Diversity at the Top: How Ethnic Composition of Management Influences Corporate Performance in U.S. Companies

Abstract

This paper aims to investigate the impact of ethnic diversity among employees and managers on firm performance, focusing on return on assets and return on equity. The analysis is conducted on a sample of 391 U.S. companies over a five-year period, 2020–2024. The quantitative framework includes a wide range of indicators related to financial performance, ethnic diversity among employees, ethnic categories of managers, and other control variables. The research methodology employs the ordinary least squares (OLS) method to highlight these effects, using fixed-effects and random-effects regression models, both linear and nonlinear. By estimating the regression models, the empirical results support the hypotheses established in the current state of the literature, indicating that ethnic diversity affects firm performance in a mixed manner, with both positive and negative effects on ROA and ROE. These findings are particularly relevant for practitioners, given the need to integrate minority representation into performance assessment, risk evaluation, and decision-making processes. Furthermore, regarding the female component within firms, this dimension contributes to the promotion of sustainability and a sound ESG-oriented approach. Consequently, social factors such as ethnicity can influence companies’ financial performance and shape how firms are perceived by investors.

1. Introduction

Ethnic diversity in a company refers to the presence of employees and managers from multiple ethnic groups, differing in socio-economic, cultural, or professional backgrounds. This diversity entails cultural, linguistic, and historical differences that can significantly influence decision-making, innovation, and communication within an organization. In the academic literature, ethnic diversity is often associated with concepts such as organizational inclusion and corporate social responsibility (CSR), and it also represents a key criterion in environmental, social, and governance (ESG) evaluations, particularly with regard to the social dimension. CSR reflects companies’ commitment to ethical and responsible practices, while the ESG framework evaluates their sustainability, with ethnic diversity being a central element of the social component.

Accordingly, numerous studies highlight that a higher level of ethnic diversity can lead to improved organizational performance, especially in more developed environments where multiple perspectives become strategic advantages.

Ethnic diversity is increasingly perceived as an important element of organizational structure, with significant implications for firm performance and internal culture, thus becoming a key indicator in assessing corporate social responsibility and human resource effectiveness.

Numerous studies in the literature support the existence of a significant relationship between ethnic diversity and firm performance. For example, Cho and Hashemi Joo (2024) show that companies with an ethnically diverse management team are more likely to achieve above-average profits compared to others, and that ethnic diversity on the board of directors reduces the risk of corporate bankruptcy. Similarly, Raina et al. (2024) highlight that diversity in top management positions can contribute to better strategic decision-making, market adaptability, and organizational innovation.

For instance, the proportions of African American, Hispanic, Asian, and other ethnic minority employees, as well as their representation in managerial positions, provide a concrete picture of the level of inclusion practiced within a firm; however, their impact on ROA and ROE appears to be mixed. Moreover, ethnic balance can influence a firm’s level of innovation and its ability to attract diverse talent. Linking these data with performance indicators therefore becomes essential in order to understand the extent to which ethnic diversity contributes to organizational efficiency.

For example, according to a study by Ha et al. (2024), ethnic diversity in management is positively associated with firm performance, and internationalization strategies should be accompanied by effective knowledge management and the promotion of gender diversity in order to maximize performance-related benefits.

Therefore, understanding the relationship between ethnic diversity within companies and their performance becomes essential in a global economic context characterized by multiculturalism and social responsibility, both in terms of ethnicity and gender. Regardless of the industry or firm size, diversity is no longer merely a social indicator but a strategic factor that can decisively influence competitiveness and long-term sustainability. Existing studies have explored diversity in various ways; however, the findings remain fragmented, and most focus on linear relationships, overlooking the complex dynamics captured by analyzing nonlinear relationships.

Research on the link between ethnic diversity and firm performance remains limited, with most studies relying on linear models that fail to capture more complex relationships. This paper addresses this gap by applying advanced econometric methods to analyze the selected companies, offering a more detailed and robust perspective on the impact of employee and managerial diversity on enterprise performance. This paper addresses these gaps, particularly the lack of studies examining the combined effects of employee and managerial diversity using advanced econometric techniques. Furthermore, existing databases often do not cover long time periods and include incomplete or previously unreported firm-level information, while the current literature remains limited in its analysis of ethnic minorities.

The study aims to highlight both the direct and nonlinear effects of workforce and managerial ethnic diversity on firm performance, providing insights for both academic research and corporate decision-making. This topic is particularly relevant in today’s context, where firms seek to understand the impact of internal factors, namely employees and managers, on organizational outcomes, and to adapt their human resources strategies accordingly. The findings of the study provide stakeholders with valuable insights regarding resource diversity, long-term trends, and support more informed decision-making on diversity and performance policies by rigorously examining how ethnic diversity affects company performance.

The topic of firm performance is widely discussed, and ethnic diversity has become an important factor in its analysis. This study aims to examine the relationship between ethnic diversity among employees and managers and company performance. The purpose of the paper is to provide a clear perspective on the impact of diversity on organizational outcomes and how it can contribute to improved performance.

The study contributes by using an extensive database covering a significant period of the last five years up to the present, allowing the observation of the evolution of ethnic diversity within companies and the identification of relevant long-term trends. This provides a more comprehensive picture of how the ethnic composition of the workforce and management influences company performance in today’s context, where racism and gender inequality should be fully eliminated.

Furthermore, the paper employs nonlinear regression models, an important contribution given that existing research largely relies on linear analyses. Nonlinear models allow for the capture of complex relationships between ethnic diversity and performance, including indirect effects, thresholds, or interactions that may not be visible in traditional models. As a result, the findings are more robust and offer a nuanced perspective on how diversity contributes to firm success.

By combining an extensive database with advanced econometric methods, the study makes a significant contribution to understanding the relationship between the ethnic diversity of employees and managers and enterprise performance, providing insights useful for both researchers and corporate stakeholders.

The structure of the paper is as follows: Section 1 offers an introduction to the topic of ethnic diversity and its importance for firm performance. Section 2 presents a literature review, highlighting the main contributions in the field. Section 3 outlines the research hypotheses, along with the methodology and econometric techniques used, while Section 4 presents the analysis results. Section 5 delves deeper into the empirical findings, and the paper concludes with Section 6, which synthesizes the study’s conclusions and its academic and practical implications. This structure ensures that the study is well-developed, systematically addresses the research questions, highlights the key findings, and discusses their practical and academic implications.

2. Prior Literature and Hypotheses Development

Regarding the current state of knowledge, there is a wide range of academic articles aligned with contemporary research, depending on the diversity of independent variables considered. The present review focuses on studies examining the impact of ethnic diversity among employees and managers on firm performance, as well as gender diversity, since in companies that are well developed from an ESG perspective it is important to understand why some teams achieve strong results while others do not.

Accordingly, the internal structure of organizations, through employees and managers, is analyzed. The findings indicate that the strongest performance outcomes emerge from teams in which ethnic and gender diversity coexist and interact naturally. Cultural differences bring new perspectives, while gender diversity contributes to more balanced decision-making. In other words, ethnic diversity fosters creativity, whereas gender diversity promotes balance. These findings suggest that sustainable and efficient performance is enhanced by individuals who think differently, collaborate effectively, and transform diversity into a genuine competitive advantage.

In practice, ethnic diversity is often perceived by organizations as a reputational risk, due to concerns about potential negative reactions from customers. However, the study by Walsh (2024) challenges this perception, demonstrating that in some cases employees from ethnic minority backgrounds are evaluated more favorably in terms of competence, satisfaction, and professional relationships. Thus, ethnic diversity can positively influence the customer–employee relationship, and the study supports the view that well-integrated ethnic diversity does not negatively affect performance. This forms the basis for the first research hypothesis: ethnic diversity positively impacts firm performance.

H1. 

Organizations with higher ethnic diversity in leadership achieve better performance.

Another important study, authored by Du et al. (2025), examines how ethnic diversity in corporate leadership influences sustainable investment and performance. The authors show that firms led by managers from ethnic minority backgrounds invest significantly more in environmental projects than those led by majority-group leaders, indicating a stronger focus on ESG considerations. The main conclusion is that ethnic diversity among managers has a tangible impact on environmental performance and corporate sustainability.

Also focusing on management positions, Adamovic and Leibbrandt (2023) investigate the existence of an ethnic ceiling in access to leadership roles, concluding that there is an ethnic glass ceiling that limits ethnic minorities’ access to managerial positions and contributes to their underrepresentation in leadership. As a result, some firms may exhibit different profitability outcomes due to varying managerial perceptions and decision-making approaches.

Furthermore, Smith and De Leon (2023) argue that firms with a higher proportion of employees and managers from racial and ethnic minority groups achieve superior financial performance. Their study shows that ethnic diversity is associated with higher profit margins, suggesting that the integration of employees from ethnic minorities can generate economic advantages. Sajadi and Vandenberghe (2025) reinforce these findings through their research, demonstrating that the management of workplace diversity influences how employees experience ethnic discrimination, and showing that a positive diversity climate reduces perceptions of racial discrimination.

A company’s profitability depends not only on managerial decisions but also on employees and the way they perceive products and build relationships with customers. Walsh (2024) demonstrates that employees from different ethnic backgrounds are evaluated differently by both customers and management, which affects professional assessments and leads to inequalities in career opportunities, even when experience and qualifications are similar.

Gender inequality is directly embedded within the broader issue of diversity, whether referring to gender diversity, racial diversity, or ethnic diversity. Ibrahim et al. (2025), in their study, conclude that women belonging to ethnic minorities are often affected by double or multiple discrimination, namely on the basis of both gender and ethnicity, which places them in a disadvantaged position compared to their male counterparts or women from majority groups. In this regard, the study contributes to the ethnic diversity literature by highlighting that structural inequalities are not exclusively gender-based, but are closely linked to ethnic and racial representation, and it supports the need for diversity, equity, and inclusion policies that simultaneously address gender and ethnicity.

Thus, ethnic and gender diversity go hand in hand, representing important elements in performance evaluation, while the presence of women within corporate structures has gained increased relevance in the analysis of financial indicators such as ROA and ROE. In the current context, a growing body of research seeks to examine the role of women, both as employees and as managers, and to integrate gender balance into ESG strategies.

Women can contribute to diversified perspectives, improved decision-making, and more complex organizational processes, which makes it necessary to assess the impact that female representation has on the return on assets and return on equity, whether in the case of female employees, managers, or board members. In this sense, analyzing the relationship between gender diversity, ROA, and ROE provides valuable insights into whether and to what extent female representation supports financial performance.

Farooq et al. (2025) argue that the presence of women on boards of directors has a significant impact on how companies hold top management accountable, showing that female directors enhance oversight of CEO performance and increase the likelihood that an underperforming CEO will be replaced. Similarly, Andreu et al. (2025) find that the presence of women in managerial positions strengthens firm performance, demonstrating that both intellectual capital and gender diversity in management reinforce financial outcomes.

Another relevant study that places diversity at the forefront is that by Mia (2022), which argues that the presence of women on boards of directors has a positive impact on employee well-being. The assertion that gender diversity on boards plays a crucial role in improving corporate governance is also supported by Mishra (2025). This study advances the hypothesis that the presence of women on boards enhances firm performance, a finding further reinforced by Huang and Lu (2025).

H2. 

A higher presence of women among employees and in managerial positions is positively associated with firm performance.

In addition, Branca et al. (2025) show that female managers encourage the hiring of other women and implement work–life balance policies, such as flexible working hours or adaptable work arrangements. In other words, female leadership can reshape workforce structures and reduce gender gaps.

Whether referring to gender or ethnic diversity, the social dimension is closely connected to performance. Diverse perspectives, varied experiences, and complementary approaches can improve decision-making. Smith and De Leon (2023) argue that gender and racial/ethnic diversity among employees and managers is positively associated with firms’ financial performance.

The discussion is further strengthened by the study of Dukhan et al. (2025), which concludes that the integration of women from ethnic minorities depends on managerial perceptions and attitudes, and that a coherent organizational culture and effective management can transform this diversity into a valuable and efficient resource.

In addition, there are earlier studies that can be considered foundational theoretical contributions. For example, Cox and Blake (1991) provide a cornerstone for a large body of research on diversity, inclusion, and diversity management by introducing the concept of the multicultural organization and explaining the mechanisms through which ethnic diversity influences performance. Similarly, Shore et al. (2009) further strengthen the theoretical framework by clarifying the mechanisms through which diversity affects organizational performance.

3. Empirical Framework

3.1. Description of the Database and Variables

The companies analyzed in this paper are U.S. based firms, and financial performance is measured using the indicators ROA and ROE, which are included in the category of performance variables. As shown in Table 1, these variables reflect the efficiency of asset utilization and the firms’ ability to generate profits for shareholders.

Table 1. Description of variables.

Variable	Acronym	Definition	Formula
Performance variables:
Return on Assets	ROA	How much profit the assets generate	Net profit/Total assets
Return on Equity	ROE	Profit generated from investments	Net profit/Shareholders’ equity
Workforce Score:
Gender Pay Gap Percentage	DSG	Difference in salaries between genders
Women Employees	FA	Women holding jobs
Women Managers	FM	Women in managerial positions at the workplace
Ethnic Minorities Employees Percentage	PA_ME	Percentage of employed ethnic minorities
Black or African American-Ethnic Minorities	N_ME	Employees of African origin
White-Ethnic Minorities Employees Percenta	A_ME	White employees
Ethnic Minorities Managers Percentage	PM_ME	Percentage of minorities in management positions
Black or African American-Ethnic Minorities Managers Percentage	N_MME	Percentage of managers of African origin
White-Ethnic Minorities Managers Percentage	A_MME	Percentage of white managers
Asian-Ethnic Minorities Managers Percentage	AS_MME	Percentage of Asian managers
Hispanic or lation-Ethnic Minorities Managers Percentage	L_MME	Percentage of Hispanic or Latino managers
Other- Ethnic Minorities Managers Percentage	AT_MME	Percentage of managers from other ethnic groups
Number of Employees	NA	Total number of employees
Announced Layoffs To Total Employees	CA_A	Percentage of announced layoffs relative to total employees
Employees With Disabilities	A_DIS	Employees with disabilities
Control variables:
Free cash flow	FCF	Money available after operating and investment expenses	Operating cash flow—Capital expenditures
Size	SIZE	Firm size
Reinvestment rate	RR	Rate at which earnings are reinvested	(Reinvested earnings)/(Net earnings)
Dividend yield	DY	Dividend return relative to share price	(Dividend per share/Share price) × 100
Effective tax rate	ETR	Actual tax rate paid	(Total tax/Taxable income) × 100

Source: Own estimates.

In addition to the dependent performance variables, the study incorporates an extensive set of independent variables from the “workforce score” category, which capture the presence of diversity and inclusion at the organizational level. These include indicators such as the gender pay gap, the share of women in total employment and in managerial positions, the distribution of employees and managers from ethnic minorities, including distinct categories such as African American, White, Asian, Hispanic/Latino, and other ethnic groups, the total number of employees, the percentage of announced layoffs relative to total staff, and the share of employees with disabilities. This approach aims to highlight ethnic diversity within firms and its effects.

Furthermore, the control variables used in the analysis include relevant financial indicators such as free cash flow, firm size, the reinvestment rate, dividend yield, and the effective tax rate.

For all quantitative variables, definitions and calculation formulas are provided in Table 1, and the data employed are drawn exclusively from recognized international databases.

3.2. Description of Econometric Methods

From a methodological perspective, ordinary least squares (OLS) method is used to highlight the impact that certain influencing factors have on the performance of companies operating within an Anglo-Saxon corporate governance system. A random sample of 391 U.S. companies was selected, and this approach was chosen in order to avoid a selection bias focused solely on either the best-performing or the weakest firms, which could have distorted the study’s results. By including in the database both highly profitable firms and firms with lower profitability, the analysis aims to capture a more comprehensive picture of how ethnic-related variables influence corporate performance.

Nevertheless, as a limitation of the study, the random selection of companies may not fully capture the entire scope of ethnic diversity among employees and managers or its correlation with the performance of U.S. firms; however, it does provide a representative overview at a general level.

Within the study, several regression models were estimated, including both linear and nonlinear specifications, as well as fixed-effects and random-effects models. The appropriate model specification was determined using the Hausman test, which allows for the identification of the optimal choice between fixed and random effects based on the correlation between individual effects and the explanatory variables. All estimations were conducted using Stata/MP 15 (64-bit) software, taking into account the characteristics of panel data.

For the nonlinear models, squared terms of the explanatory variables were introduced where curvilinear relationships were possible. In cases where these nonlinear terms proved to be statistically significant, p < 0.05, inflection points were calculated to determine the threshold at which the influence of the independent variable on the dependent variable changes sign or direction, allowing for a more nuanced interpretation of the relationships.

Additionally, during the model-building stage, multicollinearity was assessed using the correlation matrix. Variables showing correlation coefficients exceeding the conventional threshold of ±0.7 were included in separate regressions to avoid potential estimation errors and coefficient distortions.

Regarding data collection and analysis, only official sources were used. Articles relevant for the theoretical foundation and hypothesis development were sourced from ScienceDirect, while empirical data were obtained from the Thomson Reuters platform, which provides extensive financial information, ensuring the robustness of the formulated hypotheses.

Regression analysis represents the main methodological tool used to examine and highlight the relationship between ethnic diversity at both the employee and management levels and corporate performance, employing both linear and nonlinear regression models, as described previously. The companies included in the sample were selected based on data availability and transparency, since access to detailed information regarding the ethnic composition of the workforce and management teams is limited, especially in past years. Not all firms report the ethnic distribution of employees or the ethnic composition of management, which constrains the analysis, although the U.S. is among the most advanced countries in terms of ESG reporting. For this reason, the study uses the subset of companies that provide sufficient data for a solid evaluation of ethnic diversity. Managerial factors are captured through variables such as: women managers, ethnic minority managers, white/asian/hispanic ethnic minority managers, while employee diversity is measured using indicators such as: women employees, ethnic minority employees, black or african american employees, white ethnic minority employees, etc.

The calculation procedures for composite variables are presented in detail in Table 1, variable descriptions. For indicators reported directly in the Thomson Reuters platform, values were taken as provided, as no additional statistical processing was required.

To ensure the validity of the results, the variables were grouped into conceptual categories based on shared characteristics, workforce-related variables, which include both managerial and employee variables, performance variables, and control variables.

This structure facilitates the proper construction of econometric models, so that the estimated set of regressions can be generalized in the following form:

DEP_l = a₀ + a₁ × WORKF ₊ a₂ × CNTRL + ε_it

where:

• a₀ = constant;
• a₁ … a₂₀ = coefficients corresponding to the variable categories;
• i = [1, 391];
• t = [2020; 2024];
• WORKF = workforce-related variables
• CNTRL = control variables
• ε_it = error term

To illustrate the regression in detail, the following regression models were estimated for the dependent variable ROA:

• Linear regressions:

  ROA = a₀ + a₁ × DSG + a₂ × PA_ME + a₃ × NA + a₄ × FCF + a₅ × SIZE + a₆ × RR + a₇ × DY + a₈ × ETR + + ε_it

  (1)

  ROA = a₀ + a₁ × DSG + a₂ × NA + a₃ × FCF + a₄ × SIZE + a₅ × RR + a₆ × DY + a₇ × ETR + a₈ × N_ME + + ε_it

  (2)
• Nonlinear regressions:

  ROA = a₀ + a₁ × DSG + a₂ × NA + a₃ × FCF + a₄ × SIRE + a₅ × RR + a₆ × DY + a₇ × ETR + a₈ × N_MME + a₉ × N_MME_sq + ε_it

  (3)

  ROA = a₀ + a₁ × NA + a₂ × SIZE + a₃ × RR + a₄ × DY + a₅ × ETR + a₆ × A_MME + a₇ × FA + a₈ × A_MME_sq + ε_it

  (4)

4. Econometric Outcomes

As previously mentioned, an empirical study will be conducted to analyze ethnic diversity among employees and managers and its correlation with performance for a sample of 391 U.S. companies. The data are sourced from Thomson Reuters, covering a five-year period from 2020 to 2024.

Table 2 presents descriptive statistics for all variables included in the empirical analysis, providing an overview of the distribution of performance indicators, workforce score variables, and control variables. The total number of observations varies significantly depending on the variable, ranging from 165 observations for the indicator on employees with disabilities to 1930 observations for the indicator on announced layoffs, highlighting substantial differences in the availability of data reported by companies, which is also reflected in the diversity of the indicators.

Table 2. Descriptive statistics.

Variable	Obs	Mean	Std. Dev.	Min	Max
ROA	1912	0.092	0.081	−0.473	0.838
ROE	1909	0.103	6.034	−255.298	22.82
DSG	369	95.019	8.412	54.6	107.2
FA	1699	36.691	14.983	5	94.1
FM	1387	33.066	12.89	8	81.8
PA ME	1196	39.388	14.177	6.15	99.73
N ME	814	10.448	6.797	0.6	31.3
A ME	868	59.057	13.551	22.5	92.91
PM ME	861	28.917	12.647	2.4	90
N MME	479	6.355	4.25	0	24.3
A MME	529	69.449	13.131	10	97.6
AS MME	468	13.363	11.07	0	53.5
L MME	474	8.426	6.03	0	42.6
AT MME	516	5.561	9.093	0	84
NA	1929	57,159.217	150,455.82	10	2,300,000
CA A	1930	0.131	1.33	0	45.31
A DIS	165	4.722	3.583	0	20
FCF	1906	2.75 × 109	8.03 × 109	−2.68 × 1010	1.11 × 1011
SIZE	1784	1.23 × 1010	2.26 × 1010	16,991,000	1.91 × 1011
RR	1825	0.359	5.806	−16.982	225.737
DY	1355	0.075	0.146	−0.579	2.166
ETR	1743	0.128	2.128	−83	27.273

Source: Own estimates.

Regarding financial performance, ROA has a minimum value of −0.473, indicating the existence of companies that recorded significant losses, while ROE shows a large standard deviation of 6.034, reflecting substantial differences in the financial structures of the companies.

When analyzing the independent variables in the workforce score category, there is considerable dispersion in diversity data. For example, the percentage of female employees averages 36.69%, highlighting the wide variation in gender representation across different companies, while ethnic diversity variables show significant variation, with some ethnic groups having extreme values, such as African American managers or Asian American managers.

Overall, the descriptive statistics highlight a highly diverse sample, both in terms of financial performance and workforce characteristics, providing a solid basis for econometric analyses.

Table 3 presents the correlations among the variables used in the analysis, with most coefficients falling below the ±0.7 threshold, indicating weak relationships between variables and suggesting that multicollinearity is not a concern in the models.

Table 3. Correlation Matrix.

Variables	ROA	ROE	DSG	FA	FM	PA_ME	N_ME	A_ME	PM_ME	N_MME	A_MME
ROA	1
ROE	−0.002	1
DSG	−0.006	0.082	1
FA	0.002	−0.007	0.055	1
FM	−0.061	0.003	0.039	0.807	1
PA_ME	0.141	−0.02	0.17	0.312	0.282	1
N_ME	−0.197	−0.035	0.238	0.225	0.37	0.262	1
A_ME	−0.158	0.02	−0.234	−0.219	−0.198	−0.996	−0.274	1
PM_ME	0.096	−0.018	0.138	0.199	0.274	0.713	0.064	−0.662	1
N_MME	−0.112	−0.069	0.184	0.18	0.389	0.223	0.717	−0.24	0.202	1
A_MME	−0.102	0.016	−0.194	−0.178	−0.187	−0.678	−0.047	0.668	−0.993	−0.218	1
AS_MME	0.249	0.038	0.056	−0.003	−0.118	0.481	−0.39	−0.487	0.63	−0.372	−0.618
L_MME	−0.09	−0.036	0.135	0.223	0.313	0.48	0.312	−0.457	0.406	0.432	−0.435
AT_MME	−0.01	0.002	0	0.138	0.113	0.052	−0.085	−0.053	0.267	0.042	−0.265
NA	−0.021	−0.061	0.073	0.137	0.103	0.193	0.339	−0.248	0.177	0.299	−0.222
CA_A	−0.019	0.002	−0.073	−0.006	−0.008	−0.002	−0.03	0.007	−0.025	−0.028	0.018
A_DIS	−0.055	0.051	0.019	−0.159	−0.024	0.052	0.179	0.046	−0.055	−0.023	0.043
FCF	0.279	−0.009	0.06	0.001	0.022	0.157	0.017	−0.145	0.194	−0.008	−0.2
SIZE	0.038	−0.014	0.077	0.008	−0.001	0.187	0.134	−0.17	0.19	0.033	−0.241
RR	0.105	0.149	0.014	0.013	−0.01	0.034	0.006	−0.031	0.014	0.044	−0.029
DY	0.282	−0.01	0.089	−0.059	−0.113	−0.009	−0.105	0.011	−0.074	−0.089	0.072
ETR	−0.004	−0.001	−0.04	−0.005	−0.029	−0.026	0.051	0.112	−0.02	0.006	0.104
Variables	AS_MME	L_MME	AT_MME	NA	CA_A	A_DIS	FCF	SIZE	RR	DY	ETR
AS_MME	1
L_MME	−0.198	1
AT_MME	−0.043	0.033	1
NA	0.013	0.237	−0.039	1
CA_A	0.009	−0.051	−0.014	−0.001	1
A_DIS	0.064	0.178	−0.144	−0.013	−0.028	1
FCF	0.242	0.013	−0.054	0.23	−0.005	0.122	1
SIZE	0.259	0.007	−0.057	0.461	0.032	0.142	0.697	1
RR	−0.016	0.072	0.001	0.015	−0.004	0.092	0.036	0.011	1
DY	−0.081	0.108	−0.124	−0.041	0.022	0.112	0.021	−0.104	0.052	1
ETR	−0.095	−0.059	0.009	0.009	0.003	−0.045	0.004	0.007	0.001	0.017	1

Source: Own estimates.

However, some strong correlations are observed for the workforce score variables. For example, the percentage of female employees is closely related to the percentage of female managers at 0.807, while the proportion of employed minorities shows a high correlation with the percentage of minority managers at 0.713. Additionally, the ethnic distribution of employees shows a very strong negative correlation between A_ME and PA_ME, with a value of −0.996.

Regarding the control variables, the highest correlations appear between firm size and free cash flow as well as between SIZE and the total number of employees. The performance variables, ROA and ROE, exhibit low correlations both with each other and with the other variables.

Overall, the correlation matrix indicates only a few strong relationships, while the remaining variables show low correlations.

For the econometric analysis of the data, regression models without effects, as well as with fixed effects and random effects, were used. The appropriate model between fixed and random effects was chosen using the Hausman test, with the standard 5% significance threshold. If the p-value is below 5%, the fixed effects (FE) model is considered appropriate; if it is above 5%, the random effects (RE) model is used.

The results are interpreted for each table separately, taking into account the dependent variables, ROA and ROE, and the explanatory variables from the workforce score and the control variable categories. Some variables are significant only in certain models, indicating that their impact on performance depends on the combination of indicators used.

The analysis confirms several important relationships. Each results table is examined in detail, and the conclusions are compared with existing literature to explain similarities or differences and to highlight practical implications for investors, managers, and decision-makers.

Table 4 and Table 5 present multiple regression models for ROA, constructed to examine the impact of ethnic diversity among employees and managers on the firms’ return on assets. The coefficient of determination (R-sq overall) indicates that, on average, approximately 51% of the variation in ROA performance is explained by the independent variables analyzed, with regression models showing a minimum R-squared of 0.12 and a maximum of 0.91.

Table 4. Results of the regression models without effects on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROA (1).

Models	−1	−2	−3	−4	−5	−6	−7	−8	−9	−10	−11	−12	−13	−14	−15
DSG	0.0002	0.0008	−0.0001	−0.0007	−0.0006	−0.001	−0.0007	−0.0006	0.0002	0.0004	−0.0009	−0.0007	−0.0011	−0.0011	−0.0009
	−0.2004	−0.6445	(−0.0555)	(−0.9717)	(−0.5246)	(−1.0035)	(−0.6459)	(−0.6149)	−0.1333	−0.3145	(−0.8904)	(−0.7133)	(−1.1520)	(−1.1029)	(−0.9370)
PA_ME	0.0014 **									0.0023 ***	0.0015 **	0.0015 **	0.0008	0.0016 ***	0.0014 **
	−3.0958									−3.5393	−3.3258	−2.8633	−1.5784	−3.5591	−3.2514
NA	0	−0.0000 *	−0.0000 ***	0	0	0	0	−0.0000 *	0	−0.0000 **	0	−0.0000 **	0	0	−0.0000 **
	(−1.1298)	(−2.4684)	(−3.6265)	(−1.4198)	(−1.3146)	(−0.5635)	(−0.5478)	(−2.4553)	(−0.1553)	(−2.7125)	(−1.3362)	(−2.6857)	(−1.0669)	(−0.3326)	(−2.8707)
FCF	0.0000 ***	0.0000 **	0.0000 **	0.0000 ***	0.0000 **	0.0000 ***	0.0000 ***	0.0000 **	0.0000 ***	0.0000 **	0.0000 **	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	−3.5598	−3.057	−2.9695	−3.7407	−3.2536	−3.572	−3.5222	−3.3883	−4.3465	−2.765	−3.4023	−3.4852	−3.5644	−3.8302	−3.546
SIZE	0	0	0	0	0	0	0	0	−0.0000 ***	0	0	0	0	0	0
	(−1.2901)	(−0.9116)	(−0.8678)	(−0.8183)	(−0.4331)	(−1.2239)	(−0.7477)	(−0.3593)	(−3.8832)	(−0.8856)	(−0.8274)	(−0.5884)	(−1.1204)	(−1.2977)	(−0.6821)
RR	0.0444 **	0.0603 ***	0.0578 ***	0.0340 ***	0.0328 ***	0.0309 ***	0.0309 **	0.0334 ***	0.3972 ***	0.0566 ***	0.0295 **	0.0307 ***	0.0298 **	0.0266 **	0.0305 ***
	−3.2563	−3.7152	−3.7309	−4.0609	−3.4971	−3.5185	−3.2761	−3.6653	−8.7594	−3.6624	−3.3077	−3.4463	−3.3999	−2.9885	−3.5176
DY	0.2093 *	0.152	0.2221 *	0.1833 **	0.1774 *	0.1904 **	0.1592 *	0.2113 **	0.3936 **	0.1970 †	0.2262 **	0.2104 **	0.2051 **	0.2025 **	0.2494 ***
	−2.2821	−1.3696	−2.1065	−3.077	−2.3466	−2.7424	−2.1766	−2.8218	−3.2492	−1.8572	−3.1119	−2.8848	−2.9428	−2.9196	−3.472
ETR	−0.0248	−0.0361	−0.0423	−0.0036	−0.0007	0.0114	0.0033	−0.0046	−0.3800 **	−0.0287	0.0015	−0.0091	0.0056	0.0067	−0.005
	(−0.7361)	(−0.8839)	(−1.0980)	(−0.1736)	(−0.0267)	−0.5044	−0.1374	(−0.2005)	(−3.0246)	(−0.7403)	−0.0665	(−0.4096)	−0.2441	−0.2968	(−0.2301)
N_ME		−0.0028 *								−0.0028 *
		(−2.2229)								(−2.3887)
A_ME			−0.0023 ***
			(−3.6428)
PM_ME				0.0007 *
				−2.1861
N_MME					−0.0016						−0.0023
					(−0.9776)						(−1.4629)
AS_MME						0.0017 ***							0.0012 *
						−3.4975							−2.0556
L_MME							−0.0022							−0.0031 *
							(−1.5256)							(−2.2128)
AT_MME								−0.0016 *							−0.0015 *
								(−2.1693)							(−2.2514)
A_DIS									−0.0032
									(−1.5716)
A_MME												0.0002
												−0.4462
_cons	0.0063	0.0609	0.2463 *	0.1278 †	0.15	0.1506	0.1682	0.1488	0.0879	−0.0004	0.1219	0.0724	0.1393	0.1454	0.1149
	−0.06	−0.5	−2.01	−1.75	−1.46	−1.58	−1.66	−1.59	−0.52	(−0.00)	−1.26	−0.67	−1.46	−1.53	−1.29
F statistic	12.67 ***	10.51 ***	12.70 ***	16.55 ***	13.67 ***	17.06 ***	14.08 ***	15.24 ***	33.14 ***	11.76 ***	14.89 ***	14.90 ***	15.53 ***	15.73 ***	16.26 ***
R-sq	0.42	0.44	0.48	0.52	0.59	0.64	0.6	0.61	0.91	0.5	0.64	0.64	0.65	0.66	0.66
Obs	149	114	117	132	85	85	85	87	36	114	84	87	84	84	86
Mean VIF	2.19	2.22	2.2	2.2	2.7	2.69	2.9	2.52	3.66	2.11	2.55	2.51	2.69	2.75	2.39

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 5. Results of the regression models without effects on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROA (2).

Models	−16	−17	−18	−19	−20	−21	−22	−23	−24	−25	−26	−27	−28	−29	−30
DSG	−0.0006	−0.0002	−0.0005	−0.0009	0.0007	−0.0008	−0.0005	−0.0009	−0.0007		−0.0013	0	0.0007	−0.0007
	(−0.7619)	(−0.2447)	(−0.8309)	(−1.2479)	−0.4696	(−1.0885)	(−0.3987)	(−1.2855)	(−0.7011)		(−1.5284)	(−0.0207)	−0.585	(−0.7306)
PA_ME	0.0003		0.0019 ***		0.0028 ***	0.0016 ***						−0.001
	−0.58		−4.7452		−3.7336	−3.7583						(−0.6324)
NA				−0.0000 **		0	0	0	−0.0000 †		−0.0000 *	0	0	0	−0.0000 **
				(−2.9575)		(−1.1417)	(−0.2960)	−0.4138	(−1.8774)		(−2.0612)	(−0.9516)	(−1.5202)	(−1.0185)	(−2.6833)
FCF			0.0000 ***	0.0000 ***		0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
			−7.5093	−6.368		−5.7487	−6.0097	−6.2256	−3.5029	−6.995	−4.1378	−3.5694	−3.7001	−3.967
SIZE			−0.0000 ***		0.0000 *	−0.0000 †	−0.0000 **	−0.0000 **	0	−0.0000 ***	0	0	0	0	0.0000 ***
			(−4.3917)		−2.3543	(−1.8300)	(−3.0922)	(−3.2397)	(−0.7820)	(−3.3307)	(−0.3368)	(−1.1627)	(−1.5103)	(−1.0797)	−5.4576
RR									0.0290 **	0.0184 ***	0.0265 **	0.0449 **	0.0552 ***	0.0253 **	0.0260 ***
									−3.1993	−5.1006	−3.376	−3.3062	−3.6114	−2.898	−6.6778
DY					0.3640 **		0.2697 ***		0.2192 **	0.0571 ***	0.1881 **	0.2117 *	0.1325	0.1354 †	0.0667 ***
					−3.3542		−3.5444		−2.9935	−3.8117	−3.0473	−2.3191	−1.2715	−1.9485	−4.1724
ETR			0.0082	0.0259				0.0158	−0.0054	−0.0078	−0.0015	−0.0268	−0.0123	0.0152	0.0164
			−0.3786	−1.1657				−0.7286	(−0.2388)	(−0.4447)	(−0.0749)	(−0.7966)	(−0.3182)	−0.6757	−0.8896
N_ME		−0.0007		0.0012	−0.002		0.0020 †	0.0020 *	−0.0016 †				−0.0212 ***
		(−0.8253)		−1.158	(−1.4896)		−1.9112	−2.1064	(−1.7929)				(−4.3804)
A_ME		−0.0015 **		−0.001					−0.0019 ***
		(−3.1007)		(−1.5970)					(−3.4508)
PM_ME
N_MME	0.0043 *					0.0015								−0.0247 ***
	−2.6071					−1.2194								(−4.2592)
AS_MME	0.0024 ***			0.0012 †			0.0019 **	0.0023 ***			−0.0046 **
	−3.4837			−1.7998			−2.9236	−4.8796			(−3.1791)
L_MME		−0.0003	−0.0011
		(−0.2910)	(−1.3076)
AT_MME			−0.0007	−0.0006		−0.0011	−0.0015	−0.0005
			(−0.7486)	(−0.6436)		(−1.0560)	(−1.5071)	(−0.4960)
A_DIS
A_MME									0.0007	−0.0006 *					−0.0022
									−1.21	(−2.5440)					(−1.6080)
FA	−0.0009 *		−0.0006 †				−0.0005			0					−0.0001
	(−2.1037)		(−1.7420)				(−1.2390)			(−0.2495)					(−0.2383)
FM		−0.0009 †		−0.0004	−0.0008			−0.0005
		(−1.7447)		(−1.1311)	(−1.3277)			(−1.2032)
CA_A		−0.0024			−0.0144	−0.0009	−0.0107	−0.0045		−0.0012
		(−0.3833)			(−1.1307)	(−0.1832)	(−1.4333)	(−0.9372)		(−0.2807)
AS_MME_sq											0.0001 ***
											−4.5985
PA_ME_sq												0
												−1.5446
N_ME_sq													0.0007 ***
													−3.9242
N_MME_sq														0.0013 ***
														−4.1213
A_MME_sq															0
															−1.0435
_cons	0.1119	0.2559 **	0.1145 †	0.2277 *	−0.0489	0.1038	0.109	0.1523 *	0.2368 *	0.1306 ***	0.2417 **	0.0799	0.1562	0.2456 *	0.1772 ***
	−1.46	−2.64	−1.81	−2.6	(−0.36)	−1.49	−0.96	−2.24	−2.19	−6.41	−2.77	−0.7	−1.34	−2.56	−4.05
F statistic	7.68 ***	3.00 **	12.56 ***	10.23 ***	6.76 ***	11.25 ***	10.59 ***	10.45 ***	13.35 ***	21.40 ***	21.54 ***	11.63 ***	12.33 ***	16.59 ***	14.86 ***
R-sq	0.22	0.12	0.46	0.45	0.28	0.41	0.57	0.49	0.65	0.37	0.72	0.43	0.52	0.67	0.29
Obs	142	138	125	122	127	136	92	121	82	297	85	149	114	85	298
Mean VIF	1.84	1.24	1.56	1.81	1.15	2.43	2.11	2.35	2.47	1.82	5.93	5	6.32	7.96	9.38

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 6 and Table 7 present multiple regression models investigating the influence of ethnic diversity among employees and managers on firms’ financial performance, measured by return on equity. The R-squared coefficient shows that, on average, about 67% of the variation in ROE is explained by the independent variables used in these models. The R-squared values for these regressions range between 0.0313 and 0.9696, depending on the independent variables included in each model.

Table 6. Results of the regression models without effects on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROE (1).

Models	−1	−2	−3	−4	−5	−6	−7	−8	−9	−10	−11	−12	−13	−14	−15
DSG	0.0087	0.0062	0.0075	0.0075	0.005	0.0046	0.0053	0.0043	0.0056 †	0.0078	0.0051	0.0046	0.005	0.0053	0.0045
	−1.4136	−0.8155	−1.0507	−0.9883	−0.8216	−0.7581	−0.9201	−0.7716	−1.8113	−1.029	−0.8525	−0.8346	−0.8606	−0.9298	−0.8181
PA_ME	−0.0025									−0.0075 †	−0.0034	−0.0034	−0.0058 †	−0.0022	−0.0036
	(−0.9365)									(−1.8242)	(−1.2817)	(−1.0993)	(−1.8797)	(−0.8495)	(−1.3910)
NA	0	0	0	0	0.0000 *	0.0000 **	0.0000 ***	0.0000 **	0	0	0.0000 **	0.0000 **	0.0000 ***	0.0000 ***	0.0000 **
	(−0.1112)	−1.0856	−1.2328	(−0.5009)	−2.621	−3.0458	−4.2534	−3.0619	−1.306	−1.1536	−2.743	−3.3981	−3.6434	−4.2614	−3.3839
FCF	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	0	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***
	(−4.3940)	(−4.0151)	(−3.9064)	(−4.3865)	(−6.3291)	(−6.3468)	(−6.2047)	(−6.3789)	(−0.3171)	(−3.8092)	(−6.4622)	(−6.5705)	(−6.5546)	(−6.3333)	(−6.5299)
SIZE	0	0	0	0	0.0000 *	0.0000 *	0	0.0000 *	0	0	0.0000 *	0.0000 *	0.0000 *	0	0.0000 *
	−1.3995	−1.1023	−1.1219	−1.3283	−2.2674	−2.0559	−1.4731	−2.3164	(−0.3121)	−1.0785	−2.4645	−2.5139	−2.0613	−1.6404	−2.5014
RR	2.1794 ***	2.1785 ***	2.1930 ***	2.2397 ***	2.1148 ***	2.1124 ***	2.0916 ***	2.1137 ***	1.0196 ***	2.1908 ***	2.1244 ***	2.1224 ***	2.1226 ***	2.1000 ***	2.1230 ***
	−27.0005	−22.0849	−22.9587	−26.3891	−39.278	−39.268	−40.2927	−39.5812	−11.9755	−22.3988	−39.9166	−40.5191	−40.504	−40.4815	−40.3173
DY	−0.0342	−0.2672	−0.3966	−0.2671	0.5155	0.5386	0.4831	0.4693	−0.2009	−0.4173	0.429	0.4228	0.4589	0.4519	0.3902
	(−0.0632)	(−0.3969)	(−0.6110)	(−0.4427)	−1.1881	−1.2666	−1.1982	−1.068	(−0.8957)	(−0.6219)	−0.9875	−0.9859	−1.1026	−1.1161	−0.894
ETR	0.1105	0.1748	0.1525	0.101	−0.0073	0.01	0.0924	−0.0052	−0.4742 †	0.1515	−0.017	−0.0062	0.0426	0.082	−0.0069
	−0.5535	−0.7057	−0.6427	−0.4813	(−0.0525)	−0.0721	−0.6905	(−0.0387)	(−2.0217)	−0.6175	(−0.1234)	(−0.0478)	−0.3134	−0.6183	(−0.0523)
N_ME		−0.0012								−0.001
		(−0.1592)								(−0.1358)
A_ME			0.0080 *
			−2.0348
PM_ME				0.0034
				−1.07
N_MME					−0.0011						0.0011
					(−0.1193)						−0.1192
AS_MME						0.002							0.0051
						−0.6765							−1.461
L_MME							−0.0224 **							−0.0204 *
							(−2.7989)							(−2.5260)
AT_MME								0.0001							0.0003
								−0.0129							−0.084
A_DIS									0.0013
									−0.3351
A_MME												0.0003
												−0.1031
_cons	−0.6589	−0.5182	−1.0981	−0.7122	−0.522	−0.5248	−0.4066	−0.4538	−0.3093	−0.3342	−0.409	−0.3659	−0.3928	−0.3398	−0.3306
	(−1.0983)	(−0.7147)	(−1.4758)	(−0.9631)	(−0.8878)	(−0.8973)	(−0.7262)	(−0.8257)	(−1.0389)	(−0.4614)	(−0.7049)	(−0.5798)	(−0.6888)	(−0.6115)	(−0.6087)
F statistic	99.8984 ***	77.3649 ***	83.2293 ***	95.7695 ***	260.0862 ***	261.6587 ***	287.8176 ***	263.5359 ***	31.7864 ***	70.6488 ***	240.4903 ***	247.3675 ***	247.6148 ***	261.8832 ***	244.2589 ***
R-sq	0.85	0.8538	0.8593	0.8617	0.9648	0.965	0.968	0.9643	0.9008	0.8583	0.9669	0.9666	0.9679	0.9696	0.9666
Obs	150	115	118	132	85	85	85	87	37	115	84	87	84	84	86
Mean VIF	2.19	2.22	2.2	2.2	2.7	2.69	2.9	2.52	3.66	2.12	2.55	2.51	2.69	2.75	2.39

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 7. Results of the regression models without effects on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROE (2).

Models	−16	−17	−18	−19	−20	−21	−22	−23	−24	−25	−26	−27	−28	−29	−30
DSG	0.0062	0.0021	−0.0021	0.0042	0.0099	0.0064	−0.0078	0.0071	0.0041		0.0056	0.0091	0.0066	0.0053
	−0.4088	−0.1321	(−0.1299)	−0.2565	−0.4811	−0.4015	(−0.2463)	−0.4149	−0.6697		−0.9387	−1.4494	−0.8709	−0.8822
PA_ME	0.0261 *		0.0133		0.0086	0.0052						0.0011
	−2.3247		−1.3254		−0.7612	−0.5533						−0.1132
NA				0		0.0000 **	0.0000 ***	0.0000 ***	0.0000 *		0.0000 ***	0	0	0.0000 *	0.0000 †
				−1.6171		−2.9913	−4.2361	−3.3883	−2.6114		−3.554	(−0.1538)	−0.674	−2.4773	−1.748
FCF			0.0000 †	0.0000 †		0.0000 **	0	0.0000 **	−0.0000 ***	0	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***
			−1.6777	−1.9063		−3.3576	−1.2106	−3.0989	(−6.4580)	−1.2781	(−6.5092)	(−4.3785)	(−4.1805)	(−6.4909)
SIZE			0		0	−0.0000 **	0	−0.0000 *	0.0000 *	0	0	0	0	0.0000 *	0
			(−0.6854)		−0.9298	(−2.7102)	(−1.0486)	(−2.4478)	−2.5728	(−1.2953)	−1.6186	−1.358	−1.3018	−2.4767	(−1.1704)
RR									2.1331 ***	0.7821 ***	2.1247 ***	2.1787 ***	2.1909 ***	2.1316 ***	0.7700 ***
									−39.3011	−15.5142	−39.9126	−26.903	−22.2441	−39.0061	−15.2074
DY					1.7897		2.4522		0.3895	−0.0536	0.5451	−0.0375	−0.2227	0.6094	−0.0208
					−1.0988		−1.189		−0.8877	(−0.2562)	−1.3048	(−0.0691)	(−0.3322)	−1.4001	(−0.0997)
ETR			0.0657	0.0306				−0.1062	−0.0252	0.1197	0.0468	0.1133	0.1186	−0.0428	0.142
			−0.1237	−0.0588				(−0.2017)	(−0.1876)	−0.4861	−0.3405	−0.5655	−0.4752	(−0.3035)	−0.5911
N_ME		−0.0071		−0.0614 *	−0.0079		−0.0174	−0.0381	0.0051				0.0429
		(−0.4433)		(−2.4623)	(−0.3845)		(−0.6273)	(−1.6427)	−0.9308				−1.374
A_ME		−0.0124		−0.0284 †					0.0046
		(−1.4029)		(−1.9427)					−1.4083
PM_ME
N_MME	−0.0475					−0.0456 †								0.0504
	(−1.5251)					(−1.7153)								−1.3884
AS_MME	−0.0236 †			−0.0347 *			0.012	−0.006			0.0199 *
	(−1.7712)			(−2.1418)			−0.6639	(−0.5191)			−2.0521
L_MME		0.0157	0.0065
		−0.7322	−0.309
AT_MME			0.0023	−0.0031		0.0146	0.0016	0.0013
			−0.0997	(−0.1386)		−0.6796	−0.0565	−0.057
A_MME									−0.0017	−0.0063 †					0.0141
									(−0.5031)	(−1.8758)					−0.7953
FA	−0.0095		−0.0047				−0.0039			0.0007					−0.0008
	(−1.2320)		(−0.5236)				(−0.3822)			−0.2628					(−0.2928)
FM		−0.01		−0.0082	−0.0048			−0.0073
		(−1.0800)		(−0.8752)	(−0.5039)			(−0.7868)
CA_A		−0.0067			−0.0397	0.0089	0.0289	0.0004		−0.0154
		(−0.0587)			(−0.2078)	−0.0804	−0.1433	−0.0033		(−0.2660)
AS_MME_sq											−0.0003 †
											(−1.9342)
PA_ME_sq												0
												(−0.3838)
N_ME_sq													−0.0017
													(−1.4558)
N_MME_sq														−0.0029
														(−1.4704)
A_MME_sq															−0.0002
															(−1.1182)
_cons	−0.3007	1.1471	0.0231	2.9192	−0.8479	−0.3816	0.5947	0.2722	−0.6548	0.6900 *	−0.7841	−0.767	−0.7555	−0.7356	0.0414
	(−0.2088)	−0.6597	−0.0148	−1.424	(−0.4237)	(−0.2536)	−0.1929	−0.1651	(−1.0120)	−2.4215	(−1.3289)	(−1.1544)	(−1.0217)	(−1.2232)	−0.0725
F statistic	1.2821	0.7003	0.8775	1.7795 †	0.6572	2.0541 *	2.3923 *	1.8287 †	212.2975 ***	34.4258 ***	241.3905 ***	88.2780 ***	69.7304 ***	234.9625 ***	35.5114 ***
R-sq	0.0453	0.0313	0.0571	0.1251	0.0369	0.1146	0.228	0.1425	0.9676	0.488	0.9666	0.8502	0.8567	0.9657	0.4949
Obs	141	137	125	122	128	136	92	121	82	298	85	150	115	85	299
Mean VIF	1.81	1.23	1.56	1.79	1.15	2.43	2.11	2.34	2.47	1.83	5.93	5.01	6.33	7.96	9.38

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Models with effects are also constructed to test whether the results are robust and do not depend on assumptions regarding the correlation of unobserved effects with the explanatory variables. In other words, both fixed effects and random effects models are calculated to verify the robustness of the results. In simple terms, a fixed effects model examines only what changes within each company, while a random effects model also accounts for differences between companies, assuming these differences do not influence the variables.

Thus, calculating both models and checking the consistency of results provides greater confidence in the validity of the conclusions. As shown in Table 8, Table 9, Table 10 and Table 11, the same models are presented again, but the appropriate model is selected based on the Hausman test. Regarding ROA, Random Effects models appear 22 times, while Fixed Effects models appear 8 times. For ROE models, Random Effects models appear 27 times, whereas Fixed Effects models are present only 3 times.

Table 8. Results of the fixed and random effects regression models on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROA (1).

Models	−1	−4	−6	−8	−10	−12	−14	−16	−17	−20	−22	−24	−26	−28	−30
DSG	0.0028	0.0007	0.0004	0.0008	0.0006	0.0003	0.0005	0.0003	0.0007	0.0006	0.0003	0.0005	0.0002	0.0001	0
	−1.3009	−0.4188	−0.2748	−0.8056	−0.4891	−0.213	−0.3704	−0.2441	−0.4312	−0.375	−0.2568	−0.4686	−0.1419	−0.0798	−0.0405
PA_ME	0.0003									0.0011	0.0010 †	0.0012 †	0.0005	0.0011 †	0.0010 †
	−0.4158									−1.3922	−1.7676	−1.7683	−0.7541	−1.9418	−1.9331
NA	0	−0.0000 *	−0.0000 **	0	0	0	0	−0.0000 †	0	−0.0000 *	0	−0.0000 *	0	0	−0.0000 *
	(−0.2294)	(−2.1842)	(−2.6485)	(−1.2730)	(−1.2763)	(−0.9877)	(−0.7330)	(−1.9402)	−0.2558	(−2.2650)	(−1.3501)	(−2.1471)	(−1.1648)	(−0.7175)	(−2.2199)
FCF	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 **	0.0000 **	0.0000 **	0.0000 **	0.0000 **	0.0000 ***	0.0000 **	0.0000 **	0.0000 **	0.0000 **	0.0000 **
	−6.261	−5.3771	−5.1527	−3.5042	−2.9991	−3.0725	−3.1492	−3.1462	−4.0342	−5.0556	−2.9913	−3.0902	−3.0295	−3.2082	−3.1561
SIZE	0.0000 ***	0	0	0	0	0	0	0	0.0000 *	0	0	0	0	0	0
	−4.4994	(−0.9814)	(−1.0695)	(−0.6482)	(−0.3051)	(−0.7334)	(−0.5232)	(−0.2925)	−2.6181	(−1.0169)	(−0.5257)	(−0.3612)	(−0.6669)	(−0.8173)	(−0.4950)
RR	0.0348 **	0.0479 ***	0.0478 ***	0.0286 ***	0.0279 **	0.0279 **	0.0270 **	0.0285 ***	0.3347 ***	0.0478 ***	0.0274 **	0.0282 **	0.0277 **	0.0262 **	0.0281 **
	−2.9907	−3.3167	−3.3811	−3.6713	−3.1607	−3.2166	−3.0604	−3.2971	−6.6054	−3.3093	−3.1158	−3.2586	−3.1668	−2.9828	−3.2815
DY	0.0936	0.1769	0.2003 †	0.1280 *	0.1233	0.1316 †	0.1136	0.1504 *	0.0561	0.1916 †	0.1532 *	0.1487 †	0.1395 †	0.1420 †	0.1783 *
	−0.8226	−1.5228	−1.8029	−2.002	−1.5754	−1.7672	−1.4933	−1.9604	−0.4097	−1.6694	−1.9751	−1.9407	−1.8486	−1.9075	−2.3633
ETR	−0.0313	−0.008	−0.006	0.0141	0.0149	0.019	0.0157	0.0122	0.0051	−0.0051	0.0163	0.0118	0.0176	0.0174	0.0126
	(−1.4372)	(−0.2594)	(−0.1985)	−0.8614	−0.7557	−0.9679	−0.8072	−0.6298	−0.0269	(−0.1645)	−0.8192	−0.6126	−0.884	−0.8859	−0.6511
N_ME		−0.0005								−0.0008
		(−0.2940)								(−0.4308)
A_ME			−0.0014 †
			(−1.6914)
PM_ME				0.0005
				−1.0891
N_MME					−0.0013						−0.0017
					(−0.6743)						(−0.8920)
AS_MME						0.0013 †							0.0009
						−1.9237							−1.1215
L_MME							−0.0022							−0.0026
							(−1.2040)							(−1.5096)
AT_MME								−0.0013							−0.0014 †
								(−1.5915)							(−1.6903)
A_DIS									−0.0008
									(−0.1194)
A_MME												0.0004
												−0.6007
_cons	−0.3231	0.0287	0.1294	−0.0104	0.0342	0.0393	0.0578	0.0649	−0.1123	−0.007	0.0213	−0.04	0.0314	0.049	0.0427
	(−1.4793)	−0.1867	−0.8651	(−0.1126)	−0.2761	−0.3305	−0.4675	−0.5897	(−0.6767)	(−0.0460)	−0.176	(−0.3116)	−0.2604	−0.411	−0.3965
F statistic/wald chi2	31.93 ***	104.68	26.54	90.12	85.12	12.59	23.5	45.69	109.09 ***	22.33	39.1	81.03	24.8	25.5	152.65
R-sq	0.757								0.9876
Obs	149	114	117	132	85	85	85	87	36	114	84	87	84	84	86
R-sq overall	0.2289	0.3936	0.4363	0.4946	0.5759	0.6265	0.5853	0.5987	0.4417	0.4321	0.6267	0.6173	0.6358	0.6415	0.6453
Prob > chi2	0.0344	0.089	0.0935	0.2743	0.6333	0.1416	0.6658	0.5778	0	0.1095	0.2976	0.3206	0.158	0.2884	0.2647
EFECT	FE	RE	RE	RE	RE	RE	RE	RE	FE	RE	RE	RE	RE	RE	RE

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 9. Results of the fixed and random effects regression models on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROA (2).

Models	−31	−34	−35	−38	−39	−41	−44	−46	−48	−50	−52	−53	−56	−58	−59
DSG	0.0045 **	0.0009	0.0002	−0.0001	0.0041 †	0.0034 *	0.0018	−0.0001	0.0005		−0.0004	0.0027	0.0007	0.0004
	−2.9236	−0.9857	−0.1317	(−0.1318)	−1.814	−2.1672	−1.4701	(−0.1206)	−0.3918		(−0.3995)	−1.2349	−0.45	−0.3296
PA_ME	−0.0013		−0.0033 †		0.0006	−0.0017						0.0041
	(−0.8323)		(−1.8720)		−0.3939	(−1.1686)						−0.5641
NA				−0.0000 †		0	0	0	0		−0.0000 †	0	−0.0000 *	0	0
				(−1.8792)		(−1.1078)	(−0.9366)	(−0.9697)	(−1.4869)		(−1.8157)	(−0.1945)	(−1.9737)	(−1.0434)	−1.4087
FCF			0.0000 *	0.0000 ***		0.0000 *	0.0000 ***	0.0000 ***	0.0000 **	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
			−2.0404	−4.4115		−2.5643	−4.5023	−4.1129	−3.0065	−5.7664	−3.348	−6.1693	−5.4591	−3.3467
SIZE			0		0.0000 ***	0	−0.0000 †	0	0	−0.0000 †	0	0.0000 ***	0	0	0
			−1.0749		−9.5594	−0.6975	(−1.7390)	(−1.0279)	(−0.4708)	(−1.6903)	(−0.1601)	−4.4961	(−1.1963)	(−0.7386)	−1.6045
RR									0.0266 **	0.0134 ***	0.0260 **	0.0346 **	0.0478 ***	0.0255 **	0.0125 ***
									−2.9704	−5.0271	−3.1558	−2.9609	−3.308	−2.975	−3.9481
DY					0.2364		0.1809 *		0.1590 *	0.0531 **	0.1414 *	0.0836	0.164	0.1106	0.0554 *
					−1.6638		−2.3855		−1.9916	−3.2836	−2.0847	−0.7208	−1.4204	−1.4869	−2.4198
ETR			0.0188	0.0212				0.0196	0.0143	0.0094	0.0134	−0.0326	−0.0054	0.0205	0.0236
			−1.0728	−1.3093				−1.1757	−0.7247	−0.7463	−0.6961	(−1.4782)	(−0.1753)	−1.0558	−1.6493
N_ME		−0.0005		0.0013	−0.0039		0.0006	0.0015	−0.0014				−0.0086
		(−0.3627)		−0.8834	(−0.8589)		−0.4432	−1.1398	(−1.0432)				(−1.3316)
A_ME		−0.0008		−0.0002					−0.0015 *
		(−1.1368)		(−0.2821)					(−1.9853)
PM_ME
N_MME	−0.0056 †					−0.0038 *								−0.0197 **
	(−1.9391)					(−1.9953)								(−2.7735)
AS_MME	−0.0007			0.0012			0.0008	0.0016 *			−0.0043 *
	(−0.2618)			−1.3395			−0.8798	−2.4804			(−2.2263)
L_MME		0.0003	−0.0019
		−0.2205	(−0.6063)
AT_MME			0.0002	−0.0012		0.0013	−0.0018	−0.001
			−0.0509	(−0.9490)		−0.4076	(−1.3996)	(−0.8188)
A_MME									0.0008	−0.0003					0.0151 **
									−1.0698	(−0.9420)					−2.6432
FA	−0.0005		0.0056 †				−0.0003			−0.0001					0.0003
	(−0.1709)		−1.8706				(−0.5286)			(−0.2311)					−0.2567
FM		−0.0013		−0.0008	−0.0018			−0.0007
		(−1.5676)		(−1.2020)	(−0.3200)			(−1.1278)
CA_A		−0.0018			−0.0009	−0.0011	−0.0115	−0.0023		−0.0002
		(−0.5322)			(−0.0306)	(−0.3411)	(−1.5081)	(−0.7193)		(−0.0894)
AS_MME_sq											0.0001 **
											−3.0926
PA_ME_sq												0
												(−0.5215)
N_ME_sq													0.0003
													−1.2926
N_MME_sq														0.0010 **
														−2.6796
A_MME_sq															−0.0001 *
															(−2.4839)
_cons	−0.2035	0.1119	0.0006	0.1239	−0.4147	−0.1316	−0.075	0.1002	0.0859	0.1020 ***	0.1554	−0.3922	0.0634	0.1189	−0.5010 *
	(−1.0765)	−1.1098	−0.0029	−0.9773	(−1.4422)	(−0.8187)	(−0.5907)	−1.1859	−0.6393	−3.8191	−1.4149	(−1.5301)	−0.4113	−0.9949	(−2.4205)
F statistic/wald chi2	2.68 ***	140.2	2.35 ***	28.36	16.46 ***	3.57 ***	15.64	21.56	38.1	64.7	24.17	28.16 ***	46.88	64.72	4.89 ***
R-sq	0.1361		0.2191		0.6188	0.2679						0.7578			0.1768
Obs	142	138	125	122	127	136	92	121	82	297	85	149	114	85	298
R-sq overall	0.0153	0.094	0.0188	0.4267	0.1182	0.1311	0.5294	0.4549	0.6343	0.3576	0.7128	0.2169	0.4327	0.6541	0.0238
Prob > chi2	0.0012	0.1126	0.0197	0.0579	0.003	0.0488	0.2945	0.4764	0.2393	0.4349	0.1309	0.0385	0.0989	0.8409	0.0205
EFECT	FE	RE	FE	RE	FE	FE	RE	RE	RE	RE	RE	FE	RE	RE	FE

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 10. Results of the fixed and random effects regression models on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROE (1).

Models	−2	−4	−6	−8	−10	−12	−14	−16	−17	−20	−22	−24	−26	−28	−30
DSG	0.0072	0.0041	0.0051	0.0029	0.0037	0.0037	0.0039	0.0045	0.0012	0.0042	0.0031	0.0035	0.0034	0.0033	0.0041
	−0.8812	−0.4111	−0.5453	−0.3193	−0.4887	−0.4888	−0.5278	−0.657	−0.1934	−0.4194	−0.4214	−0.5332	−0.4594	−0.4629	−0.6176
PA_ME	−0.0002									−0.0023	−0.0009	−0.0018	−0.0024	−0.0003	−0.0011
	(−0.0565)									(−0.5041)	(−0.2542)	(−0.4470)	(−0.5525)	(−0.0852)	(−0.3074)
NA	0	0.0000 †	0.0000 *	0	0.0000 **	0.0000 ***	0.0000 ***	0.0000 ***	0	0.0000 †	0.0000 **	0.0000 ***	0.0000 ***	0.0000 ***	0.0000 ***
	−0.7583	−1.8432	−2.0081	−0.4725	−2.7725	−3.4013	−3.3458	−3.4129	(−0.3054)	−1.8631	−2.9889	−3.6901	−3.5088	−3.5204	−3.6589
FCF	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 **	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	0	−0.0000 **	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***
	(−3.6034)	(−3.3491)	(−3.3170)	(−2.7621)	(−4.1245)	(−4.1641)	(−4.1282)	(−4.2115)	(−0.9898)	(−3.2756)	(−4.2655)	(−4.3767)	(−4.3045)	(−4.2635)	(−4.3510)
SIZE	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
	−0.1216	(−0.1569)	(−0.1170)	−0.0461	−0.6871	−0.3872	−0.3863	−0.6175	(−1.2662)	(−0.1416)	−0.7604	−0.6844	−0.4567	−0.4366	−0.6884
RR	1.9519 ***	1.9635 ***	1.9704 ***	2.0175 ***	2.0452 ***	2.0417 ***	2.0401 ***	2.0442 ***	0.8544 ***	1.9651 ***	2.0490 ***	2.0486 ***	2.0461 ***	2.0433 ***	2.0480 ***
	−29.4303	−25.3867	−25.7588	−29.6476	−41.6859	−41.8536	−41.4516	−42.3449	−4.3738	−25.2566	−41.9642	−42.6792	−42.0981	−41.6715	−42.597
DY	0.3	0.2174	0.1546	0.2743	0.4456	0.5095	0.5009	0.483	1.1780 *	0.1816	0.4594	0.4549	0.5135	0.5287	0.4886
	−0.5551	−0.3116	−0.226	−0.4582	−0.9826	−1.1332	−1.142	−1.0594	−2.2335	−0.2586	−1.0189	−1.0216	−1.162	−1.213	−1.0822
ETR	0.0244	0.0442	0.0349	−0.0223	−0.0515	−0.029	−0.0287	−0.0454	1.0754	0.04	−0.0521	−0.0393	−0.0265	−0.0271	−0.0444
	−0.1832	−0.2801	−0.2235	(−0.1635)	(−0.4813)	(−0.2726)	(−0.2691)	(−0.4357)	−1.4715	−0.2517	(−0.4854)	(−0.3763)	(−0.2468)	(−0.2535)	(−0.4245)
N_ME		−0.0011								−0.0007
		(−0.0933)								(−0.0586)
A_ME			0.0042
			−0.8227
PM_ME				0.0061
				−1.3391
N_MME					0.0054						0.0063
					−0.4554						−0.5426
AS_MME						0.0032							0.0037
						−0.7512							−0.7346
L_MME							−0.0102							−0.0095
							(−0.9743)							(−0.9082)
AT_MME								0.0015							0.0019
								−0.2767							−0.3611
A_DIS									−0.0153
									(−0.6296)
A_MME												−0.0014
												(−0.3680)
_cons	−0.5394	−0.3004	−0.6446	−0.3151	−0.4334	−0.4641	−0.356	−0.4811	0.0733	−0.2135	−0.3651	−0.229	−0.3637	−0.3163	−0.4219
	(−0.6778)	(−0.3121)	(−0.6734)	(−0.3490)	(−0.5971)	(−0.6350)	(−0.4948)	(−0.7163)	−0.1158	(−0.2177)	(−0.5088)	(−0.3002)	(−0.5056)	(−0.4452)	(−0.6332)
F statistic/wald chi2	85.12	86.12	86.02	85.62	97.11	95.27	102.06	101.43	17.49 ***	87.34	65.36	73.35	68.74	57.73	96.85
R-sq									0.921
Obs	150	115	118	132	85	85	85	87	37	115	84	87	84	84	86
R-sq overall	0.8441	0.8432	0.8481	0.8566	0.9618	0.9622	0.9647	0.9616	0.1984	0.8454	0.9634	0.9632	0.9644	0.9659	0.9633
Prob > chi2	0.0749	0.6397	0.4562	0.2905	0.7596	0.4154	0.5927	0.6333	0.0112	0.6213	0.6946	0.5399	0.4563	0.6101	0.714
EFECT	RE	RE	RE	RE	RE	RE	RE	RE	FE	RE	RE	RE	RE	RE	RE

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01. * for 0.05, ^† for 0.10.

Table 11. Results of the fixed and random effects regression models on the relationship between the ethnic diversity of employees and managers and firm performance, measured by ROE (2).

Models	−32	−34	−36	−38	−40	−42	−44	−46	−48	−49	−52	−54	−56	−58	−59
DSG	0.0048	0.0023	−0.0004	−0.0058	0.0079	0.0077	0.0023	0.0012	0.0022		0.0042	0.0074	0.004	0.0037
	−0.2496	−0.126	(−0.0224)	(−0.2886)	−0.3163	−0.3793	−0.0681	−0.0576	−0.2987		−0.5597	−0.9062	−0.4021	−0.4877
PA_ME	0.022		0.01		0.002	0.0036						0.0043
	−1.3909		−0.6991		−0.1473	−0.2628						−0.3004
NA				0		0	0.0000 ***	0.0000 †	0.0000 **		0.0000 ***	0	0.0000 †	0.0000 **	0.0000 ***
				−1.274		−1.5516	−3.3428	−1.9447	−3.0594		−3.4478	−0.7213	−1.7904	−2.7299	−3.8304
FCF			0	0		0	0	0	−0.0000 ***	0	−0.0000 ***	−0.0000 ***	−0.0000 ***	−0.0000 ***
			−0.9699	−1.0904		−1.5133	−0.6513	−1.4767	(−4.2284)	−1.3314	(−4.1457)	(−3.5629)	(−3.3458)	(−4.0943)
SIZE			0		0	0	0	0	0	0	0	0	0	0	0
			(−0.1382)		−0.7254	(−0.9941)	(−0.5262)	(−1.0891)	−0.7259	−0.2281	−0.2134	−0.0804	(−0.1030)	−0.6915	−0.4444
RR									2.0515 ***	0.0502	2.0447 ***	1.9509 ***	1.9639 ***	2.0456 ***	0.1154 †
									−41.2989	−0.8142	−41.6003	−29.3419	−25.2646	−41.2876	−1.8613
DY					1.799		2.1515		0.402	0.0423	0.516	0.3016	0.2383	0.4489	0.0198
					−1.1147		−1.0163		−0.8653	−0.0903	−1.1455	−0.5563	−0.3387	−0.98	−0.0437
ETR			0.0293	−0.0093				−0.0194	−0.0467	−0.1349	−0.0251	0.0245	0.042	−0.0531	−0.081
			−0.076	(−0.0236)				(−0.0486)	(−0.4357)	(−0.4456)	(−0.2338)	−0.1836	−0.2644	(−0.4900)	(−0.2892)
N_ME		0.0054		−0.0424	−0.0011		−0.0299	−0.0219	0.0068				0.0113
		−0.2062		(−1.2341)	(−0.0360)		(−0.7635)	(−0.6779)	−0.8392				−0.2832
A_ME		−0.0093		−0.0242					0.0027
		(−0.6924)		(−1.2016)					−0.6165
PM_ME
N_MME	−0.0205					−0.0104								0.0106
	(−0.5849)					(−0.3434)								−0.2399
AS_MME	−0.0224			−0.0292			−0.0009	−0.0076			0.0125
	(−1.1730)			(−1.3612)			(−0.0373)	(−0.4872)			−0.8705
L_MME		0.0227	−0.0011
		−0.7315	(−0.0348)
AT_MME			−0.0001	−0.0037		0.0045	−0.005	−0.0002
			(−0.0045)	(−0.1272)		−0.1569	(−0.1389)	(−0.0068)
A_MME									−0.0036	0.0018					0.1598
									(−0.8420)	−0.1336					−1.4223
FA	−0.0149		−0.0057				−0.0066			0.0108					0.0139
	(−1.1097)		(−0.3919)				(−0.3921)			−0.4661					−0.6185
FM		−0.0183		−0.0143	−0.0089			−0.0133
		(−1.1544)		(−0.9623)	(−0.5408)			(−0.8860)
CA_A		−0.0102			−0.0442	−0.0026	0.0391	−0.0082		0.0103
		(−0.1380)			(−0.2316)	(−0.0351)	−0.1825	(−0.1050)		−0.1708
AS_MME_sq											−0.0002
											(−0.6784)
PA_ME_sq												0
												(−0.3226)
N_ME_sq													−0.0005
													(−0.3268)
N_MME_sq														−0.0003
														(−0.1220)
A_MME_sq															−0.0011
															(−1.3297)
_cons	0.0446	1.0927	0.1202	3.5841	−0.2606	−0.6044	0.0307	0.9408	−0.2769	−0.1961	−0.6	−0.6599	−0.3515	−0.4533	−6.9160 †
	−0.0233	−0.5244	−0.06	−1.2089	(−0.1032)	(−0.3181)	−0.0086	−0.4464	(−0.3498)	(−0.1541)	(−0.7916)	(−0.7467)	(−0.3597)	(−0.6041)	(−1.7019)
F statistic/wald chi2	2.65	2.15	46.25	86.57	22.2	11.85	26.54	35.63	75.25	1.34 ***	85.79	74.47	47.26	74.27	1.98 ***
R-sq										0.0148					0.0797
Obs	141	137	125	122	128	136	92	121	82	298	85	150	115	85	299
R-sq overall	0.0333	0.0253	0.0524	0.1126	0.032	0.0821	0.2177	0.1147	0.9637	0.0456	0.9635	0.8443	0.8445	0.962	0.0836
Prob > chi2	0.9352	0.9699	0.9918	0.903	0.9421	0.9705	0.9483	0.9776	0.4214	0	0.4283	0.1035	0.6872	0.6943	0
EFECT	RE	RE	RE	RE	RE	RE	RE	RE	RE	FE	RE	RE	RE	RE	FE

Source: Own estimates. *** highlight statistical significance for a threshold of 0.001. ** for 0.01, ^† for 0.10.

5. Discussion

The econometric analysis highlights significant relationships between ethnic diversity, workforce structure, and firm profitability, as reflected in the results presented in Table 4 and Table 5. The share of employees belonging to ethnic minorities shows a positive and highly significant influence on ROA, with statistical significance at the 0.01 ** and 0.001 *** levels, suggesting that a high level of ethnic diversity among employees is correlated with superior performance. The study by Altinay (2010) argues that even firms owned by ethnic minorities adapt to market changes and to the behavior of co-ethnic consumers, also showing that a higher share of employees belonging to ethnic minorities is significantly associated with superior financial performance, indicating that workforce ethnic diversity can be a key factor in increasing firms’ competitiveness and success.

Managerial diversity also has a positive impact on performance. The presence of managers from ethnic minorities, particularly those of Asian origin, is positively associated with ROA, confirming hypothesis H1 and highlighting the role of diversity at the decision-making level. In contrast, other indicators of ethnic diversity exhibit negative relationships with ROA. For example, the proportion of African American employees is negatively associated with ROA and is statistically significant at the 0.05 significance level, while the variable referring to White employees shows a negative and highly significant relationship. This finding is also reflected in the study by Tran and Krueger (2025), who argue that ethnic diversity may be associated with certain disadvantages that affect available resources and financial performance, thereby rejecting hypothesis H1.

Following the influence of ethnic diversity among employees, attention turns to managerial diversity, focusing on individuals who hold key positions within the company and are responsible for achieving profitability and high performance. Accordingly, the variables representing Hispanic/Latino managers and managers from other minority groups show negative effects on ROA, with statistical significance at the 0.05 level. For African American managers, the linear relationship displays a mixed effect, sometimes positive, sometimes negative, but remains statistically significant, while the squared term (N_MME-SQ) reveals a positive nonlinear effect, suggesting that their influence becomes favorable at higher levels. Here, the inflection point is identified at a value of 9.455121; thus, in general, a 1% increase in the variable representing African American managers leads statistically to an increase of 0.0043 in the dependent variable. At the same time, the squared term indicates the existence of a U-shaped nonlinear relationship, with the level at which the marginal effect changes sign and becomes positive. The results reveal a nonlinear relationship between the share of African American managers and ROA: below the 9.45% threshold, the effect is negative, while above this level, the impact becomes positive, indicating a minimum representation level at which managerial diversity contributes to firm performance.

With regard to nonlinear regressions, the analysis of the squared terms indicates the presence of relevant nonlinear relationships. The N_ME-SQ term shows a positive effect, signaling that at higher levels of African American employee representation, the impact on performance becomes favorable. Thus, from a statistical perspective, the effect is negative at low levels due to the calculated inflection point, but it turns positive after reaching the 15.34% threshold. By contrast, the squared terms for PA_ME and A_MME are not significant, indicating the absence of nonlinear effects in these cases.

On the other hand, several variables do not exhibit a significant influence on ROA. Gender wage gaps the share of female employees, the share of female managers, the number of announced layoffs and employees with disabilities do not exert statistically significant effects in this study, suggesting that these dimensions are not directly correlated with asset performance.

Regarding the financial control variables, the results are consistent with the existing literature. Free cash flow shows a strong and highly significant positive effect, highlighting the importance of internal financial resources. The reinvestment rate positively affects ROA and is significant at the 1% level, indicating the role of profit reinvestment in enhancing efficiency. Dividend yield also has a positive and statistically significant effect. In contrast, firm size is negatively correlated with ROA, and the effective tax rate exhibits a significant negative effect, indicating that higher tax pressure reduces asset profitability.

Overall, the results demonstrate that ethnic diversity has a complex impact on organizational performance, with positive effects for certain minority groups and managers, as well as negative relationships in other cases. Therefore, these findings resonate with the broader literature on ethnic diversity, which highlights the financial effects of diversity. For example, Frijns et al. (2016) found that cultural diversity, stemming from ethnic diversity, in corporate boards negatively affects firm performance in certain contexts. In contrast, other research, such as Covington et al. (2025), reports positive outcomes related to diversity, suggesting that the benefits of diversity may manifest through more complex mechanisms and emphasizing that promoting an inclusive work environment can enhance a firm’s innovation capacity, offering practical implications for both decision-makers and corporate leaders. Moreover, Saha et al. (2024) discuss the effects of diversity on the workforce and management, highlighting that the impact of diversity depends on institutional, structural, and contextual factors, aligning with our findings that some diversity measures were insignificant or nonlinear.

Table 8 and Table 9 present the fixed-effects and random-effects regression models, which are econometric specifications that generate important changes in both the statistical significance and the direction of the variables’ effects, indicating that firm-level heterogeneity plays an essential role in explaining performance differences. First, the gender wage gap becomes positive and statistically significant, suggesting that wage structure is associated with improved ROA. At the same time, several variables that were significant in the pooled models, such as the effective tax rate, the proportion of African American employees, the proportion of Hispanic/Latino managers, and the share of female managers, lose their statistical significance. By contrast, some variables strengthen their influence: the proportion of female employees and the share of White managers within ethnic minorities show robust positive effects on performance. Thus, hypothesis H2, which argues that the presence of female employees and female managers positively influences performance, is confirmed.

Table 6 and Table 7 show that ethnic diversity and workforce structure significantly influence economic performance. Regarding diversity-related variables, the results are heterogeneous. Gender wage differences between women and men show positive and statistically significant effects on ROE, indicating that firms with a larger gender pay gap tend to record higher levels of profitability.

Variables related to employee diversity, such as the share of ethnic minorities and the proportion of African American employees, exert a significant negative effect on ROE. This result may indicate that firms with a high proportion of ethnic minority employees operate in sectors with lower profit margins or in regions where the labor force structure differs from that of more profitable industries. In contrast, the positive association between a higher share of White employees and ROE suggests that these firms are active in more economically advantageous business environments.

With regard to management, most variables describing ethnic diversity at the managerial level indicate a significant negative effect on performance: African American managers, Asian managers, the proportion of White managers, and Hispanic/Latino managers are negatively correlated with ROE, thus rejecting H1. These results suggest that the ethnic composition of management is associated with firms’ industrial profiles or specific operational characteristics, rather than with a direct relationship between diversity and financial performance.

These findings indicate that the type of managers a firm employs depends more on the sector in which it operates and on how it is organized, rather than implying that ethnic diversity among managers directly influences firm profitability. In other words, it is not diversity per se that affects performance, but rather the sectoral context and firm-specific characteristics in which these managers operate. This interpretation is also supported by the study of Ibrahim et al. (2025), which shows that gender inequalities in the healthcare sector are not driven solely by individual competence or productivity, but by organizational structure, and that the real impact of diversity depends more on organizational and sectoral context than on individual characteristics.

Regarding organizational size and structure, the variable capturing the total number of employees is positively and significantly associated with ROE, suggesting that larger firms benefit from advantages and economies of scale. The study by Ma and Zhang (2023) on the digital economy is related to organizational size and structure through a similar mechanism: just as larger firms with more employees can benefit from economies of scale and higher financial performance, the digital economy can also generate higher revenues and reduce inequalities. These effects may stem from strong managerial capabilities, better access to resources and financing, greater bargaining power and market influence, and, importantly, enhanced innovation and adaptability. With a larger workforce, firms can develop dedicated teams for innovation and digitalization, while innovation and flexibility increase competitiveness, revenues, and consequently ROE.

With respect to control variables, the results show that firm size is positively associated with ROE, whereas free cash flow has a significant negative effect on ROE. This may indicate that firms accumulating excessive liquidity face difficulties in identifying profitable investment opportunities, implying inefficient use of available funds. The reinvestment rate is positively associated with ROE, confirming the importance of reinvestment strategies in generating superior performance. The effective tax rate negatively affects ROE, as higher taxes reduce net profits.

Variables such as female representation, employees with disabilities, other minority managers, dividend yield, and announced layoffs do not significantly affect ROE. Regarding nonlinear regressions, the results indicate the presence of a nonlinear relationship only for Asian managers. Specifically, as the proportion of Asian managers begins to increase, firm profitability initially tends to rise. For all other variables analyzed in quadratic form, the results do not reveal clear or significant patterns, indicating the absence of consistent nonlinear relationships in those cases.

In this way, the mixed relationships between diversity indicators and return on equity (ROE) align even more closely with the broader literature. Some studies report positive associations between diversity efforts and shareholder value or performance metrics, while others find limited effects, such as Filbeck et al. (2017), who argue that diversity may be associated with immediate benefits on specific accounting performance measures (ROA), but its long-term effect on ROE is unclear.

The analysis of the new fixed-effects and random-effects results for ROE, presented in Table 10 and Table 11, shows that many of the studied variables do not have a clear influence on firm profitability. Among the variables for which no impact is observed are the ethnic and gender composition of employees and managers. By contrast, dividend yield now exhibits a positive effect, indicating that firms offering more attractive dividends tend to be more profitable. The remaining variables retain their previous influences: the gender wage differential and the share of White employees are associated with higher ROE, while the share of ethnic minorities among employees contributes positively to ROE, and the effective tax rate remains a factor that significantly reduces profitability. In addition, the squared term for Asian managers indicates a nonlinear relationship in which the initial positive effects reverse beyond a certain level.

These interpretations support the research hypotheses regarding the effect of diversity but also reveal that diversity does not influence performance uniformly across all contexts. The results directly address the research question by demonstrating that ethnic diversity among employees and managers has both positive and negative effects on firm performance, and that the hypothetical effects proposed in H1 and H2 are supported for certain measures of diversity. The nonlinear effects further highlight that simple linear hypotheses may not be sufficient.

This study contributes to the existing diversity literature by showing that the impact of diversity is not uniform, and the significance of nonlinear relationships implies that firm performance can improve only beyond certain representation thresholds, thereby adding robustness to the assessment of diversity effects and providing new perspectives.

The impact of the analysis is reinforced through the inclusion of Table 12, which presents the influence of the independent variables on return on assets (ROA) and return on equity (ROE). This table provides a comparative analysis of the two profitability measures, showing the effect of each variable on the dependent variable, thereby contributing to a more concise and clear presentation of the results.

Table 12. Influence of the variables.

Variable	ROA	ROE
Ethnic minority employees (PA_ME)	+	−
African American employees (N_ME)	−/+ (positive after 15.34%)	−
White employees (A_ME)	−	+
Ethnic minority managers (PM_ME)	+	+
African American managers (N_MME)	−/+ (positive after 9.45%)	−
Asian managers (AS_MME)	+	+
Hispanic/Latino managers (L_MME)	−	−
Managers from other minority groups (AT_MME)	−	−
White managers (A_MME)	+	−
Female employees (FA)	+	+
Female managers (FM)	+	+
Gender wage gap (DSG)	+	+
Free cash flow (FCF)	+	−
Reinvestment rate (RR)	+	+
Dividend yield (DY)	+	+
Firm size (SIZE)	−	+
Effective tax rate (ETR)	−	−

Source: Own estimates.

6. Conclusions

The econometric results confirm the formulated hypotheses and highlight the role that ethnic diversity and workforce structure play in relation to firm performance.

Hypotheses confirmation: Ethnic diversity and workforce structure significantly influence firm performance.

• Impact on ROA:
  • Proportion of ethnic minority employees and managerial diversity positively influence performance.
  • African American and White employees show negative effects, indicating industry-specific dynamics.
• Impact on ROE:
  • Gender pay gaps and proportion of white employees positively associated with profitability.
  • Most ethnic managerial groups negatively affect ROE, reflecting industry-specific factors.

Overall, the results emphasize that ethnic diversity impacts firm performance in a complex manner, highlighting the importance of understanding workforce structure when designing strategies to optimize performance and sustainability. The analysis of panel regression models over a five-year period yields relevant conclusions that support the hypotheses formulated at the beginning of the study. The results demonstrate that the objectives have been achieved and are robust both theoretically and econometrically. Compared to the existing literature, our study produces findings similar to previous research while clearly explaining how the analyzed variables influence performance indicators.

Limitations: random selection of companies may affect generalizability.

The relationship between ethnic diversity and profitability is essential for investors and other stakeholders, as it allows for a more accurate assessment of performance and supports better-informed decision-making. For future research, the study could be expanded by introducing new variables, such as macroeconomic indicators or an aggregated diversity index calculated through principal component analysis, enabling the development of additional linear and nonlinear regression models. Moreover, future analyses could include a larger number of companies and consider the industry classification of firms. Companies should clearly define how and in what ways ethnic diversity influences their operations. These aspects are relevant for both the academic and business communities, providing insights that help improve decision-making and financial performance in the context of minority diversity. Another direction for future research would be to extend the study into a more complex analysis that considers a global context, including a broader range of companies with international operations across global markets. Such an approach could highlight the advantages of cultural diversity, while also revealing potential disadvantages, such as cultural conflicts, challenges in acceptance, and related issues. Therefore, analyzing the impact of cultural diversity in international markets would represent a valuable next step and an important avenue for future research.

As limitations, it should be noted that company selection was random and that results may be influenced by external factors affecting firm performance. Analyzing the impact of minority diversity on financial performance is essential for understanding how ethnic diversity influences managerial decisions and performance, helping investors, managers, and other stakeholders make better choices. The study also contributes to promoting good corporate governance and social responsibility practices, showing that a careful approach to diversity can reduce risks, enhance reputation, and support sustainable growth in financial performance.

As recommendations for investors and managers, the study suggests considering minority diversity when evaluating firm performance. Investors should integrate these aspects into investment decisions to identify sustainable and profitable opportunities, while managers and decision-makers can incorporate minority diversity into business strategies and risk management. Collaboration between investors and firm leadership is essential for promoting effective practices. From an applied perspective, companies should aim for a balanced representation of women and minorities on boards of directors and in executive leadership positions, reported on an annual basis. In addition, the integration of diversity into key performance indicators (KPIs) can be recommended by including diversity-related indicators in the performance evaluation of executives. Practical mechanisms may also be implemented, such as establishing a diversity committee to monitor these indicators and propose concrete actions.

Conclusion: Integrating minority diversity into corporate strategies enhances firm performance and supports responsible management practices.