“Ethical leadership: a dual path model for fostering ethical voice through relational identification, psychological safety, organizational identification and psychological ownership”

Ethical leadership and ethical voice

Rubin, Dierdorff, and Brown (2010) and Klein and Kozlowski (2000) described the ethical voice as the phenomena linked with a particular group; hence, it could be described as a shared unit property. Later, it was examined that ethical leadership is imperative in generating a group-level ethical voice (Huang & Paterson, 2017). Brown, Treviño, and Harrison (2005) identified and accepted the role of ethical leadership and individual behavior of ethical voice. Zheng et al. (2021) hypothesized the relationship between ethical leadership and ethical voice and explained that ethical leaders have the potential to influence the self-concepts, and interpersonal influence of followers and can affect the ethical voicing behavior of their followers. Voice is explored in empirical studies as the episodic occurrence in specific settings, a specific event and a specific coworker, and through a specific type of leader (Detert & Treviño, 2010). In this regard, it has also been defined that the influence of a leader can also be determined as person-specific. Prior studies have defined ethical voice as the employee’s/follower’s response to showing confidence about suggesting something and/or reporting any misconduct without fearing the situation. Social identity and information processing theories have widely supported the relationship between ethical leadership and ethical voice (Zheng et al., 2021). Based on the prior studies and theories, the formulated hypothesis for the said relationship is as under:

Ethical leadership and organizational identification

Ethical leadership is described as the display of normatively acceptable behavior in the workplace (Brown et al., 2005). Ethical leadership has the potential to create and maintain interpersonal relationships with followers and the propagation of such behavior to followers through two-way contact, reinforcement and decision-making. Kapur (2017) described that an ethical leader’s role is to consider the organization’s expectations, ideas, beliefs, rules and ethics. Ethical leaders are trustworthy role models for their followers, show acceptable actions, and handle them accordingly (Brown et al., 2005). Prior literature established that ethical leadership positively influences followers’ attitudes and behaviors and impacts employees’ identification with their organization (Al-Aidarous, 2021). In the organization, ethical leaders are identified as solid characters who spread compassion and integrity toward employees (Stouten, Van Dijke & De Cremer, 2012). Such types of leaders also create bonding between their followers and the organization. They have the potential to create a positive image of the organization toward their followers through the display of positive behavior, thereby creating organizational identification. Followers of ethical leadership take it as the organization’s initiative, so they tend to develop high organizational identification (Zheng et al., 2021). Social identity theory broadly defines how individual employees at the workplace distinguish themselves from the identity of the organization they belong to (Ashforth & Johnson, 2001). Therefore, we proposed H2 as follows:

Organizational identification and psychological ownership

Haslam & Reicher (2006) defined organizational identification as detrimental outcomes on the attitude and behavior of individuals. Organizational identification is recorded as the important antecedent of individuals losing their distinctiveness and objectivity (Conroy, Henle, Shore, & Stelman, 2017). It creates ownership feelings in the organization’s members, i.e. employees. Existing research implicitly defined organizational identification's positive impact on psychological outcomes in employees. Employees with high organizational identification tend to own their organization and act accordingly. Psychological ownership posits that employees with ownership feelings have a sense of possession of a particular organization; hence, it is identified as a cognitive form, and employees with organizational identification establish connections between themselves and their organization. Psychological ownership allows the owner to regard the target object as a social entity. The current study inferred that organization identification could be affected by stimulation, knowing intimately, self-identity, exercising control and owning a place.

Hence:

Psychological ownership and ethical voice

Psychological ownership is defined as the phenomenon in which an employee develops possessive feelings toward the organization (Van Dyne & Pierce, 2004). Organization-based psychological ownership is linked with the voice regarding any misconduct implemented (Ng, Leung, Chu & Qiao, 2021). This study also reported that employees who felt more contempt for the organization expressed their voices. Some prior research defined that the relationship between psychological ownership and ethical voice is not worked better in isolation; instead, it should be studied through a mechanism, such as Lu, Zhou & Chen, (2019) and Ng, Leung, Chu & Qiao, (2021) did, with an antecedent such as ethical leadership. Zheng et al. (2021) defined the context role of a relationship: an employee with high psychological ownership generates a voice at the workplace when finding any misconduct in the organization. Employees who have personally connected with their work and display more significant effort and energy toward success at work will potentially have a positive attitude and behavior. Hence:

The mediating role of organizational identification and psychological ownership

Zheng et al. (2021) identified that ethical leadership has the potential to evoke identification in followers/employees. Moreover, social identification theory also validated that identifying with an organization or individual at a personal level recorded several positive outcomes. In line with the theory and literature, Markus and Wurf (1987) also identified that any individual self-identities can positively contribute to a particular attitude or behavior. It was also identified that the said attitude or behavior worked through a psychological mechanism such as psychological ownership (Brown et al., 2005). Van Dyne and Pierce (2004) defined that psychological ownership concurs with self-identities, effecting, efficacy motives and psychological ownership. Specifically, psychological ownership aims to satisfy the need to be efficacious and in control of an individual’s environment to produce self-perceived desirable work outcomes (Olckers, 2013) through the self-identity of the organization, and it can create belongingness. Leadership literature explains that leaders can get better results from followers through identification (Malik et al., 2021). Ethical leaders can provide a mechanism to their followers, so they place themselves within their organization and own it. If they feel any misconduct in the organization, they produce a voice and whistleblow about the misconduct. Hence:

Ethical leadership and relational identification

Relational identification is potentially defined in leader-member relationships and as a novel construct (Ashforth et al., 2008). In the context of social identity theory, ethical leaders are the best source of the emergence of relational identification in their followers (Zheng et al., 2021). Under the tenet of social identity theory, it was explained how an individual in the workplace differentiate themselves from the identity of their belongingness (Tajfel & Turner, 2004). Specifically, social identity is “that part of an individual’s self-concept which derives from his knowledge of his membership to a particular leader with emotional significance through a categorization process, where I become us” (Brewer, 1991). Malik et al. (2021) explored relational identification with authentic leadership and identified that it significantly impacts the development of relational identification. Later, Zheng et al. (2021) identified that ethical leaders also could emerge with relational identification in their followers. In this regard, and by using the theoretical lens of social identity theory, the formulated hypothesis is as under:

Relational identification and psychological safety

Wang et al. (2022) investigated the relationship between psychological safety with servant leadership. The relationship between relational aspects and psychological safety was also studied. In the role relationships, it has been identified that relational identification significantly impacts psychological safety. Carmeli, Reiter-Palmon, and Ziv (2010) also investigated and reported that high-quality interrelationships influence psychological safety tendencies. Sluss and Ashforth (2007) presented the concept of relational identification, which has been identified with several psychological outcomes at the employee level.

On the other hand, Jyoti and Bhau (2015) exposed that high-quality leader-member exchange can generate relational identification in the followers so that they feel safe in the organization’s environment and act accordingly. Other researchers broadly discuss the relationship between relational identification and psychological safety. Sluss and Ashforth (2008) also argued that relational identification significantly impacts employees’ cognitive elements. The empirical evidence also suggests that relational identification is an essential determinant of employee creativity through psychological safety (Shen, Yuan, Yi, Liu & Zhan, 2019; Burhan et al., 2023). LMX theory supports the defined relationship, suggesting that employees with high relational identification feel safe in the organization and display positive attitudes and behavior. Hence:

Psychological safety and ethical voice

Social exchange theory (Blau, 1964) defines mutual exchange between organization initiatives and employees might have a positive outcome. Newman, Donohue, and Eva (2017), using the lens of social exchange theory, defined that psychological safety can give better attitudinal and behavioral outcomes at the employee level. Most of the literature that covered employee creativity and innovative behavior identified that psychological safety is the critical determinant of the employee’s positive and extra-role behaviors (Newman et al., 2017). Employees in organizations analyze and understand the situation before speaking. In this regard, psychological safety is observed as the belief that particular behaviors of the employee in the organization, i.e. generating voice, will not lead to personal harm (Detert & Burris, 2007). Edmondson (1999) categorically mentioned that psychological safety is a “shared belief that the team is safe for interpersonal risk-taking”. Therefore, it is clear that psychological safety reflects the organization’s members’ belief that they will not be punished for adverse consequences. This concept is characterized as a scenario generated by trust and mutual respect. People feel comfortable in this scenario (Liu, Liao, & Wei, 2015). Team psychological safety partly mediates the relationship between authentic leadership and internal whistleblowing. Walumbwa and Schaubroeck (2009) established that ethical leaders influence followers’ voice behavior through the mediating influence of psychological safety. Based on the literature, the formulated hypothesis is under:

The mediating role of relational identification and psychological safety

Scholars like Markus and Wurf (1987) initially identified that any self-identities of the individual employee/follower in organizational settings contribute more to adopting a particular behavior. The reason behind adopting a specific behavior is the psychological attachment of the said individuals. Though identity and integration work through the sharing of common traits in interpersonal relations, morality and ethics are directly influenced by closed relationships in the organization (Brown et al., 2005). The followers of ethical leaders with high levels of relational identification will tend to feel secure in the organization, and henceforth, they come up with ethical voicing behavior. Psychological safety is the feeling of an individual defining that the environment is safe for taking personal initiatives (Kahn, 1990). In the organization, psychological safety is a robust influential concept for altering the attitude and behavior of an individual employee. Organizations have a more focused approach to developing feelings of safety or a psychological safety climate in the workplace to achieve better performance from their most important resource, i.e. human resources.

Different researchers determined psychological safety antecedents, e.g. Ashforth et al. (2008) defined that organization identification and psychological safety are correlated. Similarly, Carmeli et al. (2010) focused on relationships with coworkers and supervisors in response to creating a psychological safety scenario where the individual feels free to take the initiative and help exhibiting positive attitudes and behaviors. Newman et al. (2017) identified that due to identification, individuals feel the positive and perceived environment as helpful and positive (Edmondson, 2004). Edmondson (2004) further explored that the psychological safety perception is based upon the qualities of a leader and group membership. Ashforth et al. (2008) relate psychological safety and identification with self-validation. Besides the thoughts about organizational personal, and relational identification, transparency and positive psychological capabilities can also lead to psychological safety. The individual must obtain a psychologically safe scenario through relational identification, which can help them perceive a psychologically safe environment. Role relationship theory, which has the basis for relational identification, can help build the perception of psychological safety (Walumbwa & Schaubroeck, 2009). High-quality relationships work as potential contributors to building a psychologically safe environment. Incorporating the arguments above, the current study proposes a sequential mediation relationship linking ethical leadership and ethical voice through relational identification and psychological safety, as follows:

Research model

Figure 1 of the research model is developed based on the above-stated hypotheses and literature. It displays the relationships among different variables from ethical leadership through ethical voice. The supporting literature and theoretical background provided sufficient support to produce the model presented below:

Research design

In line with the objectives of the current study, i.e. to identify the impact of ethical leadership on ethical voice and the role of relational identification and psychological safety as sequential mediators, the study opted for positivism as a research philosophy, a deductive approach, and a survey method. The impact of selected variables on ethical voice is determined after collecting quantitative data from the employees working in the banking sector. The causality of the variables was checked by running the regression analysis. The current study applied structural equation modeling (SEM-AMOS) techniques to test the formulated hypotheses. Other studies, such as (Mehmood et al., 2018; Shafique et al., 2018; Zheng et al., 2021), took a similar approach to test the phenomena.

Population

Methodological experts such as Sakaran (2003) described that a research population refers to the entire group of events and groups of people where problems have been identified, according to the researcher’s interest. It was also defined that the study population should be the group of individuals through which the study’s results can be generalized (Rice et al., 2020). The current study identified the ethical voice-related issue in the banking sector, so data were collected from the employees in the sector, mainly, banking officers, human resource officers and account opening officers, considered the potential population for the study. The data was collected in Pakistan, a developing country with unethical practices due to vague policies and procedures (Malik et al., 2021). Pakistan is categorized as having a high-power distance culture, hindering the employees’ ability to generate voice.

Moreover, the unemployment rate in Pakistan is significantly increasing, and the market is also dealing with an economic crunch. So, the employees retain their jobs and tend not to generate a voice. The sample size calculated for the study is 248, a moderate sample size for SEM analysis. The sample size was determined by using the G-Power formula. The purposive sampling technique was used as a sampling technique. Overall, 275 questionnaires were distributed among the employees of different branches nationwide. However, only 249 returned, with a response rate of 90%. Out of 249 questionnaires, one had unattended items and duplicate responses. So, for data analysis, 248 questionnaires were found correct in all manners.

Instrumentation

For testing the impact of variables, the Likert insert scale was identified as the best option (Miller and Brewer, 2003). The collection of responses with the help of a questionnaire is an economic tool and is widely used in a similar nature of research. The questionnaire consisted of demographic-related information, and the subsequent section contained the items for assessing the main variables of interest. The respondents rated all the items related to the variables adopted from the already developed questionnaires on a five-point Likert scale where 1 = strongly disagree and 5 = strongly agree or otherwise stated. Ethical leadership was measured through a ten-items scale by Brown et al. (2005). Relational identification-related responses were collected using a ten-items scale by Kark, Shamir, and Chen (2003). Psychological safety was also measured through the adopted questionnaire. This scale consists of six items, including three reverse questions. It is a self-reported questionnaire, and the sample items “My manager often encourages me to take on new tasks or to learn how to do things I have never done before,” “if I have a problem in this company, I could depend on my manager to be my advocate” and “often when I raise a problem with my manager, he/she does not seem very interested in helping me find a solution.” The last item, as stated, is the reverse question; after collecting the response, the researcher treated the said item and the other two questions accordingly. Miller, (2000) scale was used for organizational identification. The shortened version of the questionnaire consists of three categories. However, the current study took membership-related items and assessed them accordingly. The four-item scale by Shukla and Singh (2015) was used to assess psychological ownership. The items related to obligation have been considered and added to the questionnaire. The dependent variable, i.e. ethical voice, was measured using the scale of Zheng et al. (2021). The scale is a self-reporting questionnaire where the individuals rated the tendency of ethical voice.

Data analysis techniques

Numerical data were analyzed to study the relationship and test the hypotheses. The data analysis techniques were applied using the guidelines of Hair, Sarstedt, Matthews, and Ringle (2016) presented for SEM-AMOS. Before formal hypotheses testing, exploratory factor analysis (EFA), CFA and assumptions for regression analysis, e.g. normality, multicollinearity and data reliability, were assessed. After testing the regression assumptions, a formal path model was obtained from AMOS to test the impact of different variables. Before all this, EFA was run on the current model. EFA’s primary purpose is to identify the items’ discrepancies. In EFA, item-wise reliability, appropriateness of data, factoring method, rotation types and commonalities are determined. The rotation type selected for EFA is Promax, identified as the best-fit method for SEM.

Moreover, all other principal component matrix methods have been used to estimate model fitness and identify differences among factors. Above all, KMO and Bartlett’s tests were carried out for the appropriateness of data. Researchers like Lowry and Gaskin (2014) defined that the range of KMO 0.60 and above are in the excellent acceptance range. The result of the said test value of KMO was 0.817, which is according to the range. Furthermore, commonalities with the range set more significant than 0.4 were also identified for a good data set, and the results were within the range. Any cross-loading among items was also determined, and two items were removed from relational identification. No cross-loadings were found in the data for the remaining variables, and a smooth factor analysis surfaced.

Demographic analysis

The results related to the profile of respondents revealed that a diverse audience participated in the study. Overall responses of 248 respondents were finalized for the study. Among the respondents, gender-wise analysis notifies that balanced responses were received from males and females vis-à-vis; other categories, such as the educational level of respondents, their experience, age and social media usage, were also recorded and are presented in Table 1.

Table 1 analyzes the gender-wise values, and it notifies that 197 males, a percentage of 79.4, took part in the study. On the other hand, 51 females, with a ratio of 20.6, have filled out the questionnaires. The ratio of female respondents is identified as being lower than male respondents, which indicates that in Pakistan, males dominate the culture, and females are discouraged from taking jobs other than specific professions. Moreover, the current study adopts the purposive sampling technique, and according to this technique, most of the population is taken based on their qualification level. So, Table 1 also produced the values related to the educational level of respondents, and it has been identified that most respondents reported their academic level as master, with a percentage of 62.1 (154); 86 respondents reported a bachelor’s degree, and only eight, or 3.2% overall, reported their education level as intermediate. The table further reported the age bracket of respondents. A total of 120 respondents reported their age bracket as 26–30 years, eight as 20–25, 57 are from 31 to 35 years of age, 42 have age from 36 to 40 and 21 are 41 and above years of age, respectively. The age-wise analysis revealed that the majority of the respondents reported their age bracket as 26–30 years, and such employees are trying to learn new things from their seniors, and it is imperative to work with ethical leaders to gain the experience that would further lead them to fulfill their work in more ethical manners.

Estimation of model with structural equation modeling

The current research aims to identify the impact of ethical leadership on ethical voice by determining two paths covering relational identification and psychological safety. To achieve the objectives, the current study used the guidelines of SEM. The following analysis consisted of four sections, from data screening to path modeling. Before testing the path model, CFA was conducted to fulfill the analysis requirement. The section-wise illustration of SEM follows.

Screening of data

In the first phase of data analysis, data screening was conducted to identify whether data had any missing values, outliers, abnormality or linearity; besides, a test was run to identify the multicollinearity since these tests are prerequisites to running the formal regression and testing the hypotheses. After a thorough investigation, no missing values were found in the data; similarly, no outliers were found that would impact the construct being incorporated into the study. Table 2 displays the overall normality and descriptive stats such as minimum, maximum, mean, standard deviation, skewness and kurtosis-related values to test the normality of the data. The values under skewness are between 1 and −1, so data is interpreted as normal. Similarly, the values of Kurtosis are according to the range defined by previous researchers. Moreover, the mean and standard deviation are also according to the range. Overall results presented in Table 2 support the statement that the data are normal and allow further formal analysis.

Confirmatory factor analysis

The current study identified validity, reliability and model fitness-related tests under CFA and ran the formal regression to test the hypotheses. Table 3 shows the factor loading and reliability of the variables.

Discriminant validity

Discriminant validity in Table 4 broadly discusses the correlation among different variables. The table also included VIF and tolerance values to determine the multicollinearity-related issues, and no concerns were found of multicollinearity as well as moderate correlations among the variables being examined. Moreover, the VIF value was less than 3, confirming that the ready is fit for further analysis.

Path analysis outcomes

Path analysis was conducted after the attainment of confirmatory results from the CFA. There were no issues of common method biases; therefore, a common latent factor was not created. The following Tables 5 and 6 provide path analysis results in which all the literary foundations established were accepted.

Direct and mediation analysis

Tables 5 and 6 discuss the results related to the mediated path and provide the values for checking the direct impact of ethical leadership on ethical voice. Table 5 incorporated the path containing ethical leadership, relational identification, psychological safety and ethical voice, and the values show a significant impact of ethical leadership on ethical voice. Moreover, for mediation analysis, it is observed that the total (direct and indirect) effect of ethical leadership on ethical voice is 0.321. That is, due to both direct (unmediated) and indirect (mediated) effects of ethical leadership on ethical voice, when ethical leadership goes up by 1, the ethical voice goes up by 0.321. Furthermore, the total (direct and indirect) effect of relational identification on ethical voice is 0.106, and the total (direct and indirect) effect of psychological safety on ethical voice is 0.283. Hence partial mediation has been observed in the path.

Table 6 provides the second path, i.e. ethical leadership, organizational identification, psychological ownership and ethical voice, and the values show a significant impact of ethical leadership on ethical voice. Moreover, for mediation analysis, it is observed that the total (direct and indirect) effect of organizational identification on ethical voice is 0.103, and the total (direct and indirect) effect of psychological ownership on ethical voice is 0.239. In contrast, ethical leadership’s direct (unmediated) effect on ethical voice is 0.244. The said path also reveals partial mediation.

Figure 2 and Table 7 also contain the regression analysis values presented in the above tables. The path model was generated through SEM-AMOS, with standardized weights for each relationship. We found that ethical leadership significantly impacts organization identification and relational identification, as discussed under the direct and mediation analysis headings. In this regard, Figure 2 significantly validates all the results for direct relationships and mediation-related results.

Theoretical implications

The current study significantly contributes to the ethical leadership theory since it integrates the social information process (Salancik & Pfeffer, 1978) and self-concept perspectives (Shamir, House & Arthur, 1993). The study added unique paths that have not been tested yet in ethical leadership and followers through comprehensive mechanisms of identification, psychological safety and psychological ownership. The study found that the ethical values of followers emerged through ethical leaders and personal identities that contribute to the body of research since previous studies were conducted with social exchange theory perspectives. These mechanisms were not considered for the research. The current research defined leadership through the lens of self-concept theory and examined that ethical leaders activate follower identities that further drive individual’s positive attitudes and behaviors. The research also emphasized that leaders’ behaviors are crucial to facilitating followers’ internalization of organizational values. Leaders with ethics generate positive signals to followers that contribute to the identification (relational and organizational), further leading to psychological safety, psychological ownership and voicing behavior.

Practical implications

The current study presented several practical implications necessary for organizations. First, it confirmed that ethical leadership is a vital determinant in contributing to the ethical voice of employees/followers. Employees interacting with ethical leaders tend to share the same ethics in the organization setting. So, organizations should focus on the development of such leaders through training at the departmental level. Moreover, since ethical leadership positively influences followers’ attitudes and behavior, organizations must promote moral behavior in supervisors and subordinates.

Second, the study confirmed that relational and organizational identification contributes to the development of employees’ psychological capabilities, such as psychological empowerment and psychological safety, which further leads to identifying and reporting any misconduct in the organization. These mechanisms were not considered in previous research, and/or conducted in isolation, whereas the current study tested them collectively and offered unique insights that significantly contributed to generating an ethical voice with the help of process ethical leadership.

Limitations and directions for future research

Despite the study covering comprehensive mechanisms and identifying significant results, some limitations exist. First, it focused on the banking sector, and a moderate sample was used for the study. Future researchers should conduct in any sector where ethical misconduct prevails, such as in public sector-related organizations. Second, the article only took psychological mechanisms that potentially contribute to generating an ethical voice. Future research should take other behavioral outcomes, such as extra-role and knowledge-sharing behaviors. The literature on ethical leadership is more inclined to investigate ethical leadership outcomes concerning antecedents, so future research should study the antecedents of ethical leadership, such as personality traits.

Finally, leadership-related research cannot be conducted without boundary conditions, so future research should consider culture-related variables as moderators. For example, spiritual leadership is an emerging style recently declared the panacea for modern business issues. Future research can use spiritual leadership as an antecedent for primitive as well as prohibitive extra-role behaviors and knowledge-sharing behavior with the role of sequential mediation. Varying attitudes and emotional states such as psychological safety, perceived organizational justice, employees’ attachment, positive affect, and empathy can be used to understand the mechanism with the help of other leadership styles like empowering, and servant spiritual leadership.