The probability of ESG shareholder proposals:
The influence of socially responsible investing and corporate social
responsibility
E.R. Eding12
Thesis MSc Finance
Supervisor: Prof. Dr. L.J.R. Scholtens University of Groningen
June 2016
1 University of Groningen Faculty of Economics & Business Author: E.R. (Erwin) Eding Student number: s2596547 E-mail: erwineding@hotmail.com
2 I would like to thank my supervisor, parents, and fellow students for their help and support during the process of writing my MSc thesis.
Abstract
This paper investigates how the greenness of institutional owners and the corporate social responsibility performance (CSR) of a firm affect the probability of a shareholder proposal on environmental, social, and governance issues (ESG). The research is conducted on an U.S. data set on 1,018 unique firm-year observations between 2011 and 2014. I find that the greenness of institutional ownership is a positive factor in explaining the probability of receiving shareholder proposals on environmental issues. Employee wellbeing, which is a category of CSR performance, is positively related to the probability of receiving shareholder proposals on social issues. The performance on other categories of CSR do not seem to affect the probability of ESG shareholder proposals.
JEL codes: G32; G34; M14
2 1. Introduction
This paper investigates whether the greenness of institutional owners and corporate social responsibility (CSR) performance of a firm can be associated with the firms’ probability of receiving a shareholder proposal. The probability of receiving shareholder proposals regarding environmental, social, and governance issues (ESG) will be investigated. Shareholders are able to file proposals to be voted at on the annual general meeting of shareholders (AGM) of a firm. CSR concerns the social and environmental behaviour of a firm that goes beyond the obligations of the law (Kitzmueller and Shimshack, 2012). When an investor includes CSR performance of the firm into his investment decision making, the investor is called green and/or social responsible. The assets under management by socially responsible investing (SRI) investors experienced a vast increase in recent decades (Inderst, Kaminker, and Steward, 2012; Schueth, 2003). Because of the increasing preference for investors to include CSR performance in the investment decision making, it is interesting to examine if SRI and CSR have explanatory power in shareholder engagements. Namely, shareholder engagement is often motivated by shareholders’ preferences (Tkac, 2006). How SRI and CSR performance influence shareholder engagement is examined in this paper.
3 engagement, in the form of shareholder proposals, is more likely to target poorly performing firms. Dimson et al. (2015) add to this that firms with reputational concerns and socially conscious institutional investors are more likely to be targeted by ESG related shareholder engagements. This tendency is compliant with the increasing amount of investors who adopt social responsible investing principles.
From the previous paragraph it is clear that there are some relationships exposed between firms (non-financial) characteristics and shareholder engagement. However, it is not investigated yet if there actually does exist a significant relationship between a firms’ actual CSR performance and whether a firm is targeted by shareholder engagement on CSR issues. This paper tries to fill the gap which is not covered by other studies. The essence of this paper is reflected in the following research question:
Research question: What is the influence of socially responsible investing and corporate social responsibility performance on the probability of a shareholder proposal on environmental, social, and governance issues?
I try to relate CSR performance to shareholder engagement. Further, I try to find the relationship between the amount of SRI investors and shareholder engagement. The most common type of shareholder engagement is filing shareholder proposals at the firms’ annual general meeting of shareholders (AGM). Because of the proximity between SRI, CSR, and ESG, I will investigate the probability of shareholder proposals on ESG issues. Other types of shareholder proposals are ignored. With shareholder proposal probability is meant the chance for a firm to receive a shareholder proposal. Hence, I will investigate whether the probability of receiving an ESG shareholder proposal for firms is, among other things, related to the firms’ performance on CSR and the amount of SRI shareholders.
By answering the research question this paper expands on shareholder engagement studies from, for example: Karpoff et al. (1996) by taking modern day CSR scores into account, Jo and Harjoto (2014) by examining another pressure device (shareholder proposals instead of analyst following), and Cziraki, Renneboog, and Szilagyi (2010); Schooley et al. (2010) by not taking only governance scores in mind, but also environmental and social scores. To the best of my knowledge, no study has yet addressed the research question of this paper.
The research question is empirically addressed and based on an U.S. sample of Fortune 250 companies during 2011-2014. Whether a company received a shareholder proposal is observed from the database of Proxymonitor.org3. SRI and CSR characteristics of firms who are targeted by
4 environmental, social, or governance shareholder proposals are compared by characteristics of firms who were not targeted by ESG shareholder proposals. CSR characteristics are given by Asset4’s scores on CSR performance. SRI is measured with the percentage social responsible institutional investors. Both univariate and multivariate analysis are performed. The multivariate analysis is performed with logistic regressions. By using logistic regression the estimate for the dependent variable, shareholder proposal probability, will always lie between one and zero. Given that the logistic regression is a non-linear model, the parameters are estimated with maximum likelihood.
The analysis’ results can explain what type of firms attracts shareholder proposals and why investors4 make use of shareholder proposals. The results could explain which CSR characteristics are
important for the shareholders’ proposal targeting decision. Firms who want to avoid shareholder engagements can make use of the results to optimize relevant CSR characteristics. I find that the greenness of institutional ownership is a positive factor in explaining the probability of receiving a shareholder proposals on environmental issues. Employee wellbeing, which is a category in CSR performance, is positively related to the probability of receiving a shareholder proposals on social issues. The performance on other dimensions of CSR do not seem to affect ESG shareholder proposal probability.
This paper proceeds as follows. In Section 2 a literature review on shareholder proposal probability is provided, where I will also state what is missing in the existing literature. Next in Section 3, background information is provided on subjects examined in this paper. Those who want to know more about shareholder proposals before reading the literature review on shareholder proposal probability, might first want to read the first part of Section 3. In Section 4 I will set up hypotheses to make clear the expectations about what could influence shareholder proposal probability. Section 5, 6, and 7 report respectively how the study is conducted, the results, and robustness analysis. Finally, Section 8 concludes.
2. Literature review on shareholder proposal probability
This section presents a grasp of existing literature on shareholder proposal probability. The literature presented below is considered as most relevant to this paper. In the remainder of this paper there will be continuously reflected upon this literature to indicate the relevance and contributions of my paper. Karpoff et al. (1996) were among the first that studied shareholder proposal probability. They concluded that firms with poor prior financial performance are more likely to attract shareholder
5 proposals. Schooley et al. (2010) were the first, as far as I know, who tried to find explanatory power of shareholder proposal probability in non-financial performance measures. Schooley et al. (2010) reported a positive relationship between the level of non-corporate governance shareholder proposals and leadership tenure with corporate governance shareholder proposal probability. Karpoff et al. (1996) and Schooley et al. (2010) both estimated their models with logistic regressions with an U.S. sample during the periods 1986-1990 and 2003, respectively. Cziraki et al. (2010) confirm that the negative relationship between shareholder proposal probability and financial performance also holds between 1998 and 2008 for European firms, further they find that investors also consider ownership structure of the firm when filing a shareholder proposal. They find ownership concentration and the stake of institutional investors positively influence proposal probability. The intuition behind the positive relationship of ownership concentration is that fewer shareholders have to be convinced to vote in favour of the shareholder proposal. The positive relationship from institutional ownership suggests that proposal filers count on the voting support of institutional shareholders. Tkac (2006), Rojas, M’Zali, Turcotte, and Merrigan (2009) find activists are more likely to target large, well-known companies to maximize the impact of their campaign. Tkac (2006) and Rojas et al. (2009) came to their results with an univariate analysis of an U.S. sample over the years 1992-2002 and 1997-2004, respectively. The papers that are discussed in the above paragraph are reflected upon in the remainder of this paper. Table 1 presents characteristics of literature where will be reflected upon in my paper.
6 probability. Taking into account the interests of a broader range of stakeholders is an integral part of SRI. Investors who embrace social principles are called to conduct socially responsible investing (SRI). Just like ESG activism, the embracement of social principles into investment decision making is on the rise (Schueth, 2003). Because of the emphasis of shareholder engagement on environmental and social issues it is interesting to examine if there is a relationship between ownership of investors who embrace social principles and shareholder proposal probability.
Table 1, selected papers on shareholder proposal probability: This table provides an overview of the papers that are mainly reflected upon in my paper. The number of observations are the number of shareholder proposals that are examined by the studies. The last column presents the key results that are relevant to my paper. It is mentioned in the last column if a paper investigated the probability of a shareholder proposal on a specific issue.
Author(s) Region Period studied
Methodology Observations Key results 1. Karpoff et al. (1996) United States 1986-1990 Logistic regressions 522 Governance shareholder
proposals are attracted by firms with poor prior financial performance. 2. Schooley et al. (2010) United States 2003 Logistic regressions
182 The probability of governance shareholder proposals is positively related with firm size, combined leadership structure, board independence, and the amount of non-corporate governance shareholder proposals. 3. Cziraki et al. (2010) Europe 1998-2008 Heckman selection model
290 Shareholder proposal target firms tend to underperform and have low leverage. Further, the equity stake of institutional investors is positively related with shareholder proposal probability. 4. Tkac (2006) United States 1992-2002 Univariate analysis 2829 Non-corporate governance shareholder proposal filers target big and well-known companies to maximize their impact of their proposals. 5. Rojas et al. (2009) United States 1997-2004 Univariate analysis
2310 Several large companies are targeted systematically by shareholder proposals on social issues.
7 3. Background information
This section of the paper provides information on terms that appear frequently in this paper. Some useful information will be presented about shareholder proposals, corporate social responsibility, and social responsible investing.
3.1 What is a shareholder proposal?
Shareholders are, under rule 14a-8 from the U.S. Security and Exchange Commission (SEC), enabled to introduce proposals to U.S. publicly traded firms to be voted on at the next annual general meeting of shareholders. Under this rule it is stated that a shareholder needs 1% or $2,000 ownership in a firm to be allowed to file a proposal. Shareholders proposals may pertain to company policies and procedures, corporate governance or issues of social or environmental concern. It is not allowed to file a proposal on ordinary business issues. If the SEC accepts a shareholder proposal, the AGM can cast votes on the proposal. Whatever the outcome of the votes, shareholder proposals have a nonbinding nature.
There are several views on the use of a shareholder proposal. Schooley et al. (2010), Karpoff et al. (1996) regard shareholder proposals’ central use as an instrument to mitigate agency problems. Agency problems arise from a conflict of interest between investors and managers of a firm. Gillan and Starks (2000) argue that shareholder proposals do not mitigate agency problems because of their nonbinding nature. This is in contrast with the notion of the U.S. Social Investment Forum (USSIF) of voting behaviour on shareholder proposals. USSIF argues that, although a proposal may be rejected by a majority of votes, it can still have its value because it was favoured by a significant amount of shareholders (USSIF website5). Levit and Malenko (2011) find that the presence of an activist investor
who can discipline the manager, can enhance the impact of shareholder proposals. Disciplining management can be achieved, for example, by engaging into a proxy fight. A proxy fight consists of gathering a large group of shareholders to support shareholder proposal. A proxy fight can create additional pressure on firms to implement shareholders wishes, hence mitigating agency problems.
The goal of this paper is not to explain the use of shareholder proposals. However, the relationship that might be exposed between CSR performance and shareholder proposal probability could contribute to the understanding of what is the investors intended use of filing a shareholder proposal. A negative (positive) relationship between CSR performance could mean that shareholder proposal filer prefers to achieve a vast (marginal) over a marginal (vast) increase of CSR performance of a firm. The causality in the direction from CSR performance and SRI to ESG shareholder proposals is examined in this paper. Whether a firm did, or did not, receive a shareholder proposal on ESG issues
8 acts thereby as the dependent variable in this paper. The next two subsections provide information to facilitate further prognosis about the relationship between ESG shareholder proposals and CSR and SRI.
3.2 What is corporate social responsibility?
Traditional conflict of interests between investors and managers were driven mainly by shareholders wanting to maximize firm value (Shleifer and Vishny, 1988). Nowadays, investors also deem the interests of other stakeholders important. Corporate social responsibility (CSR) policy of a firm takes into account the interest of stakeholders, beyond the financial interests of shareholders. A common definition of CSR is: the social or environmental behaviour from a firm that goes beyond the legal or regulatory requirements of the relevant markets/economy (Kitzmueller and Shimshack, 2012). Environmental, social, and governance concerns are the three main areas which are addressed by CSR. Environmental policy concerns mainly the use of natural resources and issues related to the preservation of the earth. Social policy takes into account things like animal testing, human rights, and labour standards. Governance issues address, among other things, executive compensation and the composition of the board of directors (Meziani, 2014). Some view CSR as important, because being social responsible is a good business practice or it helps to make the world a better place (Tkac, 2006). Others, like Nobel laureate Milton Friedman, dispraise CSR because of a waste of resources. Friedman outlined his position in a famous article published by the New York Times Magazine: “The social responsibility of business is to increase its profits” (Schwartz and Saiia, 2012).
9 performance on CSR categories of a firm on firm value. A positive effect from a CSR category on firm value indicates shareholders perceive the CSR category as something positive. Since CSR is “beyond-profit”, the broader stakeholder perspective is examined as well. The stakeholder perspective is described by the effect of CSR dimensions on the reputation of the firm. According to Pérez (2015), a positive corporate reputation is established by stakeholders’ positive perception on the reported CSR performance of a firm.
10 ESG Asset4 divides governance performance into the following categories: board functions, board structure, compensation policy, vision and strategy, and shareholder rights. The first three measures are related with the board of directors of a firm. Board functions and board structure measure respectively the effectiveness, and diversity and independence of the board. Harjoto, Laksmana, and Lee (2015) find that diversity in board structure enhances a firm CSR performance. Borghesi, Chang, and Mehran (2016) find board diversity also increases firm value. Nguyen and Nielsen (2010) find that independence of the board has a positive effect on firm value. Zhang (2012) find that board independence and diversity can positively influence the firm’s reputation on social responsibility. Pfeiffer and Shields (2015) find that the choice of the compensation plan for the board is reflected in stock prices. Hence, compensation policy is valued by shareholders. Cai, Jo, and Pan (2011) find executive compensation is inversely associated with CSR of a firm. The vision and strategy measure reflects firms’ inclusion of CSR dimensions in their day-to-day decision making. Few literature can be found on CSR implementation and its effect on firm value and reputation. Shareholder rights are found to have a positive effect on firm value (Jianxin, 2015; Cremers and Ferrell, 2014). Jiraporn and Yixi (2006) find firms compensate weak shareholders rights with higher dividends to establish the reputation of the firm. It seems that the board of directors’ measures are positively valued by shareholders as well as stakeholders. Only for the vision and strategy category, it is not explicitly clear from the literature how stake- and shareholders perceive this category of governance performance.
From the literature described above, most CSR categories are suggested to be related with financial performance and corporate reputation. Hence, share- and stakeholders deem most of these CSR categories relevant. For a few CSR categories it is not clear how they are perceived by share- and stakeholders. These CSR categories are: resource reduction, human rights, and vison and strategy. Given that most CSR categories seem relevant for share- and stakeholders, it is relevant to investigate the effect of CSR on the probability of shareholder proposals on ESG issues. Section 5.1 provides further information on which CSR categories will be investigated for explaining ESG shareholder proposal probability.
3.3 What is socially responsible investing?
11 $2.2 trillion in social investments. Inderst et al. (2012) estimated global green investments in 2012 to be over $10 trillion. These are just a grasp of the available reported volumes on SRI. Whatever the definition of SRI, from the literature it is clear that socially investing is on the rise.
Investors can employ several strategies to engage in SRI. Examples of strategies are exclusionary screening, active ownership, community investing, and theme investing. One of the ways investors can make clear their support for socially responsible investing is to sign the United Nations-supported Principles of Responsible Investing (UNPRI6). The UNPRI is a covenant founded by a group
of institutional investors with the support of the United Nations. The aim of UNPRI is to put SRI into practice.
From the above there seems to be an increasing prevalence of SRI and a close relationship between SRI and ESG. These are reasons to investigate the causality from SRI to the probability of shareholder proposals on ESG issues. Therefore this paper investigates whether the greenness of investors has influence on ESG shareholder proposal probability. In Section 2 it is already made clear that Cziraki et al. (2010) find that type of ownership has influence on shareholder proposal probability. Levit and Malenko (2011) find that, despite the nonbinding nature of shareholder proposals, higher voting support increases the likelihood of management implementing the shareholders’ proposed change. It is likely that investors who adopted SRI principles provide voting support to shareholder proposals on ESG issues. Knowing that activist already consider type of ownership when filing a shareholder proposal, and higher voting support increases success of shareholder proposals, activists might take into account the greenness of investors before filing an ESG shareholder proposal.
From the entire section above, it seems that both share- and stakeholders perceive CSR as something positive. Further, SRI, CSR, and ESG seem to have elements that intertwine with each other. Given these background information, it seems appropriate to investigate if CSR and SRI do influence shareholder proposal probability on ESG issues. In Section 4, I will expand on the expected influence from CSR and SRI on probability of shareholder proposals on ESG issues.
4. Hypotheses
In this section hypotheses are set up to help answer the research question. The first hypothesis is about CSR performance and ESG shareholder proposal probability. From the above section it is clear that certain CSR characteristics have influence on the firms’ financial performance and reputation. Hence, it seems that stake- and shareholders perceive CSR as something positive. It cannot be inferred from the literature if CSR performance and ESG shareholder proposal probability have a direct relationship.
12 None of the papers presented in Table 1 paid attention to CSR in relationship with shareholder proposal probability. It is not clear what the shareholders’ goal is when filing a shareholder proposal. If the perception of Schooley et al. (2010) and Karpoff et al. (1996) regarding shareholder proposals is correct, i.e. shareholder proposals are used to mitigate agency problems, a negative relationship from CSR performance to ESG shareholder proposal probability is likely. A negative relationship means that the lower the CSR performance of a firm, the higher the probability of receiving a shareholder proposal. The agency problem in this case is that shareholders want the firm to perform better on CSR. The possible relationship between CSR performance and ESG shareholder proposal probability can also be explained in another fashion. Do investors file ESG shareholder proposals to marginally increase CSR performance of companies already performing well on CSR? Or do investors file ESG shareholder proposals to make good CSR performing firms out of bad CSR performing firms? A positive relation between CSR performance and ESG shareholder proposal probability, which is a bit counterintuitive, pleas for the first scenario. A negative relation pleas for the second scenario, this complies with the agency problem view of shareholder proposals. To test whether, and how, shareholders take into account CSR performance when filing an ESG shareholder proposal, the following (alternative) hypothesis is set up:
Hypothesis 1 (a): CSR performance of the firm positively affects the probability of receiving a shareholder proposal on ESG issues for the firm.
Hypothesis 1 (b): CSR performance of the firm negatively affects the probability of receiving a shareholder proposal on ESG issues for the firm.
Clearly, the null hypothesis for hypotheses 1(a) and 1(b) is that the CSR performance of the firm is not affecting the probability of receiving a shareholder proposal on ESG issues.
13 it necessary to file ESG shareholder proposals on firms with green institutional owners. This view could mean that proposal filers consider green institutional owners as doing a good job on monitoring the CSR performance of a firm. The presence of monitoring can reduce the conflicts arising from agency problems (Tirole, 2006). Institutional owners can be regarded as delegated monitors (Dam and Scholtens, 2012). To test the relationship from the greenness of institutional investors to ESG shareholder proposal probability, the following (alternative) hypothesis is set up:
Hypothesis 2 (a): The greenness of institutional owners of the firm positively affects the probability of receiving a shareholder proposal on ESG issues for the firm. Hypothesis 2 (b): The greenness of institutional owners of the firm negatively affects the
probability of receiving a shareholder proposal on ESG issues for the firm. Clearly, the null hypothesis for hypotheses 2(a) and 2(b) is that the greenness of institutional owners of a firm is not affecting the probability of receiving a shareholder proposal on ESG issues. How the null hypotheses are tested will become clear in the next section.
5. Model, data, and methodology
This section describes the model, data, and methodology which are used to test the hypotheses from Section 4. Variables on CSR performance and the social responsibility of institutional owners of a firm are included in the logistic regression models to explain a firm’s probability to receive an ESG shareholder proposal. The samples are based on a data set of observations between 2011 and 2014 from Proxymonitor.org. Proxymonitor.org reports the shareholder proposals received by Fortune 250 companies.
5.1 Model and variables
14 characteristics. For example, the CSR performance of a firm in 2013 might not yet be known in 2013, this forces the filer of the shareholder proposal to rely on data from 2012. Hence, if a firm received an ESG shareholder proposal for the AGM of 2014, the firm characteristics of 2013 are examined. If a firm did not receive an ESG shareholder proposal in 2012 the firm characteristics of 2011 are examined. Each firm-year observation is treated as an unique observation. A drawback of this treatment is that there may be autocorrelation in the error terms present. Autocorrelation could lead to an increase in the probability of rejecting a correct null hypothesis.
Equation (1) displays model 1, which is the most comprehensive regression model I will examine for the environmental proposal probability. Equation (2) is the second model to examine environmental shareholder proposal probability. Each variable is indicated with an abbreviation. The abbreviations may be used in tables or in the remainder of the paper.
,= + , + , + , + , ! + "" , ##!+ " , #!+ " , #!!+ $, ( 1 ) ,= + , + , + , + , ! + " , # !+ $, ( 2 )
Where , is the probability that a firm receives an environmental shareholder proposal in year t. Setting up models like this is in line with Karpoff et al. (1996); Cziraki et al. (2010); and Schooley et al. (2010). They all use the shareholder proposal probability as independent variable in their empirical research. Where these authors include control variables from prior literature respective to their study, I control for their findings, which are displayed in Table 1, in my models. The definitions of the independent variables are made clear in the sequence of the text. The models on social and governance shareholder proposal probability look similar, with their own CSR category variables, to equation (1) and (2). For models 3-4 and models 5-6 the dependent variable is social shareholder proposal probability and governance shareholder proposal probability, respectively. Including environmental CSR categories in a model to explain environmental shareholder proposal probability is equal to approach of Schooley et al. (2010). Schooley et al. (2010) included governance performance measures in a model to explain governance shareholder proposal probability. Because of the similarity of the remainder models (3-6) with models 1 and 2, it is not necessary to display them with an equation.
15 shareholder proposal probability (see Section 2, Literature review). Thus, firm size, financial performance, and institutional ownership act as control variables. The control variables reduce the possibility of the omitted variable bias. Omitting variables can cause that the explanatory power of the other variables, which are of interest to test the hypotheses, are underestimated or overestimated. The control variables reduce the probability of invalid inference about the hypotheses. Firm size is measured by the market capitalization (MCAP) and is extracted from Datastream. Datastream is a service provided by the media and information company, Thomson Reuters. Market capitalization is the share price multiplied by the number of ordinary shares outstanding. The market value is given in billions of U.S. dollars. I expect including a size variable will not explain much of the variation in which firms are targeted, because the sample group as well as the control group is based on the biggest 250 U.S. firms. However it is included, because it might be possible that certain sized firms are more or less exposed to shareholder proposals on a specific category. The financial performance of a firm is measured by the market-to-book ratio (MTB). To filter out firm size effects in financial performance there is made use of a performance ratio. Karpoff et al. (1996), Cziraki et al. (2010) both find that financial performance measured by MTB is inversely related to shareholder proposal probability. MTB is extracted from Datastream. Institutional ownership (IOWN) is measured by the sum of percentages owned by the biggest 50 institutional shareholders of a firm. So for example, if in 2014 Berkshire Hathaway’s 50 biggest shareholders have 1.5% of shares each, the IOWN variable for Berkshire in 2014 will get a value of 75% or 0.75. Studies as Cziraki et al. (2010) and Dam and Scholtens (2013) explain the implied relationship between institutional ownership and shareholder proposal probability by size of the institutional owners. Cziraki et al. (2010) find higher ownership concentration requires less shareholders to be convinced to support the proposal. Dam and Scholtens (2013) find large institutional shareholders can be convinced more easily because they account less for CSR yet. Therefore I deem it sufficient to include only the biggest 50 shareholders of a company in the institutional ownership variable. The top shareholders of a firm can be extracted from Orbis, which is a database that contains company information.
16 Responsible Investing (UNPRI) is used to verify whether shareholders include ESG considerations into their investment decision-making process. To be an UNPRI signatory, an investor has to comply with the six principles of UNPRI. UNPRI has formulated these six principles to establish SRI. These six principles can be found in Appendix A. Because only institutional investors are able to sign the UNPRI, I am not able to verify the greenness of other, for example individual shareholders. Furthermore, for the sake of practical and time considerations it would not be possible to examine the greenness of non-institutional investors. Therefore only the greenness of institutional shareholders is examined. The greenness of the top 50 institutional shareholders of a company is examined. For each firm-year observation the percentage green institutional investors is calculated. So for example, if Berkshire Hathaway has 75 institutional shareholders only the biggest 50 of those institutional shareholders are taken into account for the SRIO variable. If 25 of these 50 institutional investors are on the UNPRI signatory list, the SRIO variable will get a value of 50%, or 0.50. If the sum of shareholdings from green institutional investors will be used, multicollinearity with the IOWN variable can be present. Hence, it is important to note that impact of the greenness of institutional owners is examined, and not the impact of the percentage of shares held by green institutional investors.
17 benchmark of companies. What and how these scores, and the CSR variable scores explained in the remainder of this section, exactly measure can be found in Appendix C. In equation (1) is visually displayed that the variables on emission reduction, product innovation, and resource reduction are used to explain environmental shareholder proposal probability. In equation (2) the ECSR variable is included.
Model 3 and 4 examine social proposal probability. Social performance variables on the wellbeing of employees (EMP), product responsibility (SPR), community (SCO), and human rights (SHR) are included in model 3. In model 4 social performance is represented by a weighted social score (SCSR). The wellbeing of employees is a combined measure of scores from Asset4 on diversity and opportunity, employment quality, health and safety, and training and development. Because all these scores are closely related to employee wellbeing it is not necessary to include them all separately. The construction and source of social performance variables are the same as the variables presented in the previous paragraph.
Model 5 and 6 examine governance proposal probability. In model 5 governance performance is represented by variables on scores on board functions (GBF), board structure (GBS), vision and strategy (GVS), and shareholder rights (GSR). The governance performance in model 6 is represented by a weighted governance score (GCSR) variable. Asset4’s score on compensation policy is not included because executive compensation shareholder proposals are neither included in the governance proposal sample, see Section 5.2. The construction and source of governance performance variables are the same as the variables presented in the previous paragraph.
5.2 Data
18 timeframes I use. Proxymonitor.org tracks the shareholder proposals received by the 250 largest publicly traded companies, ranked by revenue as reported on yearly frequency by the Fortune magazine. Fortune ranks companies which are incorporated, operate and file financial statements in the United States by revenues in their respective fiscal year. Fortune’s full ranking methodology can be found at Appendix B. The database classifies the shareholder proposals in several categories. Environmental policy, social policy, and governance policy are examples of these categories. Therefore I do not have to categorize the shareholder proposals by myself anymore. From 2011 to 2014 a total of 2,483 proposals were recorded at 267 companies. The amount of unique firms observed each year from 2011 to 2014 are respectively 259, 253, 250, and 256. At some years more than 250 unique firms are observed, hence the Fortune 250 selection criteria is a bit stretched. This is because there can be several months between filing a shareholder proposal and the AGM of a firm. For example, if a shareholder proposal is filed in December 2013 for the annual general meeting in February 2014, but meanwhile the firm has been removed from the Fortune 250 list, it can still be in the database of Proxymonitor.org. Management shareholder proposals (i.e. proposals filed by the management of the firm) are excluded because these are not comparable to “ordinary” shareholder proposals. Management are inside-owners and may have other incentives to file shareholder proposals than the outside-owners, like institutional investors. After filtering out shareholder proposals submitted by the management of the target firm, 1,217 ESG shareholder proposals at 211 unique companies remain.
19 inference of the hypotheses possible. The sample for environmental shareholder proposal probability consists of 114 unique firm-year observations of firms who received an environmental shareholder proposal, and of 496 unique firm-year observations of firms that did not receive any shareholder proposal. Another possibility would be to form the environmental sample of 114 firm-year observations of firms who received an environmental shareholder proposal and 904 (=1,018-114) firm-year observations of firms who did not receive an environmental shareholder proposal. However, Schooley et al. (2010) find that the amount of non-corporate governance shareholders proposals have explanatory power in corporate governance shareholder proposal probability. This finding indicates that shareholder proposal probability among different type of shareholder proposals might be correlated. This presence of correlation might cause a lack of variation in the independent variable values among target and non-target observations. The social proposal sample and corporate governance proposal sample consists of respectively 278 and 284 observations of firms who did receive a proposal on the particular topic. Both to the social and governance sample, 496 firm-year control observations of firms that did not receive shareholder proposals are added. The sample construction is displayed in Table 2. For the corporate governance sample the shareholder proposals on executive compensation are excluded. Executive compensation proposals are excluded as a consequence of the Dodd-Frank law. The Dodd-Frank law obligates firms to allow shareholder votes on executive compensation7. The consequence of the Dodd-Frank law is observable at the data set of
Proxymonitor.org on shareholder proposal events. There are 1,004 unique firm-year observations on executive compensation shareholder proposals. If I should choose to include executive compensation proposals in the governance sample, only a control sample can be formed of 14 observations. Such control group size is insufficient to obtain reliable results. Table 3 denotes the specific topics of the environmental, social, and governance shareholder proposals. The categorical names of shareholder proposals show similarity with the names of the included variables representing CSR performance in the regression models. For example, shareholder proposals on chairman independence have a frequent presence in the governance sample. The variable board structure (GBS) measures, among other things, the independence of the board of directors. A drawback is that unequal presence of different type of shareholder proposals categories in the samples could results in overestimating the explanatory power of an independent variable. For example, it makes sense to say that the board structure variable is closer related to the probability of a shareholder proposal on chairman independence than to a shareholder proposal on proxy access. Hence, the explanatory power of the board structure variable could be overestimated if there is a high presence of shareholder proposals on chairman independence in the sample.
20 The descriptive statistics on the variables explained above of the 1,018 firm-year observations are given in Table 4. The least observed variable is still observed 958 times. This means the panel of data does not suffer from a large amount of missing observations. Also most variables do not seem to have high correlations. High correlations are mainly observed between weighted pillar scores and category scores. This makes sense because the sub categorical scores are the underlying products of the weighted category scores, see Appendix C. The variables that are included with each other into one of the models do not exhibit high correlations among each other. Therefore the results will not suffer from multicollinearity problems.
Table 2, sample construction: This table shows how the samples are constructed. The first column gives the year of observations. Column 2 till 4 give the amount firm observations of firms that received a shareholder proposal(s), in the respective year. The last column gives the amount of observations of firms that did not receive any shareholder proposal at the respective year. The bottom row gives the total amount of observations. The environmental, social, and governance samples consist of 610 (114+496), 774 (278+496), and 780 (284+496) observations, respectively.
Target firms Non-target firms
Environmental Social Governance
2011 27 64 65 136
2012 21 73 79 124
2013 28 70 71 121
2014 38 71 69 115
Total 114 278 284 496
Table 3, ESG shareholder proposal categories: This table shows the specific topics of the shareholder proposals which are part of the environmental, social and governance samples. The ‘miscellaneous’ categories contain shareholder proposals on a too broad variety of topics to categorize them separately.
Main topic Environmental Social Governance
Sub topics Environmental reporting 52 Political Spending 97
Chairman
Independence 127
Emission reduction
14
Lobbying and Political
Spending 72 Written Consent 51
Resource reduction 8 Human Rights 37 Special Meetings 36 Miscellaneous 40 Employment Rights 23 Declassify the Board 25
Animal Rights 18 Proxy Access 11
Health Care
3
Shareholder Rights
Plan (Poison Pill) 2
Miscellaneous 28 Miscellaneous 32
21
Table 4, Descriptive statistics: This table reports the descriptive statistics of the investigated variables, and the correlations between them, of the total sample. What these
variables measure can be found in Appendix C.
Firm characteristics Correlations
Firm characteristics Obs Median Mean Std. Dev Min Max 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 1. Market capitalization (billion $) 991 17.63 36.22 52.31 0.25 499.70 1.00
22 5.3 Methodology
Univariate analysis is employed first to describe the relation between shareholder proposals and each variable on its own. Tkac (2006) and Rojas et al. (2009) based their findings on univariate analysis as well. There will be a comparison made between firms who received a shareholder proposal on environmental topic and those who did not receive an ESG shareholder proposal, firms who received a shareholder proposal on social topic and those who did not, and firms who received a governance shareholder proposal and those who did not. To test for the statistical significance of differences between the target firms and their non-target peers, parametric two sample t-tests are employed. The two sample t-statistics are calculated as follows:
& = ' − ') *+ ) , + + )) ,) ( 3 )
Where ' is the characteristic mean of the target firm group and ') is the characteristics mean of the non-target firm group. + ) is the variance of the target firm sample and +)) the variance of the non-target sample. , and ,) are the sample sizes of respectively the target firm sample and non-target firm sample. Assuming the samples are normally distributed, the statistical significance of the t-statistics is given whether it absolute value is greater than the Student’s t-distribution value. Because the parametric t-test is based on a few assumptions, also a non-parametric Wilcoxon rank sum test is provided. The Wilcoxon test relaxes the assumption of normal distributions. The non-parametric test provides the possibility of inferences without the normality assumption. The Wilcoxon z-statistic is calculated as follows:
Where, "() = , (, + ,2 )+ 1)
Where, + = *, ,)(, 12+ ,)+ 1)
And is the sum of ranks assigned to observations of firms that received a shareholder proposal. The parametric and non-parametric test statistics and its respective statistical significance are provided by the software package EViews. Because the hypotheses, see Section 4, are organized in a two tailed manner, both a negative- and positive significant test statistics of the SRIO and CSR performance variables are enough to reject the null hypotheses.
5 = − "()+
23 For multivariate analysis I make use of regression models. Multivariate analysis examines the impact of a group of independent variables on a dependent variable. The dependent variable in the regression model is a dummy variable, firms did or did not receive a shareholder proposal, and thus the dependent variable is limited to take values of only zero or one. Hence, the model is dealing with a limited dependent variable. When estimating shareholder proposal probability, it is desired to get a value between zero and one (or zero and hundred percent). It is not appropriate to estimate a negative probability or a greater than hundred percent probability. Similar to Karpoff et al. (1996) and Schooley et al. (2010) a logistic regression will be estimated to investigate the effect of firms’ characteristics on shareholder proposal probability. The logistic regression is a non-linear regression model. One of the advantages of a logistic model is that the estimated independent variable will always take a value between zero and one. If instead of a logistic regression an ordinary least squares (OLS) model would be used, estimated values of the dependent variable below zero and above one should have been truncated to zero and one, respectively. A linear OLS model would result in too many observations for which the estimated probabilities would be exactly zero or one. Truncating values to zero or one will incorrectly increase the goodness of fit for the linear model.
24 6. Results
This section reports the results of both the univariate- and multivariate analysis. Whether the null hypotheses have to be rejected, or not, is discussed separately for the univariate and multivariate analysis. The final paragraph at Section 6.2 can be seen as the aggregate results of the univariate and multivariate analysis.
6.1 Univariate analysis results
Table 5 reports the mean and median test statistics for the ESG shareholder proposal target firms compared to the group of firms that did not receive an ESG shareholder proposal. The Wilcoxon rank sum statistics yield similar indications as the t-test statistics. For all three shareholder proposals types the firm size, measured by market capitalization and sales, is statistical significantly higher at the 1% level than from firms who did not receive an ESG shareholder proposal. The market capitalization and sales is for the group of firms targeted by shareholder proposal, of each issue type (E, S, or G), approximately three times higher compared to the non-target group of firms. This complies with the findings of Tkac (2006) and Schooley et al. (2010) that large companies are more exposed to shareholder proposals. The test statistics for the financial performance measures, market-to-book ratio, growth in sales and three year stock return, do not uniformly confirm the findings of Karpoff et al. (1996) and Cziraki et al. (2010). Karpoff et al. (1996) reported a negative relationship between financial performance and shareholder proposal probability. In my results the market-to-book ratio is at least by one test statistic significantly higher for all of the ESG samples. Only for firms that receive an environmental shareholder proposals, performance measured by market-to-book ratio is according to both test statistics significantly higher than non-target firms. Institutional ownership is smaller, at the 1% statistical significance level, for firms that are targeted by an ESG shareholder proposal compared to non-target firms. Firms who are targeted by either an environmental, social, or governance shareholder proposal, have approximately 10% less of institutional owners than non-target firms. This contradicts with Cziraki et al.’s (2010) finding of a positive influence of institutional ownership on shareholder proposal probability. Apparently, firms with lower institutional ownership exhibit higher exposure to ESG shareholder proposals. However not all of my univariate findings for the control variables comply to the papers’ findings in Table 1, most do seem to exhibit a relationship with shareholder proposals on ESG issues.
26
Table 5, Univariate analysis: This table reports the means of financial performance, ownership information, and corporate social responsibility scores for Fortune 250 firms
that are, and are not, targeted by ESG shareholder proposals between 2011 and 2014. The means of the non-target firms are based on year observations of firms that did not receive an ESG shareholder proposal during the respective year. Shareholder proposal target firms are separated into environmental, social, and governance groups. The means of the group are based on year observations of firms that received one or more shareholder proposal during the respective year on a respectively environmental, social, or governance topic. Firm characteristics are measured at 1 January of the year before the shareholder proposal is being voted on at the annual general meeting. The t- and
z-statistic both measure the statistical significance of the difference between a target firm group and the non-target firm mean and median, respectively (the median is not
reported). If there is a significant difference this is indicated with bolt letter type. The degree of statistical significance is indicated by asterisks, where *, **, and *** represent a statistical significance at the 10%, 5%, and 1% level, respectively.
Non-target firms Target firms
Environmental Social Governance
Firm characteristics Mean Mean t-statistic z-statistic Mean t-statistic z-statistic Mean t-statistic z-statistic
Market capitalization (billion $) 19.12 65.64 5.88*** 9.25*** 68.17 10.29*** 13.69*** 61.83 9.31*** 11.82***
Market-to-book ratio 1.89 2.81 1.65* 3.28*** 6.66 1.59 2.87*** 6.32 1.53 2.93***
Sales (billion $) 22.41 65.51 4.88*** 5.82*** 61.33 8.53*** 11.42*** 58.57 7.79*** 10.47***
Growth in sales (%) 8.95 6.81 -1.62 -0.90 7.57 -1.10 -1.92* 6.83 -1.96 -1.99**
Three year stock return (%) 9.99 10.19 0.11 0.39 9.76 -0.17 -1.07 10.23 0.16 0.9 Institutional ownership (%) 70.70 59.03 -8.24*** -7.19*** 61.38 -9.08*** -8.39*** 62.61 -7.39*** -7.07***
SRI institutional ownership (%) 47.24 49.10 2.48** 2.42*** 47.30 0.11 0.23 47.46 0.44 0.05 Emission reduction score 63.66 73.72 3.94*** 2.72***
Product innovation score 56.39 69.32 4.26*** 3.60***
Resource reduction score 67.38 76.52 3.36*** 3.62***
Weighted environmental score 64.61 76.69 4.59*** 3.74***
Employee wellbeing score 60.60 68.16 5.84*** 4.92***
Product responsibility score 53.87 54.90 0.50 0.42
Community score 66.98 67.72 0.43 0.01
Human rights score 59.02 68.14 3.60*** 3.51***
Weighted social score 64.38 73.02 5.35*** 3.98***
Board functions score 81.65 81.94 0.58 0.82
Board structure score 81.45 82.77 1.58 1.18
Vision and strategy score 60.37 71.56 5.23*** 5.17***
Shareholder rights score 61.96 63.49 0.80 0.53
27 6.2 Multivariate analysis results
Table 6 provides the results of the six regression models. All six models show for market capitalization the same positive 1% statistical significant results as the univariate analysis results. Thus, market capitalization positively influences all types (ESG) of shareholder proposal probability. The results on firm size are consistent with the findings of Tkac (2006) and Rojas et al. (2009) that firm size is positively related to shareholder proposal probability. Where they came to their results with an univariate analysis, I confirm that firm size is also positively related to ESG shareholder proposal probability in a multivariate analysis. The financial performance measured by market-to-book ratio show a positive relationship with shareholder proposal probability across all regression models. Only for models 3-6 the to-book ratio exhibits a statistical significant positive value. The positive values of market-to-book ratio are in contrast with Karpoff et al. (1996), which find that poor performance increases shareholder proposal probability. Institutional ownership is for all models negatively related to shareholder proposal probability. But in contrast to the univariate analysis only at a statistical significant level, 1%, for environmental shareholder proposal probability. Hence, institutional ownership seems to be a negative factor only in environmental shareholder proposal probability. Again, the negative relationship does not comply with the finding of Cziraki et al. (2010). The persistence of the negative relationship between institutional ownership and ESG shareholder proposal probability indicates that the finding of Cziraki et al. (2010) does not hold for ESG shareholder proposals filed at U.S. firms.
28 multivariate analysis. Regression model 5 and 6, with governance shareholder proposal probability as dependent variable, do not report significant values for the institutional ownership variables nor for the governance categories of CSR performance. Again, where Schooley et al. (2010) find that board independence and board structure is positively related to governance shareholder proposal probability, this relationship is not observed at a statistical significant level in my multivariate analysis. The pseudo ) for all the models are quite small at all below 20%. According to Brooks (2014) small pseudo ) are often the case for logistic regression models. By removing stepwise variables in the regression models I checked, with help of the pseudo ), whether there were spurious variables included in the models. Since the pseudo ) declined in all models when there were variables removed, none of the variables are spurious. Further, because the lowest percentage of correct predictions is 66,7 among all models, I regard that the regression models are doing a good job on predicting the actual ESG shareholder proposal probabilities.
29 null hypothesis cannot be rejected as well. Although the hypothesis testing did not result in a rejection of one of them, there are some factors found that seem to matter for ESG shareholder proposal probability.
7. Robustness analysis
To verify the robustness of the multivariate analysis from Section 6.2, a few other regression models and estimation methods are conducted. To verify the robustness of the control variables, firm size and financial performance, MCAP and MTB are replaced in the logistic regressions (see equation (1) and (2) for the original logistic regressions). MCAP and MTB are replaced with the amount of sales (SALES) and the three year cumulative stock return (3SR). The 3SR variable includes the change in stock price and dividend yields. The amount of sales measures the size of a firm in a different fashion. Three year cumulative stock return is a different method of measuring the financial performance. Where MTB is
Table 6, multivariate analysis results. This table reports the characteristics related to the probability of firms receiving an ESG shareholder proposal based on logistic regressions. Firm characteristic are measured at 1 January of the year before the shareholder proposal is being voted on at the annual general meeting. The dependent variable is a dummy which is equal to one if a firm is targeted by a shareholder proposal (on environmental, social, or governance topic), and equal to zero if the firm is not targeted by an ESG shareholder proposal. *,**, and *** indicate whether the parameter is statistically significant at the 10%, 5%, and 1% level, respectively. The pseudo- ) and percentage of correct predictions indicate how well the regression models fit the data.
Regressors Environmental Social Governance
Regression number 1 ) 2 ) 3 ) 4 ) 5 ) 6 )
C -3.189** -3.284*** -0.864 -0.897 -1.770 -0.956
Market capitalization (billion $) 0.016*** 0.016*** 0.025*** 0.025*** 0.020*** 0.021*** Market-to-book ratio 0.017 0.017 0.008* 0.007** 0.006** 0.006** Institutional ownership (%) -0.033*** -0.033*** -0.010 -0.011 -0.010 -0.010 SRI institutional ownership (%) 0.060*** 0.060*** -0.001 0.000 0.006 0.007 Emission reduction score 0.004
Product innovation score 0.004 Resource reduction score -0.002
Weighted environmental score 0.008
Employee wellbeing score 0.012*
Product responsibility score -0.001
Community score -0.005
Human rights score -0.003
Weighted social score 0.004
Board functions score 0.007
Board structure score 0.001
Vision and strategy score 0.005
Shareholder rights score 0.000
Weighted governance score 0.002
Correct predictions (fraction) 0.717 0.698 0.725 0.717 0.667 0.678
30 a measure of recent performance, three year cumulative stock return takes financial performance over a longer range of years into account. It is assumed that stock prices reflect the financial performance of a firm. SALES and 3SR are both extracted from Datastream. The results of the estimated logistic models with substituted performance measures can be found in Appendix D, Table 8. SALES and 3SR have the same relationship signs as MCAP and MTB. The statistical significance levels of 3SR vary from MTB in the main regression models. Also a few CSR performance dimensions now exhibit statistical significant levels. Although, the small degree of difference with the main regression models does not convince to adjust the conclusion drawn at Section 6. Ideally, the robustness of the institutional ownership and CSR performance variables is verified as well. For example, CSR performance measured by ESG Asset4 could be replaced by CSR scores from the KLD database. There are no resources available to gain access to databases to collect substitutes for the SRIO and CSR variables. Therefore I have to rely on the correctness of ESG Asset4 and Orbis. However, the SRIO can be constructed with the same source of data but with another calculation method to verify its robustness. The SRIO variable is replaced with a variable that measures the amount of percentages shares held by green institutional investors. Recall from Section 5.1 that the SRIO variable measures the greenness of institutional investors. A drawback of this substitute variable is that it has a correlation of 0.5 with the IOWN variable. Appendix D, Table 9 reports the results. Despite the correlation, the substitute variable shows for all regression models, except model 5, statistical significant values. The values from other statistical significant variable, like market capitalization and employee wellbeing, have shrunk. While the explanatory power of the institutional ownership variables is now negative at all of the models at a statistical significant level.
To test the validity of using the logistic regressions, and to provide more easily interpretable results, the six main regression models are re-estimated with the more common pooled ordinary least squares (OLS) method. For this linear model, a 1-unit increase in 6", , for example, causes an 7# increase in probability. The results can be found in Appendix D, Table 10. The results of the pooled OLS estimating models comply with the results of the logistic regression approach. Employee wellbeing is again the only statistical significant CSR category in explaining social shareholder proposal probability. The vision and strategy score is now 10% statistical significant for governance shareholder proposal probability. The signs of the control variables and SRIO variable are equal to those in the logistic regression models. Also for the pooled OLS the SRIO variable is only significantly related to environmental shareholder proposal probability.
31 shareholder proposal. Here the sample is constructed with firms that received a shareholder proposals on a specific topic and firms that did not receive a shareholder proposal on the respective topic. So for example: if Berkshire Hathaway received only a shareholder proposal on environmental topic and Apple received only a shareholder proposal on governance topic, both observations will be included in the environmental sample where Berkshire’s dependent variable (dummy) will get a value of one, and Apple’s dependent variable will get a value of zero. How the robustness sample is constructed can be found in Appendix E. The results of logistic regression estimated with an alternative sample can be find in Appendix D, Table 11.The results do not vary from the results with the original sample. The non-target part of the environmental robustness sample also included firms that received shareholder proposals on social or governance topic as control observations. This strengthens the finding that SRI institutional ownership is specifically related to environmental shareholder proposal probability.
Although some robustness models show differing results compared to the multivariate analysis in Section 6.2, the most important findings seem to hold. Employee wellbeing is the only CSR category that exhibits significant values in all robustness models, and also the greenness of institutional ownership exhibits a positive statistical significant finding in all robustness models for environmental shareholder proposal probability only. Also, the robustness analysis do not provide evidence for rejecting either of the null hypotheses.
8. Conclusion
This paper analysed whether socially responsible institutional investors and corporate social responsibility (CSR) performance of a firm can be associated with the firms’ probability of receiving a shareholder proposal on ESG issues. In particular, the probability of receiving a shareholder proposal on environmental, social, or governance issues is all examined separately. I measure the social responsibility of institutional investors by the percentage of biggest institutional owners of a firm that have signed the United Nations supported Principles of Responsible Investing. The CSR performance of a firm is measured by several environmental, social, and governance performance scores that are provided by ESG Asset4.
32 100 on employee wellbeing while firm B has a score of 50 on employee wellbeing. My results indicate that firm A has a higher probability of receiving a social shareholder proposal. Theoretically however, leaving all else equal, firm’s A score on employee wellbeing does not leave room for further improvements that are proposed by shareholders. The score of 100 is namely the highest performance score possible. Other CSR performance dimensions do not seem to have significant explanatory power in ESG shareholder proposal probability. The robustness analysis confirm the reliability of these findings. Since the vast majority of CSR categories is found to be not significantly related to different types of shareholder proposal probability, the first null hypothesis cannot be rejected. Thus, the CSR performance of the firm is not affecting the probability of receiving a shareholder proposal on ESG issues. Further, because the greenness of institutional owners is not a significant factor for all types of shareholder proposal probability, the second null hypothesis cannot be rejected either. Thus, the greenness of institutional owners of a firm is not affecting the probability of receiving a shareholder proposal on ESG issues. Further, the finding of Tkac (2006) and Schooley et al. (2010) that firm size positively affects shareholder proposal probability is confirmed for ESG shareholder proposal probability as well. The negative effect of financial performance, found by Karpoff et al. (1996) does not seem to hold for ESG shareholder proposal probability. The positive effect of institutional ownership on shareholder proposal probability reported by Cziraki et al. (2010) also does not seem to hold for ESG shareholder proposal probability. However, the effect of the greenness of institutional investors on environmental shareholder proposal probability and the effect of employee wellbeing on social shareholder proposal probability do obviously hold in the robustness analysis. Thereby, these are the main findings of this paper.
Although the scientific findings of this paper are few, it can encourage others to examine the relationship between CSR and SRI ownership of a firm and the probability of receiving a shareholder proposal on ESG issues. Where I do investigate the effect of a group of CSR’s categories on the closest respective topic of shareholder proposals, others can follow a different approach. For example, one drawback of my approach is that the samples have an unequal distribution of different shareholder proposal categories. With a more elaborate data set of observations it could be possible to construct more equal distributed samples. This would reduce the possibility of over- or underestimating the explanatory power of a CSR variable that is closely related to one type of shareholder proposal. Also, I leave it to others to explain the counterintuitive positive effect of employee wellbeing on social shareholder proposal probability.
33 References
Ba, S., Lisic, L., Liu, Q., Stallaert, J., 2013. Stock market reaction to green vehicle innovation. Production & Operations Management, 22 (4), 976-990.
Borghesi, R., Chang, K., Mehran, J., 2016. Simultaneous board and CEO diversity: does it increase firm value? Applied Economics Letters, 23 (1), 23-26.
Brooks, C., 2014. Introductory econometrics for finance. Cambridge University Press, Cambridge (United Kingdom).
Cai, Y., Jo, H., Pan, C., 2011. Vice or virtue? The impact of corporate social responsibility on executive compensation. Journal of Business Ethics, 104 (2), 159-173.
Cremers, M., Ferrell, A., 2014. Thirty years of shareholder rights and firm value. Journal of Finance, 69 (3), 1167-1196.
Cziraki, P., Renneboog, L., Szilagyi, P., 2010. Shareholder activism through proxy proposals: the European perspective. European Financial Management, 16 (5), 738-777.
Dam, L., Scholtens, B., 2012. Does ownership type matter for corporate social responsibility? Corporate Governance: An International Review, 20 (3), 233-252.
Dam, L., Scholtens, B., 2013. Ownership concentration and CSR policy of European multinational enterprises. Journal of Business Ethics, 118 (1), 117-126.
Dimson, E., Karakaş, O., Xi, L., 2015. Active ownership. Review of Financial Studies, 28 (12), 3225-3268.
Flammer, C., 2015. Does corporate social responsibility lead to superior financial performance? A regression discontinuity approach. Management Science, 61 (11), 2549-2568.
Fombrun, C., 2005. Building corporate reputation through CSR initiatives: Evolving standards. Corporate Reputation Review, 8 (1), 7-11.
Gillan, S., Starks, L. 2000. Corporate governance proposals and shareholder activism: the role of institutional investors. Journal of Financial Economics, 57 (2), 275-305.
Harjoto, M., Laksmana, I., Lee, R., 2015. Board diversity and corporate social responsibility. Journal of Business Ethics, 132 (4), 641-660.
34 Henard, D., Dacin, P., 2010. Reputation for product innovation: Its impact on consumers. Journal of Product Innovation Management, 27 (3), 321-335.
Hillman, A., Keim, G., 2001. Shareholder value, stakeholder management, and social issues: What’s the bottom line? Strategic Management Journal, 22 (2), 125-139.
Inderst, G., Kaminker, C., Stewart, F., 2012. Defining and measuring green investments: implications for institutional investors' asset allocations. OECD Working Papers on Finance, Insurance and Private Pensions, 24, OECD Publishing, Paris.
Jacobs, B., 2014. Shareholder value effects of voluntary emissions reduction. Production & Operations Management, 23 (11), 1859-1874.
Jacobs, B., Singhal, V., Subramanian, R., 2010. An empirical investigation of environmental
performance and the market value of the firm. Journal of Operations Management, 28 (5), 430-441. Jianxin, C., 2005. Understanding the endogeneity between firm value and shareholder
rights. Financial Management, 34 (4), 65-76.
Jiraporn, P., Yixi, N., 2006. Dividend policy, shareholder rights, and corporate governance. Journal of Applied Finance, 16 (2), 24-36.
Jo, H., Harjoto, M., 2011. Corporate governance and firm value: The impact of corporate social responsibility. Journal of Business Ethics, 103 (3), 351-383.
Jo, H., Harjoto, M., 2014. Analyst coverage, corporate social responsibility, and firm risk. Business Ethics: A European Review, 23 (3), 272-292.
Karpoff, J., Malatesta, P., Walking, R., 1996. Corporate governance and shareholder initiatives: empirical evidence. Journal of Financial Economics, 42 (3), 365–395.
Kitzmueller, M., Shimshack, J., 2012. Economic perspectives on corporate social responsibility. Journal of Economic Literature, 50 (1), 51-84.
Levit, D., Malenko, N., 2011. Nonbinding voting for shareholder proposals. Journal of Finance, 66 (5), 1579-1614.
Meziana, A., 2014. Investing with environmental, social, and governance issues in mind: From the back to the fore of style investing. Journal of Investing, 23 (3), 115-124.
