How identity leaders foster team innovation: a multilevel investigation

In today's rapidly evolving business landscape, organizations face an ongoing challenge of sustaining innovation to remain competitive and ensure their survival (Acar et al., 2024). Yet many struggle to translate new ideas into practice, often due to a lack of cohesion and flexibility within teams, which are essential for fostering collective innovation (Grossman et al., 2022; Huang et al., 2022). To overcome these challenges, organizations must develop teams that can swiftly adapt to changing environments while fostering a strong sense of unity to leverage collective strengths (Meyers et al., 2023). Achieving this balance is complicated by the tendency of members to operate in isolation, where individual interests may overshadow shared goals, leading to fragmentation rather than collaboration (Wu et al., 2021). Researchers seeking solutions for these issues have identified effective leadership as a critical driver that shapes team dynamics and unites members to motivate team innovation (Hughes et al., 2018).

The existing literature has examined the impact of various leadership styles on innovation, such as transformational, charismatic, authentic, and humble leadership (Born et al., 2025; Leblanc et al., 2022; Lei et al., 2020; Salloum et al., 2023). While these styles offer valuable insights, they often emphasize leaders' traits or individual followers as target of influence rather the collective nature of team dynamics (Haslam et al., 2023b). In contrast, identity leadership offers a promising framework for fostering team innovation by highlighting how leaders cultivate a shared identity that aligns team members around common goals (Haslam et al., 2023b). Identity leadership involves behaviors through which leaders represent, advance, create, and embed a shared sense of “us” within teams (Haslam et al., 2026). To date, research has begun to explore the relationship between identity leadership and innovation (e.g. Bracht et al., 2023; Smith et al., 2018). However, our understanding of how and when identity leaders drive members to innovate collectively remains far from complete (Lê and Hoang, 2025). Addressing this gap requires deeper exploration of the mechanisms underlying this relationship.

Grounded in the social identity theory of leadership (SITL), which explains how social identity processes enable leaders to create social change by engaging with a shared identity of the group they lead (Haslam et al., 2026; Hogg et al., 2012), our study proposes a theoretical framework wherein identity leadership enhances team innovation through the dual mediating mechanisms of team flexibility and social identification. Team flexibility emerges as a vital characteristic that enables teams to respond effectively to shifting external demands (He et al., 2014). This adaptive capability allows teams to identify new perspectives and improve their collaborative processes (Li et al., 2010; Liu and Lin, 2021). However, for teams to be flexible requires a strong sense of social identification among team members, wherein individuals perceive their teams as integral to their self-concept (Postmes et al., 2013). Such identification is essential for advancing collective goals rather than pursuing individual preferences (Haslam et al., 2013). By taking a multilevel perspective, our study takes a further step from prior single-level studies (e.g. Clarke et al., 2025; Figgins et al., 2024) and advances the current literature by explaining how multilevel social identity processes jointly translate identity leadership into team innovation outcomes (Carminati and Héliot, 2023; Islam, 2020).

Moreover, we incorporate leader-member exchange (LMX) ambivalence as a moderating factor at both individual and team levels to capture the complexities of social identity processes (Carminati and Héliot, 2023; Johns, 2017). While SITL posits that leaders and followers co-produce shared identity (Haslam et al., 2026; Hogg et al., 2012), this process is shaped by the dynamics of their interactions. LMX ambivalence reflects the conflicting feelings team members may have toward their leader, which can affect how strongly they identify with the group and collaborate (Lee et al., 2019; Zentgraf, 2023). Therefore, LMX ambivalence is examined as a moderator of the relationship between social identification and team innovation, and team-level LMX ambivalence as a moderator of the link between team flexibility and team innovation to explain the mechanism underpinning identity leadership-team innovation relationship.

This study makes significant contributions to the literature on innovation and leadership. First, we propose a multilevel model showing that identity leadership fosters team innovation through social identification and team flexibility. By examining these mediators, our study clarifies how identity leadership helps teams translate new ideas into practical outcomes by strengthening shared commitment and adaptability. Furthermore, we extend prior research by examining LMX ambivalence at both team and individual levels, moving beyond the dominant view of LMX as singular and stable (e.g, Ahmed et al., 2024; Zhao and Kamil, 2025). This perspective clarifies how ambivalent leader-member relationships can hinder team innovation. Additionally, our study contributes to SITL by demonstrating that identity leadership supports both continuity and change within teams. We show how leaders foster team innovation by mobilizing dynamic social identities that remain meaningful under changing social conditions, thereby clarifying how and when identity leadership operates through evolving social identity processes (Haslam et al., 2026).

The findings of this study also highlight the practical importance of developing identity leaders. For managers, this research suggests focusing more on creating and embedding shared identity within teams to help employees commit and work flexibly toward collective goals. For organizations, our study emphasizes selecting and developing leaders who are aware of shared identity and can unite members around common goals. For policymakers, it underscores the broader value of supporting identity leadership development initiatives across sectors.

Our study is grounded in SITL (Hogg et al., 2012), which extends and applies the broader Social Identity Theory (Tajfel and Turner, 1979) and Self-Categorization Theory (Turner et al., 1987) to explain leadership as a social influence process. SITL builds on the social identity analysis of social influence by emphasizing that people's social identity is shaped through their membership in meaningful groups, which influences how they see themselves and connect with others (Ashforth and Mael, 1989; Hogg et al., 2012). A core premise of SITL is that leadership effectiveness depends on leaders and followers perceiving themselves as part of a shared social identity of “we” and “us” rather than as just a personal identity of “I” and “you” (Hogg et al., 2012; Steffens et al., 2021). According to Hogg et al. (2012), group members who are most prototypical, meaning they best represent the group's shared identity, tend to be especially influential and are more likely to emerge as leaders.

Building on this idea, identity leadership, as conceptualized by Haslam et al. (2026), extends SITL beyond group identity prototypicality highlights a broader range of leader behaviors that shape and sustain shared social identity. Rather than focusing solely on being a group prototype, identity leadership emphasizes how leaders actively engage in shaping the meaning of the group to foster sense of belonging (Haslam et al., 2026). In other words, identity leadership is what leaders do to craft group social identity and make such identity matter for group members (Haslam et al., 2022; Krug et al., 2021a). In practice, this leadership approach emphasizes mutual influence between leaders and followers while advocating for debate and dialog within the group to define their shared identity and goals (Haslam et al., 2026). Evidence for the effectiveness of identity leadership is provided in a recent systematic literature review, which highlighted that identity leadership is consistently associated with positive outcomes across levels, including improved performance, engagement, trust, well-being, morale, and cohesion (Lê and Hoang, 2025).

Taken together, SITL provides a useful lens to understand how identity leadership can foster team innovation as it highlights the power of shared group identity in aligning individual behaviors with collective goals. Through its emphasis on “we-ness,” SITL explains why groups develop a common purpose and how this sense of belonging spurs members to advance group interests through innovative behaviors. Moreover, SITL helps explain the mechanisms by which identity leadership shapes and coordinates shared identity within groups (Haslam et al., 2022), and how this, in turn, drives innovation through the mediating processes of social identification and team flexibility. Furthermore, the moderating effects of individual- and team-level LMX ambivalence help illuminate relational dynamics within these identity processes. In short, through SITL, we offer a more nuanced picture of how leaders can deliberately shape and harness a shared social identity embedded in individual and team processes to drive team innovation.

Given that “leadership is by nature a multiple-level phenomenon” (Chun et al., 2009, p. 689), it is essential to view identity leadership as a group process that shapes both individual and group behaviors. Treating the group as a meaningful entity is also important when research involves individuals nested in groups such as work teams (Braun et al., 2013). By fostering a shared sense of “we,” identity leadership shapes how members collectively interpret goals and coordinate their efforts. Leaders who actively engage with and manage this shared identity are therefore better able to align the majority of members' behaviors and guide them toward actions that serve the group's interests (Haslam et al., 2022).

Despite growing evidence that identity leadership can enhance a range of individual outcomes – such as well-being (Cárdenas et al., 2024), organizational citizenship behavior (Mühlemann et al., 2022), and performance (Fransen et al., 2022) – scholars have called for more attention on how it produces outcomes at the group level (Butalia et al., 2025; Hou et al., 2021). While some recent studies suggest that identity leadership can strengthen team performance (Krug et al., 2021b) and teamfulness (Haslam et al., 2023b), these studies have only begun to uncover its broader group-level effects. Understanding how identity leaders shape group processes and outcomes remains an important focus (Lê and Hoang, 2025). To better understand this group process, there is a need to examine the multilevel pathways through which identity leadership influences team functioning and outcomes. Building on these considerations, our research contributes to the literature by investigating how identity leadership operates as a group process to influence group-level outcomes, specifically focusing on team innovation and the multilevel mechanisms that connect leaders, individuals, and the team as a whole.

Team innovation refers to the outcome of collective efforts to generate and implement new and improved ways of working (Acar et al., 2024; van Knippenberg, 2017). Team innovation involves leveraging members' diverse knowledge, skills, and perspectives to generate novel ideas and to secure the support and resources necessary for their implementation (Han and Ni, 2025; Li et al., 2018; van Knippenberg, 2017). Due to this collective effort, leadership is usually needed to coordinate the group to achieve successful team innovation (van Knippenberg, 2017).

In light of SITL, identity leadership emphasizes the role of leaders in creating a shared sense of “us” within the team, which can be a powerful driver of team innovation. Bracht et al. (2023) found that identity leadership positively influences individual innovative behavior through social identification, as individuals who identify strongly with their team are more likely to support its innovation efforts. At the group level, identity leaders can foster a shared identity that shifts team members' focus from individual behaviors to group actions (Haslam et al., 2022). By encouraging team members to define themselves as part of a meaningful ingroup, identity leaders motivate them to work toward the shared success of the group, often being willing to prioritize group goals over personal interests (Haslam et al., 2026). Considering this insight, we hypothesize the following.

Social identification refers to the process by which individuals define themselves as members of a group, adopting the group's norms, values, and goals as part of their self-concept (Postmes et al., 2013; Younis and Hammad, 2021). When people identify with a group, they are more likely to align their behavior with the collective good of that group (Butalia et al., 2025).

Identity leadership plays a pivotal role in fostering social identification. Drawing from SITL, we argue that when leaders articulate and represent what makes the group distinct, they enhance members' sense of belonging and strengthen the salience of group identity (Haslam et al., 2026; Hogg et al., 2012). By representing and advancing what the group stands for, leaders can encourage members to integrate the group's identity into their sense of self and strengthen individuals' commitment to the group (Haslam et al., 2022; Krug et al., 2021a). Supporting this mechanism, a meta-analysis by Steffens et al. (2021) shows a reliable association between leaders who represent shared identity and followers' commitment and team cohesion.

From a SITL perspective, when individuals identify with their group, they are motivated to advance the group's interests by generating new ideas and working with others to put those ideas into practice (Bracht et al., 2023). Social identification leads individuals to internalize the group's goals and values, making the group's identity part of their own self-concept (Tajfel and Turner, 1979). This identification, in turn, motivates members to work with others to put new ideas into practical outcomes, as doing so enhances both the group's status and their own sense of self-worth (Haslam et al., 2013). This view is supported by Hirst et al. (2009), who found that social identification with a work team fuels creative effort and persistence, which in turn helps individuals solve problems creatively to protect and enhance the group's standing. Similarly, Bracht et al. (2023) argue that individuals who identify strongly with their team are especially motivated to search for and develop ideas that benefit the group as a whole, as a way to sustain their group membership and positive relationships at work.

Extending this logic, we argue that individuals do not stop at developing new ideas on their own; rather, they actively seek endorsement and support from others in the group to ensure these ideas are recognized as useful and worth implementing. This group endorsement is crucial because what counts as novel and useful is shaped by group norms and shared values (Adarves-Yorno et al., 2008; Haslam et al., 2013). By engaging others in the group, individuals increase the likelihood that their ideas align with the group's identity and goals, making implementation more feasible. Therefore, individual social identification motivates group members to innovate together.

Taken together, these arguments suggest that social identification is an important mediator that translates identity leadership into team innovation. Thus, we hypothesize the following.

Team flexibility refers to a team's ability to effectively adapt to environmental changes and uncertainties (Li et al., 2010; McComb et al., 2007). Flexibility in a team involves changing the behavioral and cognitive structures of team members, allowing the entire team to respond to external pressures rather than relying on rigid routines (Ling et al., 2021; Tomer, 1995). Team flexibility plays a crucial role in achieving organizational benefits such as increased productivity, competitiveness, and profitability (Annosi et al., 2025; He et al., 2014; Li et al., 2010).

Drawing on SITL, we propose that teams led by identity leaders are likely to exhibit higher levels of flexibility. First, identity leaders often need to adjust the shared identity of the team they lead by redefining group boundaries in response to changes in social reality (Haslam et al., 2026). This social identity process can disrupt fixed ideas about who “we” are and open the way for members to question the status quo. In this way, identity leadership supports team flexibility by making group norms open to change. Second, He et al. (2014) argue that team flexibility is fostered when members feel empowered to make decisions and coordinate effectively in response to changing conditions. Identity leadership is associated with higher team cohesion (Butalia et al., 2025) and psychological safety (Clarke et al., 2025; Fransen et al., 2020). Such a cohesive and psychologically safe environment can encourage team members to challenge established norms and practices (Dar et al., 2022) and empower team members to adapt their work together.

This enhanced team flexibility, in turn, is expected to positively impact team innovation. Flexible teams tend to have a flatter hierarchy, where input flows freely and each member feels included and recognized (Dar et al., 2022; Tomer, 1995). Indeed, research has shown that such flexibility supports outcomes such as higher team performance (Günsel and Açikgöz, 2013; Lin et al., 2015) and organizational citizenship behavior (Liu and Lin, 2021). Furthermore, flexibility helps team members break through impasses, see problems from new perspectives, and even spot new problems that need solving (Georgsdottir and Getz, 2004). Through the lens of SITL, when members see themselves as part of the group, they support and build on each other's ideas. This shared identity makes teams more receptive to new ideas that extend group norms in valuable ways, giving unusual ideas a chance to be recognized and implemented (Adarves-Yorno et al., 2008; Haslam et al., 2013). Together, flexibility helps members implement creative ideas collectively to advance the group's goals.

Building on these relationships, we posit that team flexibility is an essential mediating mechanism through which identity leadership enhances team innovation. Based on the discussion, we hypothesize the following.

To examine how relational dynamics shape team innovation, we focus on LMX ambivalence as a moderator. While SITL emphasizes identification with the group as a whole (Hogg et al., 2012), Zentgraf (2023) argues that a relational perspective can shed light on how interpersonal exchanges influence social identification processes. In practice, leader-follower relationships are often ambivalent, with individuals experiencing mixed or conflicting feelings about their leader (Lee et al., 2019; Pratt, 2000). Such ambivalence may arise when leaders juggle multiple roles, send inconsistent signals (Bass and Bass, 2008), or provide uneven support across team members (Lee et al., 2019). LMX ambivalence captures this reality by reflecting the conflicting perceptions team members hold about their relationship with the leader (Lee et al., 2019). We argue that high ambivalence in relationships between the leader and individual members, or with the majority of members, may weaken the extent to which social identification or team flexibility translates into team innovation.

Individual-level LMX ambivalence refers to the mixed or conflicting perceptions an individual holds about their relationship with the leader (Lee et al., 2019). Social identification motivates members to generate ideas and work with others to implement them, but this process can be disrupted if members have ambivalent feelings about their leader. Huang et al. (2022) argue that LMX ambivalence can create uncertainty and anxiety for team members. From the SITL perspective (Hogg et al., 2012), this uncertainty makes it harder for individuals to trust the leader's role in shaping the group's shared identity or to see the leader as a prototypical member. As such, high LMX ambivalence can reduce individual members' motivation to work for the group's interests, such as implementing new ideas. In contrast, when LMX ambivalence is low, individuals are more likely to trust in the shared identity mobilized by the leader and to work with others toward group goals. Based on this discussion, the following hypothesis is formulated.

In team settings, LMX ambivalence could be conceptualized at the team level because team members are often exposed to similar leadership behaviors (Hahn and Semrau, 2023). Based on the definition by Lee et al. (2019), we refer to team LMX ambivalence as conflicting perceptions that team members collectively share about their relationship with the leader. For example, a team looks up to their leader's innovative behaviors, which often leads to exciting projects. However, taking on many projects can create stress and negatively affect employees' psychological well-being (Aksoy, 2025). This creates ambivalence, as they value the leader's innovation but feel uneasy about the potential risks (Aksoy, 2025). Furthermore, when team members see some team members receiving consistent support and opportunities from the leader while others receive less, it can also create ambivalence (Lee et al., 2019).

Huang et al. (2022) argue that LMX ambivalence may shape how team members interpret and respond to flexibility within their team. Prior research also suggests that leadership factors can influence whether team flexibility successfully translates into innovation (Günsel and Açikgöz, 2013; Jia et al., 2021). From an SITL perspective, when the majority of team members feel ambivalent toward their leader, their sense of shared identity can weaken. Unequal or inconsistent exchanges between the leader and various team members can exacerbate the dynamics (Chen et al., 2025). As a result, high shared LMX ambivalence undermines the cohesion needed to turn flexibility into coordinated efforts to implement new ideas. In contrast, when shared LMX ambivalence is low, team members are more likely to collectively identify with the shared identity, and the resulting cohesion can support members to be flexible together for innovation. Therefore, we hypothesize as follows.

Figure 1 exhibits the theoretical model and summarizes the relationships between the main constructs in this study. The following sections present the research methodology and research findings.

Our study collected data from work teams within private Vietnamese hospitality companies, such as hotels and restaurants. As an emerging economy, Vietnam's hospitality industry has expanded rapidly in recent years due to rising numbers of both domestic and international tourists (Luu, 2023). In this highly competitive market, firms often rely on their teams to generate and implement innovative ideas to stay competitive (Hoang et al., 2023). Furthermore, as part of Confucian Asia, Vietnam's strong collectivist culture can arguably support identity leadership (Haslam et al., 2022). Relational aspects such as social support and maintaining good interpersonal relationships are highly valued in Vietnamese organizations (Nguyen et al., 2025). This context is suitable for studying this leadership approach, which has been limited in research from the non-Western contexts (Lê and Hoang, 2025).

Data was gathered from two sources, managers and employees, to reduce common method bias (Podsakoff et al., 2024). We randomly selected 300 private companies from lists provided by a government agency and a business association. Human resource managers at these companies were contacted via phone and email to outline the study's objectives and request permission to survey their managers and employees. Out of the 300 contacted, 116 companies agreed to participate. With support from the human resource departments and CEOs of these firms, we distributed surveys to 650 employees and 210 managers in their teams. The surveys, in Vietnamese and labeled with codes for matching purposes, were distributed along with a letter describing the study, emphasizing voluntary participation, and assuring response confidentiality and anonymity.

Of the 650 employee surveys distributed, 345 were returned, with 315 usable responses retained after excluding 7 incomplete surveys and 23 without matching manager data. Among managers, 117 of the 210 distributed surveys were valid after removing incomplete responses and those missing employee matches. Given the multilevel design of the study, we included teams with valid responses from the manager and at least two employees, consistent with the team-level inclusion criteria used by Wang et al. (2024). This approach is also consistent with prior studies (e.g. Rego et al., 2025). The final sample consists of 294 employees (45.2%) and 92 managers (34.8%) across 92 teams.

Teams had an average size of 3.19 employees (ranging from 2 to 23, SD = 2.51). Employees were, on average, 27.4 years old (SD = 9.6), with 55.8% identifying as female, and 26.2% holding undergraduate or graduate degrees. Tenure distribution among team members was as follows: 63.6% with less than 2 years, 14.97% with 2–5 years, and 21.43% with over 5 years in their teams.

We employed back-translation method (Brislin, 1970) to translate the original English questionnaire into Vietnamese. First, a bilingual management researcher translated the items from English to Vietnamese. Next, another bilingual management researcher translated the Vietnamese version back into English. Additionally, a third management researcher reviewed both the English and Vietnamese translations. Any inconsistencies in the translation process were resolved through discussion. The response scale used five points with designated anchors: 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, and 5 = strongly agree. All survey instruments can be found in the  Appendix.

Identity leadership was evaluated using a four-item scale from Steffens et al. (2014). An exemplary item is “My manager creates a sense of cohesion within the team.”

Social identification was assessed using a four-item scale from van Der Vegt et al. (2003). The items included: “I would like to continue working with my team.”

Team flexibility was evaluated using a four-item scale from He et al. (2014). An exemplary item is “Our team members are able to make any adjustments necessary to cope with changing circumstances.”

LMX ambivalence was assessed using a seven-item scale from Lee et al. (2019). The items included: “I have conflicting thoughts: sometimes I think my manager recognizes my potential, while at other times I don't.”

Team innovation was evaluated using a six-item scale adapted from Scott and Bruce (1994). A sample item is “My team develops adequate plans and schedules for the implementation of new ideas.”

We considered team size and team tenure as control variables, because they may affect innovation outcomes (Wei and Lau, 2012; Weiss and Hoegl, 2016). To measure team tenure from individual tenure ratings (1 = less than 2 years; 2 = 2–5 years; 3 = more than 5 years), we calculated an aggregate team-level measure by averaging individual tenure responses within each team.

The team-level constructs (i.e. identity leadership, team flexibility, and team LMX ambivalence) were aggregated using a direct consensus composition model (Chan, 1998), where the team-level score was aggregated from individual ratings (Li et al., 2022). According to Bliese (2000), appropriate within-team agreement (r_wg(j)) is crucial before aggregating individual responses into team-level measures. We chose r_wg(j) because it is a widely used index that assesses whether team members' responses are similar enough to be aggregated and offers clear guidelines for interpretation (Nijstad et al., 2025; Woehr et al., 2015). Our data show that, at the team level, the median r_wg(j) coefficients were 0.87 for identity leadership, 0.86 for team flexibility, and 0.84 for LMX ambivalence. All values exceed the cutoff value of 0.70 (Woehr et al., 2015), indicating strong within-team agreement. We also calculated intraclass correlation coefficient ICC(1) for identity leadership (ICC(1) = 0.16), team flexibility (ICC(1) = 0.13), and team LMX ambivalence (ICC(1) = 0.006). These values indicate small to moderate between-group variance (LeBreton and Senter, 2008; Woehr et al., 2015). Overall, the indices reasonably justify aggregation.

In our data, employees are nested within teams/managers, and our focal variables are located at two conceptual levels. Mplus 8 program was used for analysis (Muthén and Muthén, 2017). We placed social identification and LMX ambivalence at the individual (within) level, while identity leadership, team flexibility, team LMX ambivalence, and team innovation were analyzed at the team (between) level. To test the hypotheses, we performed multilevel path analysis within a single integrated model, using the TWOLEVEL function to account for the nested data (Muthén and Muthén, 2017).

To assess our proposed multilevel indirect effects (e.g. Hypotheses 2 and 3), we adopted the analytical strategy recommended by Preacher et al. (2010). Specifically, we used the single-level estimate of the association between identity leadership and team flexibility, along with the cross-level estimate of the relationship between identity leadership and social identification for the “a” paths. We then utilized the single-level estimate of the association between team flexibility and team innovation, along with the cross-level estimate of the relationship between social identification and team innovation for the “b” paths. These estimates were used to calculate the product terms (i.e. “ab”) for the hypothesized indirect effects of identity leadership on team innovation, mediated through team flexibility and social identification. In other words, in line with Preacher et al. (2010), the indirect effect via social identification resembles a 2-1-2 multilevel mediation path, whereas the indirect effect via team flexibility is tested at the team level.

To test the moderation effects, we modeled LMX ambivalence at the individual level (Hypothesis 4a) for the link between social identification and team innovation, and team LMX ambivalence at the team level (Hypothesis 4b) for the link between team flexibility and team innovation. For significant moderation, simple slopes analysis was used to illustrate effects at higher (above the mean) and lower levels (below the mean) of the moderators.

Table 1 presents descriptive statistics and bivariate correlations. In terms of control variables, team tenure (r = −0.12, p = 0.153) and team size (r = −0.09, p = 0.216) were unrelated to team innovation. Therefore, following Bernerth and Aguinis (2016), we did not include these control variables in the final analysis when testing study hypotheses.

We assessed the reliability of each scale by calculating Cronbach's alpha. As shown in Table 1, Cronbach's alpha values ranged from 0.86 to 0.92, exceeding the recommended minimum of 0.70 (DeVellis, 2012). This indicates strong internal consistency and reliable scales. To evaluate convergent validity, we examined factor loadings, composite reliabilities (CR), and average variance extracted (AVE). Table 1 demonstrates that CR values were above the 0.60 threshold (Bagozzi and Yi, 1988) and AVE values surpassed the 0.5 cutoff (Fornell and Larcker, 1981). These results confirm convergent and discriminant validity.

We conducted a multilevel confirmatory factor analysis (MCFA) using Mplus to verify the construct validity of our study variables. The fit indices used were the comparative-fit index (CFI), Tucker–Lewis coefficient (TLI), root mean square error of approximation (RMSEA), and standardized root mean squared residual (SRMR) for both individual and team levels (Hair et al., 2013). According to Hair et al. (2013), CFI >0.90, TLI >0.90, RMSEA <0.08, and SRMR <0.08 indicate that the model provides an acceptable fit to the data. MCFA was performed with an increased number of iterations (5,000 H1 iterations) to ensure model convergence (Muthén and Muthén, 2017). This six-factor model showed a good fit with the data (χ² (212) = 335.400; CFI = 0.951; TLI = 0.938; RMSEA = 0.044; SRMR within = 0.043; SRMR between = 0.056). The hypothesized six-factor model provided the best fit compared to alternative models (see Table 2), confirming the optimal model structure. This means that all variables are distinct, including LMX ambivalence and team LMX ambivalence. As additional variables were combined, model fit worsened, further supporting discriminant validity. These results support the distinctiveness of all six constructs, including the differentiation between individual-level LMX ambivalence and team-level LMX ambivalence.

To address common method bias, we implemented several ex-ante procedures, such as guaranteeing respondent anonymity, using established and validated measurement scales, and collecting data from multiple sources. The multilevel nature of our model and the use of aggregated constructs also help reduce the likelihood of bias (Podsakoff et al., 2024). In addition, we conducted a post-hoc test using a two-factor model that separated items based on their data sources. According to Podsakoff et al. (2003), a good model fit in such a structure would suggest that observed variance is largely attributable to source differences rather than construct overlap. However, the model fit as reported in Table 2 was poor: χ²(536) = 4194.358; CFI = 0. 198; TLI = 0.176; RMSEA = 0.152; SRMR (within) = 0.399; SRMR (between) = 0.336. This poor fit of the two-factor model indicates that common method bias is unlikely to be a major concern in our data.

Having determined that our six-factor measurement model is acceptable and addressing potential common method bias, we moved forward with testing our hypothesized model. Hypothesis 1 predicted that identity leadership is positively related to team innovation. As shown in Table 3, the direct effect of identity leadership on team innovation was not significant (β = −1.08, standard error (SE) = 0.75, p = 0.054). Hence, this finding does not support Hypothesis 1.

The results indicate that identity leadership has significant positive effects on social identification (β = 0.97, SE = 0.17, p < 0.001) and team flexibility (β = 0.65, SE = 0.08, p < 0.001). Social identification (β = 1.19, SE = 0.50, p < 0.001) and team flexibility (β = 0.84, SE = 0.09, p = 0.037) both have a significant, positive effect on team innovation. Table 4 offers a summary of our mediation results. In this table, SE refers to the standard error, and the 95% CI indicates the 95% confidence interval, with LL and UL denoting its lower and upper limits, respectively; the indirect effect is considered significant if the 95% CI does not include zero. As seen in the table, there was a significant, positive indirect effect of identity leadership on team innovation through social identification, (unstandardized estimate = 1.87, SE = 0.94, p = 0.001, 95% CI [0.03, 3.71]). There was also a significant positive, indirect effect of shared leadership on team innovation through team flexibility (unstandardized estimate = 3.99, SE = 1.23, p = 0.04, 95% CI [1.58, 6.41]). Consistent with our expectations, teams that had higher levels of identity leadership subsequently reported higher levels of individual social identification and team flexibility, which, in turn, led to higher team innovation. Thus, Hypotheses 2 and 3 were supported.

Hypothesis 4a predicted that LMX ambivalence moderates the relationship between social identification and team innovation. To investigate the moderating effect of LMX ambivalence, we adopted the methodology proposed by Byun et al. (2018) by assessing the interaction term coefficients, along with observations from a simple slope test. As shown in Table 2, social identification × LMX ambivalence did not predict team innovation (β = −0.33, SE = 1.10, p = 0.639). Therefore, Hypothesis 4a is rejected.

Hypothesis 4b predicted that team LMX ambivalence moderates the relationship between team flexibility and team innovation. As shown in Table 2, interaction of team flexibility × team LMX ambivalence negatively affected team innovation (β = −1.38, SE = 0.18, p = 0.029). In addition, Figure 2 visually illustrates that team LMX ambivalence weakens the positive effect of team flexibility on team innovation. Therefore, Hypothesis 4b is supported.

First, we did not find support for Hypothesis 1, which proposed a positive direct link between identity leadership and team innovation. This result is similar to Hou et al.’s (2021) finding of a non-significant direct effect of identity leadership on team performance. One possible reason is that while strong group identity can create cohesion and psychological safety (Clarke et al., 2025; Fransen et al., 2020), this same psychological safety and group cohesion may also encourage members to conform and maintain group identity rather than change (Turner and Pratkanis, 1998). This result suggests that identity leadership needs engaged followership to enact the shared social identity in ways that drive collective goals (Haslam et al., 2023a). In other words, innovation emerges not from explicit instructions, but from followers' interpretation and enactment of identity leadership.

Second, Hypothesis 2 was supported, showing that identity leadership indirectly influences team innovation through social identification, consistent with Bracht et al.’s (2023) finding of a similar indirect link at the individual level. We also found support for Hypothesis 3, showing that team flexibility mediates the relationship between identity leadership and team innovation. Although this link is not extensively examined in the literature, our results partially align with the work of Gahan et al. (2021), which found that leadership fosters innovation outcomes through dynamic capability. These results are consistent with prior research linking different leadership behaviors to innovation-related outcomes through both individual and team mechanisms (e.g. Han and Ni, 2025; Hoang et al., 2023; Zhang et al., 2025).

Finally, our results support Hypothesis 4b, showing that team LMX ambivalence weakens the positive link between team flexibility and team innovation. However, we did not find support for Hypothesis 4a, which proposed that LMX ambivalence would reduce the positive effect of social identification on team innovation. Although this link has not been explored extensively, this pattern can be understood through prior work showing that LMX ambivalence increases uncertainty and anxiety within teams (Huang et al., 2022). When most team members feel ambivalent about their leader, this shared ambivalence disrupts teamwork and limits the benefits of flexibility for innovation. In contrast, team innovation may be less impacted when only a few team members feel ambivalent about their personal relationship with the formal leader (Fladerer et al., 2021). This suggests that the effect of LMX ambivalence may be stronger when it reflects a shared team-level condition (consensus) rather than individual variation (dispersion). We see this distinction as a useful insight for future research to explore.

Our study offers two key implications for leadership and innovation literature. First, our findings demonstrate that identity leadership fosters team innovation through a relationally contingent social identity process. Specifically, teams innovate because identity leaders foster members' identification with a shared sense of “us” and commitment to collective goals, while simultaneously mobilizing a dynamic shared identity that enables cohesion while allowing for flexibility. At the same time, our findings indicate that the social identity process underlying innovation is contingent on the clarity and consistency of relationships between leaders and group members. When most leader-member relationships within a team are marked by ambivalence, teams may struggle to translate flexibility into innovative outcomes. In essence, managing people for the implementation of new and useful ideas requires attention not only to instrumental processes of knowledge integration (van Knippenberg and Hoever, 2021), but also to how employees relate to their shared social identity and to their leader.

Additionally, our findings provide robust support for SITL (Haslam et al., 2026; Hogg et al., 2012) by demonstrating that leadership effectiveness depends upon in the capability to foster a shared social identity among group members. Leaders who engage with a shared sense of “us” can mobilize both individual members and the group as a whole to work toward collective goals, such as team innovation. Importantly, our findings are consistent with prior research which suggests that the shared identity through which leaders mobilize the group is not necessarily stable or fixed (Carminati and Héliot, 2023; Milesi, 2025). Instead, identity leadership can involve members in revising the shared identity or even producing new ones to remain relevant to evolving social realities (Haslam et al., 2026). In this way, we affirm that identity leadership can both maintain a sense of continuity and induce change within groups. However, the dynamic nature of social identity processes means they can also shift in less constructive ways, such as when shaped by ambivalent leader-member relationships.

The findings of this study offer practical suggestions for managers, organizations and policymakers. First, the present findings highlight the practical value of developing leaders who can build and sustain a shared sense of “us.” These results suggest that effective leadership is more about creating, advancing, and embedding a shared social identity within teams. In line with this, managers should focus on improving the effectiveness of their leadership by engaging directly with the groups they lead. To achieve this goal, managers could participate in leadership development programs such as the 5R Leadership Development Program (Haslam et al., 2023b), which provides structured opportunities for leaders to reflect on who “we” are and to translate shared purpose into collective action. For business owners and organizations, these findings point to the benefits of selecting and developing leaders who can unite members around common goals and sustain both cohesion and flexibility during change. Embedding identity leadership principles within human resources and training systems can help organizations build cultures that innovate collectively rather than competitively. For policymakers, these findings speak to the broader social value of supporting identity leadership development initiatives. Research on identity leadership indicates that its benefits extend well beyond innovation, encompassing improved performance and well-being (Lê and Hoang, 2025; Stevens et al., 2021). In line with this, national or industry programs that incorporate identity leadership development would help strengthen managerial capability across sectors and provide a practical means of realizing the wider benefits organizations and society alike.

Second, our study confirms that both social identification and team flexibility help explain how identity leadership drives team innovation. This suggests that managers should focus not only on building a shared sense of belonging but also on empowering teams to adapt and respond flexibly to new challenges. To strengthen social identification, managers can use inclusive language such as “we” and “us” and involve team members in shaping group goals and values (Butalia et al., 2025; Haslam et al., 2026). To enhance team flexibility, managers can encourage their teams to reflect on existing thinking and practices to support changes in shared identity in response to changing social realities (Georgsdottir and Getz, 2004; Liu and Lin, 2021).

Third, we find that when many team members have mixed feelings about their relationship with their leader, it can weaken how well flexibility leads to innovation. Managers should pay attention to these mixed feelings and try to build trust and clear expectations by being consistent and fair with team members (Lee et al., 2019). Such efforts include regularly checking in with team members, addressing concerns promptly, and ensuring that expectations are upheld consistently and equally across the team (Chen et al., 2025). Managing this ambivalence can help teams stay cohesive and innovative even when facing change.

This study has some limitations that offer opportunities for future research. First, although our multilevel design helps counterbalance common method bias (Podsakoff et al., 2024), the cross-sectional nature of the data limits our ability to make causal claims. Future research should collect data for key variables at different time points to help address this issue.

Second, we only examined how identity leadership contributes to innovation performance at the team level. Future studies should explore how identity leadership affects innovation at the organizational level, as well as how identity leadership influences innovation outcomes for both teams and individual members.

Third, our research model was tested within a specific industry and within the Vietnamese cultural context, which may limit the generalizability of the findings. To strengthen external validity, future studies should replicate this research in other industries, such as professional services, healthcare, or manufacturing, and in different cultural settings. In these contexts, the impact of identity leadership may vary, and understanding these differences could offer valuable insights. Finally, future research could examine other factors, such as organizational culture, group norms, or aspects of leader/follower identity, that may shape the link between identity leadership and innovation. By exploring these areas, future studies can build on our findings and offer a more complete understanding of how identity leadership drives group and social change.

We would like to express our sincere appreciation to the two anonymous reviewers and the handling editor, Dr. Anna Bos-Nehles, for their developmental feedback and kind guidance, which have significantly enhanced the manuscript.

Identity leadership (Steffens et al., 2014)

1. My manager is a model member of the team.

2. My manager acts as a champion for the team.

3. My manager creates a sense of cohesion within the team.

4. My manager creates structures that are useful for team members.

Social identification (van Der Vegt et al., 2003)

1. I strongly identify with the other members of my work team.

2. I would like to continue working with my team.

3. I like being a member of this work team.

4. I feel emotionally attached to this work team.

Team flexibility (He et al., 2014)

1. Our team members are flexible with respect to our team's request for changes.

2. Our team members are able to make any adjustments necessary to cope with changing circumstances.

3. When an unexpected situation arises, our team members would prefer to amend our team agreement (or regulation) rather than to hold each other to the original terms.

4. Our team members are flexible when dealing with teamwork.

LMX ambivalence (Lee et al., 2019)

1. I have conflicting thoughts: sometimes I think that my working relationship with my manager is very good, while at other times I don't.

2. I have conflicting thoughts: sometimes I think my manager understands my problems and needs, while at other times I don't.

3. I have conflicting thoughts: sometimes I think my manager would use his/her power to help to solve problems in my work, while at other times I don't.

4. I have conflicting thoughts: sometimes I think I know where I stand with my manager, while at other times I don't.

5. I have conflicting thoughts: sometimes I think that my manager would “bail me out” at his/her expense, while at other times I don't.

6. I have conflicting thoughts: sometimes I think my manager recognizes my potential, while at other times I don't.

7. I have conflicting thoughts: sometimes I think that I would defend and justify my manager's decisions if he/she were not present to do so, while at other times I don't.

Team innovation (Scott and Bruce, 1994)

1. My team searches out new technologies, processes, techniques, and/or product ideas.

2. My team generates creative ideas.

3. My team promotes and champions ideas to others.

4. My team investigates and secures funds needed to implement new ideas.

5. My team develops adequate plans and schedules for the implementation of new ideas.

6. My team is innovative.