Keywords

1 Introduction

Sustainability equality considers civilizational advance in economic, environmental, and social foundations (Elkington, 2018): an integrative model for sustainability (Hahn et al., 2015) where no dimension is prioritized (de Oliveira Neto et al., 2018). In the social dimension, sustainability concerns people and society in organizations, whether internal or external. The internal dimension addresses labor relations, training, compensation, and benefits, among other traditional issues of Human Resources Management (HRM), as well as inequalities issues like diversity, gender, race, religion, and age (Souza, 2011).

Diversity debate gains momentum in the globally connected workforce spread around different countries (Urbancová, Hudáková & Fajčíková, 2020). The rise of “anti-pluralism” politics alongside national-oriented populism (Horak et al., 2019), world conflicts, refugee crisis, extremism, and religious diversity are factors that contribute to the complexity of diversity management (Cumming et al., 2020). These matters are present in public debate that ranges from policy discussion (Ortlieb et al., 2019), step into organization dimensions (Cavico & Mujtaba, 2017), and shape the perceptions of individuals: as an employee (Buttner & Lowe, 2017), student (Ghavami & Mistry, 2019), or community members in general (Walton, 2018). In the corporate milieu, significant relevance of the subject due to evidence showing that diversity fosters creativity and innovation (Tuan, 2020) is linked to competitive advantage (Zikic, 2015; Ali, & Konrad, 2017). The Green Human Resources Management movement signals, among other issues, the relevance equity quest gains in organizations (Ehnert et al., 2014). The debate is bridged by the overlapping of strategic HRM and corporate social responsibility (CSR) in perspective to achieve competitive firm advantage and expanded to the bigger picture of business ethics (Järlström et al., 2018).

The strategic gains for the company arise when individuals express their different views and ways of developing processes, achieving goals, and creating work teams, helping companies succeed and improve, fostering innovation and change (Torres & Pérez-Nebra, 2004). The collateral gain comes when companies invest in nondiscriminatory practices, standing out of the crowd of their peers that do not have a socially responsible approach (Wailes & Michelson, 2008). Management environments with a natural level of strategic complexity, like in a hybrid (Lee et al., 2019) or an ambidextrous organization (García-Granero et al., 2018), can benefit from diversity management.

As society becomes heterogeneous, organizations seek to survive in an increasingly competitive world by addressing complex cultural hybridity that characterizes organizations, and a diverse workforce is one way to interface complexity, in and out of the organizations (Pereira & Hanashiro, 2010). Nevertheless, if the quest for diversity is carried out unchecked, companies may experience unwanted inequality and inclusion backlashes. Access is not inclusion: the company may even have half of its workforce of women, but they may receive lower wages, be the least represented in leadership positions, and experience higher turnover. Báez et al. (2018) identified significant gender gaps regarding women in dimensions like presence, salary, and seniority at tourism companies. The same frame of “present yet underrepresented” applies to individuals with special needs, different generations, ethnic groups, and so on (Ylöstalo, 2016).

Thus, organizations need to be aware of these issues and disclose how they manage diversity without putting inequality aside. This can be done through sustainability reports, aligned with frameworks, guidelines, metrics, and procedural justice (Kundu et al., 2019), to set together policy, practices, and signal commitment exposing firms to public scrutiny, therefore influencing deeper and surface levels of diversity in a decoupling-proof way.

The Global Reporting Initiative (GRI) is widely used, and it is the best-known sustainability reporting model (Alaraji & Aljuhishi, 2020). However, adherence to the GRI guidelines in some regions, such as Latin America, is still tiny. Brazil is the third country in the GRI global ranking, just after the United States and Spain (GRI, 2020), and plays a vital role in the economic and political context globally; at the same time, it is an emerging economy with high levels of unemployment, poverty, and social exclusion.

The organizational policies for diversity are still unclear in the Brazilian management literature (Saraiva & Irigaray, 2009); therefore, it is critical to understand how organizational practices are being developed toward diversity and equality. Building upon the diversity and equality, signaling and legitimacy theory, considering the literature gap and the relevance of the GRI for Brazilian context, this study comes up with the research question: “Are diversity indicators of GRI related to diversity leadership in Brazilian companies?”

We explore linkages between GRI indicators for diversity in Brazilian companies with the presence of women, indigenous, elderly, young, and Afro-descendant employees. In the GRI report, diversity policy indicator is composed of the wage ratio between genders, cases of discrimination, rape, harassment, age and gender turnover, gender composition, and other minorities. These variables together compose the leadership policy indicator. We assume companies that report GRI will present a superior performance in the leadership policy indicators. However, we want to know to what extent this influences its natural leadership diversity and how it could support our assumption that leaders act as a proxy for DEM in an SHRM approach through its potential of consistent signaling and legitimacy.

1.1 Guaranteeing the Golden Ticket Is Not Enough

The recognition of cultural diversity has been the object of numerous research. In the 1990s, diversity was crafted in the World Culture and Development Commission report, and the document denominated “Our creative diversity,” which corollary translates the requirement of the virtue of tolerance. In 2001, UNESCO considered that respect for cultural diversity is a right and an essential element for policies intended to promote dialogue among peoples (Rodrigues & Abramowicz, 2013).

Diversity is configured as the heterogeneity of race, ethnicity, gender, sexual orientation, socioeconomic status, age, physical abilities, religious beliefs, political beliefs, and other ideologies. It means understanding each individual as unique and recognizing individual differences, focusing on acceptance and respect, going beyond tolerance, and celebrating its rich dimensions (Patrick & Kumar, 2012). The way diversity and equality are framed in business gradually evolved from legal compliance and affirmative actions (Mor Barak et al., 2016) to a nonexclusive view for competitive advantage (Richard, Roh & Pieper, 2013). Workforce diversity contributes to organizational performance, net value, customer relations, innovation, creativity, and problem-solving and generates a positive image (Mor Barak et al., 2016; Ruiz-Jiménez et al., 2016).

Cultural diversity implicates respecting choices, work behavior, and the development of careers (Byars-Winston et al., 2015). Heitner et al. (2013) and Puente-Palacios et al. (2008) highlight the importance of leaders valuing diversity and recognizing abilities of workers so that individuals see themselves as equally capable. Most of the initiatives related to the success of DEM concern leadership.

Conversely, failures in diversity management lead to conflict and dysfunctional behaviors, causing severe consequences for the organization (Shore et al., 2009). Managers must establish diversity programs considering the communities’ cultural values and practices.

Diversity can be categorized into two domains: (I) the surface (or visible) level, which comprises aspects like gender, age, ethnicity, and race, and the (II) deep level, which includes the less visible ones like education, motivation, job tenure, and so on (Mor Barak et al., 2016). The simplistic approach of the diversity in companies happens due to poor conceptions that shape itself “empty vessels programs” (Ahmed, 2012) where different groups and expectations are thrown in without regarding historical or cultural aspects and power relations that influence organizational structure (Dobusch, 2017).

While diversity usually is approached in numbers, equality is addressed by its impact. It is the difference between giving golden tickets in “a statistical fashion” or truly safeguarding all hues of the diverse workforce through the whole chocolate factory. Several companies have been searching to implement inclusion business cases in their strategy, organizational process, and CSR activities (Bilimoria et al., 2008). Several scholars agree that increasing diversity is not sufficient as an HMR strategy: companies need to create policies and practices that manage diversity and promote an inclusive environment (Mor Barak et al., 2016). The workforce should feel safe to bring their authentic and unassimilated self to the workplace.

The diversity and equality management (DEM) policies and practices appear to be a proper way to address diversity. Alike what happens with HRM, research on DEM supports that to be successful, and diversity and equality practices should integrate the core of a firm’s human resources management.

From the SHRM theory, DEM relies on the assumption that a diverse workforce can improve company performance when their strategic needs are supported by diversity (Ali & Konrad, 2017). Pasztor (2019) highlights the relevance of communicating diversity in three main ways: (I) asset fostered by HRM and corporate values, (II) excellence and competitive advantage, and (III) structural mechanism supported by several initiatives like mentoring, networking, diversity training, and governance. Firms have constantly been trying to include as many stakeholders as they can to address their social contract (Carroll & Shabana, 2010) and therefore do the “right thing” (Mor Barak et al., 2016).

1.2 Consistent Signaling Diversity and Equity Through Leadership

Leaders are seen as ethical stewards of organizations. Their status is grounded in trustworthiness: good leaders honor their personal and organizational values in a way that impacts employees and society. Nevertheless, gaps between leaders and followers are related in research, and it resonates in commitment issues, wealth creation, and increased transaction costs (Caldwell, & Hayes, 2010). Companies that embed diversity in their leadership compositions communicate a consistent message, both to internal followers and external stakeholders. To see your gender and age group represented in non-predefined levels of decision-making sends a solid message to the workforce that can be converted into a commitment to the company. Through the consumer’s eyes, a public statement from an employee of your ethnic group signals that the company respects and considers you in their consumer base: “by then, for then,” is the kind of perception harder to walk by but also harder to decouple.

Leadership is a broad topic that overlaps many organizational theories. Therefore, this is carried out within the DEM approach: relations among leadership, legitimacy, and diversity have been identified in many studies (Bhattacharya et al., 2008). Legitimacy concerns are nuclear for companies since it is one of the main motivations for diversity policy implementations. Defined by Suchman (1995) and further explored by Singh and Point (2009), legitimacy can be categorized on pragmatic, moral, and cognitive dimensions: “Pragmatic legitimacy is related to the self-interest of the actors and audience, while moral and cognitive legitimacy implies the interest of the wider organization and society” (Singh & Point, 2009, p. 25).

The modern company can tackle complex challenges: absorb them in a resilient way and keep it going through times where social clashes torn apart reputations in a matter of hours. In the overwhelming world of communications, the modern organization is the one who does the right thing and is competent enough to tell the history. Usually, it is hard to make the “good deeds” reach stakeholders; information is lost among the channels and decoupled between parties. Signaling theory connects DEM with organizational performance through the way the firm communicates its commitment with values like fair treatment and inclusion to a broader spectrum of social groups: this generates interface feedback with stakeholders who value diversity and social justice and, in turn, could support firm with investments, legitimacy, and new human resources (Ali & Konrad, 2017).

Receivers read signals differently, calibrating them according to their assumptions and values and changing the strength and meaning of the messages (Suazo et al., 2009; Connelly et al., 2011). Calibration can also respond to environmental distortions: receivers experience issues in observing signals or even feel deceived by misleading or ambiguous signals. On the other end, signalers may feel tempted to false signaling when they do not have quality. The signal costs outweigh the costs of achieving the desired aspect and emitting valid signals (Connelly et al., 2011).

With proper signaling, diversity fosters a positive firm image even in controversial industry environments (Du & Vieira, 2012); nevertheless, legitimacy is a complex and multifaceted construct: addressing it without proper concern of information asymmetry gaps can lead to decoupling (Singh & Point, 2009) between what is shown in the firm website and the reports and what the stakeholders observe, through, i.e., peer relations. Organizational legitimacy is defined by the combination of technical considerations of the industry and institutional pressures of the environment. Thus the legitimacy pattern is defined by the internal strategic allocation of resources to “get the job done” but also by a necessity to comply with stakeholders by “doing the right thing” (Suchman, 1995).

2 Method

The methods of the study comprise data panel analysis to emphasize variable linkages over time. Carried out by STATA software, the approach is a multivariate statistical technique popular in social sciences. Data are analyzed over a time series. From an epistemological perspective, the researcher tries to understand how a ratio varied over time. This suggests using cross-sectional or longitudinal data analysis, with variations between 5 years in the GRI reports from the Brazilian context.

The hypotheses of the study show that the composition of leadership in companies can be influenced by contextual conditions (Kemp et al., 2015; Chawla & Sharma, 2016) such as proportion of wages between men and women, number of reported cases of discrimination, and cases law violation, in addition to the turnover of employees varying by age and gender. Therefore, it is necessary to suggest relations on the theoretical findings to fill the gaps left by the literature.

The variable components of diversity are the proportion of wages between gender, the number of cases of violation, the number of reported cases of discrimination, and the turnover of men, women, and age group. Also, in panel analysis, we consider time perspective as an independent variable to measure time variability (longitudinal effect) to compose the eight hypotheses of this study (Fig. 8.1).

Fig. 8.1
A text document. It represents hypotheses H 1 through H 8. H 1 reads, diversity indicator influences the gender ratio in organizational leadership.

Hypothesis

Full size image

The sample consisted of 305 companies (N) with 5 years of observation: results in 1525 inputs or sample observations in the function of time and variables.

3 Results and Discussion

3.1 Descriptive Data Analysis

The first command to be used in Stata is the “xtset” + “variable name.” The ideal variable presents the result as “strongly balanced,” i.e., the variable is perfect for use in a panel. Unfortunately, only the Business Year (ordinal), X2, and X3 variables were considered by Stata as strongly balanced. This command was applied to all variables, whether y or x, and all presented as “balanced,” i.e., can work but have “missing values” or values ​​not reported by the companies.

The Stata command “xset” indicates that the database is a balanced panel analysis by year variable. For each company, there is no shortage of data for the corresponding years, enabling the submission of data to panel analysis. The command “xtsum ya x10 x11 x12 x13 x14 x2 x3” is possible to describe the sample data. The analysis shows that N (sample size) was 305 companies with 1523 observations, the average of the variables remained within the expected range, and the standard deviation was considered heterogeneous. While variables such as X3 obtained a high standard deviation, others had a standard deviation of zero. This means that few companies reported any case of code of ethics violation on this sample, and none of them reported having experienced cases of discrimination in legal proceedings such as harassment and/or sexual or moral violence, obtaining no dispersion.

3.2 Fixed and Random Effects on Panel Analysis

The Hausman test is used to decide whether fixed or random effects will make the analysis. The test is required to verify the influence of time on the relationship between the variables. Even if the intercept is different between the companies, they do not vary with time; therefore, time is irrelevant. In the case of validation of this suggestion, panel analysis is of fixed effects between the data. However, in the random effects, the average intercept of companies can even be equal, but they vary over the analysis period. In the case of a statistically significant p-value (below 5%), the result of the Hausman test determines the use of fixed effects. In case of no significant p-value, time determines the variation in the data and is considered a random effect of the sample.

From Table 8.1, it is possible to observe that the probability of test f is much lower than 5%, which demonstrates the robustness of the model adopted. However, Table 8.1 shows the variables X10 and X12 are statistically significant (p-value less than 5%) and demonstrate that they are the biggest influences on ya (women in leadership). The variables X11, X14, and especially X3 do not represent influence on ya. Note that the variable x2 has been omitted from the test because its number was so low that it did not have any value to be analyzed. Only women’s turnover (X10) and the turnover of up to 30 years of age (X12) influence the female leadership composition in the analyzed companies.

Table 8.1 Random effects
Full size table

To progress the Hausman test, it is necessary to compare the relationship between the fixed effects, and the command used is “Statistics-Postestimation-ManageEstimationResults-StoreinMemory.” Thus, Stata retains in its memory the test results as “fixed.” The probability is also very high in random effects, and the variables that demonstrate statistical significance in influencing ya variation are X10 and X12.

The most suitable model is presented by the command “Specific Hausman,” and the results show that Prob>chi2 = 0.4987. The probability is very high, greater than 5%, rejecting the fixed effects model and accepting the random model. Thus, it is possible to understand that time is significant in the variation of the relationship between the variables found.

3.3 Hypothesis Results

H1 assumes that the diversity of variables influences the composition of women in the leadership of organizations that reported GRI within the period from 2009 to 2013. All are positively correlated, accepting hypothesis 1. The dependent variables show influence starting from X10 in 18.99% and X11 in 12.76%, and the one that does not correlate is X3, and all others have low correlation. Due to the results of the relationship presented, X2 was eliminated from the sample. There are very few instances through the sample, which impedes its consideration in the study and is missing.

H1 assumes that the diversity of variables influences the composition of women in the leadership of organizations that reported GRI within the period from 2009 to 2013. All are positively correlated, accepting hypothesis 1. The dependent variables show influence starting from X10 in 18.99% and X11 in 12, 76%, and the one that does not correlate is X3, and all others have low correlation. Due to the results of the relationship presented, X2 was eliminated from the sample. There are very few instances through the sample, which impedes its consideration in the study and is missing.

Hypothesis 3 is given by the composition of the diversity of variable composition of employees from 30 to 50 years of age in the leadership of organizations that reported GRI within the period from 2009 to 2013. In this case, X10 has the highest correlation of 15%, and X3 the least, being insignificant. The variable X2 was again eliminated. X11 has a reasonable improvement in the influence in which it is concluded that hypothesis 3 can be partially accepted. Thus, as shown in the analyzed theory, the diversity of leadership between 30 and 50 years of age can be influenced by the model factors and other factors that may be latent to the model. Thus, it is believed that the reasons that lead companies to structure the leadership of this age group are not the policies or actions of diversity implemented by them, being partially accepted.

The data show a high correlation from the independent variables X10 (89.8%), X11 (0.75%), X12 (85.5%), and so on. X10 and X12 represent the highest correlation. In this case, hypothesis 4 is accepted and can be said that much of the relationship that makes up the leadership of the over 50 years of age group is influenced by male turnover of employees and influenced by a growth of the value from 2009 to 2013.

Hypothesis H5 is accepted, showing more excellent distribution between the dependent variables than the variable components of the leadership of Afro-descendant or mixed-race employees in the leadership of organizations that reported GRI from 2009 to 2013. This shows that there is evidence that the variables of the composition of diversity do not affect this dimension over time.

Hypothesis 6 is supported by the decreasing values ​​from X10, i.e., it can be attributed to the variables of the composition of diversity, and their most significant influence also occurs by variable X10, showing the evolution of the indicator over time. It is possible to affirm that the variables of the composition of diversity positively influence the leadership composition concerning factors.

Hypotheses 7 and 8 were also supported, with values from 0.15 (15%) to 0.20 (20%) in X10 homogeneously distributed to X3. Again, variable X2 had no influence. This shows the leadership composition of employees and people with special needs in the leadership of organizations that reported GRI from 2009 to 2013. However, since no growth is revealed (concentration of influence in X10) in this period, it is possible to say that the diversity of leadership composition aspects relating to these assumptions remained the same.

4 Conclusions

They are saying that diversity and inclusion foster innovation and booster competitivity is quite commonplace and say that firms still struggle to comply with non-decoupled organizational diversity and equality management (Cho et al., 2017). Organizations need to be aware of the complexity to foster a heterogeneous workforce so that their quest for diversity does not end up in an announced inequality in the workplace and a decoupled rhetoric in their reports. When it comes to DEM, the real accountability is needed in the same fashion, and it is applied to sales, revenue, and other financial performance metrics.

The model used here is one more option to foster accountability in DEM. It showed that hypotheses 1, 4, 5, 6, 7, and 8 were accepted, which lead to the conclusion that investing in structural aspects of diversity composition is a vector of fostering representation in the leadership of women (hypothesis 1), over 50 years of age (hypothesis 4), Afro-descendant or mixed-race (hypothesis 5), Asian and Oriental (hypothesis 6), indigenous (hypothesis 7), and people with special needs (hypothesis 8). In a time-frame perspective, the panel analysis showed that only hypotheses 3 and 6 (30 to 50 years old and the composition of Asian and Oriental) had an evolution. This result is not necessarily related to the diversity policies but to the belief that the employee profile will change due to the qualification increase.

The influence of diversity components in employee leadership of over 50 years of age (hypothesis 3) proved to be partially accepted. Therefore diversity policies have little influence in determining the presence of this age group in the company. Combining this with the rejection of hypothesis 2 (leadership in the 30–50 years of age group), we conclude that age range is one point that diversity policies have less influence on. Conversely, ethnic and gender characterization showed more consistent results in response to the variable components of diversity proposed by the GRI.

In practical terms, organizations can leverage these results to transform the current diversity policy into a more effective policy toward leadership. The innocuousness of some policies is known, but regarding ethnic and gender issues, companies must continue investing in isonomic wage policies in the organization of committees to debate violation of code ethics and are evident in disclosing cases of discrimination.

4.1 Implications

The business case demanded nowadays presumes a good thing if your workforce looks way more similar to your customer base. In practical terms, organizations can leverage these results to transform the current diversity policy into a more effective policy toward leadership. Dobusch (2017) points out that most managerial problems are due to motivation rather than structural problems, which lead diversity to focus on the individual level and only surface touch the organizational structure. The application of metrics can fetch empirical evidence and reinforce models on both surface and deep-level diversity aspects. Practical implications rely on evidence-based evidence to guide managerial decisions to improve organizational performance and make workplace experience better for employees (Mor Barak et al., 2016).

The innocuousness of some policies is known, but concerning ethnic and gender issues, it is essential that companies continue investing in isonomic wage policies, debate committees of cases of violation of the code of ethics, and are transparent in disclosing cases of discrimination, which reinforces the studies of Ortlieb et al. (2019); it is apparent that companies or at least a minority report all cases of legal problems concerning harassment of any nature.

The limitations of the work circled the issue of missing values that had to be extracted in large quantities because the companies are not required to report all indicators every year, which meant that the accuracy of the data loses strength.

As search suggestions, this work addresses the same question found by Saraiva and Irigaray (2009) “To what extent indeed do organizations prioritize and actualize their diversity policies?” Further studies can be drawn from the GRI database and the integrated report focusing on impunity of discriminatory conduct, the effectiveness of implementing codes of ethics, and the presence of social responsibility policies.