Time Patterns in Merger and Acquisitions

Time Patterns in Merger and Acquisitions

The Effect on Shareholder Value

Master Thesis MSc BA: Specialization Finance

PETER VAN GULIK -S1622013

Abstract

In this paper I analyze time patterns in merger and acquisition activity between 2000 and 2010. I focus on the effects these patterns have on shareholder value. My sample consists out of 1580 worldwide mergers and acquisitions. Each month is classified as being either a low, normal or high merger and acquisition activity period. Compared to normal periods I find that in months with high merger and acquisition activity more often all the synergies will be allocated to the target. The total synergy gain remains unaffected by the level of merger and acquisition activity.

Document Properties

JEL Classification G14, G15, G34

Keyword Mergers and Acquisitions, Time Patterns, Shareholder value

Word Count 8031

Status Final

Compilation Date August 31, 2010

Supervisor dr. ing. N. Brunia

I. Introduction

same behavioral aspects that drive merger waves or are they unrelated? The problem around which the analyses in this paper are constructed concerns the shareholders of both the target and acquiring company and are formulated as follows; Does monthly merger and acquisition frequency, when

categorized as low, normal or high, affect the total amount of synergies gained or the distribution of those synergies? In the following section of this paper the research question will be divided into two

testable hypotheses.

The subject on which this paper focuses has not yet been researched, even though time patterns, if exploitable, can offer shareholders a considerable advantage on the stock market. One of the reasons why no prior literature exists regarding this exact subject is the lack of merger and acquisition observations. Furthermore it is also difficult to accurately classify each monthly period based upon its merger activity. I solve this by using a simplistic method, based upon a three month rolling average, to classify each of the 120 monthly periods. Taking into account the previous month's merger activity and compensating for the possibility of economic trends in the data.

In my literature analysis I will focus on two subjects: merger and acquisition waves and merger and acquisition returns. I chose the literature on merger and acquisition waves because it focuses on specific time patterns in merger and acquisition activity. Possible explanations for merger and acquisition waves might also apply to the shorter fluctuations in merger and acquisition activity on which this paper focuses. Whereas the literature on post merger announcement returns helps identifying variables that can explain abnormalities observed in the synergy gains, contributing to the robustness of my analysis.

Most scholars assume waves to be industry bound spanning over a 24-month time period (Harford, 2005 and Shleifer & Vishny, 2003). I classify each month individually based on the weighted rolling average frequency of the previous three months. This method allows for short periods of low, normal and high merger and acquisition activity to be identified, analyzed and compared with each other. By my knowledge this is the first paper that attempts to analyze and identify such patterns. Where other papers focus solely on merger and acquisition activity in the US market, this paper takes an international approach. One of the obstacles when analyzing merger and acquisition behavior on an international level is the currency in which the stock of the involved parties is denoted. The synergies cannot be estimated when the currency, in which the stock prices of the involved parties is denoted, differ. To overcome this problem the stock prices of both the target and acquiring party are converted to Euros.

methodology used to measure synergies and classify periods. In this section I will also formally define the control variables and explain the how each of the hypotheses is tested. Subsequently I will discuss my dataset and its properties. Followed by a section of empirical results and implications after which I will conclude this paper with a short summary of the results.

II. Literature

Literature on Waves

It is well known that mergers and acquisitions tend to cluster in time (Michell & Mulherin, 1996). Past literature provides various explanations as to why this clustering happens. Harford (2005) classifies these explanations in two categories; behavioral and neoclassical. The behavioral explanations are based on the observed positive correlation between stock prices and merger and acquisition activity. Whereas the neoclassical explanations relate clustering in merger and acquisition activity to global or industry level shocks. Harford (2005) identifies 35 waves in 28 different industries between 1981 and 2000. He concludes that shocks provide the best explanation for the occurrence of merger and acquisition waves and does not find significant proof for any of behavioral explanations. Harford's (2005) findings imply there should be no monthly fluctuations in merger and acquisition activity. Rhodes-Kropf and Viswanathan (2004) delve deeper into the behavioral hypothesis. Rational managers should not be eager to accept stock offers from overvalued acquirers, unless they estimate the synergies to surpass the acquirer's overvaluation. The synergies determine whether an offer is accepted or rejected in Rhodes-Kropf and Viswanathan's (2004) model. They distinguish between a firm specific and market wide overvaluation component. Rational managers are aware of the overvaluation and can correctly estimate the real value of their company but will be unable to determine which part is due to the market wide overvaluation and which part is firm specific. In an overvalued market, the target will underestimate this market wide component and overestimate the firm specific component. The target's management is thus underestimating the shared, market wide component of the overvaluation and assesses synergies to be positive accepting the acquirer's bid. Although Rhodes-Kropf and Viswanathan advocate the behavioral hypothesis, they do not rule out the possibility that merger waves are triggered by technological or regulatory shocks.

one another, a merger in the industry that increases firm size for one CEO will cause other envious CEOs to be tempted to undertake value-dissipating but size-enhancing acquisitions, thereby starting a merger wave. Under such circumstances the acquirer is to incur a loss in market value whereas the target is likely to experience a gain in market value.

Another contribution worth mentioning is that by Shughart and Tollison (1984). They analyze annual merger and acquisition data from US between 1895-1979. They find that merger and acquisition activity can be explained by a white-noise process (random process) and thereby reject the hypothesis that mergers occur in waves. This contrasts with the assumptions and findings of other scholars especially those of Golbe and White (1993), who analyze mergers and acquisitions in the US between 1919-1979. They offer a direct econometric test of the proposition that mergers occur in waves using a fitted sine curve (which is readily recognized for its wave form). The fitted sine curves provided significant explanatory power thereby confirming the assumption that mergers occur in waves.

Literature on merger and acquisition returns

Devos et al. (2009) analyze on the underlying components of synergy gains in mergers and acquisitions. They study 264 mergers between 1980-2004 and estimate the synergies using the last known Value Line forecast in the financial statement. The combined Value Line forecast of the target and acquiring company are compared to the first available forecast of the merged firm, the difference between those are determined to be the synergies. On average they find statistically significant positive synergies amounting to 10.03% of the combined equity. Only 16.35% of those can be attribute to tax gains, the remaining 83.55% where classified as operating synergies. The majority of these operating synergies consist out of cutbacks on investments.

When the acquirer and target are not located in the same country they often fall under different corporate governance regimes. Bris & Cabolis (2008) analyze the impact a cross border merger has on the abnormal returns (target and acquirer) following its announcement. They conclude that improvements in accountability and transparency are positively valued by the market. Leading to higher post announcement returns for the target only if the corporate governance regime in the acquirer's country is better.

abnormal returns on event days -1 and 0. Combination offers of stock and cash also generate significant negative abnormal returns. Travlos's sample was composed out of 167 bids in the US market.

Hypotheses formulation

Rhodes-Kropf and Viswanathan (2004) argue that overvaluation in the market causes merger waves. They argue that due to the limited information both the acquiring and target possess, with regard to overvaluations in the market, they are unable to accurately estimate the synergy gains. Assuming that the shareholders of both firms are affected by this market wide overvaluation, and thus not able to accurately estimate the synergy gains, it is reasonable to assume that the synergies of mergers and acquisitions announced in months with a high merger and acquisition frequency differ from those announced in months with a low mergers and acquisitions frequency. From this I derive the following hypotheses:

H1 The synergies of a merger or acquisition are lower in high frequency periods.

Months in which a high merger and acquisition frequency is observed can also be driven by managerial envy (Goel & Thakor, 2010), leading to irrational choices and overbidding the estimated synergies. It differs from overvaluation driven merger waves in that the market correctly values the synergies and sets stock prices accordingly. Based on this argumentation I hypothesize that in high frequency months more than 100% synergies are allocated to the target company, causing the acquirer's stock price to fall upon announcement.

H2 In high frequency periods more than 100% of the synergies are allocated to the target.

Control variables

In the analysis presented in this paper I control for variables that are known to influence the abnormal returns (and thus the synergies) of mergers and acquisitions. I control for the method of payment, distinguishing between cash, stock and offers that combine cash and stock. Stock offers are known to have a negative effect on the abnormal returns of the bidding firm (Travlos, 1987), these in turn affect the synergies which are defined as a product of the abnormal returns and market value of both parties (see equation three in section three for the mathematical expression).

Companies that operate under a common law legal system are valued higher opposed to companies operating under a civil law legal system (Anderson & Gupta, 2009). Implying that the observed synergy gains are higher when a company in a common law system takes over a company in a civil law system and lower when a company in a civil law system takes over a company in a common law system. I control for this effect using a dummy variable.

The amount of debt, or leverage, of a company affects the abnormal returns it realizes on mergers and acquisitions. Jensen (1986) purposes the theory that high debt increases the probability of positive abnormal returns for the acquiring company. It reduces the amount of free cash flow available forcing the management to make better acquisitions. There is no concrete theory as to what effect the leverage ratio of the target has on its abnormal returns. Though highly indebt companies are often valued lowered due to their reduced cash flow. It is therefore quite likely that the abnormal returns of such targets are lower.

Moeller et al. (2004) find that the size of the acquiring company has a negative effect on its post announcement abnormal returns. According to their research the effect is robust to deal and firm characteristics. Furthermore it is not reversed overtime. I control for this effect using a dummy variable that enables me to distinguish between large and small acquirers. The negative effect on the acquirer's abnormal returns reduces the overall synergies. The coefficient of this variable is therefore expected to be preceded by a negative sign when explaining the first hypothesis. This variable also changes the synergy allocation, increasing the amount allocated to the target and decreasing the amount allocated to the acquirer. A positive sign is therefore expected to proceed the coefficient of this variable when explaining the second hypothesis.

Table 1: Expected signs of the control variables for each hypothesis based upon the literature.

Variable Expected Sign H1 Expected Sign H2

Shares Negative Negative

Cash Positive Positive

Cross Border Negative Negative

Civil to Common law Positive Positive

Common to Civil law Negative Negative

Acquirer Debt Positive Positive

Target Debt Negative Negative

III. Methodology

Determining shareholder value

Gains in shareholder value will be measured using traditional event study methodology (MacKinlay, 1997). Returns on the individual securities are described as logarithmic returns, equation (1), thereby decreasing the skewness and kurtosis of the distribution.

)

/

ln(

=

P

P

R

Where 𝑃𝑡 is the price of the security at event day 𝑡, adjusted for stock splits and dividend payouts.

Abnormal returns are calculated following the market and risk adjusted method (2) (MacKinlay, 1997). According to Brown and Warner (1985) this model has the highest probability of detecting abnormal returns when present in the data.

(

β

+

α

)

−

=

t

R

RM

AR

Where 𝑅𝑡 is the return of an individual security at event day 𝑡 and 𝑅𝑀𝑡 the return of the market

(MSCI World Index). Alpha is the intercept of the security and beta the correlation and volatility in relation to the market. The parameters for the market and risk adjusted model are estimated using daily returns from the 200-day estimation period prior to the announcement. The estimation period ranges from event day -201 to -1.

Based upon the observed abnormal returns the event period is defined as a 4-day period, ranging from day -1 to 2. For each day in the event period the returns are averaged across the sample's securities (𝐴𝐴𝑅). The 𝐴𝐴𝑅s are then cumulated to yield the cumulative average abnormal returns (𝐶𝐴𝐴𝑅). Every day in the event period (𝐴𝐴𝑅) is tested for statistical significance using the cross-sectional variance (MacKinlay, 1997). The 𝐶𝐴𝑅s are used in the regressions for the calculation of the synergies and are defined as the cumulated abnormal returns over the event window per security. The synergies of each merger or acquisition are calculated as the cumulated abnormal gain or loss in market value of the involved parties (equation 3). Where 𝑀𝑉𝑡𝐴 and 𝑀𝑉𝑡𝑇 are respectively the market

values of the acquirer and the target at event day 𝑡.

The MSCI world index is used as proxy for the market because its international characteristics match those of the sample. Before any of the returns are calculated each stock price quote is converted into Euros using the exchange that prevailed that day.

Classifying time patterns

In order to test if patterns in time affect shareholder value a classification method is required. This method should be able to distinguish between months with a low, normal or high level of merger and acquisition activity. Simply dividing the sample, after ranking it by monthly merger and acquisition activity, into three equal groups does not account for cyclical effects. To tackle this problem I developed a method that, based on a three-month-rolling-average, is able to account for such cyclical effects. First the number of mergers and acquisitions in month 𝑡 are summed with the weighted number of mergers and acquisitions from the previous months (𝑡-1 and 𝑡-2). This is divided by three yielding the rolling average number of mergers and acquisitions in month 𝑡. Mathematically this can be expressed as follows:

3

+

+

=

A

W

A

W

A

RA

Where 𝑅𝐴𝑡 is the rolling average, 𝐴𝑡 the number of mergers and acquisitions and 𝑊𝑡 the weight

associated with month 𝑡. The weights on the months 𝑡-1 and 𝑡-2 are meant to reduce the impact of extremely low and high levels of merger and acquisition activity on the rolling average. The rolling average describes the normal level of merger and acquisition activity, if the actual level of merger and acquisition activity is within a 25% boundary of the calculated rolling average the month is classified as normal (see equation 5).

%

25

%

25

%

25

%

25

>

<

<

−

>

−

D

D

D

_HIGH

NORMAL

LOW

⇒

⇒

⇒

Where 𝐷𝑡 is the percentage that the actual level of merger and acquisition activity deviates from the

rolling average. 𝐷𝑡 can be calculated as follows (see equation 6).

t t t t

A

_RA

RA

D

=

−

Finally the weights associated with each month (𝑊𝑡) depend on the percentage that it deviates from

the rolling average (𝐷𝑡). A weight of one is assumed for the first two months because, for these, 𝐷𝑡

t t

_D

W

+

=

1

1

Hypothesis testing

To test whether the synergies of mergers and acquisitions in high frequency periods are lower compared to normal periods, I employ a multivariate OLS regression. The regression consists out of two explaining (the variables in bold) and eight controlling variables. In the table bellow (table 2) each of those variables is defined and where possible expressed in mathematical terms.

Table 2: The definitions of the variables

Variable Type Expression

Low Period Binary The merger or acquisition occurred in a month with a low merger or acquisition frequency

High Period Binary The merger or acquisition occurred in a month with a high merger or acquisition frequency

Shares Binary Part of the payment is made in shares

Cash Binary Part of the payment is made in cash

Cross Border Binary The target and acquiring firm both resided _{in different countries}

Civil to Common law Binary

The country where the target firm is located has a legal system based on civil law and the acquiring firm a legal system based on common law.

Common to Civil law Binary

The country where the target firm is located has a legal system based on common law and the acquiring firm a legal system based on civil law.

Acquirer Debt Ratio 𝐷𝐸𝐵𝑇𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟− 𝐶𝐴𝑆𝐻𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟 𝐴𝑆𝑆𝐸𝑇𝑆𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟

Target Debt Ratio 𝐷𝐸𝐵𝑇𝑡𝑎𝑟𝑔𝑒𝑡_{𝐴𝑆𝑆𝐸𝑇𝑆}− 𝐶𝐴𝑆𝐻𝑡𝑎𝑟𝑔𝑒𝑡

𝑡𝑎𝑟𝑔𝑒𝑡

Firm size Binary 𝑀𝐴𝑅𝐾𝐸𝑇𝑉𝐴𝐿𝑈𝐸𝑎𝑐𝑞𝑢𝑖𝑟𝑒𝑟 > 10000

The multivariate OLS regression, with which the first hypothesis will be tested, can be expressed as follows (see equation 8).

i j i j i i

V

S

=

α

+

∑

(

β

)

+

ε

Where 𝑆𝑖 is the dependent variable describing the synergies (see equation 3) of the merger or

acquisition. 𝑉𝑖,𝑗 represent the independent variables as described in table 2 and 𝛽𝑖,𝑗 their associated

coefficients. 𝛼0 is the constant term, or intercept, and 𝜀𝑖 the error term. Each of the dependent

variables will be tested on statistical significance using the Student T-Test.

I account for heteroscedasticity (inconstant variance) in the error terms by applying White's heteroscedasticity consistent standard errors (HCSE). If heteroscedasticity is present in the data it can cause the variance of the coefficients to be underestimated. The estimators will still be unbiased and consistent but it increases the chance of an type I error: incorrectly accepting the ℎ0 hypothesis.

White's method only affects the variance of the coefficients where the error terms exhibit heteroscedasticity and does not influence the variance of coefficients where the error terms do not contain heteroscedasticity (White, 1980).

The second hypothesis cannot be evaluated using the OLS technique because the dependent variable is binary. Instead I opted for a probability (probit) regression using White's robust covariance to account for heteroscedasticity in the error terms. A porbit regression assumes there is a latent variable 𝑌∗ _{describing the probability of the model's binary dependent variable 𝑌 being true. Instead}

of explaining 𝑌 a probit regression will try to explain this latent variable 𝑌∗_{. The dependent variables}

of this regression are identical to those of the multivariate OLS regression explained in the previous paragraph. Its equation can be expressed as follows (see equation 9).

i j i j i i

V

SAT

=

α

+

∑

(

β

)

+

ε

The right hand side of the equation is completely identical to the right hand side of equation 8 (the multivariate OLS equation). In the left hand side of the equation 𝑆𝑖 is replaced by 𝑆𝐴𝑇𝑖, which is a

binary variable describing whether more than 100% of the synergies are allocated to the target. 𝑆𝐴𝑇𝑖

can de expressed as follows (see equation 10).

1

_≥

+

=

MV

MV

MV

CAR

SAT

Where 𝑀𝑉𝑡𝐴 and 𝑀𝑉𝑡𝑇 are the market values of the acquirer and the target at event day 𝑡 and 𝐶𝐴𝑅𝑇

IV. Data

The mergers and acquisitions evaluated in this paper have been obtained through Bureau van Dijk's Zephyr. Stock and other firm specific data are collected from Thompson-Reuters' DataStream. The initial Zephyr export contained 22,892 mergers and acquisitions. All merger and acquisitions where the rumor, announce and completion date are identical have been removed from the sample. I have applied this procedure to eliminate the possibility of including mergers and acquisitions in my sample for which Zephyr has insufficient data to accurately estimate these dates. As it would be regulatory and humanly impossible for a listed company to announce and complete a merger or acquisition on the same day without rumors being spread. Thus I assume such entries to be erroneous and exclude them from my sample. In the following table (table 3) I present the complete list of applied constraints and how these affect the amount of observations in my sample.

Table 3: lists the number of mergers and acquisitions left after applying the specified constraint.

Constraint Result

Deals between 1-1-2000 and 1-1-2010 517,565

Deal value of at least 10 million euro 121,166

Method payment: shares, cash or a combination 30,058

Final stake of acquiring firm at least 50% 23,520

Deals status completed 23,515

Target listed 6,379

Acquirer listed 3,881

Target stock prices available on Datastream 2,135

Acquirer stock prices available on Datastream 2,098

Enough stock price data (target and acquirer) 1,727

Information on debt is available on Datastream (target and acquirer) 1,601 Exchange rates for currency conversion available (target and acquirer) 1,593 Adjusting outliners in synergy gains and abnormal returns 1,580

The final sample consists out of 1580 mergers and acquisitions over a ten year time period and has been adjusted for outliners and lack of information (see table 3 for specifics). In the following section (Descriptive Statistics) I will briefly described the characteristics of my dataset.

Descriptive Statistics

total Merger and acquisitions)1_{. Most of the mergers and acquisitions in the sample were announced} in the months April, May and June2

Graph 1: The average and median merger and acquisition frequency per month.

. These three months alone account for 28.34% of all the mergers and acquisitions in the sample. Aggregating the mergers and acquisitions for each year reveals that most of the mergers and acquisitions were announced in 2006; 16% (253).

On average each sample year contains 158 merger and acquisition observations with a standard deviation of 69. The median number of yearly merger and acquisitions amounted to 166. Most of the mergers in the sample occurred between 2003-2008. The graph 2 shows an overview of the dispersion across the sample years combined with the monthly median and average per sample year.

1 _{On average 11, 11 and 10 in January, February and August respectively.} 2 _{On average 14, 15 and 15 a month in April, May and June respectively.}

8 9 10 11 12 13 14 15 16 17 18 1 2 3 4 5 6 7 8 9 10 11 12 M er ge rs an d ac qu isti on s

Month of the year

Graph 2: The average and median merger and acquisition frequency a year over the sample period.

More than half of all target and acquiring companies resided in North America, respectively 51% and 49% of the total sample. Roughly 21% of acquiring companies were located in Europe and about 18% of targets. Meaning that during the sample period European based companies acquired more cross region assets compared to the other continents in the sample. Asia accounted for 23% of the sample's deals. Both Africa and South America represented 1% to 2% of the sample's companies (see table 1).

Table 4: Geographical dispersion targets and acquirers in absolute values and percentages of the total sample size.

Region Amount Percentage

Targets Acquirers Targets Acquirers

Asia 367 367 23.23% 23.23% Europe 277 326 17.53% 20.63% North America 808 775 51.14% 49.05% South America 19 11 1.20% 0.70% Africa 21 25 1.33% 1.58% New Zealand/Australia 88 76 5.57% 4.81%

Decomposing the method of payment into geographical regions (table 5) reveals that most of the mergers and acquisitions in Asia, Europe and North America used cash as payment method. A mixed payment method of cash and shares is relatively popular in North America, accounting for nearly 40% of all merger and acquisitions but is almost non-existent in Asia. Only 4% of mergers and acquisitions observed in Asia used a payment method that mixed of cash and shares.

Table 5: Dispersion of payment method

Asia Europe North America

Cash 49.05% 60.74% 38.06%

Shares 47.41% 27.61% 32.13%

Cash & Shares 3.54% 11.66% 29.81%

Table 6 shows the amount of cross border and local deals, where the cross border deals have been categorized by triggered changes in legal system.

Table 6: Amount of cross border mergers and acquisitions and changes in legal system

Total Percentage

Cross Border 326 20.63%

Civil To Common Law 54 3.42%

Common To Civil Law 80 5.06%

No Change 192 12.15%

Local 1254 79.37%

Only 20% of the deals crossed nation borders, out of these roughly one third of the target companies undergo a change in legal system. Noteworthy to say that most of the acquirers and targets fell under a common law legal system, accounting for 61% and 62% of the total sample respectively. The following table (table 7) describes the abnormal returns of the acquiring and target company based on the market and risk adjusted method and mean adjusted method3

Table 7: Abnormal returns calculated using the mean adjusted and market and risk adjusted method.

. Further analysis in this paper make use of the market and risk adjusted method because of its higher statistical power (Brown & Warner, 1985).

Event day -2 -1 0 1 2 3 -2 to 3 Me an Ad ju st ed Target 0.0036 0.0101 0.1051 0.0389 0.0080 0.0028 0.1685 Acquirer 0.0007 0.0003 0.0012 -0.0006 -0.0007 -0.0014 -0.0005 Ri sk a nd M ark et Ad ju st ed Target 0.0036 0.0101 0.1050 0.0387 0.0077 0.0027 0.1678 Acquirer 0.0004 0.0001 0.0011 -0.0011 -0.0012 -0.0011 -0.0018

As seen from the table (table 7) the target gains an abnormal return of 10% on the day of the announcement, whereas the acquirer loses about 0.11% on the day after the announcement. The highest abnormal returns are observed from day -1 until 2, the event window will thus be defined over this 4-day period. The abnormal returns depart from normality due to their high kurtosis, which is often the case with financial data. As a result the Jarque–Bera test is rejected at a 1% significance level.

V. Results

Before going into detail on the results of the classification method I will verify that the average monthly synergies fluctuate over time. To do this the sample is split in ten decile groups based upon the monthly level of merger and acquisition activity. Their means are tested for equality using an ANOVA (analysis of variance). As a robustness check the medians of the groups are also tested for equality using the non-parametric Kruskal-Wallis test. The results are shown in table 8.

Table 8:Results for test of equality of the synergies splitting the sample in 10 decile groups based upon the number of mergers and acquisitions

Decile Group Mean Median _deviation Standard Jarque-Bera

4 < 0.1555 0.1236 0.1448 0.63 4-7 0.1425 0.1207 0.0803 0.70 7-9 0.1528 0.1639 0.1331 0.38 9-11 0.1833 0.1657 0.0942 1.54 11-13 0.1805 0.1788 0.0419 0.54 13-15 0.1998 0.1771 0.1022 1.39 15-17 0.1246 0.1253 0.0286 0.11 17-19 0.1053 0.1166 0.0423 0.61 19-22 0.0973 0.1006 0.0416 0.29 22 > 0.1077 0.1096 0.0283 0.16 ANOVA F-Test 0.0238 ** - - - Kruskal-Wallis - 0.0030 * - -

*significant at a 1% level; **significant at a 5% level; ***significant at a 10% level

As seen from table 8 the parametric as well as the non-parametric test indicates that there is a difference in synergies between the monthly periods. According to the Jarque–Bera4

Because a classification method based on the decile groups does not take into account cyclical effects I employed a new method base upon a three month rolling average (details are discussed in the second part of section III; Methodology: Classifying time patterns). A brief overview of the outcome using this classification method is presented in table 75_{. Each cell contains the number of} mergers and acquisitions. The cells colored light gray are months classified as low periods, whereas the cells colored dark gray are months classified as high periods. In total 22 low periods were selected and 13 high periods, containing 118 and 229 merger and acquisition announcements respectively.

Table 9: Results of the employed classification method; months classified as low have light gray colored cells and months classified as high dark gray colored cells.

Month 1 2 3 4 5 6 7 8 9 10 11 12 Ye ar 2000 4 3 2 16 7 8 5 3 3 6 4 1 62 2001 2 2 5 9 8 4 5 7 7 4 4 11 68 2002 7 8 7 6 9 9 11 5 13 11 13 12 111 2003 15 14 10 19 12 16 12 13 11 14 22 22 180 2004 15 19 26 16 13 18 17 17 19 19 12 22 213 2005 15 7 16 18 16 23 22 15 16 22 15 17 202 2006 9 16 26 23 29 21 21 16 19 25 32 16 253 2007 21 20 23 21 30 25 22 13 19 15 26 14 249 2008 8 13 17 11 17 17 15 9 19 3 9 14 152 2009 13 10 11 4 12 10 8 4 9 7 1 1 90 109 112 143 143 153 151 138 102 135 126 138 130 1580

In the following table (table 10) I present the results of the Student T-Test for the abnormal returns calculated using the mean adjusted and market and risk adjusted method.

Table 10: Student T-test P-Values for returns calculated using the mean adjusted and market and risk adjusted method. Event day -1 0 1 2 -1 to 2 Me an Ad ju st ed Target 0.0000 * 0.0000 * 0.0000 * 0.0000 * 0.0000 * Acquirer 0.3527 0.2436 0.7066 0.8231 0.6969 Ri sk a nd M ark et Ad ju st ed Target 0.0000 * 0.0000 * 0.0000 * 0.0000 * 0.0000 * Acquirer _{0.4544 0.2512 0.8310 0.9385 0.8568}

*significant at a 1% level; **significant at a 5% level; ***significant at a 10% level

For the target each day in the event period, as well as the whole event period is significant at a 1% level. In contrast to the returns of the acquirer that do not experience any significant abnormalities on the individual event days nor over the whole event period. This confirms the general belief that the shareholders of target firm benefit the most from a merger or acquisition.

I continue this section with the testing the first hypothesis: The synergies of a merger or acquisition

are lower in high frequency periods. I estimate an OLS regression controlling for the variables as

presented in second table from section three of this paper. The results concerning this regression's coefficients and their associated p-values are shown in the following table (table 11).

Table 11: Regressions explaining synergies as a function of the month's classification

Expected

Sign Coefficient P-value Explaining Low Period 0.0034 _0.6402 High Period -0.0013 _0.7873 Controlling Shares - 0.0042 _0.4201 Cash + 0.0140 0.0152 * Cross Border - 0.0115 0.0202 *

Civil to Common law + -0.0150 0.0500 **

Common to Civil law - -0.0112 _0.1394

Acquirer Debt + 0.0000 0.0002 * Target Debt - 0.0088 _0.1049 Firm size - -0.0222 0.0000 * C - 0.0083 _0.1986 Statistics R-squared _0.0272 R-squared Adjusted _0.0203 Jarque-Bera 2404.3842 Number of observations 1580

*significant at a 1% level; **significant at a 5% level; ***significant at a 10% level

The high and low period dummies are not significant in the regression. Though the coefficient of the high period dummy has a negative sign as hypothesized. Out of the dummies relating to the method of payment only cash has a significant effect on market perceived synergies. Cross border has a positive effect, not matching the prediction that it would lead to cultural mismatch which in turn decreases the synergies (Brock, 2005). Cross border deals that trigger a change legal system do have a significant negative effect on the synergies, especially those that force a company to change from civil to common. This would imply that even though companies in a common law system are valued higher (Anderson & Gupta, 2009) it does not outweigh the costs of a change in legal system. In line with the findings of Moeller et al. (2004) I find that large acquirers (market value of above 10 million) make significantly worse acquisitions compared to smaller acquirers.

Based on these results I am unable to reject ℎ0 of the first hypothesis. Meaning that the synergies

are not related to merger and acquisition frequency. Neither high nor low periods in merger and acquisition activity have any effect on the total synergy gains.

Table 12: Regression of the probability that 100% of the synergies are allocated to the target.

Expected

Sign Coefficient P-value Explaining Low Period 0.0057 _0.9650 High Period 0.1985 0.0386 ** Controlling Shares - 0.3662 0.0002 * Cash + 0.4927 0.0000 * Cross Border - -0.0071 _0.9483

Civil to Common law + -0.0563 _0.7970

Common to Civil law - 0.2451 _0.1592

Acquirer Debt + 0.0020 0.0000 * Target Debt - -0.1582 0.0981 *** Firm size + 0.1400 0.0962 *** C - -0.7406 0.0000 * Statistics R-squared _0.0272 R-squared Adjusted _0.0203 Number of observations 1580

*significant at a 1% level; **significant at a 5% level; ***significant at a 10% level

Months with a high level of merger and acquisition activity have, according to the results in table 12, a significant positive effect on the probability that all synergies are allocated to the target. This implies that managerial envy could play a part in explaining the occurrence of months with a high merger and acquisition frequency. Acquiring companies make worse acquisitions in high frequency months resulting in a loss of market value upon announcement. Among the control variables only those controlling for the effects of cross border mergers and acquisitions are not significant. Omitting these variables from the regression does not change the significance of the other control or explaining variables. The significant control variables all have the expected sign except for the Shares variable, there is no acceptable economic explanation for the positive sign of this variable. Omitting it from the regression does not change the significance of the other variables but does increase the constant.

is a higher probability they are all allocated to target. This is in line with the Goel and Thakor (2010) and suggests periods of high merger and acquisition activity are triggered by managerial envy. With this the ℎ0 of the second hypothesis is rejected, implying merger and acquisition frequency has

effect on the distribution of synergy gains in favor of the target.

VI. Conclusion

In this paper I analyze time patterns in merger and acquisition frequency by classifying months as low, normal or high frequency periods. To do this I employ a three month weighted rolling average that classifies the 120 monthly periods of my sample. This resulted in 21 low and 13 high periods of merger and acquisition activity. Based on prior research I formulate two hypotheses concerning the expected effects of high and low periods of merger and acquisition activity:

I. The synergies of a merger or acquisition are lower in high frequency periods.

II. In high frequency periods more than 100% of the synergies are allocated to the target.

I test each of these hypotheses using an regression and found the second hypothesis to be confirmed. The synergies were not affected by the period's classification. But the allocation of synergies in periods of high merger and acquisition activity were significantly biased towards targets, more often allocating all of the synergies to the target. Shareholders target companies should not be worried about the level of merger and acquisition activity. In contrast to those of the acquiring companies where periods of high merger and acquisition activity encourage managerial envy, causing companies to make worse acquisitions during such periods.

Based on the results as presented in this paper it is just to conclude that fluctuations in monthly merger frequency are caused by managerial envy as found by Goel and Thakor (2010). Managers engaging in a merger or acquisition during a period of high merger and acquisition activity are likely driven by envy instead of efficiency. The overall synergies of the deal are not affected but the allocation does change. Compared to normal periods the shareholders of the target company win and those of the acquirer lose value. Implying that the level of merger and acquisition activity does leave its mark on shareholder value. Shareholders of the acquiring company should therefore be skeptical about the value they gain due to the synergies from a merger or acquisition. There is a high probability the value of these synergies end up with the shareholders of the target company, especially in periods of high merger or acquisition activity.

other months in the sample. In this paper I only analyzed 10 years of merger and acquisition data, resulting in 10 observations of the average merger and acquisition frequency for each month. Due to this small amount of observations I was unable to test this hypothesis. A dataset containing more years of data can shed light on this phenomenon and possibly relate it to shareholder value.

VII. References

Anderson, A., & Gupta, P. P. (2009). A cross-country comparison of corporate governance and firm performance: Do financial structure and the legal system matter? Journal of Contemporary

Accounting & Economics , 5 (2), 61-79.

Bris, A., & Cabolis, C. (2008). The Value of Investor Protection: Firm Evidence from Cross-Border Mergers. Review of Financial Studies , 21 (2), 605-648.

Brock, D. M. (2005). Multinational acquisition integration: the role of national culture in creating synergies. International Business Review , 14 (3), 269-288.

Brown, S. J., & Warner, J. B. (1985). Using Daily Stock Returns: The Case of Event Studies. Journal of

Finacial Economics , 14 (1), 3-31.

Devos, E., Kadapakkam, P.-R., & Krishnamurthy, S. (2009). How Do Mergers Create Value? A Comparison of Taxes, Market Power, and Efficiency Improvements as Explanations for Synergies. The

Review of Financial Studies , 22 (3), 1179-1211.

Goel, A. M., & Thakor, A. V. (2010). Do Envious CEOs Cause Merger Waves? The Review of Financial

Studies , 23 (2), 487-517.

Golbe, D. L., & White, L. J. (1993). The Time Series Behavior of Mergers. The Review of Economics and

Statistics , 75 (3), 493-499.

Gort, M. (1969). An Economic Disturbance Theory of Mergers. The Quarterly Journal of Economics ,

18 (4), 624-642.

Harford, J. (2005). What drives merger waves? Journal of Financial Economics , 77 (3), 529–560. Jensen, M. C. (1986). Cost of free cash flow, corporate finance, and takeovers. The American

Economic Review , 76 (2), 323-329.

Michell, M. L., & Mulherin, J. H. (1996). The impact of industry shocks on takeover and restructuring activity. Journal of Financial Economics , 41 (2), 193-229.

Moeller, S. B., Schlingemann, F. P., & Stulz, R. M. (2004). Firm size and the gains from acquisitions.

Journal of Financial Economics , 73 (2), 201-228.

Rhodes-Kropf, M., & Viswanathan, S. (2004). Market Valuation and Merger Waves. The Journal of

Finance , 59 (6), 2685-2718.

Shleifer, A., & Vishny, R. W. (2003). Stock market driven acquisitions. Journal of Financial Economics ,

70 (3), 295-311.

Shughart, W. F., & Tollison, R. D. (1984). The Random Character of Merger Activity. The RAND Journal

of Economics , 15 (4), 500-509.

Travlos, N. G. (1987). Takeover Bids, Methods of Payment, and Bidding Firms' Stock Returns. The

Journal of Finance , 42 (4), 943-963.

VIII. Appendix

Table A1: List of classification method parameters. Where 𝑨𝒕 is the actual number of mergers and acquisitions, 𝑾𝒕 the

associated weight, 𝑹𝑨𝒕 the three month moving average and 𝑫𝒕 the deviation from that average.