Servant leadership effect on knowledge-hiding behavior: the moderated mediation influence of hope and burnout

In today’s increasingly volatile, uncertain, complex, and ambiguous (VUCA) knowledge economy, knowledge has emerged as a key resource for both individual and organizational success (Ali et al., 2018). Organizations, therefore, expect employees to proactively acquire and share knowledge to enhance effectiveness (Meher et al., 2024). However, since knowledge resides with individuals, organizations cannot mandate its exchange. Employees often withhold knowledge, creating significant challenges for knowledge-sharing (Karatepe et al., 2019), which can contribute to inefficiencies such as duplicated work, reduced productivity, and diminished performance (Cerne et al., 2014).

Connelly et al. (2012) define knowledge-hiding as an intentional effort to withhold knowledge when requested. Although initially studied in Western, knowledge-intensive industries (Gagne et al., 2019), knowledge-hiding is also prevalent in service sectors of emerging economies like India (Ravallion, 2016). A survey of 1,700 employees found that 76% had witnessed such behavior in their workplaces (Aljawarneh and Atan, 2018). While empirical research on knowledge-hiding is expanding, knowledge management literature still primarily focuses on knowledge-sharing. Most studies emphasize the antecedents and situational drivers of knowledge-hiding (Ali et al., 2021; Connelly et al., 2012), but few examine the role of leadership behaviors in shaping this phenomenon. Understanding how leadership influences knowledge-hiding is important, given the frequency of manager–employee interactions in organizational settings (Hislop, 2002). Research has examined the impact of ethical leadership, abusive supervision, and transformational leadership on knowledge-hiding (Scuotto et al., 2022). However, studies exploring the influence of servant leadership remain limited. This gap is noteworthy, as leaders are central organizational resources whose behaviors can influence employee attitudes and actions (Hoch et al., 2018). The overall research question is formulated as: “How and under what conditions does servant leadership predict employee knowledge-hiding behavior via the mediating role of burnout and how is this indirect relation contingent upon employees’ levels of hope?”

Servant leadership—a style characterized by prioritizing followers' needs and fostering their growth and development (Liden et al., 2015)—may be relevant in understanding knowledge-hiding behaviors. Its other-centered focus and emphasis on followers’ moral and personal development position servant leaders to address both social and psychological needs in the workplace (Eva et al., 2019; Hoch et al., 2018). These leaders often cultivate prosocial behavior among employees, which may reduce tendencies to withhold knowledge. Through their support-oriented approach, servant leaders may provide resources that help employees manage stress and resist peer pressure to engage in knowledge-hiding.

Despite this potential, the mechanisms linking servant leadership to knowledge-hiding remain underexplored. Limited research has examined how servant leadership might reduce counterproductive behaviors like knowledge-hiding. To address this gap, we propose employee burnout as a potential explanatory mechanism. Burnout—characterized by emotional exhaustion, depersonalization, and reduced professional efficacy—has negative implications for both individuals and organizations (Maslach et al., 2001). It is associated with lower job performance, decreased satisfaction, diminished commitment, and higher absenteeism (Koo et al., 2020), and has been linked to counterproductive work behaviors (Smoktunowicz et al., 2015). We suggest that servant leaders may help reduce burnout through behaviors such as active listening, empowerment, and individualized support, which contribute to trusting relationships and improved employee well-being (Upadyaya et al., 2016). Their focus on personalized attention and the welfare of followers (Eva et al., 2019) may mitigate the emotional strain that fosters knowledge-hiding tendencies. Thus, burnout may serve as a mediating factor in the relationship between servant leadership and knowledge-hiding.

An important dimension of this study is its examination of hope as a moderating factor in the relationship between servant leadership and employee burnout. While prior research has typically positioned hope in leaders as a mediating construct (Belen et al., 2020), its role as a contextual moderator has received limited empirical attention (Visser et al., 2013). Drawing on hope theory, we incorporate hope—defined as goal-directed thinking involving both the motivation to pursue goals and the pathways to achieve them (Snyder et al., 2003)—to explore when servant leadership may be more or less effective in influencing employees' knowledge-hiding behaviors. Individuals high in hope are generally more capable of identifying alternative strategies to achieve goals, particularly when initial efforts are obstructed (Irving et al., 1998). Therefore, employees with high levels of hope may be better equipped to manage burnout and navigate situations involving peer knowledge-hiding requests in constructive ways. We theorize that that there exists a negative relationship between servant leadership and employee burnout, and that the latter will lead to a decrease in the level of knowledge-hiding behavior among employees. We further theorize that hope serves as a personal resource that acts as a moderator in protecting against the adverse impact of burnout. More specifically, we contend that the degree to which employees possess hope moderates the indirect effect of servant leadership on knowledge-hiding behavior through burnout, indicating that the indirect effect differs depending on the degree of hope.

We ground our model in conservation of resources (COR) theory (Hobfoll, 1989; Hobfoll et al., 2018), which posits that individuals strive to retain, protect, and build valued resources. The loss or threat of resource depletion often leads to stress and counterproductive behaviors. Within this framework, servant leadership functions as a key source of psychological and social resources—such as empowerment, emotional support, and individualized attention—that can help employees better manage workplace demands and reduce burnout (Yao et al., 2021). Figure 1 presents the proposed research model linking servant leadership to knowledge-hiding behavior via burnout at high and low levels of hope.

This study contributes to several scholarly streams, including leadership, burnout, hope, and knowledge-hiding. First, it extends the literature on knowledge-hiding by exploring how servant leadership interacts with hope to have an influence on knowledge-hiding via burnout, offering a more advanced understanding of the complex dynamics between leadership and discretionary knowledge behaviors (Eva et al., 2019; Lee and Ashford, 1996). Second, by introducing burnout as a mediating mechanism, it builds on existing research that has primarily focused on external or interpersonal factors and brings attention to an individual psychological pathway through which servant leadership may influence knowledge-hiding (Hoch et al., 2018). Third, this research contributes to the growing literature on hope by examining how hope of employees mitigates not only the relationship between servant leadership and employee burnout but also the indirect relationship between servant leadership and knowledge-hiding behavior through burnout, thereby furthering hope theory by showing that hope acts as a compensating personal resource that replaces leader’s socioemotional support when stress is present (Hobfoll, 1989; Snyder, 2002) and by providing a contingency approach to hope theory (Luthans et al., 2007; Snyder, 2002). Finally, the findings have particular relevance within the Indian context, where limited empirical work has addressed the role of hope in employee behavior. By addressing cultural differences—such as higher power distance and collectivist orientations (Hofstede, 1980)—this research responds to calls for more non-Western studies on leadership styles (Mittal and Dorfman, 2012). It highlights how servant leadership, shaped by cultural context, may operate differently across regions and sectors.

The originality of this research lies in its proposal of an alternative view on how and when servant leadership mitigates knowledge-hiding behaviors among organizational employees. While previous work has taken servant leadership as a universal positive force, this study recognizes that employee burnout plays a crucial role in linking servant leadership and knowledge-hiding. What’s even more important is that this study recognizes employee hope as a contingency factor that determines the effectiveness of servant leadership in mitigating knowledge-hiding via burnout. As employees who are less hopeful about future success are likely to rely more on leader support for overcoming burnout, servant leadership is likely to have a more pronounced impact on those employees. This study, thus, adds new knowledge by offering a contingency approach to studying the effects of servant leadership and employee burnout on knowledge-hiding behaviors.

This study is anchored in COR theory (Hobfoll, 1989; Hobfoll et al., 2018), which provides a comprehensive framework for understanding employee behavior under resource strain. COR theory posits that individuals are motivated to acquire, protect, and retain resources they value—whether personal (e.g. knowledge, skills) or organizational (e.g. supervisor support, autonomy). Resource loss, or even the threat of it, can trigger psychological stress. When such stress becomes chronic, it may manifest as burnout. In high-strain environments, employees may respond by conserving remaining resources—often through defensive strategies such as knowledge-hiding (Tian et al., 2022).

Burnout, as conceptualized in COR theory, results from prolonged exposure to resource-draining situations, including performance pressure, job insecurity, or unmet expectations despite significant effort (Hobfoll, 1989). Under such conditions, employees may perceive knowledge as a critical resource worth protecting. Empirical evidence suggests that employees experiencing burnout are more likely to withhold knowledge to avoid further depletion (Ahmad et al., 2021). Thus, knowledge-hiding can be understood as a self-protective behavior aligned with COR theory’s core propositions. In this context, servant leadership may serve as a resource-enhancing factor that buffers the effects of resource loss. By emphasizing support, empowerment, and individual development, servant leaders may help replenish employee resources and reduce emotional exhaustion.

Based on COR theory, hope is regarded as an important personal resource that influences employees’ dependency on servant leadership as a means to cope with burnout. According to COR theory, individuals who have rich internal psychological resources are more resilient against the loss of resources and hence are independent of external contextual resources (Hobfoll, 1989). As hope involves agency and pathway thinking, it equips individuals with the ability to motivate themselves and find alternate coping mechanisms when under stress (Snyder, 2002). Hence, individuals who are high in hope are well-placed to manage their emotional exhaustion through self-regulation. This makes servant leadership redundant in addressing issues of burnout among such individuals. However, individuals with low hope lack adequate internal coping resources and thus stand to gain more from the supportive nature of servant leaders. By using hope as the first-stage moderator, this study goes further than traditional servant leadership studies and proposes a resource contingency framework for the indirect reduction of knowledge-hiding behavior.

Knowledge-hiding refers to the deliberate concealment of knowledge when it is requested by others (Connelly et al., 2012; Scuotto et al., 2022). It differs from simply failing to share knowledge—knowledge-hiding is intentional and strategic. It includes behaviors such as playing dumb, evasive hiding, and rationalized hiding (Connelly et al., 2012), all of which involve varying degrees of misrepresentation or deflection. Rather than an incidental behavior, knowledge-hiding often emerges from a desire to protect personal advantage, manage workload, or avoid perceived risks. It can be both an outcome of psychological strain and a mechanism to preserve or defend one's resource base, aligning closely with COR theory's predictions.

Burnout refers to a state of emotional and physical exhaustion resulting from prolonged workplace stress (Bakker and Demerouti, 2007). It is commonly linked to high job demands and insufficient resources (Maslach et al., 2001; Bakker and Demerouti, 2007), and is associated with negative outcomes such as reduced performance, job dissatisfaction, and increased absenteeism (Koo et al., 2020). Given these consequences, scholars have emphasized the need to examine leadership approaches that may alleviate burnout, with servant leadership identified as a promising candidate (Eva et al., 2019). As a form of resource-enhancing leadership, servant leadership centers on understanding and addressing followers' needs through personal interaction and support (Greenleaf, 1970). Servant leaders aim to foster the development, motivation, and self-efficacy of their employees, often encouraging them to adopt similar values (Karatepe et al., 2019). Through this empowering approach, employees may become more engaged and less susceptible to emotional exhaustion or disengagement.

Servant leadership contributes to a psychologically safe environment where mistakes are tolerated, feedback is constructive, and interpersonal support is prioritized (Eva et al., 2019; Karatepe et al., 2019). Open communication and leader availability reinforce trust and a sense of security among employees, which may reduce the risk of burnout (Iqbal et al., 2020). Meta-analytic findings suggest that supportive workplace resources—including leadership support—are inversely associated with burnout (De Simone et al., 2021). From a COR theory perspective, servant leadership acts as a key organizational resource that helps employees offset the effects of job demands. The behaviors associated with servant leadership—such as empathy, accessibility, and concern—provide essential psychological resources that can mitigate stress and prevent resource depletion. When employees perceive their leaders as trustworthy and supportive, they may be better equipped to cope with adverse situations, reducing the likelihood of burnout (Babakus et al., 2010; Iqbal et al., 2020). Accordingly, we propose:

Knowledge-hiding, though more recently formalized as a research construct, has long existed in organizational settings (Davenport and Prusak, 1998). Unlike the simple absence of knowledge-sharing, knowledge-hiding involves a deliberate decision to withhold or obscure information, particularly when requested (Connelly et al., 2012). As knowledge is both a valuable personal asset and a source of power, employees may conceal it to maintain an advantage, especially when they perceive sharing as a threat to their position or credibility (Cerne et al., 2014).

Employees experiencing burnout may be especially inclined to protect their resources—both cognitive and emotional—by avoiding the additional demands associated with knowledge-sharing. COR theory suggests that when individuals face stressors such as emotional exhaustion or psychological strain, they engage in behaviors aimed at conserving remaining resources. In this context, knowledge-hiding can be seen as a defensive strategy used to reduce social and cognitive demands (Tian et al., 2022). Burnout may also increase disengagement, frustration, and a sense of detachment from the organization, leading to reduced motivation to cooperate with colleagues (Ali et al., 2021). Employees experiencing chronic fatigue or low psychological energy are less likely to actively contribute knowledge and more likely to withdraw or withhold it as a form of self-protection (Aljawarneh and Atan, 2018). These individuals may perceive knowledge-sharing as an energy-depleting task that provides little immediate return. In line with this reasoning, we propose:

Hope is a cognitive-motivational concept that consists of two important aspects: agency thinking (being determined in attaining goals) and pathway thinking (the perception of capability to search for ways to reach desired objectives) (Snyder, 2002). Contrary to emotion, hope serves as an internal motivational resource assisting people to stay motivated and plan their goal attainment efforts (Snyder et al., 1991, 2003). This construct can be considered as both a trait and a state depending on situational context, and it is highly relevant to overcoming difficulties and staying psychologically healthy (Colla et al., 2022; Rand and Touza, 2021). Hopeful employees tend to see obstacles from different perspectives, persevere to reach desired goals, and look for new ways to continue pursuit of objectives when faced with challenges (Irving et al., 1998; Luthans et al., 2007). Higher levels of hope have been associated with improved job performance, satisfaction, and organizational commitment (Ozyilmaz, 2020), along with lower levels of depression and burnout (Sherwin et al., 1992; Snyder et al., 2003; Surachartkumtonkun et al., 2023).

As suggested by COR theory, people attempt to acquire, keep, and defend valuable resources, and one of the basic principles of COR is that resources have the potential to be substitutes that protect people from resource loss. In light of this, it should be understood that hope is an important internal psychological resource that allows workers to manage stress, energize goal-related behavior, and avoid resource depletion. Workers with high levels of hope thus have greater internal resource reserves that help them better handle organizational challenges and reduce dependence on external contextual resources. Workers with lower levels of hope, on the other hand, do not have enough internal psychological resources; hence, they are susceptible to resource loss and exhaustion. Consequently, these workers become more reliant on external resource-giving systems like servant leadership to compensate for their deficiency in internal resources. Through its focus on emotional support, empowerment, empathy, and individualized consideration, servant leadership acts as an external resource caravan for exhausted workers.

Accordingly, the impact of servant leadership as an anti-burnout resource is contingent on the presence of internal resources in the employee (Hope). If employees have a lot of internal resources in terms of hope, servant leadership will not have much extra value, since the employees are well-endowed with enough internal resources for coping with stressful situations. Therefore, the negative association between servant leadership and burnout will be attenuated at the high level of hope (High hope → weaker negative relationship between servant leadership and burnout). On the other hand, when employees have insufficient internal resources (Hope), servant leadership serves as an important compensatory resource (Low hope → stronger negative relationship between servant leadership and burnout). Based on the same substitution principle, our claim is that hope serves as a moderator for the influence of servant leadership on burnout, implying that the ability of servant leadership to lower employee burnout is greater when hope is lessened and lesser when hope is increased. Thus, we suggest that:

According to COR theory, individuals are driven to conserve, safeguard, and accumulate valued resources, especially in situations where stress and uncertainty prevail (Hobfoll, 1989). The assumption behind the COR theory is that people who have higher resource endowments can endure losses of these resources better than those whose resources are few and are likely to fall into a cycle of resource loss (Iqbal et al., 2020; Luthans et al., 2007). Under this model, hope serves as an important internal resource for psychological resilience that consists of pathways and agency thinking, providing people with the ability to stay motivated, come up with alternative solutions, and be psychologically resilient despite the difficulties they encounter. For this reason, employees with high levels of hope have better internal resource pools from which they can draw to deal with stressful situations and remain psychologically stable in spite of the difficulty of the working conditions. Due to such adequate internal resource pools, they need less support from external contextual resources such as servant leadership to avoid resource depletion and burnout. Conversely, individuals who have low levels of hope do not have many internal psychological resources, which makes them prone to burnout due to the lack of resources.

In line with the resource substitution hypothesis of the COR approach, the importance of external resources like servant leadership increases when the internal resources are deficient. The process of providing support in servant leadership is based on emotional aspects, empathy, and empowerment, as well as individual considerations (Lemoine et al., 2019). This type of leadership can be considered an external resource caravan that helps replenish the deficient internal resources of employees and avoid burnout. In other words, in cases when individuals lack internal psychological resources (low level of hope), servant leadership becomes more valuable, as it compensates for the lack of internal resources and prevents the loss of resources. At the same time, when the level of hope is high, individuals have sufficient internal resources that allow them to manage their working conditions properly, so servant leadership adds little value.

Thus, the moderating role of servant leadership on burnout is more prominent among employees who possess low hope and less pronounced among employees who have high hope. Since burnout serves as a mediator of servant leadership’s impact on knowledge-hiding behavior, the above moderation effect is translated into a conditional indirect effect. That is, the negative indirect effect of servant leadership on knowledge-hiding behavior via burnout would be more powerful among low-hope employees and less powerful among high-hope employees, given that the latter depend more on external sources of resources to avoid burnout and associated outcomes. Therefore, we hypothesize:

We gathered data from a wide range of Indian service sector companies, including those in information technology, healthcare, business consulting, logistics, academia, research and development, human resource management, and chemistry, to test our hypotheses. Our approach involved contacting human resources managers from each company through executive MBA students enrolled in a graduate management course at the first author's university. Subsequently, we sent emails to the heads of human resources at each organization, requesting their participation in the study. Upon obtaining consent to participate, we coordinated with designated personnel from the human resources departments to facilitate the survey. These contact persons then distributed the questionnaires via email within their respective organizations. This methodology enabled us to reach a diverse array of service organizations and their employees, spread across various remote regions of the country, using a cross-sectional research design that included a self-administered postal questionnaire via email. The objectives of the survey were clearly outlined in a cover letter, which assured participants of the anonymity and confidentiality of their responses and emphasized that the data would solely be used for academic purposes. The authors confirm that the study received ethical exemption from the research ethics committees of the affiliated universities and was conducted in compliance with the ethical standards set forth in the 1964 Declaration of Helsinki and its subsequent amendments, or similar ethical guidelines.

To test the hypotheses above, we collected data from 311 employees of 53 service sector organizations in India. Specifically, 605 questionnaires were administered, of which 311 filled responses were gathered, signaling a response rate of 51.40%. Data were gathered from April 2022 to June 2022. Participants held many different positions in their respective organizations, such as CEO, program manager, manager of business transformation, India logistics manager, site director, senior business associate, digital marketing manager, senior engineer, staff nurse, project manager, product engineer manager, estimation/procurement engineer, vice president, talent scout, branch manager, data analyst, and dental surgeon among many others.

Of the 311 participants, 68.5% were male, 77.2% were married, 48.9% had a graduate-level degree, and 86.5% were non-teleworker. The participants had 13.27 (Standard Deviation (SD) = 7.91) years of work experience in the industry and 6.46 (SD = −5.77) years of organizational tenure. 28.9% of participants were less than 30 years old, 39.9% were between 31 and 40 years old, 25.7% were between 41 and 50, and 5.5% were older than 50.

We used original scales in English because the participants had substantial knowledge of the English language. We used the following response categories for the scales: 1 = “strongly disagree,” 5 = “strongly agree.” The survey instrument containing all the scale items can be obtained from the first author.

Servant leadership. We used the 7-item servant leadership scale by Liden et al. (2008). We asked our participants to report, using a 5-point scale, the extent to which they agreed with each item. An example of an item was “I would seek help from my manager if I had a personal problem.” (α = 0.84)

Hope. The 12-item individual-differences hope scale by Snyder et al. (1991) was used to measure hope. We asked each respondent to report, using a 5-point scale, the extent to which they agreed with each item. An example of an item was “There are lots of ways around any problem” for the pathway dimension and “I meet the goals that I set for myself” for the agency dimension. (α = 0.84)

Burnout. We used the 16-item Oldenburg Burnout Inventory to measure burnout (Demerouti et al., 2010). We asked each respondent to report, using a 5-point scale, the degree of their agreement with each item. An example of an item was “I can tolerate the pressure of my work very well.” (α = 0.89)

Knowledge-hiding behavior. We measured knowledge-hiding using the 12-item scale (Connelly et al., 2012). We asked each respondent to report, using a 5-point scale, the degree to which they agreed with each statement. A sample item is “In this specific situation, I agreed to help my colleagues but never really intended to.” (α = 0.92)

Control variables. Consistent with the framework proposed by Bernerth and Aguinis (2016), we incorporated several theoretically relevant control variables. Formal education and work experience (organizational tenure and total work experience) were controlled, as they may contribute to the development of knowledge, skills, and abilities, potentially mitigating burnout and knowledge-hiding behavior (Becker, 1964; Ng and Feldman, 2010). Age was included, as younger employees might prioritize speed, thereby increasing burnout risk, whereas older employees may emphasize accuracy, potentially reducing burnout (Maslach et al., 2001). Marital status was considered, as married employees often face additional home-related responsibilities, which could exacerbate burnout levels (Combs et al., 2010). Gender was included due to its possible influence on both burnout and knowledge-hiding behavior (Maslach et al., 2001; Peng, 2013; Zhao and Xia, 2017). Lastly, working pattern (teleworker vs non-teleworker) was controlled for, as teleworkers may be more susceptible to burnout due to increased social isolation and the unique nature of their work environment (Barca et al., 2025). Age groups were coded as follows: 21–30 = 1, 31–40 = 2, 41–50 = 3, and greater than 50 = 4. Gender was encoded as male = 1, female = 0. Marital status was encoded as married = 1, single = 0. Work experience and organizational tenure were measured in years. Education was coded as undergraduate = 0, graduate = 1. The working pattern was coded as non-teleworker = 0, teleworker = 1.

Table 1 presents the summary of descriptive statistics and intercorrelations among the primary variables in the study. To assess the validity of our measures, we conducted confirmatory factor analysis (CFA) using the lavaan package (version 0.6–12) in R (version 4.2.1) with maximum likelihood estimation, prior to testing the hypotheses (Rosseel, 2012). Given the significant correlations between our control variables—age, marital status, and total work experience—and the latent variables, we controlled for them in all CFA analyses (Becker et al., 2016). So, four study variables and control variables were loaded on their respective latent variables which were allowed to covary.

As suggested by previous methodological recommendations, we used item parceling to improve model parsimony and the psychometric quality of the measurement model (Williams et al., 2026). Parceling is not universally required but is considered appropriate when constructs are unidimensional and measured via multiple items because it reduces random measurement error, improves indicator reliability, and produces more stable parameter estimates in structural equation modeling (Little et al., 2002; Matsunaga, 2008; Williams et al., 2009, 2026). In addition, parceling provides a more desirable indicator-to-sample size ratio and lowers model complexity, which is particularly advantageous when estimating models containing multiple latent constructs (Bandalos, 2002; Williams et al., 2026). In the current study, all focal constructs showed conceptual and empirical unidimensionality and thus parceling was employed to represent constructs at the intended level of abstraction rather than to conceal multidimensionality.

Following the item-to-construct balance approach (Little et al., 2002), we developed parcels that well represent the content domain of each construct and have also balanced factor loadings. Parcels were created at the dimension level for theoretically multidimensional constructs. Pathway and agency indicators were used to indicate hope. Evasive hiding, playing dumb, and rationalized hiding were used to indicate knowledge-hiding behavior. For example, the pathway indicator was derived as the average of the six individual items for that dimension. The same procedure was applied for the other dimensions.

For unidimensional constructs (servant leadership and burnout), items were randomly assigned to parcels based on their standardized factor loadings to ensure the equivalence of indicator strength. Specifically, the items with highest loadings were used to anchor each parcel and the remaining items were alternately assigned in descending order following Little et al. (2002). The use of a different number of parcels for servant leadership (two parcels) and burnout (four parcels) reflects differences in scale length and content coverage, not conceptual differences. Servant leadership was measured with seven items, so two parcels were enough to generate reliable indicators and keep the model parsimonious. In contrast, burnout was assessed with 16 items and the creation of four parcels yielded a more balanced representation of item content and reduced the risk of creating overly heterogeneous parcels (Bandalos, 2002; Matsunaga, 2008). This accords with suggestions that the number of parcels should depend on the number of items and the need to balance reliability and content validity (Bandalos, 2002; Little et al., 2002; Matsunaga, 2008; Williams et al., 2026).

The two parcels for servant leadership were named servant leadership 1 (λ = 0.888) and servant leadership 2 (λ = 0.910), the two parcels for hope were named pathway (λ = 0.663) and agency (λ = 0.971), and the four parcels for burnout were named burnout 1 (λ = 0.816), burnout 2 (λ = 0.796), burnout 3 (λ = 0.877), and burnout 4 (λ = 0.866). The three parcels of knowledge-hiding behavior were labeled as evasive hiding (λ = 0.710), playing dumb (λ = 0.957) and rationalized hiding (λ = 0.712). Using these factor loadings, the average variance extracted (AVE), and composite reliability (CR) were calculated for each construct (AVE = 0.81 and CR = 0.89 for servant leadership; AVE = 0.69 and CR = 0.81 for hope; AVE = 0.71 and CR = 0.91 for burnout; AVE = 0.64 and CR = 0.84 for knowledge-hiding behavior). All factor loadings were significant (p < 0.001), positive, and greater than 0.663.

We assessed convergent and discriminant validity using standard procedures for internal model evaluation in structural equation modeling. Convergent validity was assessed based on (a) the size and statistical significance of factor loadings, (b) AVE values greater than 0.50, and (c) CR values greater than 0.70, as recommended by Fornell and Larcker (1981) and Hair et al. (2019). All factor loadings (λs) were greater than 0.660 and statistically significant; all AVE values were above the 0.50 threshold; and all CR values were above the 0.70 threshold, all of which supported the convergent and discriminant validity of the constructs (Fornell and Larcker, 1981).

Moreover, overall model fit indices (χ², SRMR, CFI, RMSEA, IFI) were examined, since good model fit lends more support to the notion that the indicators are good proxies of their latent constructs (Bentler and Bonett, 1980; Hair et al., 2019; Kline, 2016; Lambert and Newman, 2023). The four-factor model thus represented an acceptable fit (χ² = 147.13, df = 59, p < 0.001; SRMR = 0.042; CFI = 0.96; RMSEA = 0.069; IFI = 0.96), suggesting convergent validity (Hair et al., 2019; Kline, 2016; Lambert and Newman, 2023).

We developed six parsimonious models (3-factor models) to test the discriminant validity of our measures. Accordingly, we claim discriminant validity when our 4-factor model provides a stronger fit than our 3-factor models (Bagozzi et al., 1991). Our 4-factor model fits the data significantly better than all 3-factor models where servant leadership and hope were combined (Δχ² (Δdf) = 99.69 (6), p < 0.00001), servant leadership and burnout were combined (Δχ² (Δdf) = 274.51 (6), p < 0.00001), servant leadership and knowledge-hiding behavior were combined (Δχ² (Δdf) = 435.89 (6), p < 0.00001), hope and burnout were combined (Δχ² (Δdf) = 113.90 (6), p < 0.00001), hope and knowledge-hiding behavior were combined (Δχ² (Δdf) = 268.13 (6), p < 0.00001), and burnout and knowledge-hiding behavior were combined (Δχ² (Δdf) = 371.95 (6), p < 0.00001), providing evidence of discriminant validity of our constructs (Bagozzi et al., 1991; Kline, 2016). These results provide strong evidence of discriminant validity.

Although the convergent and discriminant validity are sometimes assessed by the relationships among different scales, it is well known that they can be reliably assessed in the CFA framework with these internal criteria (Hair et al., 2019; Kline, 2016). We used the CFA marker variable technique to check the results of our study in order to rule out common method variance (Williams and McGonagle, 2016).

We used PROCESS v3.3 by Hayes (2022) to test our moderated mediation model. As a decision rule, we determined that moderated mediation will be accepted if the index of moderated mediation is significant with significant b, bootstrap standard errors, and 95% confidence intervals that do not involve zero obtained through 5,000 bootstrap samples (Hayes, 2022). Huber-White corrected heteroscedasticity-consistent inference was used to deal with heteroscedasticity. Servant leadership and hope were mean-centered for construction of the product (Aiken and West, 1991), and the moderating influence of hope was investigated at low (1 SD below the mean), mean, and high (1SD above the mean) levels of hope (Aiken and West, 1991; Kline, 2016). Following Becker et al.’s (2016) recommendation, we included age (r = 0.24, p < 0.01), marital status (r = 0.16, p < 0.01), and total work experience (r = 0.28, p < 0.01) as control variables in all analyses, as they were significantly correlated with the dependent variable, burnout. In contrast, we excluded gender, education, organizational tenure, and working pattern, as they were not significantly correlated with either of the dependent variables—burnout or knowledge-hiding behavior. Importantly, the results remained consistent regardless of whether only the three significant controls or all seven potential controls were included. Additionally, we tested the direct association of servant leadership on knowledge-hiding behavior, although we did not offer a direct relationship hypothesis. Accordingly, our simple regression analysis revealed no statistically significant direct relationship between servant leadership and knowledge-hiding behavior (b = 0.00, SE = 0.05, t = −0.00, p = 0.996). We also tested the interaction influence of hope on the relationship between servant leadership and knowledge-hiding behavior (b = 0.17, SE = 0.18, t = 0.96, p = 0.334). The test results showed no interaction effect of hope on the relationship between servant leadership and knowledge-hiding, showing that any impact of leadership on knowledge-hiding will likely be via indirect mechanisms, indeed, through a mediating relationship.

Table 2 presents the analysis of the mediating role of burnout and the moderating role of hope in the relationship between servant leadership and knowledge-hiding behavior. Accordingly, the results showed that the relation between servant leadership and burnout is positive (b = 0.28, SE = 0.04, t = 8.69, p = 0.000), not providing support to our Hypothesis 1 (all burnout column in Table 2) and that the relation between burnout and knowledge-hiding behavior is negative (b = −0.47, SE = 0.09, t = −0.17, p = 0.000), not providing support to our Hypothesis 2 (all knowledge-hiding behavior column in Table 2).

The results point out that the index of moderated mediation is significant, estimate = b = 0.13, bootstrap standard error = bootSE = 0.05, 95% percentile bootstrap confidence intervals (CI: 0.02, 0.25), showing support for our moderated mediation hypothesis (Hypothesis 4). The negative conditional indirect influence for low values (-1SD) of hope was the strongest, b = −0.23 (boot SE = 0.05), 95% percentile bootstrap CI (−0.34, −0.13), it was weaker but still significant for medium values of hope, b = −0.18, boot SE = 0.04, 95% percentile bootstrap CI (−0.27, −0.11) and for high values (+1SD) of hope, b = −0.13, boot SE = 0.04, 95% percentile bootstrap CI (−0.22, −0.06).

For the a-path from servant leadership to burnout (Hypothesis 3), there was a significant interaction, but not in the expected direction, between servant leadership and hope, b = −0.28, SE = 0.11, t = −2.52, p = 0.012, ΔR² = 0.29 (Table 2). As illustrated in Figure 2 (moderating role of hope in the relationship between servant leadership and burnout), simple slope analysis showed that the conditional influence from servant leadership on burnout was strongest for low values (-1 SD) of hope, b = 0.49, SE = 0.06, t = 8.07, p = 0.000, and was weaker but still significant for medium values of hope, b = 0.38, SE = 0.04, t = 8.69, p = 0.000, and for high values of hope, b = 0.28, SE = 0.06, t = 0.4.53, p = 0.000. The b-path from burnout to knowledge-hiding behavior was significant, b = −0.47, SE = 0.09, t = −5.17, p = 0.000. The direct effect (the c’-path) from servant leadership to knowledge-hiding behavior was significant, too, b = 0.19, SE = 0.07, t = 2.47, p = 0.013.

Effect sizes (f ²) were calculated to assess the practical significance of both direct and moderating relationships (Table 2). The effect size for the relationship between the independent variables and burnout was f ² = 0.41, indicating a large to very large effect (Cohen, 1988). For knowledge-hiding behavior, the effect size was f ² = 0.16, reflecting a small to medium effect (Cohen, 1988). The moderation effect of servant leadership and hope yielded an effect size of f ² = 0.02, which is considered large (Kenny, 2018).

Even though COR theory implies that servant leadership should help decrease burnout levels among employees by giving them additional social and emotional resources, Job Demands–Resources (JD–R) theory provides a more comprehensive approach explaining why this correlation is not observed in practice (Bakker and Demerouti, 2007). According to the JD–R model, the experience of burnout happens due to the combined effect of demands and resources; therefore, the presence of a good leader alone may not make much difference if employees face certain difficulties at work. High-level demands create conditions in which even a service-oriented leader will not be able to decrease employees' exhaustion. Thus, this point illustrates that servant leadership is one of many resources employees have available in order to experience either burnout or recovery.

The unsupported relationship between burnout and knowledge-hiding can also be clarified through the lens of JD–R theory. While burnout reflects emotional exhaustion and reduced energy, JD–R theory suggests that strained employees do not necessarily engage in the same behavioral coping responses across contexts (Bakker and Demerouti, 2007). Employees experiencing burnout may conserve energy by psychologically withdrawing, limiting social interaction, or reducing discretionary involvement rather than intentionally concealing knowledge from colleagues. Moreover, knowledge-hiding behavior may depend more strongly on relational conditions such as distrust, interpersonal competition, or perceptions of injustice than on exhaustion itself. Under supportive climates, employees may continue sharing knowledge despite feeling emotionally depleted because collaborative norms remain intact. Thus, JD–R theory helps explain why burnout may not consistently predict knowledge-hiding behavior, particularly in contexts where social or organizational factors shape employee behavioral responses more strongly.

This study offers three key theoretical contributions. First, it advances the emerging scholarship on knowledge-hiding behavior by exploring its association with servant leadership, burnout, and hope—an area with limited prior research (Cerne et al., 2014; Connelly et al., 2012). Building on COR theory, the study positions servant leadership as a critical organizational resource that influences knowledge-hiding indirectly through burnout (mediator) and hope (moderator). This moves beyond previous direct-effect models and provides an advanced understanding of how leadership impacts knowledge-hiding, especially within the service sector in the Indian context. Recent studies similarly highlight servant leadership’s negative link with knowledge-hiding (Tian et al., 2022), supporting our theoretical framework.

Second, this research extends literature on servant leadership’s complex relationships with burnout and hope. It empirically tests a mediation model where burnout connects servant leadership to knowledge-hiding, addressing calls to deepen insight into knowledge-hiding antecedents (Xiao and Cooke, 2018). Our findings also emphasize the moderating role of hope—a novel contribution—that shapes the servant leadership–burnout link. Notably, burnout’s positive association with servant leadership is stronger among employees with low hope. This highlights burnout’s mediating role in diminishing knowledge-hiding, aligning with research linking negative psychological states to counterproductive behaviors (Smoktunowicz et al., 2015).

Third, grounded in hope theory, hope functions as a boundary condition that attenuates burnout in response to servant leadership. Employees high in hope tend to interpret workplace stressors as challenges and learning opportunities, thereby experiencing less burnout and reduced inclination to hide knowledge. Conversely, low-hope individuals are more vulnerable to stress, burnout, and withdrawal behaviors, which can increase knowledge-hiding behavior. This finding reinforces previous work showing hope’s buffering effects against negative psychological states (Gustafsson et al., 2010; Smoktunowicz et al., 2015). By linking hope with servant leadership, burnout, and knowledge-hiding, this study expands hope theory in organizational contexts. Finally, most knowledge-hiding research originates from Western contexts (Connelly et al., 2012). Our study addresses this gap by examining these relationships within India’s service sector, contributing valuable cross-cultural insights from a Southeast Asian emerging economy. Given India’s high collectivism and power distance (Hofstede, 1980), this integrative model enriches understanding of servant leadership’s influence on burnout, hope, and knowledge-hiding in non-Western settings, marking a novel contribution to the literature.

Based on the findings of the current study, it can be argued that companies need to be more balanced when using servant leadership. While COR theory sees servant leadership as an important emotional and social resource, our findings show that even in such situations where there is servant leadership, employees might not experience reduced levels of burnout due to the high level of task demands. Based on JD–R theory, it is important for managers to understand that their support will work only in a situation where they have sufficient time and are provided with enough role clarity and other organizational resources. In particular, service organizations should be careful not to assume that only because of servant leadership, they will be able to offset excessive job demands.

Moreover, the results call attention to the need to develop organizational climates which promote psychological safety, trust, and cooperative knowledge-sharing. The JD-R model indicates that burnout is not the only cause of knowledge-hiding since the response of employees largely depends on contextual and relational factors. Therefore, besides addressing the problem of burnout, managers need to create an environment which encourages the exchange of knowledge by providing support, fairness, and transparency to their subordinates. Such interventions can involve introducing team-based learning schemes, peer mentoring programs, rewards for collaborative behavior, and measures to prevent knowledge-hiding in organizations. Organizations should also include criteria related to knowledge-sharing when assessing employee and leader performance. Through building interpersonal trust and minimizing competitive and unfair practices within an organization, it becomes difficult for employees to engage in knowledge-hiding even under emotional exhaustion and stressful working conditions.

The third significant implication deals with the psychological resources of the employees, especially hope. As per the predictions of hope theory and COR theory, our results reveal that employees having more hope will be able to better cope with organizational stressors and less dependent on other leadership resources to preserve their psychological state. In light of the above finding, organizations are advised to adopt interventions for enhancing the pool of internal resources of their employees which include resilience building, goal setting, problem solving, and adaptive coping interventions. Human resources can conduct hope assessments while recruiting and developing their leaders within the organization to recognize those employees who can cope well in difficult situations in their services. Brief intervention programs can be used for employees for reframing their perception towards the stressor and seeing them as challenging instead of overwhelming.

Finally, the present study indicates that leaders should utilize contingency theory when implementing servant leadership principles in various organizational and cultural settings. Specifically, this study’s findings indicate that not only the leadership style, but also the availability of personal resources among employees, plays a role in the success of servant leadership principles. This means that firms in collectivist cultures with high power distance levels, like India, may find leadership principles that focus on interpersonal relationships and employee development useful. At the same time, managers are advised against applying servant leadership practices indiscriminately without taking into consideration organizational culture, work stress level, or the psychological preparedness of employees to respond to these demands. Developing tailored leadership training programs based on particular job requirements and individual employee resources might prove to be more effective for organizational outcomes than any generic application of servant leadership techniques.

These are some of the limitations associated with this study, but they equally provide key theoretical insights into future research directions. Firstly, the cross-sectional nature of the study does not allow for any determination about the causal direction implied in both COR and hope theories. While our theoretical framework suggests that servant leadership serves as a work-related resource that lowers burnout and consequently results in a reduction in knowledge-hiding behavior, there is a need for more longitudinal and experimental research to investigate whether such a process takes place over time. It would then be possible to ascertain whether servant leadership triggers the beginning of resource caravans that increase psychological resources and reduce defensive work behavior in employees. Secondly, because the data collected was primarily from self-reports, this might have impacted the findings, even with low intercorrelations.

Future research could also consider furthering the theory-building efforts of the integration of COR and JD-R frameworks adopted in this paper. The results from the current research reveal that servant leadership alone might not be enough to alleviate burnout among employees experiencing excessive demands, which implies that leadership resources work in conjunction with other demands-resources configurations. Hence, future research should consider investigating potential boundary conditions, such as workload, emotional labor, conflict, or organizational justice, to determine the conditions under which servant leadership can mitigate the risks of burnout and knowledge-hiding. Further, multilevel research in different types of service organizations would help identify how organizational climates influence the preservation of resources and sharing of knowledge. Finally, since the study has been carried out in the collectivistic and high-power distance environment of India, future cross-cultural research should explore whether the same resource-based processes apply to individualistic cultures.