Going green for the money:
Does it pay to be sustainable?
Differences in valuation of mergers and acquisitions concerning
sustainable United States companies
Master Thesis
University of Amsterdam
by
Hessel Jim Dreteler
July 2014, Amsterdam
Student number: 5921090 Thesis supervisor: dr. S.R. Arping
Study: Master’s in Business Economics Specialization: Finance
Abstract
This study investigates the relationship between sustainability and firm performance. It is found that in takeovers, the target firm benefits significantly from the level of sustainability of the acquiring firm. Shareholders of the target firm observe a cumulative abnormal return that is on average 33 percentage points higher for an increase in the acquirer’s level of sustainability by one. Besides, in takeovers between two high-level sustainable firms, the target’s CAR is more than 28% where in the case of a takeover between two low-level sustainable firms the target’s CAR is only 15%. This difference is found to be significantly different from zero. These results indicate that shareholders have a positive valuation of sustainability. An explanation for the findings might be that a learning effect takes place, or that sustainability benefits from some sort of economies of scale. Also a synergetic effect might be present in takeovers between two high-level sustainable firms.
Table of contents
Abstract ... 2
Introduction ... 4
Literature review ... 6
Data and methodology ... 10
Empirical results and discussion ... 22
Conclusion ... 34
1. Introduction
The public awareness of social sustainable issues is growing over the past decade and is more relevant now than it has been before (Ehrgott, Reimann, Kaufmann, & Carter, 2011). According to the US SIF – The Forum for Sustainable and Responsible Investment, sustainable and responsible investments have grown by 22% to a total of 3.74 trillion dollars in the period from 2009 to 2011 (2012). This illustrates that firms in the United States are increasing their investments on projects with a high level of corporate social responsibility (CSR). These firms do this either voluntarily as part of their vision and firm strategy, or due to increasing pressure from activist stakeholders (Deng, Kang, & Low, 2013). It is this tradeoff between maximizing shareholder value and maximizing stakeholder value managers have to make. To examine this tradeoff, several studies have investigated the relationship between corporate social responsibility and firm performance in an event study setting. The results of these studies are mixed. Cochran & Wood (1984) for example find a positive relation of CSR and financial performance, and McGuire, Sundgren, & Schneeweis (1988) find an increase in corporate social responsibility resulting in a reduction of firm risk. Several other studies however find no significant relationship between CSR and financial performance (McWilliams & Siegel, 2000).
Despite the extensive research on the relationship of CSR and firm performance, little research is performed on the relation between sustainability, being an important aspect of corporate social responsibility, and firm performance. This paper contributes to the existing literature by looking at the relationship between sustainability and firm performance, with the use of takeovers as a working ground. A few papers in this field of study exist. These papers are however limited in their research as they do not look at abnormal returns in the target firm. Aktas, De Bodt, & Cousin (2011) for example reveal a positive relation between the target’s social and environmental risk practices and the acquirer’s abnormal announcement return. Other research compares high versus low CSR acquirers in a takeover and finds a higher abnormal announcement return for the former group (Deng et al., 2013). The drawback of these papers can be addressed to the scope of the research, as their main focus is on the abnormal stock returns for the acquiring firm. Both papers overlook any potential abnormal returns in the stock of the target firm. This is of interest because in a takeover, it is most often the acquirer’s vision that is continued in the post-takeover firm. The target firm therefore is probably more influenced by a sustainable strategy of the acquirer than the other way around. To provide a broader view and a better understanding of the effects of sustainability, the focus of this research will be on both target and acquiring firms in mergers and acquisitions in the United States.
The reason for looking at mergers and acquisitions in examining firm performance is twofold. Firstly, a merger or acquisition is often among the largest investments a company will ever make. A merger or acquisition is therefore relevant as it will have an effect on many stake- and shareholders (Betton, Eckbo, & Thorburn, 2008). Using M&As is therefore relevant in unraveling the tradeoff between maximizing shareholder value and maximizing stakeholder value. Secondly, there is the more important fact that because mergers and acquisitions are mostly unanticipated events, the announcement of a merger or acquisition provides data with relatively little noise. This provides the possibility to examine abnormal announcement returns, which partially mitigates the reverse causality problem present in previous studies on the relationship of CSR and firm performance (Deng et al., 2013). Throughout this paper, the terms takeover and merger & acquisition will be used interchangeably for any merger or acquisition, regardless of the actual form.
The purpose of this study is to investigate whether sustainability adds value to a firm. To do this, takeovers are used as a platform to be able to extract the value of sustainability. As such, this paper answers the main question: How do shareholders value mergers and acquisitions on or by United States companies with a high level of sustainability compared to mergers and acquisitions on or by United States companies with a low level of sustainability in the period 1994 through 2013? To arrive at an answer to this question, the mergers and acquisitions are looked at in two ways. First, this paper examines the explanatory effect of corporate social responsibility and of sustainability on the value of both the target and the acquiring firm. The value of the firms is measured as the cumulative abnormal return (CAR) at the announcement period. In an efficient market, the cumulative abnormal announcement period return would capture the full change in shareholder wealth that occurs as a result of the takeover. As a parameter for CSR the MSCI ESG score is used, which is an annual data set of environmental, social, and governance ratings of publicly traded companies. As a parameter of sustainability, the environmental aspect of the MSCI ESG score is used. Second, this paper looks at whether differences exist in cumulative abnormal return between mergers and acquisitions concerning high sustainable firms compared to low sustainable firms. It does this by comparing the average cumulative abnormal returns of four different cases, i.e. where a high sustainable firm takes over a high sustainable firm, where a low sustainable firm takes over a low sustainable firm, where a high sustainable firm takes over a low sustainable firm, and where a low sustainable firm takes over a high sustainable firm.
The results of this study confirm a positive valuation of sustainability by shareholders. It is found that the environmental score of the acquiring firm is significant in determining the
cumulative abnormal announcement return of the target firm. The CAR in the announcement period of the target firm appears to be 33 percentage points higher if the target is taken over by a sustainable acquirer. The acquiring firm does not seem to benefit from a sustainable target firm. It is also found that a strong synergetic effect takes place in the case of a sustainable acquirer taking over a sustainable target, although this synergetic effect is only perceived in the cumulative abnormal return of the target firm.
The paper continues as follows. Section 2 contains a literature review that looks at previous literature on the value of corporate social responsibility and on the abnormal returns of target and acquirer. Section 3 describes the data and methodology that is used, and includes the hypotheses that will be tested throughout this paper. In section 4, the results of this paper are presented, along with an interpretation of the findings. A conclusion can be found in section 5 along with suggestions for further research.
2. Literature review
In an attempt to look at the tradeoff managers seem to make between maximizing shareholder value and maximizing stakeholder value, several studies have looked at the relationship between corporate social responsibility (CSR) and financial performance (McWilliams & Siegel, 2000). There are two understandings of investing in socially responsible projects; CSR is either value destroying or value creating. A number of studies suggest that investing in corporate social responsibility projects is value destroying. The relation between CSR and the firm’s financial performance is looked at by for instance Walley & Whitehead (1996), who suggest that “win-win” won’t work. Although the implementation of environmental standards will be beneficial to the society, it inevitably will bring higher costs for the individual companies and is therefore detrimental to competiveness. This view is shared by Vance (1975), who finds a negative correlation between corporate social responsibility and the stock price.
If corporate social responsibility is in fact value creating, there should be a positive relation between CSR and the abnormal stock return. Several studies have indeed found a positive relation between the two, and suggest that a higher level of corporate social responsibility can improve the satisfaction of the employees and the customers which leads to a higher level of productivity and a larger demand for the products (Aktas et al., 2011). Other benefits of a high level of CSR may include the opportunity to tap into new markets, and a signal to investors about the quality of the management (Fombrun & Shanley, 1990). In addition, McGuire et al. (1988) suggest that the investment in socially responsible projects
signals good and responsible behavior to financial institutions. This leads to an improved accessibility to funding. In addition, it is found that corporate social responsibility results in a reduction of firm risk (McGuire et al., 1988). In a study on the relationship between emission reduction and firm performance it is found that efforts to reduce emissions are starting to pay within two years after the implementation, justifying the initial investment (Hart & Ahuja, 1996).
Robinson et al. (2011) investigate firms that are included in the Dow Jones Sustainability Index (DJSI), which tracks stock performance of leading companies in terms of economic, environmental, and social criteria. They suggest that being added to the Dow Jones Sustainability Index results in a sustained increase in a firm’s share price, suggesting that the benefits of being included on the DJSI outweigh the costs associated with investing in CSR. They use a multivariate regression analysis to test whether this increase in stock price is the result of reputational effects or due to the firm being included in a large index, and find the former to be true, thus suggesting CSR leads to a sustained creation of value (Robinson et al., 2011).
2.1 Takeovers
Takeovers experience a great interest in the field of finance, mostly because of the large impact a takeover has on a company. For many companies, a takeover is among the largest investments the company will ever make (Betton et al., 2008). Another reason for the attention takeovers receive is the large number of takeovers each year. In 2004, one of the years where the amount of takeovers peaked, over 30,000 acquisitions were completed globally, with a combined value of $1,900 billion (Cartwright & Schoenberg, 2006).
To be able to examine the run-up period returns and the announcement period returns, first a clear separation of these event windows has to be made. The announcement period usually consists of the three days directly surrounding the merger: one day prior to the announcement until one day after the announcement. Because an event window as short as possible contributes to the statistical significance of the evidence, the statistical evidence of the announcement period return will most likely be the strongest and is therefore the most common period that is looked at (Andrade, Mitchell, & Stafford, 2001). The announcement period will be indicated with [-1, 1]. Several studies extend the period before the announcement to grasp possible insider trading, leaking of information, or speculation (Datta, Narayanan, & Pinches, 1992). This period is called the run-up period and the most common timeframe consists of two,
three, twenty, forty, or sixty days prior to the announcement, although other event windows are used as well. This study makes use of the event window of 41 to 2 days prior to the announcement as proposed by Betton et al. (2008). The run-up period will therefore be indicated with [-41, -2].
2.2 Target
Previous studies’ evidence seems to conclude that wealth creation only benefits shareholders of the target firm (Datta et al., 1992). Already in the late seventies, the impact of tender offers on the returns of stockholders is looked at. It is found that abnormal returns of 21 percent and 19 percent are realized for stockholders of the target firm in successful and unsuccessful bids respectively (Dodd & Ruback, 1977, among others). In a meta-analysis, Datta et al. (1992) analyze the empirical literature of 41 studies concerning wealth creation in takeovers. This replication analysis is realized with the use of a multivariate framework and a regression analysis. The main attention of the research is on shareholder wealth creation, gauged by looking at the prediction errors in the combined [-10, 10] time period for daily data, and month zero for monthly data. Based on 79 observations for the target firm, this results in a mean prediction error of almost 22 percent (Datta et al., 1992).
In a more recent paper, with its main goal to develop a model for the dynamics of stock returns, Hackbarth & Morellec (2008) first examine abnormal announcement returns to check whether their data is in line with past literature. To acquire the abnormal announcement period return, Hackbarth & Morellec (2008) obtain the daily abnormal return from a market model, which is then cumulated for each of the 1,086 takeovers. A three-day cumulative abnormal announcement period return of more than 18 percent is found, which showed that their data follows a pattern as described in previous literature (Hackbarth & Morellec, 2008). In a similar study, based on a sample from the period 1980 – 2005, Betton et al. (2008) find a mean announcement return of 14.61 percent.
Officer (2004) suspects that the type of payment (i.e. cash, stock, or a combination of both), and especially the inclusion of collars in the stock bid, matters for the cumulative abnormal announcement period return in takeovers. For a sample of merger bids between 1991 and 1999, the three-day cumulative abnormal announcement returns (CAAR) are measured and separated by type of payment. It is found that the median target abnormal returns are approximately 17 percent, 13 percent, and 14 percent for a bid in cash, a mixed bid, or a stock offer respectively.
2.3 Acquirer
In contrast to the large positive abnormal returns that stockholders of target firms realize, wealth creation doesn’t seem to benefit stockholders of the acquiring firm (Datta et al., 1992). When looking at 41 separate studies, Datta et al. (1992) find a mean prediction error in both the combined [-10, 10] time period and the month zero of less than a half percent for the acquiring firm. Furthermore, when the type of acquisition was a non-conglomerate merger, a significant higher return for the acquiring firm was found. By comparing the abnormal returns of the target firm with the acquiring firm, Dodd & Ruback (1977) found an abnormal return of approximately twenty percent for the target firm, while a gain of less than three percent is found for the acquiring firm. These results are supported by the findings of both Betton et al. (2008) and Hackbarth & Morellec (2008). The former find a CAAR of 0.7 percent, and the latter find a three-day announcement period abnormal return of -0.5 percent when looking at a sample of 1,086 takeovers. The research of Servaes (1991) also indicates a lower abnormal return for the bidding firm compared to the target firm. In his paper, Servaes (1991) looks at the relation between takeover gains and Tobin’s Q, and finds a negative cumulative abnormal return of one percent for the bidding firm, where the returns are larger when targets have low, and bidders have high q-ratios.
When the sample is split for method of payment in the tree types cash payments, stock payments, and mixed payments, it is found that the returns for the bidding firm are three percent, minus six percent, and minus four percent respectively (Servaes, 1991). Officer (2004) finds a similar, but less distinct result. In his research, with its focus more on announcements made in the nineties, a median bidder CAAR of zero percent, minus two percent, and zero percent is found for cash payments, stock payments, or a mixed payment respectively (Officer, 2004).
To investigate the relation between a firm’s corporate social responsibility and acquirer firm’s shareholder returns, Deng et al. (2013) use a sample of 1,556 completed United States mergers between 1992 and 2007 to calculate the cumulative abnormal return of the acquiring firm for the period surrounding the merger. It is found that the [-1, 1] period CAR is insignificantly different from zero. Subsampling this result in 786 firms with a high level of corporate social responsibility and 770 firms with a low level of corporate social responsibility gives an interesting result. For the [-1, 1] period, the mean cumulative abnormal return for high-level CSR acquirers is slightly positive, but insignificantly different from zero. The mean CAR for low-level CSR acquirers is found to be -0.48 percent, significant at the five percent level (Deng et al., 2013). More similar tests are done to support this finding and overall the results
show that a high level of corporate social responsibility of the acquiring firm leads to a higher announcement return for the acquiring firm.
2.4 Value of corporate social responsibility and sustainability in takeovers
To look at the value of corporate social responsibility in takeovers, Deng et al. (2013) divide a sample of 668 completed mergers between 1992 and 2007 in a subset in which the acquiring firm has a high level of CSR, consisting of 335 firms, and a subset in which the acquiring firm has a low level of CSR, consisting of 333 firms. When comparing the cumulative abnormal returns for the [-1, 1] announcement period using a test of difference, the mean abnormal returns show a difference of 0.64 percent in favor of the subset where the acquiring firm has a high level of CSR, significant at the ten percent level (Deng et al., 2013).
In addition of the acquiring firm’s level of corporate social responsibility, Aktas et al. (2011) include both the acquirer’s and the target’s level of social responsibility as an explanatory variable. They observe a sample of 106 takeovers announced between 1997 and 2007 and estimate the cumulative abnormal return for the acquiring firm. It is then found that the acquirer’s cumulative abnormal return is significantly higher when the target is a high social responsibility firm, than when the target is a low social responsibility firm. This suggests that the acquiring firm could benefit from socially responsible investments that the target firm has made (Aktas et al., 2011).
3. Data and methodology
The purpose of this study is to investigate whether sustainability adds value to a firm. To do this, takeovers are used as a platform to be able to extract the value of sustainability, and as such the following main question will be answered: How do shareholders value mergers and acquisitions on or by United States companies with a high level of sustainability compared to mergers and acquisitions on or by United States companies with a low level of sustainability in the period 1994 through 2013? To arrive at an answer to this question, the mergers and acquisitions are looked at in two different ways. First, this paper looks at the effect of both corporate social responsibility (CSR) and sustainability on the target and the acquiring firm separately. CSR is measured in the form of the MSCI ESG rating, and sustainability in the form of the environmental score from the MSCI ESG rating. The effect is measured through a regression on the cumulative abnormal return (CAR). Secondly, this paper looks whether
differences in cumulative abnormal return exist between takeovers concerning high sustainable firms compared to takeovers concerning low sustainable firms.
It is expected that corporate social responsibility is valued positively by shareholders, as proposed in the more recent researches done on this subject. The same is expected for the aspect sustainability. This would be revealed by an increase in the abnormal return. If instead a decrease of the abnormal return is observed in takeovers concerning high-level CSR or sustainable firms, investors vision CSR and sustainability as a costly issue, thus negatively. A positive effect of CSR and of sustainability on the abnormal return would mean that investors value CSR and sustainability positively. An explanation for this could be that a learning effect is present. This means that the low-scoring firm can learn from the practices of the high-scoring firm, thereby increasing its own knowledge and possibly even implement the CSR or environmental aspects itself. It is most likely that a learning effect will show up at the target firm. The reason for this is that it is most likely the acquirer’s vision on CSR and sustainability that is continued in the post-takeover firm. If an acquirer does value CSR and sustainability positively, the acquirer may push the target firm to meet certain CSR and sustainable standards. The target firm thereby can learn from the investments done by the acquiring firm. If an acquirer does not value CSR and sustainability positively but instead values it neutral or even negatively, the acquirer does not care about implementing the practices of the target firm on its own and thus is not willing to learn at all. Therefore, if the learning effect is present in corporate social responsibility or sustainability, it is expected that the abnormal return of the target firm reacts positively on an acquirer with a high level of CSR or sustainability.
Another explanation for a positive valuation of CSR and sustainability by shareholders is the presence of a reputational effect, as suggested by McGuire et al. (1988). In that case, the low-scoring firm gets a boost from the sustainable reputation of the high-scoring firm. If this effect is present, it is expected to be observed in the abnormal returns of both the target and the acquiring firm. This would be reflected in a positive effect of CSR and sustainability on the abnormal return.
Regardless of whether or not CSR or sustainability are found to be positively valued by shareholders, it is of interest to look at the effects of CSR and sustainability in different takeover settings. Are for example CSR and sustainability valued differently if it is a high-level firm that takes over another high-level firm than when it is a level firm that takes over another low-level firm? It is expected that cases where a high-low-level firm takes over another high-low-level firm, the abnormal return is more positive, or less negative, than in cases where a low-level firm takes over another low-level firm. If this is indeed observed, this could point to the existence of some
sort of synergetic effect in CSR or sustainability. It could for instance be costly to maintain a high level of CSR or sustainability, due to the rapidly increasing level of technology or stricter rules imposed by governments. If a higher abnormal return is observed for high-level takeovers compared to low-level takeovers, this means that these costs can be partly diminished due to an economies of scale effect.
3.1 Sample description
The sample of mergers and acquisitions is obtained from the Thomson One database, formerly known as the Securities Data Corporation’s U.S. Mergers and Acquisitions database, or SDC Platinum. The sample includes takeovers announced in the period January 1, 1994 to December 31, 2013 where both the target and the acquirer are United States companies. The final sample consists of all takeovers that meet the following criteria:
1. The acquisition is completed.
2. The acquiring firm owns less than 50% of the target firm before the takeover, and holds more than 50% of the target firm’s stock after the takeover.
3. The deal value disclosed in Thomson One is more than $1 million.
4. Both the acquiring firm and the target firm are not operational in the financial or utilities industries, which are defined by primary SIC codes between 6000 and 6999 and between 4900 and 4999 respectively.
5. Both the acquiring firm and the target firm have financial statement information available through Compustat, and stock return data available from the Center for Research in Security Prices (CRSP).
6. Both the acquiring firm and the target firm are rated by the MSCI environmental, social, and governance (ESG) stats.
This results in two workable datasets. One that is based on the target firms, which contains 310 takeovers and another that is based on the acquiring firms, which contains 477 takeovers.
MSCI publishes an annual data set of environmental, social, and governance ratings of publicly traded companies, starting from 1991, called MSCI ESG STATS (ESG). The ESG dataset builds on the achievements made by KLD, who were among the most important when it comes to sustainability ratings. The ESG dataset ranks the 3,000 largest U.S. publicly traded companies based on market capitalization based on over 60 ESG indicators in seven ESG categories for three ESG themes:
- Social: o Community o Human Rights o Employee Relations o Diversity o Customers - Governance
Next to these subjects, the ESG dataset also includes business involvement data for the controversial business themes alcohol, gambling, firearms, military, nuclear power, and tobacco. One of the main advantages of the ESG dataset is the in-depth publication of the scores. As such, the scores can be broken down by the seven ESG categories. This gives the opportunity to very specifically extract possible effects on the cumulative abnormal return. The downside of the ESG dataset is the selection MSCI makes before ranking a firm. Since it only ranks the 3,000 largest United States firms based on market capitalization, the sample is small and might suffer from a selection bias.
In table 1, a sample distribution of both the acquirer and the target sample is presented by announcement year. As can be seen, the number of takeovers gradually increases until it peaks around the years 2005, 2006 and 2007, consisted with previous literature. In the year 2008 and the years following, the number of takeovers decreases again. This is most likely due to the credit crisis that started in 2008. Table 1 also shows the market value of equity (MVE), the mean deal value, and the mean relative deal size, defined as the value of the transaction divided by the market value of equity of the acquiring firm. Distinctive peaks can be noted, both for the acquiring firms and the target firms, in the market value of equity. The acquiring firms have a spike in equity in the year 1999. This seems to correspond to the dot-com bubble that was largest at the end of the 1990s. The target firms’ market value of equity is much more even, although a small peak can be noted in 2002. When this sample is compared to for instance the sample of Masulis, Wang, & Xie (2007), it becomes clear that the sample indeed includes relatively large firms due to the selection process of MSCI in the construction of their ESG score.
3.2 Variable construction
In this subsection, the measurement of the variables is described. First, the abnormal stock returns as a dependent variable is looked at. This is followed by a description of the main
Table 1
Sample Distribution by Announcement Year
The target sample consists of 310 completed United States mergers and acquisitions that are listed in Thomson One database, the acquiring sample consists of 477 completed U.S. takeovers. Both samples include takeovers announced between 1994 and 2013, and are distributed accordingly. Variable definitions are in the Appendix.
Target sample Acquirer sample
Year Obs. Mean Market Value of Equity ($mil) (Median) Mean Deal Value ($mil) (Median) Mean Relative Size (Median) Obs. Mean Market Value of Equity ($mil) (Median) Mean Deal Value ($mil) (Median) Mean Relative Size (Median) 1994 0 - - - 1 345.50 199.21 2.04 1995 0 - - - 5 4,912.21 163.57 0.20 1996 2 2,021.81 2,821.82 0.56 11 15,299.11 683.77 0.52 1997 0 - - - 12 29,745.01 944.79 0.93 1998 2 4,888.64 6,484.42 0.46 22 42,677.21 1,169.95 -1.64 1999 2 11,813.06 1,485.26 0.23 26 94,921.82 3,041.24 2.48 2000 3 12,608.86 11,046.60 0.91 22 15,178.71 1,960.69 0.94 2001 3 8,359.29 9,500.28 0.35 23 30,310.66 1,982.75 0.37 2002 3 28,148.04 23,243.57 0.35 17 31,318.97 4,344.54 0.50 2003 14 1,802.29 2,000.36 0.32 32 42,319.98 1,145.50 0.70 2004 31 5,899.23 2,869.07 0.34 33 12,033.92 2,718.37 0.88 2005 45 4,670.24 3,507.62 0.32 42 37,544.64 3,668.75 0.88 2006 36 4,798.02 2,080.45 0.24 40 29,358.88 2,011.57 0.51 2007 43 2,220.42 1,621.55 0.31 44 29,351.92 1,548.99 1.45 2008 29 2,833.45 1,748.79 0.74 32 28,060.16 1,489.53 -1.45 2009 37 2,316.07 3,950.79 0.19 34 41,653.83 3,357.71 0.34 2010 32 2,809.85 1,516.44 0.21 37 31,163.62 1,568.83 0.20 2011 13 13,344.45 4,263.71 0.60 14 28,255.52 2,799.28 0.44 2012 15 - 935.35 0.14 23 36,407.00 1,082.84 0.59 2013 0 - - - 7 34,628.78 1,013.81 0.29 Total 310 5,296.50 2,862.08 0.34 477 33,658.42 2,111.98 0.52 (1,473.39) (798.10) (0.14) (8,180.42) (516.59) (0.23)
explanatory variable: the environmental, social, and governance score. Finally, other explanatory variables are discussed.
3.2.1 Abnormal stock return
The abnormal stock return is measured as the cumulative abnormal return (CAR), and is calculated for both the sample of acquiring firms and the sample of target firms. First, a predicted β for firm j is calculated using:
𝑟𝑗𝑡 = 𝛼𝑗 + 𝛽𝑗𝑟𝑚𝑡+ 𝜀𝑗𝑡 (1)
where 𝑟𝑗𝑡 is the logarithmic return for firm j on day t, and 𝑟𝑚𝑡 is the value weighted market return including dividends. The daily returns for each separate firm are obtained through the Center for Research in Security Prices (CRSP) database. Day zero is the day of the announcement of the bid. For the estimation period, 200 days of non-missing stock return data
is used, which is measured over the window starting at event day minus 260 and ending on event day minus 60. The results from equation (1) are then used to predict a return 𝑟𝑗𝑡𝑝𝑟𝑒𝑑 for firm j on day t. The abnormal return 𝐴𝑅𝑗𝑡 for firm j on day t follows from:
𝐴𝑅𝑗𝑡= 𝑟𝑗𝑡− 𝑟𝑗𝑡𝑝𝑟𝑒𝑑 (2) In order to calculate the cumulative abnormal return, first two event windows have to be distinguished, namely the run-up period which comprehends the days [-41, -2], and the announcement period which comprehends the days [-1, 1].
The cumulative abnormal return (CAR) for firm j over event window k is then calculated using:
𝐶𝐴𝑅𝑗𝑘 = ∑ 𝐴𝑅𝑗𝑡𝑑𝑘𝑡
𝐾 𝑘=1
(3)
where 𝐴𝑅𝑗𝑘 is the abnormal stock return of firm j in event window k as calculated using equation (1), and 𝑑𝑘𝑡 is a dummy variable that takes the value of one if day t is in event window
k, and zero otherwise. Following Betton et al. (2008), the z-values are determined as
𝑧 = ( 1 √𝑁)
𝐶𝐴𝑅𝑗𝑘
𝜎𝐴𝑅𝑗𝑘 (4)
where 𝜎𝐴𝑅𝑗𝑘 is the estimated standard error of the abnormal return of firm j in event window k.
3.2.2 Environmental, Social, and Governance score
As the main focus of this paper is the valuation of sustainability in takeovers, the main explanatory variable should capture this effect. As described earlier, MSCI publishes an annual data set of environmental, social, and governance ratings of publicly traded companies. This database is the leading authority concerning environmental, social, and governance ratings and is the most comprehensive dataset for the evaluation of a firm’s level of corporate social responsibility (Deng et al., 2013).
The MSCI ESG dataset uses a binary ranking. The ESG indicators that are used to calculate a score varies slightly over time. Due to the binary ranking system that is used and because of the variation in indicators over the years, the raw ESG-scores cannot be compared directly from one year to the next. To mitigate this issue, an adjusted ESG-score is calculated by dividing the raw ESG-score by the number of ESG indicators. Two different correlation
Table 2
Correlation Matrix
The target sample consists of 310 completed United States mergers and acquisitions that are listed in Thomson One database, the acquiring sample consists of 477 completed U.S. takeovers. Both samples only include completed takeovers announced between 1994 and 2013. Variable definitions are in the Appendix.
Panel A: Target sample
Ann’ct CAR Target ESG score Acquirer ESG score Target Ann’ct CAR Target 1.0000
ESG score Acquirer 0.3404 1.0000
ESG score Target -0.0064 0.3247 1.0000
Ann’ct CAR Target Environment score Acquirer Social score Acquirer Governance score Acquirer Environment score Target Social score Target Governance score Target Ann’ct CAR Target 1.0000
Environment score Acquirer
0.2006 1.0000
Social score Acquirer 0.2845 0.5075 1.0000
Governance score Acquirer 0.2828 0.2219 0.0697 1.0000
Environment score Target -0.0119 0.3458 -0.0135 0.1949 1.0000
Social score Target -0.1802 0.1126 0.0993 -0.1087 0.0892 1.0000
Governance score Target 0.2118 -0.0540 -0.1772 0.3445 0.2405 -0.0791 1.0000
Panel B: Acquirer sample
Ann’ct CAR Acquirer ESG score Acquirer ESG score Target Ann’ct CAR Acquirer 1.0000
ESG score Acquirer -0.1380 1.0000
ESG score Target -0.1397 0.5038 1.0000
Ann’ct CAR Acquirer Environment score Acquirer Social score Acquirer Governance score Acquirer Environment score Target Social score Target Governance score Target Ann’ct CAR Acquirer 1.0000
Environment score Acquirer
-0.0226 1.0000
Social score Acquirer -0.1831 0.5008 1.0000
Governance score Acquirer 0.1200 0.2449 0.1880 1.0000
Environment score Target -0.0459 0.3627 0.2792 0.3850 1.0000
Social score Target -0.2759 0.1844 0.4186 0.1212 0.3662 1.0000
Governance score Target 0.2246 0.0227 0.0474 0.4228 0.4017 0.1904 1.0000
matrices for both the target sample and the acquirer sample are presented in table 2, that is a correlation matrix containing the target’s ESG score and acquirer’s ESG score correlated with the announcement period CAR, and another matrix where the environmental, social, and governance part is looked at separately in correlation with the announcement period CAR. The first matrix of panel A shows the relationship between the ESG score of both target and acquirer with the cumulative abnormal return of the target firm. It shows that while the target’s ESG score is not very related to the target’s announcement period return, the acquirer’s ESG score is positively related to the target’s CAR. The first matrix in panel B displays the correlation of the ESG score and the cumulative abnormal announcement period return of the acquiring firm.
Both the ESG score of the target and the ESG score of the acquiring firm seem to be negatively related to the CAR of the acquiring firm. Interestingly, in panel A as well as in panel B, the correlation of the ESG score of the acquiring firm and the ESG score of the target firm is strongly positive with coefficients of 0.3247 and 0.5038 respectively. This could indicate a high ESG scoring acquirer having a preference for a high ESG scoring target firm. The second matrices of panel A and panel B show for all three different parts that the ESG score is built from, i.e. environment, social, and governance, the correlation with the cumulative abnormal return of respectively the target and the acquirer.
3.2.3 Control variables
In addition to the explanatory variables, a set of control variables is created to capture other effects influencing the cumulative abnormal returns. These control variables can be divided into three groups, namely bidder characteristics, target characteristics, and deal characteristics (Betton et al., 2008), and this is done accordingly. The bidder characteristics that are included are firm size, Tobin’s q, free cash flow (FCF), and leverage. When Moeller, Schlingemann, & Stulz (2004) examine a sample of 12,023 acquisitions, they find an abnormal return for the acquiring firm that is approximately two percentage points higher for smaller bidders. This finding seems to be irrespective of other firm- or deal characteristics, and looks to be a permanent effect. As an explanation for this effect, they suggest that larger firms in general pay larger premiums than smaller firms would do, and that larger firms are more tended than smaller firms to enter acquisitions with negative dollar synergetic gains. Firm size will be defined as the log of the total assets (Compustat item 6) of the acquiring firm.
In an analysis of the relationship between abnormal returns and Tobin’s q, Servaes (1991) examines a sample of 704 takeovers to find that abnormal returns for both the target and the bidder are larger when the acquiring firm has a high Tobin’s q and the target firm has a low Tobin’s q. A similar result is found by Lang, Stulz, & Walkling (1989), who interpret this result as a confirmation that takeovers of poorly managed targets by well-managed bidders have higher abnormal returns for both the target and the bidder. Tobin’s q is defined as the market value of assets divided by the book value of assets, where the market value of assets is calculated as the book value of assets (Compustat item 6) minus the book value of common equity (Compustat item 60) plus the market value of common equity (Compustat item 25 multiplied by item 199).
According to Jensen (1986), both the acquirer’s free cash flow and the acquirers leverage are of importance when explaining the abnormal return of a takeover. Because takeovers are a way for managers to spend excess cash instead of paying it out to shareholders, managers of firms with large free cash flows are more likely to commence in takeovers with little or no benefit (Jensen, 1986). Free cash flow is defined as operating income before depreciation (Compustat item 13) minus interest expenses (Compustat item 15) minus income taxes (Compustat item 16) minus capital expenditure (Compustat item 128), divided by the book value of assets. Leverage is defined as the book value of long term debt (Compustat item 9) and short term debt (Compustat item 34), divided by the market value of assets as described above.
The target characteristics that are controlled for are the target’s stock price run-up, and the ownership status of the target. Evidence of a sample of 1,814 takeovers leads Schwert (1996) to conclude that a dollar run-up in the stock price of the target firm leads to a rise in the total offer premium by roughly a dollar. Moreover, evidence is found that the pre-bid run-up is uncorrelated with the post-announcement increase in the stock price of the target firm, thus concluding that run-up is an extra cost to the acquiring firm (Schwert, 1996). The target’s stock price run-up is defined as the cumulative abnormal return over the run-up period [-41, -2]. Although the attention in the research of Schwert (1996) is focused on the target firm, and the results are thus only for the target run-up, the run-up period is also included for the acquiring firm in the regressions on the acquiring sample.
In a study on firms that made multiple takeovers during the merger wave of the 1990s, Fuller, Netter, & Stegemoller (2002) suggest that the division of gains from a takeover depends on the ownership status of the target. They find that when the target is a public firm, shareholders of the target firm benefit more, and when the target is a private firm or a subsidiary, the bidding firm’s shareholders benefit more. A possible explanation might be that the illiquidity of the private or subsidiary firm brings a discount for the acquiring firm (Fuller et al., 2002). Three dummy variables are created to capture this effect, namely public, private, and subsidiary.
The deal characteristics that are controlled for are relative size of the deal, whether both target and acquirer are operating in a high-tech industry, and the method of payment. Although often significant results are found, past literature doesn’t seem to agree fully on whether the relative size of the deal is positively or negatively related to returns (Moeller et al., 2004). Relative deal
size is defined as the transaction value of the deal divided by the market value of equity of the acquiring firm.
The Bureau of Labor Statistics has developed a list of high-technology industries based on measurements of industry employment in both R&D and technology-oriented occupations. These firms are grouped based on SIC codes. With this list a dummy variable is created that equals one if both the acquiring firm and the target firm are operative in a high-tech industry, and zero otherwise. It is suspected that both firms operating in a high-tech industry has a negative effect on the abnormal return of the acquirer, since high-tech acquirers are more likely to underestimate the costs and overestimate the synergies generated by the takeover (Masulis et al., 2007).
There exist three different methods of payment in takeovers. These include all-cash payments, all-stock payments, or a mix of a variety of debt securities and cash. The payment choice is based on several factors which are not mutually exclusive and include tax effects, deal financing costs under asymmetric information, agency and corporate control motives, and behavioral arguments (Betton et al., 2008). A payment consisting of (partly) equity seems to trigger negative abnormal returns for the acquiring shareholders (Masulis et al., 2007). Two dummy variables are therefore created, that is an all-cash dummy that equals one if the deal was fully financed by cash and zero otherwise, and a stock dummy that equals one if at least part of the deal was financed by stock.
Summary statistics of the variables are presented in table 3. As expected from previous literature, the cumulative abnormal announcement period return the target firm is strongly positive, i.e. 25.97%, while the CAR of the bidding firm for the same period is somewhat negative with a mean value of -1.02%. When the both the run-up period abnormal return and the announcement period abnormal return are included, the cumulative abnormal return for the target firm increases to 30.59%, and the CAR for the acquirer drops to -1.91%. For illustrative purposes, figure 1 displays the daily cumulative abnormal returns from day -40 through day 20 with respect to the announcement of the takeover. As can be seen in figure 1, the average cumulative abnormal return of both the target firms and the acquirer firms follow a pattern as expected. A similar pattern is found in previous literature (Betton et al., 2008).
20
Table 3
Summary Statistics
The target sample consists of 310 completed United States mergers and acquisitions that are listed in Thomson One database, the acquiring sample consists of 477 completed U.S. takeovers. Both samples include takeovers announced between 1994 and 2013, and are distributed accordingly. Variable definitions are in the Appendix.
Target sample Acquirer sample
Variable N Mean Std. Dev. 25% Median 75% N Mean Std. Dev. 25% Median 75%
Panel A: Cumulative abnormal return and MSCI Environmental, Social, and Governance score
CAR (-1, 1) 310 0.2597 0.2629 0.1127 0.2151 0.3496 477 -0.0102 0.0719 -0.0407 -0.0046 0.0202 Z-value 310 0.0937 0.0893 477 -0.0186 0.1537
CAR (-41, -2) 310 0.0462 0.1984 -0.0622 0.0419 0.1483 477 -0.0089 0.1474 -0.0777 -0.0097 0.0691 Z-value 310 0.0857 0.3595 477 -0.0234 0.2855
MSCI ESG score 73 -0.1393 0.4992 -0.4762 -0.2417 0.0083 373 0.1157 0.8148 -0.3512 -0.0417 0.3750 Environment score 76 0.0011 0.0888 0 0 0 410 0.0629 0.2088 0 0 0.0889 Social score 75 -0.1654 0.2808 -0.3333 -0.1667 0 390 0.0803 0.5810 -0.3333 0 0.4083 Governance score 76 -0.0814 0.1623 -0.1667 0 0 396 -0.0996 0.2524 -0.2857 -0.1381 0
Panel B: Bidder characteristics
Total assets ($ mil) 238 20,430.73 32,283.97 2,120.13 5,968.76 27,162.00 424 18,789.75 30,929.26 2,059.38 5,695.95 23,438.50 Market value of equity ($ mil) 238 30,976.79 49,897.00 2,071.25 6,707.32 31,858.97 424 33,658.42 53,236.58 2,290.37 8,180.42 35,834.31 Tobin’s q 238 2.0260 1.0537 1.3439 1.7732 2.3737 424 2.5260 2.0110 1.3914 1.9452 2.8265 Free cash flow 238 0.0767 0.0831 0.0487 0.0765 0.1090 424 0.0777 0.0868 0.0494 0.0796 0.1233 Leverage 238 0.1359 0.1304 0.0446 0.1007 0.1951 424 0.1244 0.1265 0.0320 0.0851 0.1838
Panel C: Target characteristics
Total assets ($ mil) 65 3,680.47 5,718.07 317.64 1,119.53 4,224.73 59 3,707.96 5,677.00 317.64 1,300.03 4,722.12 Market value of equity ($ mil) 65 5,296.50 9,185.58 474.62 1,473.39 5,120.70 59 5,422.71 9,203.07 431.15 1,694.50 6,909.61 Public (dummy) 307 1 0 456 1 0
Private (dummy) 1 1 - 5 1 0 Subsidiary (dummy) 2 1 0 15 1 0
Panel D: Deal characteristics
Relative deal size 238 0.3373 0.5568 0.0438 0.1446 0.4209 424 0.5211 4.5045 0.0531 0.2321 0.6745 High tech (dummy) 310 0.6484 0.4782 477 0.6771 0.4681
All cash 157 1 0 225 1 0
Figure 1. Percent average cumulative abnormal stock returns to targets and acquirers from day -40 through day 20 relative to the announcement day of the takeover.
3.3 Methodology
In order to look at the valuation of corporate social responsibility, and in particular sustainability, this paper uses the cumulative abnormal return as a measure of valuation. To test the hypotheses,
a regression will be used
𝐶𝐴𝑅𝑡= 𝛼 + 𝛽1𝑥1𝑡+ . . . +𝛽𝑛𝑥𝑛𝑡+ 𝛾1𝑘1𝑡+ . . . +𝛾𝑛𝑘𝑛𝑡+ 𝜐𝑛 (5) where 𝐶𝐴𝑅𝑡 is the cumulative abnormal return at time t, 𝛼 is a constant, 𝛽𝑛 is the least-squares regression coefficient of the tested explanatory variable 𝑥𝑛𝑡 at time t, 𝛾𝑛 is the regression coefficient for the tested control variables 𝑘𝑛𝑡 at time t, and 𝜐𝑛 is the error term. The direct
effect of corporate social responsibility on the cumulative abnormal return is looked at by regressing the MSCI ESG score on the cumulative abnormal return. This provides evidence to test for the effect of corporate social responsibility on firm performance. The MSCI ESG environmental rating will be regressed on the announcement period abnormal return to test for the direct effect of sustainability on the cumulative abnormal return of both the target and the acquirer. In the regressions, all of the bidder-, target-, and deal characteristics as described above will be included as control variables with the exception of the ownership status dummy
-5% 0% 5% 10% 15% 20% 25% 30% 35% -40 -38 -36 -34 -32 -30 -28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 A ve ra g e cumul ati ve a bnorma l re turn
Trading days relative to announcement date of bid
variables. As can be seen in table 3, the target sample only consists of one private and two subsidiary firms, and the acquirer sample only of five private and fifteen subsidiary firms. The ownership status will therefore not be included in the regressions as a control variable. Including these characteristics in an unpublished regression held similar results as did omitting. To test for differences in the different takeover scenarios, a two sample t-test will be used:
𝑡 = 𝑥̅1− 𝑥̅2 √𝑠12 𝑛1+ 𝑠22 𝑛2 (6)
where 𝑥̅1 and 𝑥̅2 are the mean values of respectively sample one and two, 𝑠12 and 𝑠22 are the standard deviations of the two samples, and 𝑛1 and 𝑛2 are the corresponding number of
observations. The number of degrees of freedom for this test statistic will be the smaller of 𝑛1− 1 and 𝑛2− 1.
4. Empirical results and discussion
As described above, the direct effect of corporate social responsibility on the cumulative abnormal return is looked at by regressing the MSCI ESG score on the announcement period return. Estimates of this can be found in table 4. Columns (1), (2), and (3) show the regression based on the target sample while columns (4), (5), and (6) show the regressions with respect to the acquirer sample. As can be seen in column (1) and column (2), the ESG score of neither the target firm nor the acquiring firm is significantly different from zero in explaining the cumulative abnormal return of the target firm. Including both the ESG score of the target firm and the ESG score of the acquiring firm doesn’t change this. In columns (4) and (5) the ESG scores of respectively the target firm and the acquiring firm are not significantly different from zero in the explanation of the acquirers firm CAR. The inclusion of the ESG scores of both target and acquiring firm, displayed in column (6), does not contribute statistically to the explaining of the acquirer’s CAR. This results in concluding that corporate social responsibility is valued neutral by shareholders. It does not matter if it is the target or the acquirer that is looked at. These findings are in line with the results of McWilliams & Siegel (2000), who finds corporate social responsibility to have a neutral impact on financial performance of the firm.
The results from table 4 seem to indicate that corporate social responsibility does not influences the abnormal return of either the target or the acquirer, and thus that corporate social responsibility is valued neutral by shareholders. It could however be possible that some part of
Table 4
Initial Regression Analysis
The samples consist of completed United States mergers and acquisitions that are listed in Thomson One database and are announced between 1994 and 2013. The firms are only included in the regression if they are ranked by MSCI in their ESG database. This results in a different number of firms in each regression. In the target sample, the cumulative abnormal return of the target firm is the dependent variable. In the acquirer sample, the cumulative abnormal return of the acquiring firm is the dependent variable. The explanatory variables are grouped by either ESG score, bidder characteristics, target characteristics, or deal characteristics. Variable definitions are in the Appendix. In the parentheses are t-statistics based on standard errors adjusted for heteroskedasticity.
(1) (2) (3) (4) (5) (6) Target sample Acquirer sample
MSCI ESG scores:
ESG score target -0.0541 -0.0691 -0.0031 -0.0106 (-1.21) (-1.54) -0.16 -0.56 ESG score acquirer -0.0200 0.0593 0.0036 0.0103
(-0.70) (1.28) 1.08 0.75 Bidder characteristics: Log(total assets) 0.0282 0.0202 0.0187 -0.0093 -0.0056** -0.0123** (1.49) (1.39) (0.92) -1.58 -2.36 -2.11 Tobin’s q -0.0439** -0.0359* -0.0356* -0.0049 0.0029** -0.0017 (-2.21) (-1.95) (-1.78) -0.48 1.99 -0.15 Free cash flow -0.2099 0.0203 -0.1189 0.0060 0.0683 -0.0034
(-1.23) (0.06) (-0.67) 0.05 1.31 -0.03 Leverage 0.0056 -0.1356 0.0385 -0.0328 0.0558 -0.0012 (0.02) (-0.95) (0.13) -0.29 1.42 -0.01 CAR (-41, -2) 0.0014 -0.0560* -0.0737 0.02 -1.80 -1.29 Target characteristics: CAR (-41, -2) -0.2513* -0.1528 -0.2621* (-1.82) (-1.57) (-1.78) Deal characteristics:
Relative deal size -0.0573 -0.0466 -0.0435 -0.0020* -0.0011 -0.0014 (-0.78) (-1.31) (-0.65) -1.88 -1.58 -1.21 High tech 0.0568 0.0837* 0.0453 -0.0361 -0.0207** -0.0238
(0.82) (1.75) (0.70) -1.19 -2.56 -0.84 High tech x relative deal size -0.0253 -0.0571 -0.0327 -0.0019 -0.0146** -0.0052
(-0.42) (-1.22) (-0.61) -0.19 -2.41 -0.43 Stock deal -0.0255 -0.1194 -0.0044 -0.0370 -0.0352*** -0.0590** (-0.51) (-1.41) (-0.09) -1.12 -2.92 -2.16 All-cash deal -0.0306 -0.0149 -0.0226 0.0074 -0.0118 -0.0215 (-0.50) (-0.18) (-0.38) 0.19 -1.16 -0.83 Intercept 0.0935 0.2038 0.1222 0.1190* 0.0576** 0.1508** (0.73) (1.49) (0.86) 1.89 2.25 2.36 Number of observations 56 220 55 60 341 58 Adjusted-R2 7.15% 9.37% 9.23% 14.36% 16.20% 16.07%
*** significant at the 1% level, ** significant at the 5% level, * significant at the 10% level.
the ESG score counteracts another part of the ESG score, and it is therefore of interest to examine the extent to which the different aspects the ESG score consists of influence the cumulative abnormal return. Table 5 displays the impact of the three different aspects, i.e. environment, social, and governance, in explaining the cumulative abnormal return of both the target firm in columns (1) trough (3) and the acquirer firm in columns (4) through (6). As can be seen in column (1), the corporate governance rating is highly positive and significant at the ten percent level. Both the environmental score and the social score of the target firm seem to
Table 5
Regression Analysis Environment, Social, and Governance score
The samples consist of completed United States mergers and acquisitions that are listed in Thomson One database and are announced between 1994 and 2013. The firms are only included in the regression if they are ranked by MSCI in their ESG database. This results in a different number of firms in each regression. In the target sample, the cumulative abnormal return of the target firm is the dependent variable. In the acquirer sample, the cumulative abnormal return of the acquiring firm is the dependent variable. The explanatory variables are grouped by either ESG score, bidder characteristics, target characteristics, or deal characteristics. Variable definitions are in the Appendix. In the parentheses are t-statistics based on standard errors adjusted for heteroskedasticity.
(1) (2) (3) (4) (5) (6) Target sample Acquirer sample
MSCI ESG scores:
Environment score target -0.1796 -0.3089** -0.0675 -0.1155 (-1.60) (-2.28) (-1.18) (-1.55) Social score target -0.1504 -0.1590 -0.0312 -0.0186
(-1.54) (-1.60) (-0.94) (-0.56) Governance score target 0.2482* 0.2770* 0.1667*** 0.1149*
(1.72) (1.75) (2.87) (1.97) Environment score acquirer 0.3322*** 0.0697 0.0220 0.0247
(2.87) (0.70) (1.38) (0.44) Social score acquirer -0.0896** 0.0750 0.0030 0.0044
(-2.16) (1.41) (0.57) (0.22) Governance score acquirer -0.1657 0.2040 -0.0086 0.0919*
(-1.51) (1.49) (-0.57) (1.98) Bidder characteristics: Log(total assets) 0.0375* 0.0119 0.0276 -0.0032 -0.0056** -0.0085 (2.00) (0.92) (1.37) (-0.54) (-2.38) (-1.33) Tobin’s q -0.0336 -0.0307* -0.0156 -0.0053 0.0024* 0.0018 (-1.54) (-1.82) (-0.65) (-0.63) (1.68) (0.17) Free cash flow -0.3390** 0.0549 -0.3483* -0.0358 0.0642 -0.0766
(-2.38) (0.18) (-1.92) (-0.32) (1.23) (-0.64) Leverage -0.1391 -0.1310 -0.0228 -0.0523 0.0582 -0.0148 (-0.42) (-0.89) (-0.07) (-0.47) (1.48) (-0.11) CAR (-41, -2) -0.0222 -0.0516* -0.0959 (-0.25) (-1.83) (-1.61) Target characteristics: CAR (-41, -2) -0.3198** -0.1620* -0.3773** (-2.31) (-1.67) (-2.29) Deal characteristics:
Relative deal size -0.0110 -0.0408 0.0028 -0.0022*** -0.0011 -0.0020** (-0.15) (-1.20) (0.04) (-3.62) (-1.49) (-2.21) High tech 0.0584 0.0725* 0.0536 -0.0541* -0.0220*** -0.0406
(0.90) (1.65) (0.90) (-1.79) (-2.71) (-1.36) High tech x relative deal size -0.0416 -0.0692 -0.0504 0.0094 -0.0147** 0.0048
(-0.78) (-1.39) (-0.99) (0.95) (-2.44) (0.39) Stock deal -0.0818 -0.0859 -0.0680 -0.0596** -0.0336*** -0.0819*** (-1.37) (-1.04) (-1.03) (-2.42) (-2.92) (-3.90) All-cash deal -0.1108* -0.0177 -0.1019 -0.0147 -0.0113 -0.0374* (-1.91) (-0.21) (-1.59) (-0.44) (-1.16) (-1.77) Intercept 0.0672 0.2210* 0.0972 0.1035* 0.0563** 0.1520** (0.52) (1.65) (0.65) (1.81) (2.25) (2.41) Number of observations 57 221 56 61 354 59 Adjusted-R2 12.15% 13.55% 18.19% 26.25% 15.85% 25.94%
*** significant at the 1% level, ** significant at the 5% level, * significant at the 10% level.
negatively affect the target’s cumulative abnormal return, though not statistically significant. Looking at the effect of the acquirers ESG score, column (2), in explaining the target’s cumulative abnormal return, a large and positive environmental score is observed that is
statistically significant at the one percent level. This means that investors value an environmentally strong acquiring firm as a positive influence on the target firm; an increase by one point in the environmental score of the acquirer leads to an increase of over 33 percentage points in the targets cumulative abnormal return. This is strong evidence in favor of the expectations about a learning effect being present. Surprisingly, the acquirer’s social rating seems to have a negative impact on the target firm’s valuation, as does the acquirer’s governance rating, although the latter is insignificantly different from zero. This might be interpreted as the aspect social not having any information that can be transferred through learning. In that case, the shareholders of the target firm expect investments to be made by the target firm to meet the standards of the acquiring firm, while no benefits are presented thus resulting in a decrease in stock price. In column (3), the inclusion of all three aspects of the ESG score of both the target firm and the acquiring firm in the regression is presented, and a different image occurs. When looking at the environmental score of the target firm in column (3), a negative coefficient is observed that is significant at the five percent level. An increase of one point in the target firm’s environmental score would reduce the cumulative abnormal return by almost 31%. This might indicate that the environmental score might be seen as a costly burden to some acquirers, resulting in a lower takeover price. In column (3), the environmental score of the acquiring firm does not seem to be as important anymore in explaining the target’s CAR as it was in column (2). Where in column (2) the coefficient of the acquirer’s environmental score was highly significant, it is insignificantly different from zero in column (3). This difference might be explained because of a bias through selection. In column (2) all 221 data samples are used where the acquiring firm has a MSCI ESG score at the moment of the takeover-announcement. In column (3) both the acquirer and the target have to have a MSCI ESG score at the time of announcement, and as a result leaving only 56 takeovers to observe. Since MSCI does not rank all traded firms, but makes a selection based on size of the firm, the ranked firms must be sufficiently large. This could create a bias in the results, and thus could explain the difference in importance of the acquiring firm’s ESG effects between columns (2) and (3). Looking at the ESG effects on the cumulative abnormal return of the acquiring firm in columns (4) through (6), a significant coefficient is observed for the corporate governance score of the target firm in columns (4) and (6) at respectively the ten percent and the five percent level. This is as expected and in line with the findings of Masulis et al. (2007). The environmental score does not seem to be a great influence on the cumulative abnormal return of the acquiring firm, with both the environmental score of the target as the environmental score of the acquirer being insignificantly different from zero in columns (4) through (6). These
results support again the presumption that a learning effect is in place. Besides a possible learning effect being present, the difference that the environmental score has on the target firm versus the acquirer firm might partly be explained through the omitted ownership status of the target. Fuller et al. (2002) find that the target firm benefits more if the target is a public firm, and that the acquirer benefits more if the target is a private firm or a subsidiary. The sample that is used in this research consists almost entirely of public firms, which may enhance the effect the environmental score has on the cumulative abnormal return of the target firm. The results found in table 5 are contrary to the results of Aktas et al. (2011), as they find that the acquirer’s cumulative abnormal return is significantly higher when the target is a high social responsibility firm, than when the target is a low social responsibility firm. A reason for this might be the use of a different measure of corporate social responsibility, as they use Innovest’s Intangible Value Assessment ratings where this paper uses the MSCI ESG score.
Finally, to fully grasp the effects of the different aspects of corporate social responsibility on the valuation of a firm’s stock, the social rating is split into five different facets by MSCI, i.e. community, human rights, employee relations, diversity, and product. This is of particular interest due to the statistically significant negative effect the acquirer’s social rating had on the CAR of the target firm. These indicators are included as individual variables and regressed against the cumulative abnormal return, of which the results can be found in table 6. Of these five indicators, in column (1) only community and product are statistically significant at the ten percent level with coefficients of respectively -0.9391 and -0.4860. This means that for every point added to the community or product score of the target firm, the cumulative abnormal return of the target firm, on average, decreases with 93.91 percentage points and 48.60 percentage points respectively. An explanation for this might be that investors see these aspects as burdens. Community for example, is rated higher if the company supports charitable giving programs. The indicator product receives a higher score if the company takes efforts to benefit the disadvantaged. These programs are likely to be costly to the firm, and thus harm current profit. The potential forthcoming profit of the post-takeover firm might be affected as well due to for example contracts that are to be carried out in the future. The governance score of the target is significant at the ten percent level, no different than in the previous regression. The indicator for the human rights score of the target is omitted from this regression, because it showed collinearity with the target’s CAR. When looking at the five social indicators of the acquiring firm with relation to the target firm’s cumulative abnormal return, displayed in column (2), it can be seen that both the indicators human right and product are negative and significantly different from zero, with coefficients of -0.3797 and -0.3240. An explanation can
Table 6
Regression Analysis ESG score with social score split in five
The samples consist of completed United States mergers and acquisitions that are listed in Thomson One database and are announced between 1994 and 2013. The firms are only included in the regression if they are ranked by MSCI in their ESG database. This results in a different number of firms in each regression. In the target sample, the cumulative abnormal return of the target firm is the dependent variable. In the acquirer sample, the cumulative abnormal return of the acquiring firm is the dependent variable. The social score of the ESG database is split in five categories, i.e. community, human rights, employee relations, diversity, and product. The human rights score of the target firm showed collinearity with the cumulative abnormal return of both target and acquirer, and is therefore not included in this table. The explanatory variables are grouped by either ESG score, bidder characteristics, target characteristics, or deal characteristics. Variable definitions are in the Appendix. In the parentheses are t-statistics based on standard errors adjusted for heteroskedasticity.
(1) (2) (3) (4) (5) (6) Target sample Acquirer sample
MSCI ESG scores:
Environment score target -0.2429 -0.2978 -0.0645 -0.1190 (-1.45) (-1.67) (-1.04) (-1.59) Community score target -0.9391* -1.0843 -0.1366 -0.1199
(-1.70) (-1.46) (-1.14) (-0.88) Employee score target 0.0614 0.0883 0.0128 0.0177
(0.32) (0.55) (0.16) (0.18) Diversity score target -0.1098 -0.1193 -0.0312 -0.0180
(-0.97) (-0.97) (-0.67) (-0.33) Product score target -0.4860* -0.5659* -0.0425 -0.0216
(-1.94) (-1.89) (-0.63) (-0.26) Governance score target 0.2700* 0.3565** 0.1618*** 0.1310**
(1.89) (2.11) (2.76) (2.17) Environment score acquirer 0.3312*** 0.0545 0.0190 0.0294
(3.23) (0.45) (1.11) (0.44) Community score acquirer 0.0781 -0.2458 0.0286** 0.0742
(0.45) (-0.90) (2.46) (0.80) Human rights score acquirer -0.3797* -0.2371 -0.0238 -0.1159
(-1.96) (-0.44) (-1.44) (-0.68) Employee score acquirer -0.0792 0.4941*** -0.0160 -0.0121
(-0.75) (2.99) (-1.13) (-0.13) Diversity score acquirer 0.0162 -0.0088 0.0016 0.0039
(0.26) (-0.08) (0.11) (0.08) Product score acquirer -0.3240** 0.0623 0.0113 0.0061
(-2.31) (0.73) (1.22) (0.16) Governance score acquirer -0.1445 0.1674 -0.0084 0.0801
(-1.28) (1.03) (-0.55) (1.64) Bidder characteristics: Log(total assets) 0.0237 -0.0142 0.0193 -0.0055 -0.0054** -0.0114 (1.38) (-0.95) (0.89) (-0.68) (-1.99) (-1.26) Tobin’s q -0.0351 -0.0326** -0.0226 -0.0036 0.0026* 0.0023 (-1.40) (-1.97) (-0.76) (-0.35) (1.76) (0.17) Free cash flow -0.5267*** -0.1259 -0.7715*** -0.0454 0.0683 -0.1099
(-2.95) (-0.44) (-3.19) (-0.35) (1.27) (-0.68) Leverage -0.1971 -0.0296 -0.1142 -0.0437 0.0578 -0.0074 (-0.56) (-0.20) (-0.35) (-0.37) (1.44) (-0.05) CAR (-41, -2) -0.0245 -0.0558** -0.1071* (-0.27) (-2.01) (-1.76) Target characteristics: CAR (-41, -2) -0.2992** -0.1593* -0.4429*** (-2.27) (-1.68) (-2.91) Deal characteristics:
Relative deal size -0.0121 -0.0705** 0.0261 -0.0021*** -0.0011 -0.0020** (-0.16) (-2.00) (0.39) (-3.25) (-1.48) (-2.10) High tech 0.0938 0.0581 0.1244 -0.0490 -0.0207** -0.0408
(1.28) (1.40) (1.53) (-1.37) (-2.57) (-1.01)
Table 6 - Continued
(1) (2) (3) (4) (5) (6) High tech x relative deal size -0.0430 -0.0476 -0.0612 0.0075 -0.0148** 0.0053
(-0.73) (-0.98) (-1.10) (0.67) (-2.46) (0.37) Stock deal -0.1134 -0.0489 -0.1397 -0.0595** -0.0327*** -0.0830*** (-1.67) (-0.61) (-1.59) (-2.10) (-2.86) (-2.93) All-cash deal -0.1011 -0.0099 -0.1097 -0.0110 -0.0121 -0.0354 (-1.47) (-0.13) (-1.34) (-0.31) (-1.23) (-1.37) Intercept 0.2100 0.4147*** 0.2393 0.1180 0.0526* 0.1751** (1.53) (2.85) (1.28) (1.60) (1.94) (2.36) Number of observations 57 221 56 61 354 59 Adjusted-R2 14.71% 19.00% 24.50% 22.45% 15.84% 13.50%
*** significant at the 1% level, ** significant at the 5% level, * significant at the 10% level.
be sought in the cost it brings. The human rights score becomes larger when the company has put time and resources in establishing good relations with the people near the place of operations. If the acquiring company values a high social score, it is likely that the target firm also has to meet a certain level of social responsibility. When the target company itself does not regard this to be of great importance and as a result has not invested in these social characteristics, the acquirer has to invest in these aspects when the takeover is completed. This could be costly for the post-takeover firm and thus explain the negative relation between the acquirer’s social indicators and the target’s CAR, as a smaller stock price will be paid by the acquirer. Furthermore, in column (2) the environmental score of the acquiring firm is again strongly positive and significant at the one percent level as was observed in the previous regressions. The combination of the target’s and the acquirer’s social indicators in column (3) does not show unexpected results. The only exception is the acquirer’s employee score, but this is more likely explained by the selection bias as described earlier. Column (4) shows the target’s social indicators regressed against the acquirer’s cumulative abnormal return. None of these indicators appears to be significantly different from zero, which is as expected because it is most likely the acquirer’s vision on social behavior that is continued after the takeover, and not the target’s vision. This does however indicate that the acquiring firm does not benefit from the environmental and social characteristics that the target firm might already have invested in. As in the previous regressions, the corporate governance score of the target is significantly different from zero in explaining the acquirer’s cumulative abnormal return. In column (5), out of the five social indicators only the community score is significantly different from zero, with a coefficient of 0.0286. This shows that the different aspects of the social score of the acquirer are of not much importance for the valuation of the acquirer. A result similar to the results in columns (4) and (5) is presented in column (6), where the acquirer’s and target’s scores are combined.
