Next Article in Journal
Sustainability of Commercial Banks Supported by Business Intelligence System
Previous Article in Journal
Comparison of Potential Contribution of Typical Pavement Materials to Heat Island Effect
Open AccessArticle

Identifying Corporate Sustainability Issues by Analyzing Shareholder Resolutions: A Machine-Learning Text Analytics Approach

1
Koppelman School of Business, Brookyn College of the City University of New York, Brookyn, NY 11210, USA
2
Gabelli School of Business, Fordham University, New York, NY 10023, USA
*
Author to whom correspondence should be addressed.
Sustainability 2020, 12(11), 4753; https://doi.org/10.3390/su12114753
Received: 7 May 2020 / Revised: 7 June 2020 / Accepted: 8 June 2020 / Published: 10 June 2020
(This article belongs to the Section Economic and Business Aspects of Sustainability)

Abstract

Corporations have embraced the idea of corporate environmental, social, and governance (ESG) under the general framework of sustainability. Studies have measured and analyzed the impact of internal sustainability efforts on the performance of individual companies, policies, and projects. This exploratory study attempts to extract useful insight from shareholder sustainability resolutions using machine learning-based text analytics. Prior research has studied corporate sustainability disclosures from public reports. By studying shareholder resolutions, we gain insight into the shareholders’ perspectives and objectives. The primary source for this study is the Ceres sustainability shareholder resolution database, with 1737 records spanning 2009–2019. The study utilizes a combination of text analytic approaches (i.e., word cloud, co-occurrence, row-similarities, clustering, classification, etc.) to extract insights. These are novel methods of transforming textual data into useful knowledge about corporate sustainability endeavors. This study demonstrates that stakeholders, such as shareholders, can influence corporate sustainability via resolutions. The incorporation of text analytic techniques offers insight to researchers who study vast collections of unstructured bodies of text, improving the understanding of shareholder resolutions and reaching a wider audience.
Keywords: machine learning; shareholder resolution; sustainability; sustainability reporting; text analytics machine learning; shareholder resolution; sustainability; sustainability reporting; text analytics

1. Introduction

Environmental and climate change trends, as well as a general concern about the environment and environmental impact of business operations, have provoked public demand for corporate accountability and transparency [1,2,3,4]. In light of such pressure, companies are moving sustainability to the top of the corporate agenda [5,6,7,8]. The C-Suite is recognizing the need for more transparency regarding corporate sustainability practices and their impact not only on the environment, but on society as well [9,10]. To educate the public and stakeholders, sustainability information is being integrated into corporate disclosures (e.g., annual reports). In addition, business strategies are being redrawn to result in better environmental and social performance [8,11].
Corporate sustainability encompasses a range of aspects, among them: Corporate reputation, sustenance and expansion of economic growth, customer relationships, shareholder value, prestige, and the quality of products and services [12]; implementing and demonstrating social and environmental concerns in business transactions [13]; satisfying the current needs of stakeholders without compromising the capability to meet the needs of the future [14]; and demonstrating the ability to avail opportunities and manage risk in terms of economic, environmental, and social dimensions, so as to create long-term shareholder value [15]. Despite the multiplicity of definitions, the common thread is an underlying corporate commitment to sustainable economic development and improved quality of life for employees, the community, and society [16].
Companies demonstrate commitment to sustainability by adopting and communicating practices to stakeholders. This practice is referred to as sustainability disclosure or sustainability reporting [3,17,18,19]. While financial reporting covers the financial aspects of a company’s performance, sustainability reporting covers the non-financial aspects [20,21]. The triple bottom line perspective [22] emphasizes three facets in sustainability reporting: Economic, environmental, and social growth [23,24,25,26]. This view suggests that companies have a responsibility to not only make financial profits and maintain economic growth, but also consider their environmental and societal impacts [25,27]. More and more companies have started publicly disclosing their sustainability practices [28].
Companies often communicate sustainability information along with financial information in their annual sustainability reports [29]. These reports contain information on the environmental, economic, and social impacts from an organization’s day-to-day activities [30,31]. Research has concentrated on how companies address the challenge of sustainability through disclosure in sustainability reporting. Some studies focus on the frequency of reporting [28,32], while others deploy qualitative content analysis [27] or quantitative text analysis [27,29,31,33,34,35,36] of reporting practices. Most of these studies have only considered a limited number of reports. In contrast, [31] employed text mining techniques to conduct topic modeling on 9514 sustainability reports published between 1999 and 2015. The study applied Latent Dirichlet allocation (LDA) to identify themes within large collections of documents [37]. In recent years, a trend has emerged in which shareholder activists form networks to empower shareholders and magnify their voices [38,39] to demand corporate responsibility—particularly environmental responsibility [25]. Activists continue to criticize large corporations (i.e., BP, Chevron, Exxon, Shell, Monsanto) for having a negative impact on the environment [40]. Given the pressure to improve corporate environmental performance [41], combined with the heavy price for environmental corporate misdeeds [42], it is no surprise that environmental issues are one of the most cited reasons for shareholder resolutions to be presented and voted on through the proxy vote at annual meetings [43]. Additionally, with the heightened public awareness of climate change, companies are also realizing the significance of sustainability on local economies and cultures [13,14,44,45]. Investors and stakeholders are increasingly interested in the effects of environmental, social, and governance (ESG) issues on the sustainability of business models as they focus on their ability to generate long-term value [46,47,48,49]. Therefore, it is imperative that a socially responsible company integrate both the interest of stakeholders and the well-being of the environment [50,51,52,53]. These efforts exploit growth opportunities while giving back to the community with environmental improvements [54,55,56,57]. A consequence of increased stakeholder involvement is increased shareholder activism. This is reflected through shareholder resolutions on sustainability issues, with the number of resolutions gradually increasing [58]. Shareholder resolutions offer a unique lens to study corporate sustainability practices and shed light on the most important issues [43,58,59,60].
Our motivation for the current research arises from certain premises: First, we acknowledge that there is public demand for increased corporate commitment to sustainability; second, in response to the increased public demand, companies are eager to be involved in environmental, social, and governance activities; and third, companies communicate such involvement through avenues such as shareholder resolutions and/or annual reports. Our exploratory research, therefore, deploys a machine-learning-based text analytics approach to identify and analyze the key sustainability issues in the resolutions proposed by shareholders.
We contribute to current research on sustainability reporting in many ways. First, we use current data on shareholder resolutions published as recently as 2019 and look at an extended timeframe to include multiple years. Second, we analyze a significantly large number of shareholder resolutions. By extending the timeframe and number of resolutions, we extend the coverage to many sectors and many years. This also enables us to highlight the evolution of topics over a period of time, as well as explore inter-sectoral distributions. Third, to the best of our knowledge, machine learning-based text analytics methodology has not been extensively used to analyze shareholder resolutions. This technique is essential, considering the data is mostly textual. Our study fits well in this niche. Fourth, this study looks at corporate sustainability from the lens of the shareholder (rather than from the management). Text analysis offers the ability to examine the textual information without assuming a pre-defined list of terms. In analyzing shareholder sustainability resolutions, we focused on the following: Identification and evolution of sustainability issues; incorporation of environmental, economic and social aspects in the resolutions; and the variances in sustainability resolutions (and practices) between sectors [29,31]. We explore the research question:
Do shareholder resolutions reflect corporate sustainability concerns in terms of environmental, social, and governance aspects?
The rest of the paper is organized as follows: Section 2 follows with a background on corporate sustainability and shareholder resolutions; Section 3 describes the methodology; Section 4 presents the analysis and results; Section 5 offers an integrated discussion; and lastly, Section 6 concludes with the scope and limitations of the current research and offers directions for future research.

2. Research Background

2.1. Sustainability

In its 1987 report, the Brundtland Commission defined sustainability as an initiative for meeting the needs of the present, without compromising the ability of future generations to meet their own needs [61,62]. This view of sustainability has been accepted universally [31,63].
Sustainability, applied in an organizational context, is termed corporate sustainability. This term is variably defined [63]. While some research focuses on environmental aspects of sustainability, others focus on the social aspects, and yet others adopt an integrated view by combining the environmental, social, and economic aspects [63,64,65]. We envision corporate sustainability at the junction of environmental performance, economic contribution, and social responsibility [14,66] as posited by the triple bottom line perspective [22].
Environmental practices are concerned with the rate of utilization of natural resources and of emissions in the eco-system [14]. Conservation of resources and waste management are important elements [67,68,69]. Environmental performance is indicated by resource depletion, energy use, air emissions, biodiversity, noise, solid waste, transport, and water use and discharge [66].
Social practices are those that that are directed at maintaining quality of life in the community [14,69]. Some of the aspects include corporate philanthropy [67], corporate citizenship, social partnership, and social sponsorship [66]. They also cover human capital elements like employee training programs, health programs, child-care, and others [68].
Economic practices include corporate governance, risk management, emergency management, compliance [67,69], and economic profitability, and economic equity [70]. Organizations face pressure to be sustainable from several entities such as government [2,66,71], customers and employees [72], and management [73]. Organizations need to embed sustainability within their overall strategy in order to be successful [74,75]. One of the ways in which organizations manifest their sustainable activities is through publication of corporate sustainability reports [63]. Another is to look at shareholder resolutions that concern sustainability [38].

2.2. Shareholder Resolutions

Recently, shareholders have increased their participation in managerial oversight through private communication with managers, proxy contests, and shareholder proposals [58,76]. Due to their low cost, shareholder proposals give parties otherwise left out of the corporate governance process, a chance to voice their concerns over the management of the firm [60,76].
The process of shareholder proposal is governed by the Securities and Exchange Commission’s (SEC) Shareholder Proposal Rule 14a-8, which regulates proxy voting. Rule 14a-8 gives shareholders the opportunity to submit initiatives over issues that directly affect the firm as a profit-making institution [76,77,78]. Shareholder activists can try and influence corporate behavior by trying to exercise voting rights on shareholder proposals or by entering into dialogue with companies on a particular issue [38]. They can sell shares over an issue, prepare and lodge shareholder resolutions, vote on proposals, and engage in dialogue with the management on particular issues.
Some academic research has examined the trend of shareholder activism for corporate social responsibility (CSR) and/or environmental and social issues [38,79,80]. In the United States, shareholder activism has existed for decades [38] and shareholders have historically pressured companies to address social issues through the proxy voting process for many years [81]. The movement gained popularity in 1946 with the SEC adopting rule 14a-8, requiring companies to include shareholder resolutions in their proxy statements [38] in addition to annual meetings. In the 1960s and 1970s, shareholders pressured companies about issues including employment and training, civil rights and equal opportunities, and weapons production [80]. These efforts resulted in changes to the rules for proposals to shareholder meetings. In 1970, the US courts issued a ruling against the SEC after it disallowed a proposal intending to force the Dow Chemical Company to stop producing napalm during the Vietnam War [38]. This decision paved the way for other social responsibility proposals to follow [81]. For example, Ralph Nader’s landmark Campaign GM and Project on Corporate Responsibility used shareholder proxy to pressure General Motors (GM) into being more responsible of societal needs in areas like product safety, environmental pollution, and employment discrimination [82,83,84].
During the 1980s, shareholder activism in the U.S. tended to swing toward anti-takeover activities. This was a result of public focus on making companies more competitive and adopt social and environmental responsibility [38]—an example is how, in the wake of the Exxon Valdez disaster, the investor and environmental activist alliance CERES was formed, bringing public focus on the need for improved environmental disclosure by companies to stakeholders [38]. The groups also explored how the management of these issues could be better integrated into the company [38]—an example is the mandatory implementation by companies of international standards like ISO 14001 and the International Labor Standards (ILO) [85]. Overall, in the late 1990s, the phenomenon of activism was characterized by a spirit of cooperation, rather than confrontation [86], directing the focus on CSR with mutual engagement and stakeholder dialogue [38,87].
Since most shareholder proposals are advisory in nature, management is not required to implement proposals, even if they receive majority support (see [88,89,90] for discussion of rules). However, while historically, corporate meetings were only a nominal event to symbolically implement executive intentions, with the increase in social activism, these meetings have taken on a new light [38,39,60]. The Institutional Shareholder Services (ISS) published that, in the face of corporate scandals and reforms, disclosure of social and environmental issues has come to the forefront of the corporate governance agenda [60,91]. This opened the door for shareholders to file resolutions and obligated management to distribute these resolutions with proxy statements. Resolutions became a more effective tool for shareholders to gain management attention [43,58,60]. Some management consulting firms like PricewaterhouseCoopers (PwC) even published manuals to help CEOs anticipate and deal with social and environmental issues at annual shareholder meetings [43,92].
The socially oriented shareholder activism movement has had success with several initiatives. For one, during the apartheid reign, shareholder activism forced firms to withdraw from South Africa [93]. Similarly, McDonalds was forced to stop using polystyrene clamshell packaging materials [38]. More recently, companies were pushed towards more disclosure of greenhouse gas emissions and climate change [94]. Corporations continue to face societal pressures to improve their business citizenship practices and embrace concepts like social justice, sustainability, and the triple bottom line [77,95,96,97]. The shareholder resolutions related to improvements in the performance of a company, specifically with regard to the triple bottom line, have a marked effect on corporate management [58,97,98,99,100]. Through these, shareholders urge the management to adopt a proactive rather than a reactive strategy, while still maintaining or enhancing their reputation and competitive advantage [97,99].
In addition to resolutions, activist stakeholders can deploy tactics such as media campaigns, negotiations, lawsuits, boycotts, lawsuits, and lobbying legislators, in their efforts to pressure companies into altering their social agendas [38,60,77,101]. There is interactive engagement of both shareholders and companies [38,77] in the area of CSR [38,58]. In the context of interaction between companies and shareholders, there is speculation from environmental and social activists on the power and influence of shareholders. Many shareholders are turning into active lobbyists [38,43,58,60]. They influence corporate behavior by enforcing ownership rights through the process of preparing or voting on shareholder proposals and engaging in direct dialogue with management [38,43,77]. Some academic research has examined the trend of shareholder activism for CSR and/or environmental and social issues [79,80].
Shareholder resolutions have attracted activists who are interested in influencing corporate behavior [79]. Ever since the decline of corporate takeover activity in the late 1980s, shareholder resolutions have been a venue for fostering changes by the board of directors as well as by top management. Such resolutions have become a viable instrument for activists to make demands on management for changes in corporate practices [79].
Changes desired by shareholder activists in the form of resolutions can range from managerial performance or governance issues to social issues [79]. Some examples of social issues addressed are human rights, refraining from military contracting, and changing executive compensation [79]. Considering the importance of shareholder resolutions in disseminating sustainability-related ideas, it is imperative to elicit key insights from this large corpus of information. In the following section we discuss the relevance of using text analytics as a machine learning technology, specifically for analyzing shareholder resolutions.

2.3. Text Analytics and Machine Learning

Text analytics as a machine learning approach has gained popularity with the increasing availability of electronic documents from a multitude of sources. Electronic data, in general, can be structured or unstructured. Structured data is well-defined, organized, and easily searchable; while unstructured data comprises complex data that are of free form (little to no structure), comprising multiple types including audio, video, and/or graphics. Examples of unstructured data sources include the World Wide Web, electronic corporate documents such as annual reports/resolutions, digital libraries, online forums, online news channels, chat rooms, electronic mail and blog repositories, among others [102]. Considering the ubiquity of unstructured data, extracting knowledge from these is critical as a research and application area.
Natural language processing, machine learning, and text analytics automatically classify and discover patterns from electronic documents. By using natural language processing, text analytics can transform unstructured data into a structured form that is suitable for analysis and for applying machine learning algorithms. By using text analytics, researchers can assess various dimensions of core concepts that they want to pursue in the unstructured data. Most text analytics studies use thesaurus-like dictionaries comprised of words or phrases with shared meanings [103]. In order to analyze a corpus, the method involves assessing the frequencies of entries and categories and evaluating the relative importance of central concepts in the text. The advantage of text analytics is its capacity to analyze large volume of data [104]. In the current research context, analyzing shareholder resolutions offers an appropriate domain. For one, it is a large corpus of unstructured text. It also contains important sentiments of shareholders and presents an appropriate repository to explore. Companies communicate their commitment to various public causes including sustainability through reports and resolutions. Therefore, shareholder resolutions present an important research area worth exploring. The methodology of text analytics has been used in research on analyzing unstructured textual information such as tweets on vaccination [105], legal petitions [106], or health blogs [107].

3. Methodology

This exploratory study therefore examined shareholder resolutions of companies to elicit the broad sustainability practices and trends using machine learning-based text analytic approaches. The source of the resolutions is the climate and sustainability shareholder resolutions database from CERES, a sustainability, non-profit organization with numerous sustainability-related databases [108,109,110,111,112]. The machine learning-based text analytics approach enables the processing and study of a large corpus of textual data in an efficient manner [31,36,113]. The insight gained from the study alerts companies’ management, stakeholders, ESG proponents, policy makers, and others as to the implications and future direction of ESG in companies.
We retrieved approximately 2100 records of the shareholder resolutions (https://www.ceres.org/resources/tools/climate-and-sustainability-shareholder-resolutions-database) using a Python selenium package from www.ceres.org. Our primary sample included publicly traded firms from 2009 to 2019. Out of the total of 2100 records, we had to eliminate 363 records due to missing/incomplete content or non-functional links. The final data set consisted of 1737 resolutions from 545 companies (as of 10 April 2019). We present several descriptive analyses from processing the extracted text data.
In addition to the overall analysis of resolutions for sustainability practices, we pursued an additional exploratory question relating to the potential to predict the voting status of a resolution. That is, given the features from the resolutions, can voting status be predicted? In the case of public companies in the United States, a shareholder resolution is a proposal submitted by shareholders for a vote at the company’s annual meeting. The CERES shareholder resolution database contains several categories of the voting status of resolutions—such as “Vote”, “Withdrawn”, “Omitted”, “Filed”, “No Vote”, “Sent”, and “Floor Proposal.” For the purpose of our research, we aggregated the categories other than “Vote” and “Withdrawn” into “Other”. Therefore, we considered the three voting statuses of: “Vote” to denote resolutions that were submitted for vote; “Withdrawn” to denote resolutions that were withdrawn; and “Other” to denote resolutions in all other categories.

3.1. Descriptive Data Analysis

Table 1 presents the top 10 companies with most resolutions in the database. Dominion Energy, Inc., Chevron Corporation, and Exxon Mobil Corporation contributed about 8% of total resolutions. Interestingly, Amazon.com Inc. accumulated 20 resolutions and ranked fourth in the dataset. This data is an indication that the shareholders of these companies are aware of and pay attention to climate and sustainability issues. It may also indicate that these companies are more transparent when disclosing their shareholder resolutions regarding climate and sustainability to the public.
Table 2 presents the 64 industries that filed shareholder resolutions. The top five sectors are: (1) Oil, Gas & Consumable Fuels; (2) Electric Utilities; (3) Hotels, Restaurants & Leisure; (4) Multi-Utilities; and (5) Food Products. These five industries account for 686 out of the 1737 resolutions. According to Figure 1, the Oil, Gas & Consumable Fuels sector contributed to nearly 30% of the total resolutions. The top three sectors contributed to over 50% of the total resolutions.
Figure 2 confirms that the Oil, Gas & Consumable Fuels sector filed the greatest number of resolutions during this timeframe. This sector peaked in 2016 as a result of increased awareness and political and social activism surrounding climate change during this period. Overall, resolutions started to decline after 2017. The stacked chart in Figure 3 shows the dominance of the Oil, Gas & Consumable Fuels sector over the others in terms of the volume of resolutions filed. This is followed by the Electric Utilities sector.
Table 3 displays the distribution of voting status over the study period. The voting status was extracted from each resolution. As mentioned, the voting statuses were coded as “Voted”, “Withdrawn”, or “Other.” Most of the time, the resolution is either voted or withdrawn. This implies that the resolution was voted upon or withdrawn from consideration. However, the database is limited in that we do not know the nature of the vote (whether it passed). The “Other” status comprises the miscellaneous statuses like “Omitted”, “Filed”, “No Vote”, “Sent”, and “Floor Proposal”.
Figure 4 shows the distribution of voted resolutions by year. The trend fluctuated from 2009 to 2013. However, there is a sharp upward trend from 2014 to 2018. Its dramatic rise indicates that shareholders escalated the use of resolutions as a lever to alert management to corporate climate change, ESG, and sustainability initiatives and practices. During this period of activism, resolutions tended to get a vote. As of 2019, the trend line drops due to limited data availability through April 2019.

3.2. Text Analytics

Next, we apply machine learning-based text analytics to study the resolutions. Data pre-processing was undertaken prior to applying the machine learning-based text analytics. The retrieved data for each resolution was saved as a text file prior to vectorization. A few resolutions had to be removed for lack of clarity/relevance and/or non-functional links.
Predictive models developed with text data pose several challenges to the modeling process. First, textual data cannot be used as inputs in many mathematical models. Therefore, a natural language processing (NLP) system was implemented to transform the textual content into integral elements for further analysis. Second, text-based data sets are larger in size as compared to numerical data sets. Therefore, a successful model necessitates information extraction by identifying the most relevant pieces of the data. In the data pre-processing stage, the scraped, raw summary section was transformed into plain text documents through the elimination of punctuation, spaces, numbers, and standard stop words using the Natural Language Toolkit (NLTK). Next, the text was converted to lower case using NLTK and TextBlob (https://textblob.readthedocs.io/en/dev/).
Following this, the lemmatization technique was applied to convert the words into their root form (for example, “bought” and “buying” were substituted with “buy”). Lemmatization works by grouping together the different inflected forms of a word, which enable it to be analyzed as a single term. In this way, it brings context to the words. The pandas package (https://pandas.pydata.org/) was applied to filter and manipulate scraped data into data frames for analysis. The sklearn package (https://scikit-learn.org/stable/index.html) was used to refine the results. The maximum number of features was set to approximately 4000. Tokenized words included those with more than four characters. These efforts reduced the downstream of a generated model.
The ‘term frequency-inverse document frequency’ (TF-IDF) technique was used to compute the weight of each term to signify its importance in a document. This is an information retrieval technique that assigns a weight to a term’s frequency (TF) and its inverse document frequency (IDF). Each term is given these two scores. The weight of the term is then calculated as a product of these scores. A more detailed explanation of the technique is given in the Results section. The KMeans clustering model was applied to elicit the key sustainability concepts. KMeans clustering is one of the popular machine learning algorithms deployed to classify cases based on similarity measures (that is, the distance between cases). It is typically used in pattern recognition and classification domains. Seven clusters were identified using a word cloud package. This was followed by classification with the KNN classifier (a supervised machine learning algorithmic method), wherein the data was split into two parts to measure the success of the classifier (training and testing). Table 4 summarizes the key variables extracted from the corpus of text data.
CERES tracks shareholder disclosures filed by the investor network participants pertaining to enterprises’ sustainability issues, including energy consumption, climate change, water scarcity, and sustainability initiatives reporting. These disclosures are part of broader investor efforts to address the full range of ESG issues. The disclosures are filed by many of the largest public pension funds, foundations, and religious, labor, and socially responsible investors in the U.S. Many of the investors are members of CERES’ Investor Network on Climate Risk and Sustainability. A resolution decision variable (label) is used as the response variable when building a prediction model. The “status” variable was transformed into a numerical integer to suit the machine learning algorithm. The transformed labels for “status” are Vote = 1, Withdraw = 0, and Others = 2. This was labeled as voting status indicator. A critical factor in this study is that the textual data source contributes to the forecasting of the voting status of the shareholder resolution. The shareholder can examine the initiatives that may be discussed, as well as open it for a vote at the corporate level. The study’s predictive model focused on the main text body of disclosures. It contains a narrative explanation of environmental issues, relative impacts, and future directions related to coping with such problems. The goal is to predict the voting status of a given disclosure to learn stakeholders’ sentiments toward specific ESG issues.

4. Results and Analysis

This section presents the results of the machine learning-based text analytics performed on the resolutions data set. To surface out overlapping crucial phrases, we validated the results by mapping and comparing the critical topics discovered in the word cloud analysis against the co-occurrence analysis. Subsequently, the cluster analysis was conducted on the main themes and associated topics to identify core concepts and essential factors within each cluster. Lastly, prediction of the vote decision was attempted.
The word cloud map in Figure 5 displays the words most frequently found in the corpus of resolutions. This was generated with wordclouds.com. The larger the size of a word, the more frequently it occurs in the corpus. This affirms what we know via exploratory manual reading and analysis of the key implicit issues in the resolutions. The word cloud shows that the top issues of shareholders as found in the resolutions are lobbying, disclosure, climate, risk, emissions, sustainability, and disclosure (indicated in the center of the map). Therefore, a key takeaway is that shareholders in companies understand the significance and impact of sustainability and environmental issues like gas emission. In fact, they want top management to pay attention to these matters.
Figure 6 presents the top 35 words in the shareholder resolutions’ reports. The WordCount function gives a frequency count of how often a word occurs. The most frequent words include report, shareholder, lobbying, climate, and risk. This confirms the high-level sustainability issues that are displayed in the word cloud in Figure 5. However, keywords by themselves do not contribute much to a deeper understanding of shareholders’ interests and motivations regarding sustainability.
Therefore, we next examined the co-occurrence of words. In linguistics, co-occurrence represents the likelihood of occurrence of two terms in a certain order alongside each other within a large corpus of data. In this sense, it is used as an indicator of semantic closeness of terms [114]. This model gives insight as to which issues are related.
Figure 7 displays the result of the top 30 bigram, showing the frequency of two words coming together in the data set. Interesting patterns emerge from this analysis. For instance, “climate” and “change” are found together 2300 times. This indicates that shareholder resolutions are centered around climate change. Similarly, the combination of the top five bigrams shows that stakeholders are attempting to make positive changes on issues like climate change and greenhouse gas through both direct and indirect lobbying.
Figure 8 presents the top 30 results of trigram, showing the frequency of three words used together in the data set. Resolutions appear to focus on how and in what direction shareholders are influencing the sustainability landscape in the context of companies. Shareholders are communicating their sustainability activism via direct and indirect lobbying, with a goal of mitigating problems, including greenhouse gas emissions, within reasonable cost.
Next, we generated row similarities using the term frequency-inverse document frequency (TF-IDF) technique. This algorithm searches through a corpus of documents and determines which words are favorable to use in a query. For each word in a document, a value is calculated as an inverse proportion of the frequency of occurrence of the word in the document to the percentage of documents the word occurs [115]. A high TF-IDF indicates a strong relationship with the document in which it occurs. This implies that if the word were used in a query, then the document would be relevant and of interest to the user [115]. This technique ensures efficient query retrieval with relevant words. The row similarity analysis indicates the parallels between resolutions. Cosine-similarity, a branch of row similarity, calculates a numeric quantity to denote the similarity between groups of text [115]. Different cosine-similarity values represent angles and varied orientation between the two documents. Here, we define a function that takes a resolution title as input. It then outputs a list of the 10 most similar resolutions by implementing the concept of cosine-similarity on the resolutions. Figure 9 shows the approach to identifying the top resolutions that are most relevant. For example, keywords and labels of similar resolutions are generated when a resolution pertaining to renewable energy is input to the model. This offers both shareholders and top management the insight into a resolution’s voting outcome based on other resolutions.
We next explored the potential to identify and organize the key climate change and sustainability issues shareholders are most concerned with by applying the techniques of clustering and classification in an exploratory way. Clustering helps identify the keywords that are central to the resolution; classification can help classify new resolutions into these keywords. The K-means algorithm was applied to generate the clusters. Figure 10 shows the number of keywords in each cluster.
Cluster 1, as shown in Figure 11A, has 449 sustainability resolutions with terms like carbon, climate change, emission, energy, methane, renewable, etc. This indicates that the resolutions in this cluster are associated with emission and energy. Therefore, one could tentatively label this cluster as “emission and energy.” This cluster is tentatively termed “goal.” Activism has primarily been driven by an approach where investors choose one topic, such as climate change or diversity [58].
The second cluster (cluster 2), shown in Figure 11B, has 292 resolutions. This cluster identifies key concerns of shareholders like diversity (gender, race, etc.), board, committee, compensation, executive pay, voting, etc. Therefore, this is grouped under the general term of “board.”
The issues in cluster 3 (Figure 11C), with 193 resolutions, are primarily communication, legislation, lobbying, oversight, etc. These tend to reflect “regulation”. Cluster 4 (Figure 11D), with 106 resolutions, emphasizes the communication of sustainability ideas via lobbying with the government, campaigning, disclosure, expenditure, transparency, etc. This results in the possible cluster label of “politics”.
Cluster 5, as displayed in Figure 11E, has 42 resolutions. It refers to issues like by-law, ESG, governance, metric, managing, proxy, etc. Overall, this cluster appears to describe “governance”. The governance issues in our sample comprise primarily political lobbying, corruption, and board oversight regarding environmental and social issues [58]. Therefore, this cluster is related to cluster 2, namely, “Board”.
Next, cluster 6 (Figure 11F) mentions a variety of products and services at the core of climate change and sustainability. These include deforestation, drugs, food, packaging, plastic, recycling, etc. The products, and their related activities, are essential to human life. However, the process of making them can be improved through sustainability practices like proper recycling and bio-degradable packaging. Therefore, one may assign it the label “Product Management”.
The final cluster, cluster 7 (Figure 11G), has 218 resolutions. It mentions a variety of core concepts of climate change, ESG, and sustainability practices within companies. Its key words include guidelines, environmental, reporting, responsibility, sustainability, etc. Overall, this cluster can be labeled “Accountability”.
Cluster analysis focuses on data position and the visual structure of the data within a cluster. This allows us to pinpoint important nodes (in this case, a keyword) and key findings. Next, we turn our attention to classification. This allows us to engage in predictive analytics. Given the resolution, stakeholders may ask if one can accurately classify the voting status of the resolution.
In this exploratory research, the K-nearest neighbor algorithm is used to build the classification model. Figure 12 presents the number of neighbors (x-axis) and model accuracy (y-axis). Testing accuracy peaks at 61% with five neighbors, indicating the model can predict the voting status of a given resolution with 61% accuracy. With further iteration and refinement of the models, this can be improved further.

5. Discussion

Increasingly, activist shareholders are attempting to pass resolutions on sustainability to gain the attention of management. By proposing and putting them to vote, shareholders are bringing attention to climate change, ESG, and sustainability causes. Therefore, shareholder resolutions are a key source of information on the shareholders’ sentiments regarding sustainability. By extrapolation, management and other stakeholders can gain insight into the shareholders’ sentiments.
This exploratory study used several machine learning-based methods in text analytics to analyze and elicit key information from 1737 shareholder resolutions over a 11-year period. The research capitalizes on advances in information processing technology in extracting insight from large corpuses of text, a process that was only entrusted to subjective evaluation in the past. In the process, and in line with our research question, we identified seven key clusters (topics) related to sustainability, representing shareholders’ main concerns. Our analysis shows that shareholders are concerned about topics related to emission and energy, boards, regulations, politics, governance, product management, and accountability. The identification of these key topics is consistent with results from other studies related to ESG and sustainability reports, typically from a managerial perspective. It therefore appears that both shareholders and management are interested in similar themes.
Analyzing shareholder resolutions allows a window into the landscape of issues that shareholders are passionate about. Analyzing shareholder resolutions also helps companies foresee the evolution of these issues in the public eye. In addition, it helps companies tell whether they are likely to become targets of public scrutiny related to the issues [79].

6. Conclusions

This exploratory study attempts to gain insight into the climate change and sustainability concerns of shareholders by examining shareholder resolutions. However, there are many other issues warranting investigation. For example, to what extent can text documents such as the resolutions be used to forecast sustainability trends and managerial action. Trend analysis can be performed to examine attitudes and impact over time. The impact of resolutions on the public in general and vice versa can be examined from a social media perspective (public sentiment analysis). Reproducibility and validation are key challenges in the application of machine learning and text analytics. Additionally, the intuitive labeling of clusters and interpretation of the machine learning likely make the interpretation subjective. Yet, we are confident that the results of the analysis, for example, the labeling of the clusters, generally represent what is described in the resolutions, as confirmed by prior literature. Thus, it provides a powerful methodology to study and understand sustainability. Additional limitations include the reliability of the documents and validity of data preparation techniques. Generalizability of the topics (clusters) may be uncertain by examining a small sample of resolutions. They may not characterize overall corporate efforts and initiatives regarding sustainability. Further, machine learning models are only capable of extracting a limited amount of insight. Additionally, this study used the summary section as the sole source of unstructured data. Future research should investigate the value of certain sections on a company’s 10K reports and user-generated content (e.g., tweets, blog postings, etc.).
Despite the limitations, our study contributes to policy and research in two ways. First, practitioners and researchers can utilize the results to prioritize sustainability initiatives and examine some of the clusters via different lenses like those of non-governmental organizations (NGOs) or other activists. Second, the study demonstrates the efficacy of machine learning and text analytics in understanding and gaining insight into shareholder resolutions to make informed decisions by conducting descriptive and predictive analytics. Third, we demonstrate the evolutionary trend of how shareholder resolutions can be an effective venue for fostering changes in corporate practices by corporate management.
Shareholders, being a large and prominent group of stakeholders, can effectively project a positive image of the company with transparency. In fact, regardless of whether a resolution is passed or implemented, the resolution being publicly available creates a positive impact on the dialogue regarding corporate sustainability. Future research can continue to explore and apply advanced techniques like deep learning to delve into shareholder resolutions and disclosure reports for additional content analysis. For example, prescriptive analytics can be investigated to not only predict outcomes, but also suggest likely impacts and potential strategies.
Other avenues for future research include conducting cross industry comparisons, studying global differences, and examining resolutions that affect corporate sustainability and its correlation to company performance. Sophisticated application of advanced techniques, such as artificial intelligence and deep learning, will accelerate the maturing process of gaining insight from shareholder resolutions on climate change and sustainability. Even though most proposals of shareholders in the domain of ESG obtain below 20% votes in support [58], these proposals nevertheless serve to be significant catalysts of action within companies [116] and persuade management to take effective action. Additionally, applying discovery analytics on the resolutions may shed light on innovation and new product ideas.

Author Contributions

All the authors contributed equally to all parts of the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Clarkson, P.M.; Li, Y.; Richardson, G.D.; Vasvari, F.P. Revising the Relation between Environmental Performance and Environmental Disclosure: An Empirical Analysis. Acc. Organ Soc. 2008, 33, 303–327. [Google Scholar] [CrossRef]
  2. Galbreath, J. ESG in Focus: The Australian Evidence. J. Bus. Ethic 2012, 118, 529–541. [Google Scholar] [CrossRef]
  3. Goyal, P.; Rahman, Z.; Kazmi, A.A. Identification and prioritization of corporate sustainability practices using analytical hierarchy process. J. Model. Manag. 2015, 10, 23–49. [Google Scholar] [CrossRef]
  4. Kolk, A. Sustainability, accountability and corporate governance: Exploring multinationals’ reporting practices. Bus. Strat. Environ. 2007, 17, 1–15. [Google Scholar] [CrossRef]
  5. Fauzi, H. Corporate social and environmental performance: A comparative study of Indonesian companies and multinational companies (MNCs) operating in Indonesia. J. Knowl. Glob. 2008, 1, 81–105. [Google Scholar]
  6. Goyal, P.; Rahman, Z.; Kazmi, A.A. Corporate sustainability performance and firm performance research literature review and future research agenda. Manag. Decis. 2013, 51, 361–379. [Google Scholar] [CrossRef]
  7. Klettner, A.; Clarke, T.; Boersma, M. The Governance of Corporate Sustainability: Empirical Insights into the Development, Leadership and Implementation of Responsible Business Strategy. J. Bus. Ethic 2013, 122, 145–165. [Google Scholar] [CrossRef]
  8. Velte, P. Does ESG performance have an impact on financial performance? Evidence from Germany. J. Glob. Responsib. 2017, 8, 169–178. [Google Scholar] [CrossRef]
  9. Aziz, N.A.A.; Yau, F.S.; San, O.T.; Attan, H. A Review on Green Integration into Management Control System. Procedia Soc. Behav. Sci. 2015, 172, 435–441. [Google Scholar] [CrossRef]
  10. Henri, J.-F.; Journeault, M. Eco-control: The influence of management control systems on environmental and economic performance. Account. Organ. Soc. 2010, 35, 63–80. [Google Scholar] [CrossRef]
  11. Adams, C.A.; Frost, G.R. Accessibility and functionality of the corporate website: Implications for sustainability reporting. Bus. Strat. Environ. 2006, 15, 275–287. [Google Scholar] [CrossRef]
  12. Szekely, F.; Knirsch, M. Responsible leadership and corporate social responsibility metrics for sustainable performance. Eur. Manag. J. 2005, 23, 628–647. [Google Scholar] [CrossRef]
  13. Van Marrewijk, M. Concepts and Definitions of CSR and Corporate Sustainability: Between Agency and Communion. J. Bus. Ethics 2003, 44, 95–105. [Google Scholar] [CrossRef]
  14. Dyllick, T.; Hockerts, K. Beyond the business case for corporate sustainability. Bus. Strat. Environ. 2002, 11, 130–141. [Google Scholar] [CrossRef]
  15. Lo, S.-F.; Sheu, H.-J. Is Corporate Sustainability a Value-Increasing Strategy for Business? Corp. Govern. Int. Rev. 2007, 15, 345–358. [Google Scholar] [CrossRef]
  16. Raghupathi, V.; Raghupathi, W. Corporate Sustainability Reporting and Disclosure on the Web. Inf. Resour. Manag. J. 2019, 32, 1–27. [Google Scholar] [CrossRef]
  17. Braam, G.; Peeters, R. Corporate Sustainability Performance and Assurance on Sustainability Reports: Diffusion of Accounting Practices in the Realm of Sustainable Development. Corp. Soc. Responsib. Environ. Manag. 2017, 25, 164–181. [Google Scholar] [CrossRef]
  18. Krajnc, D.; Glavic, P. A model for integrated assessment of sustainable development. Resour. Conserv. Recycl. 2005, 43, 189–208. [Google Scholar] [CrossRef]
  19. Singh, R.K.; Murty, H.; Gupta, S.; Dikshit, A.K. Development of composite sustainability performance index for steel industry. Ecol. Indic. 2007, 7, 565–588. [Google Scholar] [CrossRef]
  20. Sandberg, M.; Holmlund, M. Impression management tactics in sustainability reporting. Soc. Responsib. J. 2015, 11, 677–689. [Google Scholar] [CrossRef]
  21. Lokuwaduge, C.S.D.S.; Heenetigala, K. Integrating Environmental, Social and Governance (ESG) Disclosure for a Sustainable Development: An Australian Study. Bus. Strat. Environ. 2016, 26, 438–450. [Google Scholar] [CrossRef]
  22. Elkington, J. Cannibals with Forks: The Triple Bottom Line of the 21st Century; New Society Publishers: Stoney Creek, CT, USA, 1998. [Google Scholar]
  23. Belz, F.M.; Binder, J.K. Sustainable Entrepreneurship: A Convergent Process Model. Bus. Strat. Environ. 2015, 26, 1–17. [Google Scholar] [CrossRef]
  24. Boli, J.; Hartsuiker, D. World culture and transnational corporations: Sketch of a project. In Proceedings of the International Conference on Effects of and Responses to globalization, Istanbul, Turkey, 31 May–1 June 2002. [Google Scholar]
  25. Dahlsrud, A. How corporate social responsibility is defined: An analysis of 37 definitions. Corp. Soc. Responsib. Environ. Manag. 2008, 15, 1–13. [Google Scholar] [CrossRef]
  26. Searcy, C.; Elkhawas, D. Corporate sustainability ratings: An investigation into how corporations use the Dow Jones Sustainability Index. J. Clean. Prod. 2012, 35, 79–92. [Google Scholar] [CrossRef]
  27. Freundlieb, M.; Teuteberg, F. Corporate social responsibility reporting—A transnational analysis of online corporate social responsibility reports by market-listed companies: Contents and their evolution. Int. J. Innov. Sustain. Dev. 2013, 7, 1. [Google Scholar] [CrossRef]
  28. Kolk, A. A decade of sustainability reporting: Developments and significance. Int. J. Environ. Sustain. Dev. 2004, 3, 51. [Google Scholar] [CrossRef]
  29. Liew, W.T.; Adhitya, A.; Srinivasan, R. Sustainability trends in the process industries: A text mining-based analysis. Comput. Ind. 2014, 65, 393–400. [Google Scholar] [CrossRef]
  30. Leipziger, D. The Global Reporting Initiative. The Corporate Responsibility Code Book, 3rd ed.; Greenleaf Publishing: Sheffield, UK, 2015; pp. 516–538. [Google Scholar]
  31. Székely, N.; Brocke, J.V. What can we learn from corporate sustainability reporting? Deriving propositions for research and practice from over 9,500 corporate sustainability reports published between 1999 and 2015 using topic modelling technique. PLoS ONE 2017, 12, e0174807. [Google Scholar] [CrossRef] [PubMed]
  32. Kolk, A. Trends in sustainability reporting by the Fortune Global 250. Bus. Strat. Environ. 2003, 12, 279–291. [Google Scholar] [CrossRef]
  33. Bjørn, A.; Bey, N.; Georg, S.; Røpke, I.; Hauschild, M. Is Earth recognized as a finite system in corporate responsibility reporting? J. Clean. Prod. 2017, 163, 106–117. [Google Scholar] [CrossRef]
  34. Modapothala, J.R.; Issac, B. Evaluation of Corporate Environmental Reports Using Data Mining Approach. In Proceedings of the 2009 International Conference on Computer Engineering and Technology, Singapore, 22–24 January 2009; Volume 2, pp. 543–547. [Google Scholar]
  35. Modapothala, J.R.; Issac, B.; Jayamani, E. Appraising the Corporate Sustainability Reports—Text Mining and Multi-Discriminatory Analysis. Innov. Comp. Sci. Softw. Eng. 2009. [Google Scholar] [CrossRef]
  36. Shahi, A.M.; Issac, B.; Modapothala, J.R. Analysis of supervised text classification algorithms on corporate sustainability reports. In Proceedings of the 2011 International Conference on Computer Science and Network Technology, Harbin, China, 24–26 December 2011. [Google Scholar]
  37. Blei, D.M. Probabilistic topic models. Commun. ACM 2012, 55, 77–84. [Google Scholar] [CrossRef]
  38. O’Rourke, A. A new politics of engagement: Shareholder activism for corporate social responsibility. Bus. Strat. Environ. 2003, 12, 227–239. [Google Scholar] [CrossRef]
  39. Smith, T. Two Sides of the Coin: Shareholders Engaging Companies on Sustainability Issues/Companies Promoting CSR Leadership as Good Business. J. Invest. 2011, 20, 103–107. [Google Scholar] [CrossRef]
  40. Domenec, F. The “greening” of the annual letters published by Exxon, Chevron and BP between 2003 and 2009. J. Commun. Manag. 2012, 16, 296–311. [Google Scholar] [CrossRef]
  41. Galbreath, J. How does corporate social responsibility benefit firms? Evidence from Australia. Eur. Bus. Rev. 2010, 22, 411–431. [Google Scholar] [CrossRef]
  42. Cormier, D.; Magnan, M. The Economic Relevance of Environmental Disclosure and its Impact on Corporate Legitimacy: An Empirical Investigation. Bus. Strat. Environ. 2013, 24, 431–450. [Google Scholar] [CrossRef]
  43. Yang, A.; Uysal, N.; Taylor, M. Unleashing the Power of Networks: Shareholder Activism, Sustainable Development and Corporate Environmental Policy. Bus. Strat. Environ. 2017, 27, 712–727. [Google Scholar] [CrossRef]
  44. Baumgartner, R.J.; Ebner, D. Corporate sustainability strategies: Sustainability profiles and maturity levels. Sustain. Dev. 2010, 18, 76–89. [Google Scholar] [CrossRef]
  45. Benn, S.; Edwards, M.; Williams, T. Organizational Change for Corporate Sustainability; Routledge: London, UK, 2014. [Google Scholar]
  46. Atan, R.; Alam, M.; Said, J.; Zamri, M. The impacts of environmental, social, and governance factors on firm performance. Manag. Environ. Qual. Int. J. 2018, 29, 182–194. [Google Scholar] [CrossRef]
  47. Friede, G.; Busch, T.; Bassen, A. ESG and financial performance: Aggregated evidence from more than 2000 empirical studies. J. Sustain. Finance Invest. 2015, 5, 210–233. [Google Scholar] [CrossRef]
  48. Lee, K.-H.; Cin, B.C.; Lee, E.Y. Environmental Responsibility and Firm Performance: The Application of an Environmental, Social and Governance Model. Bus. Strat. Environ. 2014, 25, 40–53. [Google Scholar] [CrossRef]
  49. Tarmuji, I.; Maelah, R.; Tarmuji, N.H. The Impact of Environmental, Social and Governance Practices (ESG) on Economic Performance: Evidence from ESG Score. Int. J. Trade Econ. Finance 2016, 7, 67–74. [Google Scholar] [CrossRef]
  50. Chouinard, Y.; Ellison, J.; Ridgeway, R. The sustainable economy. Harvard Bus. Rev. 2011, 89, 52–62. [Google Scholar]
  51. Harjoto, M.A.; Jo, H.; Kim, Y. Is Institutional Ownership Related to Corporate Social Responsibility? The Nonlinear Relation and Its Implication for Stock Return Volatility. J. Bus. Ethics 2015, 146, 77–109. [Google Scholar] [CrossRef]
  52. Rist, L.; Moen, J. Sustainability in forest management and a new role for resilience thinking. For. Ecol. Manag. 2013, 310, 416–427. [Google Scholar] [CrossRef]
  53. Sikken, B.J. Accelerating the Transition Towards Sustainable Investing—Strategic Options for Investors, Corporations, and Other Key Stakeholders. World Econ. Forum 2011. [Google Scholar] [CrossRef]
  54. Ioannou, I.; Serafeim, G. The Consequences of Mandatory Corporate Sustainability Reporting. SSRN Electron. J. 2011, 11–100. [Google Scholar] [CrossRef]
  55. Montiel, I.; Ceballos, J.D. Defining and Measuring Corporate Sustainability. Organ. Environ. 2014, 27, 113–139. [Google Scholar] [CrossRef]
  56. Taylor, J.; Vithayathil, J.; Yim, D. Are corporate social responsibility (CSR) initiatives such as sustainable development and environmental policies value enhancing or window dressing? Corp. Soc. Responsib. Environ. Manag. 2018, 25, 971–980. [Google Scholar] [CrossRef]
  57. Waddock, S. The development of corporate responsibility/corporate citizenship. Organ. Manag. J. 2008, 5, 29–39. [Google Scholar] [CrossRef]
  58. Grewal, J.; Serafeim, G.; Yoon, A.S. Shareholder Activism on Sustainability Issues. SSRN Electron. J. 2016. [Google Scholar] [CrossRef]
  59. Herremans, I.; Nazari, J.A.; Mahmoudian, F. Stakeholder Relationships, Engagement, and Sustainability Reporting. J. Bus. Ethics 2015, 138, 417–435. [Google Scholar] [CrossRef]
  60. Lee, M.-D.P.; Lounsbury, M. Domesticating Radical Rant and Rage: An Exploration of the Consequences of Environmental Shareholder Resolutions on Corporate Environmental Performance. Bus. Soc. 2011, 50, 155–188. [Google Scholar] [CrossRef]
  61. Stoner, J. What We Want this Journal to Be: Our First Editorial Essay in which We Hope to Start a Continuing and Evolving Conversation about Why We are Now Creating this New Journal and What We Want It to Become. J. Manag. Glob. Sustain. 2013, 1, 1–6. [Google Scholar] [CrossRef]
  62. World Commission on Environment and Development Report of the World Commission on Environment and Development: Our Common Future; Oxford University Press: Oxford, UK, 1987.
  63. Linnenluecke, M.K.; Griffiths, A. Corporate sustainability and organizational culture. J. World Bus. 2010, 45, 357–366. [Google Scholar] [CrossRef]
  64. Hahn, R.; Lülfs, R. Legitimizing Negative Aspects in GRI-Oriented Sustainability Reporting: A Qualitative Analysis of Corporate Disclosure Strategies. J. Bus. Ethics 2013, 123, 401–420. [Google Scholar] [CrossRef]
  65. Montiel, I. Corporate social responsibility and corporate sustainability separate pasts, common futures. Org. Environ. 2008, 21, 245–269. [Google Scholar] [CrossRef]
  66. Azapagic, A.; Perdan, S. Managing corporate social responsibility: Translating theory into business practice. Int. J. Corp. Sustain. 2003, 10, 97–108. [Google Scholar]
  67. Belu, C. Ranking corporations based on sustainable and socially responsible practices. A data envelopment analysis (DEA) approach. Sustain. Dev. 2009, 17, 257–268. [Google Scholar] [CrossRef]
  68. Hahn, T.; Scheermesser, M. Approaches to corporate sustainability among German companies. Corp. Soc. Responsib. Environ. Manag. 2006, 13, 150–165. [Google Scholar] [CrossRef]
  69. Krug, B.A.; Burnett, S.E.; Dennis, J.H.; Lopez, R.G. Growers look at operating a sustainable greenhouse. GMPro 2008, 28, 43–45. [Google Scholar]
  70. Smith, D. Sustainable viticulture and winery practices in California: What is it, and do customers care? Int. J. Wine Res. 2009, 2, 189. [Google Scholar] [CrossRef]
  71. Daily, B.F.; Huang, S.-C. Achieving sustainability through attention to human resource factors in environmental management. Int. J. Oper. Prod. Manag. 2001, 21, 1539–1552. [Google Scholar] [CrossRef]
  72. Wilkinson, A.; Hill, M.; Gollan, P. The sustainability debate. Int. J. Oper. Prod. Manag. 2001, 21, 1492–1502. [Google Scholar] [CrossRef]
  73. Ameer, R.; Othman, R. Sustainability Practices and Corporate Financial Performance: A Study Based on the Top Global Corporations. J. Bus. Ethics 2011, 108, 61–79. [Google Scholar] [CrossRef]
  74. Hall, T.J.; Dennis, J.; Lopez, R.G.; Marshall, M.I. Factors Affecting Growers’ Willingness to Adopt Sustainable Floriculture Practices. HortScience 2009, 44, 1346–1351. [Google Scholar] [CrossRef]
  75. Seidel, S.; Recker, J.; Pimmer, C.; vom Brocke, J. Enablers and barriers to the organizational adoption of sustainable business practices. In Proceedings of the 16th Americas Conference on Information Systems: Sustainable IT Collaboration around the Globe, Lima, Peru, 12–15 August 2010; 2010. [Google Scholar]
  76. Bizjak, J.M.; Marquette, C.J. Are Shareholder Proposals All Bark and No Bite? Evidence from Shareholder Resolutions to Rescind Poison Pills. J. Financ. Quant. Anal. 1998, 33, 499. [Google Scholar] [CrossRef]
  77. Logsdon, J.M.; Van Buren, H.J. Beyond the Proxy Vote: Dialogues Between Shareholder Activists and Corporations. J. Bus. Ethics 2008, 87, 353–365. [Google Scholar] [CrossRef]
  78. Manne, H.G. Shareholder Social Proposals Viewed by an Opponent. Stanf. Law Rev. 1972, 24, 481. [Google Scholar] [CrossRef]
  79. Graves, S.B.; Waddock, S.; Rehbein, K. Fad and Fashion in Shareholder Activism: The Landscape of Shareholder Resolutions, 1988–1998. Bus. Soc. Rev. 2001, 106, 293–314. [Google Scholar] [CrossRef]
  80. Vogel, D. Trends in Shareholder Activism: 1970–1982. Calif. Manag. Rev. 1983, 25, 68–87. [Google Scholar] [CrossRef]
  81. Brancato, C.K. Institutional Investors and Corporate Governance: Beat Practices for Increasing Corporate Value; McGraw-Hill: Chicago, IL, USA, 1997. [Google Scholar]
  82. O’Sullivan, M. Contests for Corporate Control: Corporate Governance and Economic Performance in the U.S. and Germany; Oxford University Press: Oxford, UK, 2000. [Google Scholar]
  83. Monks, R.A.G.; Minnow, N. Power and Accountability; Harper: New York, NY, USA, 1991. [Google Scholar]
  84. McConnell, G.; Vogel, D. Lobbying the Corporation: Citizen Challenges to Business Authority. Politi-Sci. Q. 1979, 94, 528. [Google Scholar] [CrossRef]
  85. Social Investment Forum (SIF). Report on Socially Responsible Investing Trends in the United States; SIF Industry Research Program: Washington, DC, USA, 2001; Available online: https://www.griequity.com/resources/InvestmentIndustry/Trends/siftrensreport2001.pdf (accessed on 20 May 2020).
  86. Ryan, L.V. Shareholders and the Atom of Property: Fission or Fusion? Bus. Soc. 2000, 39, 49–75. [Google Scholar] [CrossRef]
  87. Gifford, E.J.M. Effective Shareholder Engagement: The Factors that Contribute to Shareholder Salience. J. Bus. Ethics 2010, 92, 79–97. [Google Scholar] [CrossRef]
  88. Gordon, L.A.; Pound, J. Information, ownership structure, and shareholder voting: Evidence from shareholder-sponsored corporate governance proposals. J. Financ. 1993, 48, 697–718. [Google Scholar] [CrossRef]
  89. Karpoff, J.M.; Malatesta, P.H.; Walkling, R.A. Corporate governance and shareholder initiatives: Empirical evidence. J. Financ. Econ. 1996, 42, 365–395. [Google Scholar] [CrossRef]
  90. Ryan, P.J. Rule 14a-8, institutional shareholder proposals, and corporate democracy. Ga. L. Rev. 1988, 23, 97. [Google Scholar]
  91. Taub, S. Shareholder Social Proposals Gain. 2004. Available online: https://www.cfo.com/risk-compliance/2004/09/shareholder-social-proposals-gain/ (accessed on 20 May 2020).
  92. Allen, D.; Alves, L.; Caruncho, G.; Chin, D.; Cohn, D.; Damle, B.; Dracos, B.; Friend, S.; Ha, J.; Kadar, A.; et al. The New Science of Personalized Medicine. Available online: http://www.PWC.com (accessed on 20 August 2012).
  93. Meznar, M.B.; Nigh, D.; Kwok, C.C.Y. Effect of announcements of withdrawal from South Africa on stockholder wealth. Acad. Manag. J. 1994, 37, 1633–1648. [Google Scholar]
  94. Reid, E.M.; Toffel, M.W. Responding to public and private politics: Corporate disclosure of climate change strategies. Strat. Manag. J. 2009, 30, 1157–1178. [Google Scholar] [CrossRef]
  95. Post, J.E.; Preston, L.E.; Sachs, S. Managing the Extended Enterprise: The New Stakeholder View. Calif. Manag. Rev. 2002, 45, 6–28. [Google Scholar] [CrossRef]
  96. Wood, D.J.; Logsdon, J.M.; Lewellyn, P.G. Global Business Citizenship: A Transformative Framework for Ethics and Sustainable Capitalism; Routledge: New York, NY, USA, 2015. [Google Scholar]
  97. Masud, A.K.; Rashid, H.U.; Khan, T.; Bae, S.M.; Kim, J.D. Organizational Strategy and Corporate Social Responsibility: The Mediating Effect of Triple Bottom Line. Int. J. Environ. Res. Public Health 2019, 16, 4559. [Google Scholar] [CrossRef] [PubMed]
  98. Sultana, S.; Zulkifli, N.; Zainal, D. Environmental, Social and Governance (ESG) and Investment Decision in Bangladesh. Sustainability 2018, 10, 1831. [Google Scholar] [CrossRef]
  99. Hoque, N.; Rahman, A.R.A.; Molla, R.I.; Noman, A.H.M.; Bhuiyan, M.Z.H. Is corporate social responsibility pursuing pristine business goals for sustainable development? Corp. Soc. Responsib. Environ. Manag. 2018, 25, 1130–1142. [Google Scholar] [CrossRef]
  100. Bae, S.M.; Masud, A.K.; Kim, J.D. A Cross-Country Investigation of Corporate Governance and Corporate Sustainability Disclosure: A Signaling Theory Perspective. Sustainability 2018, 10, 2611. [Google Scholar] [CrossRef]
  101. Eesley, C.; Lenox, M.J. Firm responses to secondary stakeholder action. Strat. Manag. J. 2006, 27, 765–781. [Google Scholar] [CrossRef]
  102. Baharudin, B.; Lee, L.H.; Khan, K. A Review of Machine Learning Algorithms for Text-Documents Classification. J. Adv. Inf. Technol. 2010, 1, 4–20. [Google Scholar] [CrossRef]
  103. Landmann, J.; Zuell, C. Identifying Events Using Computer-Assisted Text Analysis. Soc. Sci. Comput. Rev. 2007, 26, 483–497. [Google Scholar] [CrossRef]
  104. Shelley, M.; Krippendorff, K. Content Analysis: An Introduction to its Methodology. J. Am. Stat. Assoc. 1984, 79, 240. [Google Scholar] [CrossRef]
  105. Raghupathi, V.; Ren, J.; Raghupathi, W. Studying Public Perception about Vaccination: A Sentiment Analysis of Tweets. Int. J. Environ. Res. Public Health 2020, 17, 3464. [Google Scholar] [CrossRef] [PubMed]
  106. Raghupathi, V.; Zhou, Y.; Raghupathi, W. Legal Decision Support: Exploring Big Data Analytics Approach to Modeling Pharma Patent Validity Cases. IEEE Access. 2018, 6, 41518–41528. [Google Scholar] [CrossRef]
  107. Raghupathi, V.; Zhou, Y.; Raghupathi, W. Exploring Big Data Analytic Approaches to Cancer Blog Text Analysis. Int. J. Healthc. Inf. Syst. Inf. 2019, 14, 1–20. [Google Scholar] [CrossRef]
  108. Ceres. 2020. Available online: https://www.ceres.org/resources/tools/climate-and-sustainability-shareholder-resolutions-database (accessed on 28 February 2020).
  109. Cook, J. Political action through environmental shareholder resolution filing: Applicability to Canadian Oil Sands? J. Sustain. Finance Invest. 2012, 2, 26–43. [Google Scholar]
  110. Monks, R.; Miller, A.; Cook, J. Shareholder activism on environmental issues: A study of proposals at large US corporations (2000–2003). Nat. Resour. Forum 2004, 28, 317–330. [Google Scholar] [CrossRef]
  111. Parris, T.M. Corporate Sustainability Reporting. Environ. Sci. Policy Sustain. Dev. 2006, 48, 3. [Google Scholar] [CrossRef]
  112. Smith, J.A. The CERES principles: A voluntary code for corporate environmental responsibility. Yale J. Int. Law 1993, 18, 307. [Google Scholar]
  113. Dietterich, T.G. Machine learning in ecosystem informatics and sustainability. In Proceedings of the Twenty-First International Joint Conference on Artificial Intelligence, Pasadena, CA, USA, 14–17 July 2009. [Google Scholar]
  114. Yi, Y.; Liu, L.; Li, C.H.; Song, W.; Liu, S. Machine Learning Algorithms with Co-occurrence Based Term Association for Text Mining. In Proceedings of the 2012 Fourth International Conference on Computational Intelligence and Communication Networks, Mathura, India, 3–5 November 2012; Institute of Electrical and Electronics Engineers (IEEE): Piscataway, NJ, USA, 2012; pp. 958–962. [Google Scholar]
  115. Ramos, J. Using TF-IDF to Determine Word Relevance in Document Queries; Technical Report; Department of Computer Science, Rutgers University: Piscataway, NJ, USA, 2003. [Google Scholar]
  116. Turner, M. Here Is the Letter the World’s Largest Investor, BlackRock CEO Larry Fink, Just Sent to CEOs Everywhere. 2016. Available online: https://www.businessinsider.com.au/blackrock-larry-fink-letter-to-ceos-2017-1 (accessed on 20 May 2020).
Figure 1. Top sectors filing shareholder resolutions from 2009–2019.
Figure 1. Top sectors filing shareholder resolutions from 2009–2019.
Sustainability 12 04753 g001
Figure 2. Shareholder resolutions filing trend of top five sectors by year.
Figure 2. Shareholder resolutions filing trend of top five sectors by year.
Sustainability 12 04753 g002
Figure 3. Stacked sector breakdown of top five sectors by year.
Figure 3. Stacked sector breakdown of top five sectors by year.
Sustainability 12 04753 g003
Figure 4. Distribution of voted resolutions by year.
Figure 4. Distribution of voted resolutions by year.
Sustainability 12 04753 g004
Figure 5. Word cloud of top shareholder resolution issues.
Figure 5. Word cloud of top shareholder resolution issues.
Sustainability 12 04753 g005
Figure 6. Word count of key issues in sustainability resolutions.
Figure 6. Word count of key issues in sustainability resolutions.
Sustainability 12 04753 g006
Figure 7. Top 30 words in co-occurrence-bigram.
Figure 7. Top 30 words in co-occurrence-bigram.
Sustainability 12 04753 g007
Figure 8. Top 30 words co-occurrence-trigram.
Figure 8. Top 30 words co-occurrence-trigram.
Sustainability 12 04753 g008
Figure 9. Row similarity example.
Figure 9. Row similarity example.
Sustainability 12 04753 g009
Figure 10. Number of keywords in each cluster
Figure 10. Number of keywords in each cluster
Sustainability 12 04753 g010
Figure 11. (A) Cluster 1—Emission and Energy. (B) Cluster 2—Board. (C) Cluster 3—Regulation. (D) Cluster 4—Politics. (E) Cluster 5—Governance. (F) Cluster 6—Product Management. (G) Cluster 7—Accountability.
Figure 11. (A) Cluster 1—Emission and Energy. (B) Cluster 2—Board. (C) Cluster 3—Regulation. (D) Cluster 4—Politics. (E) Cluster 5—Governance. (F) Cluster 6—Product Management. (G) Cluster 7—Accountability.
Sustainability 12 04753 g011aSustainability 12 04753 g011bSustainability 12 04753 g011cSustainability 12 04753 g011d
Figure 12. Training and testing accuracy of KNN classifier varying number of neighbors.
Figure 12. Training and testing accuracy of KNN classifier varying number of neighbors.
Sustainability 12 04753 g012
Table 1. Top 10 companies with most resolutions.
Table 1. Top 10 companies with most resolutions.
No.RankCompanyCountProportion
11Exxon Mobil Corporation533.05%
22Chevron Corporation452.59%
33Dominion Energy, Inc.291.67%
44Amazon.com Inc.201.15%
54ConocoPhillips201.15%
66Alphabet Inc.191.09%
76Devon Energy Corporation191.09%
86McDonald’s Corp.191.09%
99Duke Energy Corporation160.92%
109Kroger Co.160.92%
1111Emerson Electric Co.140.81%
1211FedEx Corporation140.81%
Table 2. Sector breakdown from 2009 to 2019
Table 2. Sector breakdown from 2009 to 2019
RankSectorCountRankSectorCount
1Oil, Gas & Consumable Fuels31716IT Services27
2Electric Utilities11918Biotechnology26
3Hotels, Restaurants & Leisure8518Chemicals26
4Multi-Utilities8318Equity Real Estate Investment Trusts (REITs)26
5Food Products8218Internet & Direct Marketing Retail26
6Specialty Retail5618Media26
7Food & Staples Retailing5523Diversified Telecommunication Services25
8Capital Markets4724Multiline Retail24
9Banks4625Communications Equipment23
10Internet Software & Services4425Pharmaceuticals23
11Machinery3327Tobacco22
12Air Freight & Logistics3128Textiles, Apparel & Luxury Goods21
12Health Care Providers & Services3129Aerospace & Defense20
14Insurance3029Health Care Equipment & Supplies20
15Metals & Mining2929Technology Hardware, Storage & Peripherals20
16Beverages2732Electrical Equipment19
RankSectorCountRankSectorCount
33Industrial Conglomerates1849Auto Components5
34Construction & Engineering1750Professional Services4
34Software1750Real Estate Management & Development4
36Semiconductors & Semiconductor Equipment1652Airlines3
37Automobiles1552Building Products3
37Household Products1552Construction Materials3
39Diversified Financial Services1452Containers & Packaging3
39Energy Equipment & Services1452Electronic Equipment, Instruments & Components3
39Independent Power and Renewable Electricity Producers1452Life Sciences Tools & Services3
42Road & Rail1352Personal products3
43Consumer Finance1252Trading Companies & Distributors3
44Gas Utilities1060Diversified Consumer Services2
45Commercial Services & Supplies960Wireless Telecommunication Services2
46Household Durables762Distributors1
46Water Utilities762Mortgage Real Estate Investment Trusts (REITs)1
48Leisure Products662Paper & Forest Products1
Table 3. Distribution of resolution voting status by year.
Table 3. Distribution of resolution voting status by year.
YearStatusCountYearStatusCount
2009Others62015Others9
Vote24 Vote207
Withdrawn17 Withdrawn46
2009 Total 472015 Total 262
2010Others42016Others10
Vote41 Vote245
Withdrawn27 Withdrawn59
2010 Total 722016 Total 314
2011Others62017Others15
Vote17 Vote227
Withdrawn23 Withdrawn87
2011 Total 462017 Total 329
2012Others52018Others34
Vote14 Vote207
Withdrawn12 With drawn91
2012 Total 312018 Total 332
2013Others22019Others88
Vote14 Vote5
Withdrawn44 Withdrawn43
2013 Total 602019 Total 136
2014Others12
Vote33
Withdrawn63
2014 Total 108Grand Total 1737
Table 4. Description of data variables.
Table 4. Description of data variables.
VariableDescription
OrgOrganization
IndustryThe industry in which the organization operates
TitleThe title of the sustainability disclosure
Filed ByThe entity whom filed the disclosure for the organization
StatusVote/Others/Withdraw
LabelVote=1/Others=2/Withdraw=O
YFilling year of the disclosure
SummaryTextual disclosure of such report
Back to TopTop