Embedding Diversity in Sustainability Reporting

Abstract

The relevance of diversity has been recognised by academics and researchers as well as decision-makers. Diversity reporting can be perceived as the first step in addressing inequalities in organisations and potential assistance for the diversity agenda, because it allows measuring diversity and ultimately managing it. However, the recognition of the importance of diversity and diversity reporting does not necessarily contribute to a greater inclusion of diversity into sustainability reporting. The following paper attempts to determine the scope of diversity reporting, the specificity of the collected and disclosed diversity data, as well as the determinants of diversity reporting. For this purpose, a CATI (computer-assisted telephone interview) research was conducted, involving companies indexed on the Warsaw Stock Exchange. The results were analysed using the Cramer’s V contingency measure, the Kruskal–Wallis H test and ordinal regression. The results show a substantial difference in the collection of diversity information between organisations that map and that do not map their stakeholders. Furthermore, they show that, when organisations collect diversity data, their specificity is rather high, however this does not translate into an equally high level of diversity disclosure. Furthermore, the paper analyses the possible determinants of diversity disclosure, which do not necessarily overlap with the determinants of sustainability reporting.

1. Introduction

The relevance of diversity has been recognised by academics and researchers as well as decision-makers [1]. Although the research shows that diversity can have many positive outcomes for an organisation if it is managed correctly [2,3,4], previous studies show that diversity management [5] and diversity reporting [6] do not constitute very popular or widespread approaches in Polish organisations. Although increased attention is being paid towards the environmental and social impact that organisations have on their stakeholders and that increasing numbers of organisations are attempting to implement ethical, social and environmental aspects into their reporting [7], through the “triple bottom line” [8] or sustainability reporting [9], the issue of diversity is still being left out.

We can expect this to change due to the adoption of the European Parliament’s and of the Council’s Directive 2014/95/EU of 22 October 2014 amending Directive 2013/34/EU on the disclosure of non-financial and diversity information by certain large undertakings and groups, which obliges organisations embraced by its regulations to report diversity information as of 1 January 2017 [10]. The Directive specifies entities obliged to disclose non-financial and diversity information as large public interest organisations, which in the case of Poland were specifically listed in the Act of 7 May 2009 (uniform text; Polish Journal of Laws; Dz. U. of 2015, item 1011). Additionally, the Directive specifies the requirements, including the number of employees, balance sheet total and net turnover. Based on the assessment of the Polish Ministry of Finances, it was estimated that approximately 6000 large entities within the EU, including 483 Polish organisations, are affected by the new regulations. As it would seem to be very difficult to find organisations and entities beyond the influence of public interest entities, the importance of analysing particularly those organisations is very high.

Directive 2014/95/EU describes the required diversity information in a very general manner. According to Article 20, organisations will be obliged to report the following: “a description of the diversity policy applied to the undertaking’s administrative, management and supervisory bodies with regard to aspects such as, for instance, age, gender or educational and professional backgrounds, the diversity policy’s objectives, how it was implemented and the results in the reporting period. If no such policy is applied, the statement shall contain an explanation as to why this is the case” [10].

Although the Directive has signalled its understanding of diversity, the catalogue of the presented diversity dimensions has been left open. Diversity is a very complex, multi-layered and ambiguous term, which in consequence can lead to very different approaches towards diversity management and diversity reporting. Even though diversity management, as a concept and term, has been used since the 1980s, it does not have a single clear and unambiguous definition [11]. The multiplicity of diversity management’s definitions is largely a derivative of the ambiguity of diversity itself. Diversity is defined, inter alia, in terms of bio-demographic and job-related diversity [12] or, as pointed out by Christian et al. [13], is classified in taxonomies developed based on various dimensions, such as the taxonomy based on the degree of how easy the differences can be observed, taxonomy based on the degree to which the dimension is related to the job, the degree to which it is related to the task or to the relations, and taxonomy based on the division of diversity dimensions into inherent and role-related, among others. Furthermore, the taxonomy based on the so-called deep-level and surface-level dimensions was developed. These distinctions are important due to many practical implications. As suggested by Mor Barak [14], using the distinction between visible and invisible diversity can, inter alia, prove that discrimination based on the visible dimensions of diversity is more common. Gardenswartz and Rowe [15] distinguished four layers of diversity: personality; the internal dimension with age, gender, race/ethnic origin, sexual orientation and physical ability; the external dimension with religion, marital and parental status, income, educational background, work experience, among others; and the organisational dimension, including work content and field, seniority, work location, management status, etc. Apart from the above taxonomies, diversity is used to simply describe individual personal characteristics, such as gender, age, race/ethnic origin, disability and sexual orientation [11]. Organisations very often choose one or several of these dimensions without implementing the holistic concept of diversity. Furthermore, according to Harrison and Klein [16], diversity is not one, but three things: separation, variety and disparity. In the first aspect, diversity is understood as differences in peoples’ opinions, mainly due to their values, beliefs or attitudes. In this context, greater diversity is believed to reduce cohesiveness, generate a greater level of interpersonal conflict, further distrust and decrease task performance. The second definition of diversity, i.e., variety, refers to differences in knowledge and experience of people and predictably leads to increased creativity and consequently innovation, higher quality and accuracy of decisions and higher flexibility. The third definition of diversity, conceptualised as disparity, emphasises differences in proportion of valued assets and resources among people, and leads to more competition, reduced input, withdrawal. Adopting one of these meanings by researchers and practitioners can lead to different results and influence the perceived necessity of diversity reporting.

As argued by liberal feminists, gender, and thus also diversity reporting, constitutes the factor for addressing inequalities in the corporate sector [17], and therefore can be perceived as potential aid in the diversity agenda, as it allows for measuring diversity and ultimately managing it [18]. In this sense, reports are not merely passive descriptions of reality, as they rather enable it. By showing discrimination, they present accountability of a full range of stakeholders [19]. However, it should be noted that there is no guarantee whether reported diversity or discrimination will also be acted upon [19]. However, after recognising its importance, it was decided to determine the scope of diversity reporting in Polish organisations and to attempt to find the determinants of diversity reporting. Establishing the determinants of diversity reporting may be of importance for the processes of implementing it in organisations, as it can show possible factors that facilitate or hinder the implementation of diversity management.

The paper is organised as follows. The next section features a literature review with a short overview of the intersections between diversity management and corporate social responsibility (CSR), with emphasis on disclosure in both areas. Next, the research design and methodology are described. Finally, the research results are presented and followed by a discussion and conclusions.

2. Diversity Management and Corporate Social Responsibility

The connection between diversity management and CSR has already been recognised by researchers [20,21,22]. In some cases, the relation between those two strategies was not even discussed and their connection and interdependence were treated as something obvious [23]. Jamali and Dirani [24] pointed to the fact that both strategies address internal stakeholders as the target group. Although there are some tendencies to legislate the issues of CSR [25,26], the implementation of both strategies is still optional and goes beyond legal obligations [27]. Both strategies address the issues of human rights and inclusiveness [24,28]. The protection of human rights is also an important part of CSR reporting and accountability standards, such as ISO 26000 or the Global Reporting Initiative, among others, which also emphasise the need for eliminating discrimination based on gender, race, sexual orientation, disability, etc., which is the basic foundation for a successful diversity management strategy, and which already suggest a window of opportunity for embedding diversity into sustainability reporting. Finally, as Jamali and Dirani [24] emphasised, a common denominator for both strategies is the so-called business case—the logic used to argue the need and advantages arising from implementing both strategies, which include increased competitiveness, flexibility, advantages related to marketing, innovation and creativity or simply a better understanding of the customers [27,29]. Furthermore, CSR is perceived as a driver for diversity initiatives in organisations [30]. Diversity management and corporate social responsibility initiatives are very often implemented alongside one another. In such cases, CSR serves as an umbrella for diversity initiatives [31]. Due to the strong correlation between and many intersections of both strategies, it was decided to attempt to verify whether the determinants for CSR reporting also apply to diversity reporting. This is especially important, as there are other possible implementations of diversity management. Due to its nature [14], which can be described as horizontal, and the fact that it should be connected with a company’s business strategy, diversity management also has strong connections to other organisational strategies, such as human resources management, [32,33,34], which is sometimes strictly perceived as its constituent [35].

Social Reporting and Diversity Reporting

Corporate Social Reporting is perceived as an answer to stakeholders’ expectations and demands [36]. Previous analysis shows that employee-related information [37,38,39,40] is more widespread than other areas of social reporting. According to Vuontisjarvi [41], stakeholders are expected to show interest in information concerning the employment of minorities or women, equal opportunities, work–life balance or the integration of disadvantaged groups, among others, which constitute diversity topics.

Correlating diversity with corporate social responsibility in terms of inter-organisational implementation, but also in terms of reporting, is a double-edged sword. On the one hand, due to the greater universality and popularity of CSR, diversity issues can gain in importance. Several CSR reporting guidelines, including GRI, among others, in fact include sections devoted to diversity. Thus, reporting in accordance to these standards can promote diversity and diversity reporting. On the other hand, such a subordination can create the perception of diversity as simply an addendum, and consequently minimise its importance [42]. Furthermore, as suggested by Labucay [43], significant qualitative differences between CSR and Diversity Management, which influence reporting in those areas, have to be considered, namely the fact that, while CSR reporting can be based on measurable deliverables, diversity is a more complex concept, which suggests a necessity of taking more behavioural aspects into consideration.

The issue of diversity reporting and its scope, to the authors’ best knowledge, has not yet been subject to a deeper analysis. Even analyses referring to Directive 2014/95/EU disregard the issue of diversity, concentrating on labour practices and human rights [44]. Past analyses, focusing on or simply including diversity [42], due to the limited analysed samples, should be treated rather as preliminary studies. However, analyses focusing on diversity reporting as well as analyses of social reporting, with special consideration of diversity issues, demonstrate that dedicated diversity reporting [42] is being practiced by single organisations, and within social reporting, diversity is very often overlooked or in best cases handled superficially [6,45,46,47,48]. These papers were limited to analysing secondary data, i.e., the content of corporate websites, corporate social reports and corporate responsibility reports. This results in the limitation of the presented results, as they do not include primary data and, as result, could have omitted organisations undertaking diversity reporting without informing external stakeholders about the fact, which is the case, among others, in some Polish organisations, i.e., signatories of the Polish Diversity Charter, obliged to report on diversity issues. Most of them do not however inform external stakeholders about the results [49]. Furthermore, as pointed out by Holton [42], the content of and framework for diversity reporting differ from one another, thus making it difficult to conduct a comparative analysis and further limiting the previous analyses.

When disclosing non-financial information, organisations use the aforementioned Global Reporting Initiative guideline as a reporting standard [50], which includes some diversity-related indicators. As indicated by Hąbek [51], using GRI standards influences the quality of CSR reporting. However, as noted by Werkmeister [1], none of the GRI reporting guidelines specify diversity among the top-level reporting categories, making it simply one of many indicators for reporting labour aspects and some other core indicators related to diversity. The existing analyses of diversity reporting show differences in the understanding of what diversity is. Most organisations report on gender [1,52] and age diversity, disregard other dimensions [6], and present a simple diversity analysis limited to the disclosure of the age and gender structure in their workforce [41], while in most cases focusing on diversity reporting among management boards, thus linking diversity reporting with corporate governance [53]. Even if they report on the composition of the management board based on gender or age [6], only a few of them report on issues such as the gender pay gap or incidents of discrimination [54]. Furthermore, they present a limited scope for applying diversity reporting, thus limiting it mostly to the workforce and omitting suppliers and customers, among others [42]. Moreover, organisations do not present information about diversity targets. This makes it difficult to measure progress. However, even without explicit specification of a target, social comparisons of outcomes can motivate organisations to set goals and make improvements [55]. Furthermore, in contrast to CSR or environmental reports, diversity reports generally do not include key performance indicators (KPIs) [42]. Werkmeister emphasised that most organisations also do not provide juxtapositions to previous years or reporting periods [1], thus making it even harder to analyse the disclosed data in terms of targets or goals.

However, as shown by Vaughan [56], within legal services in the UK, a principle of which has imposed the obligation to collect data on workforce diversity since 2011, the percentage of organisations disclosing information on workforce diversity was estimated at over 68 per cent in 2014. Furthermore, in comparison to 2010, the disclosed information was more detailed in terms of the diversity dimensions. In 2015, all analysed organisations reported on gender, over 90 per cent reported on ethnicity, 78 per cent reported on disability and nearly 70 per cent reported on age. The least disclosed diversity dimension (58 per cent) was caring responsibilities [57]. A similar positive effect of regulating diversity reporting was noticed by Shimeld et al. [17].

Venturelli et al. [58] developed an assessment tool for analysing the reporting scope of non-financial information and diversity by combining diversity disclosure with reporting of other non-financial information. The tool’s development was a consequence of the adoption of the Directive 2014/95/EU and the resulting necessity to report on diversity issues. Using the chosen categories, the reporting and exposure level of Italian organisations in regards to non-financial information were subjected to analysis, thus demonstrating that the diversity dimension was mostly neglected by these organisations as consequence of the company management’s reluctance to engage and deal with diversity issues [59].

In view of the above, it was decided to conduct an exploratory study to determine the scope of diversity reporting, i.e., the detail of collected and disclosed diversity information, including a possible difference between disclosing diversity information towards internal and external stakeholders in organisations that map their stakeholders as well the level of overlap between the collected and disclosed diversity information.

Furthermore, due to the lack of analyses of the determinants of diversity reporting as well as the described connection and strong intersections between CSR and diversity management, it was decided to analyse the possible impact of specific CSR reporting determinants on diversity reporting. The studies suggest that organisations tend to conduct selective disclosure of non-financial information [60,61,62,63]. Many researchers analysed the determinant of this process [64], especially in regards to environmental information [65]. Among the determinants influencing the disclosure of non-financial information, the researchers pointed to the organisation’s characteristics, i.e., its size [66,67,68]. In regards to CSR, it was confirmed that compliance between the level of disclosure with sector guidelines [69] and general CSR disclosure [70] has a positive correlation with the organisation’s size. In relation to the level of disclosure of diversity information, Vaughan [57] demonstrated a positive correlation between the size of the analysed organisations (Chambers) and the willingness to publish diversity data. In relation to the above, I wanted to verify whether such a correlation also applies in regards to Polish organisations indexed on the Stock Exchange in Warsaw and representing a cross-section of almost all economic sectors. Thus, the first hypothesis was adopted:

H1. 

The organisation’s size affects the scope of diversity reporting.

Diversity reporting as a dependent variable was analysed with the division to collecting and disclosing diversity information.

Among other determinants influencing CSR reporting, researchers pointed towards the companies’ leverage [65], performance [71], operations on foreign markets [72], media visibility [67], ownership structure [67], research and development expenditures [65], industry [64,66,73,74] and the age of assets [74], among others.

To verify the possible impact of other organisation characteristics on diversity reporting, it was decided to analyse selected characteristics. Due to the sensitive nature of the information required for establishing a company’s leverage, it was decided to omit this variable, as the risk of not obtaining the necessary data through CATI research was assessed as excessively high. Similarly, the determinant of media visibility was rejected as difficult to obtain due to the nature of the conducted research. Thus, it was finally decided to adopt the following two hypotheses, which cover other organisation characteristics as variables:

H2. 

Financial performance affects the scope of diversity reporting.

H3. 

Operation on a foreign market affects diversity reporting.

The choice of the financial performance variable was dictated by the assumption that organisations in a better financial situation will be more willing to collect and disclose diversity information. The choice of the operation on foreign markets variable was dictated by the assumption that entities operating internationally act in a more diverse environment and thus have more possibilities to collect diversity data and are more pressured by diverse stakeholders to disclose this information.

Furthermore, studies analysed the impact of general contextual factors on CSR disclosure, thus demonstrating variation between CSR practices and especially CSR disclosure [75,76]. These included differences deriving from the national institutional context between the analysed regions (in the mentioned study, the authors referred to the United States and Europe) [76], but also among European countries [77]. According to Ali et al. [64], relatively little attention is paid to internal contextual factors. The existing research focuses on the attitudes of corporate executives towards CSR reporting [78]. Because of the indicated intersectionality between CSR and diversity management, it was decided to include an additional variable, which would show this correlation and the possibility to embed diversity into sustainability reporting, namely the development of non-financial reports variable. This led to the adoption of the following hypothesis:

H4. 

Development of non-financial reports affects diversity reporting.

It was assumed that organisations developing non-financial reports, i.e., already collecting and disclosing non-financial information, would be more willing to embed diversity information into the already existing process of collecting and disclosing other non-financial information.

3. Materials and Research Methodology

As already indicated, the Ministry of Finances assessed that 483 Polish organisations indexed on the Stock Exchange in Warsaw would be affected by the adoption and transposition of the Directive. Consequently, the research population was defined as all organisations listed on the Stock Exchange, which by the time of defining the population and sampling frame consisted of 493 organisations, constituting the full population. During the actual CATI (computer-assisted telephone interviews), contact was made with all organisations. The final sample consisted of 102 organisations, because other entities declined participation in the study due to various reasons, including the lack of a CSR and diversity strategy, and thus no disclosure in those areas or the necessity to obtain a permission from the Management Board to participate in any research, which was deemed as an excessively time-consuming process.

The research tool, a questionnaire, was developed based on a literature study and the operationalisation conducted for the purpose of the study. The research tool was pre-tested during a pilot study, which resulted in minor changes and further specifications of the questions. The actual study and fieldwork were conducted between September and November 2016. The time-period was dictated by the time frames included in the Directive, which states that the reporting obligation is binding as of January 2017. The reason for conducting the study in 2016 was the intention to capture the regularities in diversity reporting before the legal obligation’s entry into force and thus enabling a possible future comparison and assessment of the Directive’s effectiveness.

The questionnaire was developed using mostly the Likert scale, thus enabling the selection of the tools intended for analysing the quantitative variables. The variables used in the study are defined in

Table 1. Conceptualisation of the variables.

Variable	Definition
Stakeholder mapping	The process of identifying and analysing stakeholders. The Respondents were asked whether the organisation conducts such a process.
Internal/external stakeholders	Internal stakeholders were defined as individuals within the organisation and external stakeholders as the individuals, groups and parties outside the organisation, which however are being affected by the organisation’s actions.
Organisation’s size	The number of organisation’s employees.
Financial performance	The profit/loss that the company recorded in the last financial year or period.
Operations on foreign markets	The respondents were asked whether the organisation operates on foreign markets.
Development of non-financial reports	The respondents were asked whether the organisation develops non-financial information.

Source: own elaboration.

.

Acknowledging the fact that the choice of a particular firm size measure can affect the empirical findings due to their sensitivity [79], the decision to select the particular firm size measure was dictated by the provisions of Directive 2014/95/EU, which obliges organisations hiring more than 500 employees to report on diversity information. Thus, it was decided to use the number of employees to determine the size of the organisation.

Due to the mostly qualitative nature of the variables, it was decided to employ the Cramer’s V contingency measure as well as the Kruskal–Wallis H test. Furthermore, ordinal regression was used to verify the assumed hypotheses. The values of the model’s coefficients, together with their standard errors and statistical value t were presented. In addition, odds ratios (ORs) and their confidence intervals were calculated. The models were verified using the Brant test. A significance level of 0.05 was assumed.

4. Diversity Reporting in Polish Companies

The first analysed issue was the scope of diversity information collected by Polish organisations. Due to the possible differences between the willingness to disclose non-financial information towards internal and external stakeholders [80], organisations were divided into a group that identifies its stakeholders (28% of the sample) and a group that does not identify its stakeholders (72% of the sample). Firstly, the respondents were asked to assess the details of the collected non-financial information, i.e., diversity information. The obtained results were analysed using the Kruskal–Wallis H test and are presented in

Table 2. Details of diversity information collected by organisations that map and that do not map their stakeholders.

	H	p-Value	Median (M)
			Organisations Mapping Stakeholders	Organisations Not Mapping Stakeholders
Employee diversity	66.08508	0.00000	4	1

Source: results of own research.

. The organisations were also asked to assess the details of the collected employee diversity information on a scale from 1 to 5, in which 1 meant that the organisations do not collect the data at all and 5 meant that the collected data is very detailed.

The results show that there is a statistically significant difference between the details of the data collected by organisations mapping and not mapping their stakeholders. While organisations classified in the first group collect detailed diversity information (M = 4; arithmetic mean A = 3.965517), organisations that do not map their stakeholders generally do not collect any information on employee diversity (M = 1, A = 1.277778). Since the organisations that do not map stakeholders do not collect diversity information, further analysis was focused on organisations that map their stakeholders. This however limited the cognitive value, especially of the regression analysis, as the analysed sample was very small (28% of the original sample).

The organisations that declared the collection of diversity information were asked to assess the details of the collected diversity information in selected diversity dimensions. The results are presented in

Table 3. Details of collected information with division to selected diversity dimensions.

Diversity Dimension	A
Gender	4.758621
Age	4.758621
Disability	4.655172
Ethnic origin/race	2.964286
Nationality	4.448276

Source: results of own research.

.

In general, the organisations declared a very high level of detail of the collected diversity information, except for information regarding ethnic origin/race.

The respondents were then asked to assess the details of diversity information disclosed to internal and external stakeholders. The results are presented in

Table 4. Average declared details of diversity information disclosed to internal and external stakeholders.

	Internal Stakeholders (A)	External Stakeholders (A)
Employee diversity	3.925926	3.344828

Source: results of own research.

.

The declared details of diversity information disclosed to internal stakeholders is slightly higher than the details of information disclosed to external stakeholders. However, the details of disclosed diversity information are average. The organisations were then asked to assess the degree at which the collected and disclosed diversity information overlap. The results are presented in

Table 5. Overlap between collected and disclosed diversity information with division to stakeholder group.

	H	p-Value	Median (M)		Arithmetic Mean (A)
			Internal Stakeholders	External Stakeholders	Internal Stakeholders	External Stakeholders
Employee diversity	5.014465	0.0251	5	5	4.708333	3.793103

Source: results of own research.

.

The analysis then embraced the issue of possible diversity reporting determinants. For the purpose of this study, the following possible determinants were selected: organisation’s size, financial performance, operations on foreign markets and the development of non-financial reports. The determinants of collecting diversity information was then subject to analysis. The results are presented in

Table 6. Impact of selected variables on the details of collected diversity information.

	χ2	df	p-Value	V
Organisation’s size	12.44583	6	0.1324	0.4632314
Financial performance	13.63607	8	0.0918	0.4848759
Operations on foreign markets	6.397367	4	0.1713	0.4696795
Development of non-financial reports	6.397367	4	0.1713	0.4696795

Source: results of own research.

.

The obtained results are of no statistical significance. It is however important to note that the values of the Cramer’s V coefficient give evidence of a moderate correlation between the variables. However, it is not possible to determine a statistically relevant dependency between the analysed determinants and the details of the collected diversity information. Furthermore, due to the moderate correlation, it can be concluded that the selected variables do not necessary impact the details of the collected diversity information.

The issue of diversity information disclosure was analysed next. The results are presented in

Table 7. Impact of selected variables on the details of disclosed diversity information.

	External Stakeholders				Internal Stakeholders
	χ2	df	p-Value	V	χ2	df	p-Value	V
Organisation’s size	3.47121	8	0.9014	0.24464	11.22500	8	0.18927	0.45593
Financial performance	14.1657	8	0.0777	0.49420	12.26087	8	0.13994	0.47650
Operations on foreign markets	3.78407	4	0.4360	0.36122	2.744505	4	0.60145	0.31882
Development of non-financial reports	15.84386	4	0.00324	0.7391482	12.01648	4	0.01723	0.6671244

Source: results of own research.

.

In the case of diversity information disclosure, most of the obtained results were related to the impact of the selected variables: organisation’s size, financial performance and operation on foreign markets, is not statistically relevant either. In addition, the values obtained by the Cramer’s V coefficient indicate a rather low and moderate correlation between the variables. The only instance in which a statistically significant correlation was identified is the impact of the development of non-financial reports on diversity information disclosure. It is interesting to note the slightly higher correlation between the development of non-financial reports and the disclosure towards external, rather than internal stakeholders, which may suggest that in general non-financial reports are developed specifically for external, rather than internal stakeholders.

To verify the assumed hypothesis, the already mentioned ordinal regression was calculated.

Table 8. Impact of selected variables on the specificity of collected diversity information.

	Model Summary				Odds Ratio and Confidence Intervals
	Value	Std. Error	t Value	p Value	OR	2.5%	97.5%
Organisation’s size (50 to 249 employees)	−1.028	1.360	−0.756	0.450	0.358	0.023	5.44
Organisation’s size (over 500 employees)	0.960	1.157	0.829	0.407	2.611	0.252	26.70
Financial performance (loss)	−1.699	1.865	−0.911	0.362	0.183	0.003	6.73
Financial performance (profit)	−0.072	1.341	−0.054	0.957	0.930	0.038	11.44
Operation on foreign markets (Yes)	1.883	0.962	1.958	0.050	6.572	1.045	47.73
Development of non-financial reports (Yes)	1.194	1.050	1.137	0.256	3.299	0.409	27.95
(1) we do not collect any data; (2) we collect vey general data	−1.575	2.230	−0.706	0.480
(2) we collect very general data; (3) we collect general data	0.209	2.062	0.101	0.919
(3) we collect general data; (4) we collect rather specific data	1.347	2.090	0.645	0.519
(4) we collect rather specific data; (5) we collect very specific data	3.246	2.216	1.464	0.143

Source: results of own research.

presents the influence of the selected variables on the specificity of the collected diversity information.

Table 9. Verification of the model.

	Statistics	df	Probability
Whole model	2.160	18	1.000
Organisation’s size	0.655	6	0.995
Financial performance	0.000	6	1.000
Operation on foreign markets	0.176	3	0.981
Development of non-financial reports	0.573	3	0.903

Source: results of own research.

presents the verification of the model.

In the case of the model concerning the scope of the collected diversity information, statistically relevant dependencies can be found only in relation to the variable: operation on foreign markets. A positive answer to the question significantly increased the chances for the appearance of a higher category in the dependent variable (over 6 times).

Next, the issue of the impact of the selected variables on the specificity of diversity information disclosed to external stakeholders was analysed. The results are presented in

Table 10. Impact of selected variables on the specificity of diversity information disclosed to external stakeholders.

	Model Summary				Odds Ratio and Confidence Intervals
	Value	Std. Error	t Value	p Value	OR	2.5%	97.5%
Organisation’s size (50 to 249 employees)	−1.14	1.64	−0.695	0.487	0.319	0.008	6.9
Organisation’s size (over 500 employees)	−1.03	1.29	−0.797	0.425	0.357	0.015	3.7
Operation on foreign markets (Yes)	1.26	0.97	1.302	0.193	3.537	0.531	25.7
Development of non-financial reports (Yes)	4.24	1.45	2.917	0.004	69.443	6.008	2560.3
(1) we do not collect any data; (2) we collect vey general data	2.16	1.63	1.328	0.184
(2) we collect very general data; (3) we collect general data	2.79	1.69	1.650	0.099
(3) we collect general data; (4) we collect rather specific data	3.26	1.73	1.887	0.059
(4) we collect rather specific data; (5) we collect very specific data	4.39	1.81	2.423	0.015

Source: results of own research.

, while

Table 11. Verification of the model.

	Statistics	df	Probability
Whole model	−10.76	12	1.0
Organisation’s size	48.43	6	0.0
Financial performance	2.37	3	0.5
Operation on foreign markets	0.00	3	1.0
Development of non-financial reports	−10.76	12	1.0

Source: results of own research.

presents the verification of the model.

Due to the lack of the model’s convergence, the variable financial condition was removed. In this model, the variable development of non-financial reports had significant impact on the category of the dependent variable—the affirmative answer increased the chance of appearance of a higher category in the dependent variable. However, the assumption of the same impact of the company’s size on the dependent variable was not fulfilled.

Finally, the impact of the selected variables on the specificity of diversity information disclosed to internal stakeholders was analysed. The results are presented in

Table 12. Impact of selected variables on the specificity of diversity information disclosed to internal stakeholders.

	Model Summary				Odds Ratio and Confidence Intervals
	Value	Std. Error	t Value	p Value	OR	2.5%	97.5%
Organisation’s size (50 to 249 employees)	−1.480	1.525	−0.970	0.343	0.228	0.009	4.27
Organisation’s size (over 500 employees)	−2.076	1.375	−1.510	0.146	0.125	0.005	1.39
Operation on foreign markets (Ye)s	0.882	0.925	0.954	0.351	2.416	0.396	16.02
Development of non-financial reports (Yes)	3.448	1.174	2.936	0.008	31.435	3.837	439.92
(1) we do not collect any data; (2) we collect vey general data	−0.580	1.405	−0.413	0.684
(2) we collect very general data; (3) we collect general data	−0.275	1.393	−0.198	0.845
(3) we collect general data; (4) we collect rather specific data	0.858	1.421	0.604	0.553
(4) we collect rather specific data; (5) we collect very specific data	2.077	1.496	1.388	0.180

Source: results of own research.

, while

Table 13. Verification of the model.

	Statistics	df	Probability
Whole model	−4.62	12	1
Organisation’s size	−7.28	6	1
Financial performance	27.87	3	0
Operation on foreign markets	0.00	3	1
Development of non-financial reports	−4.62	12	1

Source: results of own research.

presents the verification of the model.

Due to the lack of the model’s convergence, the variable financial condition was removed. Similar to the previous model, the impact of developing non-financial reports turned out to be significant. However, for the variable operating on foreign markets, the assumption of proportionality was not fulfilled. This means that its impact on the dependent variable was different for each category.

The results of the ordinal regression analysis of the model and the verification of the hypothesis are presented in

Table 14. Summary of odds ratios from ordinal regression models.

	Impact of Selected Variables on the Specificity of Diversity Information
	Collected	Disclosed to External Stakeholders	Disclosed to Internal Stakeholders
Organisation’s size (50 to 249 employees)	0.358 (0.023–5.44)	0.319 (0.008–6.9)	0.228 (0.009–4.27)
Organisation’s size (over 500 employees)	2.611 (0.252–26.70)	0.357 (0.015–3.7)	0.125 (0.005–1.39)
Financial performance (loss)	0.183 (0.003–6.73)	-	-
Financial performance (profit)	0.930 (0.038–11.44)	-	-
Operation on foreign markets (Yes)	6.572 (1.045–47.73)	3.537 (0.531–25.7)	2.416 (0.396–16.02)
Development of non-financial reports (Yes)	3.299 (0.409–27.95)	69.443 (6.008–2560.3)	31.435 (3.837–439.92)

Source: results of own research.

.

The conducted analysis shows that hypothesis H3 and H4 were partially verified. Operation on foreign markets affects the specificity of the collected diversity information (H3). Organisations that operate on foreign markets are more likely to collect more specific information. The impact of this variable on the specificity of the disclosed diversity information was not verified. Additionally, the impact of the development of non-financial reports on the specificity of disclosed diversity information was confirmed (H4). Organisations that develop non-financial results are more likely to disclose more specific diversity information to internal and external stakeholders. The impact of this variable on the specificity of the collected information was not verified. Furthermore, hypotheses H1 and H2 were not verified.

5. Discussion and Conclusions

The necessity for diversity reporting is linked to the growing perception of diversity as a business issue [42]. It can therefore be expected that it will only gain in importance and become more widespread, especially with legal support, i.e., in the form of the already mentioned EU Directive, but also in view of the EU’s new attempts at establishing a quota for women in management boards [81]. However, where reporting requirements exists, they are unlikely to meet the requirements of the stakeholders [82]. It is therefore important to conduct an analysis in order to determine whether the Directive and the national law transposing it will actually affect the specificity of the collected and disclosed diversity information. This issue raises some doubts, because the Directive itself does not enforce any particular reporting guidelines or monitoring instruments. As noted by Dyduch and Krasodomska [83], Polish organisations report on CSR in order to respond to public pressure, which has also manifested in legal regulations, and in order to maintain corporate legitimacy. However, other studies suggest that legal standards can impose diversity reporting, especially in areas strictly specified by the law [17].

Furthermore, diversity reporting still seems to be a novelty among Polish organisations. Only a small group of the analysed organisations collects and discloses diversity information. This confirms the previous research results [5] and shows that there is still room and a necessity for improvement. The question of the reasons for such a small engagement in collecting and disclosing diversity information and thus diversity reporting remains open. One of the possible solutions to the relatively low level and scope of diversity reporting may be CSR-contingent compensation contracts, which correlate the executives’ remuneration with achieving non-financial goals [84]. Correlating diversity reporting to the executives’ remuneration can promote the idea of solid diversity reporting, but also prevent situations where the implementation of diversity measures into CSR reporting are sham activities aimed at improving the organisation’s image. This opens the door for a new determinant in diversity reporting. Further studies should therefore analyse whether CSR-contingent remuneration or the level of social visibility can affect the scope of diversity management.

The first issue that seems to affect diversity reporting is the stakeholder mapping, because organisations that declared not identifying nor analysing their stakeholders collect almost no diversity information. The analysed variables demonstrate that the only determinant affecting the disclosure of diversity information seems to be the development of non-financial reports. While the other variables were identified as ones affecting CSR reporting, they do not seem to have the same impact on diversity reporting. However, most of the research conducted on the determinants of CSR reporting focus on environmental reporting [85,86]. Furthermore, literature also contains examples of studies critical to the impact of on an organisation’s size [68] and financial performance [65] on the disclosure of non-financial information. The research has also shown variation between CSR practices and especially CSR disclosure [75,76]. Authors have found differences in the effects of the national institutional context between the United States and Europe [76], but also among European countries [77]. These differences can also apply in the context of diversity reporting.

It is also worth noting that differences in defining diversity are emerging and deepening. The on-going social changes and individualisation lead to an increase in the number of potential diversity dimensions, as well as the possible definitions and understandings of diversity and diversity management. Thus, it can also affect diversity reporting and further deepen the ambiguity in this area. As indicated by Holton [42], there is already a variety of content and framework for diversity reporting, which, if not regulated or even captured in a more or less widely accepted guideline, can deepen and grow into a challenge for practitioners and academics attempting to benchmark or compare diversity reporting.

The paper is not free of limitations. Since the study only embraces companies listed on the Warsaw Stock Exchange, the results cannot be treated as representative for all Polish organisations. The decision to focus on these organisations has however been dictated by the mentioned Directive 2014/95/EU. Furthermore, the rejection rate is relatively high and thus one must be cautious when trying to generalise the results. In addition, the nature of research does not allow for deeper conclusions. For this purpose, in-depth interviews, especially with organisations disclosing diversity information, would be beneficial. Another limitation is connected with the endogeneity problem. As indicated by Li [87], this issue occurs especially in connection to corporate finances. Since financial performance was chosen as one of the variables in this paper, and that there can be a reversed causality between some of the variables, the endogeneity problem is also of importance for this analysis. Another potential cause for the endogeneity problem may be the existence of variables affecting the level and scope of diversity management, which were omitted in this study. The choice of the particular variables was made based on the existing literature on the determinants of corporate social responsibility, which may not apply in the case of diversity reporting as seen in the above analysis. However, the lack of necessary information makes it difficult to address and solve the endogeneity problem. Hence, it is acknowledged that reversed causality is a limitation of this paper and should be analysed and further verified in a following study.

Despite this limitation, it is believed that the analysis has several potential theoretical and practical implications. First, since diversity reporting has not yet been thoroughly analysed, the paper provides important input into the theoretical discussion concerning diversity reporting. It shows that, although diversity reporting can be included in wider sustainability reporting, the diversity reporting determinants may not necessarily be the same as in the case of CSR reporting. As presented, we could only find statistically significant and strongest correlation for the development of non-financial reports. As for the other analysed variables, i.e., organisation’s size, operation on foreign markets and financial performance, no bases to conclude the existence of their impact on diversity reporting were found. The conducted ordinal regression analysis allowed for confirming the impact of developing non-financial reports on the specificity of disclosed diversity information. Additionally, it showed the impact of the variable operation on foreign markets on the scope of collected diversity information. Thus, hypotheses H3 and H4 were partially verified. The impact of other variables could not be verified.

Furthermore, since the analysis was conducted among Polish organisations, the paper provides results on diversity disclosure in a context that differs from traditional settings. The theoretical analyses regarding diversity reporting were conducted in Anglo-Saxon, Australian and West-European countries and we believe it to be important to include into the discussion our analysis conducted in different settings, i.e., in a country believed to be traditional and homogenous in terms of most of the basic diversity dimensions [88]. A comparative analysis can present a possibility of improving diversity reporting and embedding diversity into sustainability reporting. It would also make it possible to determine whether the national institutional context of the country in which the organisation operates affects diversity reporting.

Furthermore, if diversity is to be included in CSR reporting, the research results presented above enrich the theoretical discussion, because the determinants of a broader sustainability reporting can be subject to change.

The paper demonstrates that organisations that map their stakeholders collect and disclose more diversity data than organisations that do not map their stakeholders. It can therefore be assumed that organisations that map their stakeholders are generally more aware of the issue and benefits of diversity, which can be a result of stakeholder mapping. Therefore, it can be concluded that stakeholder mapping and analysis should become an element of process management and should be present and considered at all stages of enterprise management [89]. This however requires further analysis.

The paper also shows that the scope of diversity information collected and disclosed does not necessarily overlap, especially in terms of external stakeholders. This fact opens up an interesting research question concerning the nature of the data which is not being disclosed and the reasons for this.

The presented results are also relevant for practice, as they provide understanding of the scope of diversity reporting as well as the aspects important for enhancing diversity and sustainability disclosure. Diversity reporting allows organisations to focus on their resources and thus adapt the organisation to market changes, or even invoke changes [90]. The results can also be of interest to audit professionals, mainly because they show the possibility of embedding diversity into CSR reporting, and secondly because they can constitute an impulse for including diversity into auditing standards and guidelines. The results can also be of interest for policy makers, because they indicate areas which may require further regulation.

Even though the presented results provide an input into the understanding of diversity reporting and disclosure, the topic requires further analysis, as already indicated above. Although no strong correlation between the analysed variables could be detected, the study pushes forward our understanding of the disclosure of non-financial information, as it shows that, although diversity can and very often is implemented into CSR reporting, the motives for doing so and for the disclosure of diversity information may be different than in the case of CSR reporting. Furthermore, the impact of operation on foreign markets on the collection of diversity data and the impact of developing non-financial reports on the disclosure of diversity information to internal and external stakeholders were confirmed. Therefore, this paper shows an existing research gap as well as calls for further analysis. It also pushes forward the understanding of the determinants of diversity reporting.