The Impact of Paradoxical Leadership on Employee Knowledge-Sharing Behavior: The Role of Trust in the Leader and Employee Promotive Voice Behavior

Abstract

As the organizational environment becomes more volatile, uncertain, complex, and ambiguous, and the economy becomes increasingly knowledge-based, organizational knowledge management is key for companies’ success. This is especially important as organizational ties are weaker and job-hopping becomes a more prevalent phenomenon. As human resource mobility increases, companies must ensure that knowledge remains within the company despite employee exit. In this context, the current study sought to understand how leaders’ actions can facilitate employee knowledge sharing, focusing on paradoxical leadership. Besides examining the impact of paradoxical leadership on employees’ propensity to adopt knowledge-sharing behaviors, this study also explored the effects of one potential intervening variable (i.e., promotive voice behavior) and one potential boundary condition (i.e., trust in the leader) on this relationship. A two-wave time-lagged correlational study was conducted with a sample of 154 workers from various sectors. The results of moderated mediation analysis suggest that paradoxical leaders indirectly promote greater knowledge-sharing among subordinates by fostering their promotive-voice behaviors, but only for those with high levels of trust in the leader. The implications of these findings for current organizational challenges regarding knowledge management are discussed.

1. Introduction

As the organizational environment becomes more volatile, uncertain, complex, and ambiguous (VUCA), and the economy becomes increasingly knowledge-based, knowledge becomes an important asset for organizations. Knowledge management, including its acquisition, sharing, and use, becomes key for companies’ competitive advantage, success, and sustainability in the long-term (Rezaei et al. 2021). This is especially important as organizational ties are weaker and job-hopping becomes a more prevalent phenomenon. On this regard, a recent survey by McKinsey and Company (2022) involving workers from nine European countries found that one-third of the respondents were expecting to quit their jobs in the next three to six months. In the same line, a study by OECD (2023) revealed that changing jobs frequently has increased in OECD countries for all age groups, with younger workers being more prone to leave their jobs. As human resource mobility increases, companies must ensure that knowledge remains within the company despite employee exit to minimize human capital lost. Incentivizing employees to share their knowledge with other organizational members (e.g., leaders, coworkers, and subordinates) is, thus, an important piece for an effective knowledge management processes.

A leader’s role in facilitating the knowledge management processes within the organizational setting, including knowledge sharing, has been subject to inquiry in the past (Pellegrini et al. 2020). Several leadership styles have been related to the promotion of employee and team knowledge sharing behavior, including transformational leadership (Kim and Park 2020), transactional leadership (Hussain et al. 2017), ethical leadership (Goswami and Agrawal 2023), authentic leadership (Edú-Valsania et al. 2016), and servant leadership (Reslan et al. 2021).

Adding to the extant literature, the present study has focused on one type of leadership that increasingly is receiving more attention because of its advantages in dealing with organizational paradoxes—paradoxical leadership. This refers to the apparently competing behaviors of leaders to simultaneously meet rival workplace demands over time (Zhang et al. 2015). Evidence has shown that paradoxical leadership is associated with better leader−subordinate relationships, as well as increased positive job attitudes and behaviors, such as organizational commitment roles and extra-role behaviors (Kundi et al. 2023; Pan 2021; Zhang et al. 2015). It has also been associated with higher employee creativity and innovation at work (Lee et al. 2023a; Zhang et al. 2015).

The present study sought to analyze its relationship with employee knowledge sharing, proposing that paradoxical leaders stimulate mutual exchange of personnel knowledge, expertise, and experience, as well as knowledge co-creation amongst employees, namely by increasing their disposition to engage in promotive-voice behavior. This discretionary behavior targets the proactive implementation of constructive changes in the workplace (Liang et al. 2012). Because paradoxical leaders balance the challenges between organizational and employee development and allow employees to feel supported, autonomous, and responsible (Li et al. 2020), employees should feel more motivated to express their ideas and suggestions on how to implement improvements in the workplace and therefore be more likely to exchange their knowledge with other members. Accordingly, this study proposes employee promotive voice as a potential mechanism explaining how paradoxical leadership fosters knowledge sharing.

Trust in the leader was also included in the present study’s research model to develop a more comprehensive conceptual framework. Both voice and knowledge-sharing behaviors have some degree of risk associated, as they can be interpreted as a criticism to the current situation. The level of trust in referent people, including the leader, can inhibit or enhance the engagement in such behaviors and was therefore set as a boundary condition of paradoxical leadership’ effects on promotive voice and knowledge sharing.

By exploring the interrelationships between the abovementioned set of variables, the present study extends the existing literature on paradoxical leadership’s contribution to intensified voice behavior (Lee et al. 2023a; Xue et al. 2020) and knowledge sharing (e.g., Devi 2023; Yi et al. 2019). A careful review of the relevant literature failed to reveal any research that has focused on the relationships between these variables, as well as the moderating role trust in the leader may have.

The results provide information relevant to practitioners by clarifying which factors foster knowledge sharing, which is important to designing appropriate ways to manage knowledge, including the adoption of paradoxical leadership behaviors by managers.

This paper is structured as follows. The Section 1 presents a review of the pertinent literature and the theoretical framework underlying the research model, as well as the hypotheses’ development. In the Section 2, the methodology is described, including sample collection, procedures, and measures, after which the statistical analyses and results are presented. The Section 5 discusses the findings’ theoretical and practical implications.

2. Literature Review and Hypotheses

2.1. Paradoxical Leadership

As mentioned above, paradoxical leadership is gaining more attention, especially since it continues with the research regarding what is a paradox: in this case, in the organizational context (Batool et al. 2023). As defined by Smith and Lewis (2011, p. 386), a paradox is “contradictory yet interrelated elements that exist simultaneously and persist over time. Such elements seem logical when considered in isolation but irrational, inconsistent, and even absurd when juxtaposed”. Despite these dynamic and divergent demands, the literature highlights that to cope with these paradoxes, there is a need to juxtapose the different and opposing elements and reframe the tensions of either of them and/or even the possibilities of them both (Smith 2014). Some scholars agree that these different elements need to be accepted to be true in a simultaneous matter, and that holistic thinking is needed to achieve that (Miron-Spektor et al. 2011; Waldman and Bowen 2016).

Based on this, and considering the current challenges organizations face, the literature has been exploring how leaders can manage these conflicts and tensions and make decisions regarding the next step. Zhang et al. (2015) developed the concept of paradoxical leadership that focus on the fact that management, while in a paradoxical predicament, needs to be able to internally connect and manage the needs of both the organization and its development, as well as the needs of the employees (Li et al. 2020; Zhang et al. 2015).

The conceptualization of paradoxical leadership behavior is based on the Yin-Yang philosophy, as the two sides of the paradox are both complementary, as well as contradictory (Zhang et al. 2015). With that in mind, the authors defined that paradoxical leadership encompasses five dimensions: treating subordinates uniformly while allowing individualization; combining self-centeredness with other-centeredness; maintaining decision control while allowing autonomy; enforcing work requirements while allowing flexibility; and maintaining both distance and closeness.

However, a question can be raised, that is, how is paradoxical leadership different from other leaderships styles? Paradoxical leaders have a paradox mindset, assuming the dynamic and different possibilities of the paradox. Their role is to cope with change while maintaining an equilibrium within the dynamics of the organization by having a growth oriented, mechanistic, and flexible organizational structure. Further, and differently from other leadership styles, these leaders embrace tensions that might arise and used them as opportunities to experience and learn. Moreover, it is a leadership style that has a long-term sustainability-strategy results mindset, meaning that it thinks in long-term, not in short-term success (Batool et al. 2023; Zhang et al. 2015).

2.2. Paradoxical Leadership and Knowledge-Sharing Behavior

As established, paradoxical leadership combines two apparently incompatible yet connected behaviors and tries to emphasize work tasks and responsibilities, setting high standards for work. It also empowers and encourages employees by increasing their freedom, flexibility, and autonomy at work (Zhang et al. 2015). Furthermore, this type of leadership allows workers to make different decisions during the organizational development, is open for communication, and values different ideas (Zhang et al. 2015), which leads to knowledge sharing.

Knowledge sharing can be defined as the exchange of knowledge, expertise, and technology among different areas within an organization (Wang et al. 2009) and developing new skills or/and competencies by exchanging information or helping others (Connelly et al. 2012). However, exchanging and creating knowledge with others is not considered to be a tangible activity, meaning that it cannot be observed or enforced (Dalkir 2013; Grossman 2007). Research on the topic found that knowledge sharing has an impact on organizational learning (Swift and Hwang 2013). With knowledge sharing, research has stated that employees can experience more situations to learn by engaging with other colleagues and integrate them into practical applications on the organizational level (Farooq 2018; Swift and Hwang 2013).

Given this context, how can paradoxical leadership influence knowledge sharing in organizational settings? Since paradoxes emerge from incompatible yet interrelated elements, their management constitutes a challenge for organizations. Knowledge-sharing processes can help overcome organizational challenges by co-generating new ideas, solutions, and outputs, searching for alternatives, and optimizing decision-making processes (Razak et al. 2016; Zhang et al. 2022).

Drawing on signaling theory (Connelly et al. 2012; Spence 1978), one signal we considered is that paradoxical leadership sends signals to employees on how knowledge sharing is valued by the organization. Specifically, since paradoxical leaders embrace different ideas and promote out of the box thinking, it can be theorized that employees will be more inclined to engage in conversations and share information. Paradoxical leaders foster a culture of openness and trust, leading through this method to knowledge sharing (Devi 2023). As established by Jia et al. (2018), this will lead, possibly, to the organization developing different and efficient problem-solving and decision-making processes.

The leadership literature has found that leadership can foster knowledge-sharing behaviors. This is the case of transformational leadership, leaders can encourage employees to engage in organizational learning and allow them to experiment and communicate with each other (Kim and Park 2020), especially if those employees believe that the group values their contributions (Swift and Virick 2013).

The literature indicates that there is a relationship between paradoxical leadership and knowledge sharing. Yi et al. (2019) found a moderating effect of knowledge sharing on the relationships between paradoxical leadership and exploratory innovation in high-technologic industries. Devi (2023) found that knowledge sharing serves as a mediator in explaining the relationship between paradoxical leadership and creativity, whereas knowledge hiding does not mediate this relationship among employees. Based on the arguments provided above, the following hypothesis was developed.

Hypothesis 1: 

Paradoxical leadership is positively related to employee knowledge-sharing behavior.

2.3. Paradoxical Leadership, Employee Promotive-Voice Behavior and Knowledge-Sharing Behavior

Since paradoxical leadership is a recent topic of research, little is known about the processes underlying its effects on employees’ attitudes and behaviors at work. Regarding knowledge sharing, because it implies disclosure and/or speaking up about personal information, experiences, and expertise that help others to address their daily job demands and ultimately develop their skills and abilities and co-create further knowledge, a possible psychological intervening variable is worker’s disposition to engage in voice behaviors.

Employee voice behavior refers to discretionary behavior aiming at proactively making constructive changes (Van Dyne and LePine 1998; Morrison 2023), either by signaling potentially harmful work practices to be suppressed (i.e., prohibitive voice) or suggesting opportunities and initiatives to increase operational efficiency (i.e., promotive voice) (Liang et al. 2012). Interpersonal workplace relationships, both with leaders and coworkers, may be harmed if the receivers interpret voice as criticism (Detert and Edmondson 2011). Voicing personal ideas can be risky, so not all workers are equally willing to express their thoughts and feelings at work. Promotive voice is less risky because employees engaged in this behavior are often seen as contributing to the organization’s success (Liang et al. 2012).

Relevant antecedents of employee voice behavior include dispositional variables (e.g., proactive personality and conscientiousness) and job and organizational attitudes (e.g., autonomy, organizational commitment, and organizational justice) (Chamberlin et al. 2017; Morrison 2023). Within contextual variables, leader behavior has been identified as a key antecedent because leaders can influence workplace norms regarding voice and directly motivate or deter employee behavior (Chamberlin et al. 2017).

Indeed, voice behaviors have been positively related to distinct types of positive leadership (e.g., transformational leadership—Svendsen et al. 2018; authentic leadership—Hsiung 2012) and leader−member exchange (Carnevale et al. 2017). Recent research has also found evidence of paradoxical leadership’s positive effect on employee voice (Lee et al. 2023a; Xue et al. 2020). For instance, Li et al. (2020) found that paradoxical leadership stimulates promotive- and prohibitive-voice behaviors via the enhancement of perceived psychological safety and self-efficacy. The cited authors propose that by having good relationships with workers, paradoxical leaders can serve as role models with whom workers learn how to balance the benefits and risks of voice behavior and to express it at an appropriate time. Xiao et al. (2023) found that perceived insider status and organizational-based self-esteem-mediated paradoxical leadership effect employee voice. Rescalvo-Martin et al. (2021) found that paradoxical leadership boosted hospitality employees’ voice behavior. Kundi et al. (2023) established that work engagement intervenes as a mediating variable in paradoxical leadership and voice’s relationship using a sample from the same business sector.

Employee voice and knowledge sharing behaviors are distinct forms of discretionary behaviors at work involving risk-taking, but voice can have more unfavorable effects on interpersonal relationships than knowledge sharing (Bavik et al. 2018). Considering this, the present study has focused on promotive voice because being positive in tone seems more aligned with the prosocial nature of knowledge co-creation and disclosure. Prior studies on the relationship between variables, despite their scarcity, indicate that the two variables are intercorrelated, but do not clarify the causality nexus. For instance, Lee et al. (2023b) found that co-workers’ knowledge sharing predicts employee voice behaviors as rated by supervisors, thereby influencing employee innovation on service performance. On the other hand, Narayanan and Nadarajah (2024) found that increased employee voice led to increased knowledge productivity. Alzyoud et al. (2024) reported that increased employee voice led to increased employee innovation via the mediating role of tacit knowledge sharing. The present study proposes that promotive-voice behavior can be conducive to knowledge-sharing behavior amongst organizational members because it signals employees’ willingness to improve organizational functioning. One way to help the company to enhance its activities is sharing their knowledge, both tacitly and explicitly with others. Since paradoxical leaders welcome employees’ voice and knowledge-sharing initiatives, promotive voice is proposed as a mediating mechanism explaining how this type of leadership fosters knowledge sharing. The present study’s second hypothesis was developed to reflect the above findings.

Hypothesis 2: 

Employee promotive-voice behavior mediates the positive relationship between paradoxical leadership andemployee knowledge-sharing behavior.

2.4. The Moderating Role of Trust in the Leader

Trust is defined as a “psychological state comprising the intention to accept vulnerability based upon positive expectations about the intentions or behavior of another” (Rousseau et al. 1998, p. 395). Trust in the leader is affected by the quality of relationships supervisors have with their subordinates, including the support and genuine attention given to each other’s well-being, which is a characteristic of high-quality exchange relationships (Mitchell and Ambrose 2012; Mitchell et al. 2015). Extra-role behavior appears to become more common in the presence of stable social exchange relationships based on follower−leader trust (Agarwal 2014; May et al. 2004), including voice behavior (Carnevale et al. 2017; Van Dyne et al. 2008) and knowledge-sharing behaviors (Hao et al. 2019; Le and Nguyen 2023; Son et al. 2017).

Greater trust in the leader contributes to perceived psychological safety (May et al. 2004), which reduces feelings of worry and anxiety in followers, as they are more confident that their leader is dependable, benevolent, and will provide support if needed (Madjar and Ortiz-Walters 2009). Thus, when employees feel safe, based on the trust they have in their leaders, they are more likely to speak up and express their options openly, because trust reduces the sense that such behavior might jeopardize their situation at work. In other words, trust reduces the perceived risk of voicing, thereby fostering employee voice behaviors (Duan et al. 2019).

We propose that trust in the leader may act as a boundary condition for the positive effects of paradoxical leadership on employee voice and knowledge-sharing behaviors. As paradoxical leadership implies leaders behave in seemingly contradictory ways to successfully balance structural and relational demands over time, workers may doubt the leaders’ intentions and feel vulnerable, unless such behavior is viewed as coming from a credible source (Yang et al. 2024). If workers distrust leaders, they might doubt the leaders’ intention behind apparently contradictory behaviors and judge these as unreliable and inconsistent (Yang et al. 2024). In this situation, voicing ideas may be seen as riskier due to the manager’s inconsistent behavior—will he/she welcome the expressed ideas or not? Uncertainty brings anxiety regarding the possibility of negative consequences for the self (e.g., retaliation, punishment, etc.) (Detert and Treviño 2010; Morrison 2023) and thereby diminishes the likelihood of voice behavior. As a result, the relationship between paradoxical leadership and promotive-voice behavior will be weakened when trust is low.

If trust in the leader is high, workers will be more likely to voice because they will perceive that doing so will be safe and effective (Jiang et al. 2018). They will feel free to speak up honestly (Silla et al. 2020). Trust in leadership enables employees to accept the inherent risk to voice is based upon positive expectations of the behavior or intentions of their leaders (Detert and Treviño 2010; Silla et al. 2020). The current study’s third hypothesis was therefore written as follows.

Hypothesis 3: 

Trust in the leader moderates the positive relationship between paradoxical leadership and employee promotive-voice behaviors, such that the relationship is stronger when the level of trust is high rather than low.

It is reasonable to expect that in situations of high levels of trust in the leader, the indirect effect of paradoxical leadership on knowledge-sharing behavior via promotive voice will be stronger than when workers distrust their leaders. Trust in other organizational members has been settled as a relevant enabler of knowledge sharing (Farooq 2018; Nadeem et al. 2021; Rutten et al. 2016). When employees trust their leaders, they are more comfortable with the leader’s seemingly contradictory behavior. They are more likely to understand that their paradoxical leaders welcome ideas on how to improve the company’s activities and thus are more willing to share their expertise, knowledge, and experiences with others. If trust in the leader is low, paradoxical leaders’ apparently contradictory behavior can signal unclear messages about accepting new ideas, increasing the perceived risk of expressing voice and thereby limiting knowledge sharing initiatives. Therefore, the final hypothesis that focuses on a moderated mediating effect was developed.

Hypothesis 4: 

Trust in the leader moderates the positive indirect relationship between paradoxical leadership and employee knowledge-sharing behavior via employee promotive-voice behavior, such that the relationship is stronger when the level of trust is high rather than low.

The conceptual model hypothesized in this study is shown in Figure 1.

3. Method

3.1. Procedure and Sample

Data were collected at two times, with a 2-week interval, to minimize the possibility of encountering common method bias (Podsakoff et al. 2024) and reverse causality (De Clercq et al. 2020) and to improve power of the theoretical model (Griep et al. 2021).

Participants were recruited by social media professional platforms and by approaching organizations via e-mail or personal contacts. In this context, a non-probabilistic convenience sample was used for this study. Upon acceptance, the questionnaires were distributed via a link and sent to the participants by the opted means of communication. This communication was accompanied by a brief explanation of the study’s objective, as well as information highlighting that the participation was voluntary and anonymous and that participants could leave or ask to be removed at any moment during the data collection. To merge the questionnaires, participants were instructed to generate a personal code with the first letter of their mother’s name, number of children they have, month of birth (two digits), last letter of their last name, and the last digit of their phone number (e.g., M004A1).

In total 302 participants completed the time 1 questionnaire, and 194 participants completed the time 2 questionnaire (i.e., a response rate of 64.2%). However, due to missing the control questions or not meeting the inclusion criteria some participants were excluded (i.e., being less than 18 years old; working without a formal contract; and being under the management of the same direct manager for less than 3 months). The final sample included 154 workers from a south-European country (i.e., the final response rate was 51%). Mean age was 37.17 years old (SD = 13.18) and 68.2% were female. Education was mostly a bachelor’s degree (46.4%); however, 70.2% had some type of higher education (e.g., Master’s and/or Ph.D.). Regarding their relationship to the leader, 66.9% had a daily interaction with the leader and had worked with him/her between three months and 35 years (M = 5.01; SD = 6.74). Participants, on average, worked 9 years at their organization (SD = 11.28; min = 3 months; max = 40 years), most of them with permanent contracts (77.5%) in organizations with more than 250 employees (58.9%) in the private sector (73.5%).

3.2. Measures

The time 1 survey included the measurement scales for paradoxical leadership and trust in the leader; consequently, the time 2 survey included measurement scales for employee promotive-voice behavior and knowledge sharing. Hence, the predictor and criterion variable were separated, and this would allow us to test the tie-lagged relationship between them, as well as the role of the mediator and moderator variables. McDonald’s omega (ω) was calculated as an estimation of internal consistency for the different scales.

3.2.1. Paradoxical Leadership (Predictor Variable)

To access the employee’s perception of the leader’s paradoxical leadership, the Paradoxical Leadership scale by Zhang et al. (2015) was used. This consists of 22 items divided into five factors, as mentioned previously. A sample item is “My direct leader uses a fair approach to treat all subordinates uniformly, but also treats them as individuals”, where the participant indicates his/her response on a Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Confirmatory factor analyses (CFAs) were conducted. After removing item 20, which had a standardized estimate indicative of misalignment with the construct (i.e., −0.048), the five-factor solution still showed issues with the covariance matrix of latent variables, indicating high correlations among some factors that could lead to potential multicollinearity problems. Specifically, factors 1, 2, and 5 were highly correlated (between 1.26 and 1.67), so the model was simplified by aggregating the items of the problematic factors into one factor. Consequently, a three-factor solution was tested (χ²(154) = 308.976, p < 0.001; χ²/df = 2.01; comparative fit index [CFI] = 0.915; Tucker–Lewis index [TLI] = 0.901, root mean square error of approximation [RMSEA] = 0.068), including items related to uniformity and individualization, self-centeredness with other-centeredness, and distance and closeness in the first factor. The second factor related to decision control while allowing autonomy, and the third factor related to enforcing work requirements while allowing flexibility. This three-factor solution showed a better fit compared to the five-factor solution (χ²(154) = 325.734, p < 0.001; χ²/df = 2.11; CFI = 0.903; TLI = 0.886, RMSEA = 0.073) or the single-factor solution (χ²(189) = 432.656, p < 0.001; χ²/df = 2.28; CFI = 0.838; TLI = 0.820, RMSEA = 0.091). McDonald’s omega for total scale (ω = 0.916) was later calculated, indicating good internal consistency.

3.2.2. Knowledge-Sharing Behavior (Criterion Variable)

Employees’ self-rated knowledge-sharing behaviors using the scale from Swift and Virick (2013), which consists of 5 items. A sample item is “I often share my knowledge with my work colleagues”, where participants were asked to answer based on a Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). The measured unidimensionality was tested with a single-factor CFA which indicates a good fit (χ²(3) = 3.986, p = 0.263; χ²/df = 1.33; CFI = 0.997; TLI = 0.989; RMSEA = 0.046). The McDonald’s omega was 0.823.

3.2.3. Employee Promotive-Voice Behavior (Mediator Variable)

In order to measure employee voice behavior, the scale developed by Liang et al. (2012) was used, consisting of 5 items and ranging on a Likert response scale between 1 (never) to 7 (always). A sample item is “I proactively suggest new projects that are beneficial to the organization”. To test measure unidimensionality we ran a single-factor CFA which indicates a good fit (χ²(4) = 1.422, p = 0.840; χ²/df = 0.35; CFI = 0.1.000; TLI = 1.012; RMSEA = 0.000). The McDonald’s omega was 0.918.

3.2.4. Trust in the Leader (Moderator Variable)

To measure employee trust in the leader, the scale developed by Fehr et al. (2020) was used. The scale used a Likert response scale, ranging from 1 (strongly disagree) to 7 (strongly agree). The scale had 4 items, including “I believe that my direct management would care for my well-being”. The unidimensionality of the measure was tested and the results indicate an excellent fit (χ²(1) = 0.237, p = 0.627; χ²/df = 0.24; CFI = 1.000; TLI = 1.013; RMSEA = 0.000). The McDonald’s omega was 0.884.

3.2.5. Sociodemographic Variables

Besides all the variables mentioned above, more data were retrieved regarding the employees’ sociodemographic characteristics, such as age, gender, their relationship with the leader, such as tenure of the relationship and interaction, as well as data regarding their organization, such as type of organization and sector.

3.3. Discriminant and Convergent Validity

CFA was conducted to check whether the items of the four variables capture distinct constructs. The four-factor measurement model yielded acceptable fit indexes (i.e., CFI = 0.899; TLI = 0.890; RMSEA = 0.061). In contrast, the single-factor model and two other alternative models produced unacceptable fit statistics (Hu and Bentler 1999; Marsh et al. 2004) (see

Table 1. Fit indices.

Models	χ2 (df) p-Value	χ2/df	CFI	TLI	RMSEA
Model 1: Four-factor model (PL + Trust + EPVB + KS)	849.442 (543) p < 0.001	1.56	0.899	0.890	0.061
Model 2: Three-factor model (PL + Trust and EPVB + KS)	1273.601 (547) p < 0.001	2.32	0.761	0.740	0.093
Model 3: Three-factor model (PL + Trust + EPVB and KS)	972.010 (547) p < 0.001	1.77	0.860	0.848	0.071
Model 4: Single-factor model (all merged)	1812.064 (527) p < 0.001	3.44	0.557	0.528	0.126

Notes. df = degrees of freedom; CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error of approximation; PL = paradoxical leadership; Trust = trust in the leader; EPVB = employee promotive-voice behavior; KS = knowledge sharing.

).

To further ensure discriminant validity, average variance extracted (AVE) values were estimated and compared to the squared correlations between all pairs of variables. The composite reliability (CR) values ranged from 0.83 to 0.93, exceeding the recommended cutoff point of 0.70, as suggested by Hair et al. (2010). The AVE values are also above the threshold of 0.50, as recommended by Fornell and Larcker (1981), except for knowledge sharing (0.35) (see

Table 2. Means (Ms), standard deviations (SDs), correlations, McDonald’s omega, squared correlations, composite reliability (CR), and average variance extracted (AVE) values.

	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	CR	AVE
1. Gender (a)	--	--	--
2. Age	37.03	13.09	0.07	--
3. Education (b)	--	--	−0.18 *	−0.23 **	--
4. Tenure	9.03	11.21	0.01	0.71 **	−0.26 **	--
5. Dyad tenure	4.96	6.69	−0.04	0.60 **	−0.23 **	0.87 **	--
6. Dyad interaction (c)	--	--	−0.00	−0.16	0.09	0.01	0.02	--
7. Professional status (d)	--	--	0.03	0.42 **	−0.13	0.48 **	0.43 **	−0.17 *	--
8. Organization dimension (e)	--	--	0.01	0.19 *	0.07	0.14	0.01	−0.06	0.16 *	--
9. Knowledge-sharing behavior	5.99	0.73	−0.10	0.26 **	0.07	0.29 **	0.27 **	0.07	0.20 *	0.10	(0.823)	0.15	0.04	0.04	0.83	0.35
10. Employee promotive-voice behavior	4.57	1.07	0.16	0.15	−0.07	0.12	0.09	0.01	0.12	−0.10	0.39 **	(0.918)	0.03	0.05	0.91	0.68
11. Trust in the leader	4.40	1.38	0.02	−0.03	−0.00	−0.06	−0.13	−0.11	−0.05	0.10	0.20 *	0.18 *	(0.884)	0.49	0.87	0.64
12. Paradoxical leadership	5.07	0.91	0.11	−0.05	0.06	−0.07	−0.12	−0.02	−0.02	0.01	0.20 *	0.23 **	0.67 **	(0.916)	0.93	0.56

Note. ^(a) Gender: 0 = female, 1 = male; ^(b) Education: 1 = 9 years of schooling, 2 = 12 years of schooling, 3 = degree, 4 = master degree, 5 = Ph.D.; ^(c) Dyad interaction: 1 = daily, 2 = 4 to 6 times/week, 3 = 2 to 4 times/week, 4 = once a week, 5 = biweekly, 6 = monthly; ^(d) Professional status: 0 = fixed-term contract, 1 = permanent contract; ^(e) Organization dimension: 1 = up to 9 employees, 2 = between 10 and 50 employees, 3 = between 51 and 250 employees, 4 = more than 250 employees; McDonald’s omega in parentheses; Squared correlations in bold; * p < 0.05, ** p < 0.01.

). However, according to Fornell and Larcker (1981, p. 46), researchers can still “conclude that the convergent validity of the construct is adequate, even if more than 50% of the variance is due to error”, when the variable’s CR values surpass the recommended minimum. Overall, results indicate adequate discriminant and convergent validity.

4. Results

Table 2 presents the descriptive statistics and Spearman’s correlation coefficients for the study’s variables. The results show that these are all significantly intercorrelated producing low to moderate correlation values (all p < 0.05).

Participants’ age, tenure in the organization, and tenure in dyad are positively correlated with knowledge sharing, suggesting that older participants and those working for more years in the organization and with the same leader share their knowledge more frequently than younger workers or those working more recently in the company or dyad. Professional status also reveals a positive relationship with knowledge sharing, suggesting that workers with a permanent employment status are more prone to share their knowledge than those who are in more unstable professional situations. Hence, these four socioprofessional variables were included as covariates in subsequent analyses.

The hypothesis testing for the moderated mediation model was assessed using the PROCESS Macro for IBM SPSS Statistics 28 software (Hayes 2022; Igartua and Hayes 2021). Model 4 was used to calculate the total and indirect effect of paradoxical leadership on knowledge sharing, while Model 7 served to obtain the results of the moderated mediation analysis (see

Table 3. Moderated mediation analysis.

	Model 1 Employee Promotive-Voice Behavior (M)			Model 2 Knowledge-Sharing Behavior (Y)
	b	SE	LLCI, ULCI	b	SE	LLCI, ULCI
Constant	3.967	0.318	3.229, 4.597	4.381	0.286	3.817, 4.946
Paradoxical leadership (X)	0.316	0.139	0.041, 0.591	0.087	0.059	−0.031, 0.204
Employee promotive voice behavior (M)	--	--	--	0.248	0.050	0.148, 0.348
Trust in the leader (W)	0.065	0.088	−0.108, 0.238	--	--	--
X*W	0.160	0.054	0.053, 0.266	--	--	--
Age	0.009	0.010	−0.011, 0.029	0.008	0.006	−0.005, 0.021
Tenure	0.007	0.013	−0.017, 0.032	0.001	0.008	−0.015, 0.017
Dyad tenure	−0.004	0.017	−0.037, 0.029	0.012	0.010	−0.008, 0.032
Professional status (a)	0.120	0.217	−0.310, 0.550	0.140	0.136	−0.123, 0.409
	R2 = 0.127; F(7, 146) = 3.042, p < 0.01			R2 = 0.259; F(6, 147) = 8.546, p < 0.001
Conditional indirect effects	Effect	BootSE	95% BootCI
Low trust in the leader (−1 SD)	0.024	0.037	−0.059, 0.091
Middle trust in the leader (0 SD)	0.078	0.041	0.008, 0.168
High trust in the leader (+1 SD)	0.133	0.059	0.032, 0.262
	Index	BootSE	95% BootCI
Index of moderated mediation	0.040	0.020	0.007, 0.087

Note. ^(a) Professional status: 0 = fixed-term contract, 1 = permanent contract; b = unstandardized effect; SE = Standard error.

).

Hypothesis 1 proposed that paradoxical leadership is positively related to employee knowledge-sharing behavior. In conformity, paradoxical leadership’s total effect on the criterion variable is statistically significant (unstandardized coefficient [b] = 0.152, p < 0.05; R² = 0.107; F(5, 148) = 4.674, p < 0.001, confirming that employees’ perception that their leaders adopt paradoxical behaviors fosters their tendency to share knowledge in the workplace (Table 3). Hypotheses 1 was thus supported.

Hypothesis 2 foresaw a mediating effect of employee promotive-voice behavior in paradoxical leadership’s relationship with employee knowledge-sharing behavior. The results show that paradoxical leadership significantly increases employee promotive-voice behavior (b = 0.316; 95% confidence interval [0.041; 0.591]) and the latter variable also positively affects knowledge sharing (b = 0.248; 95% confidence interval [0.148; 0.348]). Moreover, paradoxical leadership’s indirect effect via promotive voice on reported knowledge-sharing behavior is also significant (b = 0.078; 95% confidence interval [0.008, 0.168]), which supports the proposed hypothesis. As the predictor variable’s direct effect (b = 0.09; n.s.) is no longer statistically significant, the findings indicate that the mediation is full.

As for Hypothesis 3, it proposed that trust in the leader would moderate the relationship between paradoxical leadership and employee promotive voice behavior. In other words, the higher the trust in the leader, the stronger becomes the association between paradoxical leadership and promotive-voice behavior. As seen in Table 3, while trust in the leader alone does not affect an employee’s propensity to engage in voice behaviors (b = 0.065; 95% confidence interval [−0.108, 0.238]), it interacts significantly with paradoxical leadership (b = 0.160; 95% confidence interval [0.053; 0.266]). Hypothesis 3 was thus supported, indicating that the relationship is stronger when employees’ level of trust is high rather than when their trust levels in their immediate supervisor are lower. Figure 2 depicts the conditional effect of paradoxical leadership on employee promotive voice behavior across different values of trust in the leader.

Moreover, Hypothesis 4 was also supported by the present sample’s data. The findings confirm that trust in the leader also moderates the indirect effect of paradoxical leadership on knowledge sharing via increased employee promotive-voice behavior (b = 0.040; 95% confidence interval [0.007; 0.087]). The conditional indirect effects presented in Table 3 indicate that this effect exists only when employees have middle and high levels of trust in the leader, but not when the level of trust is low. Figure 3 depicts the visual presentation of the linear function relating trust in the leader to the indirect effect of paradoxical leadership on knowledge sharing through employee promotive-voice behavior.

Overall, the full model explains 26% of the unique variance of employees’ knowledge sharing behaviors (F [6,147] = 8.546; p < 0.001). None of the covariates’ effects on the criterion variable were statistically significant, as shown in Table 3. Figure 4 depicts the theoretical model supported by results of moderated mediation analysis.

5. Discussion

5.1. Main Findings

This study sought to determine how employees’ knowledge-sharing behaviors can be explained by their leader’s paradoxical-leadership behavior via workers’ increased engagement in promotive-voice behaviors, while considering the interactive role of trust in the leader in the observed relationships. Four hypotheses were submitted to empirical validation and supported by the finding of a two-wave time-lagged correlational study.

As Hypothesis 1 suggested, these employees’ views of their leader’s paradoxical behaviors are positively related to their self-reported knowledge-sharing behaviors. This result is in line with the findings of Devi’s (2023) and underlines that paradoxical leadership is a significant facilitator of individuals’ choice to share their knowledge with others. By balancing the different aspects of a paradoxical relationship with subordinates, including maintaining decision control while allowing autonomy and ensuring compliance with work requirements while allowing flexibility, a paradoxical leader might emanate the signal that knowledge sharing is valued by the organization. This is consistent with signaling theory (Connelly et al. 2011; Spence 1978), which stresses the importance of managers communication about aspects valued by the organization. The leader’s efforts to balance different challenges may stimulate workers’ openness to share knowledge with others and thereby achieve the company’s operational and business goals.

As for Hypothesis 2, it proposed that employee promotive-voice behavior significantly mediates the positive relationship between paradoxical leadership and their knowledge sharing behaviors. Recent studies have linked paradoxical leadership to increased promotive-voice behavior (Kundi et al. 2023; Li et al. 2020; Rescalvo-Martin et al. 2021), and the latter variable has been connected to intensified knowledge sharing (Alzyoud et al. 2024; Narayanan and Nadarajah 2024). Hence, paradoxical leadership was expected to encourage employees feel more willing to speak up about improvement opportunities and suggestions and thus report more knowledge-sharing behaviors. The present study’s results support Hypothesis 2, identifying promotive voice as a relevant mediator of paradoxical leadership and knowledge sharing’s relationship.

Hypothesis 3 also received empirical support since trust in the leader proved to be a significant moderator of the positive relationship between paradoxical-leadership behavior and employee promotive voice. The prior literature suggested that follower–leader trust is conducive of extra-role behaviors (Agarwal 2014; May et al. 2004), including voice behavior (Carnevale et al. 2017; Van Dyne et al. 2008), notably because it contributes to increased psychological safety (Duan et al. 2019; Madjar and Ortiz-Walters 2009; May et al. 2004). Differently, in this study trust in the leader was proposed to interact with paradoxical leadership and reinforce its positive effect on employee promotive-voice behaviors. Accordingly, trust in the leaders would work as a shield preventing employees from doubting leaders’ intentions based on their contradictory behavior (Yang et al. 2024). The results support the proposed interactive effect, indicating that when trust in the leader is high, workers are more likely to voice rather than when it is low. In high trust situations, employees probably perceive that it is safe and feel free to speak up honestly (Jiang et al. 2018), accepting the inherent risk of such behavior based upon positive expectations of their leaders (Detert and Treviño 2010; Silla et al. 2020). In opposite ways, low trust situation should increase the risk perception, limiting the engagement in voice behaviors because of fear of negative consequences for the self.

Moreover, trust in the leader was also found to moderate the indirect effect of paradoxical leadership on knowledge sharing behaviors via increased promotive voice behaviors. Trust in other organizational members has been settled as a relevant enabler of knowledge sharing (Farooq and Sultana 2021; Nadeem et al. 2021; Rutten et al. 2016). The current study’s findings suggest that when employees trust their paradoxical leaders, they are more likely to understand that the leader welcome ideas on how to advance company’s practices and processes, and thus are more willing to share their expertise, knowledge, and experiences with others. This reinforcing effect, however, does not occur if trust in the leader is low, probably because workers cannot clearly interpret leaders’ expectations regarding these proactive extra-role behaviors, and consistent with arguments presented above, have higher risk perceptions and are fearful of negative consequences. Thus, Hypothesis 4 was supported by current study’s findings.

The effects between these variables were found to be statistically significant as they explain nearly a quarter of the variance in workers’ willingness to share their knowledge. Overall, little information is available about the boundary conditions and processes underlying how and when leaders’ paradoxical behavior shapes employee extra-role behavior, especially regarding employees’ knowledge sharing, so these findings provide added value to this field of study.

5.2. Theoretical Contributions and Implications for Management

The present study’s results expand the knowledge about paradoxical leadership as an important factor affecting employee discretionary behaviors (not only promotive-voice behavior but also knowledge-sharing behaviors) and identify trust in the leader as a relevant boundary condition that helps paradoxical leader’s behavior to be more effective. This leadership behavior has recently been found to be a key predictor of knowledge sharing (Devi 2023)—a phenomenon that contributes to companies’ effective knowledge management processes. By sharing their personal knowledge, expertise, and experiences with other organizational members, workers contribute to stronger interpersonal relationships at work, organizational learning (Farooq 2018; Swift and Hwang 2013), creativity, and innovation processes (Devi 2023; Hussain et al. 2017; Zhang et al. 2022). These are relevant factors in helping managers and their teams successfully deal with the paradoxes and challenges that organizations face today.

From a practical perspective, the above findings indicate that managers can adopt more paradoxical-leadership behavior to increase their workers’ knowledge-sharing behaviors. Managers that adopt this type of behavior can strengthen their subordinates’ promotive voice and thus enhance their knowledge-sharing behaviors. As proposed by Zhang et al. (2015), this leadership behavior implies balancing five competing needs, namely treating subordinates uniformly while allowing individualization; combining self-centeredness with other-centeredness; maintaining decision control while allowing autonomy; enforcing work requirements while allowing flexibility; and maintaining both distance and closeness. Achieving the necessary level of balance can be demanding. So, organizations should provide training opportunities to help managers to better understand how to deal with increasing uncertainties that often involve competing possibilities in terms of managerial actions (Zhang et al. 2015).

In addition to drawing managers’ attention to theories that support paradoxical thinking and action, training interventions can focus on soft skills that promote high quality leader–follower relationships. An important result of this study lies in the fundamental role of trust in leader as a boundary condition of paradoxical-leadership behavior’s effects on voice and knowledge-sharing behaviors. The effectiveness of paradoxical leadership behavior in promoting these extra-role behaviors depends on employees’ levels of trust in their leaders. When there is mistrust, employees avoid expressing their ideas and sharing their knowledge. This means that organizations and leaders must cultivate a climate of trust within the organizational environment, where people feel safe to express their voices and share their knowledge without fear of being criticized or penalized by engaging in such behaviors. Trust in the leader is a key factor in supporting the positive outcomes that this contemporary leadership style can bring to people management, including in terms of knowledge management processes.

5.3. Limitations and Future Research

The above findings provide significant added value, but they must be interpreted with care given this study’s limitations in three main areas: the research design and data collection procedure, measures, and the sample. First, despite that the risk of CMV occurrence was diminished using a two-wave time-lagged data collection procedure and other procedural (e.g., reducing evaluation apprehension) and statistical remedies (e.g., ascertaining measures’ discriminant and convergent validity), the correlational design limits definitive conclusions about causal relationships. Therefore, future studies can re-examine the relationships between variables along the time axis using a longitudinal design to further support the causal nexus here proposed.

Second, the questionnaire needed to be relatively parsimonious to encourage participation and the completion of all the items, which meant that the number of variables included was restricted. Broader results could have been obtained by using variables focused also on interpersonal relationships with coworkers, such as trust in the team and psychological safety at the team level, which can also be relevant to employee’s willingness to engage in voice (Chamberlin et al. 2017; Morrison 2023) and knowledge-sharing behaviors (Lee et al. 2023a).

Third, the sample was relatively small and non-probabilistic, so the results’ generalizability is limited. Future studies could use more representative samples to increase the findings’ stability. The origin of the sample, i.e., a south-European country, is nevertheless an advantage, since most research on paradoxical-leadership behavior has been carried out in Eastern countries, mainly with Chinese samples. As Zhang et al. (2015) argued, more research is needed to see if results can be generalized to other populations.

Besides overcoming the above-mentioned limitations, future research can further explore the effects of paradoxical leadership in knowledge management processes by including knowledge-hiding behaviors in the model. As recently proposed by Chin et al. (2024), employees’ knowledge-sharing and -hiding behaviors can be seen as having a paradoxical interrelationship, with the two types of behavior being motivated by different factors. Establishing how and when paradoxical leadership can be effective in fostering sharing behaviors while demotivating hiding ones can be relevant for a company’s improved knowledge management process.

6. Conclusions

In conclusion, organizational knowledge sharing is crucial in today’s fast-paced and highly competitive business environment. The findings in this study emphasize that paradoxical leadership can be an effective way to foster intensified willingness to speak up about opportunities for improving organizational functioning and ultimately results in better employee knowledge-sharing behaviors. However, this effect is contingent on the trust employees have in their leaders. Leaders play a fundamental role in creating environments that encourage knowledge sharing, which drive innovation, create value, and strengthen a culture of continuous learning. This, in turn, positively impacts employee engagement and motivation. This highlights the importance of establishing a precondition based on individual employee perceptions, such as trust in the leader, to create a space where successful sharing can occur within the organization.