Paternalistic Leadership Styles and Employee Voice: The Roles of Trust in Supervisors and Self-Efﬁcacy

: The purpose of this paper is to test the role of trust in supervisors in mediating the link between paternalistic leadership styles and employee voice and to investigate the contingency role of self-efﬁcacy on the relationship between trust in supervisors and employee voices. We designed the items using survey questionnaires that were assessed by prior studies and collected data from 485 highly skilled employees of manufacturing organizations. To assess the validity of the suggested hypotheses, we used a Baron and Kenny (1986) approach and conducted hierarchical regression analysis. The ﬁndings reveal that authoritarian leadership and moral leadership styles are signiﬁcantly related to trust in supervisors, which explains the association between authoritarian leadership style, and moral leadership style and employee voice. However, the results suggest that trust in supervisors does not explain the association between benevolent leadership style and employee voice. Moreover, the ﬁndings reveal that self-efﬁcacy moderates the link between trust in supervisors and employee voice. The originality of this work lies in the fact that this research is the ﬁrst to test the mediating role of trust in supervisors in the relationship between paternalistic leadership styles and employee voice and the moderating role of self-efﬁcacy on the association between trust in supervisors and employee voice.


Introduction
Voice refers to the "discretionary communication of ideas, suggestions, concerns, or opinions about work-related issues with the intent to improve organizational or unit functioning" [1]. Employee voice is important for continuous innovation [2] because employees of organizations participate in the decision-making process with their opinions and ideas [3]. By enabling employees to speak up about their opinions and concerns, organizations are able to respond to unpredicted situations [4] and avoid mistakes [5]. Moreover, voice behavior is related to higher employee motivation [6] and better working conditions [7]. Hence, it is important to understand the factors associated with employee voice.
Leadership styles play a critical role in influencing employee voice [8]. Specifically, paternalistic leadership styles play a significant role in impacting employee voice [8][9][10]. Therefore, previous research has examined the link between paternalistic leadership styles and employee voice [8][9][10]. Despite the number of studies on paternalistic leadership and voice, several questions remain unaddressed and need to be answered.
First, despite the fact that previous research studied the link between paternalistic leadership styles and employee voice [9,10], very little is known on the reasons why individuals with paternalistic leaders refrain themselves from information sharing and why they speak up on work-related issues.
Employee voice is a discretionary behavior [11]; therefore, we can consider it as a voluntary reciprocal behavior that employees engage in to facilitate improvements in their organizations. We believe that employees decide to share their concerns on workrelated issues by assessing the superiors' behavior and treatment of employees. According to social exchange theory, trust is a significant element of social exchange [11], and it lies at the heart of interpersonal relationships [12]. Thus, individuals may consider the trust factor when assessing superiors' behavior before sharing information. Hence, we assume that paternalistic leadership styles do not directly impact employees' decisions on sharing information; rather, based on assessment of the behavior and treatment of superiors, individuals first develop trust towards supervisors, and based on this trust level, individuals decide to speak up or refrain from information sharing. Thus, trust in supervisor can be an important mechanism that mediates the link between paternalistic leadership styles and employee voice.
Second, according to Bandura [13], people assess their own capabilities and competencies before making a decision to undertake certain actions. Hence, not all individuals who have trust in their supervisors can express their opinions on work-related issues.
Although several studies examined the association between trust in supervisors and employee voice [14][15][16][17], there is a lack of knowledge on when individuals who have a trust in supervisors express their opinions on work-related issues and when they do not.
Typically, individuals evaluate their own capabilities and abilities in making change and improvements in the organization before speaking up. When individuals perceive that they can handle difficult issues if they try hard enough, they are more likely to express their concerns on work-related matters. In contrast, if individuals believe that they are cannot handle challenging tasks, they are less likely to voice on work-related problems and suggest ideas to improve the company. Hence, dispositional traits such as self-efficacy can be a contingency on the relationship between trust in supervisors and employee voice.
Thus, to address these gaps in the literature, our work investigates the mediating role of trust in supervisors on the relationship between paternalistic leadership styles (such as benevolent, moral and authoritarian) and employee voice and the contingency role of self-efficacy on the link between trust in supervisors and employee voice.
Investigating the mediating role of trust in supervisors on the link between paternalistic leadership styles and employee voice enables us to understand why employees decide to speak up or stay silent when their supervisors exercise paternalistic leadership styles. Moreover, this investigation sheds light on the form of paternalistic leadership style that fosters employee voice via trust in supervisors and enables the management of companies to manage the form of paternalistic leadership style that impacts trust in supervisors, which in turn influences employee voice. Further, testing the contingent role of self-efficacy on the link between trust in supervisors and voice provides insights into when individuals who trust in their supervisors will express their opinions on work-related matters and when they will refrain from sharing information.
Previous research that examined the associations between paternalistic leadership and employee voice [8][9][10] did not consider the role of trust in supervisors on the link between paternalistic leadership styles and employee voice in their studies. Moreover, the role of self-efficacy as a moderator of the relationship between trust in supervisors and employee voice was also overlooked. Therefore, our work aims to extend the field by presenting empirical evidence on the mediating role of trust in supervisors in the relationship between paternalistic leadership styles and employee voice and on the contingency role of selfefficacy on the association between trust in supervisors and employee voice.

Authoritarian Leadership and Trust in Supervisors
Paternalistic leadership refers to "a style that combines strong discipline and authority with fatherly benevolence" [18].
A number of researchers [18][19][20][21] have investigated paternalistic leadership in last twenty years. Paternalistic leadership is perceived differently across cultures. In Western culture, paternalistic leadership is perceived negatively because it includes authoritarianism, whereas non-Western cultures (such as in Japan, China and Republic of Korea) tend to view it positively due to the protection and care of paternal authority [22]. Hence, our study examines paternalistic leadership in one of the countries, Republic of Korea, where paternalistic leadership is prevalent.
Even though paternalistic leadership shares conceptual elements with other leadership theories, such as ethical leadership and transformational leadership [23], paternalistic leadership differs theoretically from the abovementioned leadership styles. Paternalistic leadership differs from transformational leadership in a number of aspects. The main intention of transformational leaders is to transform their subordinates into leader, whereas paternalistic leaders prefer a centralized decision-making process in which subordinates obey their superiors [24]. Moreover, the emotions generated by paternalistic leaders are related to gratitude, fear, liking and admiration, whereas the emotions triggered by transformational leaders are related to stimulation, excitement and optimism [22]. The morality element of paternalistic leadership style reflects ethical leadership. However, paternalistic leadership differs from ethical leadership because ethical leaders behave as role models to their subordinates and expect followers to make ethical decisions and engage in prosocial behaviors [23], whereas paternalistic leaders do not intend to act as role models and there are no such expectations from subordinates. Moreover, paternalistic leadership exercises one-way communication and expects followers to obey and listen, whereas ethical leaders emphasize equality, in which leaders and followers can have two-way communication.
There are three forms of paternalistic leadership: authoritarian, benevolent and moral leadership styles [25]. Authoritarian leadership is defined as the leaders' behavior of asserting control and authority over employees and demanding unquestioning compliance and obedience from them. The benevolent behavior of leaders refers to a tactic in which leaders demonstrate personalized concern for familial well-being and the needs of followers. Moral leadership refers to a leader behavior that exhibits integrity and moral character through unselfish behavior.
Authoritarian leaders use strict discipline and reprimand employees for low performance [26]. As a result, the authoritarian leadership style causes negative emotions, such as fear and anger, in employees [22]. Similarly, Farh et al. [25] reported that authoritarianism tends to generate anger and fear. According to these scholars, employee fear and anger tend to reduce individuals' trust in leaders. Moreover, an authoritarian leadership style emphasizes control over employees and can make employees feel oppressed and uneasy, leading to negative social exchanges between supervisors and employees [27]. According to Lau et al. [28], an authoritarian leadership style creates a mutually distrustful working environment by not empowering subordinates in the decision-making process and closely controlling their activities. Thus, relying on the discussions above, we propose the following hypothesis: Hypothesis 1 (H1): Authoritarian leadership is negatively related to trust in supervisors.

Benevolent Leadership and Trust in Supervisors
Benevolent leaders stimulate positive emotions by showing concern to their employees, which enhances employee trust in their leaders [26]. Similarly, Colquitt et al. [29] indicated that benevolent leadership triggers positive feelings among employees by considering their familial and personal well-being. Moreover, benevolent leaders consider employee career Sustainability 2022, 14, 12805 4 of 14 development, provide feedback and offer employees opportunities to correct errors [26]. Thus, under benevolent leadership, employees feel that their leader has good intentions; with this perception, individuals tend to trust their leaders [27]. Therefore, we suggest the following hypothesis: Hypothesis 2 (H2): Benevolent leadership is positively related to trust in supervisors.
Moral leadership style has gained attention because of subordinates' trust in leaders [27]. Individuals do not trust their supervisors unless they are sure of the supervisors' good moral character, and they tend to trust leaders that show integrity and consider collective interest rather than self-interest [30]. Moral leaders tend to show unselfishness and integrity; thus, this leadership style enhances trustworthiness [31]. Similarly, Rempel et al. [32] indicated that, when supervisor altruism and integrity can assure that individuals will not be taken advantage of, employees tend to trust their supervisors. Moreover, one of the main characteristics of moral leaders is being just [28]. According to Aryee et al. [33], justice perceptions are associated with trust in leaders. Hence, a moral leadership style triggers employee trust in leaders. Furthermore, moral leaders trust and value their employees and treat them well, and therefore, individuals tend to reciprocate the goodwill of their supervisors with trust [30]. Thus, we suggest the following hypothesis:

Hypothesis 3 (H3):
Moral leadership is positively related to trust in supervisors.

Trust in Supervisors and Employee Voice
Individuals who speak up are more likely to be concerned about career development, interpersonal relationships and risks to their image [34]. Similarly, Detert and Burris [35] indicated that employees consider the risk of being labeled troublemakers who challenge the status of quo before sharing their opinions. Individuals become reluctant to express themselves when they perceive that sharing opinions or ideas may trigger retaliation and damage relationships with others [36]. According to Mayer et al. [31], individuals tend to take risks when having trust in their superiors. Employee trust in superiors reduces employees' concerns about being misunderstood and eventually leads employees to express their positions openly and freely [37]. Similarly, Dedahanov and Rhee [38] stated that the more followers trust in their superiors, the safer they feel about superiors' reactions to sharing concerns. In contrast, when there is a lower level of trust in leaders, employees cannot evaluate the risks that are related to voicing concerns and thus tend to feel unsafe [39] and refrain from speaking up. Hence, followers with trust tend to express their concerns, whereas subordinates with a lower level of trust tend to refrain from expressing thoughts. Therefore, we suggest the following hypothesis:

Hypothesis 4 (H4):
Trust in supervisors is positively related to employee voice.

Mediating Role of Trust in Supervisors
According to the principles of social exchange theory, people in exchange relationships willingly reciprocate by providing favors to other parties who provided a favor to them [11]. Social exchange theory indicates that trust is a fundamental part of social exchange and is a prerequisite for reciprocal interaction with others [11]. Hence, individuals assess the behavior of superiors and the treatment that they have received from leaders prior to speaking up on work-related problems. When followers perceive that they are treated well and receive benefit from superiors, they tend to develop trust in their supervisors, which eventually leads them to reciprocate by providing suggestions to improve the workplace. In other words, when supervisors demonstrate concerns about the daily life of their employees, behave in the manner of family members and do not take credit for employees' contributions and achievement, employees tend to develop trust in their supervisors [22]. Eventually, based on this trust, individuals are more likely to reciprocate Sustainability 2022, 14, 12805 5 of 14 the favor of superiors by developing suggestions regarding matters that affect their work group and by speaking up their opinions to better the workplace. By contrast, when supervisors require unquestioning obedience from subordinates, employees feel pressured and are less likely to trust their supervisors [27]. Consequently, they may become reluctant to provide benefits to their organizations through recommendations and suggestions to improve their workplace. Therefore, we believe that trust in supervisors is a crucial factor that explains the association between paternalistic leadership styles, namely, authoritarian, benevolent and moral, and employee voice. Hence, we propose the following hypotheses:

Moderating Role of Self-Efficacy
As mentioned earlier, previous research indicated that trust in supervisors increases employee voice [37,38]. We believe that not all employees speak up on work-related issues when they have trust in their supervisors.
According to Bandura [13], people assess their own capabilities and competencies before making a decision to undertake certain actions. Self-efficacy represents an individual's confidence in performing and organizing a certain action in the face of difficulties [13]. Perceived self-efficacy is the belief in one's ability to tackle novel or difficult tasks and to handle with difficulty in certain situations [13]. Perceived self-efficacy is generally perceived as being domain specific or task specific. Nevertheless, some scholars have also suggested general self-efficacy that is defined as a general belief in one's coping capability across wide range of new and demanding circumstances [40]. Hence, perceived self-efficacy focuses on a specific situation whereas generalized self-efficacy emphasizes on a wide range of human behaviors and handling the situations when the context is less precise [41]. Self-efficacy impact on how people think, feel and act [13]. Hence, selfefficacy is associated with individuals' perception about their competence. According to goal-setting theory, individuals with higher levels of self-efficacy have more difficult goals, whereas individuals with low self-efficacy engage in achieving less difficult objectives because they are unable to complete difficult tasks successfully [42,43]. Employee voice highlights the significance of program and policy improvements to those who responsible for these issues and challenges the status quo [35]. According to Rhee et al. [44], speaking up about work-related problems can threaten the individuals who are responsible for resolving those problems. Therefore, employees can consider exercising their employee voice as a difficult task. Following goal-setting theory, we believe that employees with high self-efficacy are more likely to perceive that they are capable of making changes and improvement in the organization by sharing their work-related concern even though it challenges the status quo and harms the subordinates who are responsible for solving those issues, whereas individuals with low self-efficacy might feel reluctant to engage in voice behavior because of their incapability to handle difficult tasks and unforeseen situations. Thus, not all individuals who have trust in their supervisors speak up their concerns on work-related matters. Individuals with higher levels of self-efficacy are likely to share their concerns on work-related issues, whereas employees with lower levels of self-efficacy tend to refrain from information sharing despite their level of trust in their superiors. Hence, self-efficacy may play a moderating role between trust in supervisors and employee voice. Thus, we suggest the following hypothesis: Hypothesis 8 (H8): Self-efficacy moderates the positive link between trust in supervisors and employee voice such that the positive link becomes stronger when self-efficacy is high.

Sample and Data Collection
To collect data, we conducted a survey in manufacturing organizations in the Republic of Korea. We met with management of 20 manufacturing companies located in Daegu, Busan, Suwon, Daejon and Incheon and explained to them the aim of our research. With the assistance of the management of companies, we acquired a list of 485 highly skilled employees and their functional managers. A convenience sampling procedure was used in data collection. We organized the data collection in two phases. In each phase, paper-andpencil surveys were distributed in small group sessions. In the first phase of data collection, we targeted the employees of organizations, and therefore, we asked the managers of organizations to invite employees, who were selected randomly, to attend in the survey. The surveys for subordinates were organized in group sessions in each company. We explained to the employees that the responses would be kept confidential. Participants answered questions related to authoritarian, moral and benevolent leadership styles; trust in supervisors; and self-efficacy. To match the answers of the subordinates to the answers of their managers, each questionnaire was coded with a researcher-assigned identification number. In the second phase of data collection, functional managers were surveyed. The functional managers were requested to evaluate the employee voice of those who attended the survey. Of the attendants, 78 were functional managers and 485 were employees. Of the 485 employee-leader dyads, 102 sets of survey questionnaires were not used in the study because of incomplete data. Thus, in total, 383 usable responses were included in this study. Of the 383 participants, 38.4 percent were female and 61.6 percent were male. According to age, 28.5 percent were between 25 and 35 years, 34.2 percent were between 36 and 45 years, 24.5 percent were between 46 and 55 years, and 12.8 percent were between 56 and 65 years. With respect to experience, 11.4 had less than 1 year, 28.2 had 1-4 years of experience, 33.2 had 5-9 years, and 27.2 had 10-15 years of experience. With respect to education level, 54.3 percent of respondents had a bachelor's degree, 36.6 percent participants had a master's degree, and 9.1 percent attendants had a PhD.

Measures
Professional translators translated the questionnaires from English into Korean, and bilingual experts backtranslated the scale items into English to ensure the accuracy of the translation [45]. In this study, paternalistic leadership style, trust in supervisors and employee voice were rated on a five-point Likert scale ranging from "1 = strongly disagree" to "5 = strongly agree." We assessed the self-efficacy items on a five-point Likert scale ranging from "1 = not all true" to "5 = exactly true".
Paternalistic Leadership Style: Cheng, Chou and Farh [46] developed a 42-item scale that assesses the three dimensions of paternalistic leadership. Several previous research [22,25,[47][48][49] used this scale to evaluate paternalistic leadership in high-tech firms, primary schools and private businesses. In 2004, Cheng and colleagues suggested an updated version of their paternalistic leadership scale that assesses the three dimensions of paternalistic leadership styles. This scale includes 26 items and it has been used by previous research [8,22] because this scale has distinctive validity from transformational leadership. Hence, in our study, we also used Cheng et al. [50]'s scale to evaluate paternalistic leadership styles. Authoritarian leadership (e.g., "My supervisor always behaves in a commanding fashion in front of employees") was tested using nine items. Benevolent leadership (e.g., "Beyond work relations, my supervisor expresses concern about my daily life") was assessed using 11 items. Moral leadership (e.g., "My leader employs people according to their virtues and does not envy others' abilities and virtues") was assessed with six items. The scales' Cronbach's α reliability values were 0.863, 0.770 and 0.941, respectively.

Trust in supervisors:
The trust in supervisor construct was measured using seven items from Robinson's [51] study. Example items include "I believe my supervisor has high integrity", "I fully trust my supervisor" and "In general, I believe my supervisor's motives and intentions are good." The scale's Cronbach's α reliability in this study was 0.861.
Self-efficacy: Self-efficacy was tested using 10 items from Schwarzer and Jerusalem [40]. Some example questions from this scale include "I can always manage to solve difficult problems if I try hard enough", "It is easy for me to stick to my aims and accomplish my goals" and "I am confident that I could deal efficiently with unexpected events". The scale's Cronbach's α reliability was 0.871.
Employee voice: The employee voice (e.g., "This employee develops and makes recommendations concerning issues that affect this work group," "This employee communicates his/her opinions about work-related issues to others in this group even if his/her opinion is different and others in the group disagree with him/her" and "This employee speaks up and encourages others in this group to become involved in issues that affect the group") was evaluated using six questions from LePine and Van Dyne [52]. The scale's Cronbach's α reliability was 0.940.
Control Variables: We measured the gender (0 = female; 1 = male), age, work experience and educational background of respondents as control variables in our study because of their potential impact on voice [8,9].

Results
Confirmatory factor analysis was used to assess the measurement model using Amos 21. The model fit was evaluated with χ 2 , CFI (comparative fit index), GFI (goodness-of-fit index), TLI (Tucker-Lewis index) and RMSEA (root-mean-square error of approximation).
The reliability was evaluated by using Cronbach's α and composite reliability (CR). According to Nunnally [55], Cronbach's α and CR values greater than 0.70 are perceived as highly reliable. The results demonstrate that all Cronbach's α and CR exceeded the recommended level of 0.70.
Discriminant and convergent validities were assessed to evaluate the validity of the measurement model. Discriminant validity is described as the degree to which the measures of different latent constructs differ from those of other constructs [53]. Discriminant validity can be demonstrated when the average variance extraction (AVE) value for each construct (within-construct variance) is higher than the squares of the correlations between the constructs (between-construct variance) [56]. The findings indicate that the AVE values exceeded the squares of the correlations between the construct and the other constructs (Table 1). Thus, the model used in our study met the criteria of discriminant validity. Convergent validity can be demonstrated when AVE values and the composite reliabilities exceed the acceptable levels of 0.50 and 0.70 [56], respectively. Our findings show that both AVE values and the composite reliabilities exceeded the threshold. Hence, in our study, these measures demonstrate convergent validity. Table 1 demonstrates the correlations among the constructs. Trust in supervisors is positively correlated with moral leadership (r = 0.266, p < 0.01), self-efficacy (r = 0.331, p < 0.01) and employee voice (r = 0.338, p < 0.01), but negatively correlated with authoritarian leadership (r = −0.341, p < 0.01). Employee voice is positively correlated with moral leadership (r = 0.178, p < 0.01) and self-efficacy (r = 0.364, p < 0.01), but negatively correlated with authoritarian leadership (r = −0.175, p < 0.01).
We assessed the validity of the proposed hypotheses by using regression analyses. The results from analyses suggest that authoritarian leadership is negatively related to trust in supervisors (β = −0.321, p < 0.01) and moral leadership is positively associated with trust in supervisors (β = 0.218, p < 0.01). Hence, H1 and H3 are supported. In H2, we hypothesized that benevolent leadership would be positively related to trust in supervisors. The findings indicate that there is no significant association between benevolent leadership and trust in supervisors (β = 0.027, p > 0.05). Thus, H2 is not supported. Moreover, the results reveal that trust in supervisors is positively associated with employee voice (β = 0.342, p < 0.01). Therefore, H4 is supported.
To evaluate the mediating role of trust in supervisors on the link between paternalistic leadership and employee voice, we used Baron and Kenny's technique [57]. According to Baron and Kenny's technique, the mediation is assumed if: independent variables (authoritarian, benevolent and moral leadership) are significantly associated with the dependent (employee voice) variable; independent variables (authoritarian, benevolent and moral leadership) are significantly related to the mediating (trust in supervisors) variable; and the mediating variable is significantly related to the dependent variable and the link of the independent variable with the dependent variable is decreased significantly (partial mediation) or remains no longer significant (full mediation) when the mediator is included in the analysis. The mediation analysis indicates that trust in supervisors fully mediates the links between authoritarian (β = −0.67, p > 0.05) and moral leadership (β = 0.088, p > 0.05) styles and employee voice because, in the third step, the mediating variable is significantly related to the dependent variable and the association between the independent and dependent variables was no longer significant when the mediator was controlled in the analysis ( Table 2). We conducted hierarchical regression analysis to evaluate the moderating role of selfefficacy on the link between trust in supervisors and employee voice. The findings indicate that self-efficacy moderates the associations between trust in supervisor and employee voice (β = 0.230, p < 0.01). Hence, H8 is also supported (see Table 3). We plotted the simple slopes of the relationships between trust in supervisors and employee voice to assess the form of interaction. The simple slope analysis revealed that the impact of trust in supervisors on employee voice is stronger for individuals with higher levels of self-efficacy (β = −0.550, p < 0.01) than for those with lower levels of self-efficacy (β = −0.012, p > 0.05) (Figure 1). Moreover, the findings demonstrate that the control variables, such as age, gender, work experience and education, do not influence employee voice. We conducted hierarchical regression analysis to evaluate the moderating role of self efficacy on the link between trust in supervisors and employee voice. The findings indi cate that self-efficacy moderates the associations between trust in supervisor and em ployee voice (β = 0.230, p < 0.01). Hence, H8 is also supported (see Table 3). We plotted the simple slopes of the relationships between trust in supervisors and employee voice to as sess the form of interaction. The simple slope analysis revealed that the impact of trust in supervisors on employee voice is stronger for individuals with higher levels of self-effi cacy (β = −0.550, p < 0.01) than for those with lower levels of self-efficacy (β = −0.012, p > 0.05) (Figure 1). Moreover, the findings demonstrate that the control variables, such a age, gender, work experience and education, do not influence employee voice.

Theoretical Discussion
In this study, we investigated the role of trust in supervisors in mediating the links among paternalistic leadership styles and employee voice and the contingency role of self-efficacy on the association between trust in supervisors and voice.
Despite the existing studies [9,10], there was a lack of a cogent theoretical framework that provided an understanding of why individuals with paternalistic leadership styles decide to express their opinions on work-related matters and why they refrain from information sharing. Hence, rather than portraying just a direct link between paternalistic leadership styles and employee voice, our study provided deeper insight into why leaders with paternalistic leadership styles impact employee voice. Therefore, we invoked social exchange theory to address the insufficient attention provided to the underlying mechanisms linking paternalistic leadership styles and employee voice by investigating trust in supervisors. The empirical findings indicate that trust in supervisors explains the relationships between authoritarian and moral leadership styles and employee voice. That is, when supervisors act in a commanding fashion in front of employees and punish employees severely for not following their supervisors, employees are less likely to believe that supervisors' intentions and motives are good. Consequently, individuals with reduced trust in their supervisors feel reluctant to improve their workplace with their recommendations and, therefore, decided to refrain from information sharing. In contrast, when superiors do not take credit for followers' achievements, employ individuals according to their virtues and do not envy others' abilities and virtues, the followers tend to perceive the supervisors as having high integrity and thus trust them. Eventually, based on this trust in supervisors, individuals decide to reciprocate to superiors and workplaces by sharing their suggestions to improve the work unit and organization. Previous researchers examined only the direct links between paternalistic leadership styles and employee voice [8][9][10] but overlooked the role of trust in supervisors in explaining the link between paternalistic leadership styles and employee voice. Hence, very little is known regarding the role of trust in supervisors in mediating the link between paternalistic leadership styles and employee voice. Thus, our work addressed this gap in the literature by presenting empirical evidence on the role of trust in supervisors to explain the link between paternalistic leadership styles and employee voice and suggests trust in supervisors as a critical mechanism that explains the relationships between authoritarian and moral leadership styles and employee voice. Our findings are in line with the perspectives of social exchange theory [11].
Moreover, the results suggest that benevolent leadership does not have a significant relationship with trust in supervisors. This insignificance can be explained in the following way: Employees might assess their supervisors' treatment and activities that are related to work issues in developing trusting relationships with them before sharing work-related concerns. Since benevolent leaders express their concerns about employees' daily lives and devote energy to taking care of employees and their family members, this behavior can be perceived by employees as unrelated to work issues. The insignificance found in the link between benevolent leadership and trust in supervisors can also be explained by the generation gap between supervisors and subordinates found in Korean organizations. The study of new generation workers in the IT industry in South Korea [58] indicated that new generation workers showed a self-centered tendency and expressed discomfort around superiors who showed concern for out-of-office issues, including their personal lives. The self-centered tendency of relatively young workers may induce subordinates to draw clear boundaries between the workplace and their personal realm. Thus, benevolent superiors that care about the familial and personal well-being of employees may not significantly affect subordinates' trust in them. Therefore, in our study, respondents did not consider a benevolent leadership style as a significant factor in shaping their trust in supervisors.
According to Bandura [13], individuals evaluate their own abilities and competencies before taking actions. Thus, not all individuals who trust supervisors express their concern about work-related problems. Previous studies [14,17,59] examined the direct links between trust in supervisors and employee voice and did not consider the contingency factors between trust in supervisors and employee voice. Hence, there was a lack of knowledge on when individuals who trust supervisors voice work-related issues and when they do not. In other words, despite the amount of research on trust in supervisors and employee voice, we have a relatively reduced understanding of when individuals who trust their supervisors share their suggestions and ideas on work-related matters and when they decide to remain silent. Therefore, our study extended the literature by investigating selfefficacy as the boundary condition on the link between trust in supervisors and employee voice. The findings suggest that individuals who have trust in supervisors tend to voice their opinions when they are confident that they can handle unexpected events efficiently. That is, the higher the self-efficacy, the stronger the link between trust in supervisors and employee voice (Table 4). In contrast, the findings reveal that individuals with lower levels of self-efficacy do not significantly change their voice, despite their trust in supervisors.

Practical Implications
The findings of this study have a number of practical implications. Our research revealed that moral leadership styles positively impact employee trust in supervisors. Thus, we recommend that human resources management understand the consequences of each leadership style and adjusts the hiring and promotion policies to better induce moral leadership behaviors from employees in supervisory positions [27]. Likewise, leadership training programs emphasizing moral leadership can help supervisors to reflect and refine their leadership behaviors so that they can effectively build trusting relationships with their subordinates while avoiding exerting excess power over employees, which can hamper trust in leadership. Moreover, we suggest that leaders show personal integrity [60] and openness to new information [9]. The findings suggest that an authoritarian leadership style is negatively associated with trust in supervisors. Organizations can avoid authoritarian leadership styles by soliciting employees' ideas to increase their feeling of importance and involvement in organizational issues [8,10]. Our study highlights the critical role of trust in supervisors in linking the relationships between authoritarian and moral leadership styles and employee voice. Hence, we recommend that practitioners understand the role of trust in supervisors as one of the key voice-supporting contexts and monitor and support the supervisors' authoritarian and moral leadership behaviors that can influence subordinates to express ideas. Moreover, our findings reveal the contingency role of self-efficacy on the associations between trust in supervisors and voice. Since a personal behavior history is regarded as one of the main drivers of self-efficacy, we recommend that organizations implement training programs that help individuals to master knowledge and skills to accomplish assigned tasks successfully, which eventually increases employee self-efficacy [50]. Moreover, to increase individual self-efficacy, we suggest that managers establish mastery modeling programs and implement supportive supervisory practices [61].

Limitations and Future Research Directions
Despite the contributions of our study, it has some limitations. Thus, we have suggestions for future research. First, in this study, we measured the unitary construct of voice, which attributes the behavior of voice rather than the intentions of respondents when speaking up. Therefore, we suggest that future research investigate the relationships among leadership styles and trust in supervisors with multidimensional constructs of voice (i.e., promotive and prohibitive voice) that express the intentions of individuals when speaking up. Second, our study examined only the paternalistic leadership style as the antecedent of trust in supervisors, which in turn fosters employee voice. Hence, we suggest that future studies assess the role of different forms of leadership styles, such as transformational and transactional styles, in influencing employee voice via trust in supervisors. Third, the data were collected only in one country, the Republic of Korea. To increase the generalizability of the study, we recommend that future research assess data from other countries where paternalistic leadership is pervasive. Institutional Review Board Statement: Akfa University Institutional Review Board issued the waivers for the research in this paper because it involves no more than minimal risk to the privacy of survey participants.