)

5 Germany is currently having a similar M&A-boom as the USA has. This ‘boom’ may be due to their relative stability, in an otherwise volatile area. Germany currently has a booming

Mittelstand, which are small to medium sized businesses. This provides interesting investment

opportunities, and is also a possible explanation of the M&A increase in Germany (Royston-Bailey, 2015).

This recent rise makes it interesting to reexamine the long-run stock price performance of merging firms. Especially the performance of the acquiring firm is interesting, because there has been an ongoing debate on the matter. The exact methodology has long been a discussion, causing results ranging from statistically significant positive returns to significant negative returns. This debate is discussed in section 2.

Overall, previous literature concludes a negative abnormal return for the acquiring firm. If the merger worsens their stock price performance in the long run, then why would the (stockholdervalue maximizing) firms want to merge? This question remains unanswered and is described as a “puzzle”. The negative returns should be a reason for acquiring companies, or at least the shareholders of the company, to rethink their strategy and perhaps completely lose the idea that a merger could increase their performance. This thesis will reexamine the puzzle to test whether it is still present in the market.

Previous literature has made measurement errors and, perhaps unknowingly, used largely overlapping datasets. The measurement errors were due to inexperience in the field of long-run performance measurement, which was under debate. However, today there seems to have been found a standard and correct methodology. This methodology is however underused; the field has since the new methodology rarely researched long-run performance. Therefore, this thesis will reexamine the methodology and use it on a more current dataset. This way it tests if the anomaly of negative post-merger returns is still in the market.

This paper will use a dataset from publicly listed firms in Germany over the period 2005 to 2010. This period is chosen to have sufficient post-merger acquisition data (which is 5 years), whilst focusing the research on firms which merged recently. Germany has not been used in previous literature and therefore has no overlap with the datasets of prior literature. The research with an unused dataset may contribute to existing literature by providing evidence for i.e. European firms. Prior research primarily researches the New York Stock Exchange (NYSE), all with similar data ranges, and only a few researchers examine the UK-market. This makes

(6)

6 their results similar, which is perhaps a good thing when debating the methodology, as it enables to easily compare results. This thesis does however not join the debate on the correct methodology, but assumes the methodology as used by Agrawal, Jaffe, and Mandelker (1992) as correct. This paper exists to examine the anomaly in a different country in order to test whether the long-run post merger performance puzzle exists outside of the USA and UK.

The foremost question under discussion in this thesis is whether the long-run stock price performance of the acquiring firm is positive or negative. To test this the market-model returns will be compared to the actual returns, where the difference is described as the abnormal return. A more detailed explanion is discussed in section 3.

This paper is organized as follows: chapter 2 is a discussion of previous literature, in which the debate on the measurement method is discussed and a summary of the most important papers is given. Then, in chapter 3, the used methodology will be explained and a detailed description of the used dataset is presented. After which the results are presented and discussed in chapter 4. Lastly, a quick summary together with an overall conclusion will be given in chapter 5.

2 Literature Review

In this section the most important literature will be discussed. First some definitions are discussed, after which general theories concerning mergers and acquisitions are explained. The theories discussed contain information about the methodology used in prior research, together with a discussion of previous methodological errors. Then the results of the leading papers will be examined more in detail in order to lay the foundation this thesis. Information on prior results are presented and discussed. Finally, a general conclusion on the literature is made. 2.1 Theories regarding post-merger performance

2.1.1 Definition

A merger or an acquisition requires two parties; an acquiring-firm and a target-firm. An acquisition is an outright purchase of the firm; in the case of an acquisition the target-firm legally seizes to exist. A merger is usually described as more friendly and is more broadly defined as the agreement between the two parties to continue as a single firm. For simplicity either of the mechanisms will be referred to as a takeover.

(7)

7 However, a distinction is made between the announcement date and the effective date. The announcement date is the date on which the merger plan is announced publicly. For example, in prior papers like Agrawal, Jaffe, & Mandelker (1992) the announcement date is chosen to equal the date on which the Wall Street Journal mentions the bid. The effective date may occur years after the announcement date, and is described as the date on which the merger takes effect or is complete. A paper which studied the post-merger returns is Mandelker (1974). Both these papers will be more thoroughly discussed in section 2.1.3.

Lastly, the definition of a tender offer is discussed. As discussed by Jensen and Ruback (1983), a tender offer is the situation where a bidding-firm offers to buy the shares of the target-firm’s shareholders. The shareholders decide individually whether they want to tender their shares. The difference between a tender offer and a merger is that a tender offer can bypass the managers of the target-firm, because the shareholders decide individually. In a merger it is common to first discuss the merging plans with the managers of the bidding-firm, before asking approval of the board of directors, after which the shareholders are asked for approval. Tender offers are therefore considered as more hostile. In situations where the target-firm’s management disapproves of the tender offers, they may participate in a so called defensive merger (Dodd, 1980). In a defensive merger the target-firm merges with a third party to increase its size, hoping to scare off the bidding-firm.

2.1.2 Reasons for takeovers

The main reason for a takeover is the belief of the bidding-firm that the two firms combined will perform better together than they will individually, this is known as ‘synergy’. Synergies may arise from numerous situations; Jensen and Ruback (1983) name economies of scale, vertical integration, adaption of more efficient production or organizational technology, the reduction of agency cost and the increase of the monopolistic power of the bidding-firm as potential sources of synergy. For example, the increase of monopolistic power allows the bidding-firm to either increase its prices or increase its output; ceteris paribus resulting in a higher equity value. Overall, synergies lead to a more efficient organization.

However, synergy potentials are not the sole reason for takeovers and not every takeover shares the benefits synergy has. Mandelker (1974) mentions that managers tend to pursue size maximization, and with it the maximization of their salaries. Brouthers, van Hastenburg and van den Ven (1998) believe that managers tend to be overly optimistic about

(8)

8 merger decisions. Another reason was discussed in section 2.1.1, namely the defensive merger. A type of merger where a target merges with another firm to increase its size and scare of the bidding-firm. These reasons are often at the expense of shareholder value.

Mergers may also produce less returns to the merged firm than expected due to a mismatch between the two corporate cultures or an overvaluation of the expected merger gains (Franks, Harris, & Titman, 1991). If not all mergers tend to increase firm performance the effects on the stock price cannot be unambiguous. This shows in the different results found in prior literature, which will be discussed in section 2.1.3 and 2.2.

2.1.3 Performance measurement, a short historical perspective

Beginning with Mandelker (1974), who is considered to be the first to examine the financial consequences of a merger in a modern fashion. Mandelker estimates the error term (which is now known as the abnormal return), using the Capital Asset Pricing Model (CAPM). Then he calculates the average of the cumulative abnormal return (CAAR). His method is later used and/or mildly adjusted in later literature, but it remains the standard for the majority of research in long-run merger performance.

For example, Langetieg (1978) uses a similar model but compares the results found to an adjusted version of the model. The adjusted versions contain at least two factors, in contrast with the one factor model used by Mandelker (1974). The factor he uses is a control firm, which should be a clearer comparison. Although they differ in their use of models, Langetieg (1978) uses stock price data for his measurement of merger performance (as well as Mandelker, 1974). This is an example of the belief that merged firms which are characterized by the objective of stockholder wealth maximization, should be examined based on their stock price performance (Langetieg, 1978).

However, according to Dimson and Marsh (1986), these models are flawed. They argue that at least the size factor should be added to the benchmark model, i.e. the CAPM. Their findings show that the factor becomes especially important when examining the long-term. The variable also seems to be important if the sample differs systematically in size. And most notably, Dimson and Marsh (1986) claim that a size factor should always be added when using the CAPM methodology, because the bias in CAPM-based abnormal returns is proportional to the magnitude of the small-firm premium.

(9)

9 The above mentioned studies of Mandelker (1974) and Langetieg (1978) are known as event studies. In these studies, the abnormal returns around the announcement date are examined. This form of study is the leading form of performance measurement and will also be used in this paper. Details are discussed in the methodology section.

Healy, Palepu & Ruback (1992) use another method, known as an accounting study. They study the post-merger cash flow performance instead of the stock prices. By doing so they examine the economic gains of the takeover rather than the stock holder gains, ignoring the claim made by Langetieg (1978). Accounting studies in general are not limited to cash flows. Any study which bases its research on reported financial results (i.e. accounting statements) is classified as an accounting study (Bruner, 2002). This method of study will not be used in this paper, because of the claim of Langetieg (1978) (cf. Lubatkin, 1983).

Two other methods of measurement are mentioned by Bruner (2002). Namely, surveys of executives and clinical studies. A survey consists of asking the managers whether the takeover created value according to them. The questions in the survey are usually standardized across the sample. A clinical study is an in-depth examination of a specific transaction. According to Bruner (2002), these studies often find new insights for the field in general by going into detail of a specific takeover. Both methods will not be further discussed in this paper, due to their lack of scientific value or their irrelevance for this paper.

Within these different methods of performance measurement another distinction must be made. Namely, the difference between studies using either the announcement date or the effective date. Mandelker (1974) and Langetieg (1978) for example use the effective date for their studies. The expected price effects will however occur before or on the announcement date. Using the effective date worsens the study’s chance of attributing the changes in stock prices to the takeover itself because there is no standard amount of days between the effective date and the announcement date (Dodd & Ruback, 1977; cf. Jensen & Ruback, 1983).

The last paper that will be discussed is a more recent one by Abhyankar, Ho, & Zhao (2005). They incorporate all noted debate in their paper and try a different method of research than done by previous event studies. Namely, they create two portfolios consisting where one of the portfolios consists of the acquiring-firm, while the other contains the selected firm with a similar size and book-to-market ratio, known as a control firm. Then the Buy-and-Hold Returns (BHR) distribution of the two portfolios are compared using the first two order of

(10)

10 stochastic dominance. Abhyankar et al. (2005) argue that the standard method of event study contains measurement and econometric problems. Therefore, they present their stochastic approach as an alternative. Their method is a more econometric way of examining potential takeover gains and exceeds the intentions of this paper. However, their results will be discussed in section 2.2 as knowledge of a different method may be helpful. For more discussion on the standard event studies see Lyon, Barber, & Tsai (1999).

2.2 Empirical Findings in the literature

The findings in the literature are as diverse as the methods of research. In this section the sample designs and the general results will be discussed.

Langetieg (1978) studies a sample of 149 mergers from a data file of 465 mergers between NYSE-listed companies. Most importantly, he deletes firms who merge more than once in a period of three years. This is a relatively mild way for data selection, as the firms should have been deleted if they occur more than once in the sample period (Lubatkin, 1983). Deletion of these firms is necessary because every additional takeover creates abnormal stock price events which cannot be contributed to the ‘original’ merger, thus a bias in the results would arise. Langetieg (1978) collected stock data from the 12 years surrounding the effective data. His general conclusion is that the four models he used all result in essentially the same answer. He finds a significant positive abnormal return for the acquired firm in the period (-72, -7), but the consolidated firm shows abnormal returns between −4,8% and −13,92% depending on the model.

Healy, Palepu, & Ruback (1992) study the 50 largest acquisitions during the period 1979 and mid-1984 in the US. They argue that their relatively small sample is sufficient because the sample represents a significant portion of the entire population based on the total dollar value. Also they explain that by choosing such a sample they reduce the probability that their sample firms are part in a similarly sized merger during and before their research period. However, by doing so they ignore the advice of Langetieg (1978) and Lubatkin (1983) and their results may potentially be biased. Their research further differs with that of Langetieg (1978) in the sense that they study the abnormal cash flow returns, as discussed in section 2.1. The main finding of their research is the significant improvement in operating cash flow

(11)

post-11 merger. Their findings of a positive effect are strongest in mergers between overlapping businesses.

Agrawal et al. (1992) find a 10% (𝑡 = −2,37) stockholder wealth loss in the following five years after the merger. They conclude this after examining a sample of 937 mergers and 227 tender offers between NYSE bidders and NYSE/AMEX-listed targets over the period 1955-1987. Their sample covers nearly the entire population of takeovers during the mentioned period. However, they fail to remove firms which merge more than once. Thereby they ignore the stock price effect of a set of mergers, which most likely causes a spike in the returns, resulting in a bias (Lubatkin, 1983). In their research they do adhere to the firm size problem discussed in Dimson & Marsh (1986) and they adjust their models accordingly.

Franks, Harris, & Titman (1991) examine a sample consisting of 399 takeovers during the period 1975-1984; again on NYSE- and AMEX-listed firms, with the restriction that the announcement date is available in the Wall Street Journal Index. Roll (1978) claims that the estimates of abnormal performance are highly sensitive to the choice of benchmark. To combat this critique Franks et al. (1991) test their sample with five different benchmark models, and include the results for each model. One of their findings is a difference between the abnormal returns based on how the takeover was financed. Overall, cash financing seems to outperform other methods of financing (see Franks, Harris, & Titman, 1991, p. 92). This is explained by the fact that tender offers typically outperform normal takeovers and tender offers are usually cash financed (Agrawal, Jaffe, & Mandelker, 1992).

Similar findings are presented by Dodd (1980) who finds returns of −1,09% surrounding the announcement date and returns of −7,22% from the announcement until the stockholder acceptance. These effects seem irrespective of whether the merger is accepted or canceled. Dodd bases these results on a sample from 1970 to 1977 consisting of 151 merger proposals, of which 71 are completed and 80 are canceled. All mergers involve a NYSE-listed firm. He designs the sample to exclude tender offers and defensive mergers. His research is focused on the short-term returns, unlike the above mentioned studies. Therefore, his exclusion of the size effect in his model may not cause a bias (Dimson & Marsh, 1986). His use of the CAPM and his calculations of the abnormal returns, which he defines as the difference between the realized returns and the predicted returns, seem just.

(12)

12 Lastly the paper of Abhyankar, Ho, & Zhao (2005) will be discussed. They create a new and more mathematical methodology of researching long-run merger performance. In their models they make use of stochastic dominance relations to compare return distributions. They focus on the UK merger market by examining a sample of 305 successful mergers completed in 1985-2000. Most notably they exclude financial and utility firms and examine only the mergers with a deal value larger than one million US dollars. Their main result is that there is no stochastic dominance of neither the benchmark portfolio nor the merger portfolio, which contradicts previous researchers who find negative returns. Secondly they conclude that there is stochastic dominance of the benchmark portfolio with respect to the merger portfolios that paid the highest premium. Therefore, they conclude that overpayment may be the reason for long-run underperformance of bidding-firms. Lastly, they find that the merger portfolio which were cash financed do outperform the benchmark portfolio. Whether the differences between their findings and those of previous literature are due to statistical errors are left for further research.

2.3 Conclusion on the literature

As seen in section 2.1 and 2.2 researchers have continuously adjusted the models used. First Mandelker (1974) introduces the market model approach. Later, his research was followed by Langetieg (1978), who uses a three-factor model rather than a one-factor model. Langetieg reports significantly negative abnormal returns together with studies like Dodd (1980) and Agrawal et al. (1992). Similarly, the review paper by Agrawal and Jaffe (2000) shows that the overall consensus is that long-run stock price performance is negative.

However, there are literature reports ranging from positive abnormal returns (i.e. Healy et al, 1992) to returns insignificantly different from zero (i.e. Mandelker, 1974). These differences are most likely due to differences in methodology, as the used samples are largely overlapping.

Overall the literature predicts negative abnormal stock returns when the correct methodology is used. Namely the critique from Dimson et al. (1986) on performance measurement is important in this field of research, together with the selection criteria of Lubatkin (1983). Agrawal et al. (1992) is an example of a paper which uses the correct methodology. They are able to use the critique of Dimson et al. (1986) whilst still separating

(13)

13 the mergers from the tender offers. Their methodology will therefore be largely replicated in this thesis.

The research done in this paper is using the model of Agrawal et al. (1992) and is thereby adhering to the critiques as presented by Dimson et al. (1986). The model itself will be discussed in section 3. Although there may be similarities, this thesis will differ from the paper of Agrawal et al. (1992) by examining a different country, namely Germany. Besides a different region, this thesis will use more recent data. It is important to use more recent data because it essentially checks whether the anomaly of negative abnormal returns is still in the market. Assuming that managers have followed the discussion on long-run merger performance, it is expected that the more recent mergers are better arranged and chosen, which may lead to positive abnormal returns. The recent rise in mergers (see Figure 1.1) does make this expectation less likely. Although recent data is important, the most important difference between this thesis and prior studies will be the country difference. As most of previous literature use overlapping samples, they are therefore likely to produce similar results. The anomaly needs testing on a different sample.

Summarizing, most studies find negative post-merger abnormal returns, which is described as anomaly in the market or a puzzle. Due to a difference in methodologies, not all literature finds the anomaly, they may even conclude a positive return. The correct methodology should incorporate a firm size factor in the models, especially when examining long-term returns (Dimson & Marsh, 1986). It is also important to differentiate between takeovers with- and without a tender offer, but this is largely explained due to the method of financing. There have been studies which incorporate these factors in their research, i.e. Agrawal et al. (1992), but the samples still largely overlap with prior research, making it likely to find similar results. This thesis will therefore examine the anomaly with an unused sample.

3 Methodology and Data

3.1 Methodology

As described in section 2, the overall consensus in performance measurement is that the performance should be measured with abnormal returns. The abnormal returns are the excess realized returns in comparison with the, by a model, predicted returns. For example, if the

(14)

14 CAPM predicts a return of 6% and the realized return was 7%, the abnormal return is equal to 1%.

However, the CAPM is described by Dimson & Marsh (1986) to be a poor measurement of predicted returns. CAPM especially underperforms in long-period examinations. Therefore, they suggest to add a size factor to the model. Agrawal et al. (1992) follow their suggestion and use a similar model as the following to estimate the benchmark portfolio return for firm 𝑖: 𝑅₀₁ − 𝑅₂₁ = 𝛽₀₄ 𝑅₄₁ − 𝑅₂₁ + 𝛽₀₆ 𝑅₆₁− 𝑅₂₁ + 𝜀₀₁ (1) 𝑅₀₁ = the return of security 𝑖 over month 𝑡.

𝑅₆₁ = the return in moth 𝑡 of a control firm, which is selected based on the market value of equity in the nearest year from 𝑡₈.

𝑅₂₁ = the risk free rate over month 𝑡, given by the one-month German treasury bill rate. 𝛽04 = the OLS-regression market beta for firm 𝑖

𝛽₀₆ = the OLS-regression size beta for firm 𝑖

𝜀₀₁ = the abnormal return generated by firm 𝑖 over month 𝑡.

The only adjustment this thesis makes to this model is a change to the definition of 𝑅61.

Where Agrawal et al. (1992) use a control portfolio consisting of all the NYSE-listed firms within the same size decile as firm 𝑖 (where size is based on the market value) to estimate the returns. This thesis uses a single control firm for each of the sample firms, where the control firm is also selected based on the market value. The main reason for this redefinition is a difference in sample selection. Unlike Agrawal et al. (1992), the sample firms in this thesis are from more than one index, which increases the difficulty to adhere to the exact methodology. This is further discussed in the section on the data (3.2).

The calculation of the betas follows the methodology of Agrawal et al. (1992) more closely. First an OLS-regression is run using monthly data from 𝑡_9: to 𝑡_9;8, thus excluding the month of the announcement period. Similar to papers like Langetieg (1978) and Agrawal et al. (1992), the resulting betas are assumed to be constant over the estimated time period. Lewellen & Nagel (2006) argue that this assumption is too strong. However, this thesis mainly focusses on the paper of Agrawal et al. (1992). To more easily compare results, their methodology is followed and the assumption is believed to be true.

The abnormal return is defined as the difference between the realized returns and the estimated returns, which is equal to the residuals.

(15)

15 𝜀₀₁ = (𝑅₀₁− 𝑅₂₁) − 𝛽₀₄ 𝑅₄₁ − 𝑅₂₁ − 𝛽₀₆ 𝑅₆₁− 𝑅₂₁ (2) Then the average abnormal return (AAR) is calculated for each month separately by adding the abnormal returns of each individual firm and taking the simple average. It is therefore equal to the average abnormal return of all the firms in month 𝑡. This generalization allows for easy interpretation and does not share the same exposure to outliers as a case study would have. 𝐴𝐴𝑅₁ = 1 𝑁₁ 𝜀01 @A 0B: (3) From the AARs the cumulative average abnormal return (CAAR) is calculated. It is equal to the summation of the AAR over the time period 𝑡: to 𝑡C. The CAAR shows the abnormal gains

over the entire period and is therefore able to show the total of the returns over the time period. In order to have an insight in the evolution of the abnormal gains of the merger the CAARs will be calculated for every of the five years following the announcement of the merger, together with the CAAR of the entire period. The results are displayed and discussed in section 4.1. 𝐶𝐴𝐴𝑅₁1_EF ₌ _𝐴𝐴𝑅 1 1F 1B1E (4) The test statistic needed to show significance of the results found is calculated with a crude dependence adjustment as described by Brown & Warner (1980, p. 251, equation A.4 and A.5). This t-test takes into account the cross-sectional dependence of the returns and is therefore believed to be the correct test. The following formula estimates the statistic over period 𝑡: - 𝑡C: 𝑡_C− 𝑡_: 1F𝐴𝐴𝑅₁ 1E 1 59 𝐴𝐴𝑅1− 𝐴𝐴𝑅HIJKL 6N4OKL C 1B;8 1B: : C (5)

Where 𝑡₀ is equal to the first month of the year under review, thus 𝑡₀ ∈ {1, 13, 25, 37, 49}. The described methodology assumes that the unexplained performance, or the abnormal returns, is caused by the merger. This assumption may however be too strong, as there are other factors which may contribute to an above predicted performance. For example, if the top competitor of firm 𝑖 loses all of its factories to a fire accident, firm 𝑖 will experience

(16)

16 some unexpected gains. The accident renders the competitor incapable of supplying stock to its stores, causing an increase in the demand of the goods of firm 𝑖. Following this reasoning, the stock price of firm 𝑖 should therefore increase, and thus the stock price returns and the abnormal returns as well. This is an example of abnormal gains which are unrelated to the merger. Thus, an important flaw of the methodology is that the abnormal returns do not differentiate between the source of gains. Therefore, the abnormal returns may lack the ability to give causal meaning to the results found. However, as described by Lubatkin (1983) the abnormal returns methodology is the accepted method in finance literature and is therefore believed to be sufficient.

However, due to this problem a different test is needed to conclude whether or not the abnormal returns are caused by the merger. Therefore, the pre-merger performance is calculated as well; following the same methodology as described above. First a regression is run on the data 𝑡_S;8 to 𝑡_S:. Then the following hypotheses are tested:

𝐻₈: 𝐶𝐴𝐴𝑅_OVL = 𝐶𝐴𝐴𝑅_OJ61 𝐻_:: 𝐶𝐴𝐴𝑅_OVL ≠ 𝐶𝐴𝐴𝑅_OJ61 The hypothesis 𝐻8 is tested with a test statistic equal to:

𝑇 =𝐶𝐴𝐴𝑅OJ61− 𝐶𝐴𝐴𝑅OVL

𝑆_Z/ 𝑛 (6)

Where 𝑆Z equals the standard deviation of the difference between the CAARs and 𝑛 is equal

to the the number of months present in the CAAR. This test is known as a dependent sample t-test. The most important assumptions of this test are that the dependent variable should be measured on a continues scale, furthermore, the test group should consist of related pairs. As the CAARs are calculated pre- and post-merger with the same group of firms these assumptions hold. Results of this test are summarized in table 4.2.

In order to examine the firms with the above mentioned methodology it is important to have access to a list of completed mergers in Germany. Also the stock price data of these merged firms, together with the stock price data of the control group, is needed to follow this methodology. How this data is obtained is explained in section 3.2 together with descriptive statistics on the sample.

(17)

17 3.2 Data

The data of merging German firms is collected over the period 2005 to 2010. This period is chosen to have sufficient post-merger acquisition data (which is 5 years), whilst focusing the research on firms which merged recently.

First a list of all publicly listed German mergers over this period is obtained from Thomson ONE. From this list all mergers without an effective date are deleted because it is unsure whether the merger is completed, therefore including these in the sample may cause a bias. Then, following the data selection process of Lubatkin (1983), all firms who merge more than once or buyback their own shares are deleted from the dataset. This data selection method prevents a bias by excluding firms with several potential spikes in the abnormal return (caused by the additional mergers) from the sample. Clean-up mergers, i.e. mergers where the remainder of the shares is bought, are also dropped from the sample, as the gains from the merger are likely to be priced in already. Lastly, following the definition of a takeover, the mergers where less than 50% of the shares is obtained are excluded.

From this dataset the 30 mergers with the largest deal value are selected for the research, but due to data availabilty issues the sample shrinks to 26. The remaining sample – together with the control firms – is presented in the appendix in table 7.1. Descriptive statistics on the sample are available in table 7.2.

This relatively small sample could cause statistical power issues, but Healy, Palepu, & Ruback (1992) claim that by focussing on the firms with the largest deal value this problem may be ignored. In their study they examine the 50 largest takeovers of their sample. They argue that a relatively small sample which still represents a significant portion of the entire population, based on the total deal value, is sufficient. The selection of firms in this thesis represents 99% (=_{C`:^_,]}C]^C_ ) of the total deal value of the remaining sample, consisting of 54 firms. Furthermore, the sample represents 94% of the sample when it was adjusted for firms occurring multiple times and private firms were deleted from the sample; which seems to adhere to the criteria of Healy et al. (1992). They also discuss that choosing a smaller sample reduces the probability of firms occurring more than once in their dataset. However, in this thesis the probability of firms occurring more than once equals zero percent because the methodology of Lubatkin (1983) excludes these firms. Furthermore, they claim that if there are

(18)

18 gains from a takeover, the gains are most likely to occur when the target firm is large (implied by a large deal value). Lastly, the public concern about the takeover performance is typically largest when the deal value is large (Healy, Palepu, & Ruback, 1992), making these mergers more interesting to examine.

Then, following the methodology of Agrawal et al. (1992), a control firm is selected from a European index for each of the 26 remaining firms, based on the market value in the nearest year. Firms were selected not solely from German indexes, because firms listed in Germany deviated more than 10% between market values. Therefore, the search window was extended to other indexes. The chosen indexes have correlation exceeding 70% (as shown in Table 3.1) with the German index, which are significant at the 1% level. They are therefore believed to be a just choice. A complete overview of the sample, together with its control firm is available in the appendix in table 7.1.

Table 3.1

Correlation matrix of used indexes based on their returns.

Indexes

DAX SDAX CAC40 CACAT AEX AMX ASCX

DAX 1 SDAX 0,7819*** 1 CAC40 0,9237*** 0,7697*** 1 CACAT 0,9243*** 0,8006*** 0,9962*** 1 AEX 0,8728*** 0,7584*** 0,9087*** 0,9154*** 1 AMX 0,7816*** 0,8147*** 0,8146*** 0,8433*** 0,8370*** 1 ASCX 0,7134*** 0,8103*** 0,7522*** 0,7783*** 0,7621*** 0,8548*** 1

*, ** and *** indicate significance at 10%, 5% and 1%, respectively.

This methodology is a weaker form of the control group method than presented in Agrawal et al. (1992) and Dimson & Marsh (1986), but is assumed to be sufficient. For example, Agrawal et al. (1992) build their control portfolio based on a firm’s size decile in the NYSE. The control group fluctuates yearly and needs to be checked continuously for potential mergers within the control group. This paper is unable to use their methodology due to the fact that the chosen firms are not clustered in a single index. Agrawal et al. (1992) focus solely on NYSE-listed firms and are therefore able to use the unselected NYSE firms as a control group. The

(19)

19 firms in this paper are not solely from the DAX, and thus either another method is needed or the firm size effect variable needs to be dropped completely. However, due to the persuasive evidence discussed in the paper of Dimson & Marsh (1986) the size effect cannot be ignored. Therefore, this thesis uses a single control firm for each of the firms instead of dropping the size factor completely. The found results may however lose statistical power due to the law of large numbers, which implies a larger probability to deviations from the true mean.

4 Empirical Results and Analysis

4.1 Post-merger CAARs

The methodology as described in section 3 is used to measure the post-merger performance of the acquiring firm. The results for the post-merger sample are shown in table 4.1. On average the acquiring firm seems to significantly underperform the benchmark model in the first four years following the merger. This is however cumulatively undone in the final year.

The results found over the first four years are similar to the paper of Agrawal et al. (1992), however the results of this paper are seemingly larger and significant in both the sub-periods and the total period. The main conclusion about these results is that the anomaly is apparently present in different markets than the American NYSE.

In contrast to their results, this paper finds that the German market does not have performance significantly different from zero in the fifth year of the merger; in fact, it is equal to zero. This result is more in line with the papers of Mandelker (1974) and Abhyankar, Ho, & Zhao (2005). Mandelker (1974) concludes that the stockholders earn ‘normal’ returns, similar to other investments with the same risk level. He also explains that – according to his findings – the market for mergers is perfectly competitive and all the information regarding the merger is efficiently added to the stock price. Perhaps this holds for the German merger market as well. Similarly, Abhyankar, Ho, & Zhao (2005) conclude that investors do not prefer a merged portfolio over an unmerged portfolio. The t-value shown in table 4.1 indicates that the overall cumulative return is equal to zero. This implies that it is not statistically different from the benchmark portfolio, as they conclude as well.

More notably, the sub-period CAAR seems to be the largest in the first year after the announcement. Furthermore, the sub-period CAAR increases every year and is positive starting in the fourth year. The increase in average abnormal returns may be explained by an adjustment

(20)

20 period following the merger. Where, for example, employees and/or managers need to adjust to the newly created situations, causing inefficiencies. When the inefficiencies are resolved the firm returns to earning normal returns.

Another possible explanation of the insignificant result in the long-run is the free rider problem, as Berk & DeMarzo (2014, pp. 950-951) discuss.

Table 4.1

Post-merger performance of the acquiring firm over different time periods.

Months after the merger announcement Sub-period Cumulative Average Abnormal Return (CAAR) Total Cumulative Average Abnormal Return (CAAR) Percent of positive Abnormal Returns (%) 1-12 -8,94%*** -8,94%*** 44,55% (-15,55) (-15,55) 13-24 -3,03%*** -11,97%*** 47,44% (-5,27) (-29,44) 25-36 -2,90%*** -14,87%*** 45,83% (-5,05) (-44,79) 37-48 4,26%*** _-10,61%*** _55,45% (7,41) (-36,90) 49-60 10,61%*** _0% _53,85% (18,45) (0,00)

They give the game-theoretical example in which the manager of firm 𝑥 is able to increase the share price of firm 𝑦 from $45 to $75 if it obtains the majority of the shares, and replaces the management of firm 𝑦. If the stockholders of firm 𝑦 know of this potential, the manager of firm 𝑥 is unable to buy the shares for less than $75, because shareholders essentially choose between selling for the tender price of i.e. $60, and the final value of $75. Thus, firm 𝑥 is unable to profit from the merger, as the minimum offer price equals the growth potential. This could

(21)

21 explain the situation as presented in table 4.1, because firm 𝑥 would need time to make firm 𝑦 profitable but end up at par with the starting situation.

Harder to explain is the difference in results with Agrawal et al. (1992), who find a statistically negative return of -10,26% over 5 years. This difference in result can be either due to the difference in control group selection or, as intended to accomplish, due to the use of a sample which does not overlap previous samples.

4.2 Comparison between pre- and post-merger CAARs

Unfortunately, the methodology used in section 4.1 fails to give causal meaning to the results found due to the inability of abnormal returns to differentiate between the source of gains, as discussed in section 3.1.

However, it is possible to compare the pre- and post-merger returns. A comparison between these abnormal returns adds to the research by testing for clear differences between the two. As explained in section 3.1 the abnormal returns cannot distinguish between the source of gains. Although tests on the differences between the two are also unable to pinpoint the source of abnormal returns, they are able to show whether the merger effects the firms positively. By comparing the pre- and post- period a better conclusion can be formed on the matter. This analysis is performed using a dependent sample t-test.

Again a test is run for each of the five years, where in this test the time periods are tested on opposite side, thus 𝐶𝐴𝐴𝑅₁1_cEcEF_{is compared with 𝐶𝐴𝐴𝑅}

1dEF

1dE _{. The results are displayed in table}

4.2. As the total CAARs are used, the number of months’ increases in each of the tests. This analysis produces test-statistics which are surprising and perhaps more meaningful than the results of table 4.1. Where the main finding of table 4.1 is the negative total CAAR in the first four years, which pars in the fifth year, table 4.2 shows that – on average – the merged firms fail to outperform their unmerged self. The t-statistics reject the null hypothesis of equal CAARs at the 1% level. Over all the time periods the difference between the total CAARs is statistically significant and negative, implying that the firm did not gain from the merger.

(22)

22

Table 4.2

Comparison of pre- and post-merger cumulative abnormal returns.

Months before/after the merger announcement Pre-merger total CAAR Difference between pre- and post-merger

CAARs Standard Deviation of Difference (%) 1-12 -3,23%*** -5,71%*** 0,0596 (-5,37) (-3,32) 1-24 -4,95%*** -7,02%*** 0,0537 (-11,65) (-6,40) 1-36 2,38%*** -17,25%*** 0,0720 (6,86) (-14,37) 1-48 -4,46%*** -6,15%*** _0,0694 (14,86) (-6,14) 1-60 3,27%*** _-3,27%*** _0,0711 (12,16) _(-3,56)

Thus, for the sample used, the merger did – on average – not increase the abnormal performance of the acquiring firm. From a stockholder’s point of view this is worrisome, because the merger did cost money (on average the mergers in the sample cost 1,1 billion euros), but the results did not increase the firm’s performance compared to its unmerged self. However, a bias may be present in these results because most of the pre-merger CAARs are measured before the financial crisis, while the post-merger CAARs are partially during this crisis. The bias probably arises due to the constant beta assumption, because the betas are likely to fluctuate over time and especially in crisis (Lewellen & Nagel, 2006).

Summarizing, first the methodology of Agrawal et al. (1992) is used to estimate the post-merger CAARs and these are tested for significance using the t-statistic as presented by Brown & Warner (1980). The main finding in this segment are statistically negative total period CAARs in the first four years, with a maximum in year three of -14,87% (𝑡 = −44,79). The

(23)

23 final two years examined have positive CAARs, resulting in a total period CAAR of 0% in year five.

These results fail to give causal meaning to the situation. Therefore, a second test is performed to compare the pre- and post-merger CAARs. This test rejects the null hypothesis which states that the pre- and post-merger CAARs are equal; the result is statistically significant over each of the five years. The results of this test indicate that the merged firm has significantly lower abnormal returns than it had prior to the merger.

5 Summary and Conclusion

This paper analyses the long-run stock price performance based on a model as presented by Dimson & Marsh (1986) and used by Agrawal, Jaffe, & Mandelker (1992). The model is an extension of the basic CAPM, where the extension is a size factor. This extension is needed because the CAPM fails to correctly estimate long-run returns (Dimson & Marsh, 1986).

In contrast to most previous literature, this paper uses data from Germany to examine the merger performance of the acquiring firm. Previous literature primarily focuses on the NYSE. This small scope of research causes largely overlapping datasets and therefore overall somewhat similar results (although the differences in the methodology used causes the opposite in some cases). To examine whether the post-merger anomaly also exists in different areas of the world or is native to the NYSE, a dataset from Germany is used. In this paper the 26 largest takeovers are examined using the model presented in Agrawal et al. (1992).

The model used is however slightly adjusted due to a difference in data selection. Ideally, the sample of merged firms consist of firms from the same index. If so, the size factor consist of a control portfolio where the control firms are all the firms which are in the same size decile as the merged firm. Unfortunately, the sample used in this thesis consist of firms from different indexes, therefore an adjustment is needed. To avoid using the basic CAPM – and thus to adhere to the critiques of Dimson & Marsh (1986) – each firm is given a control firm; the control firm is assigned based on market value of the firm at the announcement date of the merger. The firms and their assigned control firm are regressed following the Agrawal et al. (1992) model. The resulting betas are assumed to be constant over the estimated time periods, which are 𝑡_: to 𝑡_;8 and 𝑡_S;8 to 𝑡_S:. Lewellen & Nagel (2006) argue that this assumption may

(24)

24 be too strong, but because this thesis aims to compare the results with Agrawal et al. (1992) it follows their methdology as closely as possible.

This methodology calculates the simple average abnormal returns in month 𝑡, after which it uses the sum of these returns over several time periods to find if the merger outperforms the benchmark portfolio. A flaw in this methodology is that it fails to compare the pre-merger returns with the post-merger returns in order to find a statistical difference. To combat this, a dependent sample t-test is performed to test for differences between the time periods of the CAARs.

The main findings, using the methodology of Agrawal et al. (1992), are statistically negative total period CAARs in the first four years after the merger announcement, with a maximum of the negative returns in year three of -14,87% (𝑡 = −44,79). The final two years examined have positive CAARs, resulting in a total period CAAR of 0% in year five. Interestingly, the sub-period CAARs are increasing yearly, what may be explained by an adjustment period needed for the new management. Another possible explanation is known as the free-rider problem.

Although the result found is an improvement for the stockholders compared to the result of Agrawal et al. (1992), who find a significant negative total period CAAR in year five of -10,26%, it is arguably worrisome. If the merged firms fail to outperform the benchmark portfolio, the merger seems to fail its intended purpose (assuming firms are stockholder value maximizing).

However, as previously discussed, this methodology lacks a clear connection with causality. The results are interesting on their own and require further research, but they are debatable. This is illustrated by the difference in results found in previous literature. In order to further add to the debate, this paper more clearly tests for differences between CAARs by comparing the post-merger CAARs with the pre-merger CAARs.

The resulting t-statistics of this test reject the null hypothesis, which states that the pre- and post-merger CAARs are equal; the result is statistically significant over each of the five years. This result implies that the merged firm, besides from failing to outperform the benchmark, fails to outperform its unmerged self. This implication has more causal meaning than the examination of the post-merger years. Apparently management overestimates the

(25)

25 gains from the merger. However, the results found may still be due to the use of a possibly wrong model or due to the sample selection method.

Although this paper succeeds in finding different results than most previous literature by using a new sample, the results found are, following the reasoning of Healy, Palepu, & Ruback (1992), based on a relatively small sample. This may affect the outcome. Therefore, future research should use similarly large samples as Agrawal et al. (1992) to further investigate the anomaly.

Another potentially interesting addition to future research is the lengthening of the examination window. In current literature five years is the standard, however the finding of 0% total period CAAR after five years feels unsatisfying, especially considering the sub-period CAARs are increasing yearly. It could be interesting to research the effects in the very long-run, up to i.e. 10 years, because the adjustment period – according to the results found in this paper – seems to turn around after three years but it is unclear if it continues to increase after five years.

A potential flaw of previous literature may also lie in using a selective sample, namely a sample which contains only firms from the same index. This paper avoids this, but encountered problems building a control portfolio. Potentially, other results are found by first building a control portfolio based on firms from different indexes with a high correlation to the index of the merged firm, and then following the size decile methodology of Agrawal et al. (1992). Future research could also compare results using a wide range of methodologies, to test for differences in results found. However, the authors may find trouble is selecting the ‘correct’ methodology in their results.

Furthermore, the methodology could be improved by using a rolling regression. Papers, like Lewellen & Nagel (2006), argue that assuming betas to be constant over a time period as long as 5 years is unrealistic. Therefore, a rolling regression should be used with time windows changing quarterly. Assuming this methodology to be correct, the results of i.e. Agrawal et al. (1992) loses statistical power, thus in the future a rolling beta should become the standard.

Lastly, different methodologies of measuring long-run performance than the abnormal return method should not be discarded. For example, Abhyankar, Ho, & Zhao (2005) present a more econometric way of estimating the success of mergers. Instead of using abnormal returns, they test by using a stochastic dominance relation. They, similar to this paper, conclude

(26)

26 that over their entire sample investors do not prefer a merger portfolio over a portfolio which is matched based on market value and book-to-market ratio. Their methodology needs further examination and is therefore interesting.

(27)

27

6 Bibliography

Abhyankar, A., Ho, K.-Y., & Zhao, H. (2005). Long-run post- merger stock performance of UK acquiring firms: a stochastic dominance perspective. Applied Financial Economics, 679-690.

Agrawal, A., & Jaffe, J. F. (2000). The post-merger performance puzzle, in JAI Series:

Advances in Mergers and Acquisitions (Eds) G. Cooper and A. Gregory, JAI, Elsevier

Science, 1, 7-41.

Agrawal, A., Jaffe, J. F., & Mandelker, G. N. (1992). The post-merger performance of acquiring firms: a re-examination of an anomaly. Journal of Finance, 47(4), 1605-1621. Berk, J., & DeMarzo, P. (2014). Corporate Finance (Vol. 3). London: Pearson.

Brouthers, K. D., Hastenburg, P. v., & Ven, J. v. (1998). If Most Mergers Fail Why Are They so Popular? Long Range Planning, 347-353.

Brown, S. J., & Warner, J. B. (1980). Measuring security price performance. Journal of

Financial Economics, 8, 205-258.

Bruner, R. F. (2002). Does M&A Pay? A Survey of Evidence for the Decision-Maker. Journal

of Applied Finance, 48-68.

Centraal Bureau voor de Statistiek. (2016, May 30). Bedrijven; fusies en overnames,

bedrijfsgrootte, rechtsvorm, bedrijfstak. Retrieved 2016, from StatLine:

http://bit.ly/1U8lT9c

Dimson, E., & Marsh, P. (1986). Event study methodologies and the size effect. Journal of

Financial Economics, 113-142.

Dodd, P. (1980). Merger proposals, management discretion and stockholder wealth. Journal of

Financial Economics, 105-138.

Dodd, P., & Ruback, R. (1977). Tender Offers and Stockholder Returns: an empirical analysis.

Journal of Financial Economics, 351-373.

Franks, J., Harris, R., & Titman, S. (1991). The postmerger share-price performance of acquiring firms. Journal of Financial Economics, 81-96.

Healy, P. M., Palepu, K. G., & Ruback, R. S. (1992). Does corporate performance improve after mergers? Journal of Financial Economics, 135-175.

Jensen, M. C., & Ruback, R. S. (1983). The market for corporate control: The Scientific Evidence. Journal of Financial Economics, 5-50.

(28)

28 Langetieg, T. C. (1978). An application of a three-factor performance index to measure

stockholder gains from merger. Journal of Financial Economics, 365-383.

Lewellen, J., & Nagel, S. (2006). The conditional CAPM does not explain asset-pricing anomalies. Journal of Financial Economics, 82, 289-314.

Lubatkin, M. (1983, April). Mergers and the performance of the acquiring firm. The Academy

of Management Review, 8(2), 218-225.

Lyon, J. D., Barber, B. M., & Tsai, C. L. (1999). Improved methods for tests of long-run abnormal stock returns. Journal of Finance, 165-201.

Mandelker, G. (1974). Risk and Return: The case of merging firms. Journal of Financial

Economics, 303-335.

Roll, R. (1978). Ambiguity when performance is measured by the securities market line.

Journal of Finance, 33, 1051-1069.

Royston-Bailey, J. (2015, May 11). The Market Mogul. Retrieved 2016, from The Market Mogul: http://themarketmogul.com/german-ma-boom/

The Wall Street Journal. (2015, December 3). 2015 Becomes the Biggest M&A Year Ever. Retrieved 2016, from The Wall Street Journal: http://www.wsj.com/articles/2015-becomes-the-biggest-m-a-year-ever-1449187101

(29)

29

7 Appendix

Table 7.1

An overview of the sample together with information about the control firm.

Ticker Symbol of merged firm Merged firm name Announcement date Control firm name Country Control firm is listed Difference in Market value (%)

D:CON Continental 25/07/07 Post NL Netherlands 2,75%

D:VOW Volkswagen 06/05/09 Royal Dutch Shell

Netherlands 7,97%

D:FRE Fresenius SE & Co

14/10/05 Bigben France 2,72%

D:SIE Siemens 30/03/05 Groupe Gorge France 4,99%

D:DAI Daimler 30/04/08 Bayer Germany 0,85%

D:PFD4 Pfleiderer 08/02/05 Baywa Germany 6,89%

D:PSM Prosiebensat1 11/01/05 SBM offshore Netherlands 3,66% D:LHA Deutsche

Lufthansa

08/09/05 Technip France 1,93%

D:GBF Bilfinger 18/05/06 Corbion Netherlands 1,88%

D:ALV Allianz 18/01/07 Deutsche

Telekom

Germany 6,29%

D:NSU Audi 26/11/09 AKZO Nobel Netherlands 1,62%

D:RHK Rhoen klinikum

18/12/05 Royal Boskalis

Netherlands 0,80%

D:MPC MPC Capital 12/02/08 TKH Group Netherlands 7,47% D:KBU1 Colonia Real

Estate

15/09/06 Binckbank Netherlands 0,98%

D:DRI Drillisch 11/12/09 Royal

Wessanen

(30)

30 Table 7.1 (continued) Ticker Symbol of merged firm Merged firm name Announcement date Control firm name Country Control firm is listed Difference in Market value (%)

D:EVT Evotec 06/03/05 Accell Group Netherlands 3,39%

D:MRK Merck & Co 08/02/05 Valeo France 1,59%

D:BFI Berliner 18/07/07 Washtec Germany 3,19%

D:SAP SAP 25/02/06 BNP Paribas France 4,22%

D:SCUN Schuler 28/03/07 Capital Stage Germany 3,93%

D:TGT Telegate 19/02/08 Devoteam France 6,41%

D:BHS Curanum 26/11/07 Bertrandt Germany 1,80%

D:SUR Surteco 11/01/08 Biotest Germany 2,95%

(31)

31

Table 7.2

Descriptive statistics on the sample of 26 firms based on monthly returns.

Merged firm

Number of

observations Mean Min Median Max

Standard deviation D:CON 121 _0,0214 _-0,4587 _0,0296 _{0,6401 0,1401} D:VOW 121 _0,0210 _-0,4947 _0,0164 _{0,4677 0,1178} D:FRE 121 _0,0109 _-0,2851 _0,0182 _{0,6003 0,1065} D:SIE 121 _0,0017 _-0,2957 _0,0082 _{0,3991 0,1085} D:DAI 121 _0,0084 _-0,2790 _0,0081 _{0,4151 0,1009} D:PFD4 121 _0,0092 _-0,5074 _0,0000 _{0,5708 0,1661} D:PSM 121 _0,0201 _-0,5500 _0,0226 _{0,8795 0,2167} D:LHA 121 _0,0000 _-0,3123 _0,0054 _{0,2989 0,1037} D:GBF 121 _0,0163 _-0,3624 _0,0234 _{0,2916 0,1006} D:ALV 121 _-0,0033 _-0,2994 _-0,0082 _{0,3864 0,1022} D:NSU 121 _0,0117 _-0,2500 _0,0001 _{0,2725 0,0751} D:RHK 121 _0,0031 _-0,1622 _0,0059 _{0,2318 0,0732} D:MPCK 121 _-0,0004 _-0,7078 _0,0061 _{1,5279 0,2160} D:KBU1 121 _0,0375 _-0,4228 _0,0000 _{1,5762 0,2442} D:DRI 121 _0,0284 _-0,4910 _0,0340 _{0,6170 0,1434} D:DEQ 121 _0,0075 _-0,2081 _0,0021 _{0,3098 0,0619} D:WDI 121 _0,0379 _-0,3825 _0,0336 _{1,0723 0,1608} D:EVT 121 _0,0031 _-0,5646 _-0,0395 _{1,3067 0,2584} D:MRK 121 _0,0100 _-0,2248 _0,0072 _{0,3224 0,0851} D:BFI 121 _-0,0200 _-0,7727 _0,0000 _{0,5833 0,1532} D:SAP 121 _0,0069 _-0,3898 _0,0174 _{0,5068 0,1087} D:SCUN 121 _0,0125 _-0,3334 _-0,0036 _{0,6085 0,1229} D:TGT 121 _0,0158 _-0,2111 _-0,0035 _{0,9830 0,1319} D:BHS 121 _0,0086 _-0,2886 _-0,0048 _{0,6851 0,1221} D:SUR 121 _0,0085 _-0,4443 _0,0070 _{0,3228 0,1186} D:ISR 121 _0,0098 _-0,3221 _0,0081 _{0,2787 0,1121}

Long-run post-merger stock price performance : do mergers increase the stock’s performance?