The Impact of Social Responsibility on the Performance of European Listed Companies

Abstract

This research aims to analyze the impact of social responsibility (SR) on the performance of 216 European companies from 2017 to 2021. The objective of this research is to determine how the operational, financial, and market performance of companies is influenced by social responsibility practices. The methodology adopted is quantitative in nature, using the estimation of models for panel data. To quantify corporate performance, this study uses the return on assets (ROA), the return on equity (ROE), and finally Tobin’s Q ratio. Additionally, environment, social, and governance (ESG) and United Nations Global Compact (GC) scores are used to quantify SR. Our findings indicate a complex relationship between SR and corporate performance. While SR positively impacts market performance, it negatively affects operational and financial performance. This disparity becomes more pronounced when comparing companies with the highest and lowest SR scores. Further analysis reveals that the environment, social, and governance dimensions of ESG negatively correlate with ROA and ROE, but positively correlate with Tobin’s Q. The GC’s anti-corruption and environment scores exhibit a negative relationship with Tobin’s Q, the human rights dimension negatively correlates with ROE and ROA, and the labor law dimension positively influences ROE. Notably, firm size amplifies these relationships, whereas firm age has a dampening effect. This research offers significant contributions to the literature by providing a comprehensive analysis of the impact of social responsibility on corporate performance based on ESG and GC scores.

1. Introduction

The concern that organizations have regarding societal welfare and environmental protection has been increasingly discussed in recent years, gaining significant emphasis in the 21st century [1]. In short, companies have felt immense pressure from society to adopt socially responsible practices, which implies that they start viewing social responsibility (SR) as a strategy for achieving better economic results. SR, combined with effective communication, allows companies to attain important benefits for value creation [2,3]. However, several authors argue that SR only incurs high costs and can sometimes harm the economic performance of companies [3]. Other authors suggest that it has no impact on shareholder value creation and negatively influences employees [4].

Given these divergent findings, the main objective of this research is to ascertain the impact of socially responsible practices on corporate performance [5]. To quantify a company’s social responsibility, this study uses environmental, social, and governance (ESG) scores. These scores assess the company’s performance in three dimensions: environmental, social, and governance, through various indicators. This information is then disclosed by rating agencies, with scores typically ranging from 0% to 100%. It is important to highlight that a higher score indicates better results in terms of social responsibility. Although the literature primarily quantifies SR through ESG scores, the Global Compact (GC) score—based on the four fundamental principles of the United Nations Global Compact (human rights, labor law, the environment, and anti-corruption)—can also serve as a metric for this purpose. Therefore, it is also considered in this research. Regarding corporate performance, this research aims to explore the relationship between social responsibility and corporate operational, financial, and market performance. To achieve this, the values that European companies obtained between 2017 and 2021 for the ratios return on equity (ROE) (economic performance), return on assets (ROA) (operational performance), and Tobin’s Q (market performance) were collected. Additionally, data on firms’ market value and earnings for the same period were collected. The goal is to explore the relationship between social responsibility indices (ESG and GC scores) and these variables, as this relationship has not been extensively explored in the existing literature. The control variables for this study are the company’s age (measured by the number of years since its foundation), the sector of activity (Consumer Goods, Communications, Consumer Discretionary, Energy, Finance, Real Estate, Industry, Materials, Health, Technology, and Utilities), and the size of the company (measured by the logarithm of total assets). To achieve the objective of this research, models for panel data are estimated. Initially, the analysis is conducted considering the existence of a direct relationship between SR metrics (ESG and GC scores) and the operational, financial, and market performance of companies. Subsequently, the differences in performance between companies with better scores and those with worse scores are examined. The same approach is applied to the environmental, social, and governance dimensions of the ESG score. For the human rights, labor law, the environment, and anti-corruption categories of the GC score, the process differs, as the direct relationship with company performance is ascertained. The main objective here is to understand how each dimension and category influences company performance. This will allow the identification of activities where companies should focus their efforts, specifically those where there is a positive relationship between social responsibility and corporate performance. Finally, this study also aims to understand how the control variables (size, age, and sector of activity) influence company performance when associated with the ESG or GC scores. Therefore, models establishing an interaction between control variables and ESG and GC scores were estimated. Subsequently, the impact of this interaction on the performance of the sample companies was assessed. In relation to the existing literature, this study contributes with an analysis of the impact of social responsibility on operational, financial, and market performance. In short, this research uses two indices of social responsibility—ESG and GC score—to quantify the performance of companies in this area. Additionally, this research is relevant because it not only uses the overall ESG score but also addresses the scores of its individual dimensions. This broader approach allows for a more comprehensive assessment of the impact of social responsibility on company performance, making it possible to identify behaviors that positively or negatively affect operational, financial, and market performance. Furthermore, the inclusion of control variables to identify their relationship with corporate performance is another strong point of this research, given that few authors have adopted this approach. It is also important to highlight that one of the major contributions of this research is its consideration of multiple dimensions of social responsibility, in contrast to other studies that prefer a holistic approach or focus on a single dimension. Specifically, this research includes the GC score, which has been little explored in the existing literature. This score allows us to examine the impact of practices associated with SR on company performance in areas not explicitly covered by the ESG score, such as human rights, labor law, and anti-corruption. Additionally, this study focuses on Western European markets, where the literature on this topic is still scarce.

In addition to the Introduction and Conclusion sections, this research is structured into three main sections. Section 2 consists of the literature review, presenting the evolution of corporate social responsibility and its dimensions. It also summarizes the main conclusions in the literature regarding the impact of social responsibility on corporate performance, addressing its influence on stakeholders, the positive, negative, and neutral effects of social responsibility, and the importance of its disclosure and standardization. Section 3 deals with the methodology adopted in this research. Specifically, this part presents the sample characterization, the data collection process, and the formulated hypotheses. This section also explains the dependent, independent, and control variables, as well as the regression models created from these variables. Section 4 presents the results of the empirical study. Initially, the descriptive statistics and the results of the regression models are presented. Subsequently, the results obtained are analyzed and discussed.

2. Literature Review

Throughout this section, the literature review on the impact of social responsibility on corporate financial performance is presented. Initially, the evolution of the concept of corporate social responsibility and its dimensions is outlined, followed by an exploration of this topic. It is important to highlight that the impact of social responsibility on corporate financial performance is addressed from different perspectives: firstly, the relationship between social responsibility and stakeholders; then, the positive, negative, and neutral effects of social responsibility; and finally, the role of social disclosure and standardization.

2.1. The Evolution of Social Responsibility

Interest in issues related to social responsibility began to increase around the 1950s and has since grown exponentially in both the academic literature (with a primary focus on the United States of America) and organizational practice [6]. Bowen [7], cited by Carroll [6], posited that corporate behavior affects the lives of citizens in various ways, necessitating an understanding of the responsibilities companies must assume. According to this author, corporate social responsibility (CSR) entails entrepreneurs making decisions that align with society’s objectives and values. This ideology persisted through the 1960s, as emphasized by Davis [8], cited by Agudelo et al. [9]. Davis argued that social, economic, and political changes during this decade exerted significant pressure on entrepreneurs to reassess their role in society and, consequently, to define policies on social responsibility. Essentially, entities that wielded power without considering environmental impact risked damaging their relationships with stakeholders in terms of trust and respect [10].

Friedman [11], cited by Rahman [12], a renowned economist and Nobel Prize winner in economics (1976), presents a contrasting view on social responsibility. According to this economist, a company’s sole responsibility towards society is to maximize its profits within legal limits and with minimal ethical constraints—avoiding fraud or deception. Friedman argues that allocating resources to social actions instead of production can negatively impact economic performance. On the other hand, Carroll [13] presented a more comprehensive perspective, arguing that a company’s responsibilities extend beyond economic and legal obligations to encompass a broader range of societal obligations. Specifically, the author notes that there was no consensus in the literature regarding the concept of social responsibility. However, authors primarily focused on three themes. The first theme relates to the economic, legal, or discretionary aspects of a company’s social responsibility. The second theme addresses social issues such as discrimination and environmental concerns. Finally, the last theme pertains to the company’s philosophy regarding the implementation of a social responsibility policy. Recognizing the importance of these three perspectives, Carroll [13] integrates them to provide a definition of social responsibility that encompasses all obligations companies have towards society. Therefore, according to this author, social responsibility (SR) is categorized into four dimensions: economic, legal, ethical, and discretionary. Economic responsibility, considered the most crucial by the author, involves providing goods and services desired by society and earning profits from their sale. Legal responsibility occurs when companies operate within the confines of the law. Ethical responsibility encompasses ethical standards that society expects organizations to adhere to, which may exceed legal requirements. Finally, discretionary responsibility involves voluntary actions by the company that are neither mandated by law nor expected by society.

As we can see, the contributions from the 1950s, 1960s, and early 1970s helped to understand the role that social responsibility plays in fostering harmony within companies, their environment, and with stakeholders [10]. However, in the 1980s and 1990s, there emerged significant theoretical dispersion in analyzing the benefits and advantages of implementing social responsibility actions by companies [10]. During these decades, concerns about developing new or refined definitions of SR were overshadowed by alternative concepts, theories, models, or themes [6]. One model developed during this period was by Tuzzolino and Armandi [14], cited by Carroll [6]. This model is based on Maslow’s hierarchy of needs, where the authors argue that organizations have physiological, security, affiliation, esteem, and self-actualization needs similar to those of humans. According to the authors, this framework enables the assessment of socially responsible organizational performance and provides an analytical framework to facilitate the operationalization of SR. During this period, Stakeholder Theory was introduced as a dimension in corporate social responsibility (CSR) literature [15], cited by [12]. According to Freeman and Dmytriyev [16], this theory posits that the essence of business lies primarily in building relationships and creating value for all stakeholders. Freeman [15], cited by Rahman [12], emphasizes that active stakeholder participation is crucial for successfully implementing CSR. Both Stakeholder Theory and CSR underscore the idea that companies should assume responsibilities towards communities and society [16]. However, they differ in scope, with Stakeholder Theory focusing on local communities where the company operates, while CSR has a broader perspective [16]. In the 1990s, Carroll [17] introduced a highly relevant model in the social responsibility literature known as the “Pyramid Model.” This model features four levels structured in a pyramid. The bottom two levels are fundamental and mandatory for all companies. At the base of the pyramid lies the core concept of any business, which is to make a profit (economic responsibility), followed by compliance with the law, as society’s codification of acceptable behavior (legal responsibility). Moving towards the top of the pyramid, ethical responsibilities are situated, associated with what is morally right, just, and minimizing harm to stakeholders. Finally, at the apex of the pyramid, discretionary (philanthropic) responsibilities require companies to contribute financial and human resources to the community and enhance quality of life.

The 21st century is considered the era of the emergence in industry of social responsibility, marked by significant international movements, initiatives, regulations, and reports on this topic. According to Perić and Turalija [1], one of these initiatives was the Global Compact (GC), established by the UN in 2010, which outlines 10 principles of SR focusing on four main themes: human rights, labor practices, environmental sustainability, and anti-corruption. The primary goal of this pact is to encourage companies to embrace social responsibility and uphold fundamental values through their operations [1]. Another notable initiative proving the relevance of CSR was the creation of the ISO 26000 standard [18]. This standard aims to guide companies in operating responsibly, thereby contributing to societal well-being [1]. As is evident, the concept of SR has been extensively discussed over recent decades and has become increasingly pertinent in corporate environments. Companies recognize that SR is pivotal in addressing economic, social, and environmental challenges, and implementing it enables them to reap numerous benefits, as detailed in this research later.

2.2. Influence of SR on Business Performance

Companies view SR as a means to develop a competitive advantage and foster strong relationships with their stakeholders [19]. This stems from the fact that socially responsible behaviors are linked not only to enhanced reputation and visibility, but also to greater stakeholder recognition, resulting in improved relationships and organizational outcomes [20]. Given this context, the growing significance of SR in business cannot be overstated, as several authors argue that the quality of relationships a company maintains with its key stakeholders is critical to its success and longevity [20].

Among stakeholders, certain groups are more profoundly affected by a company’s SR practices than others. Investors, for instance, are typically drawn to companies with robust management practices, as the management of their assets directly impacts the returns they receive [21]. Moreover, investors are increasingly committed to supporting companies that align their strategies with social and environmental goals. According to Okafor et al. [21], such an alignment enhances the investment case and leads to significant returns. In essence, this group of stakeholders favors well-managed companies due to the effective deployment of assets that generate shareholder wealth. However, among stakeholders, employees are particularly attentive to companies’ SR practices, and their perceptions on this matter influence their behaviors such as loyalty, motivation, productivity, and workplace satisfaction [22], as cited in [20]. Duarte and Neves [23] discovered that a company’s engagement in SR initiatives not only enhances employees’ satisfaction with their organization, but also encourages their desire to participate in these initiatives. Conversely, Sameer [24] argues that in workplaces where employees are unhappy or experience discrimination, job dissatisfaction tends to prevail. This situation can adversely affect employee performance and consequently diminish company productivity. To prevent this scenario, organizations can implement an SR policy that prioritizes the interests of their employees. Indeed, Tiep et al. [19] suggest that an SR policy enables companies to allocate resources more effectively, potentially enhancing employee satisfaction and motivation. Another stakeholder group closely monitoring companies’ SR practices is consumers. Recognizing the significance of SR in shaping consumer perceptions and evaluations of a company has become increasingly important, particularly among Generation Y consumers (Millennials), who show greater concern for SR practices compared to other generations [25]. However, opinions among scholars diverge when companies adopt SR policies primarily to achieve economic benefits, as discussed further below.

2.3. The Positive, Negative, and Neutral Effect of Social Responsibility

Corporate social responsibility (CSR) can serve as a means for companies to create both economic value (profitability) and social value (benefits to society), thereby achieving sustainable success. Okafor et al. [21] observed significant revenue growth in large companies like Apple, Amazon, Cisco Systems, and Microsoft aligned with their CSR expenditures. Moreover, firms that integrate CSR are more likely to survive in competitive markets and sustain efforts to address social issues [26,27]. Le et al. [28] discovered that companies can enhance customer retention, performance, and business continuity by adopting socially responsible practices that consider societal interests, environmental concerns, and stakeholder needs. Rossi et al. [29] similarly found a positive and significant relationship between CSR measured by the ESG score and business performance in their study of 225 European companies. Cheng et al. [30] went further, asserting that CSR can improve a company’s financial performance in the current year, with certain socially responsible activities potentially yielding significant effects in subsequent years. Conversely, Wuttichindanon [31] noted that some companies prioritize the influence of CSR on stakeholders over its potential impact on economic performance.

However, not all authors argue for a positive correlation between CSR and company performance. In a study conducted by Fernández et al. [32] on 380 companies from the coastal north of Portugal, it was found that the socially responsible actions of companies do not significantly impact economic results. Consequently, the authors concluded that the market in this region is indifferent to CSR practices, with companies actively pursuing CSR policies not ranking higher than those that do not engage in socially responsible practices. Oh and Park [33], referenced by Guzman et al. [34], argue for a negative correlation between CSR and company performance, citing that the costs associated with implementing CSR practices outweigh the benefits obtained. Hirigoyen and Poulain-Rehm [35] also support a negative relationship between CSR and firm performance. Specifically, analyzing a sample of 329 companies listed in three geographic regions (United States, Europe, and Asia–Pacific) for the years 2009 and 2010, these authors found that a more active CSR policy does not lead to better financial performance and may even worsen market performance. However, they also noted that when firms perform well financially, they tend to invest less in CSR to achieve more favorable results. Buallay [36] identified a potential cause for the negative relationship between CSR and corporate performance. According to this author, the inverse relationship occurs when company managers utilize CSR for personal gain, resulting in higher costs for the entities, costs that may ultimately be borne by stakeholders and consequently reduce market value, net worth, and asset efficiency. However, other authors [34,37,38] have contradicted the notion that social responsibility harms firm performance. For instance, Guzman et al. [34] and Ali et al. [39] argue for a positive indirect relationship between CSR and firm performance. They suggest that socially responsible activities enable firms to enhance the image of their products and services, leading to significant increases in sales and overall business performance [34]. Moreover, cultivating a positive image among stakeholders can reduce overhead costs [39]. Additionally, Giannarakis et al. [40] found in their study of companies listed on the Standard & Poor’s 500 between 2009 and 2013 that firms’ engagement in CSR initiatives fosters trust within their environments, which positively impacts financial performance. Given this perspective, managers aiming to enhance business performance should consider integrating CSR not only as a strategy but also as an integral part of daily business operations [34].

It should also be noted that in the literature, there are authors who have identified a null relationship between CSR and business performance. In a study conducted by Nollet et al. [41] on companies that were part of the S&P 500 index between 2007 and 2011, it was concluded that there is no significant relationship between CSR and financial performance. Companies were found to use CSR primarily as a strategy to create additional value for their products rather than to influence financial performance. Similarly, Crisóstomo et al. [42] reached the same conclusion in their study of 73 Brazilian companies, finding no effect of CSR on company performance when measured by ROA (return on assets) and ROE (return on equity) indicators. Regarding Tobin’s Q (the ratio of the market value of a company’s assets to its replacement cost), Bannier et al. [43] found in their study of European and North American companies that they are not affected by the ESG score in terms of market performance. Madorran and Garcia [44] also obtained a similar conclusion in their study of a sample of Spanish companies from the IBEX 35 stock market index, where the relationship between CSR and financial performance was not significant in the models used.

However, the relationship between social responsibility and corporate performance should not be analyzed in isolation, as there are several variables that can mediate this relationship. Minutolo et al. [45], in a study of 467 companies in the S&P 500 from 2009 to 2015, concluded that organizational size can be one of these variables. Specifically, these authors found that when firm performance is measured by Tobin’s Q, the influence of SR disclosure is greater for large firms compared to the effects on Tobin’s Q and return on assets for smaller firms. The industry in which the firm operates and its position in the supply chain also significantly influence the adoption of socially responsible behavior [46]. In cases where customers prioritize this issue, companies are more likely to define a social responsibility policy. Conversely, when customers are not concerned with socially responsible practices, companies are less likely to show interest. Age is another variable to consider when analyzing the adoption of socially responsible practices by companies. According to Sun [47], who studied 2610 companies over the period 1992–2016, older companies may perform better in terms of SR than younger companies because they have greater flexibility to invest their financial resources in additional activities such as SR.

The impact of social responsibility on firm performance may similarly depend on the type of socially responsible activity the firm adopts. Cho et al. [37] found, for example, that not all activities have a significant impact on firm performance. Specifically, these authors noted that socially responsible activities associated with the labor market, environment, community, and workplace do not exert statistically significant effects on the ROE, ROA, and Tobin’s Q of Greek companies listed on the Athens Stock Exchange. From another perspective, Pham et al. [48] found in their study of 56 US and Chinese companies that firms with strong environmental performance can achieve higher levels of business performance. This is primarily due to lower costs and increased revenue, stemming from a positive reputation for good practices. However, not all authors support this relationship. Kamatra and Kartikaningdyah [49], for instance, argue that while both social and environmental performance contribute to better economic performance, environmental performance has a significantly smaller effect than social performance. In a more pessimistic view regarding environmental activities, Makni et al. [50] found in their study of 179 publicly traded Canadian companies that the environmental dimension of SR negatively influences corporate performance, specifically return on assets, return on equity, and market returns. Similarly, Riyadh et al. [51] found in their study of 250 energy companies between 2016 and 2018 that socially responsible behaviors such as pollution reduction, employee benefit packages, donations, and community sponsorships decrease company profits and could lead to a competitive disadvantage. This perspective is shared by some investors who see socially responsible behaviors as potentially diminishing a company’s future financial performance and thereby generating lower returns. Nejati et al. [52] argue that society-related SR practices are the least likely to increase competitive performance. From another perspective, Sameer [24] argues that a company performs negatively when it discloses its SR practices related to the environment. This is due to the high costs associated with disclosing these types of practices, which ultimately outweigh the benefits they could bring to companies. Papagrigoriou et al. [53] share this view, concluding that there is no significant correlation between SR and the financial performance of companies, even though a large part of the sample discloses SR activities. On the other hand, Alareeni and Hamdan [54] found, through an analysis of the financial (ROE), operating (ROA), and market performance (Tobin’s Q) of S&P 500 companies, that corporate disclosure of environmental, social, and governance (ESG) aspects positively affects companies’ performance measures. Li et al. [55] found the same relationship in a study of 350 companies listed in the FTSE index. Similarly, Yoo and Managi [56] reached an identical conclusion using a large sample of UK public companies from the Bloomberg database for the period 2004 to 2013.

However, it should be noted that when the analysis is conducted with each dimension of SR considered in isolation, the relationship each establishes with corporate performance varies. More precisely, Alareeni and Hamdan [54] and Li et al. [55] found that environmental and social disclosure negatively affects the operational and financial performance of companies. This may occur because these socially responsible practices imply higher costs, consequently harming operational and financial performance [55]. Nevertheless, at the market performance level, both authors found a positive relationship between Tobin’s Q and social and environmental practices. In summary, the disclosure of aspects related to a company’s management is positively related to operational and market performance but negatively related to financial performance [54]. On the other hand, Buallay [36] found a slightly different relationship in European Union companies between the same variables. Specifically, Buallay, like Alareeni and Hamdan [54] and Li et al. [55], found that the environmental dimension of the ESG score is positively related to Tobin’s Q. However, unlike the others, Buallay found that financial profitability (ROE) is positively influenced by environmental disclosures. Regarding the governance dimension of the ESG score, Buallay found it positively affects Tobin’s Q while negatively affecting financial and operational performance. This means that while disclosure of this type of information decreases asset efficiency (ROA) and return on equity (ROE), it increases market performance. Further, within the governance dimension, Pham et al. [48] found an inverse relationship between these scores and business performance. This relationship is primarily driven by the use of company resources for unjustified purposes that do not promote a good cause but merely satisfy the desires of board members. This directly affects the shareholders, potentially leading to an agency problem and consequently harming the value of the company. Elouidani and Zoubir [57], who conducted a study of 20 entities listed on the Casablanca Stock Exchange, concluded that companies engaged in areas of social welfare and environmental protection obtain lower operating results, thereby harming their value on the stock exchange. On the other hand, Elouidani and Zoubir [57] believe that investments in SR are made with a long-term perspective, meaning companies should not expect an immediate improvement in financial performance when they first adopt SR practices. Madueño et al. [58] observed this scenario in practices associated with environmental protection, noting that their effects are not immediately visible to stakeholders. Other authors argue that SR only pays off after a certain investment threshold is reached; before that point, additional SR expenses can decrease business performance [41].

Finally, it should be noted that the influence of SR on business performance is affected by the extent of a company’s disclosure. In short, companies will only reap all the economic benefits associated with SR, including improved business performance, when their practices are properly communicated and appropriate disclosure channels are used [58,59,60], as we will see in the following section.

2.4. Importance of Social Disclosure and Its Standardization

The disclosure of SR has gained particular importance in the business world over the years. In the European Union, for example, various government authorities are actively establishing and implementing sustainability reports to strengthen relationships with society and business communities [36]. Stakeholders such as investors, shareholders, creditors, and debtors are increasingly paying attention to sustainability reporting, which consequently influences their decisions regarding the company [36].

As previously mentioned, SR enables a company to boost its economic results, create value, and improve its performance. However, to achieve these benefits, the company must report its socially responsible activities [60]. According to Lys et al. [61], referenced by Faria [60], existing accounting and financial reporting standards do not adequately consider sustainability, environmental, and SR aspects. Therefore, some companies face difficulties in reporting, measuring, and recognizing social aspects, which explains the lack of communication of SR practices [32]. Given these difficulties and the desire to promote their image and social status, companies often choose to disclose their socially responsible practices through sustainability reports in parallel with their financial reporting [2], referenced by [60]. However, Huang and Wang [62] argue that companies are more likely to initiate SR reporting following announcements of regulatory violations by supervisors. Therefore, SR disclosure can be seen by these authors as a strategy to disguise a bad reputation. To assess companies’ sustainable efforts and enhance the credibility of sustainability reports, many organizations choose to follow the Global Reporting Initiative (GRI) standards [63]. The GRI standards provide a set of principles that enable organizations to rigorously define the content of their reports. Consequently, these standards allow any organization to report its economic, environmental, and social impacts in a standardized and comparable manner, making sustainability reports more consistent, higher quality, and more reliable [64].

Communicating SR practices positively impacts the relationships a company establishes by promoting a socially responsible organizational culture and building trust with stakeholders [65]. However, SR reporting should be approached with caution. Both Viererbl and Koch [66] and Chaudhri [65] assert that a high degree of SR reporting positively affects recipients’ perceptions only if the company is genuinely engaged in numerous SR activities. Over-communicating about limited activities can negate or even reverse the positive effects of SR communication. Thus, while companies recognize that SR communication can build a good image, it is crucial to maintain balance, subtlety, and modesty in reporting SR activities [65,66]. A study by the European Commission [67] concluded that companies in the European Union have weak disclosure of content related to social responsibility. Specifically, these companies do not report enough and often omit information that investors and other stakeholders consider important.

Despite the weak disclosure of SR by companies, various means, models, and support systems are available to communicate their socially responsible practices. The SA8000 (Social Accountability 8000) certification is one such means. It is a voluntary standard auditable by a third party that addresses issues related to workers’ rights, workplace conditions, and effective management systems (Social Accountability International, 2014 [68]). Another notable standard is ISO 26000, developed by APCER (Portuguese Certification Association), which aims to guide organizations on social responsibility issues. ISO 26000 assists companies in communicating their commitments, performance, and other information regarding social responsibility (SGS, 2010). Additionally, there is AA1000 (AccountAbility 1000), one of the main international standardized models for corporate social responsibility [69]. This standard guides organizations in identifying, prioritizing, and responding to sustainability challenges to improve their long-term performance [70].

To implement and disseminate SR, the United Nations Global Compact (UNGC) is a strategic choice for many company managers. This is due to the lower costs associated with participation in the UNGC compared to other SR standards and the positive evidence regarding companies’ profit, sales volume, reputation, employee satisfaction, and customer satisfaction [71,72,73]. In summary, Cetindamar and Husoy [74] argue that while participation in the UNGC may not result in significant cost advantages, it has a strong and positive influence on firms’ market performance, providing both ethical and economic benefits. However, the lack of monitoring and enforcement mechanisms (e.g., third-party audits) for companies adopting the UNGC raises questions about whether they are effectively engaging in socially responsible behaviors [73].

In the last decade, attention to sustainability issues has grown exponentially, leading to an increased focus on companies’ disclosure of environmental, social, and governance (ESG) practices [55]. Consequently, the ESG score rating market has developed considerably in recent years and is now used by leading business consulting firms worldwide [54]. ESG disclosure plays a crucial role in reducing asymmetric information between companies and stakeholders, thereby strengthening their relationship [55]. Furthermore, Okafor et al. [21] note that investors, as part of their fiduciary responsibility, are increasingly committed to the environmental, social, and sustainable development causes of companies. As a result, they often refer to ESG scores when making investment decisions [21]. In essence, companies that align their strategies with good organizational, environmental, and social management present a strong investment case and are expected to secure significant returns [55]. Consequently, organizations that disclose ESG practices tend to perform better due to enhanced reputation, investor confidence, efficient use of resources, and, ultimately, higher firm value [75].

While the mentioned standards, templates, and scores provide thorough criteria, companies have other means to disclose their SR activities, allowing them to customize content and structure. Internally, companies can establish a code of ethics to guide employees in conflicting situations, establish strategies to avoid ethical mistakes, promote positive behaviors, and improve employee performance [76]. Externally, companies can use “Social Marketing”, which involves communicating their social practices through newsletters, magazines, posters, flyers, and other means that can be shared on their website, social networks, or sent via letter or email [60]. This approach is viable as it offers excellent opportunities for interaction with stakeholders and helps maintain permanent relationships [77].

Thus, it can be concluded that disclosure carries significant weight in SR, as it enhances comparability between companies, thereby influencing their image among stakeholders. Moreover, SR holds considerable relevance and is almost a benchmark in contemporary society. Therefore, if managers of organizations, regardless of size, aim for CSR to have a meaningful impact on business performance or wish it to serve as a competitive advantage, they should prioritize disclosing their CSR practices. However, despite companies having various means of disclosure, whether standardized or customized by the organizations themselves, they often exhibit poor disclosure [46].

3. Material and Methods

Based on the literature review, there is inconsistency in opinions regarding the impact of CSR on business performance, making it relevant to address this topic. Therefore, the primary objective of this study is to analyze the influence of social responsibility practices on operational, financial, and market performance in Western European listed companies. To conduct this analysis effectively, it is crucial to emphasize that social disclosure is essential for companies to leverage CSR effectively. However, some researchers argue that this disclosure entails high costs, requires time to yield returns, and may potentially harm the company’s short-term performance [41]. The literature review also reveals that ESG scores are predominantly used to evaluate companies’ SR performance; hence, this research also employs these scores. Additionally, the GC score, based on the UN Global Compact, is utilized to quantify companies’ socially responsible activities. This score is particularly noteworthy due to its status as the world’s largest SR initiative, supported by the UN [73]. The GC score provides a normative assessment of companies based on the UN Global Compact’s four core principles: human rights, labor rights, the environment, and anti-corruption. Despite its significance, there remains a scarcity of studies in the literature using this metric to establish correlations with corporate performance; therefore, this gap is addressed in this research.

This section proceeds by detailing the process for data collection. Subsequently, it introduces the dependent and independent variables utilized in the study. Finally, the hypotheses derived from the literature review are presented, along with the corresponding regression models designed to test them.

3.1. Sample

Given that Western Europe comprises 14 major countries (Belgium, Denmark, Spain, Finland, France, Germany, The Netherlands, Ireland, Italy, Norway, Portugal, Sweden, Switzerland, and the United Kingdom), the premise of this research was to collect data from the 25 most valued companies in each country. Initially, the total number of companies intended for inclusion in the sample was 350. However, after excluding companies without data on the variables under study and performing a statistical correction to remove outliers—specifically, two companies with extremely high or low financial performance—the final sample consisted of 216 companies. The research period spans five years, from 2017 to 2021 inclusive. Choosing this timeframe necessitated excluding some companies from the sample that entered the stock market after 2017. This period was selected due to the availability of a larger number of observations and comprehensive financial and social responsibility data.

Data regarding the ESG and GC scores were sourced from the ESG Book (https://app.esgbook.com/, accessed on 1 July 2022) database provided by Arabesco S-Ray. ESG Book is globally recognized for sustainability data and technology, offering transparent and comparable ESG data from over 25,000 companies. It is a trusted data source, collaborating with reputable entities such as the Global Reporting Initiative (GRI) and the European Union Sustainable Finance Disclosure Regulation (EU SFDR). The data collection process is conducted meticulously, ensuring transparency and accuracy. Financial information for European companies was obtained from the Finbox (http://finbox.com/, accessed on 1 July 2022) database, renowned as the largest repository of valuation models and risk metrics available online. It is important to note that these financial data were collected in dollars due to the varied transaction currencies across the analyzed countries.

3.2. Variables’ Measurement

This section presents the dependent and independent variables to be used in the study, along with the rationale for their selection.

3.2.1. Dependent Variables

The evaluation of a company’s performance is typically conducted through financial statements. Therefore, one of the most common analytical approaches to assess company performance involves analyzing financial indicators [49]. In the estimated models, there are three dependent variables that measure the operating, financial, and market performance of the firms. Specifically, the variable return on equity (ROE) is used to assess financial performance. For operational performance, the variable considered is return on assets (ROA). Finally, Tobin’s Q is used to evaluate market performance. These variables are widely used by several authors in assessing the impact of SR on business performance [21,37,40,45,49,54,55,56]. ROE is a profitability ratio designed to assess the return on investment provided to the company’s shareholders [78]. This ratio is calculated by dividing the net income by the company’s assets over a specific period [54]. ROA evaluates how efficiently a company utilizes its assets to generate profit, making it a widely used metric to assess profitability [54,78]. It is calculated by dividing the net income for the period by the total assets [37]. Tobin’s Q is among the most comprehensive financial ratios used to evaluate a company’s performance, especially in the long term, and is therefore frequently employed in various studies [45,56]. This ratio reflects the firm’s management of resources and capabilities, where a higher value indicates that the firm’s assets are valued more than their replacement cost [45]. In this study, Tobin’s Q is calculated as the ratio of the firm’s market value to its total assets [55]. Similar to the other ratios, Tobin’s Q is expressed as a percentage. A value exceeding 100% indicates that the company’s market value exceeds its asset value, whereas a value below 100% indicates the opposite.

3.2.2. ESG and GC Scores

To quantify the social responsibility (SR) of the companies in the sample, we utilized ESG scores obtained from the ESG Book database. These scores serve as a measurable indicator of an organization’s socially responsible performance. The ESG score encompasses 22 sustainability-related themes that recalibrate the scores across three primary dimensions: social, environmental, and governance. The final ESG score, reflecting the company’s overall ESG performance, commitment, and effectiveness based on publicly disclosed information, ranges from 0% to 100%. It is derived from a weighted aggregation of scores from each dimension, with weights adjusted according to materiality. Scores for each dimension, also ranging from 0% to 100%, are calculated based solely on characteristics within each theme.

Additionally, the GC scores were sourced from the ESG Book database. These scores provide a normative evaluation of companies against the four core principles of the United Nations Global Compact: human rights, labor rights, the environment, and anti-corruption. The Global Compact is recognized as the largest global corporate sustainability initiative, urging companies and stakeholders not only to operate responsibly but also to pursue opportunities that advance sustainable development goals [79]. Through Arabesque S-Ray, these principles are quantified systematically for the first time. Similar to the ESG, the GC score is meticulously calculated. Both category scores and the overall GC score range from 0% to 100%, with higher scores indicating stronger performance. Initially, each category contributes 25% to the overall GC score, but adjustments occur if a category’s score falls below the 50% threshold (neutral point). Utilizing both scores is crucial for conducting a comprehensive analysis of the impact of SR on company performance.

3.3. Additional Control Variables

Additional control variables can significantly impact the dependent variable, and therefore, they should be considered when estimating models that assess the impact of SR on firms’ performance. In the academic literature on SR, several authors have utilized various additional control variables associated with firms, such as size, age, leverage, capital intensity, and industry sector [24,44,56].

In this study, the size of the sample firms is treated as a control variable. Some studies, such as Rossi et al. [29], found no significant relationship between firm performance and firm size. However, other studies have shown that firm size does influence the relationship between firm performance and SR, with larger firms typically allocating more financial resources to SR activities compared to smaller firms [45,49]. Therefore, this study uses the natural logarithm of real total assets, where nominal total assets were deflated using the inflation rate measured by the Consumer Price Index (CPI) for each country and year, as a measure of firm size, consistent with prior research [24,44,54,56,80,81]. Following the same rationale as previous research in this field, another crucial control variable for this study is age [45]. Han and Kim [82] discovered that the influence of SR on business performance diminishes as a firm’s age increases, possibly due to the embedded values generated by SR practices over time. Conversely, Sun [47] observed that older firms tend to engage in a greater number of socially responsible activities compared to younger firms, often because of greater financial resources available to them. Additionally, the sector in which a company operates is another control variable emphasized by various researchers (e.g., Waddock and Graves [83]). Madorran and Garcia [44] argue that the relationship between SR and firm performance can be influenced by the sector, and studies without this variable have sometimes failed to establish definitive connections between SR and business performance. In this study, since the focus is on listed companies, the following sectors are considered: Capital Goods, Consumer Goods, Conglomerates, Discretionary Consumer, Cyclical Consumption, Non-Cyclical Consumption, Energy, Finance, Real Estate, Manufacturing, Materials, Basic Materials, Healthcare, Services, Information Technology, Communication, Transportation, and Utilities. These sectors are classified by the Investing platform (https://pt.investing.com/, accessed on 1 July 2022) for listed companies.

3.4. Hypotheses

As the primary objective of this research is to investigate the relationship between SR and the operational, financial, and market performance of companies, the following hypotheses were formulated:

H1: 

SR has a positive effect on the operational performance of the company, as measured by the return on assets (ROA).

H2: 

SR has a positive effect on firm’s market performance, as measured by Tobin’s Q.

H3: 

SR has a positive effect on the financial performance of the company, as measured by the return on equity (ROE).

It is noteworthy that a significant number of studies have identified a negative relationship between corporate performance and environmental and social disclosure. With this in mind, the dimensions of the ESG score are also intended to validate the following hypotheses:

H4: 

Social disclosure negatively affects firms’ operating performance, as measured by ROA.

H5: 

As the company increases its score in terms of the social dimension of the ESG score, its market performance is lowered, as measured by Tobin’s Q.

H6: 

There is a negative relationship between social disclosures and ROE.

H7: 

Environmental disclosure negatively affects firms’ operating performance, as measured by ROA.

H8: 

As the firm increases its score in terms of the environmental dimension of the ESG score, Tobin’s Q (market performance) is lowered.

H9: 

There is a negative relationship between environmental disclosures and ROE.

Finally, it has also been observed that socially responsible practices that consider employees’ interests lead to an increase in companies’ financial performance. Based on this finding, the following hypothesis was formulated:

H10: 

The relationship between SR and financial performance is positive when mediated by practices associated with employees.

3.5. Regression Models

The methodology adopted in this research involves the estimation of panel data models. This approach is most suitable because it incorporates both temporal and spatial dimensions, which is essential given that the data span different countries over the period from 2017 to 2021. According to Battisti and Smolski [84], analyzing individual variables (firms) over time necessitates a sophisticated analysis, making the panel data model the appropriate choice. This model includes three types of regression: pooled OLS, fixed effects, and random effects. To determine the most appropriate regression type for each model, several tests were conducted: the F-test to identify the best model between pooled OLS and fixed effects models; the Breush–Pagan test to decide between pooled OLS and random effects models; and finally, the Haussman test to ascertain the better fit between fixed effects and random effects models [84].

To evaluate the influence of social responsibility on company performance, 10 models were estimated. For clarity and ease of interpretation in the analysis of results, each of these models has been assigned a number. It is important to note that, except for models VII, VIII, IX, and X, each of the models consists of three regression equations. The primary objective is to assess the relationship between social responsibility and operational, financial, and market performance. Therefore, each regression equation includes a different dependent variable associated with business performance (BP): ROA (operational performance), Tobin’s Q (market performance), or ROE (financial performance). Models VII, VIII, IX, and X each consist of only two regression equations: one with the dependent variable ROA and the other with ROE. Tobin’s Q was not included in these models due to its negligible impact on the results.

It should also be noted that the additional control variables are as follows:

• Firm size: measured as the natural logarithm of real total firm assets (LogSize).
• Sector of activity: represented by eleven dummy variables, each corresponding to a specific sector. Each firm’s sector is assigned a value of 1 or 0. The sectors considered are Consumer Goods, Discretionary Consumer, Energy, Finance, Real Estate, Industry, Materials, Health, Information Technology, Communication, and Utilities.
• Age: measured as the number of years from the firm’s foundation to the observation year.

The first two models (Model I and Model II) presented aim to assess the influence of the overall ESG and GC scores on the operational, financial, and market performance of Western European firms. These equations assume a direct regression between ESG/GC scores and firm performance.

$$\mathrm{I}.{BP}_{it}={\beta }_0+{\beta }_1{ESG}_{it}+{\beta }_2{LogSize}_{it}+{\beta }_3{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

$$\mathrm{I}\mathrm{I}.{BP}_{it}={\beta }_0+{\beta }_1{GC}_{it}+{\beta }_2{LogSize}_{it}+{\beta }_3{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

Model III aims to determine the impact of high and low ESG scores on firms’ performance. For this purpose, dummy variables were created based on ESG scores. Specifically, the model includes two dummy variables that categorize each firm’s ESG score as either better (score above the 3rd quartile) or worse (score below the 1st quartile). Consequently, it is assumed that scores between the 1st and 3rd quartiles will be the dummy variable omitted in this model. Regarding the interpretation of these variables, when a company scores above the 3rd quartile in overall ESG scores, the variable Best_ESG takes the value 1; otherwise, it takes 0. Conversely, the variable Worst_ESG takes the value 1 when a company scores below the 1st quartile of ESG scores, and 0 otherwise.

$$\mathrm{I}\mathrm{I}\mathrm{I}. {BP}_{it}={\beta }_0+{\beta }_1{Best\_ESG}_{it}+{\beta }_2{Worst\_ESG}_{it}+{\beta }_3{LogSize}_{it}+{\beta }_4{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

Model IV follows the same logic as the previous model, but focuses on the GC score. It replaces the dummy variables Best_ESG and Worst_ESG with Best_GC and Worst_GC.

$$\mathrm{I}\mathrm{V}.{BP}_{it}={\beta }_0+{\beta }_1{Best\_GC}_{it}+{\beta }_2{Worst\_GC}_{it}+{\beta }_3{LogSize}_{it}+{\beta }_4{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

To assess the impact of each ESG dimension (environmental, social, and governance) on company performance, Model V was estimated. In this model, the environmental dimension (Env), social dimension (Soc), and governance dimension (Gov) are represented as dummy variables. This approach follows the same rationale as the previous two models but focuses on each specific ESG dimension. The main advantage of these models lies in their ability to identify the relationships that various SR activities establish with company performance.

$$\begin{gathered} \mathrm{V}. {BP}_{it}={\beta }_0+{\beta }_1{Best\_Env}_{it}+{\beta }_2{Worst\_Env}_{it}+{\beta }_3{Best\_Soc}_{it}+{\beta }_4{Worst\_Soc}_{it}+{\beta }_5{Best\_Gov}_{it} \\ +{\beta }_6{Worst\_Gov}_{it}+{\beta }_7{LogSize}_{it}+{\beta }_8{Age}_{i,t}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it} \end{gathered}$$

Model VI assesses the relationship that each dimension of the GC score—anti-corruption (AC), human rights (HR), the environment (ENV), and labor rights (LR)—has with operational, financial, and market performance. This model follows the same rationale as Models I and II, establishing a direct relationship between GC score dimensions and business performance. While Model V applied a quartile-based categorization to ESG dimensions to assess high and low performance, a similar approach was not feasible for Model VI due to insufficient data on quartiles for the individual GC dimensions.

$$\mathrm{V}\mathrm{I}. {BP}_{it}={\beta }_0+{\beta }_1{AC}_{it}+{\beta }_2{HR}_{it}+{\beta }_3{ENV}_{it}+{\beta }_4{LR}_{it}+{\beta }_5{LogSize}_{it}+{\beta }_6{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

Finally, we also intended to evaluate how the additional control variables—size and age—influence the relationship that the ESG and GC scores establish with company performance. In this sense, four models were estimated where there is an interaction between the variables mentioned above and the SR scores (ESG and GC).

$$\mathrm{V}\mathrm{I}\mathrm{I}. {BP}_{it}={\beta }_0+{\beta }_1{ESG}_{it}+{\beta }_2{Age}_{it}+{\beta }_3{ESG\ast Age}_{it}+{\beta }_4{LogSize}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

$$\mathrm{V}\mathrm{I}\mathrm{I}\mathrm{I}.{BP}_{it}={\beta }_0+{\beta }_1{GC}_{it}+{\beta }_2{Age}_{it}+{\beta }_3{GC\ast Age}_{it}+{\beta }_4{LogSize}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

$$\mathrm{I}\mathrm{X}. {BP}_{it}={\beta }_0+{\beta }_1{ESG}_{it}+{\beta }_2{LogSize}_{it}+{\beta }_3{ESG\ast LogSize}_{it}+{\beta }_4{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

$$\mathrm{X}. {BP}_{it}={\beta }_0+{\beta }_1{GC}_{it}+{\beta }_2{LogSize}_{it}+{\beta }_3{GC\ast LogSize}_{it}+{\beta }_4{Age}_{it}+\sum {\beta }_k{Sectork}_{it}+{\mu }_{it}$$

To account for the potential impact of the COVID-19 crisis on the relationship between social responsibility and corporate performance, we included a dummy variable taking the value of one for the years 2020 and 2021, and zero otherwise, in all models. This variable was not statistically significant in all models, so it was not included in the tables with the results. Consequently, our findings suggest that the COVID-19 crisis did not significantly alter the relationship between social responsibility and corporate performance within our sample.

4. Results

This section analyzes the results obtained from the estimated empirical models, comparing them with the findings of other studies. All models in this research were estimated using R Software, version 4.2.1 [85]. Notably, due to the use of panel data regression models, each model was estimated using pooled OLS, random effects, and fixed effects methods to determine the most appropriate approach. The main issues associated with pooled OLS arise from its assumption that individual differences across entities and over time are not significant. This method combines cross-sectional and time-series data without accounting for these potential differences, which can lead to biased and inconsistent estimates if entity-specific characteristics are actually relevant. Pooled OLS does not address potential correlations within an entity over time, leading to inefficient estimates and underestimated standard errors. This, in turn, can result in misleading statistical inferences. Additionally, pooled OLS can suffer from omitted variable bias by ignoring unobserved heterogeneity, especially if the omitted variables are correlated with the included explanatory variables. Furthermore, pooled OLS assumes constant variance of errors across entities and time, an assumption that is often unrealistic in panel data, leading to inefficient and biased estimates. Despite these limitations, pooled OLS was used in this research for reference purposes only, to provide a baseline for comparison with more advanced techniques like fixed effects or random effects models, which can better handle the complexities of panel data. To ensure the absence of spurious regression, we conducted Im–Pesaran–Shin [86] and Maddala–Wu [87] panel unit root tests on all variables at levels, including firm size and age, using the purtest function from the plm package in R software. The results consistently rejected the null hypothesis of non-stationarity at an extremely low p-value threshold of less than 2.2 × 10⁻¹⁶, indicating that all variables are stationary.

4.1. Statistical Tests

When analyzing panel data, selecting the most appropriate estimation method is crucial. Three main tests help us navigate this decision process: the F-test, the Breusch–Pagan LM test, and the Hausman test. The F-test helps us choose between pooled OLS and fixed effects models. It tests the null hypothesis that there are no firm-specific effects in the data. If we reject this null hypothesis (indicating significant firm-specific effects), then fixed effects becomes a more suitable choice than pooled OLS. The Breusch–Pagan LM test aids in deciding between pooled OLS and random effects. It focuses on the error terms in the random effects model and whether they correlate with the independent variables. If we reject the null hypothesis of no correlation, this suggests a violation of the random effects assumption, and fixed effects might be preferable. Finally, the Hausman test compares the coefficients estimated by random effects and fixed effects models. It tests the null hypothesis that the random effects model is consistent (meaning its coefficients are unbiased). If we fail to reject this null hypothesis, it suggests the random effects model is a good choice. However, if we reject the null hypothesis, it indicates potential bias in the random effects model, and fixed effects might be a better option. Following a general approach, we often begin with pooled OLS estimation. The F-test then guides us towards fixed effects if firm-specific effects are evident. If the F-test is inconclusive, the Breusch–Pagan LM test can help decide between pooled OLS and random effects. Once fixed effects are deemed appropriate, the Hausman test can be used to assess the potential bias in the Random Effects model, determining if it might be a suitable alternative due to its efficiency advantages.

Table 1. Results of the statistical tests.

	Dependent Variable	p-Value
		F-Test	Breush-Pagan	Haussman
Model I (ESG—Direct relation)	ROA	1.139 × 10−13	1.903 × 10−9	1.047 × 10−9
	Tobin’s Q	<2.2 × 10−16	<2.2 × 10−16	4.474 × 10−12
	ROE	<2.2 × 10−16	<2.2 × 10−16	0.000456
Model II (GC—Direct relation)	ROA	1.181 × 10−14	6.573 × 10−10	6.474 × 10−11
	Tobin’s Q	<2.2 × 10−16	<2.2 × 10−16	8.949 × 10−9
	ROE	<2.2 × 10−16	<2.2 × 10−16	1.044 × 10−7
Model III (Best and Worst ESG)	ROA	1.084 × 10−12	1.022 × 10−09	3.058 × 10−7
	Tobin´s Q	<2.2 × 10−16	<2.2 × 10−16	<2.2 × 10−16
	ROE	<2.2 × 10−16	<2.2 × 10−16	0.00114
Model IV (Best and Worst GC)	ROA	1.286 × 10−13	4.02 × 10−10	2.923 × 10−8
	Tobin´s Q	<2.2 × 10−16	<2.2 × 10−16	9.591 × 10−9
	ROE	<2.2 × 10−16	<2.2 × 10−16	0.0002454
Model V (Best and Worst ESG Dimensions)	ROA	3.232 × 10−15	6.106 × 10−10	2.637 × 10−10
	Tobin´s Q	<2.2 × 10−16	<2.2 × 10−16	1.576 × 10−8
	ROE	<2.2 × 10−16	<2.2 × 10−16	1.737 × 10−7
Model VI (GC Dimensions)	ROA	1.437 × 10−13	2.457 × 10−08	1.734 × 10−10
	Tobin´s Q	<2.2 × 10−16	<2.2 × 10−16	2.637 × 10−5
	ROE	<2.2 × 10−16	<2.2 × 10−16	2.534 × 10−7
Model VII (ESG*Age)	ROA	6.328 × 10−15	4.356 × 10−10	1.466 × 10−10
	ROE	<2.2 × 10−16	<2.2 × 10−16	9.351 × 10−5
Model VIII (GC*Age)	ROA	<2.2 × 10−16	9.962 × 10−11	<2.2 × 10−16
	ROE	<2.2 × 10−16	<2.2 × 10−16	3.082 × 10−8
Model IX (ESG*Size)	ROA	<2.2 × 10−16	1.056 × 10−10	<2.2 × 10−16
	ROE	<2.2 × 10−16	<2.2 × 10−16	1.515 × 10−13
Model X (GC*Size)	ROA	<2.2 × 10−16	2.128 × 10−07	<2.2 × 10−16
	ROE	<2.2 × 10−16	<2.2 × 10−16	<2.2 × 10−16

shows that the F-test p-value is less than 5% for all models. This indicates the presence of significant firm-specific effects, suggesting that the fixed effects model is more appropriate than the pooled OLS model. The Breusch–Pagan test also supports this conclusion. In all models, the null hypothesis (of no random effects) is rejected at the 5% level (p-value < 0.05), implying a violation of the assumptions for the random effects model. Finally, the Hausman test reinforces the choice of the fixed effects model. With a p-value less than 5%, the null hypothesis of no systematic bias in the random effects model is rejected. This suggests potential correlation between individual effects and explanatory variables, further strengthening the case for the fixed effects model. In conclusion, based on the results of these tests, the fixed effects model appears to be the best fit for this study. However, it is important to acknowledge that this model excludes time-invariant variables (such as sector). To address this limitation and still utilize the fixed effects framework, the Hausman–Taylor estimator was employed in this analysis. It is important to emphasize that all models used the same additional control variables: size (logarithm of assets), age, and sector. Initially, the country was also considered as a control variable, but its inclusion in the models showed it to be non-significant, thus it was not included in the final estimated models. Furthermore, certain models will distinguish between the best and worst overall ESG and GC scores, as well as the dimensions of the ESG. Initially, a score scale ranging from 0% to 100% is presented, with class intervals of 5%. Subsequently, the worst scores are those below the 1st quartile, and the best scores are those above the 3rd quartile (noting that scores are rounded to the nearest class and subject to adjustment to enhance model outcomes).

4.2. Discussion

Model I consists of three econometric regressions aimed at analyzing the direct relationship between the dependent variables (ROA, Tobin’s Q, and ROE) and the ESG score. The results are presented in

Table 2. Model I regression results.

Dependent Variable	Model I
	ROA	Tobin´s Q	ROE
(Intercept)	56.6850 (35.3542)	−102.0110 (123.8263)	94.9243 * (47.4151)
ESG	−0.5298 *** (0.1350)	1.1801 ** (0.4328)	−0.2904 (0.2029)
LogSize	−13.5375 *** (2.6388)	−32.8235 *** (8.4502)	−12.2805 ** (3.9665)
Age	1.1849 *** (0.3493)	4.8769 *** (1.1184)	0.5883 (0.5250)
Communications	44.8720 (30.5345)	193.1312 (123.3742)	18.4098 (28.4757)
Discretionary Consumer	17.2946 (23.8982)	102.7982 (101.7762)	4.1714 (16.1214)
Energy	54.6479˙ (29.1671)	186.7155 (117.2823)	25.4302 (27.7558)
Financial	37.8169˙ (21.3231)	55.1876 (88.8265)	28.2434˙ (16.9835)
Real Estate	80.3482 (49.7674)	314.7512 (201.2277)	32.7161 (46.2695)
Industry	17.9012 (21.8046)	96.3245 (92.3830)	6.0264 (15.3812)
Materials	17.1660 (22.9892)	96.6927 (97.8568)	5.7203 (15.5779)
Health	17.9463 (24.1481)	268.6576 ** (103.5142)	17.4696 (15.2822)
Technology	47.6893 (31.0592)	314.0329 * (123.7365)	6.4870 (30.6473)
Utilities	64.9141 * (31.3108)	203.7387 (124.3850)	36.7744 (31.2203)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘˙’ 0.1 ‘ ‘ 1.

. Concerning Model I, it is observed that when the dependent variable is ROA, the ESG variable (representing the ESG scores obtained by firms), LogSize, and Age are statistically significant at least at the 0.1% level. This indicates that these variables influence the operational performance of European firms. At the sector level, it can also be noted that the Energy, Financial, and Utilities sectors are relevant in this regression. Specifically, the Financial and Energy sectors are statistically significant at least at the 10% level, and the Utilities sector at least at the 5% level. From the results, it is evident that the additional control variables exhibit varied relationships with ROA. Age and the Energy, Financial, and Utilities sectors demonstrate a positive relationship with operational performance. In contrast, LogSize shows a negative relationship with this economic-financial ratio. As for the ESG variable, it demonstrates an estimated negative impact on the operational performance of European firms. More precisely, a one percentage point change in the ESG variable results in a decrease of approximately 0.53 percentage points in firms’ operational performance, ceteris paribus. Based on these findings, there is evidence to reject Hypothesis 1 (H1), which posited a positive relationship between SR and ROA.

Contrarily, in the literature review, Kamatra and Kartikaningdyah [49] and Cheng et al. [30] identified a positive relationship between SR and operational and financial performance. Rossi et al. [29], who also focused their study on European firms using the ESG score to measure SR, found a positive relationship between ESG and the ROA ratio, thus contrasting the relationship identified in this study. However, these authors used companies from the major European economies (France, Spain, Germany, and Italy) with quite large results and excluded companies from the financial sector. In contrast, the sample of this research included listed firms from smaller countries, such as Portugal or Ireland, as well as firms belonging to other large economies, such as the United Kingdom. These differences may be the cause of the contrast between these two studies. One possible justification for this contrast in results may be associated with the fact that the companies in the sample adopt socially responsible behaviors to have a positive impact on stakeholders, without emphasizing the relationship that SR establishes with business performance, as argued by Wuttichindanon [31]. However, it should be noted that the results of the first regression equation of Model I support several studies that have also confirmed the existence of a negative and significant association between operational performance and SR practices [[33,35,36], as referenced by [34]]. An interesting finding from Buallay’s [36] study is that this author also used a sample of companies from various European economies (including smaller countries) and identified an inverse relationship between SR and ROA. This inverse relationship could be justified for several reasons. While Buallay [36] attributes it to managers’ personal benefits from SR, Oh and Park [33], as referenced by Guzman et al. [34], highlight the high costs associated with SR. Despite this negative relationship, Western European companies are likely to continue with an SR policy due to the significant societal pressure they face [38]. Regarding the second regression equation of Model I, it is noteworthy that the independent variable ESG is statistically significant at least at the 1% level, and the additional control variables Age and LogSize are statistically significant at least at the 0.1% level. Additionally, the Healthcare and Technology sectors are significant for assessing market performance, with both sectors being statistically significant at least at the 1% and 5% levels, respectively. It should be noted that the additional control variables exhibit a positive relationship with Tobin’s Q, except for the LogSize variable, which shows a negative relationship. Regarding the ESG variable, a positive estimated impact on market performance can be observed. Specifically, when ESG increases by one percentage point, market performance increases by 1.18 percentage points, ceteris paribus. Based on this observation, it is possible to accept Hypothesis 2 (H2) posited earlier. This positive relationship identified in this study contrasts with findings in the literature. Specifically, Bannier et al. [43] found no relationship between market performance and ESG, while Hirigoyen and Poulain-Rehm [35] identified a negative relationship between SR and market performance. However, identifying a positive relationship between SR and market performance is not unprecedented in the literature, as other authors have also observed this relationship [21,37,40,55,56,75]. The rationale provided by most authors for this positive relationship with the market primarily revolves around the favorable image companies project to stakeholders, particularly investors, whose decisions are influenced by ESG scores. As SR becomes increasingly integral to investors’ considerations, socially responsible companies enhance their market value through improved reputation and investor trust, thereby boosting their Tobin’s Q ratio [75]. Another perspective supporting this positive correlation between the ratio and ESG scores is articulated by Yoo and Managi [56], suggesting that companies with higher Tobin’s Q values in the sample exhibit greater market value and financial capacity, allowing them to undertake more socially responsible initiatives and consequently elevate their ESG scores. Conversely, the final regression equation of Model I is less pertinent in explaining the relationship between social responsibility and financial performance, as ROE does not achieve statistical significance at a reasonable confidence level. Hence, Hypothesis 3 (H3) cannot be accepted or rejected. Indeed, aside from the Financial sector, which is statistically significant at least at the 10% level and demonstrates a positive relationship with operational performance, only LogSize is statistically significant in this equation at the 1% level, indicating a negative correlation with ROE. This finding aligns with conclusions from several other studies that similarly found no association between SR and ROE [35,41,42,44,49]. The absence of a relationship between SR and ROE in Western European companies may stem from these companies using SR primarily to enhance product value without expecting it to directly influence financial performance, as argued in Nollet et al. [41].

Model II, the results of which are presented in

Table 3. Model II regression results.

Dependent Variable	Model II
	ROA	Tobin´s Q	ROE
(Intercept)	57.1423˙ (34.3489)	−4.7509 (122.8456)	121.7688 ** (46.1777)
GC	−0.7182 *** (0.1532)	−0.0132 (0.4944)	−0.8302 *** (0.2296)
LogSize	−11.6821 *** (2.6054)	−35.9915 *** (8.4114)	−11.0023 ** (3.9060)
Age	1.1253 ** (0.3481)	4.9057 *** (1.1237)	0.5275 (0.5218)
Communications	43.3392 (29.4106)	191.6631 (123.7006)	16.2484 (27.8473)
Discretionary Consumer	15.8547 (22.7791)	100.3521 (102.0192)	1.8521 (15.3881)
Energy	51.1334˙ (28.1217)	188.2435 (117.6029)	21.7984 (27.1780)
Financial	34.4318˙ (20.4152)	58.2617 (89.0431)	25.1791 (16.4424)
Real Estate	78.4639 (47.9294)	307.1653 (201.7597)	28.4946 (45.2416)
Industry	16.9234 (20.8019)	97.3427 (92.5961)	5.1764 (14.7338)
Materials	17.7666 (21.9117)	99.1242 (98.0830)	7.0696 (14.8654)
Health	19.1517 (22.9789)	266.4072 * (103.7435)	18.2480 (14.4603)
Technology	46.9970 (29.9921)	311.9650 * (124.0692)	5.1306 (30.0506)
Utilities	62.2473 * (30.2462)	207.8913˙ (124.7107)	34.8285 (30.6178)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

, comprises three regression equations designed to explore the relationship between GC score and financial, market, and operational performance. It is worth noting that this score of social responsibility is derived not only differently from that for ESG, but also incorporates other dimensions (anti-corruption, environment, human rights, and labor rights). In essence, this score is underpinned by normative principles, specifically those of the UN Global Compact. Regarding the results of Model II, in the first regression equation, GC score and LogSize are statistically significant at the 0.1% level. The control variable Age, on the other hand, is statistically significant at least at the 1% level, while the energy and financial sectors are significant at the 10% level. Notably, the Utilities sector is statistically significant at least at the 5% level. Concerning the relationship these variables establish with operational performance, it is observed that the control variables have a positive estimated impact on ROA, except for LogSize, which shows a negative correlation. Despite GC score assessing aspects of social responsibility that the ESG does not cover (such as human rights, labor law, and anti-corruption), it still leads to decreased asset efficiency, indicating a decline in operational performance. Specifically, a one percentage point increase in this score results in a 0.72 percentage points decrease in ROA, ceteris paribus. This observation further supports the rejection of Hypothesis 1 (H1). Unlike Model I, the regression equation of Model II, which assesses the impact of GC score on market performance, does not yield results of significant relevance. Therefore, Hypothesis 2 (H2) cannot be accepted or rejected based on these findings alone. The only notable findings indicate that the additional control variables LogSize, Age, and the Health, Technology, and Utilities sectors are sufficiently significant to influence Tobin’s Q. Consequently, it can be concluded that market performance is not influenced by GC score, contrary to the perspective of Cetindamar and Husoy [74], who argue for a strong and positive influence of the UN Global Compact on market performance. The lack of monitoring and auditing in this SR measure may contribute to the absence of a relationship between GC score and market performance. Finally, the last regression equation in Model II assesses the impact of GC score on financial performance. Unlike Model I, in this case, the independent variable GC score is statistically significant at least at the 0.1% level. LogSize, on the other hand, is statistically significant at least at the 1% level. Both LogSize and GC score show a negative relationship with ROE. More precisely, a one percentage point increase in the GC score corresponds to a 0.83 percentage points decrease in ROE, ceteris paribus. Given these results, Hypothesis 3 (H3), which assumed a positive relationship between SR and financial performance, is rejected. This identified relationship between GC score and market performance contradicts the findings of Alareeni and Hamdan [54] and Li et al. [55], who found a positive relationship between ROE and SR. Additionally, it also contradicts the views of authors who advocate a neutral relationship between these variables [42,44]. In conclusion, it is possible to determine that GC score negatively influences both operational and financial performance and has no influence on market performance. This is an important contribution since the consulted literature only mentioned that participation in the UNGC positively influences profit, sales volume, reputation, employee satisfaction, and customer satisfaction [71,72,73]. Moreover, even though participation in the UNGC entails lower costs compared to other SR standards [73], this benefit is ultimately not sufficient to alter the relationship between GC score and business performance. In short, when comparing the results of the estimates from Models I and II, there is evidence that companies implement SR strategies through the GC score dimensions (significant estimates in ROA and ROE), while investors (the market) place greater value on social responsibility (SR) activities developed in the ESG dimensions. This can indicate how management should align with stakeholder expectations.

Model III was developed to ascertain the impact high and low ESG scores have on firms’ financial, market, and operational performance. Specifically, two dummy variables were defined for the overall ESG scores. The Best_ESG dummy variable takes the value 1 when the firms in the sample score above 65% in the ESG scores (3rd quartile) and 0 otherwise. The dummy variable Worst_ESG takes the value 1 when the firm scores below 50% in the overall ESG scores (1st quartile) and 0 otherwise. The results of Model III are presented in

Table 4. Model III regression results.

Dependent Variable	Model III
	ROA	Tobin´s Q	ROE
(Intercept)	25.4677 (33.3778)	−65.2301 (122.9641)	78.5097˙ (44.7991)
Best_ESG	−5.7485 ** (1.8950)	22.6326 *** (6.0188)	−5.2034˙ (2.8341)
Worst_ESG	−0.4642 (1.7136)	−6.5064 (5.4427)	−2.9070 (2.5628)
LogSize	−12.4194 *** (2.6343)	−33.8288 *** (8.3669)	−11.4910 ** (3.9397)
Age	1.0834 ** (0.3520)	5.2947 *** (1.1180)	0.5134 (0.5264)
Communications	41.2441 (28.9145)	211.1546 (131.0249)	15.6861 (27.8769)
Discretionary Consumer	17.3157 (22.1718)	107.0221 (109.1107)	4.5415 (15.1782)
Energy	50.3620˙ (27.6533)	204.0253 (124.4084)	22.2954 (27.1842)
Financial	35.9464˙ (19.9420)	59.4542 (94.8259)	26.8366 (16.3193)
Real Estate	76.5860 (47.1238)	341.7755 (213.7552)	30.0397 (45.3060)
Industry	16.7269 (20.2658)	101.2146 (98.9399)	5.4430 (14.5649)
Materials	15.1838 (21.3260)	103.1289 (104.8858)	4.4524 (14.6551)
Health	17.6587 (22.3418)	272.6021 * (111.0946)	17.1711 (14.2195)
Technology	43.6776 (29.5651)	334.8863 * (131.0653)	3.5718 (30.1482)
Utilities	58.8781 * (29.8051)	225.3619˙ (131.6125)	32.2937 (30.6722)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

. As seen from the results obtained in the first equation of Model III, the variables Best_ESG and Age are statistically significant at the 1% level. On the other hand, the variable LogSize and the Utilities sector are statistically significant at the 0.1% and 5% levels, respectively. The Energy and Financial sectors are statistically significant at the 10% level. This means that all these variables influence ROA. It should be noted that the dummy variable Worst_ESG has no statistical relevance in this regression equation. In terms of interpreting the results of the variables, it can be seen that all the significant control variables, except LogSize, positively relate to ROA. LogSize shows a negative relationship. As for the Best_ESG variable, it can be seen that operating performance worsens when firms score higher than 65% on the ESG score. When the Best_ESG dummy variable assumes the value 1, ROA decreases by 5.75 percentage points, ceteris paribus. Therefore, it can be concluded that European firms aiming for high levels of operational performance should pay attention to high ESG scores, as these may impair operational performance. This inverse relationship between ESG scores and ROA was already found in the direct relationship but was more pronounced for high scores. The results of the second regression equation show that the variables Best_ESG, LogSize, and Age are statistically significant at the 0.1% level, while the healthcare and technology sectors are statistically significant at the 5% level. The Utilities sector is statistically significant at the 10% level. In this case, contrary to the results of the first equation, the Best_ESG variable exhibits a positive relationship with market performance. More precisely, when firms score above 65% in ESG, the Tobin’s Q ratio increases by 22.63 percentage points. Given this, firms in the sample that want to increase their market performance should target their efforts toward achieving better ESG scores. In addition to the direct positive relationship between these scores and market performance, it was also found that high scores significantly promote higher performance. Finally, the results of the last equation of Model III show that only the control variable LogSize is statistically significant at the 1% level. This means that neither scores above 65% on the ESG score nor scores below 50% can influence the financial performance of the sample firms. This is a natural conclusion since it has been previously observed that there is no direct relationship between ROE and ESG scores.

Model IV follows the same logic as the previous one, except that in this case, instead of defining the scale of better and worse scores with the ESG, the GC score was used. For the GC score, two dummy variables were created to distinguish the best and worst scores. The Best_GC score takes the value 1 when the scores of the companies are above 65% (class closest to the 3rd quartile) and 0 otherwise. On the other hand, the Worst GC score variable takes the value 1 when the scores obtained by the companies in the sample are below 55% (class closest to the 1st quartile) and 0 otherwise. For this SR score, three regression equations were estimated, each with a different business performance variable—ROA, Tobin’s Q, and ROE—whose results are presented in

Table 5. Model IV regression results.

Dependent Variable	Model IV
	ROA	Tobin´s Q	ROE
(Intercept)	4.2688 (34.4523)	−0.7624 (119.1825)	68.0488 (44.8250)
Best_GC	0.4448 (1.5574)	−1.3215 (5.0034)	−0.4556 (2.3395)
Worst_GC	7.2873 *** (1.9790)	−2.1575 (6.3575)	3.4467 (2.9727)
LogSize	−12.1977 *** (2.6261)	−35.8136 *** (8.4364)	−11.4498 ** (3.9447)
Age	1.2681 *** (0.3538)	4.8413 *** (1.1367)	0.6061 (0.5315)
Communications	49.7883 (32.1831)	188.8763 (122.8293)	19.9095 (28.8412)
Discretionary Consumer	19.0678 (25.4825)	99.5246 (100.9910)	4.7173 (16.3721)
Energy	57.3168˙ (30.7036)	185.7538 (116.7827)	25.7765 (28.0915)
Financial	36.7175 (22.6061)	57.5403 (82.2144)	27.2433 (17.1685)
Real Estate	90.3023˙ (52.4873)	302.3842 (200.3971)	36.0392 (46.9275)
Industry	18.8434 (23.2168)	96.4384 (916834,)	6.1583 (15.5933)
Materials	17.3833 (24.4977)	98.6364 (97.0714)	5.7009 (15.7747)
Health	20.5831 (25.7871)	265.5477 ** (102.6737)	18.5830 (15.5290)
Technology	52.9767 (32.6243)	309.0968 * (123.2648)	8.1592 (30.9946)
Utilities	68.1234 * (32.8812)	204.6540˙ (123.9672)	37.1732 (31.6011)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

. The goal is to understand how corporate reputation in terms of SR affects business performance. As can be seen, the results of Model IV are not very compelling, even though several adjustments were made. In fact, the variables Best_GC score and Worst GC score do not influence the market and financial performance of the sample companies, as these variables never achieve statistical significance. For Tobin’s Q, this was an expected result, given that a direct relationship between market performance and GC score was not previously identified. However, the same cannot be said for ROE, as this variable was previously found to have a negative relationship with GC score at a reasonable significance level. In both the equation with Tobin’s Q as the dependent variable and the equation with ROE, only the additional control variables show statistical significance. More precisely, in the second equation, the variables LogSize and Age are statistically significant at the 0.1% level, and the Health, Technology, and Utilities sectors are statistically significant at the 1%, 5%, and 10% levels, respectively. Except for LogSize, all variables have a positive relationship with Tobin’s Q. In the third regression equation, only the LogSize variable is statistically significant at the 1% level and shows a negative relationship with ROE.

In the case of the first equation, where ROA is the dependent variable, the scenario is different. Here, the independent variable Worst_GC score is statistically significant at the 0.1% level, indicating that this variable affects operating performance. Interestingly, low or reasonable scores lead to better operational performance. When the dummy variable Worst_GC score takes the value 1, operational performance increases by 7.29 percentage points, ceteris paribus. This reinforces the theory that focusing on the UN Global Compact standards may not be the best means for companies to improve their operational performance. In addition to previously identifying an inverse relationship between ROA and GC score, Model IV shows that scores below 55% can be beneficial for operational performance. It should also be noted that in the first regression equation of Model IV, the additional control variables LogSize and Age are statistically significant at the 0.1% level, while the Energy and Real Estate sectors are statistically significant at the 10% level. The Utilities sector is statistically significant at the 5% level. Regarding the relationship these variables have with ROA, except for LogSize, all the statistically significant control variables show a positive relationship with operating performance.

Model V was estimated to assess the impact of obtaining better and worse scores in the dimensions of the ESG on the financial, market, and operational performance of the sample companies. In this case, the division between high and low scores was conducted as follows:

• In the environment dimension, a score less than 60% (1st quartile) is considered worse and a score greater than 70% (3rd quartile) is considered better.
• In the social dimension, a score lower than 55% (1st quartile) is considered worse and a score higher than 65% (3rd quartile) is considered better.
• In the governance dimension, a score lower than 35% (1st quartile) is considered worse and a score higher than 70% is considered better (for this variable, the third quartile value was 60%, but considering the 70% score improved the model results).

The results for Model V are shown in

Table 6. Model V regression results.

Dependent Variable	Model V
	ROA	Tobin´s Q	ROE
(Intercept)	30.0137 (32.5927)	−33.9364 (122.5424)	88.4181 * (43.9611)
Best_Env	−3.5957 * (1.4995)	−0.6359 (4.9055)	−4.2539˙ (2.2623)
Worst_Env	2.9824 (2.9895)	−1.5049 (9.7799)	3.7681 (4.5103)
Best_Soc	−2.6973˙ (1.4626)	9.1115˙ (4.7849)	−3.9687˙ (2.2067)
Worst_Soc	6.1360 ** (1.8846)	−0.9396 (6.1655)	7.0687 * (2.8434)
Best_Gov	−17.4641 *** (3.1917)	20.6839 * (10.4414)	−18.9519 *** (4.8154)
Worst_Gov	0.9694 (1.6599)	−0.4061 (5.4302)	−2.9699 (2.5043)
LogSize	−12.5375 *** (2.6035)	−36.1656 *** (8.5173)	−11.0552 ** (3.9280)
Age	1.0561 ** (0.3499)	5.1818 *** (1.1446)	0.3821 (0.5279)
Communications	39.8719 (28.3366)	205.6913 (129.6186)	8.9162 (27.1195)
Discretionary Consumer	16.0835 (21.6174)	105.2875 (107.3724)	1.2879 (13.9641)
Energy	48.2089˙ (27.1118)	202.1909 (123.1348)	14.7930 (26.4997)
Financial	34.7587˙ (19.4897)	63.8061 (93.5167)	23.0921 (15.5492)
Real Estate	71.6216 (46.1870)	331.6589 (211.4639)	15.6712 (44.0798)
Industry	15.6708 (19.7651)	102.1279 (97.4000)	2.4972 (13.5320)
Materials	14.8590 (20.7700)	102.7417 (103.1677)	2.7098 (13.4122)
Health	17.1489 (21.7520)	270.8463 * (109.2250)	15.5432 (12.7747)
Technology	42.4919 (28.9570)	326.2290 * (129.7329)	−3.2068 (29.3777)
Utilities	58.1542 * (29.2750)	222.6474˙ (130.4945)	25.7619 (30.0866)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ‘ 1.

. From the results of the first regression equation of Model V, it is possible to observe that the best scores in the environment, social, and governance dimensions, as well as the worst score in the social dimension, influence the operational performance of Western European companies. This is because the variables Best_Gov, Best_Env, and Best_Soc are statistically significant at least at the 0.1%, 5%, and 10% levels, respectively. The variable Worst_Social is statistically significant at least at the 1% level. Regarding the additional control variables, it can be seen that LogSize, Age, and the Utilities sector are statistically significant at least at the 0.1%, 1%, and 5% levels, respectively. The Energy and Financial sectors are also statistically significant, but this time at the 10% level. In terms of the interpretation of the variables, high scores in the environment dimension impair the operational performance of the companies in the sample, thus confirming Hypothesis 7 (H7) raised earlier. More precisely, scores above 70% in this dimension decrease the ROA ratio by about 3.60 percentage points, ceteris paribus. Regarding the social dimension, the relationship between the variable Best_Soc and operational performance is similar, as scores higher than 65% imply a 2.67 percentage points decrease in ROA. This finding confirms Hypothesis 4 (H4), which argues for a negative relationship between social practices and operational performance. Interestingly, scores below 55% positively influence operational performance; if the Worst_Social variable assumes the value 1, the ROA ratio increases by about 6.14 percentage points, ceteris paribus. Regarding the governance dimension, scores above 70% significantly harm the operational performance of European companies.

When the Best_Gov variable takes the value 1, the ROA ratio decreases by 17.46 percentage points, ceteris paribus. It should also be noted that all the additional control variables are positively related to ROA, except for the LogSize variable, which shows an inverse relationship. The results of the second regression equation indicate that only the Best_Soc and Best_Gov variables influence the market performance of the sample companies, as both are statistically significant at least at the 10% and 5% levels, respectively. The additional control variables LogSize and Age are statistically significant at least at the 0.1% level, while the Healthcare and Technology sectors are significant at the 5% level. The Utilities sector is statistically significant at the 10% level. The variables associated with the environmental dimension do not show any significance, so no relationship between environmental practices and Tobin’s Q was identified, making it impossible to validate Hypothesis 8 (H8). Interestingly, the relationships that the Best_Gov and Best_Soc variables establish with market performance are contrary to those they establish with operational performance. Specifically, in the second equation’s results, scores higher than 65% in the social dimension increase the Tobin’s Q ratio by 9.11 percentage points, ceteris paribus. Scores above 70% in the governance dimension imply an increase of about 20.68 percentage points in the Tobin’s Q ratio, ceteris paribus. Given these results, Hypothesis 5 (H5) is rejected, as it posited an inverse relationship between social practices and market performance. All additional control variables are positively related to the dependent variable, except for LogSize, which shows an inverse relationship. Regarding the last regression equation of Model V, which contains ROE as the dependent variable, it presents two statistically significant variables at least at the 10% level (Best_Env and Best_Soc), one statistically significant variable at least at the 5% level (Worst_Soc), and another variable at the 0.1% level (Best_Gov). Note that among the additional control variables used, only LogSize is statistically significant at the 1% level, and as observed in previous models, this variable is negatively related to ROE. From the results of the last regression equation of Model V, several conclusions can be drawn. First, an inverse relationship was observed between the independent variables Best_Env, Best_Soc, and Best_Gov and the dependent variable ROE. More precisely, scores higher than 70% in the environment dimension result in a decrease of about 4.25 percentage points in financial performance, ceteris paribus. In the case of the social dimension, a score higher than 55% implies a decrease of 3.97 percentage points in the ROE variable, ceteris paribus. The governance dimension exhibits the most pronounced inverse relationship; when scores in this dimension are higher than 70%, ROE decreases by 18.95 percentage points, ceteris paribus. It should also be noted that the Worst_Soc variable is positively related to ROE. This means that scores below 55% in the social dimension imply an increase of 7.07 percentage points in financial performance, ceteris paribus. These results support both Hypothesis 6 (H6) and Hypothesis 9 (H9), which argued for a negative relationship of social and environmental practices with ROE. As seen from the results presented above, obtaining high scores in the environment dimension hurts the operational and financial performance of Western European companies. Congruent with the results of this research, Alareeni and Hamdan [54], Elouidani and Zoubir [57], Faria [60], Makni et al. [50], Riyadh et al. [51], and Sameer [24] also identified an inverse relationship between environmental practices and operational and financial performance. This is a recurrent conclusion in the literature, justified by the high costs associated with these practices, whose benefits in business performance are only reflected in the long term and from a certain level of investment [41,57,58,88]. On the other hand, the positive relationship between the environmental dimension of SR and business performance is also advocated in the literature [36,48,49,54]. It is important to highlight the study by Alareeni and Hamdan [54] that verified both interactions; initially, as in this research, they observed an inverse relationship between environmental practices and operational and financial performance. However, they also found a positive relationship between these practices and market performance, thus contradicting the absence of a relationship identified in this research between these variables. Regarding the social dimension, the results of Model V reveal two key relationships: a negative relationship between social practices and both operational and financial performance, and a positive relationship between these practices and market performance. Similar findings were previously reported by Makni et al. [50] and Li et al. [55]. In the literature, several authors argue that social practices harm company performance due to the high costs associated with these activities, which ultimately outweigh their economic benefits [24,51,52,57]. However, some studies also highlight a positive relationship between social practices and business performance [49,55]. Specifically, Kamatra and Kartikaningdyah [49] and Li et al. [55] found that the social dimension positively relates to market performance, attributing this to the substantial investment attracted by companies that excel in social practices. Investors increasingly seek socially responsible companies to build their investment portfolios, aiming to meet their fiduciary responsibilities [21]. In the governance dimension, high scores are associated with a decrease in asset efficiency (ROA) and return on equity (ROE) for Western European companies. This finding is particularly interesting because this dimension pertains to corporate and management policies. The evidence suggests that European company management may be prioritizing the interests of board members, creating agency problems between board members and shareholders, and consequently decreasing company value [48]. Similar relationships were observed by Pham et al. [48] and Buallay [36] in North American, Chinese, and European companies. Conversely, Model V results indicate that a high score in the governance dimension enhances market performance, which aligns with expectations. Investors are attentive to quality management policies, and managerial decisions can significantly influence shareholders’ investment choices [21,55]. Regarding the GC score, because it has higher scores in its quartile categories, it would not be sensible to distinguish between the best and worst scores, as most of the worst scores are significantly above the neutral point. Tests revealed that distinguishing between good and bad scores was not beneficial for the GC score categories, as the estimated models did not show statistically significant variables. Therefore, it was preferable to understand the direct relationship each variable has with financial, market, and operational performance, with the results presented in

Table 7. Model VI regression results.

Dependent Variable	Model VI
	ROA	Tobin´s Q	ROE
(Intercept)	54.8000 (34.1714)	−1.7901 (124.5119)	116.0716 * (46.3439)
AC	0.1187 (0.1365)	−0.7691˙ (0.4623)	0.0023 (0.2049)
HR	−0.6389 ** (0.2063)	0.5214 (0.6655)	−0.9672 ** (0.3084)
ENV	−0.3379 * (0.1373)	0.2563 (0.4451)	−0.3116 (0.2062)
LR	0.2310 (0.2047)	−0.2090 (0.6635)	0.5545˙ (0.3075)
LogSize	−11.2944 *** (2.5974)	−36.9243 *** (8.4191)	−10.5772 ** (3.9013)
Age	1.0861 ** (0.3472)	5.0304 *** (1.1253)	0.5114 (0.5214)
Communications	40.3138 (28.7069)	198.9143 (126.2443)	13.8654 (27.6502)
Discretionary Consumer	14.4340 (22.0702)	103.0411 (104.4375)	0.1386 (15.1107)
Energy	48.5755˙ (27.4653)	194.8564 (119.9835)	19.7636 (26.9965)
Financial	31.4390 (19.8631)	62.9150 (91.0837)	20.9118 (16.3071)
Real Estate	71.1137 (46.8194)	324.1796 (206.0159)	22.1528 (45.0149)
Industry	14.6811 (20.1798)	101.8910 (94.7862)	2.8401 (14.5288)
Materials	16.6173 (21.2374)	101.9393 (100.4193)	5.9662 (14.6206)
Health	16.7443 (22.2534)	269.0575 * (106.2685)	14.6071 (14.1940)
Technology	43.9828 (29.3256)	319.4327 * (126.5060)	2.8600 (29.8715)
Utilities	59.4909 * (29.5927)	215.3915˙ (127.1578)	32.7002 (30.4600)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

.

From the results of Model VI, it is possible to determine which GC score dimensions affect ROA, Tobin’s Q, and ROE. The practices associated with human rights (HR) are the only ones that affect more than one dependent variable. Specifically, the HR variable is statistically significant at the 1% level when the dependent variables are ROA and ROE. The remaining independent variables uniquely influence each of the business performance variables. The anti-corruption (AC) and labor rights (LR) variables are statistically significant at the 10% level when analyzing market performance for the AC variable and financial performance for the LR variable. The environment variable (ENV) is statistically significant at the 5% level when the dependent variable is ROA. Regarding the additional control variables, only the variables LogSize, Age, and the Energy, Health, Technology, and Utilities sectors are shown to be relevant in the estimated regression equations.

Specifically, the following are observed:

• The variable LogSize, in the first two regression equations, is statistically significant at least at 0.1%, and in the last equation at 1%.
• The Age variable is statistically significant at least at the 1% level when the dependent variable is ROA and 0.1% when the dependent variable is Tobin’s Q.
• When the dependent variable is ROA and Tobin’s Q, the Utilities sector is statistically significant at least at the 5% and 10% level, respectively.
• The Health Care and Technology sectors are statistically significant at the 5% level when the dependent variable is Tobin’s Q. When the dependent variable is ROA, the Energy sector is statistically significant at least at the 10% level.

In terms of interpreting the results, we find that anti-corruption activities slightly decrease market performance. Specifically, an increase of one percentage point in the AC variable implies a decrease of 0.77 percentage points in Tobin’s Q. Human rights activities show an inverse relationship with operational and financial performance, meaning an increase of one percentage point in the HR variable leads to a decrease of about 0.64 percentage points in ROA and 0.97 percentage points in ROE. Environmental practices negatively affect financial performance, with an increase of one percentage point in the environment dimension of the GC score leading to a decrease of 0.34 percentage points. Finally, labor rights are the only variable that shows a positive relationship with business performance, as an increase of one percentage point in the LR variable implies an increase of about 0.55 percentage points in ROE. All additional control variables, except for LogSize, show a positive relationship with the business performance variables. Anti-corruption practices negatively influencing market performance might be justified by the notion that transparency reduces illicit benefits that could otherwise enhance business performance. Regarding environmental practices, the results of the ENV variable indicate a negative influence on corporate performance, consistent with the inverse relationship identified in Model V between the environment dimension of the ESG and the performance of European companies. Social responsibility (SR) practices that consider employees’ interests can positively affect their loyalty, motivation, productivity, and workplace satisfaction, subsequently benefiting firm performance [20,23,24]. By adopting an SR policy, companies can save resources intended to boost employee satisfaction and retention [19]. In this research, the LR variable, representing employees, encompasses factors like workplace health and safety, diversity, compensation, training and development, and job quality. The positive relationship between ROE and the LR variable aligns with previous findings, supporting the hypothesis that European companies improve financial performance by implementing SR policies that consider employee interests. Therefore, Hypothesis 10 (H10) is accepted. Conversely, an inverse relationship was found between human rights practices and business performance. Investments in human rights-oriented CSR policies may exceed their economic benefits, thereby harming business performance. The human rights dimension was not used to represent employees alone, as it also affects other stakeholders, including the community and consumers, through aspects like product quality and safety, access to products, and community relations. The literature review indicates that SR can indirectly affect company performance, mediated by other factors. Thus, the models estimated next verify how Size and Age can influence the impact of SR on firm performance. Other variables, such as Sector, were not considered, as only size and age present significant results in mediating the relationship between social responsibility and firm performance. Not all performance variables yield significant results, so only the impact on ROA and ROE is presented.

Table 8. Model VII and VIII regression results.

Dependent Variable	Model VII		Model VIII
	ROA	ROE	ROA	ROE
(Intercept)	17.1288 (35.7997)	29.9940 (48.6963)	11.5713 (33.3677)	72.5952 (46.9258)
ESG	0.1134 (0.2054)	0.7653 * (0.3082)
GC			0.2724 (0.2301)	0.2387 (0.3481)
Age	1.6215 *** (0.3618)	1.3050 * (0.5428)	1.7931 *** (0.3614)	1.2481 * (0.5467)
ESG*Age	−0.0093 *** (0.0023)	−0.0153 *** (0.0034)
GC*Age			−0.0147 *** (0.0026)	−0.0159 *** (0.0039)
LogSize	−12.1204 *** (2.6366)	−9.9545 * (3.9555)	−10.7674 *** (2.5636)	−10.0152 ** (3.8783)
Communications	39.4215 (28.7309)	9.4632 (26.8562)	30.1799 (25.3046)	2.0488 (26.2017)
Discretionary Consumer	15.0079 (22.0954)	0.4179 (13.9582)	12.0967 (18.5047)	−2.2029 (12.8129)
Energy	48.0322˙ (27.5013)	14.5708 (26.3043)	38.9267 (24.2939)	8.6268 (25.6583)
Financial	31.4784 (19.9129)	17.8390 (15.6306)	29.3689˙ (17.0042)	19.7159 (14.7501)
Real Estate	75.1030 (46.7856)	24.1063 (43.5379)	61.6716 (41.1450)	10.3748 (42.3945)
Industry	15.7211 (20.1956)	2.4477 (13.5348)	12.6058 (17.0034)	0.5175 (12.5744)
Materials	14.1959 (21.2641)	0.8450 (13.5300)	13.2376 (17.8155)	2.1825 (12.4263)
Health	15.1098 (22.2799)	12.8136 (12.8969)	14.9835 (18.5235)	13.7504 (11.5587)
Technology	41.9149 (29.3479)	−2.9915 (29.1629)	33.7954 (26.1184)	−9.1146 (28.5588)
Utilities	56.6500˙ (29.6437)	23.2091 (29.8287)	47.4489˙ (26.4257)	18.8603 (29.1986)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

presents the results of the models that use age as a mediator between corporate performance and ESG/GC score.

By analyzing Model VII, it can be seen that the ESG*Age variable is statistically significant at the 0.1% level, whether the dependent variable is ROA or ROE. In both regression equations of this model, the additional control variables LogSize and Age are statistically significant but at different levels. Specifically, when the dependent variable is ROA, LogSize and Age are statistically significant at least at the 0.1% level. Conversely, when the dependent variable is ROE, these variables increase their significance level to 5%. Notably, only when the dependent variable is ROA do the Energy and Utilities sector variables show a reasonable significance level (10%). Similarly, in Model VIII, the GC*Age variable is also statistically significant at the 0.1% level, indicating that the interaction between the GC score and Age influences the operational and financial performance of Western European companies. Regarding the significance of the additional control variables LogSize and Age, these variables behave identically to those in Model VII. Additionally, in the first equation of this model, the financial and utilities sectors are statistically significant at least at the 10% level. The results of Models VII and VIII contradict Sun’s [47] view that, as the sample companies age, the impact of SR on operational and financial performance diminishes. This is due to firms failing to keep up with SR trends over the years and not taking necessary actions, which negatively affects their performance. In the literature, Han and Kim [82] also identified a negative relationship between SR and business performance when mediated by age. It should also be noted that the statistically significant additional control variables show a positive relationship with ROA and ROE, except for LogSize.

Finally,

Table 9. Model IX and X regression results.

Dependent Variable	Model IX		Model X
	ROA	ROE	ROA	ROE
(Intercept)	378.6380 *** (51.7652)	497.8850 *** (73.4000)	415.8267 *** (49.6044)	514.2321 *** (70.1674)
ESG	−6.5303 *** (0.6980)	−7.8007 *** (1.0635)
GC			−7.2911 *** (0.6418)	−8.0221 *** (0.9909)
ESG*LogSize	0.6070 *** (0.0694)	0.7597 *** (0.1057)
GC*LogSize			0.6656 *** (0.0633)	0.7283 *** (0.0978)
LogSize	−48.6457 *** (4.7433)	−56.2225 *** (7.2273)	−51.7429 *** (4.5325)	−54.8357 *** (6.9978)
Age	1.4610 *** (0.3362)	0.9339˙ (0.5123)	1.5145 *** (0.3297)	0.9533˙ (0.5091)
Communications	59.7874˙ (35.4758)	37.0782 (32.3465)	57.2864 (36.4946)	31.5091 (32.4877)
Discretionary Consumer	23.6657 (29.0286)	12.1455 (21.5562)	21.9692 (30.1397)	8.5425 (21.9051)
Energy	63.0479˙ (33.7224)	35.9438 (31.1931)	65.2342˙ (34.6924)	37.2272 (31.3864)
Financial	45.9411˙ (25.4196)	38.4118˙ (20.7517)	34.9555 (26.2815)	25.7520 (20.9278)
Real Estate	100.1484˙ (57.8338)	57.4983 (52.5717)	106.6371˙ (59.5465)	59.3210 (52.9209)
Industry	23.0322 (26.3657)	12.4484 (20.0350)	24.1593 (27.3548)	13.0938 (20.3378)
Materials	21.6080 (27.9083)	11.2800 (20.7615)	25.0276 (28.9793)	15.0144 (21.1097)
Health	22.3720 (29.4940)	23.0089 (21.2246)	23.9703 (30.6536)	23.5205 (21.6058)
Technology	58.8328˙ (35.6352)	20.4343 (33.9523)	65.4083˙ (36.6017)	25.2757 (34.1272)
Utilities	77.3022 * (35.8422)	52.2795 (34.4544)	79.9815 * (36.7835)	54.2329 (34.5902)

Note: The values in parentheses correspond to the estimated standard error of each coefficient in the model. Significance Codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1.

presents the last two estimated models, each with two regression equations, which assess the impact of GC score on firms’ performance through size. In Model IX, the variable ESG*LogSize has a positive estimated impact on ROA and ROE and is statistically significant in both regression equations at the 0.1% significance level. Regarding additional control variables, in the first regression equation, all variables except for the Industry, Materials, and Health sectors are statistically significant. Specifically, Age and LogSize are significant at least at the 0.1% level, the Utilities sector at least at the 5% level, and the remaining sectors at least at the 10% level. Notably, only the LogSize variable shows a negative relationship with ROA and ROE, while the remaining additional control variables are positively related to the dependent variables. In the second regression equation, some additional control variables are no longer statistically significant, such as the Communications, Energy, Real Estate, Technology, and Utilities sectors.

In Model X, the results are very similar to the previous model, with the GC*LogSize variable also showing a positive relationship with operational and financial performance and maintaining the same significance level (0.1%). Compared to the previous model, the additional control variables in both regression equations retain the same statistical significance, except for the Financial sector, which is no longer statistically significant at a reasonable level. The relationship established by the statistically significant additional control variables with ROA and ROE remains the same as in Model IX: a positive relationship, except for the LogSize variable, which establishes a negative relationship.

Overall, we can conclude that regardless of the variable used to measure SR (ESG or GC score), the size of European companies, measured by the total number of assets, positively influences the impact of SR on operational and financial performance. This indicates that companies with more assets have more resources available to develop socially responsible practices. The results of Models IX and X align with the arguments of Kamatra and Kartikaningdyah [49] and Minutolo et al. [45], suggesting that the size of firms positively influences the relationship between their performance and SR. This contradicts the view of Rossi et al. [29], who argued for an inverse relationship.

5. Conclusions

Over the last decades, social responsibility has taken on a significant role in society, with many companies committing to becoming more socially and environmentally conscious. However, this commitment incurs costs, which can be high for some organizations and may negatively affect business performance. The literature presents studies supporting this perspective, but there are others that show a positive relationship between social responsibility and business performance, primarily mediated by its positive impact on stakeholders such as consumers, employees, and investors. In fact, there is no consensus on the relationship between social responsibility and corporate performance, making it crucial to seek new findings. Therefore, this research analyzed the impact of social responsibility on the performance of 216 listed companies in Western Europe from 2017 to 2021. To evaluate corporate performance, the ROA ratio (operating performance), Tobin’s Q (market performance), and ROE (financial performance) were used. Social responsibility was quantified using the ESG and GC scores. This impact was analyzed by estimating ten models for panel data, where six of them include three regression equations and four include only two. Initially, the intention was to understand the relationship between SR and the ESG and GC score indices. The results indicate that this relationship cannot be observed in a linear fashion. Depending on the type of performance being analyzed, SR can take on different meanings. Specifically, when SR is measured by ESG scores, it shows a negative relationship with operational performance and a positive one with market performance. No relationship was found with financial performance. When GC score is used to measure SR, a negative relationship with both operational and financial performance was identified. However, no relationship was found with market performance. Importantly, even when a distinction is made between the best and worst overall ESG and GC scores (where the best scores are higher than the third quartile, and the worst scores are lower than the first quartile), the results only confirm that obtaining high scores on these two indices hurts both operational and financial performance. High ESG scores only benefit companies in terms of positively influencing market performance. Subsequently, the influence of the ESG and GC score dimensions on business performance was investigated. For the ESG score dimensions, a distinction was made between the best and worst scores (adopting the same process used for the overall scores). For the GC score dimensions, this process was not adopted due to the lack of significant results, and thus the direct relationship was to analyzed. Regarding the ESG score, the findings show that high scores in the environmental dimension hurt the operational and financial performance of European companies. The social dimension also negatively impacts operational and financial performance, with high scores leading to decreases in ROA and ROE, while lower scores lead to increases in these financial ratios. However, the relationship with market performance is different, as high scores in the social dimension positively influence Tobin’s Q. Finally, high scores in the governance dimension show a negative relationship with operational and financial performance but a positive relationship with market performance. For the GC score, it is evident that practices associated with anti-corruption negatively influence market performance. Practices associated with human rights also show a negative relationship, but with operational and financial performance. In line with the environmental dimension of the ESG, the environmental category of the GC score also presents a negative relationship with operational performance. Notably, practices related to labor laws positively influence the financial performance of European companies. Finally, we sought to determine the relationship between the additional control variables—age and size—with the ESG and the GC scores. Due to a lack of relevant results, this analysis was not extended to sectors of activity and focused solely on operational and financial performance, excluding market performance. The results show that as firms age, SR, when measured by ESG and GC scores, negatively impacts financial and operational performance. This finding provides evidence that older companies are not keeping up with trends in social responsibility, which has negative effects on business performance. On the other hand, the results of this research indicate that the variable “size” positively influences the relationship that SR establishes with operational and financial performance. Specifically, since the size of European companies is measured by total assets, it is possible to conclude that companies with more assets have more resources available for developing socially responsible practices.

The findings of this study make a significant contribution to the literature on this topic. This research focuses on the Western European market, including countries that have been little addressed in this context, such as Portugal. This study innovatively uses two metrics to assess social responsibility—the ESG and the GC scores—providing a more complete and comprehensive analysis of the impact of socially responsible practices on corporate performance. Furthermore, it evaluates the performance of companies in three areas: operational, financial, and market performance. These findings also help identify how the dimensions of ESG scores and GC score categories relate to the business performance of the sampled companies. This is crucial for managers of European companies, as it helps them determine which SR activities are most beneficial for operational, financial, or market results. Additionally, this study is one of the few that examines how the control variables of age and size influence the relationship between the ESG, GC, and business performance.

The limitations of this research are primarily related to the fact that the sample is not representative of the European market, as it mainly consists of small and medium-sized companies. Moreover, not all listed companies could be considered, as several were not listed in the ESG Book database and the Finbox database. This study included only those companies assigned ESG and GC scores, which does not imply that the excluded companies are not socially responsible. In the future, it would be interesting to conduct studies addressing these limitations. For example, research could explore the impact of social responsibility on the performance of European small and medium-sized enterprises, identifying the SR activities that effectively improve business results. These findings would be valuable for SME managers, who often operate with limited budgets, making any investment with a medium- or long-term financial return or an uncertain immediate return a considerable risk.