Valuation impact of currency depreciations on cross-border mergers and acquisitions: evidence from the Eurozone, UK and USA

Universidade do Minho Escola de Economia e Gestão Diogo Emanuel Queirós Meneses Valuation impact of currency depreciations on crossborder mergers and acquisitions – evidence from the Eurozone, UK and USA Master’s Dissertation: Master in Finance Trabalho realizado sob a orientação do: Professor Doutor Gilberto Loureiro July 2019 I Direitos de autor e condições de utilização do trabalho por terceiros Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial-SemDerivações CC BY-NC-ND II Agradecimentos Concluindo uma importante etapa no meu percurso académico, são várias as pessoas a quem quero deixar um agradecimento especial. Antes demais, quero agradecer ao meu orientador, o professor Gilberto Loureiro, que foi fundamental para a elaboração e conclusão deste trabalho, agradeço pela sua paciência, disponibilidade e partilha de conhecimentos. De seguida, quero deixar um agradecimento especial à minha mãe e às minhas irmãs que sempre me motivaram em todo o percurso académico e não me deixaram desistir. Por fim, quero agradecer às minhas colegas de mestrado, Catarina Pires e Sofia Rebelo por todos os momentos que passamos juntos que acabaram por tornar esta experiência inesquecível, e que certamente deixará saudades. III Statement of integrity I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. IV Abstract In this study I analyze the impact of currency shocks on the Cumulative Abnormal Returns (CAR) resulted from cross-border Mergers and Acquisitions (M&As) around the announcement day. Cross-border M&As represent almost one third of the deals. As there is not much information about it, this makes this analysis relevant. Furthermore, the scarce of information increases when we talk the ways that currency movements affect this kind of deals. Throughout this study, I analyze three different samples, with acquirers from Eurozone, United Kingdom and United States of America, each of them in relation to the rest of the world to see how the CARs are affected by currency shocks, which might or not, create incentives to pursue cross-border deals under depreciation effects. I believe that companies only would follow a cross-border deal under depreciation shocks, only if, they believe that those currency changes are temporary. They must be aware that a depreciation in the target’s currency decreases the cost of the deal, but it also reduces the revenues when converted to the local currency. To make this analysis, I rely on three clean samples with 1204 deals for the Eurozone sample, 763 deals for the UK sample and 2126 for the US sample, all the deals occur between 2009 and 2018. Attending the main target’s currencies, I find that Brazilian real and Australian dollar are currencies that depreciate more frequently in relation to Euro and the US dollar. On the UK side, the currencies that depreciate more frequently in relation to the pound are the Australian and the Canadian dollar. Despite I do not find many significant variables related to currency depreciations, the ones that I find when isolated from other variables reveal positive effects on the CARs. In the hypothesis 1a) and 1b) the Eurozone sample reveals a positive impact of 1.84 percentage points for the CAR(-2,+2) and 0.677 percentage points for the CAR(-1,+1), respectively, when the deal occurs 4 quarters after a currency shock. Moving to hypothesis 2b) the UK sample reveals a positive impact of 4.23,4.57 and 5.35 percentage points for CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5), when there were at least 2 currency shocks in 4 quarters before the deal. V Keywords: cross-border M&As, Cumulative Abnormal Returns, currency shocks VI Resumo Neste estudo vou analisar o impacto de choques cambiais nos Cumulative Abnormal Returns (CAR), resultantes de fusões e aquisições internacionais, na altura do anúncio do negócio. Este tipo de negócios representa quase um terço do número total de fusões e aquisições. Esta análise torna-se relevante, porque apesar de haver muitos estudos para fusões e aquisições, não há muitos que abordem fusões e aquisições internacionais. Essa escassez aumenta quando falamos de efeitos de câmbio em fusões e aquisições. Ao longo deste trabalho, vamos analisar 3 amostras diferentes, cada uma delas com adquirentes da Zona Euro, do Reino Unido e dos Estados Unidos da América, em relação ao resto do mundo, para ser possível analisar como reagem à volatilidade das taxas de câmbio, o que poderá criar incentivos a seguir uma fusão ou aquisição internacional. Acredito que as empresas apenas sigam um negócio deste tipo quando houver efeitos de câmbio, apenas se acreditarem que as mudanças cambiais são temporárias. Por um lado, a depreciação da moeda da empresa alvo decresce o custo do negócio, mas por outro, também reduz as receitas, quando convertidas para a moeda do adquirente. Para realizar este estudo, vou analisar três amostras com 1204 negócios para a amostra da zona Euro, 763 para a amostra do Reino Unido e 2126 para a amostra dos Estados Unidos. Todos os negócios estão compreendidos entre o período temporal 2009 a 2018. Atendendo às moedas dos principais alvos dos adquirentes, o real do Brasil e o dólar australiano são as moedas que se depreciam mais frequentemente em relação ao euro e ao dólar americano. Do lado do Reino Unido, as moedas que se depreciam mais frequentemente em relação à libra são o dólar australiano e canadiano. Apesar de não ter encontrado muitas variáveis significativas relacionadas com o câmbio da moeda, as que encontrei e isoladas de outras variáveis, revelam efeitos positivos nos CARs. Na hipótese 1a) e 1b) a amostra da zona Euro revela um impacto positivo de 1.84 pontos percentuais para o CAR(-2,+2) e 0.677 pontos percentuais para o CAR(-1,+1), respetivamente, quando o negócio ocorre nos 4 trimestres depois de um choque cambial. Na hipótese 2b) a amostra do Reino Unido VII revela um impacto positivo de 4.23, 4.57 e 5.35 pontos percentuais para o CAR(-1,+1), CAR(-2,+2) e CAR(-5,+5), quando houveram pelo menos 2 choques cambiais nos 4 trimestres antes do negócio. Palavras-Chave: fusões e aquisições internacionais, CARS, desvalorização da moeda VIII Tables of contents 1. Introduction ............................................................................................................................ 1 2. Literature Review .................................................................................................................... 3 2.1 M&A Motivations ................................................................................................................... 3 2.2 Market Value and Currency Movements................................................................................. 3 2.3 Financial Crisis and its impact on M&A ................................................................................. 6 2.4 Barriers and catalysts to M&As ............................................................................................. 6 2.5 Determinants of acquirer’s returns ........................................................................................ 7 3. Hypotheses .............................................................................................................................. 11 4. Methodology ............................................................................................................................. 13 4.1 Event Study ........................................................................................................................ 13 4.2 Regressions ........................................................................................................................ 14 5.Data .......................................................................................................................................... 17 5.1 Currency Shocks ................................................................................................................ 26 6. Empirical Results ...................................................................................................................... 32 6.1 Hypothesis 1a) ................................................................................................................... 38 6.2 Hypothesis 1b) ................................................................................................................... 38 6.3 Hypothesis 2a) ................................................................................................................... 46 6.4 Hypothesis 2b) ................................................................................................................... 46 7. Conclusion ............................................................................................................................... 48 References ................................................................................................................................... 50 Appendix ...................................................................................................................................... 53 6 2.3 Financial Crisis and its impact on M&A Financial crisis is another factor that can affect M&As, the crisis from 2008 to 2009 demanded struggling industries to consolidate in the economic downturn, and had effects on the global wealth distribution leading to shifting locations of potential growth; cross-border deals increased and were expanded to other areas such as Brazil, Russia, India, and China (Grave et al., 2012). In a financial scenario, it is important to join forces to improve expansion, growth, and innovation to overcome difficulties. Financial crisis brought the sub-prime impact and put the global financial sector in need of a revolutionary reform, with risks of bankruptcy, growing deficits and loss of confidence on financial institutions. As a result, small companies were really struggling and only a few large companies could take advantage of consolidating the market by acquiring firms in their core industries (Grave et al., 2012). China prospered in this financial crisis buying companies that suffer effects of sovereign-debt, focusing essentially on natural resources industry that has pioneered M&As innovation (Grave et al., 2012). Africa, Latin America and parts of Asia companies have also made some steps to grow and started buying labor force to specific industry skills in some regions and then proceed to educate and build skills and offer good conditions to them to keep the key workers (Grave et al., 2012). 2.4 Barriers and catalysts to M&As Mergers and Acquisitions involve a complex process, especially in cross-border deals. There are several features to pay attention to that could facilitate or make the deal harder. Differences or shocks in terms of language, culture, believes and geographic differences are some of the turnoffs that we could find in the process, while, differences in valuation, governance standards and access to lower costs of production could make the process smoother. But the intention of the acquirer could be beyond 7 that, the technology, communication, travel improvements have facilitated deals for firms which want to enhance its global position and wide diversity (Grave et al., 2012). Most of the studies agree that cultural differences are barriers to cross-border M&As. However, it was found that in contrast with the announcement effect, in the long run, acquisitions are more successful if acquirer and target come from countries with different cultures (Chakrabarti, GuptaMukherjee, & Jayaraman, 2009). A merger or acquisition abroad could also be motivated by the acquirer’s will to get a quicker access to the markets where the target is located (Sharma, 2016). 2.5 Determinants of acquirer’s returns Target status classify companies as public, private, joint ventures, subsidiaries and government companies, which has influence on bidder’s returns, essentially through two hypotheses. The first related to managerial motive, which says that bidder is not willing to pay high premiums for smaller and less-known unlisted firms motivated by synergies that could be created with the deal, which does not increase the acquisition price. It is important to keep on mind that is easier to integrate small private targets than large listed targets. For these reasons some authors believe that bidders could have more benefits by acquiring private targets than public targets (Draper & Paudyal, 2006). The second hypothesis is related to liquidity. Good standards of information are required in order to not have liquidity issues. However, usually there is lack and poor information for unlisted targets, which arises problems of liquidity. Due to illiquid targets, acquirer’s bargaining power is enhanced, which also suggests that bidders of private targets are more likely to have higher gains, increasing the likelihood of underpaying (Chang, 1998; Draper & Paudyal, 2006). Draper and Paudyal (2006) split their sample into two, one for listed targets and the other for unlisted. They find that acquirers of public targets deal with a significant loss, on average, of 0.4% around the deal announcement. On the other hand, a positive return up to 2,19% is observed when 8 unlisted targets are involved. These results seem to support the hypotheses of managerial motive and liquidity. Focusing now on means of payment, there are many studies which supports the idea that the payment choices can influence the stock excess returns, and that happens due asymmetric information and, also, insufficient corporate monitoring. Apparently, the means of payment could transmit some signs about the market price and consequently influence the stock gains. If the bidder company sees its stocks undervalued the acquirer could pursue a merger or acquisition paid in cash, otherwise, the acquirer’s manager has incentives to choose a stock payment (Chang, 1998). It is expected a positive impact on returns if the payment is made in cash and a negative impact if it is made in stock (Loughran & Vijh, 1997). Still, there is some studies that find no abnormal returns from a cash offering (Chang, 1998). Other studies seem to agree with the idea that if the acquirer does the payment in stock suffers a decline in its share price, and the opposite happens if the payment is made in cash. The effect is expected to be the same also for the target’s stock price (Draper & Paudyal, 1999). A study by Draper and Paudyal (2006) shows that bidders that use cash as mean of payment, reveal excess returns around the announcement date, by approximately 2%. On the other hand, and contrary to other studies, they did not find a significant loss, on average, regarding payments in shares. The results are slightly different when the analysis is made for listed and unlisted targets. Regarding listed targets, the results are sensitive concerning cash payments and there is a significant drop in case of a stock payment, empowering the hypothesis of asymmetric information, and this payment could be comparable to the act of issuing equity to the public. As far as unlisted targets are concerned, a cash payment has a significant impact on returns, and so does a payment in stock, however, in the last method of payment the results are not significant post-event period. In this study, the impact on bidder’s shareholders for unlisted targets are positive either for cash or stock payment (Draper & Paudyal, 2006). 9 Privately held targets acquisitions paid in stock may create blockholders, since private targets are managed by a small group of people. Blockholders can have two different effects, it could serve as a group of people who cares about monitoring managerial performance (Shleifer & Vishny, 1986). However, the reverse effect could happen if that ownership concentration make the blockholders acting to maximize only their wealth or if makes deals more expensive (Fama& Jensen, 1983), Stulz (1988), (Morck, Shleifer, & Vishny, 1988). A payment in stock can also have a negative impact on bidder’s returns if the acquirer pays in stock to a publicly traded target with many shareholders who are dealing with asymmetric information (Myers & Majluf, 1984). Some studies have found that companies that are used to make many mergers and acquisitions, called serial acquirers, earn lower returns up a certain number of deals (Boubakri, Chan, & Kooli, 2012; Fuller, Netter, & Stegemoller, 2002). This result may be explained by “anticipation effect”, which means that the market is able to identify serial acquirers, and, therefore, the announcement returns are reduced with additional M&As (Karolyi, Liao, & Loureiro, 2015). Another explanation to the lower announcement returns could be given by agency costs required to monitor serial acquirers, that can destroy investment value to create an “illusion of growth” (Jensen, 2005), that illusion could be motivated by acquirer’s overconfident after successful previous deals – management hubris (Billett & Qian, 2008). Management hubris is many times related to the firm’s size, the bigger the company is, more confident the managers are. Usually large firms start from negative dollar synergies due to the higher acquisition premiums that they are willing to pay. That could explain why small firms overperform large firms, exceeding the large firms’ abnormal returns by 2.24 percentage points (Moeller, Schlingemann & Stulz, 2004). Serial acquirers are gaining importance as far as M&As are concerned, to reinforce this idea it was found that 1/5 of listed acquirers is a serial acquirer. Serials acquirers act not only in the domestic market and industry, but, they pursue more and more in cross-border deals and in different industries, which is motivated by changes of industrial competition (Karolyi et al., 2015). 10 Other factor that has impact on bidder’s returns in this kind of deals is differences between the bidder and target corporate governance standards. If acquirer is from a country that offers better protection, then, the synergies generated could be higher, and it may be anticipated that target assets would be better managed, leading to abnormal returns (Martynova & Renneboog, 2008). Also, macroeconomic factors, such as GDP (that can be related to corporate governance) may lead to international expansion, through mergers and acquisitions, for example (Boateng, Hua, Uddin, & Du, 2013). 11 3. Hypotheses As it was mentioned, there is lack of studies that relate the effects of currency depreciations with the impact of returns on cross-border deals. That’s why I will try to establish a link between depreciation variables with target status and acquirer’s and target’s industries. Based on literature review a depreciation event should bring positive returns to the bidder (Goldberg & Koistad, 1994). Plus, if the company is seeking a specific asset that is able to generate returns in multiple currencies, it may take advantage from a currency depreciation to get access to get access to that asset at a lower cost (Guo & Trivedi, 2002). On the side of FDI (which includes crossborder M&As), the higher is the exchange rate volatility the higher is the investment (Guo & Trivedi, 2002). Given this, it is expected that depreciation leads to higher CARs. Regarding target status, if the target is private, the announcement returns are expected to be higher regardless the mean of payment used (Draper & Paudyal, 2006). So, in interaction with currency depreciation variables it is expected a positive impact on the CARs from this interaction. At last, if the acquirer and target play in the same industry the impact of currency shocks on bidder’s returns is expected to be positive (Harris & Ravenscraft, 1991), which makes me believe that the interaction of the industry variables with depreciation variables would once more lead to a positive impact on the CARs. In this study I have to lower the recommended level of depreciation to consider a currency shock, that’s why besides the variable “aftershock” I also use the variable “multipleshocks” to see if the effect of the second is stronger or more significant. For this reason, I will test two hypotheses, one of them related to the variable “aftershock” and the other with the variable “multipleshocks”. Each hypothesis will be divided into two parts, according to the interactions above mentioned. I am going to test the following hypotheses: H1) bidders obtain higher announcement returns when they acquire a target from countries that have suffered a currency depreciation. H1a) The effect of H1 is enlarge when the target is private 12 H1b) The effect of H1 is enlarge when acquirer and target play in the same industry H2) bidders obtain higher announcement returns when they acquire a target from countries that have suffered multiple currency depreciations. H2a) The effect of H2 is enlarge when the target is private H2a) The effect of H2 is enlarge when acquirer and target play in the same industry 13 4 Methodology 4.1 Event Study This study aims to measure the impact of currency shocks on announcement returns in crossborder M&As for acquirers from Eurozone, the UK and the USA. In order to measure those returns, I will use the same methodology used by Frankel and Rose (1996) event study methodology, that is the most used method to assess the value creation resulted of M&As. An event study is made in order to measure a specific event on the value of a given firm, in this case, we will measure the impact of a cross-border merger or acquisition on the firm value. First, we have to define an event window, that is typically very short, a few days before the event and a few days after. We also, need to define the estimation window, that will end a few days before the event window, to not take the risk of contamination. The estimation window (by average) will help us estimating the expected stock price if the event doesn’t occur. Then, we just need to compute the difference between the expected stock price without the event and the actual stock price to get the expected returns. The abnormal return (AR) for any company j in day t is given by: 𝐴𝑅𝑖𝜏 = 𝑅𝑗𝜏 − 𝛼𝑗  − 𝛽𝑖 𝑅𝑚𝜏, where 𝑅𝑖𝜏 represents the realized stock return at day t, the predicted return is given by 𝛼𝑗  − 𝛽𝑗 𝑅𝑚𝜏 that is the estimate of the shareholder return if the takeover hadn’t happened, finally 𝜏 may assume different values, is equal to zero at the time of the announcement, is equal to 1 one day after the announcement, is equal to -1 one day before the announcement and so on. It is important to say that the control return is estimated from α and β that are computed from the regression of the individual firm return, based on the market return (𝑅𝑚𝜏) for the 255-day period ending 25 days before the announcement. Finally, the cumulative abnormal return - (CAR) between any 2 dates (𝜏𝑎 and 𝜏𝑏) is given by: 14 𝐶𝐴𝑅𝑗(𝜏1,𝜏2)=∑𝐴𝑅𝑗𝜏 𝑖=𝜏𝑎 𝑖=𝜏𝑏. In this study I use three window lengths, ±1 (𝜏𝑎= −1 and 𝜏𝑏= 1),±2 (𝜏𝑎= −2 and 𝜏𝑏= 2) and ±5 (𝜏𝑎= −5 and 𝜏𝑏= 5). The CARs will be computed for each deal. 4.2 Regressions In order to test if a currency depreciation in the target’s country leads to positive abnormal returns for the bidder, being the dependent variable the Cumulative Abnormal Returns (CAR), I will test the null hypothesis - the event has no impact on firm value. As currency shocks would take time to be noticed, under this hypothesis, I use the variable “aftershock”, which is the variable of greatest interest in this analysis, it is a dummy variable and is equal to one in the four quarters after the shock. Moreover, there will be an interaction of this variable with acquirer and target’s industries and target public status. Under the first hypothesis the following regressions are used for each of the three samples in this study: 1𝑐𝑎𝑟 (𝜏1, 𝜏2)=𝛽0+ 𝛽1𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡+𝛽2𝑐𝑎𝑠ℎ+𝛽3𝑠𝑡𝑜𝑐𝑘+𝛽4𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽5𝑠𝑒𝑟𝑖𝑎𝑙𝑎𝑐𝑞 + 𝛽6𝑎𝑐𝑞𝑔𝑑𝑝(𝑡) +𝛽7𝑙𝑛𝑎𝑐𝑞𝑠𝑖𝑧𝑒(𝑡 − 1) + 𝛽8𝑀𝐵𝑉(𝑡 − 1) + 𝛽9𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒(𝑡 − 1) + 𝛽10𝑎𝑓𝑡𝑒𝑟𝑠ℎ𝑜𝑐𝑘 + 𝛽11𝑎𝑓𝑡𝑒𝑟𝑠𝑐ℎ𝑜𝑘_𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡 + 𝛿1𝑎𝑐𝑞𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿2𝑡𝑎𝑟𝑔𝑒𝑡𝑛𝑎𝑡𝑖𝑜𝑛 + 𝜀 1𝑎) Where, 𝑐𝑎𝑟 (𝜏1, 𝜏2) is the dependent variable that represents the cumulative abnormal return for a given event window, which is computed for each deal; “privatetarget” is a dummy variable, which is equal to 1 if the target is private; “cash” is a dummy variable, which is equal to 1 if the deal payment is made exclusively in cash; “stock” is a dummy variable, which is equal to 1 if the deal payment is made exclusively in stock; “sameindustry” is a dummy variable, which is equal to 1 if acquirer and target operate in the same industry; “serialacq” is a dummy variable, which is equal to 1 if over the all sample the bidder acquires more than five times; “acqgdp” is the GDP growth rate of the acquirer’s nation in the year of the deal; “lnacqsize( t-1)” is the logarithmic variable of the acquirer’s total assets, one year before the deal; “MBV t-1” is the logarithmic variable of the acquirer’s market-to-book value, 15 one year before the deal, in the acquirer’s currency ; “leverage t-1” is the ratio of Total Liabilities/Total Assets, one year before the deal, in the acquirer’s currency; “aftershock” is a dummy variable, which is equal to 1 in the four quarters after the currency shock; “aftershock_privatetarget” is a dummy variable, which is equal to 1 for private targets four quarters after a currency shock ; “acqindustry” represents dummies for acquirer industries (classified by the 2-digit SIC-codes); finally “targetnation” represents dummies for the target’s Nation . 𝑐𝑎𝑟 (𝜏1,𝜏2)=𝛽0+ 𝛽1𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡+𝛽2𝑐𝑎𝑠ℎ+𝛽3𝑠𝑡𝑜𝑐𝑘+𝛽4𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽5𝑠𝑒𝑟𝑖𝑎𝑙𝑎𝑐𝑞 + 𝛽6𝑎𝑐𝑞𝑔𝑑𝑝(𝑡) + 𝛽7 𝑙𝑛𝑎𝑐𝑞𝑠𝑖𝑧𝑒(𝑡 − 1) + 𝛽8 𝑀𝐵𝑉(𝑡 − 1) + 𝛽9 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒(𝑡 − 1) + 𝛽10 𝑎𝑓𝑡𝑒𝑟𝑠ℎ𝑜𝑐𝑘+𝛽11𝑎𝑓𝑡𝑒𝑟𝑠ℎ𝑜𝑐𝑘_𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿1𝑎𝑐𝑞𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿2𝑡𝑎𝑟𝑔𝑒𝑡𝑛𝑎𝑡𝑖𝑜𝑛+ 𝜀 1𝑏) All the variables are defined above, except “aftershock_sameindustry” that is a dummy variable, which is equal to 1 for acquirers and targets that play in the same industry (classified by the 2-digit SIC-codes) four quarters after a currency shock. Under the second hypothesis the following regressions are used for each of the three samples in this study: 𝑐𝑎𝑟 (𝜏1,𝜏2)=𝛽0+ 𝛽1𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡+𝛽2𝑐𝑎𝑠ℎ+𝛽3𝑠𝑡𝑜𝑐𝑘+𝛽4𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽5𝑠𝑒𝑟𝑖𝑎𝑙𝑎𝑐𝑞 + 𝛽6𝑎𝑐𝑞𝑔𝑑𝑝(𝑡) +𝛽7𝑙𝑛𝑎𝑐𝑞𝑠𝑖𝑧𝑒(𝑡 − 1) + 𝛽8𝑀𝐵𝑉(𝑡 − 1) + 𝛽9𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒(𝑡 − 1) + 𝛽10𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑒𝑠ℎ𝑜𝑐𝑘𝑠 + 𝛽11𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑒𝑠ℎ𝑜𝑐𝑘𝑠_𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡 + 𝛿1𝑎𝑐𝑞𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿2𝑡𝑎𝑟𝑔𝑒𝑡𝑛𝑎𝑡𝑖𝑜𝑛 + 𝜀 2𝑎) All the variables are defined above, except “multipleshocks” that is a dummy variable, which is equal to 1 if in the last four quarters there were at least two currency shocks; and also the variable “multipleshocks_privatetarget that is a dummy variable, which is equal to 1 for private targets if in the last four quarters there were at least two currency shocks. 𝑐𝑎𝑟 (𝜏1,𝜏2)=𝛽0+ 𝛽1𝑝𝑟𝑖𝑣𝑎𝑡𝑒𝑡𝑎𝑟𝑔𝑒𝑡+𝛽2𝑐𝑎𝑠ℎ+𝛽3𝑠𝑡𝑜𝑐𝑘+𝛽4𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛽5𝑠𝑒𝑟𝑖𝑎𝑙𝑎𝑐𝑞 + 𝛽6𝑎𝑐𝑞𝑔𝑑𝑝(𝑡) +𝛽7 𝑙𝑛𝑎𝑐𝑞𝑠𝑖𝑧𝑒(𝑡 − 1) + 𝛽8 𝑀𝐵𝑉(𝑡 − 1) + 𝛽9 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒(𝑡 − 1) + 𝛽10 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑒𝑠ℎ𝑜𝑘𝑠+𝛽11𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑒𝑠ℎ𝑜𝑐𝑘𝑠_𝑠𝑎𝑚𝑒𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿1𝑎𝑐𝑞𝑖𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + 𝛿2𝑡𝑎𝑟𝑔𝑒𝑡𝑛𝑎𝑡𝑖𝑜𝑛+ 𝜀 2𝑏) 22 Finally, the leverage ratio in the year before the deal, “leverage(t-1)”, which is computed by the ratio Total Liabilities (t-1) / Total Assets (t-1), is once more quite similar for the Eurozone, the UK and the USA, regarding medians, that display values of 0.58,0.50,0,52, respectively. 23 Table 3 Descriptive statistics (1) (2) (3) (4) (5) (6) VARIABLES N Mean p50 Sd min max Privatetarget 1,195 0.618 1 0.486 0 1 Cash 1,195 0.120 0 0.325 0 1 Stock 1,195 0.0117 0 0.108 0 1 Sameindustry 1,195 0.520 1 0.500 0 1 Serialacq 1,195 0.521 1 0.500 0 1 Aftershock 1,195 0.244 0 0.430 0 1 Multipleshocks 1,195 0.0770 0 0.267 0 1 Shock 1,195 0.0770 0 0.267 0 1 Acqgdp 1,195 1.338 1.551 2.782 -9.132 25.12 Totalassets(t-1) million€ 1,195 8.595 2.699 1.1940 67,709 3.8790 MBV (t-1) 1,195 2.321 2.040 1.228 0.890 4.680 Leverage (t-1) 1,195 0.579 0.582 0.143 0.338 0.786 Table 3 shows descriptive statistics for the main variables of each sample, the variables “acqgdp”, “totalassets” and “MBV” were winsoried at 1% level of each tail, in order to remove extreme values that could bias our results. The meaning of each variable can be consulted in the appendix. Panel A: Eurozone sample 24 (1) (2) (3) (4) (5) (6) VARIABLES N Mean p50 Sd min max Privatetarget 760 0.626 1 0.484 0 1 Cash 760 0.272 0 0.445 0 1 Stock 760 0.0250 0 0.156 0 1 Sameindustry 760 0.513 1 0.500 0 1 Serialacq 760 0.546 1 0.498 0 1 Aftershock 760 0.133 0 0.340 0 1 Multipleshocks 760 0.0224 0 0.148 0 1 Shock 760 0.0461 0 0.210 0 1 Acqgdp 760 1.397 1.823 1.830 -4.188 2.948 Totalassets(t-1) million £ 760 2.340 0.690338 3.637 33,128 1.2930 MBV(t-1) 760 2.730 2.280 1.640 0.890 5.980 Leverage (t-1) 760 0.502 0.504 0.185 0.208 0.799 Panel B: UK sample 25 (1) (2) (3) (4) (5) (6) VARIABLES N mean p50 sd min max Privatetarget 2,118 0.656 1 0.475 0 1 Cash 2,118 0.152 0 0.359 0 1 Stock 2,118 0.0165 0 0.128 0 1 Sameindustry 2,118 0.500 1 0.500 0 1 Serialacq 2,118 0.416 0 0.493 0 1 Aftershock 2,118 0.431 0 0.495 0 1 Multipleshocks 2,118 0.234 0 0.424 0 1 Shock 2,118 0.183 0 0.387 0 1 Acqgdp 2,118 1.923 2.250 1.272 -2.537 2.881 Totalassets(t-1) million$ 2,118 2.700 1.030 3.638 66,460 1.2410 MBV 2,118 2.847 2.430 1.675 0.930 6.270 Leverage 2,118 0.524 0.523 0.193 0.219 0.827 Panel C: US sample 26 5.1 Currency Shocks In order to identify the most relevant variable in this work, is necessary to identify currency shocks as a result of depreciation episodes, which will be crucial to understand how shocks affect our CARs. Usually currency shocks episodes are characterized as a depreciation of at least 25%, in relation to another currency, for differences computed in relation to the same quarter of the previous year (Desai, Foley, & Forbes, 2008; Frankel & Roseb, 1996). However, following the same values I was not able to find any currency shocks for the target’s currencies for the UK sample, regarding the Eurozone sample I only found shock for Brazil and in the US sample I found 8 shocks for Brazil and 1 for France and Australia (no shocks for the remaining 3 targets). Lowering the reference value to 15% the currency shocks improve significantly for the US sample, and the Eurozone sample, however in the last sample I did not find shocks for Canada and Switzerland, for the UK sample I only got 1 shock for Canada and 2 shocks for Australia (no shocks for the three remaining target’s countries). In order to assess better results, I will consider that there is a currency shock when there is a depreciation of 10% at least, in relation to the same quarter in the year before – homologous analysis. Plus, I compute currency changes computing real exchange rates, using for that effect the CPI for the required countries. In this analysis was required to extract from Datastream the quarterly nominal exchange rates 4 from the main target’s currencies to acquirer’s currencies. For the Eurozone sample, I obtained ( x US $ to 1), ( x £ to 1€), ( x Swiss franc to 1€), ( x Brazilian real to 1€), ( x Swedish krona to 1€), ( x Canadian $ to 1€) and ( x Australian $ to 1€). Regarding the UK sample, I obtained ( x US $ to 1£), ( x Australian $ to 1£), ( x € to 1£) and ( x Canadian $ to 1£). 4 Exchange rates were collected from 4 quarters before the beginning of the sample untill the end, in order to identify shocks from the beginning if necessary. Those 4 quarters are necessary, because it is made a homologous analysis, and for that reason we will get missing values from the first 4 quarters. 27 Finally, for the last sample, I obtained ( x £ to 1 US $), ( x Canadian $ to 1 US $), ( x € to 1 US $), ( x Australian $ to 1 US $) and ( x Brazilian real to 1 US $). Then, those exchange rates were adjusted to the rate of inflation, using for that purpose the Consumer Price Index (CPI) 5 , of each target country, which was also gathered from Datastream. Table 4 Depreciations against acquirer's currency 5 Following the same line as exchange rates, also, the CPIs were gathered from 4 quarters before the sample untill 2018. Panel A: Depreciations against € US $ 3 shocks 2009(Q4) 2011(Q2) 2018(Q1) UK: £ 3 shocks 2009(Q1) 2016(Q3,Q4) Swiss franc 0 shocks - Brazilian real 10 shocks 2009(Q1) 2013(Q3,Q4) 2014(Q1) 2015(Q3,Q4) 2016(Q1) 2018(Q1,Q2,Q3) Sweden krona 2 shocks 2009(Q1,Q2) Canadian $ 3 shocks 2013(Q3,Q4) 2014(Q1) Panel B: Depreciations against £ US $ 1 shock 2009(Q4) Australia $ 4 shocks 2013(Q4) 2014(Q1,12) 2015(13) Canadian $ 4 shocks 2014(Q1,Q2),2015(Q3,Q4) Germany: € 1 shock 2015 (Q1) Netherlands: € 1 shock 2015(Q1) Table 4 shows the number of currency shocks found in the entire period for each sample, for year and quarters. 28 Looking first at panel A, we can see that Brazilian real is the currency that suffers more from currency depreciation shocks against euro, having 11 shocks, the second currency with more shocks (6) is the Australian dollar, then we have the pound, the US dollar and the Canadian dollar with 3 shocks, Sweden krona with 2 shocks and at last we have the Swiss franc without any shocks in the sample. Moving to panel B, we can observe that there are not many currency shock depreciations against £, we have 4 shocks for Australian and Canadian dollar and only 1 shock for the US dollar and for euro (from France and Netherlands). Finally, in panel C, we can observe the currency depreciations shocks in relation to the US dollar, as had happened in the Eurozone Sample, Brazil is once more the currency with more shocks (11), followed by Australian dollar, and euro (from Germany, France and Netherlands) with 9 shocks, then we have £ with 8 shocks and Canadian dollar with 7 shocks. The US sample is the one with more currency shocks in relation to the main target’s currencies. We can also highlight that either Eurozone sample or US sample display a devaluation of the pound in the 2 quarters after the Brexit referendum 2016(Q3,Q4), in the US Sample those shocks are also followed by another 2 shocks, 2017(Q1,Q2). Panel C: Depreciations against the US $ UK: £ 8 shocks 2009(Q1,Q2,Q3) 2015(Q2) 2016(Q3,Q4), 2017(Q1,Q2) Canadian $ 7 shocks 2009(Q1,Q2,Q3) 2015(Q2,Q3,Q4) 2016(Q1) Germany: € 9 shocks 2009(Q1,Q2) 2010(Q3) 2012(Q3) 2015(Q1,Q2,Q3,Q4) 2016(Q1) Australian $ 9 shocks 2009(Q1,Q2,Q3) 2013(Q3) 2014(Q1) 2015(Q2,Q3,Q4) 2016(Q1) France: € 9 shocks 2009(Q1,Q2) 2010(Q3) 2012(Q3) 2015(Q1,Q2,Q3,Q4) 2016(Q1) Netherlands:€ 9 shocks 2009(Q1,Q2) 2010(Q3) 2012(Q3) 2015(Q1,Q2,Q3,Q4) 2016(Q1) Brazilian real 11 shocks 2009(Q1,Q2,Q3), 2012(Q3), 2014(Q1), 2015(Q2,Q3,Q4), 2016(Q1), 2018(Q3,Q4) 29 Panel A: Eurozone sample Table 5 CARs analysis CAR(-1,+1) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 903 0.009 0.003 -2.345*** -1.071 1 292 0.017 0.004 diff (1-0) 1195 0.008 0.001 CAR(-2,+2) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 903 0.010 0.004 -1.075 0.256 1 292 0.011 0.005 diff (1-0) 1195 0.001 0.001 CAR(-5,+5) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 903 0.008 0.004 -0.2906 0.273 1 292 0.010 0.004 diff (1-0) 1195 0.002 0.000 Table 5 shows tests of equality of means and medians, it also shows t-stats and Wilcoxon-stats from the tests, respectively. *, **, *** stand for statistical significance at 10%, 5%, and 1% levels, respectively. 30 Panel B: UK sample Panel C: US sample CAR(-1,+1) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 659 0.012 0.005 -0.638 -0.266 1 101 0.016 0.008 diff (1-0) 760 0.004 0.003 CAR(-2,+2) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 659 0.012 0.005 -0.0228 0.438 1 101 0.012 0.004 diff (1-0) 760 0.000 -0.001 CAR(-5,+5) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 659 0.010 0.005 -0.423 0.077 1 101 0.014 0.006 diff (1-0) 760 0.004 0.001 CAR(-1,+1) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 1205 0,006 0,002 0,678 1,215 1 913 0,003 0,000 diff (1-0) 2118 -0,003 -0,002 CAR(-2,+2) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 1205 0,008 0,002 0,9166 0,567 1 913 0,002 0,001 diff (1-0) 2118 -0,006 -0,001 CAR(-5,+5) Aftershock obs Mean Median t-stats means Wilcoxon z-test 0 1205 0,003 -0,001 0,4164 0,156 1 913 -0,002 0,000 diff (1-0) 2118 -0,005 0,001 31 Looking at panel A from table 5 we can observe that the tests of equality of means and medians on CARS, that are computed for each deal. The results are only significant for the Eurozone CAR(- 1,+1), that is the cumulative abnormal return between one day before and one day after the announcement, which is statistically significant at 1% level and reveals that the mean for The CAR(- 1,+1) is always positive either if aftershock is equal to zero (0.9%) or to one(1.7%), but the impact is higher under the effect of aftershock=1, the difference between the means of CAR(-1,+1) when “aftershock”=1 and “aftershock”=0 is of 0.008, or 0.8 percentage points. The remaining insignificant Eurozone CARS are positive independent of the value that aftershock assumes, and there’s a higher impact for the deals that occur 4 quarters after a currency shock. The insignificant Eurozone medians are always positive and higher for deals that occurs 4 quarters after a currency shock, expect for CAR(- 5,+5). Looking at panel B, we can say that the UK Sample is not significant regarding the means and medians for the CARS. The table shows that all the CARS are positive either under the effect of depreciation or not, but they are higher for deals that occur 4 quarters after a currency shock, except for CAR(-2,+2) that the difference between the means is approximately zero. The medians are also positive for all situations, but higher for CAR(-1,+1) and CAR(-5,+5) when “aftershock”=1, and lower for the CAR(-2,+2). Moving to panel C, we can see that once more there is no significance for means and medians. The means for the CARS are always positive, except for CAR (-5,+5) when aftershock=1. Looking at the differences, we can also see that CARS are higher when “aftershock”=0, for all the CARS, which leads to better results when there is not a depreciation effect. Finally, looking at the medians they have always positive values whatever the value that “aftershock” assumes, except for CAR(-5,+5) when “aftershock”=0. The medians are higher when “aftershock”=0 for CAR(-1,+1) and CAR(-2,+2) and lower for CAR(-5,+5). 38 6.1 Hypothesis 1a) Regarding our hypothesis H1 a) we can see that the variable “aftershock” is only significant for the Eurozone sample, at 1% level, for CAR (-2,+2), displaying a beta of 0.0184, which means that, on average, a deal made 4 quarters after a currency shock, has a positive impact of 1.84 percentage points on the CAR(-2,+2), ceteris paribus. Despite all the remaining “aftershock” variables are insignificant for 1a) regressions, they have a positive impact for the Eurozone and UK CARs, and a negative impact on all the US CARs. The interaction between aftershock and privatetarget, variable “aftershock_privatetarget”, is only significant for the Eurozone sample, for the CAR(-2,+2) and CAR(-5,+5), at 1% level. Then, we can say, that, on average, a deal that involves a private target in the 4 quarters after a currency shock, has an impact on cars of -1.97 percentage points and -2.30 percentage points, respectively. Although the remaining “aftershock_privatetarget” variables are insignificant all of them have a negative impact on the car, except for the US CAR(-5,+5), ceteris paribus. Looking at the significant values I am not able to say that the hypothesis H1 is enlarged with the interaction between the variables “aftershock” and “privatetarget” for the Eurozone CAR(-2,+2) that is the only one with significant coefficients for the variable “aftershock” and “aftershock_privatetarget”. In this case a private target does not contribute to higher CARS. 6.2 Hypothesis 1b) Moving to the hypothesis 1b) related to 1b) regressions, we can see that once more, we only can observe a significant coefficient of 0.00677 for “aftershock”, significant at 1% level, from the Eurozone sample, CAR(-1,+1). Then, on average, when a deal occurs 4 quarters after a currency shock there’s a little positive impact of 0.677 percentage points on the CAR(-1,+1), ceteris paribus. All the remaining aftershock variables for this hypothesis are positive, expect the US CAR(-2,+2). 39 The interaction between aftershock and same industry, variable “aftershock_sameindustry”, is only significant for the US CAR(-1,+1) and CAR(-5,+5), at 5% and 1% levels, respectively. This means that on average, a deal that involves an acquirer and target that play in the same industry, have a negative impact of 2.27 percentage points and 4.4 percentage points on CAR(-1,+1) and CAR(-5,+5), respectively, ceteris paribus. The remaining “aftershock_sameindustry” for regresions 1b) are only positive for all the Eurozone sample, being consistent to the “aftershock” positive betas, and also for the UK CAR(-5,+5). To sum up, due to the insignificant coefficients I am not able to say that the bidder’s return in the 4 quarters after a currency shock are enlarged when target and acquirer play in the same industry. 40 Table 7 Regressions to test the hypotheses H2a) and H2b) H2a) H2a) H2a) H2b) H2b) H2b) VARIABLES car(-1,+1) car(-2,+2) car(-5,+5) car(-1,+1) car(-2,+2) car(-5,+5) Privatetarget 0.0002 0.0007 -0.0059 -0.0034 -0.0029 -0.00883** (0.0633) (0.211) (-1.424) (-1.167) (-0.903) (-2.152) Cash -0.0022 0.0016 0.0000 -0.0028 0.0010 -0.0007 (-0.614) (0.412) (0.00665) (-0.752) (0.246) (-0.127) Stock 0.0092 0.0174 0.0112 0.0081 0.0162 0.0116 (0.389) (0.694) (0.321) (0.341) (0.643) (0.332) Sameindustry 0.00847* 0.00927** 0.0121*** 0.00631** 0.00723** 0.00846** (1.896) (2.189) (2.708) (2.092) (2.184) (2.100) Serealacq -0.0003 -0.0003 -0.0006 0.0008 0.0008 0.0008 (-0.0872) (-0.0819) (-0.122) (0.209) (0.186) (0.153) Acqgdp(t) 0.0001 0.0003 0.0000 0.0000 0.0002 0.0000 (0.156) (0.532) (0.0565) (-0.0224) (0.368) (-0.0472) Lnacqsize(t-1) -0.0058*** -0.0059*** -0.0063*** -0.0061*** -0.0061*** -0.0066*** (-3.221) (-3.442) (-3.353) (-3.072) (-3.330) (-3.409) MBV(t-1) 0.0037 0.0034 0.0034 0.0036 0.0032 0.0032 Panel A: Eurozone regressions of CARS2nd hypothesis Table 6 displays the outputs for the regressions 2a) and 2b) for the CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5) for each sample. 41 (1.522) (1.477) (1.419) (1.503) (1.440) (1.374) Leverage(t-1) 0.0379** 0.0383** 0.0319 0.0389** 0.0394** 0.0320 (2.179) (2.228) (1.627) (2.181) (2.240) (1.621) Multipleshocks 0.0416 0.0366 0.0241 0.0029 -0.0014 -0.0200 (1.301) (1.353) (1.059) (0.389) (-0.165) (-1.472) multipleshocks_privatetarget -0.0459 -0.0463* -0.0376 (-1.491) (-1.739) (-1.501) multipleshocks_sameindustry 0.0266 0.0250 0.0475** (0.956) (1.018) (2.001) Constant 0.0498 0.0431 0.0855** 0.0549 0.0482 0.0906** (1.484) (1.342) (2.276) (1.550) (1.429) (2.360) Observations 1,195 1,195 1,195 1,195 1,195 1,195 R-squared 0.110 0.116 0.108 0.102 0.107 0.111 Acquirer Industry Dummy YES YES YES YES YES YES Target Nation Dummy YES YES YES YES YES YES Note: The estimations from the regression 2a) and 2b) presented in section 4 are presented in this table. The dependent variables are the CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5), all variables definition can be consulted in the appendix. The variables “acqgdp”, “totalassets” and “MBV” were winsoried at 1% level of each tail, in order to remove extreme values that could bias our results. . T-statistics are in parentheses. ***, ** or * indicates that the coefficient estimates are significant at 1%, 5% or 10% level, respectively. 42 H2a) H2a) H2a) H2b) H2b) H2b) VARIABLES car(-1,+1) car(-2,+2) car(-5,+5) car(-1,+1) car(-2,+2) car(-5,+5) privatetarget -0.0064 -0.0061 -0.0132 -0.0061 -0.0056 -0.0128 (-1.137) (-0.924) (-1.645) (-1.093) (-0.862) (-1.610) cash 0.0021 -0.0019 -0.0059 0.0022 -0.0019 -0.0057 (0.425) (-0.335) (-0.734) (0.434) (-0.326) (-0.718) stock -0.0034 0.0159 0.0007 -0.0028 0.0168 0.0014 (-0.107) (0.378) (0.0144) (-0.0880) (0.400) (0.0288) sameindustry -0.0068 -0.0074 -0.0034 -0.0062 -0.0068 -0.0023 (-1.499) (-1.418) (-0.473) (-1.358) (-1.272) (-0.313) serealacq -0.0016 -0.0060 -0.0095 -0.0013 -0.0056 -0.0090 (-0.277) (-0.857) (-1.030) (-0.224) (-0.791) (-0.971) acqgdp(t) -0.0029 -0.0030 -0.0043 -0.0029 -0.0030 -0.0043 (-1.613) (-1.480) (-1.563) (-1.616) (-1.486) (-1.561) lnacqsize(t-1) -0.0060*** -0.0053** -0.0036 -0.0060*** -0.0054** -0.0037 (-2.842) (-2.202) (-1.109) (-2.855) (-2.223) (-1.126) MBV(t-1) -0.0012 0.0003 0.0017 -0.0012 0.0003 0.0017 (-0.845) (0.155) (0.834) (-0.849) (0.150) (0.827) leverage(t-1) 0.0211 0.0182 0.0005 0.0217 0.0190 0.0014 (1.155) (0.816) (0.0183) (1.184) (0.854) (0.0490) multipleshocks 0.0005 -0.0149 -0.0108 0.0423** 0.0457** 0.0535** (0.0287) (-0.808) (-0.394) (2.530) (2.129) (2.169) multipleshocks_privatetarget 0.0377* 0.0571** 0.0509 (1.729) (2.218) (1.562) multipleshocks_sameindustry -0.0178 -0.0216 -0.0401 Panel B: UK regressions of CARS2nd hypothesis 43 (-0.832) (-0.703) (-1.524) Constant 0.128*** 0.135*** 0.155*** 0.128*** 0.135*** 0.154*** (4.288) (4.078) (3.560) (4.277) (4.066) (3.545) Observations 760 760 760 760 760 760 R-squared 0.107 0.085 0.094 0.106 0.084 0.094 Acquirer Industry Dummy YES YES YES YES YES YES Target Nation Dummy YES YES YES YES YES YES Note: The estimations from the regression 2a) and 2b) presented in section 4 are presented in this table. The dependent variables are the CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5), all variables definition can be consulted in the appendix. The variables “acqgdp”, “totalassets” and “MBV” were winsoried at 1% level of each tail, in order to remove extreme values that could bias our results. . T-statistics are in parentheses. ***, ** or * indicates that the coefficient estimates are significant at 1%, 5% or 10% level, respectively. 44 H2a) H2a) H2a) H2b) H2b) H2b) VARIABLES car(-1,+1) car(-2,+2) car(-5,+5) car(-1,+1) car(-2,+2) car(-5,+5) privatetarget -0.0040 -0.0037 -0.0245 -0.0048 -0.0050 -0.0258 (-0.474) (-0.320) (-0.988) (-0.605) (-0.458) (-1.084) cash -0.0026 -0.0010 -0.0098 -0.0026 -0.0010 -0.0098 (-0.658) (-0.176) (-1.035) (-0.656) (-0.175) (-1.035) stock 0.0009 0.1320 -0.0016 0.0017 0.1320 -0.0001 (0.00819) (0.736) (-0.00537) (0.0143) (0.738) (-0.00046) sameindustry 0.0065 0.0004 0.0158 0.0089 0.0009 0.0211 (0.912) (0.0325) (0.684) (1.088) (0.0727) (0.829) serealacq 0.00507** 0.0052 0.0100** 0.00523** 0.0052 0.0103** (1.986) (1.499) (2.097) (2.013) (1.503) (2.143) Acqgdp(t) -0.0005 -0.0019 -0.0037 -0.0005 -0.0020 -0.0037 (-0.380) (-1.094) (-1.501) (-0.396) (-1.135) (-1.508) lnacqsize(t-1) -0.0057** -0.0062** -0.0045 -0.0057** -0.0062** -0.0044 (-2.459) (-2.186) (-1.000) (-2.458) (-2.183) (-0.983) MBV(t-1) -0.0030 -0.0025 -0.0001 -0.0030 -0.0025 -0.0002 (-1.410) (-0.883) (-0.0266) (-1.417) (-0.885) (-0.0371) leverage(t-1) 0.0310 0.0441 0.0208 0.0312 0.0438 0.0212 (0.875) (0.888) (0.231) (0.879) (0.884) (0.237) multipleshocks -0.0023 -0.0044 0.0039 0.0007 -0.0071 0.0117 (-0.412) (-0.566) (0.310) (0.134) (-0.781) (0.768) multipleshocks_privatetarget -0.0029 -0.0055 -0.0046 (-0.448) (-0.677) (-0.399) multipleshocks_sameindustry -0.0102 -0.0020 -0.0226 Panel C: US regressions of CARS2nd hypothesis 45 (-1.151) (-0.168) (-1.353) Constant 0.0781*** 0.0714*** -0.0095 0.0824*** 0.0754*** -0.0012 (4.914) (3.276) (-0.281) (5.170) (3.515) (-0.0360) Observations 2,112 2,112 2,112 2,112 2,112 2,112 R-squared 0.037 0.086 0.095 0.038 0.086 0.095 Acquirer Industry Dummy YES YES YES YES YES YES Target Nation Dummy YES YES YES YES YES YES Note: The estimations from the regression 2a) and 2b) presented in section 4 are presented in this table. The dependent variables are the CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5), all variables definition can be consulted in the appendix. The variables “acqgdp”, “totalassets” and “MBV” were winsoried at 1% level of each tail, in order to remove extreme values that could bias our results. . T-statistics are in parentheses. ***, ** or * indicates that the coefficient estimates are significant at 1%, 5% or 10% level, respectively. 46 6.3 Hypothesis 2a) Moving now to the hypothesis 2a), it was not found any significant coefficients for the variable “multipleshocks”, that has a positive impact on all the CARS for the Eurozone sample, and for the UK CAR(-1,+1) and the US CAR(-5,+5). The interaction between multipleshocks and privatetarget, variable “multipleshocks_privatetarget”, is only significant for Eurozone CAR(-2,+2) and UK CAR(-1,+1) and CAR(-2,+2), at 1% of significant level for the first two, and at 5% for the last one. On average, a deal that involves a private target when there were at least two currency shocks in the last four quarters, has a negative impact of 4.63 percentage points for the Eurozone CAR(-2,+2) and a positive impact of 3.77 and 5.71 percentage points for the UK CAR(-1,+1) and CAR(-2,+2), ceteris paribus. The remaining insignificant coefficients for the CARS are negative, except for the UK CAR(-5,+5), which means that this variable is always positive for our CARs in the UK sample. Due to the lack of significant values this hypothesis cannot be confirmed. 6.4 Hypothesis 2b) Finally, in our last hypothesis 2b), we can observe that the variable “multipleshocks”, is only significant for the UK CARs, significant at 5% level. When there were at least 2 currency shocks in the 4 quarters before the deal, on average, there is a positive impact of 4.23,4.57 and 5.35 percentage points, for the UK CAR(-1,+1), CAR(-2,+2) and CAR(-5,+5), respectively, ceteris paribus. At last, the interaction between multipleshocks and sameindustry, variable “multipleshocks_sameindustry”, is only significant for the Eurozone CAR(-5,+5) at a 5% level of significance. The coefficient of 0.0475 means that, on average, when there were at least 2 currency shocks in the 4 quarters before the deal that involves an acquirer and a target that play in the same industry, there is, on average, a positive impact of 4.75 percentage points on the Eurozone CAR(-5,+5), 47 ceteris paribus. The remaining insignificant “multipleshocks_sameindustry” have a positive impact for the Eurozone CARs and a negative impact for all the UK and the US CARs. As there are no significant UK CARs for the variable “multipleshocks_sameindustry” we cannot confirm that the positive effects on bidder’s CARS for the UK displayed in the variable “multipleshocks” are enlarged. Then, I am not able to confirm this hypothesis. 54 Aftershock_privatetarget Dummy variable, which is equal to 1 for private targets four quarters after a currency shock ´ Aftershock_sameindustry Source: Datastream/SDC Platinum Dummy variable, which is equal to 1 for acquirers and targets that play in the same industry (classified by the 2-digit SIC-codes) four quarters after a currency shock Source: Datastream/SDC Platinum Multipleshocks Dummy variable, which is equal to 1 if in the last four quarters there were at least two currency shocks Source: Datastream Multipleshocks_private target Dummy variable, which is equal to 1 for private targets if in the last four quarters there were at least two currency shocks Source: Datastream/SDC Platinum Multipleshocks_sameindustry Dummy variable, which is equal to 1 for acquirers and targets that play in the same industry (classified by the 2-digit SIC-codes) if there were at least 2 currency shocks in the last 4 quarters Acqindustry Source: Datastream/SDC Platinum Dummies for acquirer industries (classified by the 2-digit SICcodes) Source: SDC Platinum Targetnation Dummies for the target’s Nation Source: SDC Platinum