The effect of corporate acquisitions on stock
performance of bidder firms in the US
Hamied Safay
S3185443 M.Sc. Finance Supervisor: Dr. A.G. Schertler January 10th, 2019 University of Groningen Faculty of Economics and Business AbstractThis paper examines the impact of acquisition announcements on the bidder firms’ stock performance, by means of an event study as described in MacKinlay (1997). I analyze the post-announcement stock returns of 265 publicly quoted US firms subsequent to 606 acquisitions, during the period 2003-2018. Employing the Fama-French three-factor as the expected return model, I find that bidder firms, on average, earn significantly positive abnormal returns at the order of 0.21% on the day of the announcement of an acquisition and significantly positive cumulative abnormal returns at the order of 0.46% in a three-day window around the announcement day. Utilizing multivariate ordinary least squares regressions on the entire sample and subsamples of domestic, cross-border, horizontal and conglomerate acquisitions, I find that these (cumulative) average abnormal returns can partially be explained by bidders’ return on assets and whether cross-border acquisitions were paid for with non-cash instruments. Cross-border acquisitions during the 2007-2008 financial crisis additionally, on average, negatively impacted bidders’ abnormal returns.
1 Introduction
by Hackbarth and Morellec (2008) affirms negative abnormal returns (during the event window) with respect to the bidder firms, whereas the contrary is stated by Wu et al. (2016) and Boateng et al. (2008). The lack of a consensus regarding bidder performance is thus seen as the main problem. Obtaining insights on the post-facto effect of acquisition announcements on the short-term stock performance of bidder US firms (confined to the period 2003-2018), to contribute to the clarity of literature on this topic, will therefore be the aim of this study. This aim is achieved by means of an event study and additionally examining the determinants of the (cumulative) abnormal returns utilizing multivariate OLS regressions. The multivariate regression analyses subsequently allow for the more detailed disjoint examination of bidder performance, to produce answers to questions with regards to the deviations and relations that exist between domestic and cross-border acquisitions, and horizontal and conglomerate acquisitions. This study differs from other studies on the equivalent topic on grounds of it utilizing expected return models based on both the return of the aggregate market (Fama-French three-factor model) and the individual firms (constant-mean model), which I have not yet seen happen before. This research will thus, all aspects considered, contribute to the ever-expanding knowledge on mergers and acquisitions in the realm of corporate finance.
The outline of this study is structured as follows: section two provides an elaboration on the relevant literature and subsequently states the hypotheses that follow from the literature review. Section three describes the applied methodology. Section four provides insight on the used data with regards to the variables. Section five presents the empirical results w.r.t the event study and regression analysis. This study is concluded in section six by dint of concluding remarks on the empirical results and their implications.
2 Literature review
grounds of being subjected to less risk. Managerial synergies are conclusively related the firms’ management and their ability to enhance its performance (Trautwein, 1990). Halpern (1982), as elucidated in his study, affirms that synergetic M&A transactions are subsequently subjected to positive bidder performance. One of the main motives against mergers and acquisitions is based on Jensen’s free cash flow theory (1986), which hypothesizes that, contingent upon retaining free cash flows, a firm’s management rather invests in negative net present value investments (i.e. value destroying acquisitions) than distributing cash in form of dividends. High free cash flows and relatively low investing opportunities are consequently presumed to conjointly be related to negative abnormal returns, on the bidder firms’ parts. Low investment opportunities are often proxied by a Tobin’s Q ratio between 0 and 1, implying that a specific firm’s market value is lower than its replacement costs (Tobin, 1969). Tobin’s Q can thus subsequently be computed as the ratio of a firm’s market value divided by the replacement costs of its assets (Lang et al., 1989). An increase in this ratio implies that firms are more likely to attract positive net present value projects, such that the free cash flow may be utilized in a more productive manner. The expected impact of the interaction between a low Tobin’s Q ratio and cash holdings is presumed to negatively affect (cumulative) average abnormal returns. In addition to Tobin’s Q, another common approach to measuring performance is to examine bidder’s return on assets (see, e.g., Adams and Mehran, 2005, and Yermack, 1996), as this ratio provides an objective outlook on past performance. The return on assets is computed as the ratio of net sales over total assets and is hypothesized to have strong prediction power with regards to firms’ future performance (Yermack, 1996). The following hypothesis is based on the interaction between a low Tobin’s Q ratio and excess cash holdings (see Lang et al., 1991), and is constructed as follows:
H1: Low investment opportunities conjoint with free cash holdings are, on average,
negatively related to the (cumulative) abnormal returns
2.2 Bidder performance
Taking the implications of the theoretical literature into consideration, it is difficult to determine whether announcements of acquisitions are indeed performance enhancing in the short-term. An overview of empirical studies is therefore examined to uncover the nature of this effect regarding bidder performance.
(2008) are however contradicted by Dodd (1980), Hackbarth and Morellec (2008), Gupta and Misra (2007), and Ben-Amar and Andre (2006), whom all, in their respective samples, found negative (cumulative) abnormal returns, on the bidder firms’ part. The negative abnormal returns in the study by Dodd (1980), however, were not significant. Given the abovementioned empirical studies, it is safe to assume that there does not seem to be a definite consensus on whether or not announcements of acquisitions affect bidder performance positively or negatively. The results vary geographically. The main hypotheses with regards to the short-term event study are formulated on basis of the literature on bidder performance. Due to the lack of a consensus, the effect of acquisition announcements can be ambiguous. I therefore construct the hypothesis as such:
H2: Announcements of corporate acquisitions, on average, do not generate significant
(cumulative) abnormal returns for bidder firms in the short term
2.2.1 Domestic vs. cross-border M&A
Since the empirical literature on bidder performance proved that there does not seem to be a consensus on whether or not announcements of acquisitions affect bidder performance positively or negatively, it may be interesting to investigate whether a consensus exists in relation to cross-border M&A, as the results of the abovementioned studies varied geographically.
plea that, although the post-announcement abnormal returns in their study were not significant, the three-year post-acquisition returns were highly significantly negative at -32%. Even when cross-border deals are compared to domestic deals, as was the case in the study by Aw and Chatterjee (2004), it was found that cross-border acquisitions yielded lower returns for the bidder firms, though these abnormal returns were not significant. Gugler et al. (2003) support this finding, by affirming not finding significant differences in cross-border and domestic acquisitions, performance wise. Despite using other estimation techniques, such as Fama-French’s three-factor model, both Ang & Kohers (2001) and Moeller et al. (2004) found no evidence of abnormal returns in the long run (a three-year period post-acquisition). Empirical evidence thus, conclusively, leans towards relating cross-border acquisitions to negative abnormal returns. The third hypothesis is therefore formulated as:
H3: Cross-border acquisitions are negatively related to the (cumulative) abnormal
returns 2.3 Determinants of bidder performance This subsection provides a discussion on common determinants of bidder performance, which are bidders’ market capitalization and transaction value, method of payment and M&A hostility. 2.3.1 Market capitalization and transaction value The cumulative abnormal returns to the bidder are positively correlated with the relative size (market capitalization) of the bidder firms, as the gains to bidder firms increase with an increase in firm size (Asquith et al., 1983). This finding is supported by Franks et al. (1991), whom found similar results by examining bidder firms’ post-merger performance. As for the value of a transaction, Hogholm (2016) reported that the size of M&A deals has a significantly negative impact on the bidders’ abnormal returns on the day of an acquisition announcement, as a negative relationship exists. In the long term, large sized deals tend to underperform (Dutta & Jog, 2009). The following two hypotheses are constructed in regards to the market capitalization of bidders and the transaction values of acquisitions:
H4: Bidder market capitalization is positively related to the (cumulative) abnormal
issue of adverse selection (information asymmetry), whereby a firm issuing equity implies the overvaluation of its stock. Furthermore, Loughran and Vijh (1997) affirms the existence of adverse selection and that argues transactions are only financed by equity, when the bidder firms’ stock is overvalued and hence relates equity financing to negative abnormal returns. This finding is self-evidently accordant with the signal theory as formulated by Leland and Pyle (1977). Theoretical implications thus consensually relate equity financing to negative (cumulative) abnormal returns. These theoretical implications are subsequently supported by empirical evidence by Moeller and Schlingeman (2005), Franks et al. (1988) and Travlos (1987), whom, in their respective studies, found that equity financed corporate acquisitions, contrary to cash financed acquisitions, were related to significantly negative abnormal returns, on the bidders’ part. As both the theoretical and the empirical literature imply that non-cash payments yield negative abnormal returns to the bidder firms. The sixth hypothesis is thus: H6: The payment for an acquisition by equity is negatively related to the (cumulative) abnormal returns 2.3.3 M&A hostility The attitude of an acquisition principally takes two forms, it may either be friendly or hostile. Friendly acquisitions are axiomatically subjected to negotiations between the bidder and targets firms, whereas the target firms’ management is deliberately disregarded upon hostile acquisitions. The level of hostility is proxied by whether a tender offer was made by the bidder firm, as tender offers generally coincide with hostile acquisitions (Holmen and Nivorozhkin, 2009). A tender offer is executed by bidder firms and is formally defined as a public invitation to the target firm’s stockholders to tender their shares, for a pre-specified price, which is usually above the current market price (Sridharan and Reinganum, 1995). Bids, with respect to hostile acquisitions, are commonly greater than their analogous counterpart in friendly acquisitions, on grounds of the higher premium that must directly be paid to the targets’ stockholders, to bypass its management (Asquith and Kim, 1982). This relatively large premium is intuitively only proposed when it is outweighed by the potential synergies that may arise from the acquisition (Travlos, 1987). Frank and Harris (1989) specified that these hostile acquisitions, outperformed friendly acquisitions, with regards to bidder performance. Successful tender offers are consequently found to be firm value enhancing, for both the bidder and the target (Bradley et al., 1988), and positively related to abnormal returns (Dodd and Ruback, 1977). The relationship between tender offers and abnormal returns is therefore expected to be positive. This conclusively results in the final hypothesis:
H7: Acquisitions in which a tender offer is made are positively related to the (cumulative)
abnormal returns
3 Methodology
This section delves deeper into the methodology, which is based on MacKinlay (1997). I describe the framework of an event study and provide a quantitative analysis of the applied parametric and nonparametric statistical significance tests that are utilized to acquire the (cumulative) abnormal returns. In addition, I construct distinct regression equations on the basis of the literature review to examine the determinants of the (cumulative) abnormal returns.
3.1 Event study framework
The estimation window is expressed as !" up and till !#, whereas the event window is expressed as !# up and till !$.
Now that the event windows and the estimation window are determined, the computations regarding the normal return and expected return have to be constructed. The continuously compounded return of a security follows from:
%&' = )* + ,&'
,&'-./ (1)
Where %&' is the return of security i on day t, ln is the natural logarithm, ,&' is the adjusted stock price of security i on day t and ,&'4#is the adjusted stock price of security i on day 54#.
The general computation of abnormal returns, given an event, follows from:
6%&' = %&'− 8(%&'|:') (2)
Where 6%&' is the abnormal return of security i on day t, and 8(%&'|:') is the expected return of security i on day t, given an event X. I utilize the Fama-French three-factor model (1993) with regards to computing the expected return. The traditional capital asset pricing model (abbreviated as CAPM) estimates the return of a security by measuring its sensitivity to the market portfolio, in excess of the risk-free rate of interest. Contrary to the CAPM, the Fama-French three-factor model employs two additional factors, namely the outperformance of small-cap firms relative to high-cap firms, and the outperformance of high book-to-market firms relative to low-book-to market firms. Fama and French (1993) constructed various portfolios on basis of firms’ market capitalization and book-to-market ratios, and found that, in the long term, small-cap firms outperform large-cap firms, whereas value firms (high book-to-market ratio) outperform growth firms (low book-to-the long term, small-cap firms outperform large-cap firms, whereas value firms (high book-to-market ratio). This model therefore explains the return on a security on basis of its sensitivity to the market portfolio, a size-related portfolio and a portfolio on grounds of the book-to-market ratio. The inputs for the additional factors are obtained from Kenneth French’s data library. The expected return can thus be computed by means of the following equation:
8(%&') = <& + >?+ @&A%B− >?C + D&EFG + H&IFJ + K&' (3)
Where 8(%&') is the expected return on security i on day t, <& is the constant term, >? is the risk-free rate of interest, @& is the sensitivity of security i to the market, D& is the sensitivity of security i to the size premium, H& is the sensitivity of security i to the value premium, and K&' is the error term with zero mean and constant variance.
The abnormal returns utilizing the Fama-French three-factor model then follow from:
6%&' = %&'− (<M&+ @N&A%B− >?C + DN&EFG + Ĥ&IFJ + K&') (4)
Where the constant term <M and the coefficients @N, DN and Ĥ are estimated by dint of a multiple ordinary least squares (OLS) regression. The individual abnormal returns must be aggregated in order to compute the average abnormal returns across all securities. This follows from equation: 6%QQQQ'= 1 RS 6%&' T &U# (5)
Where 6%QQQQ' are the average abnormal returns on day t, obtained from the equally weighted summation of the abnormal returns across the totality of securities on the same day t and N is the total number of observations.
The significance of the obtained (cumulative) average abnormal returns can conclusively be determined by means of conducting parametric and nonparametric significance tests. Parametric significance tests require normally distributed data, whereas nonparametric tests are not restricted by this condition. The central limit theorem (abbreviated as CLT) affirms that non-normality should not be an issue for sufficiently large sample sizes (n > 30, (Cohen and Levinthal, 1990)), as its mean will be well-approximated by a normal distribution, even if the data are non-normally distributed (DasGupta, 2010). The sample I use consists of 614 observations, which is sufficiently large to satisfy the conditions of the CLT. However, in addition to the application of the CLT, I evaluate the normality of the data by conducting the Jarque-Bera test. Rejecting the null hypothesis of this test implies that the data are non-normally distributed.
3.1.1 Significance tests
Where W&' is the ranking of security i on day t, F& is the number of matched returns in the estimation window and J& is the number of matched returns in the event window. The test statistic is subsequently calculated as the ratio of the average excess rank over its standard deviation.
3.2 Regression analysis
In addition to the short-term examination of the stock performance of bidders, by means of an event study, I furthermore run multivariate regressions on the determinants of the (cumulative) abnormal returns. These determinants are thoroughly discussed in the literature review (section 2) and have proven to (significantly) affect stock performance of bidder firms. To ensure the utilization of the best linear unbiased estimators (abbreviated as B.L.U.E.) in the regression analyses, I perform the White test (1980), where the squared residuals in the regression model are regressed on the independent variables, to determine whether the variance of the error terms is constant over time. Heteroscedasticity is countered by applying White’s heteroscedasticity consistent standard errors. Another complication may be multicollinearity, which implies the existence of a linear relationship between the independent variables, leading to biased estimators. Multicollinearity will be checked for by the application of variance inflation factors. In case multicollinearity is detected in independent variables, these variables will be removed, as is advised by O’Brien (2007). I will conclusively rely on the central limit theorem, contingent on the violation of the assumption of normally distributed residuals, given the relatively large sample size. It is conclusively imperative to manage the abovementioned complications, as inferences based on the regression output may be erroneous when the estimators are not B.L.U.E..
The (cumulative) average abnormal returns are regressed on the explanatory variables, which are chosen on basis of a review of the literature on M&A and its determinants. These independent variables have proven to significantly affect bidder performance. The two baseline regression equations initially capture the individual effects and are constructed as:
6%
QQQQ" = < + @#F[&+ @$!\& + @][I& + @^%_6&+ @`!a&+ @bR[&+ @c!_&+ @dI6& + @e[G& + K& (7)
[6%
QQQQQQ[4#,#] = < + @#F[& + @$!\& + @][I&+ @^%_6& + @`!a& + @bR[& + @c!_& + @dI6& +
@e[G&+ K& (8)
Where <" is the constant term in the equation and @i corresponds to the coefficients of the variables. Furthermore, I set up nine explanatory variables, consisting of five dummy variables, where:
I. F[& equals the logarithm of the bidder firm’s market capitalization.
III. [I& equals the ratio of cash holdings over the bidder firm’s total assets. IV. %_6& equals the ratio of net sales over the bidder firm’s total assets.
V. !a& is a dummy variable assigned the value “1” if the ratio of Tobin’s Q is between zero and one, and “0” otherwise.
VI. R[& is a dummy variable assigned the value “1” if a transaction is paid for with non-cash instruments, and “0” otherwise.
VII. !_& is a dummy variable assigned the value “1” if a tender offer is made regarding the transaction, and “0” otherwise.
VIII. I6& is a dummy variable assigned the value “1” if the acquisition is in a related industry (i.e. horizontal and the same two-digit SIC codes), and “0” otherwise.
IX. [G& is a dummy variable assigned the value “1” if the acquisition concerns a cross-border firm, and “0” otherwise.
The specifications of the baseline equations are subsequently altered by including interaction terms between the low Tobin’s Q dummy and cash holdings (see Lang et al., 1991), between the non-cash dummy and cross-border variable, and between the market capitalization and transaction value. These additional terms are constructed on grounds of an examination of the literature (see, e.g. Jensen, 1986, and Lang et al., 1991) and consulting the correlation matrix and checking for high correlations between variables, and self-evidently capture the interaction effects between the sets of variables. Ramsey’s RESET test (1962) is furthermore conducted to verify the correct specification of the models.
4 Data
This section describes the data platforms (Zephyr and DataStream) that are utilized with regards to this study. It furthermore amplifies on the restrictions with respect to the dataset, and summarizes the descriptive statistics on the (in)dependent variables in the regression analysis.
Zephyr is an online database, created by Bureau van Dijk, containing information regarding corporate transactions (i.e. mergers and acquisitions and initial public offerings). All information regarding M&A will thus be extracted from the Zephyr database. The steps in the data extraction process are described as follows:
I. The bidder firm is listed (publicly quoted), whereas the target firm is either listed, delisted or unlisted. Intuitively, the acquirer must be listed, since stock price information is imperative. This step resulted in 317,978 transactions.
III. The bid is for a majority stake in the target firm. Thus, the percentage of final acquired stake is minimally 50%. This results in 96,719 transactions. IV. The current deal status is completed. This step results in 82,436 transactions. V. The bid for an acquisition is announced and completed within the 1st of January 2003 and 1st of January 2018. This step results in 52,599 transactions. VI. The bidder firm is a constituent of the S&P500. This resulted in 5,488 transactions. VII. The transaction has a minimum value of 100M USD. This step resulted in 640
transactions.
Table 1. Descriptive statistics on the (in)dependent variables
This table provides summary statistics on the dependent and independent variables. The values of the mean, median, standard deviation, minimum, and maximum are denoted in percentages, except for the values of the logarithm of bidder’s market capitalization, which are denoted in logs.
Variable Obs. Mean Median Std. dev Min Max
Dependent 6% QQQQ" 614 0.205 0.049 2.604 -18.886 12.719 [6% QQQQQQ[4#,#] 614 0.456 0.264 3.663 -32.798 20.931 Independent Logarithm of bidder’s market cap 614 24.35 24.16 1.120 21.70 27.40 Transaction value 614 3.806 1.357 7.035 0.013 68.84 Cash holdings 614 9.875 7.358 9.102 0.026 70.41 Return on assets 614 7.140 6.750 6.537 -42.88 31.43 Low Tobin’s Q dummy 614 4.780 0 20.74 0 1 Non-cash dummy 614 52.64 1 49.97 0 1 Tender offer dummy 614 4.785 0 21.36 0 1 Horizontal acquisition dummy 614 76.57 1 42.39 0 1 Cross-border dummy 614 28.22 0 45.04 0 1 The count of firms with a Tobin’s Q ratio between zero and one is notably low, implying that the bulk of the firms in this sample are likely to attract positive NPV investment opportunities. The same low count holds for the number of tender offers. The majority of acquisitions are thus presumed to have been executed ‘peacefully’, in collaboration with the target firms’ management. The count of domestic acquisitions and acquisitions in related industries are relatively high, compared to their respective counterparts.
5 Results
This section provides an overview of the empirical results concerning the short-term event study. I initially present the results of the (non)parametric significance tests and subsequently elaborate on the results with respect to the determinants of bidder performance. 5.1 Short-term event study This subsection amplifies on the results with regards to the short-term event study. Table 3 provides the p-values of the parametric cross-sectional t-test and nonparametric Corrado rank test (1989). The results show that the average abnormal returns were at their highest levels on the day of an announcement and the day subsequent to an announcement. This implies that bidder firms, on average, earned a significantly positive return of approximately 0.21%, in excess of the expected return, on the day of announcing an acquisition. This translates to yielding a return of roughly 67.54% on an annual basis, implying its economic significance in addition to statistical significance. The shock is directly incorporated into the stock prices of bidder firms and diminishes relatively quick. This finding is therefore in line with the semi-strong of market efficiency. The high abnormal returns on the day subsequent to an announcement may be explained by time zone differences between the bidder and target firms, since acquisitions may have been announced when the US markets were not trading (i.e. after 5 P.M). This finding furthermore corresponds to Wong and Cheung (2009), Wu et al. (2016), and Boateng et al. (2008), whom all, in their respective studies, found that bidder firms were subjected to positive abnormal returns subsequent to announcing an acquisition. The hypothesis of no average abnormal returns on the event day is consequently rejected on a 1% and 5% alpha level, on grounds of the t-test and Corrado rank test, respectively. Table 3. Results of the t-test and Corrado rank test w.r.t the daily average abnormal returns This table provides the daily p-values with regards to the results of the t-test and the Corrado (1989) rank test. The average abnormal returns are calculated on grounds of the Fama-French three-factor model (1993). The total count of observations equals 606 acquisitions.
Day !"#### in % t-test Corrado rank test
4 0.002 0.975 0.618 5 0.076 0.214 0.006*** 6 -0.153 0.013** 0.002*** 7 -0.024 0.692 0.663 8 -0.048 0.429 0.218 9 0.034 0.582 0.442 10 -0.061 0.318 0.179 Note: *, **, *** indicate statistical significance at the ten, five and one percent alpha level, respectively. 5.1.1 Robustness of the results The main disadvantage of the cross-sectional t-test I performed in the previous subsection is that it does not account for cross-sectional correlations in the abnormal returns and is prone to event-induced volatility (Brown and Warner, 1985). Given the fact that stock returns are typically positively correlated, this may lead to the underestimation of the variance in the abnormal returns, and thus lead to an increase in a type I error, where the null hypothesis may be unjustly rejected when it is actually true (Kolari and Pynnonen, 2010). It is therefore imperative to take this issue into consideration when testing for the significance of the (cumulative) abnormal returns. Under the condition of no cross-correlation among the abnormal returns, Kolari and Pynnonen (2010) found that standardized significance tests (see, e.g., Patell (1976); Boehmer et al. (1991)) have proven to outperform non-standardized tests, such as the traditional t-test. However, the abovementioned standardized significance tests rely on the assumption of cross-sectional independency in the abnormal returns. Kolari and Pynnonen (2010) therefore constructed an adjusted standardized t-test on basis of the tests by Patell (1976) and Boehmer et al. (1991), to account for cross-correlation in the abnormal returns. The adjusted standardized test, however, is not robust to event-induced changes in volatility, whereas the standardized t-test by Boehmer et al. (1991) is. I thus conduct both the standardized as well as the adjusted standardized t-tests, as both of these tests have their advantages. The abnormal returns can be standardized by means of the following equation:
$!"%& = !"%&
$()*+ (9)
returns following an event. The fraction of positive cumulative abnormal returns is computes as follows: 0̂ = 1 34 1 56 4 7%& 89 &:8; < %:6 (10) Where 0̂ is the fraction of positive cumulative abnormal returns in the estimation window, N is the count of observations and where 7%& equals “1” if the sign of the abnormal returns is positive and “0” otherwise.
The results of the robust (non)parametric significance tests on the (cumulative) average abnormal returns are presented in table 4. The results of the (adjusted) standardized t-tests are conforming to the result of the cross-sectional t-test.
Table 4. Results of the (robust) non(parametric) significance tests w.r.t. the short-term event study This table provides the p-values with respect to the significance of the average abnormal returns on the event day (day 0) and the cumulative average abnormal returns in three distinct event windows. The parametric tests consist of the cross-sectional t-test, the standardized t-test (Boehmer et al., 1991), and the adjusted standardized t-test (Kolari and Pynnonen, 2010). The nonparametric test consists of the Cowan generalized sign test (1992). The (cumulative) average abnormal returns are calculated on grounds of the Fama-French three-factor model (1993). The count of observations equals 606 acquisitions.
Window (")$%&&&&&&&& in % t-test Std. t-test Adjusted std. t-test Cowan gen. sign test
Event day 0.205 0.001*** 0.055* 0.103 0.021** [-1,1] 0.456 0.000*** 0.002*** 0.001*** 0.004*** [-5,5] 0.373 0.066* 0.069* 0.122 0.011** [-10,10] -0.000 0.995 0.997 0.980 0.313 Note: *, **, *** indicate statistical significance at the ten, five and one percent alpha level, respectively. As an additional and final means of robustness, I also utilize the constant-mean model to compute the expected returns and subsequently the abnormal returns. Self-evidently, this model aggregates the return on a security over the pre-specified estimation window of 120 trading days and computes the mean. This is expressed quantitatively as: %'( = *(%'(|,() + .'( = 0' + .'( (11)
5.2 Regression analysis
This section delves deeper into the determinants of M&A performance, to examine whether bidder performance may be explained by theoretical implications.
Table 5. Regression output w.r.t the baseline regression equations
returns, logarithm of bidder’s market capitalization, transaction value, cash holdings, and the return on assets. Dummy variables are not winsorized, since these variables are binary. I conclusively add an additional dummy variable to the regression equation to examine whether the credit crisis had a significant impact on bidder performance. This dummy variable takes on the value “1” if the acquisition was announced during the 2007-2008 recession, and “0” otherwise. However, acquisitions occur less frequently during periods of economic hardships, even though they may be performance enhancing (Rhodes and Stelter, 2009) and result in better deals on grounds of a decline in firm value. The baseline regression equations is therefore transformed into: #$ %%%%&'()*+',-.= 0 + 2345' + 2"67'+ 2859'+ 2:$;#'+ 2<6=' + 2>?5' + 2@6;' + 2A9#'+ 2BC5'+ 23D5E'+ 2336='∗ 59'+ 23"45'∗ 67'+ 2385E' ∗ ?5'+ G' (12) #$ %%%%.*M-)N'O = 0 + 2345'+ 2"67' + 2859'+ 2:$;#'+ 2<6='+ 2>?5'+ 2@6;'+ 2A9#' + 2BC5' + 23D6=' ∗ 59' + 23345' ∗ 67' + G' (13) #$ %%%%O+*))QR*+.-+ = 0 + 2345' + 2"67' + 2859' + 2:$;#' + 2<6=' + 2>?5' + 2@6;' + 2A9#' + 2BC5' + 23D6=' ∗ 59' + 23345' ∗ 67' + G' (14) #$ %%%%T*+',*(NUV = 0 + 2345'+ 2"67' + 2859'+ 2:$;#' + 2<6=' + 2>?5' + 2@6;'+ 2AC5' + 2B5E'+ 23D6='∗ 59' + 23345'∗ 67' + 23"5E'∗ ?5'+ G' (15) #$ %%%%O*(XV*M-+UN- = 0 + 2345'+ 2"67' + 2859'+ 2:$;#' + 2<6='+ 2>?5'+ 2@6;'+ 2AC5' + 2B5E'+ 23D6='∗ 59' + 23345'∗ 67' + 23"5E'∗ ?5'+ G' (16)
Variable (8) (9) (10) (11) (12) Constant 0.033 0.033 -0.022 0.046 -0.020 (0.279) (0.498) (0.657) (0.335) (0.746) Logarithm of bidder’s market cap -0.001 -0.001 0.001 -0.002 0.001 (0.337) (0.583) (0.569) (0.435) (0.728) Transaction value -0.068 -0.346 0.407 -0.707 0.767 (0.702) (0.780) (0.761) (0.566) (0.246) Cash holdings -0.014 -0.026 0.006 -0.032* 0.008 (0.341) (0.129) (0.876) (0.082) (0.752) Return on assets 0.012 -0.001 0.026 0.002 0.069 (0.592) (0.950) (0.385) (0.906) (0.217) Low Tobin’s Q dummy -0.001 -0.014 0.010 -0.016 0.012 (0.916) (0.373) (0.373) (0.436) (0.105) Non-cash dummy -0.004 -0.005 -0.013*** -0.004 -0.006 (0.231) (0.196) (0.005) (0.355) (0.388) Tender offer dummy 0.005 0.016 -0.003 0.013 -0.012** (0.415) (0.204) (0.718) (0.189) (0.047) Horizontal acquisition dummy 0.005 0.004 0.003 (0.124) (0.273) (0.616) Financial crisis dummy -0.003 -0.008 0.000 -0.006 -0.003 (0.535) (0.359) (0.976) (0.517) (0.716) Cross-border dummy 0.004 0.003 0.010 (0.376) (0.608) (0.315) Low Tobin’s Q x Cash holdings 0.018 0.064 0.056 0.164 -0.111** (0.837) (0.597) (0.585) (0.381) (0.015) Log market cap x transaction value 0.003 0.015 -0.018 0.031 -0.036 (0.659) (0.765) (0.767) (0.546) (0.192) Cross-border x Non-cash -0.007 -0.004 -0.015 (0.188) (0.537) (0.235) Number of observations 606 435 171 464 142 Adjusted R" 0.024 0.026 0.064 0.030 0.095 P-value (F-statistic) 0.192 0.419 0.087* 0.303 0.139 Note: *, **, *** indicate statistical significance at the ten, five and one percent alpha level, respectively. Regressing subsamples of the cumulative average abnormal has led to a decrease in the count of significant coefficients of the independent variables, as was expected. This occurs mainly due to the fact that new information will be incorporated into the stock prices instantly, which implies that the impact of a shock will less likely be able to persist over longer event windows. However, the cash holdings variable gained significance with regards to the cumulative average abnormal returns of horizontal acquisitions. This finding is not supported nor rejected in the literature, as I have not read any papers that examine the impact of this specific variable in combination with horizontal acquisitions.
6 Conclusion
determinants of this impact, and whether the determinants had a positive or negative effect on the stock performance. Empirical and theoretical studies in the field of mergers and acquisitions do not provide a consensus upon the definite sign of the impact of acquisition announcements, as this impact has proven to vary geographically.
The empirical component of this study is partially conducted by means of an event study as described in MacKinlay (1997). The sample consists of US bidder firms in the period 2003-2018. The totality of the sample consists of 265 firms and 614 announcements of acquisitions. The overall results of the short-term event study imply that US firms, on average, earned a significant amount of roughly 0.21% daily returns, subsequent to announcing an acquisition. This yields a return of 67.54%, on an annual basis. Cumulative abnormal returns in a three-day event window were also proven to be significantly positive, at the order of 0.45% (46.54% on an annual basis). The multivariate ordinary least squares regression subsequently provided evidence that these returns can partially be explained by the bidder’s return on assets (positive impact), by whether the (cross-border) acquisition is paid for with non-cash instruments (negative impact) and whether the acquisition was announced in the period of the 2007-2008 financial crisis (negative impact). Other explanatory variables did not significantly affect the (cumulative) abnormal returns. All of the hypotheses, with the exception of hypothesis 6 regarding the negative impact of non-cash payments, are rejected. The availability of time is considered a great limitation, since additional theories could have been implemented into this thesis, in a greater time span. The next few limitations are based on the relatively small sample size. Although this sample is perfectly suitable for a short-term event study, it may have fallen short with regards to the regression analyses, as variables such as the tender offer dummy and the low Tobin’s Q dummy were subjected to relatively low observations, especially in the specified subsamples. Increasing the number of observations for these variables would surely lead to increasingly reliable inferences. This limitation furthermore did not allow for the implementation of additional variables, such as private firms or firms in emerging markets (see Chari et al., 2004), into the regression analyses, as the count of these observations was also remarkably low. It would subsequently also not have been possible to categorize them into the specified subsamples. Including additional years into the sample would not improve this limitation, as the database Zephyr only provides data on M&A from 2001 onwards for the American market and 1997 for the European market. These few additional years did not significantly increase the number of private acquisitions. Other limitations may include the arbitrary chosen minimum transaction value of 100M USD and only including completed events in the dataset. A sample selection bias may have arisen on grounds of these limitations, as these values were chosen non-randomly.
The suggestions for further research are mainly based on the limitations of this study. It is imperative to include a sufficiently large amount of observations into the data set, in order to be able to make unbiased and just inferences. In the context of this study, this may have been possible by expanding the geographical location of the bidder firms to, for instance, Canada or other countries with high levels of M&A activity. Increasing the sample size by including other countries/regions would furthermore allow for the implementation of private acquisitions into the regression analyses, allowing for a more detailed examination and understanding of acquisition performance. Another solution is applying less restrictions on the dataset, such as decreasing the minimum transaction value. This may, however, affect the outcome of the event study significantly. Conclusively, as I only included completed and confirmed acquisitions into my sample, it may be of interest to additionally include rumored acquisitions into the sample and examine the differences between confirmed and rumored announcements of corporate acquisitions.
References
Adams, R., Mehran, H., 2005. Corporate performance, board structure and its determinants in the banking industry. Unpublished working paper. University of Oxford, Oxford. Andrade, G., Mitchell, M., Stafford, E., 2001. New evidence and perspectives on mergers. Journal of Economic Perspectives 15, 103-120. Ang, J., Kohers, N., 2001. The take-over market for privately held companies: the US experience. Cambridge Journal of Economics 25, 723-48.Asquith P., Bruner R., Mullins, D., 1983. The gains to bidding firms from merger. Journal of Financial Economics 11 121-139.
Asquith, P., Kim, E., 1982. The impact of merger bids on the participating firms’ security holders. Journal of Finance 37 1209-1228.
Aw, M., Chatterjee, R., 2004. The performance of UK firms acquiring large cross-border and domestic takeover targets. Applied Financial Economics 14, 337-349.
Barkema, H. G., Schijven, M., 2008. How do firms learn to make acquisitions? A review of past research and an agenda for the future. Journal of Management 34, 594-634.
Bertrand, O., Zuniga, P., 2006. R&D and M&A: are cross-border M&A different? An investigation on OECD countries. International Journal of Industrial Organization 24, 401-423. Bieshaar, H., Knight, J., van Wassenaer, A., 2001. Deals that create value. McKinsey Quarterly 1, 64-73.
Boateng, A., Qian, W., Tianle, Y., 2008. Cross-border M&As by Chinese firms: an analysis of strategic motives and performance. Thunderbird International Business Review 50, 259-270.
Boehmer, E., Masumeci, J., Poulsen, A., 1991. Event-study methodology under conditions of induced variance. Journal of Financial Economics 30, 253-272.
Bradley, M., Desai, A., Han Kim, E., 1983. The rationale behind interfirm tender offers: Information or synergy? Journal of Financial Economics 11, 183-206.
Bradley, M., Desai, A., Kim, E., 1988. Synergetic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms. Journal of Financial Economics 21, 3-40. Buckley, P., Casson, M., 1976. The Future of Multinational Enterprise. Homes and Meier Press, London.
Cohen, W., Levinthal, D., 1990. Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly 35, 128-152.
Conn, R., Cosh, A., Guest, P., Hughes, A., 2005. The impact on UK acquirers of domestic, cross-border, public and private acquisitions. Journal of Business Finance & Accounting 32, 815-870.
Corrado, C., 1989. A nonparametric test for abnormal security-price performance in event studies. Journal of Financial Economics 23, 385-395.
Cowan, A., 1992. Nonparametric event study tests. Review of Quantitative Finance and Accounting 2, 343-358.
DasGupta, A., 2010. Normal Approximations and the Central Limit Theorem. In: Fundamentals of Probability: A First Course. Springer, New York.
Dutta, S., Jog, V., 2009. The long-term performance of acquiring firms: a re-examination of an anomaly. Journal of Banking & Finance 33, 1400-1412.
Fama, E., 1970. Efficient capital markets: a review of theory and empirical work. Journal of Finance 25, 383-417.
Fama, E., French, K., 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial Economics 33, 3-56.
Franks, J., Harris, R., 1989. Shareholder wealth effects of corporate takeovers: the U.K. experience 1955-1985. Journal of Financial Economics 23, 225-249.
Franks, J., Harris, R., Mayer, C., 1988. Means of payment in takeover: results for the United Kingdom and the United States. University of Chicago Press, Chicago, IL. Franks, J., Harris, R., Titman, S., 1991. The postmerger share-price performance of acquiring firms. Journal of Financial Economics 29, 81-96. Gaughan, P., 2007. Mergers, Acquisitions, and Corporate Restructurings. John Wiley & Sons, Inc., Hoboken, New Jersey. Gersdorff, N. von, Bacon, F., 2009. Mergers and acquisitions: a test of market efficiency. Journal of Finance and Accountancy 1, 1-8. Grimpe, C., Hussinger, K., 2008. Pre-empting technology competition through firm acquisitions. Economic Letters 100, 189-191. Gugler, K., Mueller, D., Yurtoglu, B., Zulehner, C., 2003. The effects of mergers: an international comparison. International Journal of Industrial Organization 21, 625-653. Gupta, A., Misra, L., 2007. Deal size, bid premium, and gains in bank mergers: the impact of managerial motivations. The Financial Review 42, 373-400. Hackbarth, D., Morellec, E., 2008. Stock returns in mergers and acquisitions. Journal of Finance 63, 1213-1252. Halpern, P., 1982. Corporate acquisitions: a theory of special cases? A review of event studies applied to acquisitions. Journal of Finance 38, 297-317. Hogholm, K., 2016. Bidder’s gain in Public M&A transactions: does size matter? International Journal of Economics and Finance 8, 1-12.
Holmen, M., Nivorozhkin, E., Rana, R., 2009. Selection bias and event studies: the case of takeover likelihood and takeover premium. Unpublished working paper. University of Gothenburg, Gothenburg.
Institute for Mergers, Acquisitions and Alliances., 2018. M&A Statistics. Available at: https://imaa-institute.org/mergers-and-acquisitions-statistics/ [Accessed 2018:10-11]. Jensen, M., 1986. Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review 76, 323-329. Jensen, M., Ruback, R., 1983. The market for corporate control: The scientific evidence. Journal of Financial Economics 11, 5-50. Kolari, J., Pynnonen, S., 2010. Nonparametric rank tests for event studies. Unpublished working paper. Texas A&M University, College Station. Lang, L., Stulz, R., Walkling, R., 1989. Managerial performance, Tobin’s Q, and the gains from successful tender offers. Journal of Financial Economics 24, 137-154. Lang, L., Stulz, R., Walkling, R., 1991. A test of the free cash flow hypothesis. Journal of Financial Economics 29, 315-355.
Leland, H., Pyle, D., 1977. Informational asymmetries, financial structure, and financial intermediation. Journal of Finance 32, 371-387. Loughran, T., Vijh, A., 1997. Do long-term shareholders benefit from corporate acquisitions? Journal of Finance 52, 1765-90. MacKinlay, A., 1997. Event studies in economics and finance. Journal of Economic Literature 35, 13-39.
Moeller, S., Schlingemann, F., 2005. Global diversification and bigger gains: a comparison between cross-border and domestic acquisitions. Journal of Banking & Finance 29, 533-564. Moeller, S., Schlingemann, F., Stulz, R., 2004. Firm size and the gains from acquisitions. Journal of Financial Economics 73, 201-228. Myers, S., 1984. The capital structure puzzle. Journal of Finance 39, 575-592.
Myers, S., Majluf, N., 1984. Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics 13, 187-221.
O’Brien, R., 2007. A caution regarding rules of thumb for variance inflation factors. Quality & Quantity 41, 673-690
Park, N., 2004. A guide to using event study methods in multi-country settings. Strategic Management Journal 25, 655-668.
Rhodes, D., Stelter, D., 2009. Seize advantages in downturn. Harvard Business Review 87, 50-58. Rock, R., Rock, M., Sikora, M., 1994. The Mergers & Acquisitions Handbook. McGraw-Hill, New York. Schleifer, A., Vishny, R., 2003. Stock market driven acquisitions. Journal of Financial Economics 70, 295-311. Shah, P., Arora, P., 2014. M&A announcements and their effect on return to shareholders: an event study. Accounting and Finance Research 3, 170-190. Shimizu, K., Hitt, M., Vaidyanath, D., Pisano, V., 2004. Theoretical foundations of cross-border mergers and acquisitions: a review of current research and recommendations for the future. Journal of International Management 10, 307-353
Sridharan, U., Reinganum, M., 1995. Determinants of the choice of the hostile takeover mechanism: an empirical analysis of tender offers and proxy contests. Financial Management 24, 57-67. Sudarsanam, S., Mahate, A., 2003. Glamour acquirers, method of payment and post-acquisition performance: the UK evidence. Journal of Business Finance & Accounting 30, 299-341. Tobin, J., 1969. A general equilibrium approach to money theory. Journal of Money, Credit and Banking 1, 15-29. Trautwein, F., 1990. Merger motives and prescriptions. Strategic Management Journal 11, 283-295.
Travlos, N., 1987. Corporate takeover bids, methods of payment, and bidding firms’ stock returns. Journal of Finance 42, 943-963. Tuch, C., O’Sullivan, N., 2007. The impact of acquisitions on firm performance: a review of the evidence. International Journal of Management Reviews 9, 141-170. Walsh, J., 1989. Doing a deal: merger and acquisition negotiations and their impact upon target company top management turnover. Strategic Management Journal 10, 307-322.
Weston, F., Siu, A., Johnson, B., 2002. Takeovers, restructuring and corporate governance. Prentice Hall, New Jersey.
White, H., 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48, 817-38.
Wong, A., Cheung, K., 2009. The effects of mergers and acquisition announcements on the security prices of bidding firms and target firms in Asia. International Journal of Economics and Finance 1, 274-283. Wu, X., Yang, X., Yang, H., Lei, H., 2016. Cross-border mergers and acquisitions by Chinese firms: value creation or value destruction? Journal of Contemporary China 25, 130-145. Yermack, D., 1996. Higher market valuation of companies with a smaller board of directors. Journal of Financial Economics 40, 185-211.
Appendices
Appendix A Table 8. Results of the t-test and Corrado rank test w.r.t the daily average abnormal returns This table provides the daily p-values with regards to the results of the t-test and the Corrado (1989) rank test. The average abnormal returns are calculated on grounds of the constant-mean model. The total count of observations equals 606 acquisitions.Day #$%%%% in % t-test Corrado rank test
Table 9. Results of the (robust) non(parametric) significance tests w.r.t. the short-term event study
This table provides the p-values with respect to the significance of the average abnormal returns on the event day (day 0) and the cumulative average abnormal returns in three distinct event windows. The parametric tests consist of the cross-sectional t-test, the standardized t-test (Boehmer et al., 1991), and the adjusted standardized t-test (Kolari and Pynnonen, 2010). The nonparametric test consists of the Cowan generalized sign test (1992). The (cumulative) average abnormal returns are calculated on grounds of the constant-mean model. The count of observations equals 606 acquisitions.
Window (")$%&&&&&&&& in % t-test Std. t-test Adjusted std. t-test Cowan gen. sign test
