Corporate Social Responsibility and Acquisition Performance:
The bidder and target perspective on sustainable takeovers.
University of Groningen, Faculty of Economics and Business MSc Finance Thesis - Semester II 2019/2020
Supervisor: dr. A. Plantinga
Jan Elbers (S3597547) j.h.elbers@student.rug.nl
Date: 04.06.2020
Abstract
Many studies have researched the effect of CSR on firm performance. However, its role in the context of acquisitions is not very well researched yet. This study extends the literature on CSR by examining the relationship of CSR and acquisition performance from the bidder and target perspective. Recent studies find significant positive abnormal returns for acquiring shareholders in relation with CSR. This distorts the general picture on bidder announcement returns in M&A. To supplement the research area around CSR, this study combines the event study methodology with a multivariate regression. Other than most studies, it finds no significant positive relationship between bidder’s level of CSR and acquisition announcement returns. If any relationship exists, it is negative. It further shows that targets are negatively affected by bidder’s level of CSR.
Overall, a relationship between CSR and acquisition performance seems to prevail only under certain circumstances, where the perspective of bidding shareholders is more complex compared to target shareholders.
Word count (incl. tables, footnotes, and references): 11,092 JEL classification: G32, G34, M14
2 1. Introduction
Corporate social responsibility (CSR) has gained a lot of importance over the recent years. This can be seen by changes in regulation concerning non-financial disclosure (e.g. in Europe; see Directive 2014/95/EU)1 as well as by investors, who pay more attention to sustainable investments. According to the Global Sustainable Investment Review 20182 (GSIR), sustainable investment assets in the five major markets3 amounted to $30.7 trillion at the start of 2018. This is an increase of 34% in two years. These numbers prove the relevance of socially responsible investment (SRI) and introduce the question of how this impacts corporations. ESG integration, which describes the degree to which investors include environmental, social and governance matters in the financial analysis process, has become one of the most important sustainable investment strategies (GSIR, 2018).
Due to the growing importance of socially responsible investment, the aim of this study is to investigate the role of CSR in M&A deals, by examining the differences in value creation comparing high CSR firms (high ESG rating) and low CSR firms (low ESG rating). The reason why it is focused on M&A transactions, and specifically on acquisitions, is because it is one of the main growth strategies and can have huge impact on shareholder value (Deng, Kang, and Low, 2013). Furthermore, the relationship between CSR and M&A is not very well researched yet (Tampakoudis and Anagnostopoulu, 2020; Arouri, Gomes, and Pukthuanthong, 2019). Hence, this study contributes to the existing literature on CSR, by putting it in the context of acquisitions.
Since corporate social responsibility is a complex and far reaching concept, the increasing demand on assessing corporations’ activities and their impact on society, has led to several non-financial performance metrics that can be compared across companies. Being able to combine financial as well as non-financial performance measures allows investors to make better evaluations and predictions on firms’ performance. Both, in the short- and long- term. This is because risk exposures to environmental, social and governance matters can be assessed objectively through quantified metrics such as ESG.
Since CSR can be understood as a reflection of a firm’s interaction with society and certainly plays a role from a strategic perspective (see stakeholder theory), M&As conducted by firms with similar CSR activities are expected to be more successful than firms with contradictory perspectives. This intuition is derived from the perception that firms with similar CSR activities also share similar corporate values and hence face less challenges, internally and externally, during the different stages of a M&A transaction. Of course, success can be defined in various ways. The characteristic for success in this study is whether
1https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0095 2http://www.gsi-alliance.org/trends-report-2018/
3 acquiring and target shareholders earn significant abnormal returns (ARs) around the announcement and whether this is positively associated with a firm’s level of ESG.
Recent studies concerning CSR and M&A show that CSR oriented bidders earn significant positive abnormal announcement returns (Krishnamurti, Shams, Pensiero, and Velayutham, 2019), relish larger increases in post-merger long-term operating performance, and obtain positive long-term stock returns (Deng et al., 2013). They even need less time to complete the deal (Deng et al., 2013) and are less likely to fail (Arouri et al., 2019; Deng et al., 2013).
This finding contrasts with earlier studies on M&A, which show that on average the gains to acquiring shareholders tend to be negative (see, e.g., Aktas, de Bodt and Cousin, 2011; Andrade, Mitchel and Stafford, 2001; Aktas, de Bodt and Roll, 2004)4 or at most zero (see, e.g., Jensen and Ruback, 1983; Betton, Eckbo and Thorburn, 2008)5. Moeller, Schlingemann and Stulz (2005) find that value destruction for acquiring shareholders is related to large and public deals. Other reasons are provided by Bradley, Desai, and Kim. (1988), who find that competition among bidding firms increases the total synergistic gains, which however flow entirely to target shareholders. They show that acquirers experience an average gain that is statistically not different from zero. A side effect of competition is systematic overpaying as shown by Akdogu (2011) and hence embodies a further consequence for low bidder returns. The so-called free rider problem (Grossman and Hart, 1980) that encourages target shareholder not to tender their shares and the hubris hypothesis (Roll, 1986) arguing that targets gain through a wealth transfer from acquiring shareholders and not via synergies, are two other reasons that might explain why acquirers tend to break even on average. A further but short-lived observation related to low bidder returns is attributed to short selling of merger arbitrageurs and index fund rebalancing (Mitchell, Pulvino, and Satfford, 2004).
As shown by these examples, a wide range of literature on acquirer announcement returns exists. Although the literature seems to provide a clear picture on value creation for bidders, recent findings on the CSR-M&A relationship distort the picture, wherefore the main research question of this study is:
Does a firm’s level of CSR have a positive impact on acquirer’s and target’s abnormal announcement returns?
4The results of Andrade et al., 2001 and Aktas et al., 2004 are not significant at conventional levels.
5 Jensen and Ruback (1983) highlight that takeovers are value creating activities that are mainly driven
4 To investigate the existence of such a relationship, this study takes a closer look at the impact of CSR on the value creation of acquisitions. As most of the recent studies only focus on one aspect of CSR, such as corporate governance (Tampakoudis, Nerantzidis, Soubeniotis and Soutsas, 2018) or look at it holistically, this study evaluates the influence of the three individual components of ESG (environmental, social, and governance). To my knowledge only the studies of Arouri et al. (2019) and Xie, Nozawa, Yagi, Fujii, and Managi (2018) make this distinction. Furthermore, some studies only concentrate on national samples (Krishnamurti et al., 2019; Deng et al., 2013) or specific regions such as the EU and emerging market economies (Tampakoudis and Anagnostopoulu, 2020; Yen and Andre, 2019). Those that only focus on national samples cannot capture country effects, which is a disadvantage, since CSR practices are very different across countries due to distinctions in legal origins. (Liang and Renneboog, 2017). To account for country effects, and provide results that are as broad as possible, this study employs an international sample. Hence, by investigating the CSR-M&A relationship on an international sample and distinguishing between the three elements of CSR6, a more nuanced perspective is provided.
The question whether CSR activity is value creating or destructing links back to the stakeholder value maximization and shareholder expense view (see, e.g., Deng et al., 2013). The stakeholder value maximization view claims that by focusing on other stakeholders, firms’ operations gain more support, which increases shareholder value (Deng et al., 2013). This claim is supported by Porter’s hypotheses expounding that especially environmental activities are a source of innovation that create extra revenue and make up for the additional costs (Porter and van der Linde, 1995). Contrary, the shareholder expense view posits that investments in CSR are inefficient and transfer wealth from shareholders to stakeholders (Deng et al., 2013). The additional costs associated with CSR activity are not only caused by inefficient resource allocation, but also by agency problems that consequently put the firm at a competitive disadvantage (Friedman, 1970; Sternberg, 1997).
By relating CSR to acquisition performance it is hoped to provide an answer to these two perspectives and the associated agency problem. This helps corporate managers to make better strategic decisions in terms of sustainability and enhances the target selection process. The remainder of this paper is structured the following way. Section two provides the most recent insights on CSR. Section three displays the research methods and data used to test the hypotheses formulated in section two. Section four presents the empirical results, and sections five and six finish the paper by concluding and pointing out limitations.
6 There is no clear definition of CSR. However, literature and practice show that CSR and related
5 2. Literature review
There are many measures to assess the success of an acquisition. Most studies that investigate the relationship of CSR and M&A, assess the short- and/or long-term performance by calculating abnormal returns (see, e.g., Krishnamurti et al., 2019; Yen and Andre, 2019) or the change in Tobin’s Q (Tampakoudis and Anagnostopoulu, 2020). Some other studies take a different approach and look at measures such as the time it needs to complete the deal, the premiums that are being paid (Aktas et al., 2011; Deng et al., 2013) or how likely it is that the deal is completed (Arouri et al., 2019). This study measures the success by calculating the abnormal return at the announcement day of the acquisition and tests for the relationship with ESG.
6
H1: Higher levels of CSR lead to higher abnormal announcement returns for both acquirer and target.
By looking at the findings of Deng et al. (2013) and Tampakoudis and Anagnostopoulu (2020), it seems that the stakeholder value maximization view has more validity than the shareholder expense view. However, Yen and Andre (2019) provide evidence that none of the perspectives alone can explain the effect of CSR on M&A. They claim that the cost-benefit concerns of investors are driving the effect of CSR on announcement returns. Although they stress the positive potential of CSR, especially in cross border deals, the agency costs associated with CSR are the crucial underlying factor defining the signal of CSR in M&A announcements. This implies, CSR is only associated with value creation, where stake- and shareholder perspectives are aligned and are not a result of management entrenchment. Neither very high nor very low investments in CSR seem to be reasonable approaches to balance both perspectives and simultaneously minimize agency costs concerns. Consequently, a high level of information asymmetry could fuel the cost-benefit concerns of investors, and hence diminish the positive effects of CSR shown by Deng et al. (2013), Krishnamurti et al. (2019), and Tampakoudis and Anagnostopoulu (2020). Because of this, it is hypothesized that only moderate investments in CSR activities align the stake- and shareholder perspectives efficiently and lead to positive market reactions reflected by an increase in acquirer’s and target’s announcement returns. Therefore, the second hypothesis is:
H2: Only acquisitions by firms with moderate ESG ratings experience positive abnormal announcement returns.
7
H3: Firms with high financial performance prior to the acquisition announcement are not rewarded by high CSR performance.
The consideration of agency problems related to CSR seem to be an important factor influencing the effectiveness of CSR in M&A. Recent studies show that independent managers are more prone to extensive inefficient investments in CSR and thereby putting the company at a competitive disadvantage (Becchetti et al., 2008; Cespa and Cestone, 2007). Yen and Andre (2019) claim that an increase in ownership structure could mitigate the agency problem. This is in line with the findings of Arora and Dharwadkar, (2011) and Barnea and Rubin, (2010), constituting that not only strong governance mechanisms, but also an effective legal system needs to be in place to incentivize managers to balance the interests of insiders and outsiders. Since the ESG governance score, should consider potential agency problems, it is expected that higher governance scores signal lower agency problems. Because of this, the fourth hypothesis is:
H4: Governance has a positive impact on abnormal announcement returns and a higher signaling power than the environmental and social dimensions.
By answering all four hypotheses, this study contributes to the current literature in the following way. First, it delivers additional evidence towards the uncertain relationship between CSR and M&A from the acquirer and target perspective (H1). Second, it tackles the discussion around the stake- and shareholder theories on CSR, by investigating whether a more balanced approach towards CSR (moderate ESG score) outperforms acquisitions of companies with high/low ESG scores (H2). Third, it seizes the claim of Becchetti et al. (2008) and Groening and Kanuri (2013) by testing whether high financial performance prior the acquisition announcement diminishes the effect of CSR (H3). Fourth, it provides insights on the three dimensions of CSR, by specifically looking at governance, because it is most prone to agency problems once no effective mechanisms are in place (H4).
3. Data and methods
8 Tamapakoudis et al., 2018). The database is updated every two weeks. This update entails the addition of new companies, fiscal year updates or new controversy events. Where earlier studies had to resort to the ASSET4 database the introduction of Eikon came with an improved ESG rating process. Thomson Reuters now applies an ESG controversies overlay, industry and country benchmarks, size- and impact adjusted category scores, as well as an improved rank scoring methodology. It is considered as one of the most comprehensive ESG databases by drawing on 400 different ESG metrics dating back to 2002. The overall ESG score incorporates 10 categories whose performance is assessed based on publicly reported information. (Thomson Reuters, 2017)
Table 1: ESG dimensions and sub-categories
The overall score is based on 178 data points and grouped into ten categories.
Pillar Category Indicators in Scoring Weights
Environmental Resource Use Emissions Innovation 20 22 19 11% 12% 11% Social Workforce Human Rights Community Product Responsibility 29 8 14 12 16% 4.50% 8% 7% Governance Management Shareholders CSR Strategy 34 12 8 19% 7% 4.50% Total 178 100%
Source: Thomson Reuters
9 Table 2: Sample selection criteria
M&A deals are obtained from Thomson Reuters Eikon and filtered based on the five criteria below. Selection Criteria
Sample Size
Events Lost 1. Completed M&A deals between public firms over the
period 2000 - 2019 31,773 -
2. Acquirer needs to have an ESG rating 8,566 23,207
3. M&A deals where >50% of target’s equity is acquired 2,434 6,132
4. Only acquisitions (no mergers) 652 1,782
5. Acquirer needs to have an ESG rating one year prior
the announcement 290 392
The value that is created by the acquisition is measured with the event study methodology. Event studies are a popular tool in economics and finance to assess the effect of events on firm performance. The first paper using the event study methodology dates to Fama, Fisher, Jensen, and Roll (1969).
This study follows the findings of Campbell, Cowan and Salotti (2010) and applies the market model adjusted to national market indexes. They point out that in multi-country event studies this is the primary choice and that adapting to national market indexes delivers good results and does not require conversion of currencies.
Although the market model underlies quite strong assumptions, such as jointly multivariate normal and i.i.d. distributed returns, MacKinlay (1997) highlights that these assumptions are reasonable, and inferences tend to be robust. The market model follows a linear specification to estimate a normal return that is not influenced by an event. The normal returns are estimated over a period of 250 days prior the 61-day event window. Both windows are considered long enough to estimate event-unaffected market model parameters. To avoid biased estimates, only securities that at least exhibit 165 matched returns7 within the estimation window are included.
After estimating the normal returns for each security i, the abnormal returns within the event window are calculated based on an out of sample basis as shown in equation one.
𝐴𝑅𝑖𝑡 = 𝑅𝑖𝑡− 𝛼𝑖 − 𝛽𝑖𝑅𝑚𝑡 (1)
As important information might be leaked before or even after the announcement, it is of interest to investigate pre- as well as post-announcement effects. To control for such circumstances and make correct inferences, it is necessary to aggregate daily abnormal returns through time and across securities. Before relevant sub-windows can be created, though, first a cross-sectional aggregation on a daily and then on a multi-day level is performed.
7 A matched return of a security is a return that can be linked to a contemporaneous market return. A
10 To evaluate the significance of those returns, the event study literature offers a range of tests that account for different circumstances. Typically, a parametric and a nonparametric test are conducted to check results on robustness. Most important in a correct test selection are sample characteristics that determine the probability distribution of variables.
Parametric tests assume returns to be normally distributed, which is typically not the case. Cowan (1992) shows that stock returns and abnormal returns are highly non-normal. This non-normality is characterized by right skewness and leptokurtosis (see, e.g., Cowan 1992, Fama, 1976; Brown and Warner, 1985). Therefore, nonparametric tests might be preferred.
In line with Campbell et al. (2010) three tests are conducted. First, the standardized cross-sectional test, henceforth BMP test, of Boehmer, Musumeci and Poulsen (1991) is applied. In a second step, the generalized sign test (Cowan, Nayar, and Singh, 1990; Sanger & Peterson, 1990) and Corrado’s rank test (1989) are added. The advantages of the BMP test are that it does not require an insignificant variance increase around the announcement and that it corrects for the misspecification problem that the ordinary cross-sectional test might face in heterogeneous samples. The sample underlying this study is neither restricted by geographic location nor by industry, size, or trading volume. Hence, it is heterogeneous by construction, what needs to be accounted for. Boehmer et al. (1991) eliminate this problem by standardizing event period residuals and adjusting for the forecast error produced through the market model parameters.
To address variance increases around the announcement, they use the contemporaneous cross-sectional standard error in their test statistic and do not rely on standard errors from the estimation window. As acquisition announcements tend to happen independent form each other, event-date-clustering, and a possible violation of uncorrelated residuals across securities is rather unlikely.
The test statistic of the BMP test is given by equation two, where the standardized abnormal return (SAR) for security i is calculated by dividing the abnormal return at time t by the standard deviation of residuals in the estimation window.
𝑇𝑐 = 1 𝑁∑ 𝑆𝐴𝑅𝑖𝑡 𝑁 𝑖=1 √ 1 𝑁(𝑁−1)∑ (𝑆𝐴𝑅𝑖𝑡− ∑ 𝑆𝐴𝑅𝑖𝑡 𝑁 𝑁 𝑖=1 ) 2 𝑁 𝑖=1 ⁄ (2)
As mentioned above, the generalized sign and rank test are conducted to accommodate the BMP test.
11 increases, wherefore the BMP test provides a good robustness check under such conditions (Campbell et al., 2010).
It is further highlighted that the BMP test neither loses its power nor its specification under increased non-normality. Also, the generalized sign test continues to perform well, and with higher power than the BMP test on non-normal return distributions.
Overall, the tests do not only perform well under the mentioned conditions, they also complement each other’s weaknesses. Whether it is the BMP test under variance increases or the generalized sign and rank test under non-normality. Combining all three tests allows to neither sacrifice power and correct specification nor depart from conservatism and overstate results.
The test statistic for the generalized sign test is based on equation three, where parameter w represents the number of positive abnormal returns in the event window and 𝑝̂ the fraction of positive abnormal returns in the estimation window across all securities respectively.
𝑍𝐺 = 𝑤−𝑁𝑝̂
[𝑁𝑝̂(1−𝑝̂)]0.5 (3)
Corrado’s (1989) rank test statistic is outlined in equation four. The rank test considers the whole 311-day time series of abnormal returns, that is estimation and event window together, and assigns a rank to each of them. 𝐾̅𝐷 reflects the average rank across all securities in d days of the event window and 𝐾̅𝑡 the average rank across all securities on day
t of the 311-day time series.
𝑍𝑅 = 𝑑0.5 𝐾̅𝐷−156
[∑280(𝐾̅𝑡−156)2/311
𝑡=1 ]
0.5 (4)
Both the generalized sign and rank test are distributed asymptotically as unit normal. Contrary, the BMP test follows a student t-distribution.
After measuring the acquisition performance, a multivariate regression is conducted to test the impact of ESG on acquirer and target gains, respectively. However, acquisition performance is influenced by many factors, wherefore it is not enough to regress (cumulative) abnormal returns on ESG alone. Based on theory and previous literature, specific firm- and deal characteristics are chosen as control variables. Consequently, the following regression equation has been established.
𝐶𝐴𝑅𝑖 = 𝛽0+ 𝛽1𝐸𝑆𝐺𝑖+ 𝛽2𝑀𝐶𝑖+ 𝛽3𝑅𝑂𝐸𝑖 + 𝛽4𝐷𝑆𝑖+ 𝛽5𝑅𝑆𝑖 + 𝛽6𝐷𝑖 + 𝛽7𝐶𝑖 +
12 Apart from ESG, it is controlled for the size and the profitability of the bidder, the deal value, relative size, and several dummy variables. The dummy variables are accounting for industry relatedness, year, geographic region, and the method of payment. A detailed overview of the control variables is delivered in Table 3.
Table 3: Variable explanation
This table explains the dependent and independent variables used in the multivariate regression.
Variable Explanation
CAR (Cumulative abnormal return) Cumulative abnormal return of acquirer and target. For acquirer only the announcement returns at t = 0 and for targets the CARs over event period (-2, 1) are applied. This variable has been winsorized to obtain normally/nearly normally distributed residuals.
ESG (Environmental-, Social- and Governance scores)
Thomson Reuters Environmental, Social and Governance scores. Both the overall score as well as the individual dimensions are tested. The score is measured on a continuous scale between 0 and 100 and broken down into quartiles (see Table A11).
MC (Market capitalization) Natural logarithm of acquirer’s market capitalization one month prior the announcement.
ROE (Return on equity) Return on equity of acquiring firms one year before the announcement.
DS (Deal size) Natural logarithm of the deal size.
RS (Relative size) Relative size of the acquisition as measured by the fraction of deal size to acquirer’s market capitalization.
D (Diversification dummy) A dummy variable being 1 if a target of another industry is acquired.
C (Cash dummy) A dummy variable being 1 if the transaction was paid in cash. S (Share dummy) A dummy variable being 1 if the transaction was paid with
shares.
CB (Cross border dummy) A dummy variable being 1 if a target in another country is acquired.
R (Vector of region dummies) A vector of target’s geographic region as defined by the World Bank. These are East Asia & Pacific, Europe & Central Asia, Latin America & Caribbean, Middle East & North Africa, North America, South Asia, and Sub-Saharan Africa.
13 To obtain a robust OLS model, the primary concerns are related to heteroscedasticity and multicollinearity. To avoid heteroscedasticity, heteroscedasticity-consistent standard errors according to White (1980) are being applied. Multicollinearity is tested by an informal and a formal test. First, a correlation matrix of the independent variables is plotted to directly grasp possible problems. Second, the variance inflation factors (VIF) are calculated after each regression.8 The correlation matrix indicates no multicollinearity problem between the main explanatory variables, except of course for the overall ESG score and its individual dimensions. Because of this, the overall score is never combined with its individual dimensions within one model. For the more formal test, typical threshold values indicating a multicollinearity problem are VIFs of five and ten. For reasons of conservatism, only parameters with VIFs lower than five are included in the model. The parameter with the highest VIF is excluded first. If multicollinearity persists, the parameter with the next highest VIF coefficient is excluded. This process is followed until every parameter is below five. To further improve the model, dependent variables are winsorized and large continuous variables transformed into a logarithmic scale.9 This approach was preferred, because excluding observations might omit important information. Furthermore, as the sample is already heavily restricted by the availability of ESG ratings, losing observations through exclusion did not seem reasonable.
To get an idea about the sample, descriptive statistics are provided in Table 4. Panel A reports on all continuous- and Panel B on all dummy variables used in the multivariate regression. Further statistics on the sub-samples are illustrated in Table 5.
A quick look at these tables shows that most of the acquirers have ESG scores between 25 and 75 (quartiles B and C) and that these are the events where bidder and targets earn the highest average abnormal returns. Also, bidding firms with moderate financial performance prior the announcement earn positive ARs, while from the target perspective bidder with high financial performance create more value. Furthermore, most of the deals tend to be small and paid with cash. Diversifying and focus acquisitions are quite balanced but slightly geared towards diversification. The same can be observed for national versus cross border deals. Considering geographic regions, acquisitions are clustering in East Asia & Pacific and Europe & Central Asia. The yearly overview displays that most acquisitions in this sample are made between 2006 and 2012.
8 A correlation matrix is included in the appendix. The VIF coefficients are not separately reported. 9 ARs and CARs were typically winsorized at the 5th and 95th percentile. However, for the sub-samples
14 Table 4: Sample statistics
Panel A provides an insight on the continuous variables and Panel B on the dummy variables used in the multivariate regression. Type of payment displays a category “other”, which includes payments made with liabilities as well as those that could not be determined with certainty, where the latter constitutes the bigger portion.
Panel A: Continuous variables
Variable Max Min Med Ave
ESG 90.8 2.4 49.5 48.6
Market cap ($m) 237,177 168 9,498 24,937 Deal size ($m) 39,463.7 1.8 187.2 1,400.5
Relative size 5.8% 0.00% 0.02% 0.1%
Return on equity 181.8% -78.3% 11.1% 13.8%
Panel B: Dummy variables
Type of deal Observations Acquisitions
Diversification 150 Year Observations
Concentration 124 2019 9
Cross border 112 2018 11
National 155 2017 11
2016 9
Type of payment Observations 2015 14
Cash 159 2014 18
Share 21 2013 18
Other 94 2012 21
2011 31
Region Observations 2010 19
East Asia & Pacific 111 2009 20
Europe & Central Asia 123 2008 18
Latin America & Caribbean 8 2007 28
Middle East & North Africa 2 2006 30
North America 14 2005 13
South Asia 4 2004 2
Sub-Saharan Africa 3 2003 2
Table 5: Sub-sample statistics 1
Two categories of sub-samples have been created. The first category of sub-samples is based on bidder’s ESG scores, the second on bidder’s return on equity. The first category shows how many acquirers reside in each ESG quartile (A, B, C or D) and provides descriptive statistics for each of them.
15
Panel A: ESG sub-samples for acquirer
Abnormal return (%) Acquirer ESG
ESG
quartile Obs.
Max Min Med Ave
Max Min Med Ave
A 37 3.6 -7.3 0.02 -0.2 90.8 75.1 81.9 82.1
B 94 12.5 -13.8 0.1 0.3 74.8 50.2 60.5 61.8
C 92 6.6 -7.1 0.2 0.2 50.0 25.1 38.5 38.2
D 49 13.8 -11.9 0.05 -0.2 24.8 2.4 19.3 17.4
Panel B: ROE sub-sample for acquirer
Abnormal return (%) Acquirer ESG
ROE Obs. Max Min Med Ave Max Min Med Ave
High 85 5.3 -9.0 0.04 0.0 89.6 2.4 47.3 50.3
Medium 86 12.5 -11.9 0.4 0.5 90.8 6.1 53.9 50.0
Low 86 13.8 -13.8 -0.02 -0.2 87.2 5.0 49.9 47.7
Panel C: ESG sub-samples for target
Abnormal return (%) Acquirer ESG
ESG
quartile Obs.
Max Min Med Ave
Max Min Med Ave
A 25 147.1 -20.8 11.5 13.9 90.8 75.1 81.4 81.1
B 63 84.6 -44.7 13.0 17.8 73.9 50.2 63.5 62.5
C 63 128.2 -22.1 12.1 19.5 50.0 26.0 38.9 38.7
D 27 97.2 -5.6 10.4 16.5 24.8 6.1 19.3 18.0
Panel D: ROE sub-samples for target
Abnormal return (%) Acquirer ESG
ROE Obs. Max Min Med Ave Max Min Med Ave
High 60 147.1 -4.9 12.8 21.5 89.3 10.4 44.7 47.0 Medium 61 77.8 -20.8 10.2 15.6 90.8 6.1 49.8 48.7 Low 57 97.2 -44.7 18.0 15.9 87.2 10.7 54.4 54.3 4. Results 4.1 Event study
16 Panel A of Table A1 presents the average abnormal returns for bidding and Panel B for target firms. Overall, results show the picture provided by previous literature on M&A performance. Acquirers earn significant but economically not relevant abnormal returns of 0.111% at the announcement on average. Taking a closer look at the three-day CAR (-1, 1) reveals that abnormal returns become virtually zero (0.002%). This is in line with the observation that acquiring shareholders tend to break even on average (see, e.g., Jensen and Ruback, 1983; Betton et al., 2008).
Contrary, targets realize significant average abnormal returns of 9.31% at the announcement and between 15% and 19% over longer event periods. This illustrates that most of the value is transferred to target shareholders and confirms the findings of Aktas et al. (2004, 2011), Andrade et al. (2001) and Mulherin and Boone (2000), who find cumulative abnormal returns around 10%, 16% and 20%, respectively.
While acquiring firms only seem to earn significant abnormal returns at the announcement day, target shareholders benefit from significant returns already two days before until one day after the announcement. The results are robust as indicated by the parametric and nonparametric test statistics. The result for bidders is somewhat less significant with levels of 10% and 1% compared to 1% levels across all three tests for target shareholders.
Table 6: Snapshot of abnormal return statistics for bidder
Average abnormal returns and the corresponding significance level have been calculated for event window (-30, 30). Only a snapshot is presented here. Statistics on the whole event window can be obtained in Table A1 in the appendix. Three statistical tests have been employed. The standardized cross-sectional test (TC,) the generalized sign test (ZG), and the rank test (ZR). Additionally, statistics on
variance, kurtosis and skewness are provided. Significance is reported for 10% (*), 5% (**) and 1% (***) levels.
t N AAR (%) TC ZG ZR Var Kurt Skew
17 For target firms, several sub-windows, ranging from two-day to six-day CARs, have been tested. This is because some days before and after the announcement show high significant results that are not robust, however. The results of the sub-windows are shown in Table 7. Not surprisingly, but reassuring, all CARs are highly significant at the 1% level across all test statistics. Even if the six-day CAR is highly significant, it was decided to choose the CAR over event period (-2, 1). This is because all days over this period are also individually proven significant across at least two tests.
Table 7: Cumulative abnormal return statistics for target
Sub-windows ranging from two- to six-day CARs have been tested. TC, ZG, and ZR represent the test
statistics for the standardized cross-sectional, generalized sign, and rank test, respectively. Additionally, statistics on variance, kurtosis and skewness are being reported. Significance is highlighted for 10% (*), 5% (**) and 1% (***) levels.
t CAAR (%) TC ZG ZR Var Kurt Skew
(-3, 2) 18.675 10.079*** 9.897*** 2.654*** 621.883 5.728 1.736 (-2, 2) 18.300 9.835*** 9.897*** 2.968*** 627.277 5.797 1.790 (-2, 1) 16.917 9.468*** 9.751*** 3.728*** 580.269 7.099 2.056 (-1, 2) 17.378 9.358*** 8.283*** 3.168*** 627.671 6.055 1.863 (-1, 1) 15.995 9.011*** 8.429*** 4.248*** 574.579 7.591 2.170 (0, 2) 16.647 8.976*** 8.283*** 3.835*** 618.749 6.226 1.919 (0, 1) 15.264 8.610*** 8.429*** 5.788*** 566.008 7.924 2.248 4.2 Multivariate regression
The multivariate regression consists of three models that were applied to the whole sample as well as sub-samples. These models built upon each other by adding control variables to the preceding equation. Model 1 is the simplest and just regresses the (cumulative) abnormal announcement return on the variable of interest (either the overall ESG score or its individual dimensions). Model 2 adds the main control variables, and Model 3 supplements year dummies.10 To get a better understanding on how ESG is received by the market in the context of acquisitions, two categories of sub-samples with four and three sub-groups, respectively, are created. The first sub-sample category is based on acquirer’s ESG ratings and serves the purpose of assessing whether a nonlinear relationship between ESG and announcement returns exists (hypothesis two).11 The second is based on acquirer’s return on equity to assess the possibility of decreasing marginal returns from SRI as claimed by hypothesis three.
10 The main control variables are defined as the firm- and deal specific characteristics. Year dummies
control for macroeconomic conditions. Regressions on target firms, exclude region dummies as those are irrelevant from their perspective.
11Thomson Reuters ESG scores are categorized into quartiles A, B, C, and D. A breakdown of these
18
4.2.1 Results hypothesis one
The regression results show that the overall score does not have any influence on the value creation for bidding shareholders. All three regression models show insignificant results for the ESG coefficient and therewith support the findings of Chen and Gavious, (2015), Aktas et al. (2011), McWilliams and Siegel (2001) and Moore (2001). However, other control variables reveal significant results. Models 2 and 3 of Table A3 show that relative size, geographic location as well as the year of the announcement positively impact bidding announcement returns. Model 3 predicts a 0.65% increase in abnormal announcement returns per unit increase in relative size. This means that the interplay of being a large firm and paying low prices has a positive effect on acquisition performance, which is not surprising since relative size counts to the most important factors driving acquirer’s announcement returns (see, e.g., Betton et al., 2008). The analysis of the three individual dimensions environmental, social and governance, reveals the same result as the overall score.
Target announcement returns are significantly negative affected by the overall ESG score as well as by deal size. While the ESG score imposes a rather weak impact on targets’ CARs with 0.190% per unit change12, a one unit increase in the deal size almost decreases the CAR by 2% on average. Contrary to the overall score, none of the individual dimensions alone exhibits a significant impact on targets’ CARs and hence mirrors the finding for acquirers. A significant constant in Model 1 exposes that, independent of acquires ESG rating, targets earn significant CARs of 18% on average.
4.2.2 Results hypothesis two (ESG sub-samples)
Because hypothesis two of this study expects a nonlinear relationship between ESG and (cumulative) abnormal returns, a more nuanced approach is applied by splitting the sample into four sub-groups that are based on the ESG quartile ratings.
The analysis of the different groups indicates an interesting result. Acquisition performance of firms in the high ESG group (quartile A) do not seem to be impacted by ESG. While Model 2 in Table 6 predicts a significant negative impact of -0.062% per unit increase in the governance dimension, Model 3 reports the respective coefficient as insignificant. Obviously, the year effect drives this result and seems to be more important in acquisitions than the degree of a firm’s level of governance. The same applies to the bidding firm’s financial performance prior the announcement. While being significant and positive when ignoring year effects, it becomes insignificant when year dummies are introduced. Instead of governance and financial performance, size and deal type become significant
12 The overall ESG score can obtain values between 0 and 100 on a continuous scale. Assuming an
19 factors. This result shows that the acquisition performance of firms with high ESG ratings is mainly driven by size and deal type. The size effect considered alone is not necessarily in line with theory, because high valued firms tend to do inferior acquisitions compared to small firms (see, e.g., Dong, Hirshleifer, Richardson, and Teoh, 2002). However, it is more difficult to explain the significant positive effect of cross border deals. This result contradicts with the home bias theory. Nevertheless, it makes sense when thinking about large mature firms that need to expand to new markets to prolong their growth face, which could be one plausible explanation that is in line with the growth opportunities signaling theory (McCardle and Viswanathan, 1994; Jovanovic and Braguinsky, 2002). Looking at the average market capitalization of firms within the different ESG categories reveals that indeed those firms with higher ESG ratings tend to be bigger than those in other categories. Since size alone is of course not a sign for maturity, it does indicate a potential truth in the expectation formulated above.13
Model 2 further reveals a negative impact of transactions paid with shares. This is not unexpected regarding signaling theories, predicting that firms issue shares when overvalued and pay cash otherwise. Overall, the result for high ESG firms is not entirely clear, but when trusting Model 3, other factors such as size, type of deal, the target’s location and economic conditions, as proxied by the year dummies, seem to matter more. If governance has an impact it is very little and of negative sign.
Contrary to firms with a very high ESG rating, the group with B ratings provides a clearer result regarding ESG. Across all three models, the environmental dimension reveals a negative significant result at the 5% level and the effect increases from Model 1 to Model 3. Other than that, relative size and cash payment exhibit significant positive effects. The fact that the A group reveals a significant negative effect for share payments and the B group a significant positive effect for cash payments is not only in line with signaling theory, but also with the perception that large firms tend to be overvalued more often compared to small firms (Moeller, Schlingemann, and Stulz, 2004).
Groups C and D do not reveal any significance for the individual ESG dimensions. The C group shows a significant negative result for deal size and a significant positive result for bidder’s financial performance. The latter, however, seems to be driven by adding irrelevant control variables to Model 3. A relevant deal characteristic that has a significant impact on low ESG firms is the method of payment. Transactions paid with shares exert a negative impact of -3.5% on bidder’s announcement returns on average. Again, this is in line with signaling theory.
The target perspective is similar to that of acquirers but reveals one essential difference. While for acquirers the environmental dimension exhibits a significant negative result, targets acquired by firms with B ratings are exposed to a negative association with
13The average market capitalization measured in millions of USD of firms in quartile A amounts to
20 the overall ESG score. Furthermore, this association is much stronger with -0.988% compared to the result obtained for the whole sample. Nevertheless, the significant effect of ESG on CAR for targets acquired by B rated firms vanishes for acquisitions by C rated firms. Unfortunately, the sample sizes for targets whose bidder belong to the A or D group are too small to conduct a proper regression analysis. To provide an insight on these groups, Table 5 lists some descriptive statistics. Comparing target’s CARs between these four categories shows that targets acquired by firms of category C earn on average the highest CAR with 19%. Those acquired by firms with an A, B, or D rating earn 14%, 18%, and 17%, respectively. This hints towards the fact that acquisitions by small firms with low ESG ratings create more value than those of large firms with high ratings. This observation is in line with the shareholder expense view as well as the size effect frequently observed in M&A.
Striking is that targets acquired by firms in ESG quartile B reveal significant constants for all three regression models considering the overall ESG score (Table A6). The coefficients exhibit that targets earn significant positive CARs around 35% irrespective of the acquirer CSR level. Adding control variables increases target’s CARs to 55% on average. This is extreme but as predicted by the model only the case for focus acquisitions within the same country, holding everything else equal and factoring in region and year fixed effects. Table A2 shows that most of the acquisitions are indeed within a nation’s borders but tend to be diversifying. Specifically, for targets acquired by firms in quartile B the relations are 62:38 (same country – cross border) and 65:35 (diversification – focus). A closer look at the distributions of the cumulative abnormal returns for diversifying and focus acquisitions reveals that the extreme return is probably a result of right skewness and an underrepresentation of focused acquisitions.
4.2.3 Results hypothesis three (ROE sub-samples)
Because of the expectation of decreasing marginal returns from investing in SRI, a second sub-sample consisting of three categories has been created. The sample has been split into high, medium, and low return on equity firms to differentiate between the levels of firms’ financial performance prior the announcement.
21 the ROE coefficient itself crystallizes as a significant factor. Striking is however the sign of the coefficients. It is negative for high ROE firms and positive for moderate ROE firms. A further noticeable result is that for firms with very high financial performance, size and ROE are both significant at the 10% level and negative, while relative size is significant at the 1% level and positive. This shows that firms with high financial performance are rewarded by making good acquisitions that are characterized by big spreads in bidders’ size and targets’ value. Essentially, this translates into buying cheap. Hence, firms that already perform financially well prior the announcement are not rewarded by further increasing performance nor by increasing size, but by paying low acquisitions prices.
Significant constants in models two and three (Table 7) show that after adding control variables, acquirer seem to earn quite high abnormal returns (between 3.5% and 4.9%) when doing focused acquisitions within the same country. The constant also reflects the control group for the payment method. Unfortunately, no concise statement can be made here, because the control group constitutes a big portion of deals that could not be assigned to either cash- or share payment (see Table A2 payment “other”). According to theory, however, cash payments are more likely. Assessing whether significance is driven by outliers requires to look at the frequencies of the respective control groups. Table A2 shows that focused acquisitions within the same country are quite balanced with their counterparts (diversifying cross border acquisitions), which reduces the likelihood that potential outliers or a numerical mismatch between control and treated groups cause significance. Hence, ceteris paribus, financially healthy firms that do focused acquisitions within the same country, earn significant positive returns.
The same observation, with even higher average abnormal returns, can be made for firms with moderate financial performance prior the acquisition. Yet, the exception here is that year effects reverse the sign of the constant, highlighting the strong effect of macroeconomic conditions persistent within the respective years.
This finding is a further indication that bidding shareholders only earn significant and economically relevant abnormal returns when the conditions described above come together. Obviously, this is rarely the case as proven by this as well as many other studies investigating acquirer announcement returns.
22 announcement returns (Betton et al., 2008), it becomes a detrimental factor for targets when bidder experience poor financial performance prior the acquisition announcement.
The regressions further uncover that target shareholders earn significant CARs between 16% and 21% across all profitability groups when ESG as well as all other control variables are ignored.
Table 6: Bidder regression results - ESG sub-samples (hypothesis two)
Cardinals (1), (2), and (3) represents the three models that were estimated per ESG quartile rating. The numbers in parentheses report robust standard errors and the stars indicate significance at the 10% (*), 5% (**) and 1% (***) levels. The sub-sample regressions were only performed on the individual dimensions as the overall score did not reveal any significant result throughout the analysis. Region FE and Year FE are controlling for target region and year fixed effects, respectively.
A Rating B Rating VARIABLES (1) (2) (3) (1) (2) (3) Constant 1.471 (3.079) 0.887 (4.673) -9.437 (8.140) 0.713 (1.207) -0.020 (1.847) -0.468 (2.058) Environmental 0.021 -0.001 0.051 -0.014** -0.017** -0.021** (0.023) (0.025) (0.034) (0.006) (0.007) (0.010) Social -0.027 -0.042 -0.022 0.008 0.004 0.006 (0.035) (0.047) (0.052) (0.011) (0.012) (0.013) Governance -0.013 -0.062** -0.019 -0.002 -0.013 -0.014 (0.020) (0.025) (0.027) (0.011) (0.012) (0.013) Market cap 0.513 0.677** 0.130 0.166 (0.311) (0.266) (0.174) (0.181) Return on equity 0.017** -0.007 -0.000 0.001 (0.007) (0.008) (0.008) (0.010) Deal size 0.211 (0.235) 0.005 (0.115) -0.011 (0.140) Relative size -0.436 (3.362) 0.785 (2.785) 2.954*** (0.956) 2.767** (1.212) Diversification 0.165 1.094 -0.054 0.169 (0.692) (0.793) (0.320) (0.368) Cash 0.084 -0.357 0.481 0.766** (0.529) (0.604) (0.373) (0.346) Shares -3.746*** (1.028) -0.013 (0.806) 0.110 (0.921) Cross border 0.941 1.657** 0.163 0.487 (0.787) (0.699) (0.386) (0.438)
Region FE No Yes Yes No Yes Yes
Year FE No No Yes No No Yes
Observations 37 36 36 94 89 89
23 Table 7: Bidder regression results - ROE sub-samples (hypothesis three)
Cardinals (1), (2), and (3) represents the three models that were estimated per category. The numbers in parentheses report robust standard errors and the stars indicate significance at the 10% (*), 5% (**) and 1% (***) levels. The sub-sample regressions were only performed on the individual dimensions as the overall score did not reveal any significant result throughout the analysis. Region FE and Year FE are controlling for target region and year fixed effects, respectively.
High ROE Medium ROE Low ROE
VARIABLES (1) (2) (3) (1) (2) (3) (1) (2) (3) Constant 0.036 (0.450) 3.515** (1.539) 4.851** (1.933) -0.059 (0.479) 4.907*** (1.785) -4.631** (1.942) 0.028 (0.541) -0.994 (2.080) -2.788 (2.366) Environmental -0.009 -0.007 -0.009 0.005 -0.007 -0.015 -0.009 -0.019* -0.014 (0.009) (0.008) (0.009) (0.008) (0.009) (0.012) (0.009) (0.011) (0.013) Social 0.001 0.012 0.017 0.005 0.006 0.008 -0.011 -0.012 -0.024* (0.010) (0.010) (0.011) (0.009) (0.011) (0.013) (0.011) (0.013) (0.014) Governance 0.007 0.010 0.016* 0.000 -0.002 0.000 0.014 0.027** 0.028** (0.008) (0.008) (0.009) (0.008) (0.007) (0.008) (0.012) (0.011) (0.012) Market cap -0.357* (0.183) 0.555** (0.212) 0.375 (0.228) 0.407 (0.264) 0.139 (0.290) 0.351 (0.272) ROE -0.008* -0.007 0.172** 0.158* 0.010 0.030 (0.005) (0.006) (0.074) (0.087) (0.020) (0.030) Deal size -0.116 (0.103) -0.120 (0.104) -0.147 (0.154) -0.068 (0.154) -0.189 (0.204) -0.296* (0.175) Relative size 0.607*** (0.201) 0.566** (0.268) 2.864*** (1.051) 2.110** (1.008) -0.538 (1.586) 0.436 (1.497) Diversification -0.026 0.042 -0.437 -0.478 0.302 0.487 (0.350) (0.380) (0.344) (0.361) (0.370) (0.438) Cash -0.012 -0.185 0.667* 0.607 0.401 0.233 (0.336) (0.373) (0.374) (0.411) (0.586) (0.565) Shares 0.403 0.578 -0.299 0.269 -1.320* -1.269 (0.473) (0.616) (0.775) (0.746) (0.711) (0.828) Cross border -0.341 -0.165 -0.178 -0.013 -0.442 -0.687 (0.383) (0.443) (0.417) (0.463) (0.597) (0.508)
Region FE No Yes Yes No Yes Yes No Yes Yes
Year FE No No Yes No No Yes No No Yes
Observations 86 85 85 86 82 82 85 84 84
Adjusted R2 -0.008 0.026 0.002 -0.014 0.123 0.223 0.013 0.166 0.313
4.2.4 Results hypothesis four
Hypothesis four expects governance to be the most important dimension. From the bidder perspective, it is not entirely clear whether governance plays a role in acquisitions at all. Given the results, governance emerges as significant only for firms with high ESG ratings and high financial performance. However, the first one becomes insignificant when year dummies are added and the second raises issues because of a poorly fit model and a low significance level.
From the target perspective, the individual dimensions never reveal significant results. One exception exists for targets acquired by firms with B ratings. Here the social dimension exhibits a weakly significant result with -0.378%.
24 higher signaling power nor exposes a positive relationship with bidder or target abnormal returns.
5. Conclusion
This paper investigates the relationship between CSR and acquisition performance. Recent studies on the CSR-M&A relationship show that acquirer announcement returns are positively influenced by a firm’s level of CSR. Essential to the topic of corporate social responsibility are stake- and shareholder theories. Two broad categories exist that are described as the shareholder expense and stakeholder value maximization view. Shareholders apprehend investments in CSR as a wealth transfer to stakeholders, while stakeholders consider it a forward-looking value enhancing investment in society’s interest. Which of these perspectives prevails in the context of acquisitions is assessed via a two-step analysis that measures acquisition performance by means of an event study in a first step, and then assesses the influence of CSR in a second step. After controlling for firm- and deal specific characteristics, the following can be concluded:
• Acquirer’s level of CSR does not exhibit a positive effect on bidder announcement returns but poses a significant negative effect on target shareholders.
• A nonlinear relationship as expected according to hypothesis two cannot be observed. The environmental dimension is found to exhibit a significant negative impact on bidding shareholders and the overall ESG score on target shareholders. However, only for firms with ESG ratings in quartile B.
• The assumption that financially healthy firms are not rewarded by investments in ESG can neither be denied nor affirmed. Investors of financially healthy firms seem to be indifferent when it comes to CSR as reflected by insignificant results for both acquirer and targets.
• The governance dimension does not carry more signaling power than the other two dimensions. Which dimension prevails seem to depend on firm-, deal- and macroeconomic characteristics.
The findings reveal that CSR is a complex concept and as shown by this study only of relevance in very specific circumstances. Firm- and deal characteristics as well as macroeconomic conditions predominantly drive acquisition performance for bidding and target shareholders. If any relationship between CSR and announcement returns exists, it is most likely negative and hence, in line with shareholder theory. However, for the most part, investors seem to believe that CSR is not a value enhancing factor in acquisitions.
25 acquisitions target firms remain separate entities, the strategic focus might be completely different as opposed to mergers where firms become one entity. CSR effects such as “learning” described by Aktas et al. (2011) might become of importance in mergers but not in acquisitions.
Because of this, further research could be devoted to the question whether CSR poses a significant different effect in mergers compared to acquisitions and whether the different strategic outlooks (merger or acquisition) can explain inconsistencies in the CSR-M&A relationship.
6. Limitations
The fact that CSR is a complex and far reaching concept makes it difficult to measure its impact on acquisition performance. The ESG database of Thomson Reuters is considered as one of the most complete ones, wherefore it was chosen in this study to proxy for differences in firms’ level of CSR. Unfortunately, even if it is one of the largest databases, at the point of this study only 7,220 firm ratings were available. Due to additional selection criteria imposed on the data, a rather small sample remained. This might lead to biases in the true relationship of CSR and acquisition performance.
Furthermore, there is a tendency that rather large firms are rated on ESG, wherefore the sample lacks some randomness. Additional consequences are missing ESG ratings for the respective target firms, which might cause an endogeneity problem. To mitigate this problem, geographic region dummies were introduced. The idea is to account for cultural differences and attitudes that might be related with a firm’s perception on CSR dimensions. Studies on CSR find significant relationships with national cultures, wherefore this approach is considered the best alternative without losing additional observations.
The fact that target firms are much smaller and not that frequently traded has the consequence of less robust data. Some of them are lacking consistent price data, wherefore a minimum requirement for matched returns within the estimation window was introduced. This minimum value was set to 165. Of course, this has led to the exclusion of further events especially form the target perspective. Whether the selection criteria were too restrictive is difficult to judge. The criteria are certainly arguable, while no guarantee for better results. The maxim was to allow losing observations for potentially more powerful results.
26 7. References
Akdogu, E., 2011. Value-maximizing managers, value-increasing mergers and overbidding. Journal of Financial and Quantitative Analysis 46(1), 83-110
Aktas, N., De Bodt, E., J. G., Cousin, 2011. Do financial markets care about SRI? Evidence from mergers and acquisitions. Journal of Banking and Finance 35(7), 1753-1761.
Aktas, N., de Bodt, E., R., Roll, 2004. Market response to European regulation of business combinations. Journal of Financial and Quantitative Analysis 39(4), 731-757.
Andrade, G. Mitchell, M., E., Stafford, 2001. New evidence and perspectives on mergers. Journal of Economic Perspectives 15(2), 103-120.
Arora, P., Dharwadkar, R., 2011. Corporate governance and corporate social responsibility (CSR): The moderating roles of attainment discrepancy and organization slack. Corporate Governance: An International Review 19(2), 136-152.
Arouri, M., Gomes, M., K., Pukthuanthing, 2019. Corporate social responsibility and M&A uncertainty. Journal of Corporate Finance 56, 176.198.
Barnea, A., Rubin, A., 2010. Corporate social responsibility as a conflict between shareholders. Journal of Business Ethics 97(1), 71-86.
Becchetti, L., Di Giacomo, S., D., Pinnacchio, 2008. Corporate social responsibility and corporate performance: Evidence from a panel of US listed companies. Applied Economics 40(5), 541-567.
Betton, S., Eckbo, B.E., K. S., Thorburn, 2008. Corporate takeovers. In: B. E. Eckbo (Ed.), Handbook of corporate finance: Empirical corporate finance, vol. 2. North Holland, Amsterdam, pp. 291-429.
Boehmer, E., Musumeci, J., A. B., Poulsen, 1991. Event study methodology under conditions of event-induced variance. Journal of Financial Economics 30(2), 253-272.
Bradley, M., Desai, A., E. H., Kim, 1988. Synergistic gains from corporate acquisitions and their division between the stockholders of target and acquiring firms. Journal of Financial Economics 21(1), 3-40.
Brown, S. J., Warner, J. B., 1985. Using daily stock returns: The case of event studies. Journal of Financial Economics 14(1), 3-31
Campbell, C. J., Cowan, A. R., V., Salotti, 2010. Multi-country event-study methods. Journal of Banking and Finance 34(12), 3078-3090.
Cespa, G., Cestone, G., 2007. Corporate social responsibility and managerial entrenchment. Journal of Economics and Management Strategy 16(3), 741-771.
27 Cheung, Y., Tan, W., Ahn, H. J., Z., Zhang, 2010. Does corporate social responsibility matter in Asian emerging markets? Journal of Business Ethics 92(3), 401-413.
Corrado, C. J., 1989. A nonparametric test for abnormal security-price performance in event studies. Journal of Financial Economics 23(2), 385-395.
Cowan, A. R., 1992. Nonparametric event study tests. Review of Quantitative Finance and Accounting 2(4), 343-358.
Cowan, A. R., Nayar, N., A. K., Singh, 1990. Stock returns before and after calls of convertible bonds. Journal of Financial and Quantitative Analysis 25(4), 549-554.
Deng, X., Kang, J., B. S., Low, 2013. Corporate social responsibility and stakeholder value maximization: Evidence form mergers. Journal of Financial Economics 110(1), 87-109. Directive 2014/95/EU of the European Parliament and of the Council. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0095, last accessed 29.05.2020.
Dong, M., Hirshleifer, D., Richardson, S., S. H. Teoh, 2002. Does investor misevaluation drive the takeover market? Unpublished working paper, The Ohio State University, Columbus, OH. Fama, E. F., 1976. Foundations of Finance. Basic Books, New York.
Fama, E., Fisher, L., Jensen, M., R. Roll, 1969. The adjustment of stock prices to new information. International Economic Review 10(1), 1-21.
Friedman, M., 1970. The social responsibility of business is to increase its profits. In: Zimmerli W. C., Holzinger M., Richter K. (Ed.), Corporate Ethics and Corporate Governance. Springer, Berlin, Heidelberg, pp. 173-178.
Global Sustainable Investment Review. http://www.gsi-alliance.org/trends-report-2018/, last accessed 29.05.2020.
Groening, C., Kanuri, V. K., 2013. Investor reaction to positive and negative corporate social events. Journal of Business research 66(10), 1852-1860.
Grossman, S. J., Hart, O. D., 1980. Takeover bids, the free-rider problem, and the theory of the corporation. Bell Journal of Economics 11(1), 42-64.
Jensen, M. C., Ruback, R. S., 1983. The market for corporate control: The scientific evidence. Journal of Financial Economics 11(1), 5-50.
Jovanovic, B., Braguinsky, S., 2002. Bidder discounts and target premia in takeovers. NBER Working Paper 9009, NBER, Cambridge, MA.
Krishnamurti, C., Shams, S., Pensiero, D., E. Velayutham, 2019. Socially responsible firms and mergers and acquisitions performance: Australian evidence. Pacific-Basin Finance Journal 57, 101193.
28 MacKinlay, A. C., 1997. Event studies in economics and finance. Journal of Economic Literature 35(1), 13-39.
Margolis, J. D., Walsh, J. P., 2003. Misery loves companies: Rethinking social initiatives by business. Administrative Science Quarterly 48(2), 268-305.
McCardle, K.F., Viswanathan, S., 1994. The direct entry versus takeover decision and stock price performance around takeovers. Journal of Business 67(1), 1-43
McWillimas, A., Siegel, D., 2001. Corporate social responsibility: A theory of the firm perspective. Academy of Management Review 26(1), 117-127.
Mitchell, M., Pulvino, T., E., Stafford, 2004. Price pressure around mergers. Journal of Finance 59(1), 31-63.
Moeller, S. B., Schlingemann, F. P., R. M., Stulz, 2004. Firm size and the gains from acquisitions. Journal of Financial Economics 73(2), 201-228.
Moeller, S.B., Schlingemann, F.P., R. M., Stulz, 2005. Wealth destruction on a massive scale? A study of acquiring-firm returns in the recent merger wave. Journal of Corporate Finance 60(2), 757-782.
Moore, G., 2001. Corporate social and financial performance: An investigation in the U.K. supermarket industry. Journal f Business Ethics 34(3-4), 299-315.
Mulherin, J., Boone, A., 2000.Comparing acquisitions and divestitures. Journal of Corporate Finance 6(2), 117-139.
Porter, M. E., van der Linde, C., 1995. Toward a new conception of the environment-competitiveness relationship. Journal of Economic Perspectives 9(4), 97-118.
Roll, R., 1986. The hubris hypothesis of corporate takeovers. The Journal of Business 59(2), 197-216.
Sanger, G. C., Peterson, J. D., 1990. An empirical analysis of common stock delistings. Journal of Financial and Quantitative Analysis 25(2), 261-272.
Sternberg, E., 1997. The defects of stakeholder theory. Corporate Governance: An International Review 5(1), 3-10.
Tampakoudis, I., Anagnostopoulu, E., 2020. The effect of mergers and acquisitions on environmental, social and governance performance and market value: Evidence from EU acquirers. Business Strategy and the Environment, 1-11.
Tampakoudis, I., Nerantzidis, M., Soubeniotis, D., A., Soutsas, 2018. The effect of corporate governance mechanisms on European mergers and acquisitions. Corporate Governance 18(5), 965-986.
Thomson Reuters ESG Scores.
29 White, H., 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48(4), 817-838.
Xie, J., Nozawa, W., Yagi, M., Fujii, H., S., Managi, 2018. Do environmental, social, and governance activities improve corporate financial performance? Business Strategy and the Environment 28(2), 286-300.
30 8. Appendix
Table A1: Bidder and target abnormal return statistics
Average abnormal returns and the corresponding significance level have been calculated for the event window (-30, 30). Three statistical tests were conducted. The standardized cross-sectional test (TC), the
generalized sign test (ZG), and the rank test (ZR). Additionally, statistics on variance, kurtosis and
skewness are provided. Significance is reported for 10% (*), 5% (**) and 1% (***) levels. Panel A illustrates the results for the bidder and Panel B for the target.
Panel A: Average abnormal returns of bidder
t N AAR (%) TC ZG ZR Var Kurt Skew
31 7 274 0.091 0.578 -0.108 -0.213 5.549 104.753 7.896 8 274 -0.033 -0.366 0.134 -0.542 3.559 18.644 0.500 9 274 0.043 0.412 0.739 0.046 2.210 5.246 0.799 10 274 0.019 1.110 0.739 0.527 1.920 3.126 -0.103 11 274 -0.119 -1.430* -0.592 -1.847** 2.259 9.272 0.666 12 274 -0.193 -1.191 0.739 -1.223 3.175 29.288 -3.263 13 274 -0.140 -1.251 -1.075 -1.700** 3.539 11.736 -0.553 14 274 0.017 -0.184 -0.350 -1.035 4.554 14.876 0.284 15 274 -0.060 0.008 0.618 -0.155 2.128 5.150 0.418 16 274 -0.101 0.151 -1.196 -0.978 2.475 12.171 -0.725 17 274 -0.002 0.461 1.464* 0.020 2.591 2.774 -0.095 18 274 0.174 0.866 1.464* 0.829 4.013 30.588 3.521 19 274 -0.051 -1.031 -0.471 -1.251 2.086 11.542 0.176 20 274 0.091 1.096 0.860 0.440 2.220 5.358 0.401 21 274 -0.010 -0.555 0.134 -0.453 1.735 1.895 -0.233 22 274 -0.016 0.174 0.013 -0.584 1.813 2.847 0.112 23 274 0.048 0.485 1.464* -0.051 1.962 4.323 0.541 24 274 0.035 0.124 1.948** 0.135 2.201 5.188 0.541 25 274 -0.088 -0.484 0.255 -0.717 2.881 14.496 -1.806 26 274 -0.114 -1.120 0.739 -0.756 2.161 7.075 -0.842 27 274 0.010 -0.163 0.013 -0.283 2.205 2.927 -0.019 28 274 0.038 -0.200 -0.713 -0.757 1.746 4.825 0.695 29 274 -0.040 -0.027 0.255 -0.529 3.248 5.207 -0.616 30 274 -0.028 -0.507 -0.592 -1.087 2.493 4.713 0.599
Panel B: Average abnormal returns of target
t N AAR (%) TC ZG ZR Var Kur Skew
33 Table A2: Sub-sample statistics 2
These statistics represent the relations of control- and treated groups for the three dummy variables “diversification”, “cross border” and “method of payment”. They help to better understand whether an under- or overrepresentation exists and as such provide insights on potential outliers driving significance. Payment “other” includes deals paid neither with cash nor shares and those that could not be identified with certainty, where the latter constitutes the biggest portion.
Panel A: Acquirer
ESG quartile Same industry Same country Payment "other"
A 41% 42% 46%
B 39% 54% 36%
C 52% 63% 27%
D 49% 69% 37%
ROE Same industry Same country Payment "other"
High 42% 45% 42%
Medium 47% 57% 31%
Low 49% 73% 31%
Panel B: Target
ESG quartile Same industry Same country Payment "other"
A 32% 54% 38%
B 35% 62% 31%
C 46% 77% 25%
D 52% 80% 28%
ROE Same industry Same country Payment "other"
High 35% 61% 34%
Medium 44% 66% 28%
