The Long Run Performance of Mergers and Acquisitions
Euro Area Acquirers from 1997-2005 tested with Calendar Time Models.
University of Groningen Faculty of Economics and Business MSc Business Administration: Finance Profile: Corporate Financial Management
Abstract
This study examines the long term performance of mergers and acquisitions (M&A) for an Euro area acquiring firm during the time period 1997-2005. We look into four different geographical locations in where the target firm is located. To test the long term performance in each geographical location we look at the abnormal returns using the Fama and French three factor model and the mean calendar time abnormal return (MCTAR) method. To create monthly acquisitive portfolios we will make use of a Weighted Least Squares (WLS) method, where the weights are taken as the square root of the amount of firms per calendar month and the Ordinary Least Squares (OLS) method where only the months with a minimum of ten firms are considered. Domestic Euro area acquirers show insignificant positive abnormal returns while cross-border Euro area acquirer show insignificant negative abnormal returns. The inter-continental and intra-continental Euro area acquirers show a insignificant positive abnormal returns. The method of payment is adding value for a domestic and cross border Euro area acquirer except when shares are used in a cross-border M&A transaction. The inter-continental acquisitions within the Euro area all show significant positive abnormal returns while intra-continental Euro area acquirers show that the M&A activity is destroying value in the long run.
Keywords: Mergers and Acquisitions, domestic and cross border, inter-continental and
intra-continental, Calendar time study, Fama and French three factor model.
TABLE OF CONTENTS
1.Introduction………...5
2.Empirical literature..………....6
2.1 Long term studies………...6
2.2 Long term performance of Domestic Acquisitions………7
2.3 Long term performance of Cross-border acquisitions ………...8
2.4 Hypotheses – Geographical diversification……….9
2.5 Hypotheses - Method of payment………...10
3. Methodology………...12
3.1 Improved methods for long term abnormal return………..12
3.2 Calendar time Studies……….………..……..15
3.3 Fama and French three factor model………...16
3.4 Empirical evidence Fama and French three factor model………..18
3.5 Mean monthly calendar time abnormal returns………...……...…19
4. Data………...20
4.1 Summary statistics domestic acquisitions………..21
4.2 Summary statistics cross-border acquisitions………...….21
4.3 Summary statistics of inter-continental acquirer……….……..22
4.4 Summary statistics of intra-continental acquirer………...……23
5. Results ……….………....23
5.1 Results of Domestic acquisitions………..…..23
5.2 Results of Cross-border acquisitions………...24
5.3 Results of Inter-continental acquisitions………...24
5.4 Results of Intra-continental acquisitions………...25
5.5 Method of Payment - domestic / cross-border………25
5.6Method of Payment – intra-continental / intra-continental……….26
5.7 Robustness check………...….27
6. Conclusions………..……28
7. References………. .30
Overview of Tables
2.1 Studies of long term performance domestic acquisitions 2.2 Studies of long term performance cross-border acquisitions 3.1 Construction of the acquisitive portfolio
3.2 Major indices used for Fama and French three factor model
3.3 Overview of studies of domestic Mergers and Acquisitions on a Fama and French three factor model
4.1 Summary statistics Domestic acquirer 4.2 Summary statistics Cross-border acquirer 4.3 Summary statistics Inter-continental acquirer 4.4 Summary statistics Intra-continental acquirer
5.1 Domestic acquisitions on a Fama and French three factor model 5.2 Cross-border acquisitions on a Fama and French three factor model 5.3 Inter-continental acquisitions on a Fama and French three factor model 5.4 Intra-continental acquisitions on a Fama and French three factor model 5.5 Method of payment domestic and cross-border acquisitions
5.6 Method of payment inter- and intercontinental acquisitions
1. Introduction
The ongoing debate on the abnormal returns of mergers and acquisition (M&A) activities has generated a large amount of empirical evidence over the years. The results of these long term performance studies of M&A’s have been mixed. Moreover, the results vary between the types of acquisitions. Research on domestic acquisitions, for example Agrawal, Jaffe and Mandelker (1992), Asquith (1983), Magenheim and Muller (1988), Loderer and Martin (1992) , Gregory (1997), Loughran, Vijh (1997) and Rau and Vermaelen (1998) are among others who found significant negative long term abnormal returns. Loderer and Martin (1992), however, found significant positive abnormal returns for domestic acquisitions. For cross-border acquisitions this is not different. The studies of Andrè, Kooli and L’Her (2004), Kang (1993), Black, Carnes and Jandik (2001), Gugler, Mueller, Yurtoglu and Zulehner (2003) found a significant negative long term abnormal. Conn, Cosh, Guest and Hughes (2001) do not find any evidence of negative long term post acquisition performance for cross-border acquisitions.
It remains an interesting question why firms look for opportunities for growth outside their home country, especially since empirical evidence has shown mixed results on the long term performance of these M&A’s. Despite these mixed results the volume of cross-border M&A activity in Europe increased significantly in the latter part of the nineties. After nearly doubling in 1997 and 1998, the volume of European M&A peaked in the year 1999. European merger activity significantly declined over the next two years to a total of 532 billion dollars in 2001. According to Moeller and Schlingeman (2005) the explanation for this cross-border growth is that product and capital markets are more integrated and new markets emerge, as a result these global markets have become an important strategic issue for companies. The integration of European countries in the European Union (EU) has triggered firms to invest more cross-border. Also the establishment of the European Monetary Union and the introduction of the Euro had a positive influence on cross-border activity in the European Union. The European Union has developed a single market through a standardized system of laws which apply in all member states, guaranteeing the freedom of movement of people, goods, services and capital.
These new circumstances, i.e. the unity of one currency and opportunities to infiltrate new markets and gains from firm specific assets, leads us to test if these different geographic locations where the M&A activities took place have been profitable for a Euro area acquirer. It is interesting to test long term abnormal return because cross-border, inter-continental and intra-continental M&A’s are more complex than domestic acquisitions due to differences in political and economic environment, corporate organization, culture, tradition, tax and accounting rules.
This study will intensively test the long run performance of M&A activity of domestic, cross-border, inter-continental and intra-continental acquiring firms in the Euro area in the period of 1997-2005. There are a total of 160 domestic acquirers, 355 cross-border acquirers, 188 inter-continental acquirers and 116 intra-continental acquirers. The main question in this paper is whether the M&A performed by Euro acquirers are value destroying activities in the long run. We will also test whether the method of payment has any influence on the abnormal return in each geographic location. In a perfect market without uncertainty it seems that investors are indifferent between the various methods of payment, because there will be no differences in their wealth outcomes.
robustness check to see if the data would fit another model better. In order to construct the Fama and French three factor model we will follow the articles from Mitchell and Stafford (2000) and Andre, Kooli and L’Her (2004). For the mean calendar time abnormal return model as a robustness check on the inter- and intra-continental M&A’s. We will follow the articles from Andre, Kooli, and L’Her (2004) and Lyon, Barber and Tsai (1999) for the construction of this model.
This paper is organized as follows. In the second section we will give an overview of the previous literature that was used for testing the long term abnormal return in M&A activities. We will show the reasons why the best fit for our research is the calendar time Fama and French three factor model. This section also shows previous research on the geographical diversification and the method of payment used in M&A’s. In the third section we will explain the methodology used for the calendar time Fama and French three factor model and the mean calendar time abnormal return method. Section 4 shows the data and descriptive statistics we used. In section 5 the results from the regression analysis are shown. Also in this section a robustness check is performed on the inter- and intra-continental sub-sample. In section 6 we will draw conclusions.
2. Empirical literature
In a long term study the profitability of stock returns is in general examined for 12, 24 or 60 months. Abnormal returns are in general calculated as the difference between the actual returns and the expected returns. The expected returns can either be calculated with the asset pricing model like the Capital Asset Pricing Model (CAPM) or a benchmark portfolio of firms with the same characteristics of the acquiring firm but with no M&A activity. This way the average of all abnormal returns of a firm that had an M&A activity before, during and after the announcement of an M&A can be empirically tested. In these studies conclusions can be drawn of how the stock price of acquiring as well as target firms on average behave before, during and after the announcement of a merger. The long term performance studies are explained in section 2.1.
2.1 Long Term Studies
Most long term studies focus on domestic deals because M&A history has showed us that most deals are performed domestically. However, the last decade there has been an increase in both the number and value of cross-border deals. But what are the advantages and disadvantages of performing a cross-border M&A. Empirical literature has shown mixed results when looking at the results of cross-border M&A deals. Why do firms look outside of their ‘own’ market if empirical evidence has shown that negative long run returns can be obtained. In this section we will discuss the empirical literature on domestic and cross-border M&A’s.
structure flexibility can be created by switching the production facilities to countries where the marginal costs are the lowest. The third way in which a cross border deal can create firm value by geographical diversification is that geographical diversified companies will be more valuable because of their ability to arbitrage institutional restrictions such as tax codes and financial restrictions (Bodnar et al 1997). An example is where a geographically diversified company can transfer profits/losses to countries where the tax advantage is the lowest. Also a geographically diversified company has the option to pick the low cost location to raise capital, the low cost location to declare profits and in which market to concentrate market power.
Along with Bodnar et al (1997), Aw and Chatterjee (2004) find that cross bidder returns are higher if the bidder can apply one’s country superior management techniques. Besides the reasons stated by Bodnar (1997) the main reason to invest in a foreign company is market access. In order to avoid trade barriers cross-border M&A’s may be motivated by a need to operate locally (Danbolt 2004). With the passing of the Single European act in 1995 and the introduction of the single European market in 1992 the number of cross border M&A did rise to new heights. Investors are willing to pay a higher premium for geographically diversified companies. This premium will increase the value of the geographically diversified companies relative to that of domestic companies (Bodnar et al 1997).
These positive circumstances for cross-border M&A, i.e., economies of scale, positive market conditions, favorable arbitrage conditions, superior management techniques from acquiring country, and market access should all lead to a positive value creation for cross-border M&A in the long run. If the characteristics mentioned above are unique to a geographically diversified company and cannot be otherwise acquired by investors than the value of a geographically diversified company should be higher then a pure domestic company. These earlier positive influences on the value creation of geographical diversified companies suggest that they are more valuable than pure domestic companies. Next we will look at the empirical literature on the long run abnormal return on domestic acquisitions.
2.2 Long term performance of Domestic Acquisitions
complete stock mergers earn significantly negative excess returns of -25 % whereas firms that complete cash tender offers earn significantly positive excess returns of 61.7 %. Rau and Vermaelen (1998) examined the long run performance of 3169 mergers and 348 tender offers with acquirers listed on the NYSE/AMEX during the period between January 1980 and December 1991. They find that acquirers in mergers under perform in the three years after the acquisition.
Another example of a long term study of domestic deals is the article of Agrawal and Jaffe (2000). They give a review on long term performance of M&A’s until 2000. They state that in the eighties of the last century there seemed to be negative abnormal returns in the years after M&A’s. In the article of Agrawal and Jaffe (2000) the long run performance of M&A’s can be divided into two periods : The period before and the period after the article of Frank, Harris and Titman (1991). The article of Frank, Harris and Titman (1991) was a major breakthrough in the research on long term performance of M&A’s. The articles that appeared before the article of Frank, Harris and Titman (1991) are all event studies using the CAPM model to calculate expected returns. Agrawal and Jaffe (2000) conclude that there seems no strong evidence of significant negative abnormal returns. The results from the studies of Frank, Harris and Titman (1991) only imply that the efficient market hypothesis does not hold and an anomaly exists. Agrawal and Jaffe (2000) found that new methodologies can explain the long term abnormal returns in M&A’s better and conclude that the methods of Frank, Harris and Titman (1991) are outdated. Jensen and Rubach (1983) observed in their article that the observed negative abnormal returns in the long run seem to violate the efficient market hypothesis. If a M&A activity is underperforming in relation to a benchmark over the long run, one can say that the asset price return at the announcement of the merger and acquisitions are overestimated and not efficient.
To sum up these previously discussed papers we can conclude that the long run cumulative abnormal returns for domestic acquisitions is negative for different sample sizes, sample periods and event windows. Next we will go into the empirical literature on the long run abnormal return on cross-border M&A’s. An overview of these previous papers is shown in table 2.1
Table 2.1
Studies of long term performance domestic acquisitions
This table shows an overview of studies on the long term performance of domestic acquisitions.
Study Sample Event Cumulative
size window abnormal returns
Agrawal, Jaffe Mandelker (1992) 765 -0,125 -10.26%
Gregory (1997) 452 -0,5 0,8
Loughran, Vijh (1997) 434 -1,125 -14.2 %
Rau, Vermaelen (1998) 3968 (0,36 months) -4%
Domestic acquistions Sample period 1955-1987 1984-1992 1970-1989 1980-1991
2.3 Long term performance of Cross-border acquisitions
run performance of acquiring Canadian firms. In their research they found that the intercepts for cross-border and domestic M&A’s are negative. The intercepts achieve significance only for cross-border transactions. The difference in the abnormal returns between cross-border and domestic portfolios is negative and significant at the 5% level. They concluded that cross- border M&A performed worse then do domestic M&A’s.
In line with Andrè, Kooli and L’Her (2004) are Gugler, Mueller, Yurtoglu and Zulehner (2003) find that cross-border acquisitions result in a significant decrease in the market value of the acquiring firm over the five post acquisitions years. They research 429 transactions on cross-border mergers from United States, United Kingdom, Continental Europe, Japan and Australia/New Zealand/Canada. Cross-border acquisitions of UK firms did not generate significantly larger changes in sales and profits than was true for other cross-border acquisitions, and the same was true for all other countries. Although individual mergers can have quite different consequences in terms of efficiency and market power, their effect do not appear to depend on the country of origin of the merging firms.
In contrast to previous results is the study of Kang (1993) shows that in acquisitions conducted by Japanese companies in the United States, there have been significant benefits for the shareholders of both companies but they also found a negative long term abnormal return.
To sum up these previous studies, we can conclude that most studies on the long term abnormal return of cross-border M&A show a negative impact across time periods and research methods. Results are shown in table 2.2 Next we move into section 2.4 where we show our formed hypotheses to be tested.
Table 2.2
Studies of long term performance cross-border acquisitions
This table shows an overview of studies on the long term performance of cross-border acquisitions. The cumulative abnormal return is calculated as the return of the acquisitive firms minus the risk free rate in the same month.
Cumulative
Study Sample Event abnormal returns
size window (t-statistic)
Andre, Kooli and L'Her (2004) 267 (0,36 months) -1.146 % (3.01)
Kang (1993) 119 (0,36 months) -0.124 % (2.04)
Gugler, Mueller, Yurtoglu 429 (1-5yrs) Not significant
Cross-border acquistions Sample period 1977-1983 1975-1988 1981-1998
2.4 Hypotheses – Geographical diversification
The first hypotheses we test is if there is a significant difference in the long run abnormal return for Euro area acquirers of domestic or cross-border M&A.
Hypotheses 1 : Cross-Border acquiring firms long run abnormal returns differ significantly from domestic acquiring firms long run abnormal returns.
specific assets are capitalized in a higher company value (Bodnar et al.,1997). Also cross-border M&A’s can create value with a multinational corporate system. Geographic diversification gives the company the possibility to take advantage of market conditions. An example is cost structure flexibility, where the company can switch their production facilities to countries where the marginal costs are the lowest. Also, cross-border M&A’s can create value because it will be more valuable because of their ability to arbitrage institutional restrictions such as tax codes and financial restrictions (Bodnar et al., 1997). According to Aw and Chatterjee (2004) cross-border returns are higher if the bidder can apply one country’s superior management techniques. The downside of this search for international growth through M&A is explained by Danbolt (2004). He argues that acquiring companies who do not have a foothold in another country and see entering this market as a value creating opportunity, are willing to pay a higher premium than companies who are already active in that market. This will have a negative impact on the value creation for cross-border M&A. Also, if a market is higher integrated, the market for corporate control will be more competitive. More competition means that it is less likely that bidders earn synergetic gains so the returns will be lower (Moeller and Schlingeman, 2005). The positive circumstances for cross-border M&A, i.e., economies of scale, positive market conditions, favorable arbitrage conditions, superior management techniques from acquiring country, and market access should all lead to a positive value creation for cross-border M&A in the long run.
The second hypotheses we test is there is a difference in the long run abnormal return of Euro area acquirers of a inter-continental M&A and a intra-continental M&A.
Hypotheses 2 : Inter-continental acquiring firms long run abnormal returns differ significantly from intra-continental acquiring firms long run abnormal returns.
We expect the long run abnormal return for a intra-continental Euro area acquiring firm to be significantly lower then the inter-continental Euro area acquiring firm. The reasoning behind this is that a intra-continental M&A activity is even more complex and time consuming then a cross-border M&A due to currency difference, cultural obstacles, and legal rules. As we can see in the summary statistics there are less intra-continental M&A’s then inter-continental M&A’s. Is this because it is a very expensive and time consuming activity or is it because less value is added to an intra-continental M&A in the long run?
2.5 Hypotheses - Method of payment
an adverse selection problem. Myers and Majluf (1984) see the method of payment as an information signal. According to Rau and Vermaelen (1998) managers that are better informed about the long-term prospects of their firm than is the market, they will tend to pay for their acquisitions with shares when they believe that their stock is overvalued and use cash otherwise. Hence, the means of payment hypothesis predicts that, on average, long-run abnormal returns to bidders will be negative in share financed acquisitions and positive in cash financed acquisitions. Note that such a timing strategy will only work if the market (and especially the target shareholders) underestimates the extent of over- or undervaluation of the bidder.
The third hypotheses we will test is whether the method of payment has a significant influence on the long run abnormal return of Euro acquirers of domestic and cross-border firms.
Hypotheses 3 : Domestic (cross-border) acquisitions abnormal returns where the method of payment is cash do not differ significantly from domestic (cross-border) acquisitions where the method of payment is shares.
Cross-border M&A’s are more difficult to value accurately because of assymetric information. Hansen (1987) and Moeller and Schlingemann (2005) state that acquiring firms will finance the deal with shares if the asymmetric information about targets is high. Although if the acquiring firm possesses private information concerning the value of their equity, the bidder has to deal with an adverse selection problem; the equity of the bidder may be overvalued. For an acquiring firm it is profitable to finance a merger with overvalued equity. According to Myers and Majluf (1984) the method of payment acts as an information signal. Bidder shareholders see a merger paid with shares as a signal that the equity is overvalued. This signal influences the bidder returns in a negative way. Bidder shareholders see a merger or acquisition paid with cash as a positive signal and influences bidder returns in a positive way. In cross-border M&A’s the target is frequently unwilling to accept foreign equity, which forces the bidder to pay with cash (Gaughan, 2002). Because of these differences in the method of payment we will test if this variable has any significant influence on the abnormal return of an Euro area acquiring firm.
The fourth hypotheses looks whether the method of payment has a significant effect on the long run abnormal return for Euro area acquirers of inter- and intra-continental firms.
Hypotheses 4 : Inter-(intra-) continental acquisitions abnormal returns where the method of payment is cash differ significantly from inter- (intra-) continental acquisitions where the method of payment is shares.
We expect the long run abnormal return for an inter-continental Euro area acquiring firm paid with cash to be significantly higher then the inter-continental Euro area acquirer paid with shares. In cross-border and even more with intra-continental M&A’s the target is frequently unwilling to accept foreign equity, which can force the acquiring firm to pay with cash. This can have an influence on the positive signalling effect when paying with cash and will be non existent in cross border and intra-continental M&A transactions. We also expect the long run abnormal return of a intra-continental acquisition paid with cash to have a higher long term abnormal return then intra-continental acquisitions paid with shares.
3. Methodology
3.1 Improved methods for long term abnormal return
The methodology of the studies on the calculation of the long term abnormal returns on M&A’s have improved over time. There is still much debate on which is the best method to use. In the following section we show how the methods have evolved and improved over time. The first method we discuss are the size controlled event studies. The paper of Frank, Harris and Titman (1991) can be seen as the first study that focuses on the long term performance of M&A’s. In their article, they use the returns of firms with the same characteristics, but without these events as benchmarks to calculate expected returns. This later portfolio is called the reference portfolio. They construct four different benchmarks, but claim that an eight-portfolio is most important. This reference portfolio is made up of four portfolios based on firm size, three based on dividend yield and one based on past returns. The article of Frank, Harris and Titman (1991) is the first article that controls for firms size, next to the firm’s beta to calculate expected returns. The authors do not find any significant abnormal returns. They conclude that the efficient market hypotheses should not be rejected and that there is no mispricings at the announcement dates of M&A’s. Further, they conclude that earlier found significant abnormal returns exist because of the results of errors in the benchmark chosen.
Agrawal, Jaffe and Mandelker (1992) calculate abnormal returns in the five years after acquisitions and tender offers, where they also control for firm size and beta. Where the study of Frank, Harris and Titman found no significant abnormal returns the study of Agrawal, Jaffe and Mandelker (1992) found a significant underperformance of -10% up to five years after the researched data set. They state that the difference in results is due to the fact that they researched different time periods. The study of Loderer and Martin (1992) and Kennedy and Limmack (1996) are other studies that control for size in looking at the abnormal return in the long run performance of mergers M&A’s. Loderer and Martin (1992) find that the abnormal returns are significantly negative in the period of 1966-1969 but in the later periods are insignificantly different from zero. Kennedy and Limmack (1996) find insignificantly underperformance of mergers in the period 1980-1989 in the United Kingdom up to twelve months after the announcement dates.
negative abnormal return across different datasets and time periods. Within these size controlled event studies, there was a lot of space for improvement. This improvement came with the book to market and size controlled event studies, which will be explained in the next section.
According to the research of Fama and French (1992,1993), expected returns should be controlled for book to market ratios, size and firm beta and not only firm size. In the 1960s Sharpe introduced the Capital Asset Pricing Model (CAPM). In this model the expected excess returns are calculated by taking the excess market returns and multiplying them with a beta factor of the analyzed portfolio. The beta of the portfolio is a measure of the risk of the portfolio and is measured as the covariance of returns of that portfolio with the returns of the market portfolio divided by the variance of the stock returns of the market portfolio.
Fama and French (1992, 1993) studied the returns of a very long history of American stocks. They concluded that the CAPM model of Sharpe is not an adequate model to calculate excess returns. They came to the conclusion that other factors had to be taken into account in combination with beta and showed with their model that the factors book to market and market capitalization together with the beta factor explain stock returns much better. In the theory of Fama and French a higher return should imply higher risk. A volatile stock has a high percentage of a very high or a very low return and thus has a high risk factor.
Other studies that have taking book to market into account to calculate abnormal returns to a benchmark portfolio return include Anderson and Mandelker (1993) and Gregory (1997). In their study the returns of the appropriate size and book to market control portfolios are subtracted from the returns of the acquiring firms up to five years after the acquisitions. Anderson and Mandelker (1993) find significant negative abnormal return over five years. Gregory (1997) researches the stock market in the UK market. He calculates abnormal return with six different models, including CAPM and the Fama and French three factor model. Gregory (1997) reports a significant underperformance of -12% to -18% of five year post merger returns with the Fama and French three factor benchmark. The conclusions that are made by the authors from these earlier studies on long term abnormal return are debatable for the following reasons.
Barber and Lyon (1997) and Kothari and Warner (1997) state that the measurement in the long run abnormal return event studies are biased and unreliable. The reference portfolio contain only firms that will survive during the period analyzed therefore the benchmark will be tilted toward a survivorship biased portfolio. There are some issues with this methodology though. For example there is a so-called rebalancing bias. This bias stems from the fact that the portfolio being analyzed is not rebalanced monthly while the reference portfolio is monthly rebalanced. Lyon, Barber and Tsai (1999) recommend to use either and approach based on a traditional event study but with carefully constructed monthly balanced reference portfolios together with taking other t-test into account or to use and approach in calendar time instead of event time.
The studies that took book to market and firm size into account were an improvement from the studies that only took firm size into account. The book to market and size controlled studies found negative long term abnormal returns in line with the size controlled studies. An improvement on the size and book to market controlled event studies was the buy and hold methodology which will be explained in the next section.
on an investment in a portfolio under test can be directly compared with the returns of an investment in a reference portfolio without these events. This method controls for the survivor and rebalancing bias. Lyon, Barber an Tsai (1999) state that he buy and hold methodology solves the survivor, rebalancing and skewness bias. However, according to Mitchel and Stafford (2000) there still remains a discrepancy. The sample is cross-sectionally dependent and therefore not random; the buy and hold methodology will give biased results. This is constructed by keeping an unadjusted reference portfolio through the whole time period and not include new listed firms in the reference portfolio. Concerns arising from the skewness of individual firm long horizon abnormal returns hampered statistical inference in many initial studies.
Ikenberry and Lakonishok (1995) introduce a bootstrapping procedure for statistical inference that stimulates an empirical null distribution of the estimator, relaxing the assumption of normality. Brav and Gompers (1997) state that the pervasiveness of underperformance may be misleading because the returns may be correlated in calendar time. When an economy is in a bear market for a whole year then all the share prices will show underperformance for the next three or five years while it is just a single year of a downward market. According to Fama (1998) the buy-and-hold abnormal return is not the correct measure. The systematic errors that arise with imperfect expected return proxies – the bad model problem – are compounded with long horizon returns.
According to Mitchell and Stafford (2000) there are various problems with the buy and hold methodology approach. The first is that when the sample is cross-sectionally dependent and therefore not random, the buy and hold methodology will give biased results. If a shock to the economy has the result of the whole economy is in a bear market then the underperformance of those firms in that particular month will be misleading as they will be in the acquisitive portfolio for 36 months. Mitchell and Stafford (2000) show that M&A activity is clustered in time and also in industry and thus returns are cross-sectionally dependent. According to Mitchell and Stafford (2000) there exist more problems with the buy and hold methodology. There exist a bad model problem in that it is difficult to make a correct expected asset pricing model to calculate abnormal returns. The results of the abnormal returns may be a due to a combination of the underlying model to calculate expected returns and the abnormal returns from an event. This is known as the joint hypotheses problem.
To sum up, the buy and hold methodology shows a negative long term abnormal return across various time periods. The size controlled event study, book to market and size controlled event study as well as the buy and hold method all show a negative long term abnormal return in the long run. An improvement for this method is the Fama and French three factor model, which we will use for our study.
3.2 Calendar time Studies
Calendar time series are different from event time studies in that the focus point are the calendar months in which the M&A activity took place. In each calendar month the sample of firms is scanned on firms that have announced an M&A activity within a last predefined period prior to that calendar month. These firms will be defined as acquisitive firms. The abnormal returns of these firms is calculated on a monthly basis. This acquisitive portfolio can be made equally weighted (EW) or value weighted (VW). In the equally weighted method, all the firms in the acquisitive portfolio that had an M&A activity are equally weighted while in the value weighted method they are value weighted with respect to their firm size.Like Mitchell and Stafford (2000), Brav and Gompers (1997) and Fama (1998), we think the calendar time portfolio procedure has more power to identify reliable evidence of abnormal performance than the buy-and-hold abnormal return methodology.
For our research we use the article of André, Kooli and L’Her (2004) and Mitchell and Stafford (2000). We test the equally and value weighted abnormal return of the acquisitive portfolio on an ordinary least squares regression (OLS) as well as a weighted least squares regression (WLS) for the following reasons. Loughran and Ritter (2000) state that there is a potential problem for heteroscedacticity due to the fact that the variance of the residuals of the regressions is dependent on the amount of firms per calendar month. This may lead an ordinary least regression to be inefficient. Loughran and Ritter (2000) argue that a weighed least regression mitigates this problem. The calendar months are in this case weighted by the square root of the number of firms in the acquisitive portfolio in that month. The regression will then be effectively transformed such that all observations will be equally weighted, both dependent and independent variables in the regression are multiplied by the square root of the amount of acquisitive firms. The weighted least squares regression will weigh months with more activity (hot markets) heavier then months with low activity (cold markets). If abnormal return are more present in hot markets than in cold markets the ordinary least squares method will average over these and abnormal returns be hard to detect. Andrè, Kooli and L’ Her (2004) prefer to use the weighted least squares method for these reasons mentioned above.
There are also advocates of the weighted least squares method. Mitchell and Stafford (2000) state that the use of weighted least squares methods contradicts with the purpose to use the calendar time regression which is the fact that firms are cross related in time. They show that taking a minimum of ten firms per month into account in a calendar time portfolio should be enough to take care of the possible heteroscedastic behavior of the residuals. Mitchell and Stafford (2000) also find that there is no evidence that abnormal returns are hard to detect by using ordinary least squared (OLS) method. We will test the WLS as well as the OLS to get the best overall view of the long term abnormal return.
3.3 Fama and French three factor model
The calendar time Fama-French three-factor model is given by the following expression:
)
SMB
, , 0 , ,t−
F t=
+
t+
1 t+
2(
m t−
F t+
Σ
pR
HML
R
R
R
α
β
β
β
The dependent variable (Rp,t−RF,t) is the monthly excess return of the acquisitive portfolio over the Risk Free rate (RF) for a given calendar month (t). For the risk free asset we use the return on the one-month euro mark deposit quoted in London and will be downloaded from Datastream. The betas β0, β1 and β2 are the loadings of the portfolio on
each factor. The
α
is the intercept and is a measure of the average monthly abnormal returns of the M&A sample of Euro area acquiring firms, which is zero under the null of no abnormal performance. If the Fama and French model provides complete description of expected returns, then the intercept represents the combined effects of mispricing and model misspecification. The dependent variable in the regression is a portfolio consisting of firms that had an M&A activity in a preceding defined period. This portfolio will be called the acquisitive portfolio. To form the acquisitive portfolio, for each month from January 1997 till December 2005, the average of returns of the firms that announced an M&A activity with up to three years after the month they made an acquisition is calculated. We refresh formed portfolios each month as new M&A might have occurred in that particular month and other M&A that have been part of a portfolio for 36 months are excluded. For example, the firm ‘Casino Guichard ‘ from France announced an M&A activity on September 3 1997 and will be in the acquisitive portfolio until August 2000. In table 3.2 we report a extract of a part of an acquisitive portfolio used in our research.Table 3.1
Construction of the acquisitive portfolio
This table shows an overview of the construction of the acquisitive portfolios The portfolios are either equally or value weighted each calendar month for the entire period (Jan 1997-Dec 2005). A firm stays in the acquisitive portfolio for a period of 36 months. Each month the average return is calculated with equally or value weighted multipliers for the size of the firms.
Announcement 1997 1998
Acquiror name date September October November December January February
Casino Guichard 3-9-1997 19,50% -12,70% 3,90% 1,50% 2,40% 9,10%
Bank of Ireland 6-11-1997 16,70% 11,50% 10,10% 26,20%
Banco Santander 20-2-1998 31,20%
Average 19,50% -12,70% 10,30% 6,50% 6,25% 22,17%
Monthly return
The mimicking portfolios for our research dataset needs to contain as many representative stocks for the Euro area. This way we will get the best risk premiums for size, book to market and market excess returns. We had to construct our own portfolios the same way Fama and French constructed their portfolios on the NYSE, AMEX and NASDAQ to get the best risk premium for size, book to market and excess market return for American stocks. Their data library contains the universe on stock returns for the United States of America from 1927 until today. In the next section their methodology is explained and how we constructed our own risk premiums for size, book to market and market excess returns for Euro area stocks.
Fama and French (1992, 1993) construct the SMB and HML portfolios as follows. In June of each year from 1963 until 1991, all NYSE, AMEX and NASDAQ firms are ranked on size. The median size is used to split the dataset in a NYSE, AMEX and NASDAQ stocks into two groups. They call these the BIG (B) and SMALL (S). Since the introduction of the CAPM by Sharpe (1964) and the extension by Litner (1965) asset pricing models have been an intensive topic of research. The extension of the CAPM was suggested by Fama and French (1992, 1993, 1995, 1997). In their research they include two zero-cost portfolios : A Small minus Big (SMB) portfolio based on the market capitalization of the firm considered and a High-minus-Low (HML) portfolio which is based on the book-to-market value of the stock. In a later study, Fama and French (1998) provide the international evidence by investigating the model for a number of countries. Fama (1998) specifically recommends the construction of monthly portfolios in calendar time for measuring the average abnormal long run performance. There are three reasons why Fama (1998) prefers the calendar time approach over the event time approach for measuring abnormal performance. First, monthly returns are less subject to the bad model problem. Second, monthly portfolios allow the researcher to examine the cross-correlation between firms in the sample. Third, the portfolio returns allow better statistical inferences.
Then the two groups are also split into three groups based on the book to market ratios. The breakpoints for these groups are the bottom 30 % (LOW), middle 40 % (MEDIUM) and the top 30% (HIGH) for the year t-1book to market ratios. Now six portfolios are formed on the size and book to market ratios. The B/H portfolio contains the firms in the BIG group for size and in the HIGH group book to market ratio The B/M portfolio contains the firms in the BIG Group for size and in the MEDIUM group for market to book ratio. The B/L portfolio contains the firms in the BIG group for size and in the LOW group for book to market ratio. The SMALL portfolios are formed the same way. Monthly value weighted returns are calculated for the from July of year t to June of year t+1. The portfolios are reformed every year at year t+1. This is done to assure that the financial ratios are known from the year before. The mimicking portfolio SMB is the difference per month between the average of the three portfolios with the SMALL size group (SH, SM, SL) and the average of the portfolios with the BIG group (BH, BM, BL). The mimicking portfolios of the HML group is the difference between the average of the returns between the between the two portfolios with high book to market ratios ( SH, BH) and the portfolio with the low book to market ratios (SL, BL).
part of the large universe of Euro area stocks. The median size is used to split these stocks into the BIG (B) and SMALL (S) groups. Then we used the book to market ratios to split these 2 groups into 6 portfolios. Monthly value weighed returns are calculated for each group (BH, BM, BL) and (SH, SM, SL).
Table 3.2
Major indices used for Fama and French three factor model
This table shows an overview of the major indices of the Euro area countries. These indices were used to construct the Fama and French portfolios to get best risk premiums for excess market return, size and book to market ratios. All firms from these indices were included in the loadings of the Fama and French three factor model. The return, market to book and size of these firms were used to construct the Fama and French three factor model.
Country Austria Belgium Finland France Germany Spain
Stock index Austrian Traded Index Euronext Brussels Helsinki Stock Exchange Euronext Paris Frankfurt Stock Exchange Madrid Stock Exchange
Country Greece Ireland Italy Netherlands Portugal
Stock index Athens Exchange Ireland Stock Exchange Italian Stock exchange Euronext Amsterdam Portuguese Stock Index
3.4 Empirical evidence Fama and French three factor model
In the article of Mitchell and Stafford (2000) they find a small underperformance of M&A with their calendar time study in the United States for stocks from 1961 until 1993. They found a significant negative abnormal return for an equally weighted and an R-squared of 0.97. For a value weighted portfolio they found an insignificant underperformance and a R-squared of 0.95. They conclude that with taking the bad model problem into account they find no evidence of underperformance for a domestic acquiring firm. Gregory and Matako (2004) researched the long run performance of domestic takeovers in the UK. Gregory and Matako (2004) found a significant negative average monthly abnormal return of -0.30 % in five years of an equally weighted portfolio of acquiring firms. They found a R-squared of 0.90 for their test which was reliable to draw conclusions for this model. They used a bootstrapped t-test for event studies but also a calendar time Fama and French three factor model. Andrè, Kooli and L’Her (2004) research the long run performance of Canadian acquiring firms. In their research the find for an equally weighted portfolio of Canadian acquirers a significant monthly average underperformance of -0.8 % ( t = -2.77) in the three years after the acquisition. They found a R-squared of 0.62. For the value weighted portfolio they found a significant negative abnormal return of -0.8 % . For this model they found a R-squared of 0.70. They concluded that the acquiring firms in Canada are significantly underperforming in the long run. An overview of these studies is shown in table 3.3.
Table 3.3
Overview of studies of domestic M&A’s on a Fama and French three factor model
This table shows an overview of the long run performance of domestic M&A's on a Fama and French French Three Factor model.
Study Equally Value Sample Event
Weighted Weighted Period Window
Domestic Acquirers
Mitchell and Stafford (2000) -0.20% -0.03% 1961-1993 (0,60 months) Gregory and Matatko (2004) -0.75% -0.84% 1984-2000 (0,36 months) Andre, Kooli and L'Her (2004) -0.30% -0.25% 1977-1983 (0,36 months)
3.5 Mean monthly calendar time abnormal returns
A variation of the Fama and French calendar time model is the Mean Calendar Time Abnormal Return model (MCTAR). This method is different then Fama and French in that the performance of acquisitive firms is not compared to the expected returns of the Capital Asset Pricing model (CAPM) but compared to a reference portfolio. These reference portfolio consist of firms that did not have a M&A activity in the last three years and are of comparable size and book to market ratios. In our paper we formed our own reference portfolios. These portfolios consisted of all firms from our thirteen indices that did not have an M&A activity in the years of research. This way we present the best method to create ‘expected’ returns for firms that did not have an M&A activity. We used this method because our data on inter- and intra-continental acquisitions did not show a good fit for the Fama and French three factor model. With this method the return from the acquisitive firm is directly compared to the average return of the non acquisitive firm in the same book to market ratio and size group.
For each calendar month, we calculate the abnormal return as the difference between return for each security (Ri,t) and the return on the corresponding reference portfolios. The return data from this firm in that month is subtracted form the return in the reference group to calculate the abnormal returns. Each acquisitive firm is put in line with on of the nine reference portfolios, corresponding to book to market ratio and size. For the construction of the reference portfolio the methodology of Barber and Lyon (1997) and Kothari and Warner (1997) will be used to form the reference portfolios. We formed the non acquisitive portfolio as follows. All the publicly traded firms from the 13 Euro area indices that did not have an acquisition in the calendar year are included in the reference portfolio. All non acquisitive firms that are in the top 30% of the highest market capitalization of all firms in the sample will be grouped. The value weighted returns of this group will give the average return of this highest market capitalization group and will be called SB as in the article of Barber and Lyon (1997). The same is done for the middle 40% and the last 30% which we call SM and SS groups respectively. The size and book to market ratios groups are formed with end of the year data. The same is done for the book to market ratios. All non acquisitive firms with the highest 30% of book to market ratio will be grouped and named BH. All non acquisitive firms with a book to market ratio of the middle 40% and the last 30% of the whole sample are put in a group and are called BM and BL respectively. These groups are refreshed monthly, just the same as the acquisitive portfolio. Nine portfolios will be constructed this way.
The formula for the calendar time abnormal return is as follows :
t cpi, t i, t i, R -R CTAR =
Here
R
i,tis the return on the acquisitive stock andR
cpi,tis the return on the referenceportfolio. Then for each calendar month we calculate a mean return CTAR across the firms in the portfolio: i t i i CTAR w, Nt 1 t CTAR
∑
= =WhereNtis the number of firms in the calendar month t and Wi,t =1/Nt in case of equally
weighted abnormal returns and , 1
1 1 , , / − = −
∑
= i t N i t i t i size sizet-The next step we do is to define a grand mean monthly calendar time abnormal return (MCTAR). This can be defined as the average of the monthly average abnormal return as can by the following formula :
t t CTAR
∑
= = T 1 T) (1/ MCTARIn this formula T can be defined as the total number of calendar months and the sum is taken over all calendar months.
) / ) ( ( /
t(MCTAR)=MCTAR
σ
CTARt TIf this test statistic is significantly different from zero, it indicates a significant abnormal return.
4. Data
The Zephyr database is used to collect the dataset of M&A’s with an Euro area acquirer. In this study we search for all Euro area acquiring firms that had any M&A activity in the period of January 1997 till December 2005. We will research four sub-samples. The first sub-sample consist of domestic Euro area acquirers. The second sub-sample consists of cross-border Euro area acquirers. The third sub-sample consists of inter-continental Euro area acquirers. The fourth sub-sample consists of intra-continental Euro area acquirers with a target firm located in the U.S.
The acquiring firms had to be located in one of these Euro area countries : Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal and Spain. These countries are part of the European Union (EU) that have the Euro as the main currency of payment. The introduction of the Euro as the main form of payment for countries that are part of the European Monetary Union (EMU) made it easier for firms to look for M&A’s outside their home borders.
We require that the bidder is listed on a stock exchange. The acquiring company has to be listed on one of the Euro area exchanges because in order to select our benchmark index we a need financial data on those firms that have not had any M&A activity to compare with our acquisitive portfolio of firms. We are looking for a merger or acquisition that has a deal value of 50 million or more. We use this value to isolate the biggest M&A’s in this time period. Furthermore the acquiring company has to have an acquired stake with a minimum of 50% and a maximum of 100%. The deal type has to be either a merger or acquisition. The last requirement is that the form of payment has to be either cash or shares. For the risk free asset we use the return on the one-month euro mark deposit quoted in London.
4.1 Summary statistics for domestic acquisitions
First we came up with a sub-sample of firms that met all the requirement as mentioned before and had all the financial information needed to calculate abnormal returns. The first sub-sample consists of domestic Euro area acquirers. The sub-sample consists of 160 domestic M&A activities. The descriptive statistics of domestic M&A’s are shown in table 4.1. For the total 160 domestic M&A that took place in the period of 1997-2005 there were a total of 108 occasions were the method of payment was cash, 29 shares payments and 23 mixed payments.
Table 4.1
Summary statistics Domestic acquisitions
This table shows the summary statistics of domestic acquisitions performed by a Euro area acquirer. The sample consist of 160 Mergers and acquisitions for an domestic M&A.
Aggregate Average Method of payment Dollar Dollar
Total Value Value
Number of of All Per Number Number Number Year Acquisitions Acquisitions Acquisition of Cash of Share of Mixed
(millions) (millions) Payments Payments Payments
1997 3 5.134 1.711 2 0 1 1998 8 4.917 614 5 1 2 1999 20 37.775 1.888 12 5 3 2000 23 21.771 946 12 7 4 2001 17 54.788 3.222 8 3 6 2002 27 13.466 498 23 2 2 2003 11 19.723 470 7 4 0 2004 25 78.507 2.121 20 3 2 2005 26 18.810 723 19 4 3 Total 160 254.895 108 29 23
4.2 Summary statistics for cross-border acquisition acquisitions
The cross-border M&A activities within the Euro area show a much higher frequency. The sub-sample consists of 355 cross-border M&A activities. The descriptive statistics of the cross-border M&A’s are shown in table 4.2. Of the total sample of 355 acquisitions that took place within the Euro area during the period January 1997 and December 2005 there were 277 occasions were the method of payment was cash. This indicates that around 78 % of our total sample the acquisition was financed with cash. This is a bit higher then the domestic acquisitions were 68 % of the total sample was financed with cash. There were 41 occasions were the method of payment was shares, this is about 12 % of the total sample and is lower then the domestic acquisitions sub-sample (18%). There were 37 occasions were the deal was financed with a combination of cash and shares. This is about 10% of the total sample of 355 cross-border acquisitions.
Table 4.2
Summary statistics Cross-border acquirer
This table shows the summary statistics of cross-border acquisitions performed by a Euro area acquirer. The sample consist of 355 Mergers and acquisitions for cross-border deals.
Aggregate Average Method of payment
Dollar Dollar
Total Value Value Number Number Number
Number of of All Per of Cash of Shares of Mixed
4.3 Summary statistics for inter-continental acquisitions
The sub-sample of inter-continental M&A’s consists of 188 firms where we were able to find sufficient financial information to be included in the sub-sample. Of the 188 cross-border M&A’s there are 106 domestic deals and 82 cross-border deals. This indicates that around 56 % of the sample consist of domestic deals and around 44 % of the sample consist of cross-border deals. Of the whole sample of 188 deals there were 121 M&A’s were the method of payment was cash. This is about 65% of all deals were the method of payment is cash. There were 27 deals were the method of payment was shares, which is about 14% of the whole sample of 188 firms. There were 40 payments of M&A of an Euro acquiring firm with a Euro area target were the method of payment was mixed, this is about 21% of the whole sample.
Table 4.3
Summary statistics Inter-continental acquirer
This table shows the summary statistics of inter-continental acquisitions performed by a Euro area acquirer. The sample consist of 188 Mergers and acquisitions for inter-continental M&A’s.
Aggregate Method of Payment Dollar
Total Value Number Number Number M&A M&A Number of of All of Cash of Share of Mixed Cross Within Year Acquisitions Acquisitions Payments Payments Payments Border Border
(millions) 1997 4 5.201 2 0 2 2 2 1998 7 19.861 4 1 2 2 5 1999 17 68.963 11 4 2 8 9 2000 26 33.716 14 5 7 14 12 2001 19 55.155 8 3 8 10 9 2002 30 14.862 25 2 3 13 17 2003 16 6.992 8 6 2 10 6 2004 28 6.742 21 2 5 12 16 2005 41 14.935 28 4 9 11 30 Total 188 226.432 121 27 40 82 106
4.4 Summary statistics of intra-continental acquirer
The sub-sample of intra-continental Euro area acquiring firm with a United States target consist of 116 M&A acquirers. Of these 116 M&A’s there are 81 occasions we were the method of payment is cash. There were 19 occasions were the method of payment was shares which is about 15% of the whole sample. This is about the same as the inter-continental acquiring firm of an Euro target firm sample which had 14% of the M&A’s paid with shares. There were 17 occasions of an Euro area acquiring firm taking over a U.S target were the method of payment was mixed. The summary statistics for an intra-continental acquirer are shown in table 4.4.
Table 4.4
Summary statistics Intra-continental acquirer
This table shows the summary statistics of intra-continental acquisitions performed by a Euro area acquirer. The sample consist of 160 Mergers and acquisitions for intra-continental M&A’s.
Aggregate Average Method of payment
Dollar Dollar
Total Value Value
Number of of All Per Number Number Number
5. Results
Most studies are examined with an ordinary least regression (OLS). In this method the total sum of the squared differences between a model and the data points are examined. The statistical R-squared is a direct measure of how well the regression fits the model. This variable measures the amount of variance that is explained by the model. A high R-squared means that the data points fit the model well.
In our research we will make use of the Weighted Least Square (WLS) regression as well as the Ordinary least squares (OLS) regression. These are the SMB, HML, Excess market returns and abnormal return. The loadings of the SMB portfolio shows the risk premium for the variable of size. The loadings of the HML portfolio show if there is extra risk for the book to market ratios. The loadings of the excess market return show us if the portfolio is riskier then the market (>1.00). The most interesting variable is the long run abnormal return. When this test-statistic is significantly different from zero we can reject the hypothesis of zero abnormal returns.
5.2 Domestic Acquisitions
Table 5.1 reports the results of domestic Euro area acquisitions. For domestic firms across all our research methods we found an adjusted R-squared between 42.3% and 48.7%. We can conclude that the data does not fit the Fama and French three factor model very well, at least not as good as the original Fama and French (1993) results with a fit around 95%. The loadings of the SMB portfolio show a significant negative value across all research methods, therefore there is a low risk premium for this portfolio. The loadings of the HML is insignificantly different from zero; there is no extra risk for book to market ratios on our portfolio. The excess market loading is significantly positive across all research methods varying from 0.770 to 0.826. This means that the acquisitive portfolio is less riskier then the market. The abnormal return shows an insignificant positive number varying from 0.2% to 3.7% across all research methods. We can conclude from this regression that the hypothesis of zero abnormal returns cannot be rejected for our sample. This positive value is in line with Loderer and Martin (1992) and Gregory (1997) who found a significant positive value of 1.50% and 0.80% respectively. Our findings of insignificantly positive results were higher for the WLS method (3.7% and 3.1%) and lower for the OLS method (0.5% and 0.2%) compared with the studies of Loderer and Martin (1992) and Gregory (1997).
Table 5.1
Domestic acquisitions on a Fama and French three factor model
This table shows the regression results of domestic Euro area acquisitions on a Fama and French three factor model. The results are the loadings of the HML, SMB and Excess market return loadings on cross border acquisitions within the Euro area during the time period of January 1997 until December 2005. There are a total of 160 completed acquisitions within this time period.
SMB HML Excess market return
Abnormal return
R-squared Adjusted R-squared F-statistic
Fama and French three factor model
Weighted Least Squares Ordinary Least Squares
Equally weighted (1) Value weighted (2) Equally weighted (4) Value weighted (4)
-0.314 * -0.056*** -0.250* -0.442** (-1.58) (-2.57) (-1.32) (-2.26) 0.008 0.083 -0.047 -0.001 (0.05) (0.63) (-0.36) (-0.01) 0.826*** 0.815*** 0.770*** 0.797*** (4.36) (4.24) (4.04) (4.51) 0.037 0.031 0.005 0.002 (1.17) (0.91) (0.09) (0.28) 0.468 0.487 0.423 0.475 0.452 0.471 0.400 0.454 0.000 0.000 18.559 0.000
5.3 Cross border acquisitions
Table 5.2 reports the results of cross-border Euro area acquisitions. The value of the adjusted R-squared varies from 48.7 % to 54.6 %. This is not as good a fit as the original Fama and French (1993) of 95 % but around the same values as the domestic Euro area acquisitions. The loadings of the SMB portfolio show a significant negative value. The value of the value weighted portfolios is higher compared to the equally weighted portfolios. In a value weighted portfolio the bigger firms are weighted heavier and therefore is a lower risk premium. The loadings of the HML is insignificantly different from zero, there is no extra risk for book to market in our portfolio. The excess market return is significantly positive across all research methods but not riskier then the market itself. The abnormal return is insignificantly negative varying from -0.3 % to -1.5 %. We can conclude from this that the hypotheses of zero abnormal returns cannot be rejected for our sample. Our results are in line with Andre, Kooli and L’Her (2004) and Kang (1993) who found a significant negative result of -1.14 % and -0.12 %.
Table 5.2
Cross-border acquisitions on a Fama and French three factor model
This table shows the regression results of cross-border Euro area acquisitions on a Fama and French three factor model.
SMB HML Excess market return
Abnormal return
R-squared Adjusted R-squared F-statistic
Fama and French three factor model
Weighted Least Squares Ordinary Least Squares
Equally weighted (1) Value weighted (2) Equally weighted (3) Value weighted (4)
-0.425** -0.600*** -0.344** -0.537 (-2.12) (-3.03) (-2.03) (-3.26) -0.008 -0.010 0.053 0.117 (-0.07) (-0.08) (0.55) (1.01) 0.888*** 0.822*** 0.808*** 0.800*** -0.003 -0.001 -0.001 (5.38) -0.015 (6.23) (6.22) (5.95) (-0.07) (-0.33) (-0.08) (-0.16) 0.546 0.526 0.487 0.494 0.533 0.512 0.472 0.480 0.000 0.000 0.000 0.464
* Significant at the 0.10 level; ** significant at the 0.05 level; *** significant at the 0.01 level. (t-values are shown in parentheses)
5.4 Inter-continental Acquisitions
Table 5.3
Inter-continental acquisitions on a Fama and French three factor model
This table shows the regression results of a inter-continental Euro area acquisitions on a Fama and French three factor model.
SMB HML Excess market return
Abnormal return
R-squared Adjusted R-squared
F-statistic
Fama and French three factor model
Weighted Least Squares Ordinary Least Squares
Equally weighted (1) Value weighted (2) Equally weighted (3) Value weighted (4)
0.535*** 0.094 0.548*** 0.103 -5.412 (0.45) (5.57) (0.49) 0.134* 0.079 0.136** 0.080 (1.62) (0.66) (1.64) (0.66) 0.934*** 0.855*** 0.947*** 0.868*** (19.79) (9.71) (20.60) (9.92) 0.003 0.007 -0.002 0.005 (0.16) (0.20) (-0.08) (0.14) 0.779 0.567 0.804 0.581 0.772 0.554 0.797 0.566 0.000 0.000 0.000 0.000
* Significant at the 0.10 level; ** significant at the 0.05 level; *** significant at the 0.01 level
5.5 Intra-continental acquisitions
Table 5.4 reports the intra-continental Euro area acquisitions. These acquisitions have a Euro area acquirer with a United States target firm. The adjusted R-squared of these portfolios of intra-continental acquisitions are very low varying from 3.9% to 9.2 %. This means that the data points in our sample have almost no explanation for the dependent variable. The loadings of the SMB portfolio show a negative value where the value weighted WLS method is significant. The loadings of the HML portfolio show a insignificant positive value. The Excess market return show a negative value, where the equally weighed OLS method is significant. The abnormal return shows a significant positive value for equally weighted portfolios performed with an OLS regression. The other methods show a insignificant value for abnormal returns.
Table 5.4
Intra-continental acquisitions on a Fama and French three factor model
This table shows the regression results of a intra-continental Euro area acquisitions on a Fama and French three factor model.
SMB HML Excess market return
Abnormal return
R-squared Adjusted R-squared
F-statistic
Fama and French three factor model
Weighted Least Squares Ordinary Least Squares
Equally weighted (1) Value weighted (2) Equally weighted (3) Value weighted (4)
0.354 -0.916** -0.042 -0.927 (-0.93) (-1.80) (-0.81) (-1.79) 0.016 0.082 0.010 0.078 (0.05) (0.22) (0.04) (0.21) 0.209 -0.258 -0.349* -0.274 (0.98) (-0.95) (-1.15) (-0.99) -0.031 0.012 0.203* 0.006 (-0.64) (0.14) (1.23) (0.07) 0.092 0.039 0.090 0.040 0.065 0.010 0.058 0.007 3.352 1.348 0.042 0.314
* Significant at the 0.10 level; ** significant at the 0.05 level; *** significant at the 0.01 level (t-values are shown in parentheses)
5.6Method of Payment - domestic / cross-border
