Mergers & Acquisitions and Market Concentration in the
European Market for Beer:
An Event Study
Sascha Leurs – 5876958
Master’s Thesis Economics – Track: Markets and Regulation Faculty of Economics and Business - University of Amsterdam
July 9 2017, Haarlem Supervisor: Jo Seldeslachts
Statement of Originality
This document is written by Student Sascha Leurs who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
This thesis examines the relationship between market concentration and abnormal stock market returns for companies involved in a merger or acquisition as well as their rivals. The results from previous studies give no clear evidence on the relationship between the two variables and can therefore be seen as ambiguous. We use the event study method to assess the sign of this relationship. It focuses on the European beer industry and uses data from M&A’s from 2002 up to and including 2015.
We find that the sign and size of the abnormal returns estimated mostly align with the returns found in many other event studies, but the relationship between market concentration and these returns differs for the firms involved: there seems to be no relationship when it comes to acquiring firms, a positive one for targets and a negative one for rival firms.
This finding largely underwrites the conclusions from several previously conducted studies but also shows that it is not a simple task to specify the connection between market concentration and abnormal returns.
Table of Contents
1. Introduction 4
2. Mergers and Acquisitions 6
3. The Beer Market 6
4. Methodology 8
4.1. Event Study 8
4.2. Market Model 9
5. Literature Review 11
5.1. Previous Event Studies: Abnormal Returns 12 5.2. Market Concentration and Stock Market Returns 14 5.2.1. Industrial Organization-based Theory and Studies 14
5.2.2. Finance-based studies 18
5.2.3. Conclusion 20
6. Research Question 21
7. Data 21
8. Results 25
8.1. Acquirers, Targets and Rivals 25
8.2. Correlation Results 29
9. Conclusion 35
1. Introduction
Over the past decade, mergers and acquisitions have been big business in the European beer market, leading to an increase in concentration in the market. Production in Europe currently accounts for more than twenty-five percent of the world’s beer production and the market consists of breweries of many sizes, ranging from microbreweries to global market leaders. (Jírovec and Calleja, 2013)
Recently, the two largest brewing companies in the world, SABMiller and Anheuser-Busch InBev, announced that they were planning to merge into one firm. Worldwide, AB InBev currently has a market share of approximately 20%, SABMiller 10% and third in the list of largest brewers is Heineken, with about 9% (TIHM Horecamakelaardij, n.d.). This type of merger, and the increase in concentration that inevitably goes with it, leads to questions concerning the brewing industry, considering the link between market concentration and stock market returns.
This thesis uses the event study method to analyse investor responses through abnormal stock market returns as reactions to announcements of mergers and acquisitions in the European beer market for the years ranging from 2002 up to and including 2015. The aim is to investigate whether there is a link between investor responses to merger and acquisition announcements and the market concentration in this industry.
Based on the examined literature we expect market concentration to have a
relationship with stock market returns. Which direction this relationship takes differs among theories and studies. Economic theories from the industrial organization field (Bertrand and Cournot) give a positive sign to the relationship, while many conducted studies focusing on finance state a negative relationship. Several studies have been conducted to assess the link between market concentration and (abnormal) stock market returns in the context of mergers, although most have analysed the United States market, and have not focused the research to one particular industry.
We therefore aim to add to existing research not only by focusing on one specific industry, but also by looking at the possible relationship between market concentration and investor responses (stock returns) in this industry. The main question we address is: How has the degree of market concentration in the European beer market influenced the effects of merger & acquisition announcements on abnormal stock market returns?
The remainder of the paper is organized as follows. Section 2 gives some information on the theoretical background on M&A’s. Section 3 gives an introduction to the European beer market. Section 4 describes the methodology used in the paper. Section 5 explores the theoretical framework that lays the foundation for this thesis, followed by the research question in section 6. Section 7 describes the sample construction and all the data used in the analysis. Section 8 documents the main results of the paper. Section 9 concludes the paper.
2. Mergers and Acquisitions
This section gives some theoretical background to mergers and acquisitions (M&A’s). Mergers and acquisitions is the general term for the consolidation of companies or assets. The difference between the two is as follows: a merger refers to a combination of two or more firms to form a new firm, while an acquisition is the purchase of one firm by another with no new company being formed. These transactions are generally divided into different types, based on how they are financed or their legal nature (Coates, 2014).
3. The Beer Market
This section serves as an introduction to the beer market itself and the developments that have characterized it over the years.
As mentioned in the introduction of this paper, many breweries of all sizes are active in Europe. According to Beer Statistics 2015 from The Brewers of Europe, the number of active brewing companies in Europe 3786 was in 2009 and grew to 7094 in 2014. In 2014, production in Europe totalled 400.619.000 hectolitres, compared to 409.315.000 hl in 2009. Total exports, intra-EU as well as extra-EU, amounted to 78.298.000 hectolitres (up from 64.732.000 hl in 2009). Most beer is traded within the Single Market, but exports to outside the EU have been rising continuously since 2000. Meanwhile, consumption in the EU has fallen during the same period from 378.531.000 hl to 369.576.00 hl. (Jírovec and Calleja, 2013).
In 2004, the top ten brewing companies in the world controlled about 51% of the market, by volume (Roach, 2015). In 2014, the four leading brewing companies were already up to 45.7% (Statistica, 2016). The global concentration rate (by volume), measured by CR5 increased from 25.4% in 2000 to 46.3% in 2009. Using instead the CR10 shows growth from 37.3% in 2000 to 59.3 % in 2009. It can thus be concluded there has been a great increase in concentration in the market. On the European market, the picture doesn’t differ much from that on the global market. In 2008 the three largest companies, InBev, SABMiller and Anheuser-Busch, controlled about 1/3 of the European market. Including the market shares
from Heineken and Carlsberg, that added up to almost 50%. (Smoluk-Sikorska and Kalinowski, 2013)
For a little over a decade now, the beer industry has been particularly prone to horizontal mergers between brewers. Over the past years, the industry has seen announcements of mergers and acquisitions almost on a monthly basis. Many of these consolidations happen between smaller brewing companies, but these are not the events gathering attention. This consolidation tendency has also been especially present among the large brewing companies. The observed increase in market concentration can be largely attributed to the consolidation behaviour of firms already enjoying relatively larger market shares. Even the top ten brewing companies in the world form no exception, a few extreme examples being InBev & Anheuser-Busch in 2008 and Molson Brewery & Coors Brewing Company in 2005. These merged companies, in 2014, had global market shares of 20.8% and 3.2%, respectively (Statistica, 2016). As the market concentration has increased drastically, despite the emergence of countless small and microbreweries, the consolidation wave can be seen as having played an important role in this development (Madsen, Pedersen and Lund-Thomsen, 2011).
4. Methodology
4.1. Event Study
An event study uses financial data, often share prices, to assess the effect of a specific economic event on the market valuation of a firm and its rivals (Davies and Ormosi, 2012). The methodology used in this assessment is based on the assumption that the discounted value of a firm’s future profit stream is represented by its share prices. It is an ex-ante analysis, which could be useful to predict future profitability, as financial markets are supposed to be forward-looking. (Duso, Gugler and Yurtoglu, 2010).
Event studies are typically used to analyse merger announcements. It analyses events by measuring abnormal returns that have resulted from the event (merger). An abnormal return is the difference between the actual change in share price that resulted from the event and the counterfactual, i.e. what would have happened if the event had not occurred (Davies and Ormosi, 2012). Herein lies the difficulty: identifying the right counterfactual. An event study tries to predict this
An important aspect of the event study methodology is the assumption of rational market, or the efficient market hypothesis (EMH). This states that if the market is efficient, share prices accurately reflect a firm’s discounted profits. A change in share prices would thus immediately show the effect of any event on discounted profits. (Davies and Ormosi, 2012; Duso et al., 2010)
Duso et al. (2010) compare the use of an event study using stock market reactions to a merger announcement to using accounting data as an ex-post assessment (meaning after the event has happened). They conclude using an event study may be more attractive to competition
authorities. Using accounting data may cause bias in the assessment, as the parties involved are the ones providing the information. Using share prices is a way to get rid of this
information asymmetry (when one party involved in a transaction has more or superior information than the other). Event studies are also argued to be easy to carry out due to the data needed being easily accessible. This, however, depends strongly on the market and firms being analysed, as they need to be listed on stock exchanges.
When it comes to analysing the results from event studies, Davies and Ormosi identify that there might be some ambiguity. For example, if returns for the merging parties increase there can be two explanations: anticompetitive effects (i.e. collusion, exclusion) or
pro-competitive effects (efficiency gains). Diepold, Feinberg, Round and Tustin (2008) also notice that abnormal returns of the firms involved do not shed any light on the competitive impact of M&A’s. They state: “Positive abnormal returns to the firms involved simply tells us that investors expect the merger to reflect well on profitability, but not why”.
This is where analysing rivals of the merging firms becomes useful: the effect of the event on the rival firms can mostly remove this ambiguity. When looking at returns as an indication of efficiencies, a decrease in rivals’ stock prices/abnormal returns would show the expectation in the market that the merger/acquisition will have pro-competitive effects: there are efficiency gains present putting the rival at a competitive disadvantage as the marginal costs of the merged firm will decrease. Due to the subsequent increased competition, it would lead to a decrease in the rivals’ price and profitability. On the other hand, if rivals’ share prices/abnormal returns increase it would indicate anti-competitive effects and the event generating market power.Market power increases will likely cause the merged firm to raise its price, which means gains to its rivals as they will be able to increase their prices as well (Rabello, Perobelli and Vasconcelos, 2015). This is why not only looking at the firms
involved but the rivals as well gives more insight into the effects M&A’s have in the industry. Sonenshine (2011) does add some critique to the reliance on analysis of rivals. Rival firms may often not be similar in size to the merging firm of might be a multi-product firm. It can therefore be the case that only a fraction of their revenues are derived from the market affected by the event. Other events in the window investigated may also influence their response which is then unjustly attributed to the event analysed. Insignificant results for rivals can thus possibly indicate that competitors were unaffected by the event instead of
interpreting them as efficiency-enhancing. The use of a static model should also be taken into account, as this gives a time-independent view and thus does not change with time. Some care in interpreting results is thus in order.
All in all, based on the studies performed by Duso et al. and Davies and Ormosi, it seems that the event study methodology can be useful to analyse mergers.
4.2. Market model
There are several models used in estimating abnormal returns when conducting event studies. A study performed by Armitage (1995) compares, among other things, several of these
studies that aimed to test the performance of these models conducted by, among others, Brown and Warner in 1980 as well as 1985.
Comparing these models, Amitage comes to the conclusion that the most widely used model in event studies, the market model, also performs the best.
The model I will use is therefore the market model, which can be written as follows:
ARit = Rit − E(Rit|Xt) (1)
where ARit, Rit, and E(Rit|Xt) are the abnormal return, the actual return, and the expected returns for period t. The Xt is the conditioning information for the market model that is shown below to estimate normal returns.
Rit =αi + βiRmt + εit (2)
with E (εit) = 0 and var (εit) = s2εi where Rit and Rmt are the period-t returns on the stock i and the market m and with εit an i.i.d. normally distributed error term.
The market model uses the return relationship (formula 2) between the actual return and the return on a chosen market index, for a specific time frame before the event, to estimate, via regression analysis, the counterfactual: the returns on the stock that one would expect to see were there no event. The expected return (the “normal return”) is then calculated based on these estimates. Ordinary Least Squares estimation is generally a consistent
procedure to estimate the parameters of the market model (MacKinlay, 1997).
A limitation when using this basic market model is the prediction of the investor returns for the counterfactual. To do this we take share prices from a number of trading days before the announcement and use these in a regression analysis. However, share prices can be quite volatile so it is difficult to say if they can be predicted accurately. To limit this bias, I will use control variables. Control variables are related to the dependent variable and thus may affect the outcome of the research. By adding these variables as control variables, they are kept constant and their effects are removed from the equation. Finally, a significance test will be performed to assess the significance of the abnormal returns retrieved from the model.
5. Literature Review
Mergers, acquisitions and market concentration can undoubtedly be seen as linked. The concept of two firms joining forces causing an increase in market concentration needs no further explanation. It is thus with good reason that concentration, as well as market share, play a significant role in antitrust regulation. The European Commission states the following in their Merger Guidelines: “Market shares and concentration levels provide useful first indications of the market structure and of the competitive importance of both the merging parties and their competitors. “
In their assessment of the competitive implications of M&A activity the EC follows two steps. The first is to define the relevant product and geographic markets. The second is a competitive assessment. This assessment focuses on market share and concentration (given by the Herfindahl-Hirschman Index). Based on the market shares and concentration levels they decide whether or not to further consider a potential merger. Generally, they find that a merger with a market share of 50% or more and HHI levels between 1 000 and 2 000 and a delta above 250, or a merger with a post-merger HHI above 2 000 and a delta above 150, competition concerns are likely to be identified (with several exceptions).
Economic theory focused on industry structure states that market concentration is based on the link between the size of the market and optimal firm size (the size at which a firm operates at the lowest possible average costs). If one of the two changes, so does the concentration in the market (Ornstein, 1981). With a larger optimal firm size relative to the size of the market, concentration will be higher as well, and vice versa. So, when the size of the market changes, for example due to product innovation or buyers’ tastes, while optimal firm size remains the same, market concentration will change.
According to Ornstein (1981), mergers can be seen as a way to obtain minimum efficient scale (when the size of a firm is such that it is fully exploiting economies of scale and thus operating at the lowest possible average costs). He argues that a change in
concentration itself is therefore not the direct result of a merger, but of exploitable economies of scale relative to the size of the market that cause merging to be attractive to firms in the first place.
A difference between the view of the EC and that of Ornstein when it comes to the direct link between mergers and concentration can be noticed here. Ornstein attributes this difference to governments and antitrust authorities believing potential gains to productive
efficiency are generally overshadowed by the potential allocative efficiency losses (meaning losses for consumers due to increased prices are higher than the gains for producers from lower production costs).
In the following subsections, we will delve deeper into the existing literature
pertaining to the link between market concentration and (abnormal) stock market returns as well as some previously conducted event studies to assess abnormal return results.
5.1. Previous Event Studies: Abnormal Returns
Event studies have been used in economic and financial research many times over the years. This trend started with Fama et al (1969), who conducted and event study analysing the effect of a stock split has on the stock market.
Since then, many event studies have been conducted, with their focus ranging from country-specific research to entire continents. When it comes to mergers and acquisitions, most of these studies have aimed to analyse the effects of involvement of competition
authorities on investor responses. In this section, we shortly discuss event studies to use their results to later compare to our own.
We have used the results from four event studies to compare and it seems that they generally come to the same conclusions despite their different focus/set-ups (models, countries and types of firms included time frames within which the research was conducted).
Acquiring firms seem to produce insignificant returns, ranging from slightly positive to slightly negative. Sonenshine (2011) mentions that the small returns can be seen as a result of many target firms being small and thus not very significant compared to the acquirer’s value. Diepold et al. (2008) find that the reaction from acquiring firms changes depending on whether a merger is raised with the Australian antitrust enforcement or not.
Campa and Hernando (2002) find that acquirers more often earn negative returns in cross-border mergers than domestic ones.
The results for target firms are found to be positive and often statistically significant in all the included studies. Clougherty and Duso (2009) and Sonenshine (2011) both find that three-day cumulative average abnormal returns are around 3.5%. Contrary to findings in previous studies, Diepold et al. (2008) notice that in their study of the Australian firms, the abnormal returns to targets on the day after the announcement are almost as large as on the announcement date. As an explanation, they bring forward the possible slow reaction of
Australian investors. Campa and Hernando (2002) add that targets in non-regulated industries tend to receive higher returns compared to regulated industries.
Rival firms’ abnormal returns are included in three of the four studies, and are mostly found to be small but positive and statistically significant. This finding appears to fit the theory that abnormal returns to rivals are based on investor expectations that rivals would gain pricing power as a result of an increase in market concentration. As the abnormal returns found are small there could be other factors influencing investor behaviour in this case (Sonenshine, 2011). Clougherty and Duso (2009) conclude that rivals generally gain from merger activity by their competitors.
Sonenshine (2001) adds to this research by also running a separate regression of the three-day CAR regressed on the change in market concentration measured by the Herfindahl Index. Looking at the effect of a change in market concentration, the coefficient on the change in HHI is positive and significant. He sees this as possible evidence of investors in a more concentrated industry believing the merger to be of larger value.
Covin, Daily, Dalton and King (2004) conduct a meta-analytic review of M&A performance. Meta-analysis uses multiple studies to try to establish the best estimate for a true population relationship. The authors use data on abnormal returns from 93 previously conducted event studies. They find that acquiring firms have small but significant abnormal returns on the day of the announcement of the merger/acquisition. After the announcement, the abnormal returns become insignificant and/or negative. The acquired firm abnormal returns are only considered on the day of the announcement, and found to be positive and significant (0.70%).
Covin et al. conclude that the true population relationship between M&A activity and the performance of acquiring firms is close to zero or negative beyond the day a merger or acquisition is announced. On average, completed acquisitions do not seem to improve the financial performance of acquiring firms after the day of the announcement. Rather, they seem to have either insignificant or slightly negative effects on their performance.
Taking the literature reviewed into account, we find that most of the results from the studies follow the same direction: target and rival abnormal returns are generally found to be positive and significant, whereas abnormal returns to acquirers range from negative to slightly positive (and generally insignificant). We do not expect to find significantly different results for the calculated abnormal returns in this study.
5.2. Market Concentration and Stock Market Returns
This section reviews existing literature on the relationship between market concentration and stock market returns. We look at theory and studies from the industrial organization field as well as several studies from the finance field to assess this relationship.
5.2.1. Industrial Organization-based Theory and Studies
In this section, we try to delve deeper into the theories behind the relationship between market concentration and returns (stock prices) and include some additional studies.
As mentioned in section 4.1., a merger can have two opposing effects on stock prices due to efficiency gains and market power. Which effect dominates will determine what happens to the prices of the merged firms, as well as the rivals. If a merger gives the merged firm the opportunity to utilize efficiency gains, and thus save costs without facilitating price increases, the merged firms benefit but rival firms will not. Prices will likely decrease and the rival will lose profit, leading to lower stock prices for the rival.
If, however the market power effect dominates the merged firm will increase its price, giving the rival(s) the opportunity to increase their product price as well. The stock prices of either the rival or all firms go up (Lindsay and Berridge, 2012). Röller, Stennek and Verboven (2006) find that the market power effect tends to dominate.
To assess the relationship between market concentration and stock returns, we look at the link between the following variables:
market concentration à market power (market prices) à stock prices (returns) We look at theories and studies used in the industrial organization field that focus on the link between these variables in general as well as specifically for mergers. First, we focus on the relationship between market concentration and market power. Secondly, we look at the relationship between market power and stock prices.
Before we look at specific models, it is important to define “market power”. Market power is defined as the ability of a (dominant) firm to profitably raise its price above competitive levels. Whether a firm has market power and to which extent can be measured in different ways. One of the most well-known measures (also used by the European Commission) is the Hirschman-Herfindahl Index (HHI), which is a measure of concentration based on the number of firms in a market and their market shares. (OECD, 2002)
Another important measure is known as the Lerner index: 𝐿 = 𝑃 − 𝑀𝐶
𝑃 = −
1
𝜀
where P is the market price set by the firm and MC is the firm's marginal cost. S is the firm’s market share,
ε
is the firm price elasticity of demand. The index ranges from 0 to 1, with higher numbers implying greater market power. In a perfectly competitive situation where P=MC, the index will be equal to 0: the firm has no market power. When MC=0, Lerner Index is equal 1, indicating the presence of monopoly power. (ICT Regulation Toolkit, n.d.) Several theoretical economic models focus on competition. Two of the most well-known are the Bertrand and the Cournot models. The Bertrand model focuses on price competition, while the Cournot model focuses on output competition.The Bertrand model states that each firm in the market faces the same market demand and has a profit maximization goal, given the price of its competitors. Assuming homogeneous goods and symmetric costs, in equilibrium firms will price at marginal cost, meaning that no firm will make a profit. The only profitable mergers in this situation of price competition are mergers to a monopoly situation. With more firms, price competition will lead to the
equilibrium price, as firms will undercut each other to increase their market share until price equals marginal costs. This means that the amount of competition (and concentration) in the market has no effect on prices and merger profitability when it does not lead to a monopoly situation (so as long as post-merger n>1), goods are homogenous and costs are symmetric among firms.
With asymmetric costs, only mergers to monopoly or which include the two most efficient firms are profitable. When the two lowest-cost firms merge, they can charge a price equal to the marginal costs of the next lowest-cost competitor, instead of the marginal costs of the least efficient merging firm. In these versions of the model, we can say that when market concentration is highest (monopoly or duopoly between the most efficient firms), prices increase. This ability to increase prices shows that the merging firms gain market power (as also seen in the Lerner Index). (Tremblay and Tremblay, 2009)
Adding a degree of product differentiation, each firm faces a demand that is downward-sloping for all levels of their own price, but an increase in a competitor’s price increases the firm’s demand curve. Say there are three firms in the same relevant market. Firm A sells good A, firm B sells good B and firm C sells good C. Any firm wanting to raise the price of their good would lose sales as consumers will switch to another of the gods or leave the market. Now firms A and B merge, and thus increasing market concentration. Now, if the merged firm increases the price of good A, sales that would be lost to good B before the merger are no longer lost as sales for good B stay within the new merged firm, and vice versa. This may give an incentive to increase prices, and increases market power (as the market concentration increases due to the merger). (Baker and Bresnahan, 1985)
The Cournot duopoly model is a model of imperfect competition that states that each firm has a profit maximization goal, given the output of its competitors, and have the same view of market demand. Each firm believes (correctly) that its competitors behave optimally. Each firm’s output decision is assumed to affects the product price. Cournot’s approach defines optimum prices by maximising both market share and profitability, to avoid the problem of firms losing market share due to higher prices. The optimum price will thus be the same for both companies, as otherwise the firm with the lower price will obtain full market share.
In a situation of homogenous goods and symmetric costs and with two firms in the market, both will produce half of the market output in equilibrium and the market price will be in between marginal costs and the monopoly price. With asymmetric costs, both will produce but the firm with the lowest costs will produce a larger share of the market output than the other. When the number of firms increases and market concentration decreases, the equilibrium heads more and more in the direction of the competitive equilibrium, and thus lower prices and less market power. (Tremblay and Tremblay, 2009)
Many price-concentration studies have been performed over the years. Singh and Zhu (2008) as well as Weiss (1979) look at several of these studies performed in different industries and find that the general finding among them is that higher concentration is associated with higher prices.
Newmark (2004) collects information from many studies concerning the price-concentration relationship, ranging from 1989 to 2002 and also finds that most studies conclude that they are positively related. This thus aligns with the economic theory from the Bertrand and Cournot models.
To conclude, there seems to be a positive relationship between concentration and product price. This is the case with concentration changes due to mergers (Bertrand and Cournot) as well as concentration of the market in general (price-concentration studies).
Next, we focus on the relationship between market power (product price increases) and stock prices/returns, in general and in the case of mergers. First, we note that an increase in stock prices indicates a higher return (Lorette, n.d.).
As we have seen from the Bertrand and Cournot models and the price-concentration studies, an increase in concentration generally leads to an increase in prices and market power. As a firm behaving rationally would not increase its product price if it did not lead to any higher profitability (under the assumption that firms behave profit-maximizing), expected
profitability increases. When investors expect a higher profitability in the (near) future, the stock price (and subsequent return) increases. Expected future cash flows for rivals increase as well as they also increase their product price, leading to an increase in stock prices as investors expect a higher profitability.
Singal (1996) studies airline mergers and integrates stock market and product market data. He finds evidence of a link between the stock market and the product market:”
Abnormal stock returns are correlated with profit changes signifying that the stock market anticipates profit changes and adjusts accordingly. Overall, enhancement of market power by airline mergers is supported both by stock prices and product prices.” He also finds that for rival firms, abnormal returns are positively correlated with changes in concentration,
consistent with increased market power. The merging firms do not benefit as much from these increases in concentration. This is explained by the free rider problem associated with a dominant firm model: those who benefit do not pay for it.
This conclusion is backed up by Sonenshine (2011), who used the event study methodology to assess the effects of R&D and (changes in) market concentration on merger outcomes in the United States. He uses the three-day cumulative abnormal return regressed on the change in concentration measured by the Herfindahl Index.
Looking at the effect of a change in market concentration, he finds the coefficient on the change in HHI is positive and significant for target firms. Sonenshine sees this as possible evidence of investors in a more concentrated industry believing the merger to be of larger value. The coefficient for rival and acquiring firms was however insignificant.
The event studies from section 5.1. also underwrite the finding that there is generally a positive relationship, as mergers increase concentration and most firms’ stock prices have increased.
To conclude, we find that there is a positive relationship between product
prices/market power and stock prices, as well as a positive relationship between concentration and product price/market power. This leads to the conclusion that there is a positive
relationship between concentration and stock prices. 5.2.2 Finance-based studies
The literature discussed in this section all focuses on the specific relationship between market concentration and stock returns.
Hou and Robinson (2006) identify several potential risk-based channels through which market structure may affect stock returns, drawing on classic industrial organization work from Schumpeter (1912) and Bain (1954) such as the Structure/Conduct/Performance
paradigm. This paradigm explores the cause-and-effect relationship between market structure (number and size of firms, barriers to entry etc.), conduct (pricing, promotion, R&D) and performance (degree of production and allocative efficiencies, technological progress etc.) The paradigm makes the following predictions: that concentration will facilitate collusion, whether tacit or explicit, and that as barriers to entry rise, the optimal price-cost margin of the leading firm(s) likewise will increase (Weiss, 1979).
Product market structure is known to affects managers’ equilibrium operating
decisions. If these decisions in turn affect the risk of a firm’s cash flows, then these decisions should impact stock returns. They focus on two channels through which this happens:
Hou and Robinson mention on innovation: “If innovation is risky, and this risk is priced, then this predicts that competitive industries or firms on the competitive fringe of established industries earn higher returns, all else equal.” As suggested by the S/C/P paradigm, barriers to entry also affect expected returns when differences in the number of competitors in an industry, or in the pricing practices they observe, change the risk
characteristics of the firms. When firms that are, for example, on the competitive fringe or potential entrants innovate, this poses a certain risk that established firms do not face. As this risk is priced, higher returns are made.
To assess these statements, they run regressions on data ranging from 1963 to 2001 from US based stock exchanges to find that firms in highly concentrated industries earn up to 4% lower annual returns than the firms in the most competitive industries. This result holds for industry portfolio as well as individual firm-level returns, even after controlling for known return predictors such as size and book-to-market ratio. They argue that concentration affects the cross-section variation (the variation between returns from different stocks or portfolios) of investor returns and conclude that industries in which innovation and distress risk is higher, which are found here to be competitive industries, have higher expected returns. The
explanation for this would be that investors in more competitive industries (associated with more innovation and distress risk) face higher risk and thus expect a higher return premium. Hashem and Su (2015) do fairly the same analysis but for the UK to see if the results from the US analysis hold in another, similar, market. They use data from the London Stock exchange and, just as Hou and Robinson, use this data to run regressions. They find results consistent with the findings of Hou and Robinson: a negative relationship between
concentration and stock returns. They too rely on the SCP paradigm to explain these results. They argue, “Concentrated industries engage less in innovations and face lower innovation risk compared with competitive industries. Second, concentrated industries have higher barriers to entry, which protects their firms from distress risk. Therefore, investors should anticipate lower risk-adjusted stock returns associated with lower innovation and distress risks in concentrated industries.”
Sharma (2010) follows the research conducted by Hou and Robinson as well. This continuation of the research is based on the premise that market concentration cannot be the only factor that captures product market competition and thus affect asset prices. They argue that stockholders of firms with greater product substitutability will demand higher returns. This is due to the diminished pricing power of easily substitutable products. They use approximately the same data sources, but add different measures for product market
competition. The first is based on the Lerner Index: the PCM: price-cost margin. The second is based on industry market size, using average of aggregate sales.
Their conclusions align with that of Hou and Robinson but add several new findings. Firstly, firms with higher product substitutability earn higher returns than those with lower substitutability. Secondly, firms in industries with relatively larger market sizes earn higher returns than those in smaller markets. Third, they find some evidence, although very weak, that firms in more competitive markets are riskier during a recession, since they earn higher abnormal returns during a recession compared to a boom. Lastly, they find that both product substitutability and market size have explanatory power with respect to stock returns. The results are statistically more reliable for product substitutability. They mention their results “highlight the multi-dimensional structure of product market competition and its impact on asset prices. “
Gallagher, Ignatieva and McCulloch (2013) again follow in the footsteps of Hou and Robinson, but analyse the Australian equity market over the period of 1993 to 2007. In contrast to the other findings mentioned above here, they do not find evidence for higher returns in more competitive environments. In fact, they find the opposite: firms in more concentrated industries earn higher returns. This finding may be due to the nature of the Australian market: it is a relatively small open economy where larger companies with higher market power dominate concentrated industries. They hypothesize the higher returns are generated by companies with monopoly rents investing these rents in product and process innovation. As the outcomes of these investments are uncertain and thus increase risk, investors require higher stock returns.
5.2.3. Conclusion
Comparing the literature on the link between market concentration and stock returns, we find that most results from the finance-based studies generally align with the results from Hou and Robinson that the relationship is a negative one, with the exception of the study for Australia, which can be explained by the specific nature of the market there. However, the industrial organization-based theory and studies suggest otherwise. This leaves us with the notion that there is indeed a relationship between market concentration and (abnormal) stock returns. What direction this relationship takes exactly, remains ambiguous.
6. Research Question
The research question this thesis aims to answer through these hypotheses is the following: • How has the degree of market concentration in the European beer market influenced
the effects of merger & acquisition announcements on abnormal stock market returns?
7. Data
To answer our research question and hypothesis, we look at 18 mergers/acquisitions between brewing companies that took place between 2002 and 2015. All target firms were based and active in Europe. All acquiring firms were also already active on the European beer market prior to the M&A event. For some events, there is no or not enough data available to include the target as well as the acquirer in the analysis, so our sample size consists of 16 acquirers and 12 targets. The percentage of shares of the target company acquired in all events included is at least 50%. The list of acquirers, targets, the country out of which they operate, the
announcement date and the percentage of shares acquired in/owned after the transaction can be found in table 1. Companies in brackets are not included in the analysis due to lack of data.
Per merger we define a number of rivals, depending on the availability of data. We do so by using the rivals (or through the definition of the relevant market) mentioned by competition
authorities in their analysis of the merger. All companies involved must be listed on a public stock exchange. This specific criterion, although necessary to conduct the study, affects the number of usable M&A’s immensely.
We use daily closing prices retrieved from Datastream to determine the actual return as well as estimate the normal return. The information on announcements dates was retrieved through the Thomson ONE database.
As a market index, we will use the Standard & Poor 500 index, which is value weighted. As mentioned by Armitage (1994), it does not seem to matter much whether an equally weighted or value weighted market index is used. Davies and Ormosi mention that the trade-off when using a general index is that it is less likely to be influenced by market-specific exogenous events, but more independent of the event studied. In this case we chose
independency of the events analysed.
To calculate the counterfactual, we use closing prices from 120 to 20 trading days before the announcement date of the merger. This way there are 100 observations. According to the study on event study methods by Armitage (1994), 100 days is sufficient for an
accurate estimation of a and b when using daily data. This way there will be a lower chance of bias due to possible leaking of information prior to the announcement. Trading days may differ between the stock and the market index. To make sure we have the same number of observations for all estimations when data is not available for all variables on the same date, we use the next date where data is readily available for the stock price, control variables and the market index.
To assess the abnormal returns, we will use two different time frames. The abnormal return (AR) will be calculated on the announcement date (event date). For the cumulative abnormal return (CAR) we will use three separate event windows: (-1,1), (-3,3) and (-5,5). These windows for the cumulative abnormal returns let the estimation capture effects of information that may be leaked prior to the official announcement and the effects of slow information distribution (Diepold et al., 2008).
The two control variables used in the estimation of abnormal returns per firm are firm size and the book-to-market ratio. These variables have been widely used in literature
pertaining to market concentration and stock market reactions, for example in Hou and
Robinson (2006) and Gallagher and Ignatieva (2013). These two variables are considered risk factors and return determinants. As mentioned before, in part 3.1, risk is seen as an important
factor in the relationship between concentration and returns. This would make company characteristics directly affecting risk as well as return useful as control variables.
The variable firm size will be measured by market capitalization: the market value of a company’s outstanding shares. Market capitalization is a widely used measure to classify companies’ sizes. This is a useful variable in two ways: stock prices alone tend to
misrepresent the actual value of a company and capitalization gives insight into the growth and risk potential of a stock. Firms with large capitalization have, historically, experienced lower risk accompanied by slower growth. Firms with a small capitalization have a tendency for more potential to grow, but with higher risk. (Investopedia, 2015) The data used for this variable is retrieved from Datastream. The data is in millions of US dollars.
Lastly, the book-to-market ratio will be included. The book-to-market ratio is calculated by dividing a firm’s book value by its market value. It tries to identify over- or undervalued stocks. If the ratio exceeds a value of 1, the stock is assumed to be undervalued and vice versa. In Datastream the inverse ratio can be retrieved: the market-to-book ratio. By using the following the needed ratio will be calculated: 1/market-to-book-ratio).
Due to including these control variables and the variable for concentration, instead of equation (2), specified in section 3.2, we now get:
Rit =αi + β1iRmt + β2i ln(MCit) + β3i ln(BMit) + εit (3) with E (εit) = 0 and var (εit) = s2εi where Rit and Rmt are the period-t returns on the stock i and the market m. MCit and and BMit are the market capitalization and the book-to-market ratio for firm (stock) i in period t, respectively. Equation (1) will then be used to calculate the abnormal return.
To test the significance of the found results we will run several t-tests. We will run these for the average of the abnormal returns as well as the cumulative abnormal returns for all acquirers, targets and rivals separately. In addition to this, we will also look at the
t-statistic for the average AR and the average CAR for all targets together, as well as for all the acquirers and all the rivals.
When assessing the average abnormal returns for all firms combined, we add two control variables: whether the merger/acquisition was domestic or cross-border and the percentage of shares that was acquired (less than 75% versus more than 75%).
As this thesis focuses on the relationship between investor responses and market
concentration, we need a measure for this market concentration. For this, we will use the Herfindahl-Hirschman index (HHI):
N
H = S si2 (4) i=1
Where si2 is the squared market share of firm i in the market, and N the number of firms. The HHI can range from zero to 10.000. The higher the index, the higher the concentration level in the market.
We will calculate the HHI separately for each merger, based on the rivals and/or relevant market defined per merger by competition authorities. First, we calculate the market shares by dividing the total net sales of the target and the rivals defined for each merger by their separate net sales. Then we add these into equation (4) to get the HHI. Data on net sales is retrieved from Datastream and the Orbis database.
To assess the impact of concentration in the industry on returns we will run several pair wise correlations. We will look at the correlations between the HHI and the abnormal returns at announcement date, AR (0), the average abnormal return over the (-1,1) window and the cumulative abnormal returns over all three event windows. This will be done separately for the targets, acquirers and rivals.
If the correlation between the HHI and the abnormal return is positive and significant for the target and the rival firms during the event, this would accommodate the hypothesis that market power is generated by anti-competitive events, since as market concentration
increases, the target firm and rivals will be able to price above marginal cost and increase their profitability. (Sonenshine, 2011)
8. Results
Presented here are the results derived from the equations that were explained in section 6. In section 7.1 are the abnormal returns for the sample of mergers and acquisitions. All estimated abnormal returns are derived from the market model, which is represented by equations (1) and (3). These serve the purpose of comparing the sample of M&A’s in the European beer industry to the results from previous studies. The found correlations between the estimated abnormal returns on the concentration index can be found in section 7.2. 8.1. Acquirers, Targets and Rivals
Table 2 represents the share-price response to the announcement of an M&A event found for all acquirers and targets as separate groups. The first three rows represent the abnormal returns for all targets and acquirers for the day prior to the event, the event day and the day after the event. In the fourth through seventh row we find the average abnormal returns over the event window (-1,1) for all targets and acquirers, as well as the cumulative effect over the three previously specified event windows: (-1,1), (-3,3) and (-5,5). Significance is assessed by t-statistics, where significance indicates the abnormal returns are “different from zero”.
Looking at the CAR’s, the targets’ abnormal returns are positive for all CAR windows and significant for the (-3,3) window. This is in line with the studies discussed in section 5.2. The CAR’s of the acquirers are negative, contrary to previous studies. In keeping with most other studies however, the acquirer CAR’s are also found to be insignificant. The same applies to the AAR’s. The target CAR’s are comparable to those found in the study performed in Europe by Campa and Hernando, but lower than results in studies performed in other
countries (Diepold et al., 2008). The positive returns on the announcement day for the target firms can be explained by investor behaviour. As Arora and Shah (2014) mention: “Investors are aware about the high amount of premiums paid to the target firm if the deal is closed. This leads to high returns to the target firm’s shareholders if the merger is underwritten. Thus, an announcement generally attracts investors to buy the shares of the target firm in hopes of earning higher returns in the future. “
Acquirer CAR’s are small as in other studies, but negative. The size of the returns may be attributed to the size of the targets. It is generally found that when large firms acquire small firms, the large firms’ investor responses are negligible, as the target will only have a small
effect on the acquiring company’s business prospects. Another factor to include is the control variables used in the calculation of the abnormal returns for all firms combined. As many of the mergers where the percentage shares acquired in the transaction had very negative abnormal returns, this may have skewed the results.
The AR’s for the acquirers for all three separate days are all small and not statistically significant, which also fits with the returns found in other studies.
Noticeably, the AR (-1) for targets and acquirers turn out to be negative (but not significant). This could indicate that there is, on average, no information leakage prior to the announcements, as that would lead to already positive returns on the day prior. The returns for targets on the day of the announcement are found to be 4.57% and significant. For acquirers, the returns are small and insignificant. The AR (1) is positive for both parties involved, but not statistically significant for the acquirer. We can conclude that he response for the firms thus takes place on the day of the announcement. This can possibly be explained by there not being any delay in information distribution.
Table 3 gives the merger-specific results for all three CAR event windows. As can be seen in the columns depicting the CAR (-1,1) for the merging companies, target reactions are found to be almost all positive, with more than half of the results (seven out of 12) significant. For the acquiring companies this number is much lower: only 5 out of 16 companies have statistically significant abnormal returns. Returns also differ among firms, ranging from -8.39% to 11.31%. Interesting is the obvious increase in abnormal returns, be it positive or negative, when targets are larger in size (compared to the acquirer). This could be due to investors
having higher expectations regarding these mergers’/acquisitions’ possible (un)successfulness when it comes to future profitability increases.
Comparing these results to the other two event windows, most returns seem to increase with the number of days included, but do not change signs (excluding some
exceptions). Including more event days seems to have little to no effect on significance either. Most returns’ significance level remained the same throughout all three event windows, Including more event days generally does not seem to have much of an impact for the analysis.
Table 4 represents the share-price response to the announcement of an M&A event found for all rival firms with enough available data. As can be seen in the table, most returns are found to be small, negative and insignificant. Only AR (0) and CAR (-3,3) are positive. Abnormal returns increase with the number of event days, but the event windows do not have much of an impact on the significance of the results. The positive outcome of the rivals’ abnormal returns on AR (0) can be an indication that the mergers in the sample were, on average, perceived as anti-competitive when the announcement was made. As mentioned in earlier sections, investors expecting anti-competitive effects and pricing power generated by events leads to positive returns to the rival companies as they are expected to be able to increase their profitability as the merging companies will likely raise their prices.
The returns found here are however very small and not statistically significant. On average one could conclude that investor behaviour may have been influenced by other factors. As mentioned before in section 3.1, rival firms may, for example, only derive a fraction of their revenues from the market in which the event takes place or simply be affected by other events during the investigation window. The small and insignificant results can thus mean the rivals were not affected by the event at all.
To give some more insight, table 5 shows the rivals’ CAR (-1,1), (-3,3) and (-5,5) for each merger in the sample. Investor reactions are shown to differ to a great extent among the events, with the CAR (-1,1) ranging from -2.88 to 14.99. Including more days in the event window had differing results. As with the targets and acquirers, some returns increased, others decreased. In the (-3,3) and (-5,5) windows, only one extra rival CAR was found to be
significant compared to the (-1,1) window. Interesting is that the rival abnormal returns found to be significant in the two larger event windows are not the same as the ones found to be significant in the (-1,1) window, except for one. This shows that adding more event days does change merger specific results, though not the overall results.
A negative abnormal return would be a sign of pro-competitive effects and efficiency gains for the companies merging. This would lead to an expected decrease in profitability for the rival firms and thus a reason for the negative reaction to the event. The positive cumulative abnormal returns are on average found to be higher than the negative ones, indicating several events to be perceived as highly anti-competitive. The t-statistics are generally low. This can be attributed to the small number of rival firms included in the analysis. Further analysis with a larger sample size is therefore recommended.
Based on the abnormal returns for the CAR window (-1,1), 56% of the M&A’s were seen as anti-competitive by rival firms’ investors. From these perceived anti-competitive events, only two met the standard significance levels. Of the perceived pro-competitive events, there were also only two meeting these standards. As both the pro-competitive and the anti-competitive events include mergers between firms of all shapes and sizes, the results cannot be attributed to firm sizes. They also include domestic a well as cross-border mergers so that does not seem to show any effect on these results either. Of course, the sample size in this analysis is quite small so no definitive conclusions can be made.
8.2. Correlation Results
Table 6 presents the results found when using pairwise correlation between the previously estimated abnormal returns and the concentration (by HHI) calculated for the relevant market from each merger/acquisition. The first part looks at all mergers, the second at all mergers weighted by size of the firm. Column 2 and 5 show the correlations between the HHI and the abnormal returns estimated for the acquiring firms for the announcement date, AR (0), the
cumulative abnormal returns for the event windows (-1,1), (-3,3) and (-5,5), and the average abnormal returns for the (-1,1) window. Below the correlations the corresponding p-value (to indicate significance) can be found. Column 3 and 6 show the same for targets, 4 and 7 for rivals.
The correlations for the acquirers for all mergers range from zero to -0,20. This indicates a negative relationship between the concentration in the relevant market for acquiring firms. None of these results were however found to be significant, far from actually. There thus seems to be no relationship between them. Using weights, the results remain negative and insignificant, albeit more negative with lower p-values.
HHI and target firms show negative correlations for all event windows specified when looking at all mergers. The correlations found for AR (0), CAR (-1,1) and CAR (-3,3) are not significant. With the exception of AR (0), the p-value is quite lower than those found for the acquirers. The highest negative values found for the correlations are also the ones found to be significant: CAR (-5,5) with a correlation of -0,58 and AAR (-1,1) with -0,61 are both found to be significant at a 5% level. When the mergers are weighted according to firm size, the results change drastically: all correlations become positive. None of these results are however
significant. The fact that the correlations become positive would indicate that larger target firms tend to show a higher correlation between returns and concentration in the market. This would seem to show that they tend to have higher returns when the market concentration increases.
The rival firms show, unlike the targets and acquirers, a positive correlation for all event windows. The correlations for CAR (-1,1) and AAR (-1,1) are found to be significant at a 5% level. The results from the weighted correlations are fairly similar in size but now in all but one event window they are found to be significant.
Table 7 shows the correlations when only using cross-border M&A’s and only domestic ones. The target and rival coefficients do not change much with cross-border activities. Returns for targets remain negative, with one less significant event window. Returns for the rivals are still positive, also with one event window no longer being found significant. With domestic events, the coefficients are more negative (with three event windows now displaying
significant results) for target firms and more positive for rivals compared to the results using all firms. It would seem target and rival firms thus tend to react more strongly to domestic events.
Acquirer results have changed: all coefficients expect for AR (0) have become
positive. They are still all found to be insignificant. Acquirer correlations from only domestic events are, on the other hand, more negative than with all mergers (except for CAR (-5,5)). Now, the results for three event windows are also found to be significant. This gives the indication that cross-border mergers tend to have a positive effect on abnormal returns with higher market concentration, while domestic events tend to have negative returns with an increased concentration index.
Table 8 shows the correlation coefficients for all mergers/acquisitions where 100 percent of shares were acquired in the transaction. Acquirer coefficients are all found positive expect for the AR (0) window. None are found to be significant. The coefficients for the rival
correlations are again positive, with the exception of the AR (0) window. Target correlations are positive and insignificant. One explanation for the positive correlation for targets here could be that most events included here were cases where two already large firms merged or a large firm acquired a relatively small one, which positively influenced investor expectations. As can be seen in section 8.1, these events showed positive target returns to the
announcement. An increase in the concentration index thus tends to coincide with lower abnormal returns for target firms (except for the cases when the results are weighted and when only M&A’s where 100 percent of shares were acquired are included) but higher abnormal returns for rival firms. Concentration does not seem to affect acquirer abnormal returns, considering the very high p-values, and thus insignificance of the results, displayed. It should be noted that acquirer correlation coefficients change drastically depending on the events included.
The results regarding the acquiring firms, in accordance with the varying results found in section 5.2. regarding the relationship between investor responses and concentration, are insignificant in all cases and inconclusive. They differ with every change made to the included events.
This inconclusiveness could be explained partly by the low, insignificant abnormal return estimates for acquiring firms. One could argue that the low abnormal returns in response to M&A activity in this study as well as others, which encompass a wide array of
industries and thus concentration indices, already indicate that the firms are not really affected by any circumstance, therefore understandably leading to these results for the correlations.
Another factor to keep in mind is the size of acquiring firms. Most firms in this study on the acquiring side are quite large (i.e. Heineken, Carlsberg, AB-Inbev etc.). As these firms are generally already at the top of their market (assessed per merger/acquisition) in terms of market share they are probably not as affected by the distribution of sizes of the other firms and thus the concentration in their particular market. Merging with or acquiring a small firm will probably not affect them in a significant way. They also simply may have less to fear from their competitors and how competitive the market is relative to the (smaller) targets and rivals. Again, this also leads back to the low abnormal returns: larger firms may be less affected by their engagement in mergers/acquisitions and thus investor responses would be negligible.
The inconclusive and insignificant results do coincide with the statement brought forward by Singal (1996) that the merging firms do not benefit much from these increases in concentration, compared to rivals. Sonenshine (2011) also finds an insignificant relationship between concentration and stock returns for acquirers.
The results regarding target firms follow what one would expect when looking at the material covered in section 5.2.2. The mostly negative relationship between the abnormal returns and the market concentration results from the theory that firms in concentrated industries tend to face lower risk as they are less inclined to innovate and have higher barriers to entry and are thus not as likely to face new competitors compared to firms in competitive industries. The positive results for the correlations between rival firms and concentration found in all instances regardless of included events do not follow the results as outlined in section 5.2.2. As mentioned previously, the theory states that a negative relationship between the returns and the HHI comes from the idea that firms in concentrated industries tend to face lower risk as they are less inclined to innovate and have higher barriers to entry and are thus not as likely to face new competitors compared to firms in competitive industries. However, it could be the case that the rival firms are small compared to the acquiring and/or target firms and/or wish to increase their market share. They could then be explicitly motivated to engage in innovation to achieve this goal. This would then lead to the firm facing more risk.
Innovation is also mentioned in this respect by Syrneonidis (1996), who argues firms in more concentrated industries may have more incentive to innovate due to their greater
market power. This is because market power might make it easier for them to collect the returns from innovation activities. Important here is that patents become more valuable with greater market power. He also mentions that with market power, firms can benefit from certain mechanisms making innovation easier: learning by doing, control of the distribution channel, secrecy, and marketing investments. Looking at the correlation from this perspective, a positive result instead of a negative one would make some sense.
The positive relationship does coincide with the theory and studies addressed in section 5.2.1. Singal (1996) and Sonenshine (2011) both find that for rival firms abnormal returns are positively correlated with changes in concentration, consistent with increased market power.
Sonenshine (2011) argues the following: “If the correlation between the HHI and the abnormal return is positive and significant for the target and the rival firms during the event, this would accommodate the hypothesis that market power is generated by anti-competitive events, since as market concentration increases, the target firm and rivals will be able to price above marginal cost and increase their profitability.”. As mentioned in the previous section, 56% of the M&A’s in the CAR (-1,1) window were seen as anti-competitive by rival firms’ investors. It was also noted that the negative returns were on average greater than the positive ones. These results coincide with the correlation results found here when following the reasoning given by Sonenshine, at least for the rivals. A positive relationship between concentration and returns for rival firms can be seen as evidence of an increase in perceived value of a merger by investors as the market becomes more concentrated.
We thus find that the correlation results follow some of the material covered in section 5.2, although the role of the firm in the M&A transaction makes a difference as to which previously discussed material the results coincide with.
9. Conclusion
In this thesis, we examine the relationship between market concentration and abnormal stock market returns for companies involved in a merger or acquisition as well as their rivals. We use the event study method to answer the research question posed:
How has the degree of market concentration in the European beer market influenced the effects of merger & acquisition announcements on abnormal stock market returns?
Data is used from eighteen mergers/acquisitions, of whom at least one is based in the EU. We use this data in the market model to estimate abnormal returns: the difference between the actual change in share price that resulted from an event and the counterfactual, i.e. what would have happened if the event had not occurred.
Assessing the literature on previously conducted event studies, the expectation is that target and rival abnormal returns are generally found to be positive and significant, whereas abnormal returns to acquirers range from negative to slightly positive. Looking at the literature reviewed on the relationship between market concentration and stock returns, we conclude that market concentration has an ambiguous effect on stock returns.
We find that our estimates for the abnormal returns largely follow the established theory. The rivals show slightly positive returns, albeit not statistically significant. The positivity can be seen as a sign of anti-competitiveness of the merger. Taking the lack of significance into account one could however conclude that investor behaviour may have been influenced by other factors. Targets firms show almost only positive results, with most being significant. This coincides with the literature. Acquiring firms also show mostly positive results. They are however very small with a large number not found to be significant. These results all mostly align with the expectations formed when analysing previously conducted studies in other geographical areas and industries.
The concentration in the relevant market for each merger is based on the
Herfindahl-Hirschman index (HHI) using total net sales. The relationship between market concentration and abnormal stock market returns (and thus investor responses) is assessed by using pairwise correlations.
The correlations found for target firms mostly coincide with the finance-based studies: they are negative and some significant. Using weights according to firm size or only including events where the percentage of shares acquired was 100 percent turns the correlations positive but none of these results are significant.
Acquiring firms show differing correlations: every change from included events or weights gives different results. However, no result in any of the cases assessed was found to be significant. Factors to be taken into account regarding these results are the size of the acquirers and the already close-to-zero abnormal returns. These findings follow the ambiguous results on the relationship in previous theoretical work and studies.
Rival firms follow the industrial organization-based theoretical framework and studies. Possible explanations for the negative correlations found are the incentives that might be present for the rivals to innovate despite being in a concentrated industry, and the idea that the anti-competitiveness of the events creates market power for the rivals.
We find that the relationship between market concentration and (abnormal) stock market returns is a very complicated one. In the industry we have assessed here, we find that the relationship differs with respect to the role of the firms: acquirers, targets and rivals. One can thus not simply state that there is a specific positive or negative connection between the two. We can, however, see that for target and rival firms, there indeed is a relationship.
Further research is needed to assess the results found in this thesis. As the sample size is quite small here, an expansion is needed to check the findings on a larger scale. Due to the small sample size, there might be some bias towards the more extreme estimations for the abnormal returns or certain characteristics specific to firms included. Possibilities to expand on the research conducted here are for example to still focus on the beer industry but include not only European but worldwide M&A’s or to focus on a different merger prone industry with more available data. An increase in the sample size would also give the possibility to use regression analysis on the relationship between abnormal returns and the concentration index instead of correlations, which could give more definitive results and might give the possibility to look at causality instead of only correlation.
