1
Stock Market Driven Acquisitions: Does theory hold for
cross-border mergers & acquisitions?
Maurice Huisman 10005072 Supervisor: Dr. F. Peters MSc Business Economics Master Thesis Second Semester 2013-2014
2
Abstract
This study elaborates on 398 completed all cash- and equity-financed deals between 1985 and 2012 over the countries Canada, France, Germany, the United Kingdom, and the United States. It tests the SMDA theory in a cross-border M&A context, thereby focusing on the return of the acquirer’s country equity index and exchange rate, as previously used in the study of Erel et al. (2012). The statistical approach enables this study to assess differences between both types of payment methods. Overall, this study finds limited support for the SMDA theory, indicating no significance for the main variables of interest. Furthermore, subsample analysis of the dot.com bubble is consistent with the previous mentioned result. After performing additional robustness checks, results indicate again limited support. This could indicate that other factors could drive the choice of cash or equity as a payment method for cross-border M&A activity.
3 Table of contents
1. Introduction 4
2. Literature Review 6
2.1 Stock Market Driven Acquisitions Theory 6
2.2 Alternative Theories 9
2.3 Cross-Border M&A Activity 10
3. Hypotheses and Methodology 12
3.1 Hypotheses 12
3.2 Methodology 14
4. Data and Descriptive Statistics 17
4.1 Data 17
4.2 M&A Activity 1985-2012 19
4.3 Descriptive Statistics 20
5. Results 28
5.1 Takeover Premiums 28
5.2 Cumulative Abnormal Returns 30
5.3 Buy-and-Hold Abnormal Returns 32
5.4 M&A Activity during the Dot.com Bubble 33
6. Robustness Checks 38
6.1 Descriptive Statistics 38
6.2 Takeover Premiums 39
6.3 Cumulative Abnormal Returns 41
6.4 Buy-and-Hold Abnormal Returns 42
7. Conclusion 44
Reference list 47
Appendix 1 50
4 1. Introduction
In the vast amount of literature on mergers and acquisitions, the focus is primarily on domestic mergers & acquisitions (hereafter referred to as M&A). However, cross-border M&A activity increased from 23% of total volume in 1998 to 45% of total volume in 2007 (Erel et al., 2012), indicating the importance of this field of research. Companies usually engage in cross-border M&As for the same reasons as domestic ones. For example, to increase internal growth and profitability or other management reasons. Though going across country borders creates additional frictions in the acquisition process, due to regulation (e.g. accounting standards) or cultural differences (e.g. languages). Recent studies show that these frictions are of significant effect on the volume of M&A (Rossi & Volpin, 2001; Ahern et al., 2012). Besides the previous mentioned effects, on an international scale, macro-economic factors of valuation also play a role. The return of the exchange rate and equity index, and the market-to-book ratio of a country prove to have a significant effect on the probability to engage in a cross-border merger (Erel et al., 2012).
The merger waves of the 1960s and 1990s were characterized by high stock market valuations and equity as an acquisition method. Shleifer & Vishny (2003) propose the stock market driven acquisitions (SMDA) theory, as the neoclassical theory does not explain for acquisition payment methods and additional stock market evidence. This theory argues that overvalued firms issue equity as acquisition method to benefit from their valuation advantage. Prior studies connected market timing and abnormal merger activity, as Loughran and Vijh (1997) find evidence of equity acquirers underperform to cash acquirers on the long-run, suggesting that firms try to take advantage of their overvaluation. Choe et al. (1993) finds that, with respect to seasoned equity offerings, firms tend to issue equity during prosperous economic times. Loughran et al. (1994) shows that firms time their initial public offerings during high market valuations and these firms face negative returns on a long-term horizon.
Beside the empirical studies, a number of transactions seem to indicate valuation driven behavior. For example, the $203 billion equity-financed purchase of Mannesmann AG by Vodafone AirTouch PLC (announced November 1999; completed June 2000). Vodafone’s stock price was approximately £204 pence at the time of the announcement, as the share price hit all time high in September 1999. Vodafone initially offered $132 billion, which was a premium of 4.50% per share at the time. Ultimately, Vodafone paid $203 billion (increase of 54% with respect to their original offer). After the acquisition, Vodafone performed poorly, as share prices dropped to £67.57 pence (loss of almost 67%) 3 years after the acquisition. Alongside the large increase in takeover price, there could be other reasons that drove the poor performance. However, if the board knew share prices were at an all time high, this serves as a clear illustration of the SMDA theory.
5 Though there are several empirical studies that tested the SMDA theory, none of them
analyzed if the theory holds in context of cross-border M&A activity. Therefore, this study addresses the following research question:
Does the Stock Market Driven Acquisitions theory hold for cross-border M&A?
The theory predicts differences between cash and equity acquirers, and therefore, this study uses a dummy variable to assess any significant differences between the 2 acquisition methods. In addition, this thesis tests whether the macro-economic valuation factors, previously identified by Erel et al. (2012), are significant determinants of takeover premiums, cumulative abnormal returns (hereafter CARs), and buy-and-hold abnormal returns (hereafter BHARs). The results of these variables indicate whether the SMDA theory holds in cross-border M&As.
This study firstly hypothesizes predictions regarding the dependent variables. More precisely, equity acquirers should pay higher takeover premiums, as stock payers are willing to bid more aggressively to take advantage of their overvaluation. Furthermore, equity acquirers should underperform around the announcement date and in the long-run, which corresponds to prior empirical evidence. In addition, this study predicts several relationships pertaining to the independent variables. Firstly, the return of the acquirer’s country main equity index should be significantly positive for stock acquirers with respect to takeover premiums. Secondly, in the context of CARs and BHARs, this study expects equity index return significantly negative for equity acquirers. The exchange rate return should have the same effect for both types of acquirers because all acquisition methods benefit from currency appreciation. Regarding the robustness checks, this study expects the relative valuation measures to be of the same relationship as the equity index returns.
The analysis focuses on 398 completed cross-border M&A deals between 1985 and 2012 over the countries Canada, France, Germany, the United Kingdom, and the United States. Consistent with previous studies, the dummy variable for stock acquirers proves significantly negative for CARs and BHARs, supporting the hypotheses. However, this study is unable to find concrete evidence that supports the SMDA theory. Based on past returns of the equity indices, takeover premiums do not rise for stock acquirers if performance increases. Moreover, the interaction term proves positively related with long-run returns, which was unexpected. Although past returns indicate a negative relationship for equity acquirers around the announcement date. No coefficients of the equity indices past return appear to be of any significance. With respect to the exchange rate past year’s return, this study expects no difference between the types of acquirers. However, the coefficient correlates positively with takeover premiums. A possible reason could be that the factors identified by Erel et al. (2012) are no overvaluation measures, as they could measure changes in the fundamentals.
6 Consequently, difficulties arise to establish evidence that supports the SMDA theory. However, the subsample analysis of the dot.com bubble (during this period the market was aware of the overvaluation) shows limited support of the theory’s predictions.
When testing for robustness, the country difference M/B ratio is positive regarding takeover premiums for equity acquirers. Indicating that stock acquirers pay more than equal valued cash acquirers. However, Shiller’s PE10 ratio for equity acquirers proves significantly negative. Both valuation measures indicate for stock acquirers positive coefficients with respect to CARs and BHARs, though this study hypothesizes the opposite relationship.
This study proceeds as follows. Section 2 discusses the relevant literature regarding market timing, the SMDA theory and empirical evidence, alternative theories, and existing literature on cross-border M&A activity. Section 3 elaborates on the proposed hypotheses and methodology to analyze if the theory holds. Section 4 describes the data sources, and the descriptive statistics of this study. Section 5 presents empirical results of the SMDA theory with respect to takeover premiums, CARs, and BHARs. Section 6 presents additional robustness checks. Lastly, section 7 consists of concluding remarks.
2. Literature review
Although a large part of M&A activity contains cross-border mergers, academic literature primarily focuses on domestic M&A activity within the USA. Understanding domestic mergers is relevant concerning international M&A, however, it does not highlight cross-country
differences. Previous studies show that these differences are significant factors in the cross-border M&A process. In addition, some empirical evidence seems related with the Stock Market Driven Acquisitions (hereafter SMDA) theory; therefore, the first section discusses relevant literature regarding the SMDA theory. The second part consists of alternative theories
concerning M&A empirical evidence. Finally, this section reviews empirical evidence of cross-border M&A activity.
2.1 Stock Market Driven Acquisitions Theory
The idea that market timing and equity issuances are related exists for some time now, and an extensive amount of literature supports this view. Firms should want to issue equity during periods of high market valuation in order to receive the best possible price. Initial studies on samples of firms in the US and the UK show evidence of market timing with respect to seasoned equity offerings. The results show that management delays their decision until their stock is highly valued, as management bases their decision of equity issuance on historical security prices when choosing between debt and equity (Taggart, 1977; Marsh, 1982).
7 stock and the business cycle, arguing that firms sell their equity in times of lower adverse
selection costs. In periods of economic growth, firms face better investment opportunities as well as more certainty about the true value of the firms’ assets. The findings show that firms tend to issue equity during good times of the economic business cycle, consistent with the predictions. In addition, Spiess & Affleck-Graves (1995) find that, with respect to seasoned equity offerings, firms who issued equity underperform on the long-run against non-issuing firms of matching characteristics.
Relatedly, empirical evidence also indicates the existence of market timing at initial public offerings. Loughran et al. (1994) shows that firms successfully time their initial public offerings during high market valuations, and subsequently, these firms face negative returns on a long-run horizon. Furthermore, over a 3-year time window, firms who go public relatively underperform to private firms of same size and industry (Ritter, 1991). The evidence supports the idea that market timing exists at initial public offerings and seasoned equity offerings, as share prices seem to revert to their fundamental value over time (Loughran & Ritter, 1995).
Consistent with previous evidence, Baker & Wurgler (2002) show evidence of firms issuing their equity when the market valuations are relatively high. Ikenberry et al. (1995) finds that firms tend to engage in share repurchase programs when equity valuations are low. In addition, firms with share repurchase programs achieve a 12% abnormal return over a control portfolio using buy-and-hold strategies over a 4-year time window. Graham & Harvey (2001) survey 392 CFOs to identify the most important factors influencing equity emission decisions. The CFOs identify as most important the under- or overvaluation of their stock, and recent rise of their stock prices. Conversely, in another survey conducted by Graham (1999b), over two-thirds of the CFOs feel that their equity is undervalued and only 3% indicates stock
overvaluation.
Shleifer & Vishny (2003) propose the stock market driven acquisitions theory (hereafter SMDA theory). The main assumption is that financial markets are inefficient in pricing of firms, and consequently, some firms are overvalued and others are undervalued. The second
fundamental assumption, management of a firm is completely rational and knows when their firm is under- or overvalued. In addition, the market does not have this information, and therefore, management could take advantage of its overvaluation by engaging in a merger project. This process can be viewed as a form of arbitrage, where rational managers take advantage of inefficient markets using stock as a payment method. The result is that overvalued firms tend to acquire relatively less-overvalued firms with equity as an acquisition method. However, the latter assumption should be taken with caution, as it is unrealistic to assume that a firm’s management team always has superior information over the market.
The existing neoclassical theory sees mergers as an efficiency-improving response to industry shocks, and predicts that mergers increase profitability. Although this theory has some
8 explanatory power, it is not able to clarify the chosen payment method of firms. In addition,
only if multiple industries experience shocks at the same time, the neoclassical theory is able to explain aggregate merger waves (Shleifer & Vishny, 2003). The SMDA theory relates to the existing neoclassical theory and is capable to explain additional stock market evidence. Why firms choose for a specific acquisition method, why cash acquirers obtain positive long-run abnormal returns, and why stock acquirers do not perform similarly. According to the SMDA theory, acquirers choose equity as a payment method when their stock is overvalued, taking advantage of the overvaluation. As the overvaluation is only temporary, stock prices revert to their fundamental value over time, and consequently, equity acquirers have negative abnormal returns over a long-term horizon. This effect does not play a role for cash bidders (Friedman, 2006).
Although the SMDA theory predicts that equity acquisitions should have negative long-term abnormal returns, this prediction does not have to hold in all situations. For example, an acquisition made by a fair-valued acquirer for an undervalued target firm, using equity as a payment method, would expect positive returns. Since the undervalued target reverts to its fundamental value over time, it should generate positive returns on the long-run. Why would a target firm accept a bid when it is undervalued? A rational target firm would not, however, in the context of SMDA, a target firm only knows about their own undervaluation. They cannot assess whether the acquirer is making a bid because of synergies or because of overvaluation. Rhodes-Kropf & Viswanathan (2004) propose the rational model of mergers and relates to the SMDA theory of Shleifer & Vishny (2003). Again, management has private information whether their firm is under- or overvalued as well as information on the potential value when merging with another firm. The paper identifies 2 effects that could cause the mispricing, a firm-specific component and a market-wide component. Management of a firm knows when they are mispriced but cannot assess where this effect comes from, possibly a firm-specific or market-wide effect. Managers should act to maximize value of the firm, and therefore, accepting propositions that create more value than the stand-alone value of the firm. Consequently, the management team bases their decision on the information of mispricing and assessment of expected synergies. The market-wide component affects the perception of management on synergies, possibly estimating the synergies too high in times of high market valuation due to an estimation error. Hence, target firm managers could accept acquisition bids even when they are undervalued. This hypothetical situation seems plausible, however, according to the theory, only overvalued acquirers would choose to issue stock as payment method. Fair valued firms would be acting irrational when they choose equity, as acquirers know that this way of financing is more expensive. In turn, only cash acquirers could benefit from positive long-run returns of undervalued targets.
9 The theoretical propositions of valuation driven acquisitions are followed by several
empirical studies that tested whether both theories hold. Examples of overvaluation proxies are market-to-book ratio, value-to-price ratio and earnings-to-price ratio. Friedman (2006) tests the SMDA theory, estimating the influence of overvaluation on merger premiums. The findings of Friedman (2006) are robust to explanations of the neoclassical theory. Using overvaluation proxies, Friedman (2006) finds that equity acquirers pay a 4.5% higher merger premium compared to cash acquirers. Ang & Cheng (2006) find that acquiring firms are relatively overvalued compared to non-acquiring firms and targets, and in addition, overvalued acquirers prefer using stock as a payment method. Both studies’ findings correspond with the valuation driven merger theories. However, Rhodes-Kropf et al. (2005) disentangles the market-to-book ratio in 3 parts. The results indicate that equity valuation of a firm affect the probability to engage in a merger, but no significant evidence whether this is driven by misvaluation. Hence, empirical evidence regarding stock valuation driven theories is ambiguous. The above studies concentrated only on domestic M&A activity, where this study focusses on cross-border M&As and seeks to find whether the valuation driven acquisition theories hold.
2.2 Alternative Theories
Alongside empirical studies of the SMDA theory, other studies focus on a relationship between Tobin’s Q and merger activity. Andrade et al. (2001) finds that since 1973 more than two-thirds of all mergers consist of an acquirer with a higher Q than the target. Servaes (1991) perceives the Q ratio as an indication of management performance, and shows that target, acquirer and total takeover returns are higher when the target firm has a low Q and the acquirer a high Q. An alternative theory that corresponds to M&A empirical evidence is the Q-theory of mergers by Jovanovic & Rousseau (2002). The Q-theory treats M&A activity as used-capital-market transactions and predicts that a firm’s investment rate should rise with its Q, where a firm’s Q is a function of capital stock. Results show that firms’ M&A investments respond more to its Q than for direct investments. Firms with high market valuations could expect more future growth, and therefore, could wish to expand by engaging in a M&A project. A drawback of the Q-theory is that it predicts the same relationship for cash and stock deals even though empirical evidence proves otherwise (Shleifer & Vishny, 2003).
Harford (2005) tests whether market timing or industry-level shocks cause merger waves. Using a sample of industry-level merger waves in the 1980s and 1990s, this study supports the neoclassical view on merger waves. The results indicate that mergers occur due to industry shocks. An industry shock could initiate a merger wave if there is an adequate level of capital liquidity in the market to facilitate asset reallocations. When testing the cause of an industry merger wave, Harford (2005) finds the industry market-to-book ratio to lose its significance after adding economic variables. However, the coefficient of the industry-level shock proves
10 significant when there is sufficient capital liquidity in the market. In addition, results show that industry-level merger waves cause aggregate merger waves. According to Harford (2005), previous studies assume a relationship between high stock market valuations and merger waves, while the actual cause of merger waves is capital liquidity. Because high levels of capital liquidity, induces lower transaction costs, and combined with an industry shock, this could produce numerous transactions. In addition, Harford (2005) shows that equity acquirers have high probabilities of also being a partial-firm acquirer using cash as acquisition method. These findings contradict predictions of the SMDA theory by Shleifer & Vishny (2003). Hence, existing literature thus far does not fully support the SMDA theory.
2.3 Cross-Border Mergers and Acquisitions
As mentioned previously, the amount of cross-border M&As is a significant percentage of total M&A volume. The industry grew from 23% in 1998 to 45% of total volume in 2007 (Erel et al, 2012). The rationale behind cross-border acquisitions is the same as for domestic ones. Most frequently used explanations are synergies or economies of scale. M&A activity between 1992 and 2000, also known as the fifth merger wave, is characterized by cross-border M&As. In addition, comparing to the fourth merger wave, Andrade et al. (2001) shows a 50% increase of stock as acquisition method.
Although cross-border mergers occur for the same reasons as domestic mergers, in an international setting one would expect additional frictions in the M&A process. For example, firms face cultural (e.g. language) and regulation differences (e.g. shareholder protection). Prior studies show that these additional frictions have significant effects on the probability or volume of cross-border M&As. Cultural differences are important in the context of cross-border mergers, because the process of merging firms involves teamwork of both parties. Extensive psychological research shows that people demonstrate intergroup biases. Hewstone et al. (2002) argues that a firm’s employee is more likely to prefer working with other employees that share the same cultural values, even if this would result in efficiency losses. In addition, Akerlof (1997) argues that social distance is negatively correlated with respect to communication within a firm, and consequently, this affects a firms’ ability to make important decisions. Hoffman et al. (1999) and Glaeser et al. (2002) find empirical evidence that supports this theoretical argument. Ahern et al. (2012) finds significant negative correlation between cultural distance and cross-border M&A volume. Large cultural distance between 2 countries enhances
difficulties in coordination of both parties’ employees, and this increases the costs of merging the firms.
The law and legislation of a target firm’s country are also important in the cross-border M&A process. Empirical findings show that the level of shareholder protection is of significant effect. According to Rossi & Volpin (2004), if minority shareholders have fewer rights, the
11 probability increases that they will be subject to expropriation, and consequently, this leads to less developed capital markets. Firms that are located in less developed countries face more difficulties when raising external funds for company takeovers. La Porta et al. (2002) shows that higher shareholder protection is correlated with higher firm valuation. When creditors and shareholders are better protected, the investors’ willingness to pay for financial assets rises.
Alongside shareholder protection, improvement of accounting standards enhances transparency of corporate accounts. More specifically, this reduces the extent to which an entrepreneur can expropriate shareholders. Both previously discussed factors, level of shareholder protection and accounting standards, have a significant effect on the volume of M&A activity across countries (Rossi & Volpin, 2004). In addition, Ferreira et al. (2010) finds that weaker legal institutions of a country increase the likelihood of being a target firm and more foreign institutional ownership within a company. In cross-border mergers, target firms usually adopt the corporate governance structures (e.g. accounting standards) of the foreign acquirer, thereby giving them the opportunity to change their level of shareholder protection. The result: increase of the industry’s Tobin’s Q (Bris & Cabolis, 2008). Chari et al. (2009) finds that for cross-border M&As multinational acquirers tend to be located in developed-markets and targets in emerging-markets. Post-acquisition stock performance is significantly higher for developed-market acquirers than those of emerging-developed-market acquirers. These findings indicate that
developed-market acquirers are better able in overcoming problems as incomplete contracting, and consequently, enhance firm value.
More related to the aim of this study, fluctuations in the equity index level or exchange rate affect the decision to engage in cross-border M&A projects. A foreign firm can be relatively inexpensive after depreciation of its currency, and therefore arise as a potential target. Alternatively, related to the SMDA theory, Erel et al. (2012) argues that positive NPV investments can arise for acquiring firms when they are overvalued and/or undervaluation of target firms. As in the SMDA theory, where mispricing takes place in financial markets, Baker et al. (2009) argues that cross-market mispricing could occur because of irrational expectations of investors.
The previous mentioned arguments focus on temporary fluctuations in valuation, however, permanent valuation differences can also affect the likelihood of merger decisions. A firm that produces goods for cross-continental companies can become a potential target after permanent exchange rate devaluation. Kindleberger (1969) argues that expected earnings rise or cost of capital decreases after a cross-border acquisition. According to Froot and Stein (1991), after permanent valuation changes, foreign firms’ cost of capital relatively declines compared to domestic firms. This is mainly due to less informational problems, because lower cost of capital allows foreign firms to bid more aggressively compared to domestic firms, and thereby affecting the likelihood of engaging in a merger project.
12 Prior empirical evidence shows that currency movements and stock market performance affects the propensity of firms in 2 different countries to merge. Indicating a valuation effect, where high valued firms tend to buy firms of low valuation. With respect to a short-term horizon, acquiring firms tend to be in countries where currency has appreciated and targets are located in countries where currency has depreciated. In addition, Erel et al. (2012) finds that acquiring firms tend to be located in countries where the country stock market performance significantly outperformed those of the target firms. Other studies support the above results with respect to foreign direct investments; currency movements and temporary upward stock market valuations significantly affect foreign direct investment cash flows (Baker et al., 2009; Klein et al., 2002).
3. Hypotheses and Methodology
This section consists of 2 parts. The first part elaborates on the constructed hypotheses and clarifies the expected differences between cash and equity acquirers. The second part discusses the proposed methodology to test the hypotheses.
3.1 Hypotheses
This study tests if the stock market driven acquisition theory holds for cross-border M&As, and therefore, analyses the effect of the past 12-months return of the acquirers’ country stock market and the past 12-months return of the acquirers’ exchange rate. The previous mentioned variables proved significantly in the study of Erel et al. (2012) regarding the likelihood of cross-border M&A decisions.
With respect to takeover premiums, this study expects:
H1: If the past 12-months return of the acquirers’ country equity index is positive, cross-border
takeover premiums tend to increase more for equity acquirers than for cash acquirers.
As stated previously, the SMDA theory assumes that financial markets are inefficient in pricing of companies, and consequently, some firms are overvalued and other firms are undervalued. In addition, the theory assumes that managers have private information (the market does not have this information) if their company is mispriced. Management also knows that current overvaluation reverts to its fundamental value over time. Hence, to take advantage of the overvaluation, managers are willing to pay more for a target firm to obtain the benefit. Higher equity index valuations increase the likelihood of firms being overvalued or drives up the overvaluation of firms even further. The above stated expectation is not of any effect
13 regarding cash as an acquisition method and therefore takeover premiums of equity acquirers should exceed those paid by cash acquirers, respectively.
The second hypothesis of this study is the following:
H2: If the past 12-months return of the acquirers’ country exchange rate is positive, takeover
premiums tends to increase for both cash and equity deals.
If the acquiring firm’s currency appreciates against the target firm’s currency, a foreign target firm becomes relatively cheaper against comparable domestic firms. This allows the acquiring firm to bid more aggressively compared to bidders of the target firm’s country, and consequently, paying a higher takeover premium in order to close the deal. As stock prices are denominated in local currency, takeover premiums should increase for both payment methods if the acquirer’s currency appreciates.
This study’s third hypothesis is the following:
H3: Equity acquirers tend to underperform compared to cash acquirers around the
announcement date.
Management of a firm has private possession of the information if their firm is overvalued. Hence, when a firm issues equity as payment method, the market is unable to assess if the acquirer is overvalued. Therefore, this expectation is consistent with prior evidence and does not focus on predictions of the SMDA theory. An equity acquisition announcement is interpreted as a combination of a merger announcement and an equity issuance. Andrade et al. (2001) finds that the financial markets tend to respond negatively to equity issues.
The fourth and final hypothesis of this study is:
H4: After the merger is completed, long-run performance of equity acquirers should
underperform to the performance of cash acquirers.
According to the SMDA theory, mispricing is temporary and reverts to its fundamental value over time. Management of an acquiring firm will only issue equity in case of overvaluation. In case of fair value or undervaluation, cash is used as an acquisition method, since there is no potential benefit of issuing equity. This study expects that equity acquirers will underperform cash acquirers on the long-run, as prior empirical evidence already proved that cash acquirers outperform equity acquirers on a long-term horizon (Loughran & Vijh, 1997).
14 3.2 Methodology
This study tests whether the SMDA theory holds for cross-border M&A activity and therefore tries to measure significant differences between the 2 types of acquirers. These tests are implemented for the dependent variables as well as important independent variables. In this manner, this study is able to assess whether initial results show support towards the hypotheses. For differences in mean, a two-sample t-test is executed with unequal sample size and variances. In addition, the Wilcoxon rank-sum test is performed to test for significant differences in the population of both groups.
To test the first 2 hypotheses, takeover premiums are measured at 3 time intervals at the announcement date. Differences between cash- and equity-financed deals are than analyzed. Andrade et al. (2001) argues that takeover premiums should be estimated over a short time window, to obtain reliable statistical evidence. To establish the effect of the independent variables of interest on takeover premiums, OLS is used.
The first regression this study estimates is the following:
1
= + 12! + " 12! ∗ $ % + &'( ) 12! + *'( ) 12! ∗ $ % + + $ + + , + +-
For the dependent variable, the superscript T represents target firms, subscript i represents target firm i, in country j, at time t. The takeover premium is constructed as the difference between the purchase price offered by the acquirer and the pre-acquisition stock price of the target firm, divided the pre-acquisition stock price of the target firm. The pre-acquisitions stock price is observed at [-1,0], [-5,0], and [-20,0] days at the announcement date. This construction is consistent with Andrade et al. (2001).
The first independent variable, 12! , represents the past 12-months return of the country stock market of the acquiring firm’s country, for acquiring firm x, in country y, at time
t. The variable $ % represents a dummy which indicates 1 if the deal is equity-financed, and 0
otherwise. The second coefficient represents an interaction term between the acquiring firm’s country stock market return and the dummy variable. By adding the interaction term, this study is able to see whether there are significant differences between cash and equity bidders. In addition to the acquirer’s country stock market level, this study uses the exchange rate return ('( ) 12! ) between the target’s country (j) and the acquirer’s country (y) at time t. Moreover, making use of macro-economic variables induces less exposure to endogeneity. To establish a causal relationship for the variables of interest, control variables ( for targets and , for acquirers) are added on a firm-specific level. Where subscript i represents the target firm
15 and x the acquiring firm at time t. Only firm size is used as a control variable with respect to the target firms. For the acquirers the following control variables are used: firm size, book leverage and the cash-to-assets ratio. Firm size is measured as the logarithm of total assets, book leverage is measured as total debt divided by total assets, and cash-to-assets ratio as the current value of marketable securities and cash divided by total assets. The control variables are computed consistently with the study of Friedman (2006).
Finally, country fixed effects ( ) are included by using dummy variables for differences between countries that do not vary over time, j represents the target’s country and y represents the acquirer’s country. In this manner, this study controls for differences in culture or language, thereby establishing a better causal relationship between the dependent variables and the returns of the exchange rate and equity index. If there is interaction between countries, for example between Canada and France, the dummy variable indicates one and zero otherwise. Such a dummy is created for all possible interactions between countries, except between the United Kingdom and the United States. To make sure no dummy trap arises.
If equity index returns are positive, it is relatively easier to obtain external financing and therefore both types of acquirers benefit. This study expects a positive relationship between takeover premiums and equity index returns. However, to be consistent with the SMDA theory, this study predicts that the coefficient of the acquirer’s equity index return for equity bidders should be more positive than the coefficient of cash bidders. As equity index valuations increase, the likelihood increases of firms being overvalued or becoming more overvalued. Regarding the exchange rate returns, this study expects the same relationship for cash and equity acquirers. Since both acquisitions methods benefit from a currency appreciation.
As for the control variables, target firm size is expected to have a negative coefficient. Since the dependent variable is relative, the larger a target firm is the higher the premium must be in absolute value as compared to smaller deals. For the acquirer firm size, this study expects a positive relationship. Larger firms are probably more mature and have easier access to external financing compared to younger firms, and therefore can afford to pay higher takeover premiums. Furthermore, if a firm has much cash at hand, they could bid more aggressively since they are not reliant on external funding, and consequently leading to a positive relationship between takeover premiums and the cash-to-assets ratio. Concerning book leverage, this study expects a negative relationship. When a firm is highly leveraged, bidding more than necessary means a waste of money which could have been spend on paying off loans. This would mean that firms take the risk of not meeting future obligations.
16 Continuing with the third hypothesis, the following regression is estimated:
2 , ! = + 12! +
" 12! ∗ $ % + &'( ) 12! + *'( ) 12! ∗ $ % + + $ + + , + +-
This regression is estimated with cumulative abnormal return (CAR) of acquiring firms as a dependent variable, measured with a time window [-1,1] at the announcement date. To establish abnormal returns a market-adjusted return model is used, where the benchmark is the return of the acquirer’s country main index. This regression analyses whether there are
differences in abnormal returns by the choice of payment method. The same variables are used as in the first regression.
If the stock exchange has performed well over the past 12-months of the acquiring firm, then CARs are expected to be higher as the market sentiment is better. However, equity acquirers are expected to underperform to cash acquirers even if the financial market has performed well. The use of stock as acquisition method could be interpreted as a signal of overvaluation after good performance of the equity index; therefore, this study predicts that the overall effect of equity index return is lower for stock bidders regarding CARs. A positive relationship is predicted regarding the exchange rate return of the acquirer. If performance was good over the past year, investors will recognize that the firm takes advantage of the currency appreciation. Moreover, there are no significant differences expected between cash and equity bidders with respect to the exchange rate return.
As for the control variables, target firm size is expected to have a negative relationship with the CARs. Acquiring large firm means spending a lot of money with the risk of the acquisition being successful. Acquirer firm size should be positively correlated; a larger and more mature firm could easier handle the risk induced by the acquisition. Regarding the cash-to-assets ratio, this study expects a positive effect, as the firm has a buffer against any unpredicted downfalls. When the acquiring firm has a high level of debt, investors are predicted to respond negatively on the announcement date, because the acquisition could cause more risk.
To test the fourth hypotheses, this study estimates the following regression:
3 /), ! = + 12! + " 12! ∗ $ % + &'( ) 12! + *'( ) 12! ∗ $ % + + $ + + , + +-
The dependent variable represents buy-and-hold abnormal returns of the acquiring firm. Measuring the abnormal return with a time window of [0,36] months at the completion date. The abnormal return is measured as the return of the acquiring company minus benchmark
17 returns. As a benchmark this study uses the returns of the acquiring firm’s country main
exchange. In addition, for firms located in Canada or the United states, portfolios of the Fama-French website are used. For France, Germany, and the UK, the Fama-Fama-French portfolios are not used as a benchmark. This because of cultural differences compared to the US, since these portfolios consist of American companies. Canadian firms are used because of cultural similarities with the US, and most stock listed firms in Canada have business operations in the US. Finally, the same independent and control variables are used as in the previous regressions.
As stated in the previous section, equity acquirers are expected to underperform compared to cash acquirers. Therefore, the coefficient of the dummy for stock acquirers is expected to be negative. This prediction is consistent with empirical evidence of Andrade et al. (2001), and Loughran and Vijh (1997). Furthermore, the SMDA theory predicts that overvalued stock reverts to its fundamental value over time. Consequently, this study predicts a negative relationship between the BHARs and the interaction term of equity index returns and the dummy variable. The exchange rate returns are expected to have the same relationship with BHARs for cash and equity bidders.
With respect to the control variables, target firm size is expected to be positively related. Large target firms probably have proven themselves in the past and therefore are less risky. The same reasoning holds for the acquirer firm size. Furthermore, large acquirers could purchase target firms to gain market share or new products. When a firm has much cash at hand, it could protect itself against unforeseen events in the future, therefore expecting a positive relationship for the cash-to-assets ratio and BHARs. Relatedly, high levels of book leverage are expected to have a negative impact on long-run returns, as firms bear the risk of not being able to meet future interest payments.
4. Data and Descriptive Statistics
This section discusses the process of data selection and the performed selection criteria. Furthermore, the M&A activity of this sample and the descriptive statistics are discussed. The first part starts by elaborating on the data process and how the representative data sample is obtained. The second part gives an overview of the number of M&As per country over the years 1985-2012. The last section discusses the descriptive statistics.
4.1 Data
In order to test the hypotheses, this study collects data of all variables described in the previous section. The database Thomson One is used to obtain data on completed cross-border mergers and acquisitions for the countries Canada, France, Germany, the United Kingdom, and the United States. Data is acquired for the years between 1985 until 2012, as data of Thomson One
18 only goes back to the year of 1985. All selected companies are publicly listed since stock prices are needed for the dependent variables. Moreover, only deals that are completed by all cash- or equity-financing are selected, this in order to elaborate on the differences between the types of acquirers. As for the acquisition techniques, only mergers, tender offers and exchange offers are chosen. Any other acquisition methods, for example a leveraged buyout, are excluded to obtain a representative sample. In addition, data is also collected for transaction value, price paid per share by the acquiring firm, target firms’ total equity and assets values, and acquiring firms’ total cash, debt, equity and assets values. Furthermore, this study removes M&As where the market capitalization of the target firm does not exceed $1 million, consistent with the study of Friedman (2006). Also, observations are removed with CARs higher than 100%, which is unrealistic in a 3-day time window. This criterion removes 1 observation that consists of a CAR of 462.42%.
Alongside the database Thomson One, Datastream is used to acquire missing target and acquirer stock prices, and for missing balance sheet items. In addition, Datastream is also used to collect data on exchange rates between the US dollar, Canadian dollar, Pound Sterling (UK), Deutsche Mark(till 2002), French Franc(till 2002) and the Euro(from 2002 onwards for France and Germany). Country stock market levels of Canada, France, Germany, the UK, and the US are also obtained through Datastream.
The main indices used as benchmark to calculate abnormal returns are the German Dax 30, the French CAC 40, the English FTSE 100, the Canadian S&P/TSX 60, and the American S&P 500. Furthermore, BHARs are also calculated using Fama-French portfolios as a benchmark. This benchmark is only used for long-run returns of companies located in Canada or the United States. Since these portfolios consist of only American companies, Canada is also chosen because of the similarities in culture. The Fama-French benchmark consists of value- and equal-weighted portfolios based on the size of market equity. To measure abnormal returns, this study divides the acquiring firms in 5 groups based on size of market equity to have the appropriate benchmark.
After successively applying the previous methodology and criteria, this study’s sample consists of 398 completed cross-border M&A deals. Statistical evidence for the first 3
regressions is elaborated on the 398 observations. However, the number of observations for the BHAR regressions is somewhat lower. Since the long-run returns for this study´s time window were only available till the end of 2010. In addition, some firms liquidated or filed for
bankruptcy before the end of the 36 months window and therefore these observations were unavailable. Using the acquiring firm’s country main exchange as benchmark, this study remains with 310 BHAR observations. As for the Fama-French benchmarks this study has 189 observations, because of the fact that only observations of Canada and the US are used. For France, Germany and the UK no such portfolio was available.
19 4.2 M&A Activity 1985-2012
Table 1. M&A Activity 1985-2012
This table gives an overview for all border M&A activity over 1985-2012. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States. For each country the number of target and acquiring firms per year is shown. Furthermore, the total amount of M&As per year is also shown.
Year
M&A Activity 1985-2012
N
Targets Acquirers
Canada France Germany UK USA Canada France Germany UK USA
1985 1 1 0 0 0 0 0 0 0 0 1 1986 2 0 0 0 1 1 0 0 0 1 1 1987 7 0 0 0 1 6 0 0 0 6 1 1988 8 1 0 0 0 7 1 0 0 6 1 1989 10 0 0 0 4 6 0 0 0 6 4 1990 4 0 0 0 2 2 1 1 0 1 1 1991 4 1 0 0 2 1 0 1 1 0 2 1992 2 1 0 0 1 0 0 0 0 0 2 1993 1 1 0 0 0 0 0 0 0 0 1 1994 10 2 0 0 3 5 2 0 1 2 5 1995 10 4 0 0 2 4 3 0 0 1 6 1996 11 5 0 0 2 4 5 0 0 0 6 1997 20 5 2 0 11 2 0 2 3 1 14 1998 38 9 1 0 16 12 6 3 1 4 24 1999 35 8 1 1 7 18 4 5 5 7 14 2000 41 12 0 1 12 16 3 7 4 11 16 2001 23 11 1 0 3 8 2 2 1 5 13 2002 14 3 0 0 5 6 4 0 1 1 8 2003 20 7 0 2 5 6 6 1 0 4 9 2004 15 7 0 0 4 4 3 0 1 1 10 2005 14 3 1 0 6 4 1 3 2 1 7 2006 18 11 0 0 2 5 1 4 2 2 9 2007 24 7 0 1 5 11 6 3 3 6 6 2008 15 5 0 0 6 4 3 2 1 2 7 2009 10 3 0 0 3 4 3 1 0 1 5 2010 20 10 1 0 3 6 8 0 2 2 8 2011 10 5 0 0 1 4 0 1 0 4 5 2012 11 2 0 0 3 6 4 0 1 2 4 Total 398 124 7 5 110 152 66 36 29 77 190
An overview of all M&A activity of this study’s data sample is shown in table 1. After applying the restrictions discussed in the previous paragraph, this study successively remains with 398 M&A observations. Most of the target firms are located in Canada, the UK, and the US. Only 3% (12 out of 398) of all target firms are located in France and Germany. A possible reason could be language differences, since most acquiring companies (e.g. 84%) are from Canada, the UK, and the US. People speak the same language in these countries and this imposes fewer
20 frictions in the merger’s organizational process. The United States has the highest number of target and acquiring firms, while Germany has the lowest number of target firms as well as the lowest acquiring firms. Moreover, in figures 3 and 4 of appendix are shown the number of target and acquiring firms per year. These graphs show that the US almost has the highest number of target and acquiring firms per year.
Furthermore, as visible in table 1, the total number of observations per year varies quite a lot. For example, in 1985 and 1993 this study only contains 1 acquisition which is the lowest amount observed per year. While in 2000 the highest number of cross-border M&As is measured, namely 41 acquisitions. These findings are consistent with results of Erel et al. (2012). The year 2000 was the final year of the 5th merger wave, which was characterized with cross-border M&As. Figure 1 of appendix 2 shows an overview of the fluctuation in the number of M&As per year over 1985-2012.
Figure 1 also shows that during the fifth merger wave the number of cross-border M&As increased per year, where after the dot.com crash the number of cross-border M&As dropped to a significant lower amount per year. The number slightly picked up again towards the global financial crisis, after which the amount of acquisitions dropped again. It is only logical that we observe more activity during prosperous economic times. During these periods it is more easy to gain access to external funding (e.g. debt) and firms would probably also obtain a better price for their shares if they would issue equity. During harsh economic times firms face problems with access to external financing. Since share prices are lower during these periods, firms would have to issue more shares to obtain the necessary funds for the acquisition. This, in turn, dilutes shareholder ownership within a company. In addition, external funds become more difficult to obtain, as most investors are capital restrained during these periods.
4.3 Descriptive Statistics
This section discusses the descriptive statistics of all variables. The first part of this section discusses all independent variables, where the parts that follow compare the differences in dependent variables of interest between cash and equity acquirers. The last section discusses the descriptive statistics of the variables for the robustness checks. As previously mentioned in the methodology section, tests for differences in mean and median are performed to test for significant differences between the 2 types of acquirers. For differences in the mean, a two-sample t-test is implemented. In addition, the Wilcoxon rank-sum test measures any significant differences in the values of the 2 subsamples.
21 4.3.1 Independent Variables
Table 2A. Descriptive Statistics
This table shows descriptive statistics for this study’s sample. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. For definition of all variables refer to the appendix. The table provides an overview for the number of observations, mean, standard deviation, median, minimum and maximum for each variable of interest. The variables Transaction Value, Target Total Assets, Target Equity Value, Acquirer Total Assets, and Acquirer Equity value are denominated in dollar millions($mil). The variables Acquirer Cash-to-Assets ratio and Acquirer Book Leverage are denominated in percentages.
Descriptive Statistics for Cross-Border M&A Activity 1985-2012
(1) (2) (3) (4) (5) (6)
N Mean Standard Deviation Median Minimum Maximum
Transaction Value 398 1,253.78 4,050.27 179.66 1.05 48,174.08
Target Total Assets 398 1,540.37 8,069.32 98.15 0.01 107,793.60
Target Total Equity Value 398 1,293.84 4,094.71 187.64 1.05 48,714.31
Acquirer Total Assets 398 13,004.70 42,249.81 1,571.70 1.10 626,933.00
Acquirer Total Equity Value 398 4,030.89 9,301.06 674.84 0.30 78,272.00
Acquirer Cash-to-Assets Ratio 398 17.47 19.37 10.67 0.00 97.49
Acquirer Book Leverage 398 17.66 15.78 15.22 0.00 85.05
As shown in table 2A, the average transaction value of this study’s sample is roughly $1,248 million. However, the standard deviation is quite large. This is due to some high value transactions, as visible by the maximum transaction value. The median shows a value of $180 million, which is a more representative number of the sample. Since 80% (317 out of 398) of all cross-border deals in this sample are below $1 billion. Furthermore, 60% (241 out of 398) of these transactions were financed by cash. The remaining 40% of all transactions was financed with equity. This finding is almost equal to the ratio of cash and equity deals in the domestic M&A sample of Rhodes-Kropf et al. (2004), where 44% is equity deals and the remaining cash deals. As visible in figure 2 of the appendix, the number of equity deals increased during the fifth merger wave, which is consistent with previously documented empirical evidence.
If one would compare the firm size of the target and acquiring firms, it is easily noticed that the acquirers are substantially larger. The mean of the acquiring firms’ total assets is more than 8 times as large as the mean of the target firms’ total assets. Although the standard deviation of the acquirers is almost 5 times as high, the difference between the median of both variables is substantial. Furthermore, as shown in figure 5 of the appendix , the average of the acquiring firms total assets exceed the average of the target firms total assets every year. Comparing with the studies of Friedman (2006) and Rhodes-Kropf et al. (2004), they find similar results
22 regarding US samples. This indicates that acquirers identify potential targets domestically as well as cross-borders of the same relative size.
The acquirers’ cash-to-assets ratio, used as a control variable, show that firms keep a relatively low level of cash compared to their total assets value. The average is around 17% and the median is almost 11%. However, in absolute value, the total value of cash could be quite large. Acquirer book leverage shows the approximate same value, indicating that firms primarily consist of equity before engaging in an acquisition.
Table 2B. Descriptive Statistics
This table shows descriptive statistics for this study’s sample. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. CSML R12 represents the acquirer’s country equity index return over the past 12 months. EXCH R12 represents the acquirer’s country exchange rate return over the past 12 months against the target’s currency. The table provides an overview for the number of observations, mean, standard deviation, median, minimum and maximum for each variable of interest. All variables in this table are denominated in percentage terms. The symbols ***, **, * denote statistical significance for the two-sample t-test (mean) and Wilcoxon rank-sum test (median) at the 1%, 5%, and 10% level, respectively.
Descriptive Statistics for Cross-Border M&A Activity 1985-2012
(1) (2) (3) (4) (5) (6)
N Mean Standard Deviation Median Minimum Maximum
CSML R12 398 11.68 19.00 12.79 -44.57 57.53 CSML R12 241 12.00 18.79 14.20 -44.57 57.53 Cash Acquirers CSML R12 157 11.19 19.36 11.33 -41.84 50.33 Equity Acquirers EXCH R12 398 -0.51 8.20 -0.41 -27.77 39.30 EXCH R12 241 -0.19 8.11 0.00 -19.94 39.30 Cash Acquirers EXCH R12 157 -1.00 8.35 -0.58 -27.77 32.92 Equity Acquirers
Focusing on the variables of interest, table 2B shows that the mean of the past 12-months equity index return for stock acquirers is lower than for cash-financed deals. Although the difference is small, based on the hypotheses of the SMDA theory, this study expects that the returns of equity acquirers should be substantially higher than of their counterparty. Results indicate that
acquisitions decisions are made after relative good performance of the financial market. Since long-run nominal returns of financial markets vary 6%-8.50% per year (Jorion and Goetzmann, 1999). The two-sample t-test and the Wilcoxon rank-sum test show no significant differences between the 2 groups with scores of -0.41 and -0.38, respectively.
Regarding the exchange rate return over the prior 12-months between the acquiring firm’s country and the target firm’s country, the mean for both groups is slightly negative. If one
23 would look at the minimum value measured of this variable, it shows a negative number of -27.77%. This indicates that an acquiring firm still has an incentive to acquire another firm, even after a substantial drop in currency value. Hence, the target firm becomes more expensive. Possibly, it could be that the acquiring firm wishes to take over the target firm before the currency would lose even more value. Logically, acquiring firms should be more likely to acquire targets across borders when these firms become relatively cheaper after currency appreciation. This study finds no significant difference with respect to the two-sample t-test, as the t-statistic shows -0.96. Also, the Wilcoxon rank-sum test shows no significance.
4.3.2 Takeover Premiums
Table 3. Takeover Premiums 1985-2012
This table shows descriptive statistics for takeover premiums. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. Takeover premiums are defined as the difference between the price paid by the acquirer minus the pre-acquisition target price at time -t, divided by the pre-pre-acquisition target price at time -t. The table provides an overview for the number of observations, mean, standard deviation, median, minimum and maximum for each variable of interest. All results in this table are denominated in percentage terms. The symbols ***, **, * denote statistical significance for the two-sample t-test (mean) and Wilcoxon rank-sum test (median) at the 1%, 5%, and 10% level, respectively.
Descriptive Statistics for Takeover Premiums
(1) (2) (3) (4) (5) (6)
N Mean Standard Deviation Median Minimum Maximum
Takeover Premium[-1] 398 40.85 52.21 30.49 -66.95 465.22 Takeover Premium[-1] 241 40.97 46.34 31.79 -53.73 465.22 Cash Acquirers Takeover Premium[-1] 157 40.66 60.27 28.89 -66.95 447.65 Equity Acquirers Takeover Premium[-5] 398 47.08 52.13 34.93 -69.28 454.32 Takeover Premium[-5] 241 48.03 43.17 37.83 -41.15 306.26 Cash Acquirers Takeover Premium[-5] 157 45.62 63.60 29.42** -69.28 454.32 Equity Acquirers Takeover Premium[-20] 398 53.10 56.46 40.81 -97.90 461.25 Takeover Premium[-20] 241 51.69 42.67 43.84 -97.90 213..04 Cash Acquirers Takeover Premium[-20] 157 55.26 72.82 38.20 -68.22 461.25 Equity Acquirers
24 In table 3 are the descriptive statistics for the takeover premiums of this study’s data sample. Measured at 3 different intervals, acquirers tend to overpay for the target firm between 40% and 53% on average. Furthermore, there is a clear trend over time. The average of the takeover premiums rises if we move further away from the announcement date. This is consistent with prior evidence, showing a run-up in share prices as it becomes apparent that the acquiring firm is going to make an acquisition announcement.
Relating to the SMDA theory, this study does not find much evidence regarding that equity acquirers pay higher premiums compared to cash acquirers. Only for takeover premiums 20 days prior to the announcement date do equity acquirers overpay cash acquirers 3.50% on average. At the 1 day prior to the announcement date, equity acquirers pay roughly the same on average as cash acquirers. Where measured at 5 days prior to the announcement, equity
acquirers pay about 2.50% less on average. Overall, the evidence does not support the hypotheses. Using the tests described in the methodology part, this study does not find any significant differences in the takeover premiums between the 2 groups. Their two-sample t-test values are -0.05, -0.42, and 0.55, respectively. However, for the Wilcoxon rank-sum test, the value is significantly negative for takeover premiums at the second time interval. This supports the previous results that stock acquirers seem to pay less than cash acquirers.
The average takeover premium per year is shown in figure 6 of appendix 2. When analyzing, it is visible that there is a rising trend towards the end of the fifth merger wave and dot.com bubble. After the burst of the bubble, takeover premiums decreased and subsequently started to rise again towards the global financial crisis. Thereafter, takeover premiums decreased again and starting to rise again towards 2012.
25 4.3.3 Cumulative Abnormal Returns
Table 4. Cumulative Abnormal Returns 1985-2012
This table shows descriptive statistics for cumulative abnormal returns. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. CARs are defined as the cumulative returns of the acquiring firm minus the cumulative returns of the acquirer’s equity index, measured over a -1 to +1 day at the announcement date. The table provides an overview for the number of observations, mean, standard deviation, median, minimum and maximum for each variable of interest. All results in this table are denominated in percentage terms. The symbols ***, **, * denote statistical significance for the two-sample t-test (mean) and Wilcoxon rank-sum test (median) at the 1%, 5%, and 10% level, respectively.
Descriptive Statistics for Cumulative Abnormal Returns
(1) (2) (3) (4) (5) (6)
N Mean Standard Deviation Median Minimum Maximum
CAR[-1,1] 398 -0.68 8.59 -0.68 -35.49 96.87
CAR[-1,1] 241 0.48 8.37 0.33 -35.49 96.87
Cash Acquirers
CAR[-1,1] 157 -2.47*** 8.65 -2.05*** -30.87 37.00
Equity Acquirers
In table 4 an overview is provided for the cumulative abnormal returns on a 1 day prior till 1 day after the announcement date time window. The CARs for this study´s sample show in general a small negative return around the announcement date. The standard deviation is almost equal across the sample in general as well as the subsamples. If we would compare the cash and equity acquirers, you would notice that cash acquirers perform on average almost 3% better than equity acquirers. Furthermore, the finding that the CARs of equity acquirers tend to be negative is consistent with hypothesis 3. Based on prior empirical evidence, the hypothesis stated that CARs of equity acquirers are negative around the announcement date.
The findings are consistent with prior documented evidence which focused on domestic M&As. Andrade et al. (2001) finds that cash acquirers also have a positive return of 0.4% and equity acquirers a negative performance of -1.5% around the announcement date. The results indicate that cross-border cash acquirers have the same performance as domestic ones, and that cross-border equity acquirers perform worse relative to domestic acquirers.
When testing for a significant difference in the mean CAR of both groups, the results indicate that equity acquirers significantly underperform. Using the two-sample t-test described in the methodology part, the t-value is -3.36 and is significant at the 1% significance level. This finding is consistent with the third hypothesis and with prior empirical evidence. In addition, the score of the Wilcoxon rank-sum test is also significant, thereby supporting the previous result.
The average CAR per year is shown in figure 7 of appendix 2. It is visible that, for most years, the average is around 0%. What is more interesting is that the average of the CARs does not rise during prosperous economic times. Around 2000, the end of the fifth merger wave and the dot.com bubble, average is still around 0% and even negative.
26 4.3.4 Buy-and-Hold Abnormal Returns
Table 5. Buy-and-Hold Abnormal Returns 1985-2012
This table shows descriptive statistics for buy-and-hold abnormal returns. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. BHARs are measured as the monthly holding return from 0 to 36 months after the completion date, against 3 different benchmarks. The table provides an overview for the number of observations, mean, standard deviation, median, minimum and maximum for each variable of interest. All results in this table are denominated in percentage terms. The symbols ***, **, * denote statistical significance for the two-sample t-test (mean) and Wilcoxon rank-sum test (median) at the 1%, 5%, and 10% level, respectively.
Descriptive Statistics for Buy-and-Hold Abnormal Returns
(1) (2) (3) (4) (5) (6)
N Mean Standard Deviation Median Minimum Maximum
Benchmark: Market BHAR[0,36] 310 -10.52 73.82 -11.63 -210.27 609.18 BHAR[0,36] 197 0.88 73.04 2.42 -210.27 609.18 Cash Acquirers BHAR[0,36] 113 -30.41*** 71.23 -38.57*** -192.21 317.08 Equity Acquirers
Fama-French Portfolios based on size of Market Equity Benchmark: Value Weighted Portfolio
BHAR[0,36] 189 -21.71 81.35 -31.40 -171.42 530.38
BHAR[0,36] 105 -3.27 80.28 -2.57 -140.55 530.28
Cash Acquirers
BHAR[0,36] 84 -44.47*** 77.10 -62.18*** -171.42 327.84
Equity Acquirers
Benchmark: Equal Weighted Portfolio
BHAR[0,36] 189 -22.66 81.05 -34.83 -228.63 496.41
BHAR[0,36] 105 -3.85 77.95 -4.80 -138.48 496.41
Cash Acquirers
BHAR[0,36] 84 -46.18*** 79.10 -60.47*** -228.63 326.34
Equity Acquirers
Table 5 shows the descriptive statistics for the BHARs on the 0 to 36 months time window after the completion date. The number of observations is lower compared to the previous tables, because not all long-run returns could be obtained. On average, the BHARs are all negative against the different benchmark returns. Against the performance of the acquiring firm’s country main exchange, the firms perform the best. Cash acquirers show a small positive return on a 36 months window, while equity acquirers clearly underperform. This finding is consistent with hypothesis 4, where is expected that equity acquirers should underperform cash acquirers on the long-run. This result supports the SMDA theory, which predicts that overvalued shares should revert to their fundamental value on the long-run.
27 Regarding the BHARs against the portfolios based on size of market equity, the number of observations is lower than against the market. This is due to that the portfolio returns were only measured against long-run returns of firms in Canada and the US. Based on these portfolio returns, cash acquirers also have negative performance on the long-run. However, equity acquirers perform much worse. These results also correspond with hypothesis 4, supporting prior evidence regarding long-term performance of equity acquirers.
Using the two-sample t-test described in the methodology section, this study finds significant long-run underperformance of equity acquirers compared to cash acquirers. Their t-values are -3.69, -3.58, and -3.68, respectively. These scores are significant at the 1%
significance level. In addition, the Wilcoxon rank-sum test shows corresponding results with scores significant at the 1% level. The values indicate -4.68, -4.48, and -4.38, respectively.
In figure 8 of the appendix 2 is the average BHAR shown per year, measured against different benchmarks. It is clearly visible that towards well-doing economic times, BHARs rise on a 36 months horizon. After the dot.com bubble and the global financial crisis there is a clear decrease in BHARs against the portfolios based on size of market equity. However, using the financial market as a benchmark, BHARs increased after the burst of the dot.com bubble. This indicates that acquiring firms still outperformed the market during harsh financial times.
28 5. Results
This section discusses the regression results regarding takeover premiums, CARS, and BHARs. In addition, this section discusses if the results are consistent with this study’s hypotheses.
5.1 Takeover Premiums
Table 6. Regression Results Takeover Premiums 1985-2012
This table shows the regression results for takeover premiums. The sample consists of cross-border M&A activity between Canada, France, Germany, the United Kingdom, and the United States between 1985 and 2012. Takeover premium is defined as the difference between the price paid by the acquirer minus the pre-acquisition price at time t, divided by the pre-pre-acquisition price at time t. For the definition of the other variables, refer to the appendix 1. Standard errors of the regressions are robust and the T-statistics are shown in the parentheses. The symbols ***, **, * denote statistical significance at the 1%, 5%, and 10% level, respectively.
Regression Results for Takeover Premiums
Takeover Premium[-1,0] Takeover Premium[-5,0] Takeover Premium[-20,0]
(1) (2) (3) (4) (5) (6) CSML R12 -0.126 -0.108 -0.261 -0.241 -0.351** -0.338** (-0.78) (-0.65) (-1.60) (-1.43) (-2.20) (-2.04) EXCH R12 -0.714* -0.792** -0.604* -0.695** -0.193 -0.266 (-1.88) (-2.00) (-1.78) (-2.00) (-0.57) (-0.77) CSML R12 * -0.254 -0.310 -.076 -0.144 0.104 0.030 Dummy Stock (-1.01) (-1.21) (-0.27) (-0.51) (0.30) (0.09) EXCH R12 * 0.460 0.556 0.363 0.447 0.018 0.105 Dummy Stock (0.65) (0.79) (0.48) (0.59) (-0.03) (0.13) Dummy Stock -0.338 1.353 -5.151 -4.116 -1.425 -0.664 (-0.05) (0.19) (-0.69) (-0.53) (-0.16) (-0.07)
Control Variables Yes Yes Yes Yes Yes Yes
Constant Yes Yes Yes Yes Yes Yes
Include dummies for fixed effects between countries
No Yes No Yes No Yes
Observations 398 398 398 398 398 398
R2 0.0529 0.0707 0.0514 0.0723 0.0412 0.0575
Regarding the relationship between takeover premiums and the acquirer’s equity index return, this study predicts a positive relationship. Since relative good performance of the financial markets could enhance the probability of firms being overvalued, and according to the SMDA theory, overvalued firms issue equity as payment method. Therefore, these firms should be willing to pay a higher takeover premium and thereby take advantage of their overvaluation. For cash acquirers is also a positive relationship predicted, because good performance of the equity indices improves the access to external financing. When analyzing the results, it appears that the
