Synergies and announcement return
An empirical analysis of the relationship between cost and revenue synergies and the return for the shareholders of the acquirer.
Robert van den Breemer
July 2008
University of Groningen, Faculty of Economics and Business MSc Business Administration
Synergies and announcement return
An empirical analysis of the relationship between cost and revenue synergies and the return for the shareholders of the acquirer.
Student Robert van den Breemer
s1333836
Blasiusstraat 15-IV 1091 CJ Amsterdam S1333836@student.rug.nl
University University of Groningen
Faculty of Economics and Business
Educational program MSc Business Administration, Specialization Corporate Finance
Company Duff & Phelps LLC
University supervisors Dr. J.H. von Eije (first supervisor) Dr. L. Dam (second supervisor)
Company supervisors Drs. R. Meiberg Drs. A. Bregonje
Synergies and announcement return*
An empirical analysis of the relationship between cost and revenue synergies and the return for the shareholders of the acquirer.
Abstract
This research studies the effects of synergies on the return to the acquiring shareholder. A distinction is made between revenue and cost synergies. Using a sample of 243 acquisitions, the transcripts of analyst conference calls are examined for information on the expected synergies arising from the transaction. Since revenue synergies are generally viewed as being more unique to a particular buyer than cost synergies, it is likely that the acquisition premium paid to the target reflects the full amount of expected cost synergies and only a part of the revenue synergies. The remainder of the revenue synergies can therefore be a return to the acquiring shareholder. This research examines the rumor, the announcement and the completion return. It is hypothesized that transactions that expect revenue synergies will generate a higher announcement return than transactions that present only cost synergies. The results, however, do not suggest such a relationship and do not provide the significant results to be able to accept the hypothesis.
JEL classification: G14, G34
Keywords: cost and revenue synergies, shareholders of the acquirer, announcement return, event study
Table of contents Abstract...3 Table of contents...4 Introduction ...5 I. Background...8 A. Synergies ...8
A1. Distribution of synergies ...9
B. Return for the acquiring shareholders ...11
C. Acquisitions...13 C1. Premium ...14 D. Hypothesis ...15 II. Data ...16 A. Sample selection ...16 B. Other information...16 III. Methodology...18 A. Event study ...18 B. Univariate analysis ...21
C. Multivariate regression analysis...22
IV. Results...25
A. Results of the event study...25
B. Results of the univariate analysis...26
C. Results of the multivariate analysis...29
V. Conclusion...32
References...35
Appendix I Cost synergies ...37
Appendix II Revenue synergies ...42
Appendix III Acquisitions in sample...46
Appendix IV Results of the event study...50
Appendix V Results of the univariate analysis...52
Introduction
The merger and acquisition activity has erupted in the late nineties. This explosion of transactions is mainly due to the profound conviction that corporate takeovers create value. Jensen and Ruback (1983) explained that acquisitions, in general, create economic value. This evidence is consistent with the synergy theory of acquisitions, presented by Bradley, Desai and Kim (1988). The synergy theory posits that the acquisition of control over the target enables the acquirer to redeploy the combined assets of the two firms toward higher-valued uses. Seth, Song and Pettit (2000) use the synergy theory to explain that acquisitions take place when the value of the combined firm is greater than the sum of the values of the individual firms.
A number of opportunities can be defined that contribute to this creation of synergistic value. The first opportunity, presented by Damodaran (2005), is called financial synergy. Financial synergies are the result of a reduction of the cost of capital of the combined firm. Operational synergies arise from combining operations of hitherto separate units and knowledge transfers. These operational synergies are suggested by Porter (1985). The third opportunity for the creation of synergistic value is called collusive synergies. These arise from increased market power of the combined company. Collusive synergy is found to be, on average, associated with the highest value according to Chatterjee (1986). Trautwein (1990) later added managerial synergies that arise when the bidder’s managers possess superior planning and monitoring abilities that benefit the target’s performance.
Synergies in general have been said to be one of the main reasons for acquisitions to take place according to Goergen and Renneboog (2004). Firms use synergies as a reason to acquire and to explain the acquisition to their shareholders. Shareholders take notion of the expected synergies and evaluate these opportunities. Moreover, if the synergies are a reason for the firm to acquire, shareholders of the acquiring firm might show a positive reaction to these synergies. So by that rationale, if the amount of synergies increases the reaction of the shareholders should become more positive as well.
However, do shareholders actually take the amount of synergies into consideration when they judge the announcement of an acquisition of their company? Management can be very enthusiastic about a certain deal. According to them the deal might provide them a unique opportunity. However, despite the conviction of the management, shareholders might judge this opportunity differently. This might have to do with the premium an acquirer pays to the shareholders of the target. Perhaps all the synergies are included in the premium and can therefore not be seen as a gain or opportunity for the acquiring company. Perhaps also the type of acquisition is a reason for shareholders to value synergies differently. It might be that acquisitions in the same sector or country are much more appreciated than acquisitions in a foreign county or a different sector. Or perhaps some types of synergies are indeed included in the acquisition premium while some other types of synergies are not.
announcement. Next to the announcement date, the rumor date and the date of completion are used in this research. The synergies are split in cost-based and revenue-based synergies. Capron (1999) uses this distinction, though the description that focuses on the areas of synergy origination (financial, operational and collusive) is used most often. On the other hand, the separation between cost and revenue synergies is one that is used in practice very often. Cost based synergies are cost savings like redundant staff or administrative facilities. These cost based savings are therefore often referred to as bottom line synergies. The revenue based synergies are top line synergies, originating from new market, product and geography entries.
The difference between cost and revenue synergies is their level of uniqueness. Cost synergies are generally attainable; every firm can fire the CEO of the acquired company and save money. If a firm acquires a certain company, a premium has to be paid to convince the shareholders of the target to sell their shares. This premium is likely to contain all cost synergies; if the firm does not include these cost synergies in the premium, it runs the risk of another firm bidding a higher premium. The general availability of cost synergies does not apply for revenue synergies. When the acquiring firm has operations in the United States and the target firm also operates in Europe, this can form example provide an opportunity for the acquirer to sell its products in Europe. However, not every acquirer has this same geographic scope. So overall, revenue synergies are more unique for the acquiring firm. Therefore, the total amount of revenue synergies does not necessarily have to be included in the premium (i.e. be transferred to the shareholders of the target). Revenue synergies can consequently be seen as a potential signal for the acquiring shareholders that the acquisition can return in a gain for them as well.
Based on this rationale, this research hypothesizes that revenue synergies generate higher announcement returns for the shareholders of the acquirer than cost synergies. Announcement return is the abnormal return around a certain event, in this case the announcement of an acquisition. By examining the relationship between announcement return and synergies, information can be gathered that can be valuable to both science and practice. When research shows the announcement return is a reflection of the presented synergies, acquiring firms might be more careful in their formulation of the expected synergies as their share price might react. This research can provide evidence on the value of the synergies for the acquiring shareholders. Furthermore, because firms give a lot of different examples of cost and revenue synergies, combining all these examples could lead to a profound distinction between cost and revenue synergies. Finally, this research, that uses transcripts of analyst conference calls is, to my knowledge, the first of its kind. By extracting data from the conference call transcripts this research uses unique quantitative data. The database that results from this data gathering is therefore rather exclusive and can provide a refreshing look at synergies and the return for the acquiring shareholder.
I. Background
This section describes the literature to develop the hypothesis for empirical testing. Furthermore, the literature is used to provide guidance on the explanatory variables that should be used in the regression analysis. The structure of this section is as follows; first the research regarding synergies is discussed. Several types of synergies are examined and their characteristics are described. The second section discusses the return for the acquiring shareholder. The section determines whether the research shows that shareholders of the acquirer actually gain from acquisitions. Furthermore, the factors that determine this possible gain are discussed. Finally, the third section concerns acquisitions, focusing on horizontal and non-horizontal acquisitions. Not all forms of acquisitions lead to a gain for the acquiring shareholders, therefore also other factors (method of payment and domestic or cross-border) that can be recognized in research are examined.
A. Synergies
Synergies arise when the acquisition of control over the target enables the acquirer to redeploy the combined assets of the two firms toward higher-valued uses, according to Bradley, Desai and Kim (1988). Damodaran (2005) adds to this that the additional value created through the combination of two firms is an opportunity that would not have been available to these firms operating independently. Synergies can thus be seen as the additional value that arises through an acquisition. This additional value is a result of the summation of both positive and negative synergies. Following Fulghieri and Hodrick (2006) the negative synergies arise in the presence of diseconomies of scale and scope and other coordination costs1. The positive synergies can be derived from a variety of sources of which a few will be discussed in this chapter. In general the positive synergies are a result of the combination of complementary resources and economies of horizontal or vertical integration. Therefore, total synergy value is the total firm value after the acquisition, less the value of the hitherto separate entities.
Synergies can be categorized in different ways; one way to do this is by origin. Damodaran (2005) distinguishes financial and operational synergies. Operational synergies affect the operations of the combined firm and include economies of scale, increased pricing power and higher growth potential. Operational synergies can have an effect on the margins, return and growth of the acquirer. Financial synergies are more focused and include tax benefits, a higher debt capacity and additional uses for excess cash. Financial synergies result in a lower cost of capital. The categorization of Damodaran is more or less followed by Chatterjee (1986) as they both use a rather broad categorization of synergies. Chatterjee (1986) however adds the collusive synergies to the operational and financial synergies. Chatterjee (1986) differs with Damodaran when he mentions collusive synergies to represent the class of scarce resources
1
that lead to market power. Chatterjee (1986) describes operational synergies as opportunities that lead to production and/ or administrative efficiencies.
Another typology of synergies is given by Capron (1999). He distinguishes cost-based synergies from revenue based synergies. Cost synergies are mostly achieved through the elimination of redundant activities and inefficient management practices and lead to cost savings. Revenue-based synergies are achieved through resource redeployment through an increased market coverage or enhanced innovation capability and lead to revenue enhancing capabilities. Damodaran (2005) uses the same typology, however he uses the terms cost and growth synergies, with the growth synergies representing the revenue synergies of Capron.
Next to a distinction on type, synergies can also be recognized in terms of achievability, as not all firms are able to achieve the same synergies. Pursche (1989) suggests a classification into universal, endemic and unique synergies. Universal synergies include economies of scale such as management information systems, shared administrative activities such as senior management, and other shared facilities. Endemic synergies are those only available to a few acquirers, typically those in the same industry as the acquirer. These would include most economies of scope, such as a broadened geographic coverage or a redundant sales force in a horizontal acquisition. Final type is unique synergies, opportunities that are distinctive to a particular acquirer. These synergies are usually tied to a unique skill the buyer has. James, Mendonca, Peters and Wilson (1997) continue using this distinction and give a better understanding of the unique synergies by explicitly mentioning special skills or assets (such as distribution channels or databases); leveraging a company’s existing business base to create new business opportunities; and perhaps even changing the industry structure to seize a competitive advantage. James, Mendonca, Peters and Wilson (1997) further point out that the unique synergies are often the deciding factor in most deals.
When examining the announcement presentations, firms mention various forms of extra cost and revenue synergies. For both types a lot of different examples are presented to the investors. Appendices I and II list an overview of the different presentation of synergies. The synergies that are most apparent and frequently mention are a delisting of the acquired company, aligning of the assets to maximize efficiencies or asset rationalization, elimination of redundant sales force, avoiding of having to go to the market to hire new people, elimination of selling, general and administrative (SG&A) and procurement (activities related to purchasing goods, services and works) expenses, higher utilization of current capacity and combined media exposure on the cost side. The most apparent and frequently mentioned revenue synergies are an expanded geographic reach, better penetration of new and existing markets, cross-selling initiatives and the expansion of customer and product opportunities.
A1. Distribution of synergies
acquiring or acquired shareholders receive most of the added value that arises from an acquisition?
James, Mendonca, Peters and Wilson (1997) mention the decisive nature of unique synergies. Looking more into the distribution of the synergies Damodaran (2005) adds that since synergy requires skills and strengths contributed by both the acquiring and target firms for its existence, the acquiring company’s share of the synergy dependsupon the uniqueness of the strengths it brings to the mix. In the limiting case, if only the acquiring firm has the components necessary to achieve the synergies, it should receive a large share of the synergy benefits. If the acquiring firm’s strengths are not unique and could be offered by other firms as well, the bargaining power shifts to the target firm and the target firm’s stockholders should receive the bulk of the benefits. In the end the competition is most likely to raise the price of the acquisition to the point where synergistic savings of the second best acquirer are paid to shareholders of the target company. This is also pointed out by Goldberg and Godwin (2001). If the synergistic benefits are fully unique to the acquiring firm, there is no competing acquirer. In this case, the purchaser may receive some or all of the unique synergistic benefits.
Berkovitch and Narayanan (1993) find the following on the distribution of synergy between the acquirer and the target. First, the higher the cash component, the lower the fraction of synergy captured by the target. Second, the dollar amount of synergy captured by the target increases with the level of total synergy and the level of the cash component. Third, an increase in competition among the acquirers raises the fraction of synergy that is captured by the target. Fourth, the target’s payoff is higher when there is actual (rather than potential) competition. Fifth, the fraction of synergy captured by the target decreases with the level of total synergy. Ideally acquiring shareholders would thus like to see acquisitions without competition, full cash payment and a high total level of synergies.
There is one factor that does not contribute to this difference between cost and revenue synergies; the attainability of the synergies. Houston, James and Ryngaert (2001) explain that cost synergies are easy to attain synergies as the areas where cost can be cut can be explained, quantified and supported by detailed analysis. However, the revenue synergies are much more uncertain and therefore harder to attain than cost synergies. This is also the explanation of their research results that show cost synergies to have a better effect on the announcement return of the acquirer than revenue synergies. One remark that must be made is that Houston, James and Ryngaert (2001) use a sample that consists of only mergers between banks.
B. Return for the acquiring shareholders
Overall a large body of evidence indicates that, in general, target firm shareholders obtain most of the gains in acquisitions. Jensen and Ruback (1983) show that in general target firm shareholders gain from an acquisition whereas acquiring firm shareholders do not lose. Bradley, Desai and Kim (1988) show that both in single and multiple-bidder cases, target firm shareholders have greater gains than the acquiring firm shareholders. In general can be said that competition among prospective acquirers pushes the price of the target company to the point where the target company shareholders receive the expected value added resulting from the acquisition. The gains for the shareholders of the acquirer are therefore not statistically different from zero according to Goldberg and Godwin (2001).
Empirical research by Moeller, Schlingemann and Stulz (2005) shows that over the period 1998 to 2001, acquiring-firm shareholders lost 12 cents around acquisition announcements per dollar spent on acquisitions for a total aggregated loss of $240 billion. However, in this sample of 4,136 acquisitions, 87 are large loss deals. The total wealth loss associated with this 2,1% of acquisitions in the sample is $397 billion, about 43.4% of the total money spent on acquisitions. The conclusion could therefore be that on average the acquirers’ shareholders may not earn or might even lose money in acquisitions. However, this is mainly due to a small number of acquisitions that generate very large losses.
Acquiring shareholders do not earn or even lose from acquisitions. Knowing this, it is remarkable that the number of acquisitions has been increasing rapidly in the last ten years. Therefore, it would be interesting to know what factors determine a possible gain for the shareholders of the acquirer when they do an acquisition. Copeland, Koller and Murrin (2000) conclude that several factors influence the value of acquisitions for the acquiring shareholders. First, only if the deal leads to unique and significant synergies is the market likely to perceive it as increasing the acquirer’s value. Otherwise, it is likely that the target company’s shareholders capture the value created by the acquisition in the premium paid above the pre acquisition stock price. Second, acquirers that pay lower premiums (less than 10 percent) are three times more likely to see their stock price favorably affected by the announcement.
acquisition for acquiring shareholders. The results of Singh and Montgomery (1987) suggest that related acquisitions have greater total dollar gains than unrelated transactions. This is observed as merely a result of the availability of opportunities for gains, with related acquisitions having more of these opportunities. However, Churyk and Baker (2004) show that the premiums are not related to the degree of industrial association between the companies. Therefore can be concluded that although related acquisitions do generate higher total dollar gains, this is not reflected in the premium paid
The announcement return is said to be influenced by the amount of premium paid, the total level of unique and significant synergies and the relatedness between the target and the acquirer. Berkovitch and Narayanan (1993) mention the importance of cash payment for the announcement return of the acquirer. Travlos (1987) and Goergen and Renneboog (2004) show that acquisitions of public firms paid with equity are accompanied by lower announcement returns than payments with cash.
Moeller and Schlingemann (2005) find evidence for a significant difference in announcement return by domestic and cross-border acquisitions. In their large sized sample of US acquisitions, cross-border acquisitions lead to less return for the acquirer compared to domestic acquisitions. They conclude that stock returns are negatively associated with an increase in both global and industrial diversification. This would confirm that companies should not attempt to do what shareholders can do themselves, creating a diversified portfolio. Goergen and Renneboog (2004) use a European sample of acquisitions and obtain the same results. They also find that domestic acquisitions and acquisitions trigger higher wealth effects than cross-border operations.
One last factor that might be considered is the relative size of the acquirer compared to the target. Jarrell and Poulsen (1989) argued that as the target increases in size relative to that of the acquirer, the impact of the acquisition would be more readily observed in the acquirer's return. Thus, if acquisitions are on average wealth-increasing projects for acquiring firms, the largest positive return should be observed when the target is large relative to the acquiring firm. However Goergen and Renneboog (2004) and Travlos (1987) find no evidence for this relative size effect on target and bidder wealth effects. The reason Travlos (1987) mentions is that because the study focuses on large M&A deals, the average relative size is fairly homogeneous and therefore does not present a significant result.
affect the announcement return are therefore also suggested for the rumor and the completion return.
C. Acquisitions
As has been shown in the previous section, relatedness provides a basis for positive returns for the acquiring shareholders in an acquisition. Next to theories considering the returns from acquisitions, another part of the acquisitions research focuses on the motives for acquisitions. Cited many times as reasons to acquire is ‘we need to be in this game’, ‘we must be big to survive and thrive’ and ‘if we don’t buy it, someone else will’ according to Darragh, Dodig and O'Hanley (1997).
There are three general motives for takeovers: synergy, agency and hubris according to Berkovitch and Narayanan (1993). Goergen and Renneboog (2004) mention the synergy motive to be the prime motivation in their study of a large sample of European acquisitions. Therefore, when ignoring the effects of agency and hubris, the synergy motive, to engage in a takeover only when it results in gains to shareholders of both acquirer and target, suggests that gains to acquirer and target shareholders are positive and shared between acquirer and target. For value creation Lubatkin (1983) emphasizes the necessity of strategic fit. This statement from Lubatkin (1983) contributes to Damodaran (2005) who mentions the necessity of specific skills and strengths for realizing the value of the acquisition. The better the strategic fit between the acquiring and the acquired firm, that is, the more the respective environments of the two firms have unifying features, the greater should be the performance gain to the acquired firm. Next to strategic fit, some competitive advantage should exist that make certain acquisitions successful. Chatterjee (1986) mentions scarce resources to be important for creating economic value and has three suggestions on where this competitive advantage should be derived from. First, the acquiring firm may have an information advantage in terms of unique access to information regarding the viability of the acquisition. Second, the acquiring firm may have a competitive advantage in terms of how it manages the acquisition process. Third, the acquisition may lead to a unique business model that is difficult for competitors to copy.
characteristics. The resources should be valuable, rare, inimitable and non-substitutable to be able to generate and sustain competitive advantage.
This uniqueness of a certain deal that could create economic value for both acquirer and target can be obtained in four different areas according to Darragh, Dodig and O'Hanley (1997). The first area is the addition of successful new products to an existing distribution platform. The second is the acquisition of a strong brand name or operational excellence in sales or marketing. Third is gaining access to an attractive geographic or product market. And the fourth is gaining access to a broader range of distribution channels. These four sources of strategic value can lead to the creation of economic value for the acquirer.
Trautwein (1990) elaborated on a number of theories concerning acquisition success trying to explain a firm’s motive for acquisitions. For this research the theories that explain acquisitions as a rational choice benefiting the acquiring shareholders are key, the previously mentioned ‘motives’ are therefore not considered. The first theory is the efficiency theory explaining acquisitions as a result of net gains through synergies. The second is the monopoly theory that looks at acquisitions as aiming to obtain wealth transfers from customers. The third theory is the raider theory, which aims at obtaining wealth transfers from target’s shareholders. Fourth theory is the valuation theory, which aims at achieving net gains through private information.
In general, most acquisitions fall into one of these four categories. For this research the emphasis is put on the first motive, the efficiency theory. Net gains through synergy are attempted to be achieved by the acquirer. To maximize this value that can be created, the acquirer and target need to have a strategic fit and a set of scarce resources. Most of these opportunities are available with horizontal acquisitions according to Chatterjee (1986). Horizontal acquisitions are able to generate collusive synergies to a certain extent as antitrust policies have been created to reduce these opportunities. Next to the collusive synergies, horizontal acquisitions generally have more overlap ergo more opportunities for creating value when joining forces. Finally, with horizontal acquisitions, the acquirer has better knowledge of the company he buys. Consequently, he is better aware of the opportunities and the pitfalls that come with the sector. In general the acquirer is better able to realize possible gains through the acquisition as he better understands what he is buying. This is supported by the theory on the benefits of related diversification and the application of core competencies of Barney (1988). C1. Premium
According to Berkovitch and Narayanan (1993) and Barney (1988) the acquisition premium should reflect the value to the acquirer of the expected synergies. On the other hand, Slusky and Caves (1991) show in their research that only the financial synergies are related to the premium. They found no association between the real synergies and the premium. However, Slusky and Caves (1991) did not have actual synergy information, they used proxies based on acquirer and target financials.
D. Hypothesis
Concluding the section that provides the background for this thesis a hypothesis is posed. The previous section discussed that uniqueness is a decisive factor for acquisitions to succeed and add value. An acquiring firm needs to have a unique asset or a unique combination with the target to make the acquisition add value to both parties. The presented synergies are an expectation of the value that can be extracted when the target and the acquirer are combined. Whereas the cost synergies presented are rather general, the revenue synergies more often have a unique nature, making them available for only one or a few acquirers. It is therefore expected that the presented revenue synergies are not necessarily all paid to the target. This would create an opportunity for a gain for the acquiring shareholder. Based on this rationale the following hypothesis has been constructed that will be tested with a large sample of acquisitions.
II. Data A. Sample selection
The sample consists of companies from North America, Canada, Western Europe and Scandinavia2, that did an acquisition in the period between August 2002 and the end of December 2007. This sample is obtained from the Zephyr database from Bureau van Dijk. The Zephyr database contains half a million public and private takeovers. The database contains information on both the acquisition itself, but also on the acquirer, target and a possible vendor. The information is both quantitative (for example company financials and deal value), and qualitative (for example company background and sector information).
Several criteria have been used to obtain the acquisitions in the sample. First of all, the sample is restricted to the larger transactions, since the return resulting from the acquisition is more apparent with the larger transactions. Therefore a minimal deal value of US$1 billion is chosen. Zephyr converts Euro into Dollar and vice versa for all deal values based on the exchange rate on the date of the announcement. Second, the sample requires acquirers to have stock return data available on Capital IQ so that the market reaction to the acquisition can be measured. This restriction eliminates acquisitions by private buyers. Third, acquisitions by financials and utilities are eliminated as these operate in regulated industries. Fourth, since all the information on the bid should be in the measured announcement return, only acquisitions that received just one bid are considered. Fifth, given that this research aims at examining the actual return as a result to an acquisition, only winning bids in completed acquisitions are examined. Sixth, acquisitions where the acquirer does not obtain the complete ownership of the target are eliminated. With all the selection criteria combined the sample yields 291 acquisitions. Due to some missing extra information on certain transactions that turns out to be necessary for further examination of the data, the final sample yields 243 acquisitions. These excluded acquisitions are for example acquisitions in which the acquiring company was takeover itself in a later stage. Also firms that bought a private firm, which made it impossible to determine the acquisition premium, are excluded. Appendix III lists the acquisitions that have been used for this research.
B. Other information
Zephyr provides the acquisition sample and includes deal specifics as deal value, premium paid and the dates of rumor, announcement and completion. The used premium is the difference between the total offer price per target share and the closing price of the target share on the day prior to the official announcement. Furthermore, Zephyr provides US SIC (Standard Industrial Classification) codes for both acquirer and target, indicating the main industry of the parties involved. Following Eckbo (1983), by comparing these 4-digit codes one can determine
2
whether the acquisition is horizontal, that is when the SIC code of the acquirer is the same as the one of the target3. In the sample of this research 198 (82%) acquisitions can be characterized as horizontal. Consequently 45 (18%) acquisitions can be characterized as non-horizontal.
Calculating the announcement return creates a need for share prices and other financial data of the acquirer around the three important dates of the acquisition. Capital IQ, a division of Standard & Poor’s, provides global private and public capital market data. Capital IQ can be used to generate share prices of both markets and individual securities. Moreover, Capital IQ also provides the EBITDA (Earnings Before Interest, Taxes, Depreciation and Amortization) for the acquirer for the year of the acquisition. This EBITDA is used to convert the total amount synergies into a better comparable figure. Finally Capital IQ also provides the official filing dates for annual or quarterly reports. With these dates it is possible to find whether the announcement of the acquisition coincided with the announcement of other company information, that is, quarterly and annual reports. If this is the case for some announcement, rumor or completion dates, this can be taken into account with the regression analysis.
Final information source is Reuters Knowledge. Reuters provides information on all the larger companies in the world. They gather broker reports, analyst research and annual reports and present them on one platform. Among the sources of information presented on Reuters are transcripts of analyst conference calls. Companies announcing an acquisition, give conference calls for their investors and interested analysts to provide them with comments on the deal and give them the opportunity to ask questions on the deal rationale. For each acquisition in the sample the transcript is searched for information on the synergies that result from the announced transaction. The search terms used are “cost synergies” and “revenue synergies”. If these searches do not return a result, a new search is done using the terms “cost savings” and “growth opportunities.” These terms are used interchangeably by the firms presenting the acquisition announcements. For all firms that reported synergies a capitalized value for these expected synergies is obtained. This value is the present value for all the expected future synergies. Next to synergy information, the premium paid in the transaction (the percentage of the purchase price above the market price) is also obtained from the conference call transcripts. For all transactions the selected premium is the premium over close price on the last trading day prior to the announcement of the acquisition.
For the 243 acquisitions in the sample the information on both cost and revenue synergies is collected. There are two special cases. One is when the companies announced synergies but are not yet able or allowed to publish a fixed number. These acquisitions got the label, “synergies not quantified.” The second case is when the company does not mention any cost or revenue opportunities arising from the transaction. These acquisitions got the label “synergies not mentioned.”
3
III. Methodology
This section discusses the methodology of the event study, the univariate analysis and the multivariate analysis. The section is separated into three different parts. The first part discusses the event study and discusses the calculation of the announcement, rumor and completion return. Furthermore the methods that are used to calculate the abnormal return and the selection of the event window and the estimation period are described. For the description of the event study the methodology of Campbell, Lo and MacKinlay (1997) is used. The second part of this section discusses the univariate analysis, the individual analysis of certain sub samples. Finally the multivariate analysis is discussed.
A. Event study
Event studies measure the effects of economic events on the value of firms by measuring the stock price reaction. A fundamental requirement for any event study is a suitable deep market for the instrument that is under observation. If the instrument is seldom traded, the price may not fully reflect all changes in economic value on an efficient basis. Following Campbell, Lo and MacKinlay (1997) seven steps are identified to conduct a typical event study.
(1) Event definition
The initial task when conducting an event study is to define the event of interest and identify the period over which the price of the relevant financial instrument, in this case the stock price, is examined. Assuming perfect capital markets, one would expect the stock price to immediately reflect all the value affecting events. However, in practice the market may acquire information prior to the event, speculate on the content of the information or take some time to assimilate information and react to it. For this reason it is common to use an event window of two or three days prior to and after the actual event date4. Using an 11 day event window increases the risk of events other than the one that is actually studied having an effect on the final results. For this research a seven day event window is chosen, similar to Moeller, Schlingemann and Stulz (2005) to capture both lead and lag effects.
(2) Selection criteria
The second step in the typical event study is to select criteria for choosing firms to include in the sample. The different requirements for the selection have already been discussed in the section data. This section elaborates on the criteria and sources used to select the acquisition sample for this research. Appendix III lists the acquisitions that have been used for this research.
(3) Normal and abnormal returns.
4
To asses the impact of a specific event on the return of a certain stock price one must first find out what the return would have been in the absence of the event. This return is called the ‘normal return.’ The return in excess of the normal return that the stock generates during the event period is called the ‘abnormal return.’ This abnormal return therefore reveals the return that is generated because of the event. Brown and Warner (1985) describe three procedures available to estimate this normal and abnormal return.
Mean adjusted return
This procedure calculates the abnormal return on a certain security by assuming that the mean return of that security is constant over time. The formula for the abnormal returns is as follows;
i t i t i
R
R
A
,
,
(1)Where
R
i,t is the return on a security i on day t,A
i,t is the abnormal return for security i at dayt and
R
i is the average of security i’s daily returns over the estimation period. Market adjusted returnThe second procedure is potentially superior as it removes a portion of the return that is related to movement in the market. The formula for the abnormal return with the market adjusted returns is as follows; t m t i t i
R
R
A
,
,
, (2)Where
R
m,tis the return on a certain selected market index. In this model the return of the stock is expect to be equally reflected in the market return, i.e. this formula assumes a beta of 1.OLS market return
The third procedure is probably the most widely used and provides the best result when calculating the abnormal return. This model uses the estimation period to estimate both alpha and beta, increasing the chance of isolating the effect of a specific event. The formula used for the OLS market returns procedure is as follows;
t m i i t i t i
R
R
A
,
,
ˆ
ˆ
, (3)Where
ˆ
iis the estimated intercept and
ˆ
iis the estimated OLS regression coefficient between the stock price i and the market portfolio over the estimation period.(4) Estimation procedure
The parameters are estimated using a set of the data referred to as the ‘estimation window.’ This estimation window is used to estimate the parameters of the model, beta and alpha. This estimation window is a certain period of approximately 120 trading days prior to the event and generally does not include the event window itself5. Authors often use daily stock data for their estimation period, however there are certain disadvantages to using daily stock data according to Lo and MacKinlay (1988). Since the estimation period should provide well estimated parameters, a large number of observations are appropriate. While daily sampling yields many observations, the biases associated with non-trading, the bid-ask spread and asynchronous prices are troublesome. Daily fluctuations would therefore overstate the fluctuation of the stock price relative to weekly returns whereas monthly data would understate the realistic movement of the stock price. Daily data shows too much information, monthly data too little. Weekly sampling can provide the compromise, yielding a large number of observations while minimizing the biases inherent in daily data. For this reason an estimation window of 104 weeks, meaning the close price of each week, mostly Friday, is chosen, providing weekly stock and market returns to the parameters of the models.
(5) Testing procedure
After the estimation of the parameters and the calculation of the abnormal return for each day in the event window a testing framework needs to be defined. Furthermore, techniques have to be chosen for the aggregation of the results over time and across individual firms. The hypotheses presented after the review of the literature is used in this phase of the event study. The cumulative abnormal return (CAR) is defined as the sum of the abnormal returns for each day (d) of the event window;
iti
AR
CAR
, (4)Under the null hypothesis, the given event has no impact on the mean or variance of returns; hence the abnormal return is expected to be zero. To be able to draw conclusions from the CAR, a test statistic, the students-t (ti), must be calculated. There are various ways to calculate
ti, but for this sample with estimated cumulative abnormal returns the following method can be
used according to Salinger (1992);
N
CAR
t
i i i
(5)In this model
iis the standard deviation of the distribution and N is the number of days in the sample. To draw conclusions from the event study the likelihood that an event study test correctly rejects the null hypothesis has to be known. Therefore, the significance of the statistical test is set on 1%, 5% and 10% levels.
5
The cumulative abnormal return for all transactions can subsequently be compounded into one number, the cumulative average abnormal return (CAAR). This CAAR is simply the average of the cumulative abnormal returns for each transaction in the sample. The same formula is used to calculate the students t-statistic, only the N is not the number of days but the number of transactions in the sample. The final two steps involve the presentation and interpretation of the results and the conclusions. These final two steps are elaborated on in the sections to come.
B. Univariate analysis
Univariate analysis is the assessment of the distributional properties of a variable. It serves the purpose of description and preparation for multivariate analysis. These functions correspond to the two primary forms of univariate analysis, the assessment of central tendency and of dispersion. What is a typical value, and to what extent do the measured values differ from this typical value? Although descriptive research often focuses on identifying the characteristics of a set of transactions, variation from this typical value usually is the most important concern with regard to subsequent multivariate analysis.
In some instances, univariate analysis itself is the research goal. This is the case in descriptive research that is concerned solely with defining the occurrence of some phenomenon. However, most research has an explanatory focus, and univariate analysis serves primarily as a foundation for the multivariate analysis. For this research, the question is asked whether presenting certain synergies influences the return for the acquiring shareholders. Although answering such a research question eventually requires multivariate analysis, univariate analysis is a first step.
Separating these groups could explain if synergies in general matter for the judgment on the value of the acquisition by the shareholders of the acquirer.
For all mentioned sub samples the mean and the standard deviation in their CAARs are calculated for each of the three events (rumor, announcement and completion). Whether the results for both groups differ significantly is measured using a students t-test. Because the samples are of very different sizes, the variance is likely to be extremely different for the two groups, therefore a t-test that does not assume equal variances is chosen to calculate the significance of the different sets of variables. The following formula is adapted from Kanji (1993) for the test statistic;
2 1 2 2 2 1 2 1 2 1 2 1
n
s
n
s
x
x
t
i
(6)With population means µ1 and µ2, sample sizes n1 and n2, sample means x1 and x2 and
variances s1 2
and s2 2
. Since the t-test for the univariate analysis compares the means of two groups that can be of a different sample size, the following formula is used to calculate the degrees of freedom;
2
1
1
2 2 2 4 2 1 2 1 4 1 2 2 2 2 1 2 1
)
(
)
(
n
n
S
n
n
S
n
s
n
s
df'
(7)This formula is adopted from Kanji (1993), with sy representing the standard deviation of
sample y and ny the number of items in sample y.
C. Multivariate regression analysis
The final analysis for this research is a multivariate regression analysis. The event study and the univariate analysis provided the descriptive statistics for the individual sub samples. What still needs to be determined is whether the abnormal return that results from an acquisition announcement is influenced by different acquisitions characteristics in general and the synergies in particular. The total sample includes 243 transactions; however, not all transactions have presented cost and/or revenue synergy information. Therefore the samples used for the multivariate analysis differs with the number of acquisitions that presented the certain type of synergies.
paid is the fourth dependent variable. The premium is used as a dependent variable as it might provide new insights into the relationship between synergies and the premium paid. Literature, using proxies for the available synergies, hardly shows any relationships, however the actual synergy numbers used with this research might provide a new perspective.
The model used for the multivariate analysis includes three independent (or explanatory) variables. The first independent variable is the total amount of synergies (TS) presented. The second is the total amount of cost synergies (CS) presented and third is the total amount of revenue synergies (RS) presented. All these synergy values are converted using the acquirers EBITDA from the year of the acquisition. This provides a number between zero and one displaying the amount of synergies that can be gained compared to the earnings of that year. Next to these three variables, four indicators are used to represent dummy variables that provide the control variables for this test. The first dummy variable is for the type of acquisition (HOR), whether it is a horizontal (1) or non-horizontal (0) acquisition. The second dummy variable depicts the method of payment (CASH), either cash (1) or equity (0). The third dummy variable shows whether other information (OINFO) is provided to the market during the seven day event window. If this is the case the indicator is a (1), if this is not the case this indicator is (0). With other info quarter or annual results are meant. Final dummy variable (CROSSB) indicates whether the acquisition is cross-border (1) or domestic (0). DVi represents the
dependent variables i (i=1,2,3,4). With i = 1 representing announcement return, i = 2 is the rumor return, i = 2 is the completion return and i = 4 is the acquisition premium. This leads to the following models that are tested;
TS HOR CASH OINFO CROSSB
DVi 1 2 3 4 5 (8)
CS RS HOR CASH OINFO CROSSB
DVi 1 2 3 4 5 6 (9)
CS HOR CASH OINFO CROSSB
DVi 1 2 3 4 5 (10)
RS HOR CASH OINFO CROSSB
DVi 1 2 3 4 5 (11)
IV. Results
This section discusses the results of the event study, the univariate analysis and the multivariate analysis.
A. Results of the event study
For the rumor, the announcement and the completion date the abnormal return is measured over the seven day event window. Three methods are used; the constant mean, the market and the OLS market method. The latter is used as the primary indicator of the abnormal return for reasons discussed in the methodology section. The constant mean and the market method are used as checks for the robustness of the OLS market method. The results of this robustness check are presented in table IV.1.
Table IV.1: Robustness check OLS Market method Paired two sample for means t-test (p-value) Rumor Announcement Completion
2.09** (0.04) 2.78*** (0.01) 2.08** (0.04) Market method -1.16 (0.25) -0.69 (0.49) -0.76 (0.45) Constant mean method
As can be seen in table IV.1, the OLS market method is robust with the market model for the rumor, the announcement and the completion return on a five percent significance level. The OLS market model is not robust for the results of the constant mean model.
Table IV.2 reports the descriptive statistics of the cumulative average abnormal returns over the seven day event window for the three different dates of the event study6. Also the descriptive statistics of the acquisition premium are included in table IV.2. A graphical representation of the distribution of the rumor, the announcement and the completion return is presented in figures A.IV.1, A.IV.2 and A.IV.3 respectively.
Table IV.2: Descriptive statistics of the total sample
CAAR (-3/+3) (OLS market method)
Mean Median Max Min Kurtosis Skewness Bera Jarque St Dev t-statistic Announcement return N = 243 -0.72% -1.04% 28.7% -25.1% 1.51 0.02 22.38 0.08 -1.47 Rumor return N = 100 0.71% 1.14% 25.8% -17.2% 3.35 0.56 5.74 0.07 1.07 Completion return N = 243 1.04% 1.26% 26.8% -18.9% 3.40 0.42 8.81 0.06 2.65*** Acquisition premium N = 243 26.4% 22.0% 102.0% -4.9% 3.10 1.42 82.31 17.65 0.69 This table shows the descriptive statistics for the cumulative average abnormal return following the OLS market method for a seven day window surrounding the announcement date. Statistical significance at a 10%, 5% and 1% level is denoted with *,** and ***.
As can be seen the average announcement return to the acquirer for the acquisitions in the used sample is negative and thereby similar to for example Moeller, Schlingemann and Stulz (2005), however this result is not significant. The average rumor return is positive but not significant. The average return on the completion date however, is positive and significant at a
6
one percent level. This might be a signal of speculative intents from investors, aiming for a profit when the deal turns out to be unsuccessful. This might occur because shareholders wait for the completion announcement and start buying shares as soon as they know the acquisition is successful. The table also shows that the distribution of the error terms for the three event dates do not exactly follow a normal distribution. A normal distribution is supposed to have a skewness close to zero and a kurtosis close to three. The rumor and the completion return more or less follow this rule. Because these three all have a fairly large sample size this does not cause any problems according to Brooks (2002). The announcement return and the acquisition premium do not follow the normal distribution. The Bera Jarque is too large. With the premium there are many outliers that could be removed. For the announcement return the kurtosis is leptokurtic. This is a signal of a flat distribution where the variance is the result of a large group of less extreme values. Therefore, the null hypothesis of residual normality is rejected, implying that the conclusions that are drawn about the coefficient estimates could be wrong, although the sample is probably sufficiently large to not give great cause for concern. B. Results of the univariate analysis
Five sub samples have been tested in the univariate analysis7. First sub sample is the horizontal versus the non horizontal transactions. As can be seen in table A.V.1 there are relatively more horizontal acquisitions. The results show a minor and insignificant difference for the rumor dates. There is a difference for the announcement date of almost three percent (-2.77%) at a significance level of one percent. For the total sample of 243 acquisitions the non-horizontal acquisitions clearly have a better announcement return than the non-horizontal. This result is remarkable since Singh and Montgomery (1987) suggest horizontal acquisitions to generate a better announcement return.
Considering the return around the completion, the CAAR’s are positive and significant for both horizontal as for non-horizontal transactions. This may be due to the investors speculative intents mentioned earlier, however this speculative intent does not differ between the two samples as the difference between the horizontal and the non-horizontal transactions is not significant.
The second sub sample is the comparison between domestic and cross-border transactions. The results are presented in table A.V.2. For the rumor return both sub samples generate a positive, although not significant CAAR. The difference between the domestic and the cross-border transactions for the rumor date is small and not significant. For the announcement date the difference is just over one percent (1,08%), however the result is not significant. The announcement return for the domestic transactions is minus one percent (-0.99%) and significant at a ten percent level. The difference in the return around the completion date is
7
almost two percent (1,87%) in the advantage of the domestic transactions and significant at a five percent level. Moreover, the domestic transactions generate a positive completion return of 1,5 percent on a one percent significance level.
The third sub sample differentiates on method of payment. There are three types of payment distinguished; cash, equity or a mixture of cash and equity. Different things can be noticed when considering table A.V.3. First of all, cash transactions appear to generate a positive rumor, announcement and completion return. However, none of these returns are significant. For transactions that are paid with equity only the completion return is positive and significant at a one percent level. The rumor and the announcement return for equity transactions are both negative. The announcement return for equity transactions is clearly negative (-2,58%) at a significance level of ten percent. The mixture of cash and equity show a positive but not significant return on the rumor date. The announcement returns for transactions paid with a mixture of cash and equity are negative (-1,50%) and not significant. The completion return for these transactions is large and positive (2,26%) on a five percent significance level.
When putting the returns to the different groups within this sub sample together, the following results are obtained. The difference between the returns on the cash transactions and the equity transactions is significant at a five percent level for both the completion and the announcement return. This difference is in favor of the cash transactions for the announcement return; 2,88% better. This means that for this sample the cash transactions generate a significantly better announcement return than the equity transactions. The remarkable thing is that the return around completion is also significant and more than two percent (2.19%), but it is in favor of the equity transactions. The difference between the cash and the equity transactions is thus significant but the inverse for both events. Perhaps deals that are financed with cash are more likely to succeed and are therefore less prone to speculative intents from investors. The difference between cash and the transactions with a mixed payment is significant for the completion return. The result is in favor of the mixed payment for the completion return. The relationship between the equity and the mixed payment transactions shows no significant results.
The most interesting univariate analysis is the sub sample that distinguishes between the presented synergies. In this sub sample four types are distinguished. Synergies not presented, synergies not quantified, only cost synergies and both cost and revenue synergies. It is hypothesized that the group with the transactions that present revenue synergies generates the largest abnormal return. The results of the univariate analysis for this sub sample are displayed in table IV.3.
The group where the synergies are not quantified generates a far better result for these two events. However, due to insignificance no conclusions can be drawn. The completion return for both the not mentioned and the not quantified synergies is positive but only significant for the group of the not quantified synergies.
Table IV.3: Results univariate analysis by type of synergies in transaction
Syn. not presented Syn. not quantified Cost synergies Revenue synergies Difference (NP/NQ) Difference (C/C&R) Rumor N = 25 N = 10 N = 43 N = 22 CAAR (-3/+3) -0.70% 3.87% 0.99% 0.35% -4.57% 0.64% Standard Dev. 0.05 0.09 0.07 0.06 t-statistic (df') -0.65 1.34 0.95 0.25 -1.48 (12) 0.37 (44) Announcement N = 60 N = 31 N = 98 N = 54 CAAR (-3/+3) -1.34% 0.74% -0.47% -1.31% -2.08% 0.85% Standard Dev. 0.09 0.07 0.07 0.08 t-statistic (df') -1.19 0.61 -0.68 -1.19 -1.26 (75) 0.65 (95) Completion N = 60 N = 31 N = 98 N = 54 CAAR (-3/+3) 1.34% 1.98% 0.83% 0.55% -0.64% 0.28% Standard Dev. 0.06 0.06 0.07 0.05 t-statistic (df') 1.62 1.71* 1.26 0.84 -0.45 (61) 0.31 (138) This table shows the CAAR for the sub sample of synergies resulting from the transaction for the different event dates of these acquisitions.'Synergies not presented' consists of those acquisitions that did not mention any synergy opportunities arising from the transaction. 'Synergies not quantified' are those acquisitions of which the firm mentioned some synergy opportunities, but was not able to quantify them. Cost synergies are the transactions that only mentioned some cost savings. Revenue synergies consists of those transactions that presented revenue opportunities resulting of the merger. Statistical significance at a 10%, 5% and 1% level is denoted with *,** and ***.
It is assumed that revenue synergies have a more unique character than cost synergies. Because of this characteristic, revenue synergies are not fully paid to the target in an acquisition. This would therefore lead to a better return for the acquiring shareholders. The rumor, announcement and the completion return are lower for the transactions with cost and revenue synergies compared to transactions that present only cost synergies. Perhaps the explanation of Houston, James and Ryngaert (2001) might be valid for the results of this analysis as well. Houston, James and Ryngaert (2001) explain that since revenue synergies are much harder to attain than cost synergies, shareholders appear to be skeptical about the revenue projections. However, since there are no significant results, it is not possible to draw any definite conclusions. Therefore, these results do not provide support for the posed hypothesis but can neither explain whether the explanation of Houston, James and Ryngaert (2001) is true for this case.
Table IV.4: Results univariate analysis by synergy number given No synergy number given Synergy number given Difference (No Number/ Number) Rumor N = 35 N = 65 CAAR (-3/+3) 0.61% 0.77% -0.16% Standard Dev. 0.07 0.07 t-statistic (df') 0.52 0.93 -0.11 (68) Announcement N = 91 N = 152 CAAR (-3/+3) -0.63% -0.77% 0.14% Standard Dev. 0.08 0.07 t-statistic (df') -0.74 -1.30 0.13 (174) Completion N = 91 N = 152 CAAR (-3/+3) 1.56% 0.73% 0.83% Standard Dev. 0.06 0.06 t-statistic (df') 2.32** 1.51 1 (179) This table shows the CAAR for the sub sample of acquisitions that either present or did not present a synergy number. These CAAR's are given for the different event dates of these acquisitions. Statistical significance at a 10%, 5% and 1% level is denoted with *,** and ***.
A few explanations are available for the lack of results. First is the use of acquisitions that received only one bid. It might be the case that a knockout premium is presented to the target shareholders in a lot of these cases. In this case the acquiring shareholders might regard the premium as an overpayment despite the presented synergies. And second, the used synergy numbers are predictions of the management that has to convince its own and also the target shareholders of the value that can be achieved when the firms are combined. The numbers are therefore quite subjective, contributing to a skeptical attitude of the acquiring shareholder towards synergies.
Because of the insignificant results of the univariate analysis considering the synergies presented in a transaction, it is interesting to look at the results of the multivariate analysis. The results of the multivariate analysis might shed some light on whether presenting synergies in a transaction has any effect on the value acquiring shareholders expect to arise from an acquisition.
C. Results of the multivariate analysis
Table IV.5: Results regression, total synergies (equation 8)
Rumor Announcement Completion Premium
Coefficient Prob. Coefficient Prob. Coefficient Prob. Coefficient Prob.
Intercept -0.006 0.828 0.008 0.632 0.049 0.001 0.195 0.000 Horizontal -0.001 0.966 -0.033 0.027** -0.019 0.132 0.015 0.641 Cross-border 0.005 0.801 -0.009 0.526 -0.014 0.213 0.016 0.588 Cash 0.004 0.869 0.017 0.273 -0.030 0.014** 0.049 0.128 Other info -0.014 0.697 0.051 0.179 0.018 0.393 Total synergies 0.069 0.129 0.009 0.596 -0.030 0.026** -0.033 0.347 Observations N = 65 N = 152 N = 152 N = 152 Adj. R-squared -0.360 0.025 0.065 0.008 F-statistic 0.557 1.764 3.114 1.289 Prob. (F-statistic) 0.733 0.124 0.011** 0.277
This table shows the results of the multivariate analysis with four dependent variables. The first three dependent variables are the cumulative average abnormal return for a seven-day event window around the rumor, the announcement and the completion date of an acquisition. Final dependent variable is the acquisition premium paid to the target. All independent variables, except the ones concerning synergies, are dummy variables. The first dummy variable is for the type of acquisition 'Horizontal', indicating a horizontal (1) or a non-horizontal (0) acquisition. The dummy variable 'Cash' is on the method of payment, either cash (1) or equity (0). The third dummy variable shows whether other information 'Other info' was provided to the market on the date of the announcement/ rumor/ completion, if this is the case the indicator is a (1), if this is not the case this indicator is (0). Final dummy variable 'Cross-border' indicates whether the acquisition is cross-border (1) or domestic (0). 'Total synergies' is the total amount of synergies presented with an acquisition, divided by the acquirers EBITDA from the year of the acquisition. Statistical significance at a 10%, 5% and 1% level is denoted with *,** and ***.
Final result from the first regression is that horizontal transactions have a significant negative influence on the announcement return. This does not correspond with Singh and Montgomery (1987) who suggest horizontal transactions to generate a better return for the acquirer.
Table IV.6 reports the results of the regression of equation 9. The multivariate analysis of equation 9 presents one variable that influences a dependent variable. Cross-border transactions appear to have a negative relationship with the completion return. As no apparent reason for this relationship can be given, it is hard to draw any conclusions. The cost and revenue synergies do not show any significant relationship with the dependent variables. Also the other control variables do not have coefficients that are significantly different from zero.
Table IV.6: Results regression, cost and revenue synergies (equation 9)
Rumor Announcement Completion Premium
Coefficient Prob. Coefficient Prob. Coefficient Prob. Coefficient Prob.
Intercept -0.044 0.313 -0.008 0.820 0.025 0.201 0.200 0.000 Horizontal 0.004 0.918 -0.041 0.124 -0.001 0.956 -0.021 0.589 Cross-border -0.022 0.525 -0.005 0.857 -0.029 0.058* 0.058 0.140 Cash 0.063 0.123 0.018 0.534 -0.019 0.247 0.028 0.524 Other info -0.009 0.866 0.054 0.329 0.004 0.885 Cost synergies 0.256 0.278 0.164 0.338 0.097 0.327 -0.004 0.987 Revenue synergies -0.123 0.784 -0.075 0.703 -0.126 0.262 0.075 0.801 Observations N = 22 N = 54 N = 54 N = 54 Adj. R-squared 0.245 -0.017 0.059 -0.014 F-statistic 0.812 0.856 1.550 0.849 Prob. (F-statistic) 0.577 0.534 0.183 0.522
For an explanation of the dependent and independent variables in this regression see table 11. 'Cost synergies' is the amount of cost synergies presented with an acquisition divided by the acquirers EBITDA from the year of the acquisition. 'Revenue synergies' is the amount of revenue synergies presented with an acquisition divided by the acquirers EBITDA from the year of the acquisition. Statistical significance at a 10%, 5% and 1% level is denoted with *,** and ***.
