FACULTY OF ECONOMICS AND BUSINESS
Herding Behavior in Merger Waves
An analysis of abnormal returns and the method of payment across
merger waves
Bachelor Thesis
July 1, 2014
First version
Name: Marijke van Ruijven
Student number: 10248595
Specialization: Finance and Organization
Field: Finance
Abstract
This research investigates whether herding behavior occurs in merger waves that took place between 2002 and 2012 and what the consequences are of this behavior. To test the occurrence of this behavior, the short – and long-term cumulative abnormal returns to bidders at the first stage for all mergers and acquisition of publicly traded U.S. companies that took place in a merger wave are compared to the ones at the second stage of a wave. Herding predicts that the abnormal returns are lower for mergers initiated in the second stage of a merger wave. In the short-term the
difference is indeed positive (0,112%), but in the long-term the difference is negative (-6,847%). By comparing the returns for the sample that only consists of bid
announcements that are financed by the most-commonly used payment method, the same results are found. This study finds evidence that herding behavior only occurs in the Retail Trade industry.
1. Introduction
Are mergers and acquisitions value-increasing or value-decreasing when they occur in waves? And is there a difference in returns between the two stages of a merger wave? That they occur in waves and that they cluster by industry (Mitchell and Mulherin, 1996) is, in fact, assumed by almost all researchers who study mergers and
acquisitions. But whether mergers increase or decrease value is a thing on which they have different meanings. You could say that there is nowadays a debate between value-increasing theories and value-destroying theories. The value-increasing theories can again be subdivided into other theories like the theory of efficiency, the market power theory and the theory of corporate control. The theory of efficiency suggest, for instance, that mergers will only occur when they are expected to generate enough realisable synergies to make the deal beneficial to both parties in a way that they both create positive returns (Weitzel and McCarthy, 2011).
The value-destroying theories can be divided into two different groups. The first group assumes that the management of the bidding firm could make mistakes because they are ‘boundedly rational’ (Weitzel and McCarthy, 2011). What in general are value-increasing purposes are now due to informational constraints purposes that could lead to the incurring of losses. The other group assumes managers who are rational, but self-serving. They therefore maximise their own private utility function, which will, in most of the cases, not be harmful to the firm value.
Most of the value-destroying theories will be explained in the next section, because they are an explanation for the fact that managers, who know that bids later in a wave leads to lower and/or negative returns, still choose to merge in a later stage of a wave. There are, in fact, some studies, like Floegel et al. (2005) and Harford (2003), that investigate the dynamics of a merger wave, and both find some evidence that at a later stage of a wave the abnormal returns are lower than at the beginning of a wave for both bidder and target. So it is strange that managers still undertake mergers in a later stage when they know this could decrease the value of the firm. There are however explanations for, and one of it is the hubris hypothesis of Roll. Roll (1986) suggests that the managers of the bidding firms overpay for their targets, because they have to much confident about the quality of their estimation of the targets’ values. Another explanation is the agency costs of free cash flows (Jensen, 1986).
The aim of this research is to identify another explanation, herding behavior, to answer the research question: ‘Is there a positive difference between the abnormal returns of a merger occurring at the first stage and a merger occurring at the second stage of a merger wave and how could this difference be explained?’ If this herding behavior occurs, than managers will ignore their private information about the attractiveness of other possibilities and they will just mimic the investment decision of other managers (Scharfstein and Stein, 1988). Suppose, for instance, a person on the street who has to decide in which of the two restaurants he is going to dine. When both restaurants look attractive, but empty because it is early in the evening, this person will random choose one restaurant. When other people pass by, they will choose the restaurant in which the first person is dinning because a restaurant that having customers makes it a better choice. This will go on, resulting in the restaurant that is chosen by the first person will do more business. This herding behavior will probably lead to successful mergers at the beginning of a wave, suggesting that there is a first-mover advantage. But for the reason that at the end the managers will ignore their own private information, the quality of the mergers occurring at the later stage of a wave will be worse.
To test this, I compare the short- and long-term abnormal returns for the two stages in a merger wave for all the mergers and acquisitions of publicly traded U.S. companies between January 2002 and December 2011. In the short-term, I find evidence that late wave bids have lower CARs, cumulative abnormal returns, (-0,013%) compared to the early wave bids (0,099%) for the whole sample. I find however no evidence for the whole sample that the late wave bidders perform worse than the early wave bidders in the long-term. The measurement of the CARs in the long-term with an event-window of (-1, +252) in the first stage and the late stage reveals in fact a negative difference of -6,847%.
The lower quality of the mergers occurring at the later stage of a wave may, however, also be due to other reasons like competition, managerial hubris and agency problems. I therefore look at the most commonly used payment method.Because when the most commonly used methods of payments are the same in both stages, this will suggest that the managers not only mimic the previous managers by initiate a merger, but also in the way to finance the deal. So this will give stronger evidence that herding behavior occurs because the second mover managers mimic the first-movers in two ways. To test this, I again, compare the short- and long-term abnormal
returns for the two stages in a merger wave for all the mergers and acquisitions of publicly traded U.S. companies between January 2002 and December 2011. But now I only use the mergers that are financed by the most commonly used payment method. I find the same results for this new sample. The same industries (Transportation & Public Utilities, Wholesale Trade and Retail Trade) have lower abnormal returns in the short-run at the late stage. And the same industries (Transportation & Public Utilities, Wholesale Trade and Finance, Insurance & Real Estate) have no lower abnormal returns at the late stage in the long run.
This research is organized as follows. Section 2 briefly reviews the most related literature. Section 3 presents the hypotheses and the establishment of them. Section 4 describes the date set, the process of identifying merger waves and the method how the research question will be answered. Section 5 presents the empirical tests and the results. And the last section will conclude, will discuss the findings, will describes the limitations of the research and will make some suggestions for further research.
2. Literature Review
2.1 Merger waves: drivers and consequences
There are different reasons provided by economic theory for why a merger might occur. Andrade et al. (2001) mentioned efficiency-related reasons as economies of scale and other synergies; the creation of market power by forming monopolies or oligopolies; market discipline; self-serving attempts; and the advantage of taking opportunities for diversification. Mitchell and Mulherin (1996) also tried to explain why a merger occurs based on the features that mergers occur in waves and that they strongly cluster by industry within a wave. They suggest that a merger might occur as a reaction to unexpected shocks to industry structure.
That mergers are preceded by technological or industrial shocks is also a conclusion that Martynova and Renneboog (2008) make based on the evidence they found. They also provide evidence that mergers occur in a positive economic and political environment, as in periods of economic recovery that correspondents with rapid credit expansion and stock market booms. Another finding is that mergers are driven by regulatory changes, such as anti-trust legislation and deregulation of
markets. They finally show that merger activity can also be influenced by managers’ personal objectives.
Martynova and Renneboog (2005) classified the theories about the drivers of merger waves into three groups. The first group consists of neoclassical models, which suggest that the waves emerge due to industrial, economic, political or
regulatory shocks. The second group suggests that they are caused by self-interested managerial decisions, predicated upon herding, hubris and agency problems. The last group proposes that waves occur as a result of (over)valuation-related timing by management. Shleifer and Vishny (2003) and Rhodes-Kropf and Viswanathan (2004) are examples of this last group: their theory suggests that waves are triggered by stock market overvaluation. Also Martynova and Renneboog (2008) state that there is growing evidence that overvaluation of the acquiring firms is an important
determinant of an increase in mergers, especially for mergers that are paid with equity or a combination of equity and cash.
Hence, there are several drivers of a merger wave, but what are the
consequences? As previously mentioned, there is a debate between value increasing- and value decreasing theories. Andrade et al. (2001), for instance, suggest that mergers seem to create value, with most of the gains acquiring to the target company. This correspondent to the evidence Jensen and Ruback (1983) find: corporate
takeovers generate positive gains that do not appear to come from the creation of market power; the shareholders of the target firm benefit, especially in the
announcement period (Jarrell, Brickley and Netter, 1988);and the shareholders of the bidding firm do not lose.
On the other hand, Banal-Estañol and Seldeslachts (2005) state that when the integration of different cultures is not successful, than this could lead to value-destroying mergers. They also find that merger failures are more likely when transaction costs are lower, which is the case in economic booms, because lower merging cost induce firms to merge more, but to integrate less. Duchin and Schmidt (2012) suggest that worse mergers may be caused by agency-driven behavior, which is encouraged by reduced monitoring and lower penalties for mergers that tend to be inefficient. The fact that managers maximize their own utility, and not those of the shareholders, is typically correlated with the firm size (Banal-Estañol and
This brings us to Moeller et al. (2003), who investigate the size effect and examine the cross-sectional variation in the announcement returns of mergers. They conclude that the size effect is more important than how a merger is financed or than what the organizational form of the acquired assets is. They also find evidence that small firms gain from mergers and large firms lose and that shareholders of small firms earn systematically more when the merger is announced. Martynova and Renneboog (2006) show that not only the size effect causes the estimated
shareholders wealth effect. It depends, for instance, also on the type of merger, the bid attitude, the payment method, the legal status of the target firm and the takeover strategy.
2.2 Wave effect
Not only the total wealth effect has been studied: there is also some research on the wave effect. Martynova and Renneboog (2006) for instance, show that mergers occurring at a later stage of the merger wave trigger lower gains to shareholders than mergers at the beginning of the wave. They demonstrate that in the wave of 2000-2001 the lowest 6-months CAARs1 are realized in mergers that occur at the end of the wave for both bidder and target firms. They also show that mergers undertaken at the late stage of the wave of 1990s destroyed bidders’ value and therefore conclude that the majority of value-destroying mergers occur in the second half of a merger wave. This conclusion is consistent with the findings of Moeller et al. (2005), who show a similar decline in merger profitability over the wave of 1990. Harford (2003) also did some research on the wave effect and find evidence that mergers occurring at the later stage of a merger wave trigger lower abnormal returns,2 than those at the beginning of
the wave.
That bidder gains fall during merger peaks is a conclusion of Shelton (2000), and he suggests that bidders tend to bid more aggressively, because they have greater tendencies to overpay for target firms and undertake more risky merger deals. That
1 Cumulative average abnormal returns for N securities over different event windows (from day t1 to
day t2) as follows:
2 An abnormal return reflects the unexpected future economic rents arising from the transaction. In
other words, an abnormal return of zero reflects a fair rate of return on the merger investment from the acquirer’s point of view (Andrade, Mitchell, and Stafford, 2001).
bidders often overpay for their targets at the end of a wave is also a result that Floegel et al. (2005) find. They explain this by saying that bidders want to avoid that one of their competitors buys the target and thereby gets a competitive advantage. They also find some evidence that managers of bidding firms become to overconfidence towards the end of a wave. Graham (1999) explains the difference in returns over the wave by the first-mover advantage theory. He says that the first one who merger presumably buys the best target, which will result in better short- and long-term returns. Even though, he mentioned that there is a distortional component in a merger wave because the managers do not stop at their optimal stopping point. The last firms that merge lose money suggest Persons and Warther (1997), so they conclude that merger waves will always end badly.
2.3 Explanations
So there are several researchers who conclude that mergers occurring at a later stage of a wave are worse than those at the beginning of the wave. But what are the explanations for this wave effect? As mentioned before, Floegel et al. (2005) show that bidders often rationally overpay for their targets at the end of a wave and do not find evidence for it at the beginning stage of a wave. So they suggest that rational overpayment is an explanation for the wave effect on abnormal returns. Martynova and Renneboog (2005) however, say that personal objectives of the managers will influence the merger activity. They show that managerial hubris and herding behavior increases towards the wave, which usually leads to worse mergers. They also
mentioned the limited information processing as an explanation for the wave effect. Herding (which will be explained in the next section), in combination with hubris or agency problems, is also an explanation Harford (2003, 2004) gives on his findings that mergers that occurring at a later stage of the merger wave trigger lower abnormal returns than those at the beginning of the wave. Managerial hubris is a common explanation for value-destroying mergers, because this behavior suggests that managers of the bidding firm are overconfident in their abilities to estimate the value of the target firm and, therefore, overestimate the creation of synergic value and eventually overpay their target (Roll, 1986). Roll (1986) explains this by suggesting that an individual manager will not abstain from bidding because he would have learned from his own past errors. He, therefore, may convince himself that his valuation is right, and that the market valuation is wrong, for the reason that the
market valuation does not reflect the true economic value of the merged firms (Roll, 1986). That managers may over-valuate the future results of the merged firms because they are overconfident about the synergy gains, is also a conclusion Banal-Estañol and Seldeslachts (2005) make, based on the fact that the managers may not foresee post-merger difficulties.
In order to come back to the other explanation for the existence of value destroying mergers, the agency problems, we need to look at the study of Jensen (1996). There is, in fact, a conflict of interest between shareholders (principles) and managers (agents) about pay out policies, in particular when the firm generates substantial free cash flow.3 The managers of those firms do not want to pay out these
cash flows to their shareholders, because this would decrease their control on
resources and so reduces their power (Jensen, 1996). They could increase their private benefit when they use the cash to finance mergers. The problem is, however, that the managers are self-interested, which will lead to investments below the cost of capital or a waste of cash on firm inefficiencies. This agency problem, therefore, implies that managers of firms with substantial free cash flows and unused borrowing power are more likely to attempt value-destroying mergers. Also Harford (1999) shows that bidders with large free cash flows generate significantly lower abnormal returns at the announcement of a merger. Duchin and Schmidt (2012) suggest that this agency-driven behavior, together with herding, could result in worse mergers. They indeed find evidence that the average long-term performance of mergers undertaken during merger waves is significantly worse than those initiated outside waves. They also looked at the corporate governance and find that it is weaker in mergers inside a wave, which suggest that agency problems exist in mergers occurring in merger waves.
Another explanation why managers initiated unsuccessful mergers is the possibility of ‘sharing the blame’ (Duchin and Schmidt, 2012). Decisions on mergers have, in fact, unpredictable components, which according to Duchin and Schmidt (2012) implies the possibility that ‘good’ managers may be unlucky. They also explain the lower returns in merger waves by the fact that mergers initiated inside a merger wave may be non-elective ‘mergers of necessity’, whereas the merger outside a wave may be elective. When mergers are undertaken by necessity, the higher levels
3 Free cash flow is the excess cash flow of that required to fund all projects that have positive net
present values when discounted at the relevant cost of capital, often generated by industrial shocks or by booming financial markets (Jensen, 1996).
of uncertainty are driven by industry shocks. This will create a new economic environment that triggers merger waves, but will give lower results.
From a different perspective, Gorton et al. (2009), show that
value-destroying mergers can precede a value-creating merger wave. Managers, in fact, use an active merger policy to protect themselves in order not to be taken because they prefer to keep their firm independent. When managers anticipate an effective merger wave in the near future, then they just want to remain independent. Therefore, they are going to use this active policy that is to a certain level inefficient, which will lead to an inefficient defensive merger wave.
2.4 Herding
There are several studies that mentioned herding behavior as an explanation for value-destroying mergers. Martynova and Renneboog (2008) suggest, for instance, that herding in combination with hubris predicts that efficient mergers are followed by inefficient ones. This suggestion is bases on the fact that herding predicts that managers tend to mimic the actions of a leader. The decision model analysed by Banerjee (1992) discuss that each decision maker takes into account the decision made by previous decision makers before taking its own decision, because previous decision makers have some information that could be important. Also Persons and Warther (1997) assume that the only way decision makers can learn about the quality of their investment is from the experience of early movers. Banerjee (1992) shows that herding behavior occurs in the decision-making process, so that the decision makers will do what others are doing rather than using their own private information.
Undertaking a merger in a merger wave is also a decision-making process and different researchers suggest that herding behavior will appear in this process.
Martynova and Renneboog (2008) are one of those researchers, and they show that the successful mergers at the beginning encourage other firms to do similar. So the managers mimic the action of the leader, who initiated a successful merger, instead of taking their actions based on clear economic rationale. The model of Persons and Warther (1997) predicts that this will continue until the experience of the mimicking managers is poor enough. This implies that managers who undertake a merger at the later stage of a wave will perform poorly, because they mimic a predecessor who already performs not very efficient (Persons and Warther, 1997), or because their private information, which they do not use, supports a non-merging strategy (Cabral,
2002).
Herding could be categorised into three different groups: informational cascades, reputational herding and investigative herding (Cabral, 2002). The basic cascade model of Devenow and Welch (1996) occurs when only actions are visible instead of private information and when there are finite limits to manager’s private information and possible actions. Managers will observe the decisions previously made by managers, and they will gain useful information from this. At some point, the observed information become so overwhelming that the manager’s private information is not strong enough to reverse the previously made decisions (Graham, 1999). The individual managers will, therefore, completely ignore their own private information and will mimic the actions of the crowd. Graham (1999) says that this creates a domino effect because this scenario also holds for anyone acting after the mimicking individual manager.
The second category, reputational herding, shows that managers have a desire to protect or signal their reputation (Devenow and Welch, 1996). When there is asymmetric information in the market, managers may prefer to ‘hide in the herd’ or to ‘ride in the herd’ in order to protect, signal or gain reputation. When they hide, the managers could not be evaluated, and when they ride, they could reveal and/or prove their quality (Devenow and Welch, 1996). So when managers acting as part of a group and, therefore, mimic their predecessors, they could obtain positively
reputational externalities (Graham, 1999). This reputational behavior could however lead to value-destroying mergers because the managers are only self-interested. In the study of Milbourn et al. (1999) they find that the CEOs of banks want to enhance their reputation, which could lead to herding behavior. To enhance their reputations they, in fact, increase the size and scope of the banks, not taking into account that this could waste shareholders’ wealth. Other CEOs in the rest of the industry will follow these actions because otherwise their reputation will not be enhanced (Milbourn, Boot and Thakor, 1999). That career concerns may trigger managers to follow their
predecessors and undertake mergers of worsening quality is also a conclusion Duchin and Schmidt (2012) make.
Investigative herding is the last category and occurs, according to Graham (1999), when a manager chooses to investigate something that he thinks other
managers will also investigate. The manager could only profit from this investigation if the other managers push the price of the assets in the manager’s anticipated
direction. So the first manager wants that other managers are going to herd, otherwise he may be stuck to an asset that he could not profitably sell. Thus, the conclusion is that the three categories of herding could lead to systematic wrong decision-making by entire populations because they make the sub-optimal choice instead of the best choice (Devenow and Welch, 1996).
Graham (1999) developed a model, which implies that when a manager has a high reputation; or low ability; or when there is strong public information that is inconsistent with the manager’s private information; or when the correlation of informative private signals across managers is positive, the manager is likely to herd. So when one of these conditions is met, there is according to the model of Graham (1999) an incentive for second-movers to mimic the leader and to do not use their own private information. Also Scharfstein and Stein (1990) developed a model, which explains that the inefficient behavior of managers could be rational in their point of view, since they are concerned about their reputation on the labour market. Because when a manager mimics the behavior of the previously decision makers, this will suggest that there is a positive correlation across the managers’ informative private signals. For that reason, the manager will more likely to be smart. A manager would, therefore, undertake an investment if others before him did it too, even though the investment has a negative return (Scharfstein and Stein, 1990), because otherwise he will be considered as dumb. This correspondent to the statement of Keynes (1936): “it is better for reputation to fail conventionally than to succeed unconventionally.” So Keynes already suggest that when managers are concerned about their reputation, they follow the herd.
3. Hypothesis
This research will investigate whether herding behavior occurs in merger waves and what the consequences are of this behavior. Banerjee (1992), Persons and Warther (1997) and Cabral (2002) all suggest that managers will not use their own private information in a decision-making process and that they mimic the action of the leader. This herding behavior will not always give negative results, because when the leader makes a good decision it is reasonable that the first mimicking manager will also make a good decision. However, when this behavior continues there is a high
probability that a turning point will occur that results in a destruction of the good decisions. This implies that managers who undertake a merger at a later stage of the wave will perform poorly, because they mimic a predecessor who already performs not very efficient, or because their private information, which they do not use,
supports a non-merging strategy. Although it is difficult to test for herding, I suggest, based on this implication, that herding behavior occurs when both the short-term and the long-term performance of mergers are worse when they are initiated at the late stage of a merger wave compared to those that are initiated at the first stage. I, therefore, formulate the following hypotheses:
Hypothesis 1: The short-term abnormal returns of a merger initiated at the second
stage of a merger wave are lower than those of a merger that occurred at the first stage of a wave.
Hypothesis 2: The long-term abnormal returns of a merger initiated at the second
stage of a merger wave are lower than those of a merger that occurred at the first stage of wave.
The first hypothesis is based on the conclusion that mergers occurring at a later stage of the merger wave trigger in the short-term lower gains to shareholders than mergers at the beginning of the wave, which is made by several researchers like Martynova and Renneboog (2006); Moeler et al. (2005), Shelton (2000); Harford (2003) and Floegel et al. (2005).
The second hypothesis is based on the intuition that when a manager want to make a prediction on whether his investment will give positive returns in the long run in order to decide whether to invest or not, he has to use his own private information. When he does not use it, there is more chance that his investment will not perform very well in the long run because he made the investment without taking into account his own analysis. Herding behavior causes that, at some point, managers do not us their own private information and I, therefore, suggest that this behavior occurs when the long-term performance are lower at the later stage of a merger wave.
By using both the short- and long-term returns, it will give more insight about the quality of a merger then using only one time period. There is also less chance that the long-term returns are influenced by unexpected events during the announcement
date compared to short-term returns. So by evaluating the short- and long-term returns I can make a better conclusion on whether the merger is value-destroying and whether herding behavior occurs.
The difference in abnormal returns at the first and the second stage in a merger wave are, however, also caused by other reasons as mentioned in the previous section. One of those reasons is for instance the concern of competition. In a merger wave there are, in fact, many companies that want to merge. Bidders want to avoid that one of their competitors buys the target and thereby gets a competitive advantage. Especially at the end of wave, in which there are less valuable targets. When two firms merge, a competitive advantage can be achieved because they together could get a greater market power (Feinberg, 1985). Firms with greater market power could charge higher price for their products/services, and they will, therefore, earn greater margins trough the appropriation of consumer surplus. To avoid that one of the bidders competitors gets a competitive advantage, they often overpay for their targets. This overpayment at the end of a wave reduces the gains to shareholders. So herding behavior is not the only possible explanation for the difference in the returns for mergers occurring at different stages of a merger wave.
I, therefore, will look at the method of payments. Because when the most commonly used methods of payments are the same in both stages, this will suggest that the managers not only mimic the previous managers by initiate a merger, but also in the way to finance the deal. So when the most commonly used method of payment in the first stage is the same as in the second stage, this will gives stronger evidence that herding behavior occurs because the second mover managers mimic the first-movers in two ways. When this behavior occurs, as mentioned before, the returns will be lower at the second stage of a merger wave. I, therefore, formulate the following hypothesis based on my own intuition:
Hypothesis 3: When the most commonly used payment method is the same at both
stages, the abnormal returns of mergers that are paid by that method will be lower when they are undertaken in the second stage.
4. Methodology
4.1 Data set
I will focus on all mergers and acquisitions of publicly traded U.S. companies in the 10-year period from January 2002 till December 2011. The reason is that the financial crisis occurred in this period. I, therefore, can study whether this crisis affects the number of initiated mergers.In previous studies, the researchers all say that merger waves tend to be 24-months long (Mitchell and Mulherin, 1996; Harford, 2003), so in this research merger waves will be 24-months long as well.
I will start looking for all merger and acquisition bids recorded by Thomson One, which previously had been known as SDC (Security Data Company) Platinum. My sample is restricted to the deals where:
- The minimum deal value is 50 Million Dollars. - The deal is completed.
- Before the announcement, all bidders had to be listed on the centre for research in security prices (CRSP).
This specification leads to 1286 bid announcements. I will assign each bid to their industry based on the SIC code recorded by Thomson One at the time of the announcement, in such way that I can assign each bid to 11 different industries. 4
These industries are; Agriculture, Forestry & Fishing; Mining; Construction; Manufacturing; Transportation & Public Utilities; Wholesale Trade; Retail Trade; Finance, Insurance & Real Estate; Services; Public Administration; and
Nonclassifiable Establishments.
4.2 Identification of merger waves
To investigate whether there is a merger wave, I have to calculate for each industry the highest 24-months concentration of merger bids. This highest 24-months
concentration can be seen as a potential wave. To test if this potential wave is really a merger wave, I will compare the highest months concentrations to the average 24-months concentrations. If the highest 24-24-months concentration deviates from the average less than the standard deviation, it is seen as no wave. Likewise, if the highest 24-months concentration deviates from the average less than 30% of the average, it is
seen as no wave as well.5 This results in 5 merger waves in 5 different industries that contain 184 bid announcements,which is shown in Table 1. The merger wave of the Construction industry can however not be used, because all the bid announcements are in the first year of the merger wave.
4.2.1 Analysis of the merger waves
The graphs of the course of the 24-months highest concentration of the eight different industries are shown in Appendix 2. When looking at the graphs, there is no peak visible in the Mining, the Construction, the Manufacturing and the Services industries, and you can only see some small fluctuations in the Mining and Services industries. This implies that there are no merger waves in those industries, and when looking at the criteria I used, there are indeed no merger waves in that period for those industries.
There are however large declines in all the industries, which could be explained by the financial crisis that starts in 2007. In fact, the 24-months
concentrations begin to fall in 2006, which takes into account the period from 2006 till 2008. So from this point you can analyse the effect of the crisis, and it indeed affects the number of mergers and acquisitions initiated. This decline has a lot of impact on the average. If I correct for the crisis, and only measure the average highest 24-months concentration until 2007, only the highest 24-months concentration of the Transportation & Public Utilities industry meets the criteria to be a merger wave. This is also shown in Figure 4, based on the fact that in the crisis the level of mergers declines to its initial level and decreases no further than this. This is in contrast to the other industries, in which the level of mergers decreases further than their initial levels.
After the drop caused by the financial crisis, you see for the Mining, the Manufacturing, the Wholesale trade, the Retail trade, the Finance, Insurance & Real Estate and the Services industry an increase in the 24-months concentration. This is probably due to the economy that is recovering and gradually picks up. This is consistent with the evidence Martynova and Renneboog (2008) find. They conclude that many mergers occur in a positive economic and political environment, as in periods of economic recovery that correspondents with rapid credit expansion and
5 This method differs from the most commonly used method by Harford (2003). He simulate the
test statistic 1000 times from the model with new random seed values and tests for significance based on the empirical distribution function of the simulated test statistics.
stock market booms. In the Retail trade industry the increase in the 24-months highest concentration is so large, that it could be seen as a merger wave. Another thing that is remarkable is the Finance, Insurance & Real Estate industry in which two peaks are shown. According to Harford (2003), it is however not possible to have two merger waves in one decade. I, therefore, used the peak with the highest concentration and denoted the second peak as ‘no merger wave’.
4.3 Method
To answer the research question I have to split the sample into two different time periods. I have to assign each announcement of a bid that take place in the first 12-months of an industry merger wave to the ‘early wave bids’ and the announcements made in the second 12-months to the ‘late wave bids’. To compare these two periods I have to look at the short and long-term abnormal returns. Like Floegel et al. (2005), I will calculate the cumulative abnormal return (CAR) for the short-term with a three-day event window around the announcement date (-1,1). To calculate this, I will use the following formula:
Whereas Rit the daily stock-return of firm i on date t is and Rmt the return for the
equally weighted CRSP index on date t. The daily stock-returns can me measured by the daily stock prices, which are available from the daily stock security files at CRPS in WRDS.6
To calculate the long-term CARs, I have to include the abnormal returns for one day prior to the announcement and the whole year after the announcement (-1, +252).
When I calculate the short-term and long-term CARs for the two different periods, I can compare the means to see if there is a difference in the first stage and the second stage of a merger wave. To test for significance of the means I can use a one sample t-test, whereas H0: µ = 0 and H1: µ ≠ 0. To test whether the differences between the first and the second stage of a merger wave are significant I can use a two-sample independent t-test with unequal variances, whereas H0: µEARLY - µLATE =
0 and H1: µEARLY - µLATE > 0. If the abnormal returns are lower at the second stage
than at the first stage, this will be a good sign that herding behavior occurs in a merger wave.
This sign is however not enough to make a good conclusion, as mentioned before, and I therefore have to look at the most commonly used payment method. According to Thomson One, the payments could be done by stock; by cash only7; by a combination of stock and cash (hybrid); and by other combinations. I have to look for the most commonly used payment method at the first stage of an identified merger wave in a specific industry and the one that is most commonly used at the later stage. When these payment methods are the same in both stages, I have to look at the cumulative abnormal returns of those mergers that are made by that payment method. Then I have to compare these returns of the different stages to see if there is a
difference. With this information I can verify or falsify my hypothesis. The same method is used to test for significance.
5. Results
5.1 Short-term performance
Based on the three-day event window, I find an average abnormal return of 0,43% to the bidder within a merger wave. Floegel et al. (2005), who did the same research, find an average abnormal return of 0,3559%. This difference can be explained by the smaller sample size I used. Other earlier studies, for instance, the one of Moeller et al. (2003), show a return of 1,1% and the research of Netter et al. (2011) reveal an
average of 1,02%. The different outcomes of those two studies compared to the one of Floegel et al. (2005) can be explained by the fact that Floegel et al. (2005) used a sample that only includes transactions with a minimum value of $100 million. The sample is, therefore, much smaller compared to the ones of Moeller et al. (2003) and Netter et al. (2011), which results in different findings. The difference between my research and the researches of Moeller et al. (2003) and Netter et al. (2011) can also
7 Transactions in which the only consideration offered is cash, earnout or assumption of liabilities, or
any combination of the three
be explained by this phenomenon. I set, in fact, many more restrictions on the sample, which is therefore much smaller, and gives other results. The difference between the results of Martynova and Renneboog (2006), who find an average abnormal return of 0,5% to the bidders announcements, and my research, can partly be explained by the fact that they used a one-day event window instead of the three-day event window I used.
After splitting the sample into two time periods (early wave bids and late wave bids), new information reveals: the late wave bids have lower average abnormal returns (-0,013%) compared to the early wave bids (0,099%). These returns and the differences are however not significant, probably caused by the small sample. In Table 2 all results, including the tests for significance, are summarized. It is shown that mergers that are initiated at the late stage of a merger wave in 3 out of 4
industries perform worse. This wave effect is also be found by Floegel et al. (2005), who split the sample into two time periods as well. They find an average abnormal return of 1,5562% for announcements at the early wave stage and an average
abnormal announcement return of -1,1079% for late wave bidders. The difference is much greater than my results and is highly significant, which is probably due to the bigger sample size.
5.2 Long-term performance
Based on the one-year event window, I find an average abnormal return of -8,184% to bidding firms (-8,824% if the sample is corrected for missing values). This is
consistent with the value-destroying theories and the behavioral hypothesis of Harford (2004), which implies that long-term returns should be poor following merger waves. Harford (2004) mention that other earlier studies also find evidence of poor long-term performance of bidders. There are, however, studies that do not find evidence of this underperformance in the long run. Moeller et al. (2003), for instance, find an
insignificant abnormal return of -0,041% for the whole sample. They explain the different conclusions by the sensitivity of the estimation method. There are, in fact, several methods, for example the event-time approach and de calendar-time portfolio approach, to estimate the long-term returns (Moeller, Schlingemann and Stulz, 2003). These different approaches go along with different outcomes and conclusions.
Besides the average abnormal return of the whole sample, I also measured the returns for the split sample. I find an average abnormal return of -11,587%
(-13,067%) for early wave bidders and -4,740% (-4,725%) for late wave bidders. These differences are however not significant. In Table 3 all results, including the tests for significance, are summarized. It is shown that mergers that are initiated at the late stage of a merger wave in 3 out of 4 industries perform not worse than those initiated at the early stage. This is inconsistent with my hypothesis and previous studies and could be explained by several things. The first one is the sensitivity of the estimation method as mentioned above. The second one is the sample size, which is much smaller in my research. The last one is the sample itself, because I put some other restrictions on it, and I used another time-period. This part will be further explained in the section ‘limitations’.
5.3 The method of payment
Under the behavioural hypothesis of Harford (2004), the only reason why a merger wave occurs is the fact that managers use overvalued stock to acquire assets of a less overvalued firm. If the managers are better informed about the long-term performance of their firm than the market and think that their stock is overvalued, they will use stock to pay for their acquisitions (Rau and Vermaelen, 1998). The payment by stock also signals that the shareholders of the bidding firm want to keep the shareholders of the target firm involved and that they will share the risk together (Martynova and Renneboog, 2006). It is, therefore, an important signal of quality of the target firm, because it suggests that the target firm may be less valuable. On the other hand, mergers and acquisition that are paid by cash will signal that the target firm is valuable because the managers want to pay off the shareholders of the target firm in order to not share future benefits (Martynova and Renneboog, 2006). This suggests that the long-term abnormal returns to bidders will be negative for mergers that are financed by stock and positive for those that are paid by cash (Rau and Vermaelen, 1998).
There are indeed several studies that find evidence for this. Martynova and Renneboog for instance, report that cash only bids, as well as hybrid bids, has
substantially higher abnormal returns than stock bids. This is consistent with the study of Loughran and Vijh (1997), which concludes that firms experience long-run
underperformance when they use stock as their method of payment, and the study of Travlos (1987), which shows that stock mergers will have negative short-term
announcement returns and cash mergers will have normal returns at the announcement period.
In my research, I first investigate which method of payment is most commonly used in both stages of the merger wave. I find that in the industries Transportation & Public Utilities, Wholesale Trade and Retail Trade for both stages the most commonly used method op payment is cash and for Finance, Insurance & Real Estate it is hybrid. This suggests that managers mimicking their processors in two ways, because they follow the herd by initiate a merger and by using the same method to finance the deal. This will predict that the new sample, which only consist of bid announcements that are financed by the most-commonly used method for that industry, will have lower short- and long-term abnormal returns at the late stage compared to the first stage because herding behavior occurs.
Based on the same methodology of the previous sections and the new sample, I find that in the short run 3 out of 4 industries have lower abnormal returns at the second stage of a merger wave. These industries are Transportation & Public Utilities, Wholesale Trade and Retail Trade and the difference are respectively 3,676%,
3,168% and 3,836%. In Table 4 Panel A all results, including the tests for significance, are summarized. The differences between the first stage and the second stage are however not significant, probably due the small sample size.
In the long run, however, only the Retail Trade industry has lower abnormal returns at the second stage. This is probably due to the methodology I used, which is the same as in the previous section, and therefore this inconsistent evidence could be explained by the same reasons as mentioned in the previous section. This could also explain why the industries Transportation & Public Utilities, Retail Trade, and Finance, Insurance & Real Estate overall have negative abnormal returns in the long run instead of positive returns, which is predicted by previous studies as mentioned at the beginning of this section. The results for the long-term, including the tests for significance, are summarized in Panel B of Table 4.
6.
Conclusion
that when this behavior occurs, the performance of mergers initiated at the late stage of a merger wave will be worse. There are several reasons why herding behavior occurs, for instance by the concern of the manager’s reputation in the labour market or in informational cascades, in which managers cannot deal with the overwhelming existing information. This suggests that managers will not use their own private information and that they mimic the action of the leader. This implies that managers who undertake a merger at a later stage of the wave will perform poorly, because they mimic a predecessor who already performs not very efficient, or because their private information, which they do not use, supports a non-merging strategy.
This research finds, however, no evidence that this behavior occurs in all of the merger waves between 2002 and 2011. Although the short-term abnormal returns are lower at the late stage for 3 out of 4 industry merger waves, the long-term
abnormal returns are not. I find instead evidence that the long-term abnormal returns are higher at the late stage, which is contrary to what I expected and what herding behavior predicts. Only the Retail Trade industry has both in the short- and long-term lower abnormal returns at the late stage of the wave compared to the first stage.
When analysing the new sample that consists only the bid announcements that are financed by the most commonly used method of payment, I find the same results. The same industries (Transportation & Public Utilities, Wholesale Trade and Retail Trade) have lower abnormal returns in the short-run at the late stage. And the same industries (Transportation & Public Utilities, Wholesale Trade and Finance, Insurance & Real Estate) have no lower abnormal returns at the late stage in the long run. Based on these findings I can conclude that herding behavior only occurs in the Retail Trade industry. This industry has at the early stage a short-term abnormal return to bidders of 4,405% and at the late stage 1,185%. In the long-term, the abnormal returns are respectively for the early and late stage, -5,127% and -14,632%. For the new sample, which only consists of bid announcements that are financed by cash, the short-term abnormal return to bidders at the early stage is 4,541% and 0,705% at the late stage. The longterm abnormal return to bidders is 4,271% for the early wave bids and -19,44% for the late wave bids. The differences between the two stages are however not significant, so I can actually draw no conclusions on these findings. This brings us to the limitations of this research, which will be discussed in the next section.
6.1 Limitations and future research
As mentioned before, most of the results are not significant. This is probably due to the small sample size I used. Eventually, I investigate 181 bid announcements, which is much smaller compared to the 12023 transactions of Moeller et al. (2003), the 5278 mergers of Martynova and Renneboog (2006) and the 1025 bid announcements of Floegel et al. (2005). This smaller sample can be explained by the fact that I put some extra requirements on it. The restriction that deviates the most from other studies is the one that says that both the bidder and the target are U.S. publicly traded firms. In other studies, they put, most of the time, only the restriction that the bidder is a publicly traded firm and that the target firm could be public, private of subsidiary. Using only publicly traded firms reduce the sample almost to the half.
Another limitation occurs in the assignment of the bid announcements to the different industries. In most other studies, for example the ones of Harford (2003) and Floegel et al. (2005), they assign the bid announcements to one of the 48 industries defined by Fama and French (1997). I did not use this division of the industry and assigned to 11 different industries based on two-digit SIC codes provided by McKimmon Center for Extension & Continuing Eduction. This results in other, bigger industries, which makes it less precise.
Another thing that I did not do compared to other researchers, is the method to test whether the potential wave is indeed a merger wave. They commonly use the method provided by Harford (2003), in which the test statistic is simulated 1000 times from the model with new random seed values and the tests for significance is based on the empirical distribution function of these simulated test statistics. I used the criteria that if the highest 24-months concentration deviates from the average less than 30% of the average, it is seen as no wave.
The last limitation of this research is that I did not correct the sample for buybacks and bids made by the same bidder that occur within less than 300 trading days. When not excluding those bids, it can cause biases in the estimation of the parameters according to Martynova and Renneboog (2006). Floegel et al. (2005), however, say that the modified market model I used to estimate the abnormal returns to bidders, already correct for these frequent bidders. So that is why I did not exclude them.
I only looked at the abnormal returns to bidders and did not measure and analyze those to targets. Herding behavior could however occur on every part of the
population. To test whether it occurs on the target firms is, therefore, my suggestion for further research. Another suggestion is to investigate the returns outside the
merger waves for the period I tested, because then it is possible to compare mergers in and outside a wave. This will give more inside about the impact of a merger wave.
References
Andrade, G., Mitchell, M., Stafford, E., (2001). New evidence and perspectives on mergers. Journal of Economic Perspectives 15, 103–120.
Banal-Estañol, A., and Seldeslachts, J., (2005). Merger Failures. Discussion papers //
WZB, Wissenschaftszentrum Berlin für Sozialforschung,
Forschungsschwerpunkt Markt und Politische Ökonomie, Abteilung Wettbewerbsfähigkeit und industrieller Wandel SP II 2005-09.
Banerjee, A.V., (1992). A Simple Model of Herd Behavior, The quarterly Journal of
Economics 107(3), 797-817.
Cabral, I., (2002). A Herding Approach to Merger Waves. European University
Institute EUI ECO 26.
Devenow, A., and Welch, I., (1996). Rational Herding in Financial Economics,
European Economic Review 40, 603-615.
Duchin, R., and Schmidt, B., (2012). Riding the Merger Wave: Uncertainty,
Reduces Monitoring, and Bad Acquisitions. Journal of Financial Economics,
forthcoming.
Fama, E.F., and French, K., (1997). Industry Costs of Equity. Journal of Financial
Economics 43, 153–193.
Feinberg, R.M., (1985). Sales at Risk: A Test of the Mutual Forbearance theory of Conglomerate Behaviour, Journal of Business 58, 225-241.
Floegel, V., Gebken, T., and Johanning, L., (2005). The Dynamics within Merger Waves – Evidence from industry Merger waves of the 1990s -. Union
PanAgora Asset Managemnt, WHU - Otto Beisheim School of Management and Source For Alpha AG.
Gorton, G., Kahl, M., and Rosen, R.J., (2009). Eat or Be Eaten: A Theory of Mergers and Firms Size. Journal of Finance 64(3), 1291-1344.
Graham, J., (1999). Herding among Investment Newsletters: Theory and Evidence.
Journal of Finance 54(1), 237-268.
4, J., (1999). Corporate Case Reserves and Acquisitions. Journal of Finance
54(6),1969-1997
Harford, J., (2003). Efficient and Distortional Components to Industry Merger Waves. AFA 2004 San Diego Meetings.
Harford, J., (2004). What drives merger waves? Journal of Financial Economics 77, 529–560.
Jarell, G.A., Brickley, J.A., and Netter, J.M., (1988). The Market for Corporate Control: The Empirical Evidence Since 1980. Journal of Economic
Perspectives 2(1), 49-68.
Jensen, M.C., and Rubakc, R.S., (1983). The Market for Corporate Control: The Scientific Evidence. Journal of Finance 11, 5-50.
Jensen, M., (1986). Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers. American Economic Review 76 (2), 323-329.
Jensen, M., (2004). Agency Costs of Overvalued Equity. Harvard NOM Working
Paper 26; ECGI Finance Working Paper 39.
Keynes, J.M., (1936). The General Theory of Employment, Interest and Money
Macmillan, London.
Loughran, T., and Vijh, A.M., (1997). Do Long-Term Shareholders Benefit From Corporate Acquisitions? Journal of Finance 52(5), 1765-1790.
Martynova, M., and Renneboog, L., 2005, Takeover Waves: Triggers, Performance and Motives. ECGI - Finance Working Paper 97.
Martynova, M., and Renneboog, L., 2006, The Performance of the European Market for Corporate Control: Evidence from the 5th Takeover Wave. ECGI – Finance
Working Paper 135.
Martynova, M., and Renneboog, L., (2008). A century of corporate takeovers: What have we learned and where do we stand? Journal of Banking &
Finance 97.
Moeller, S.B., Schlingemann, F.P., and Stulz, R.M., (2003). Firm Size and the Gains from Acquisitions, Journal of Financial Economics 73, 201-228.
Moeller, S.B., Schlingemann, F.P., and Stulz, R.M., (2005). Wealth Destruction on a Massive Scale? A Study of Acquiring-Firm returns in the Recent Merger Wave. Journal of Finance 60(2), 757-782.
Milbourn, T.T., Boot, A.W.A., and Thakor, A.V., (1999). Megamergers and expanded scope: Theories of bank size and activity diversity. Journal of Banking &
Finance 23, 195-214.
Mitchell, M.L., and Mulherin, J.H., (1996). The impact of industry shocks on takeover and restructuring activity. Journal of Financial Economics 41(2), 193-229.
Netter, J., Stegemoller, M., and Wintoki, M.B., (2011). The implications of Data Screens on Mergers and Acquisitions Analysis: A Large Sample Studsy of Mergers and Acquisitions from 1992 to 2009. The Review of Financial
Studies 24(7), 2316-2357
Persons, J.C., and Warther, V.A., (1997). Boom and Bust Patterns in the Adoption of Financial Innovations. The Review of Financial Studies 10(4), 939-967. Rau, P.R., and Vermaelen, T., (1998). Glamour, Value and the Post-Acquisition
Performance of Acquiring Firms. Journal of Financial Economics 49(2), 223-253.
Rhodes-Kropf, M., and Viswanathan, S., (2004). Market valuation and merger waves.
Journal of Finance 59(6), 2685-2718.
Roll, R., (1986). The Hubris Hypothesis of Corporate Takeovers. Journal of Business
59, 197- 216.
Scharfstein, D., and Stein, J., (1990). Herd behavior and investment. American
Economic Review 80, 465–479.
Shelton, L.M., (2000). Merger Market Dynamics: Insight into the Behaviour of Target and Bidder Firms. Journal of Economic Behavior and Organization 41(4), 363-383.
Shleifer, A., and Vishny, R.M., (2003). Stock market driven acquisitions, Journal of
Financial Economics 70, 295-312.
Travlos, Nicolas G., (1987) Corporate Takeover Bids, Methods of Payment, and Bidding Firms' Stock Returns. Journal of Finance 42, 943-963.
Weitzel, U., and McCarthy, K.J., (2011). Theory and evidence on mergers and acquisitions by small and medium enterprises. International Journal of
Table 1
Industry Merger Waves
The table contains all industry merger waves of the U.S. that took place between 2002 and 2012. The first column represents the 11 different industries based on the two-digit SIC code.8 The second shows whether it is a merger wave or not. The third reports the number of bid announcements in the industry merger wave, which is the highest 24-months concentration. The fourth column shows the average 24-months concentration for a particular industry. The fifth column displays the difference between the highest 24-months concentration and the average of the 24-months concentrations multiplied by 30% of the average. The sixth reports standard deviation and the last column shows the starting dates of the industry merger waves.
Industry Wave identification Number of announcement bids in wave* Average 24-months concentration ** Standard Deviation Starting date
Agriculture, Forestry & Fishing
No Wave (too few announcement bids)
2 - - - -
Mining NoWave 13 10,167 -0,2171 1,483 -
Construction Wave 3 1,875 0,5625 0,641 7 Jan 2002 Manufacturing No Wave 92 72,928 -2,8064 13,565 -
Transportation & Public Utilities
Wave
32 20,704 5,0484 6,119 14 Dec 2004 Wholesale Trade Wave 10 6,318 1,7866 1,81 5 Aug 2009 Retail Trade Wave 19 14,157 0,5959 3,646 1 Jul 2003 Finance, Insurance &
Real Estate
Wave 120 88,058 5,5246 10,648 12 Mar 2003 No wave9 118 88,058 3,5246 10,648 -
Services No Wave 76 63,147 -6,0911 26,387 - Public Administration No Wave (No
announcement bids) 0 - - - - Nonclassifiable Establishments. No Wave (No announcement bids) 0 - - - -
* = The highest 24-months concentration
** = The number of announcement bids in wave minus (1,30 * the average 24-months concentration)
8 See appendix 1.
9 No merger wave, because according to Harford (2003) it is not possible to have two merger waves in
Table 2
Short-term Cumulative Abnormal Returns to bidders in different stages of industry merger waves
The table contains all industry merger waves of the U.S. that took place between 2002 and 2012. The first row represents the 4 different industries based on the two-digit SIC code, in which a merger wave occurs. These industries: Transportation & Public Utilities, Wholesale Trade, Retail Trade, and Finance, Insurance & Real Estate are denoted by (1), (2), (3) and (4) respectively. Column (5) represents the total. The CAR denotes the cumulative abnormal return to bidders for a three-day event window (-1, +1) around the announcement date. ‘Early’ states for bid announcement that take place in the first 12-months of an industry merger wave and ‘Late’ for those that are made in the second 12-months. ‘Difference’ shows the difference between the early- and late stage. Standard deviations are in parentheses. Respectively *, **,
and *** indicate statistical significance at 1%, 5% and 10% level.
(1) (2) (3) (4) (5) CAR .02030** .00155 .02710*** -.00918** .00043 (.05043) (.02831) (.06144) (.04502) (.04902) Early .02895*** .02328 .04405 -.01243** .00099 (.05975) (.03678) (.07808) (.04914) (.05722) Late .01318 -.00776 .01185 -.00545 -.00013 (.04182) (.02035) (.03980) (.03990) (.03947) Difference .01577 .03104 .03220 -.00699 .00112 Observations 31 10 19 118 178 Early 14 3 9 63 89 Late 17 7 10 55 89
Table 3
Long-term Cumulative Abnormal Returns to bidders in different stages of industry merger waves
The table contains all industry merger waves of the U.S. that took place between 2002 and 2012. The first row represents the 4 different industries based on the two-digit SIC code, in which a merger wave occurs. These industries: Transportation & Public Utilities, Wholesale Trade, Retail Trade, and Finance, Insurance & Real Estate are denoted by (1), (2), (3) and (4) respectively. Column (6) represents the total. Not all the long-term returns are available for the Finance, Insurance & real Estate industry, so there is added an extra column that is corrected for those missing values. In this column the bid announcements for which no one-year event window returns measurement is possible are deleted and this column is denoted by (5). Column (7) represents the corrected total. The CAR denotes the cumulative abnormal return to bidders for a one-year event window (-1, +252) around the announcement date. ‘Early’ states for bid announcements that take place in the first 12-months of an industry merger wave and ‘Late’ for those that are made in the second 12-months. ‘Diff’ shows the difference between the early- and late stage. Standard deviations are in parentheses. Respectively *, **, and *** indicate statistical significance at 1%, 5% and 10% level.
(1) (2) (3) (4) (5) (6) (7) CAR .03225 .12465 -.10129 -.10893* -.12024* -.08164* -.08824* (.34318) (.21999) (.32445) (.22451) (.22108) (.26423) (.26421) Early -.06322 -.11221 -.05127 -.13698* -.15973* -.11587* -.13067* (.21258) (.10724) (.41790) (.23874) (.22837) (.25330) (.24890) Late -.00675 .22616** -.14632*** -.07680* -.07711* -.04740 -.04725 (.42729) (.16937) (.22482) (.20445) (.20635) (.27184) (-.78570) Diff. -.05647 -.33837 .09505 -.06018 -.08262 -.06847 -.08342 Obs. 31 10 19 118 113 178 173 Early 14 3 9 63 59 89 85 Late 17 7 10 55 54 89 88
Table 4
Short- and Long-term Cumulative Abnormal Returns to bidders in different stages of industry merger waves
The table contains all bid announcements of the U.S. that are financed by the most commonly used method of payment for that specific industry and that took place between 2002 and 2012. Panel A shows the average short-term CAR, which denotes the cumulative abnormal return to bidders for a three-day event window (-1, +1) around the announcement date. Panel B shows the long-term cumulative abnormal return to bidders for a one-year event window (-1, +252) around the announcement date. The first row, in both Panel A and B, represents the 4 different industries based on the two-digit SIC code, in which a merger wave occurs. These industries: Transportation & Public Utilities, Wholesale Trade, Retail Trade, and Finance, Insurance & Real Estate are denoted by (1), (2), (3) and (4) respectively. ‘Early’ states for bid announcements that take place in the first 12-months of an industry merger wave and ‘Late’ for those that are made in the second 12-months. ‘Difference’ shows the difference between the early- and late stage. Standard deviations are in parentheses. Respectively *, **, and ***
indicate statistical significance at 1%, 5% and 10% level. Not all the long-term returns are available for the Finance, Insurance & real Estate industry, so there is added an extra column in Panel B that corrects for those missing values. In this column the bid announcements for which no one-year event window returns measurement is possible are deleted and this column is denoted by (5).
Panel A: Short-term Cumulative Abnormal Returns (-1, +1)
(1) (2) (3) (4) CAR .02397*** .00216 .02495 -.01091 (.05722) (.02996) (.06402) (.05030) Early .04718 .02328 .04541 -.01323 (07905) (.03678) (.08366) (.06368) Late .01042 -.00841 .00705 -.00799 (.03750) (.02222) (.03759) (.02622) Difference .03676 .03168 .03836 -.00524 Observations 19 9 15 52 Early 7 3 7 29 Late 12 6 8 23
Panel B: Long-term Cumulative Abnormal Returns (-1, +252)
(1) (2) (3) (4) (5) CAR -.08206 .08214 -.12361 -.14954* -.17422* (.36148) (.18469) (.33445) (.20268) (.18642) Early -.10721 -.11221 -.04271 -.15639* -.19931* (.22664) (.10724) (.47321) (.25956) (.23655) Late -.06739 .17931** -.19440* -.14090* -.14456* (.43031) (.12642) (.13964) (.09667) (.09730) Difference -.03982 -.29151 .15168 -.01549 -.05475 Observations 19 9 15 52 48 Early 7 3 7 29 26 Late 12 6 8 23 22
Appendix 1
The two-digit SIC codes provided by McKimmon Center for Extension & Continuing Eduction.
A. Agriculture, Forestry, & Fishing
01 Agricultural Production Crops
02 Agriculture production livestock and animal specialties 07 Agricultural Services
08 Forestry
09 Fishing, hunting, and trapping
B. Mining
10 Metal Mining 12 Coal Mining
13 Oil And Gas Extraction
14 Mining And Quarrying Of Nonmetallic Minerals, Except Fuels
C. Construction
15 Building Construction General Contractors And Operative Builders 16 Heavy Construction Other Than Building Construction Contractors 17 Construction Special Trade Contractors
D. Manufacturing
20 Food And Kindred Products 21 Tobacco Products
22 Textile Mill Products
23 Apparel And Other Finished Products Made From Fabrics And Similar Materials 24 Lumber And Wood Products, Except Furniture
25 Furniture And Fixtures 26 Paper And Allied Products
27 Printing, Publishing, And Allied Industries 28 Chemicals And Allied Products
29 Petroleum Refining And Related Industries 30 Rubber And Miscellaneous Plastics Products 31 Leather And Leather Products
32 Stone, Clay, Glass, And Concrete Products 33 Primary Metal Industries
34 Fabricated Metal Products, Except Machinery And Transportation Equipment 35 Industrial And Commercial Machinery And Computer Equipment
36 Electronic And Other Electrical Equipment And Components, Except Computer Equipment 37 Transportation Equipment
38 Measuring, Analyzing, And Controlling Instruments; Photographic, Medical And Optical Goods 39 Miscellaneous Manufacturing Industries
E. Transportation & Public Utilities
40 Railroad Transportation
41 Local And Suburban Transit And Interurban Highway Passenger Transportation 42 Motor Freight Transportation And Warehousing
43 United States Postal Service 44 Water Transportation 45 Transportation By Air
46 Pipelines, Except Natural Gas 47 Transportation Services
