A multilevel study of the dual effects of leader political skill on follower performance

Abstract

Using a sample of 61 teams (i.e., 305 employees and 61 leaders) from multiple banks, this study examines a multilevel model of the influence of dual effects of leader political skill (LPS) on followers’ performance at two levels. We examined the effect of followers’ perception of individual-focused LPS on followers’ individual performance rated by their leaders and the effect of followers’ perception of group-focused LPS on leader-rated team performance. In addition, we also examined a cross-level effect from group-focused LPS at team level to follower performance at individual level. The results revealed that individual-focused LPS predicts followers’ performance at the individual level and the group-focused LPS predicts team performance at the team level. Results also supported the cross-level effect from group-focused LPS at team level to followers’ performance at individual level. Strengths, limitations, and implications for both theory and practice as well as for future research are also discussed.

Keywords:

• individual-focused leader political skill
• group-focused leader political skill
• follower performance
• team performance
• multilevel and cross-level effect

PUBLIC INTEREST STATEMENT

Ensuring employees’ motivation towards the accomplishment of their individual as well as collective goals is fundamental to effective leadership, and therefore, many scholars have attempted to examine the impact of leadership on the performance of their followers. This study endeavors to investigate the dual effects of leader political skill (LPS) on followers’ performance at two levels. The results show that LPS at the individual level positively predicts followers’ performance and at the team level, predicts team performance. Thus, it answers the question of how leaders in modern organizations can motivate their followers towards both the individual and team goals simultaneously.

1. Introduction

Changes in the design and dynamics of organizational structures have resulted in numerous implications for the way organizations now function and respond (Ahearn, Ferris, Hochwarter, Douglas, & Ammeter, 2004). One of those implications is the shift in the focus of the organizations from bureaucratic structures to alternative mechanisms of coordination and control (Cascio, 1995). Modern organizations rely heavily on team-based structures for the accomplishment of their goals and vision (Gully, Incalcaterra, Joshi, & Beaubien, 2002; Kozlowski & Bell, 2003; Mathieu, Hollenbeck, van Knippenberg, & Ilgen, 2017). These team-based structures elucidate the fact that social interactions in modern organizations are a very significant part of their day to day workings.

By their very essence, all types of organizational structures are social entities consisting of various individual members collaborating with each other to achieve collective goals and objectives (Fulmer & Ostroff, 2016; Katz & Kahn, 1978). However, it is not very unusual that such collaborations and social interactions are often characterized by political decisions (Mintzberg, 1983; Pfeffer, 1992). The political perspectives on organizations suggest that organizations are alliances of various individuals with competing interests. It assumes that multiple interests, scarce resources, and uncertainty associated with the organizational outcomes make organizational interactions and decision-making a political phenomenon (Ferris & Judge, 1991).

Assuming that businesses are inevitably political arenas (Mintzberg, 1983, 1985), leaders in organizations must be politically skilled so that they can not only understand the political work settings around them but also thrive in demanding times and challenging scenarios. Thus, like all organizational settings, these team-based structures cannot reap their desired benefits without politically skilled team leaders (Ahearn et al., 2004). This is also important because these politically skilled leaders are not only responsible for their own performance but also for their followers, who derive inspiration directly from their leaders. These leaders have the foremost, yet a challenging responsibility of not only developing and motivating their followers for the accomplishment of their individual responsibilities, but also the responsibility of building an environment of trust and collaboration for their team members to achieve their collective goals (Chun, Cho, & Sosik, 2016; Hackman, 2002). Hence, in order to overcome this challenge politically skilled leaders must have both the individual and the team leadership abilities.

Political theory of leadership was proposed by Ammeter, Douglas, Gardner, Hochwarter, and Ferris in 2002. Later, Ahearn et al. (2004) proposed and showed that leaders who are politically skilled are well-equipped to influence their teams positively. Since then researchers have explored the favorable outcomes of leader political skill at both the individual and the team-levels separately (e.g., Treadway et al., 2004; Yang & Zhang, 2014). The research measuring the impact of leader political skill on follower outcomes at the individual-level is well documented but remains very limited about the impact of political skill in the context of teams (Ferris, Perrewé, Daniels, Lawong, & Holmes, 2017). Similarly, it has been argued several times in literature that leadership inherently is a multilevel phenomenon (Batistič, Černe, & Vogel, 2017; Yammarino & Dansereau, 2008), and leadership models should integrate individual-level processes with the team-level processes. Thus, a failure to integrate lower-level and higher-level processes will ultimately result in a limited understanding of effective leadership (Kozlowski & Bell, 2003).

The existing leadership models of leader political skill (LPS) have neither integrated these levels nor have drawn a clear distinction between leader-follower effects and leader-team effects. Thus, our study contributes to the existing literature in a number of ways. First of all, this research investigates the leader political skill as a multilevel phenomenon by integrating the individual- and team-levels. Furthermore, the distinction of LPS at two levels will also allow researchers to examine the cross-level relationships between higher and lower-level variables. Such effects will not only help us understand the dynamic interplay among followers within a team at the individual-level but also the team as a whole, at the team-level.

The analysis of cross-level effects goes beyond the traditional testing of same-level effects and elucidates an interactive and complimentary way through which both the team and the individual-level variables mutually affect and predict individual outcomes (Batistič et al., 2017). Moreover, the cross-level approach explains the additional variance in individual-level variables, over and above the inputs of the individual-level, which consequently allows us to comprehend employee attitudes in a better way. The purpose of this study is to explain the variance in individual level outcome because of the group level variable. The dominant perception in management research is that the higher levels within which lower-level processes are nested usually exert a stronger downward influence than the lower-level variables exerting an upward influence. This downward influence is generally assumed to be greater, higher and more significant than a weaker upward influence (Hitt, Beamish, Jackson, & Mathieu, 2007; Kozlowski & Klein, 2000).

Specifically, we endeavor to examine the dual effects of LPS at both the individual and the team-level and to explore cross-level effect from the higher-level to the lower-level variable. We are proposing that individual-focused LPS behavior builds individual followers’ capabilities and expertise, and consequently enhances their individual performance. The individual-focused LPS is directed towards the individual followers, implying that the leader takes an interest in all followers, recognizes their individual skills and capabilities, and mentors and coaches them accordingly. In other words, the leader may set diverse targets or offer mentoring for the unique abilities of different followers according to their knowledge and aptitudes. Thus, the techniques and behaviors of individual-focused LPS may differ around different followers.

On the contrary, group-focused LPS is targeted on the unified efforts of the team as a whole. Specifically, leaders articulate a shared vision for the followers and emphasize the shared efforts of the team. It inspires followers to follow a shared vision and to develop shared beliefs in order to achieve those unified goals (Wihler, Frieder, Blickle, Oerder, Schütte, 2016). The target of the group-focused LPS is the team, and thus leaders communicate the same expectations and similar team values towards different members of the team. Thus, the content of group-focused LPS now remains the same towards the team, unlike individual-focused LPS behavior which is tailored to individual members of the team. Hence, it is the first attempt to investigate the dual level effects of leader political skill as well as the investigation of the cross-level effects in a multilevel study.

Moreover, we also argue that despite being targeted at the whole team, group-focused LPS will still have a trickle-down effect on followers’ individual performance. Group-focused LPS, while focused on leading teams also primes the individual outcomes of the individual members of the team. Therefore, an investigation of this cross-level effect can reveal how group-focused LPS impacts the motivation of individual team members in addition to the individual-focused LPS.

Finally, our research also endeavors to respond to the managerially relevant questions like, how can politically skilled leaders motivate their individual followers for their individual performance as well as motivate their whole teams to achieve team goals? Are individual-focused LPS and group-focused LPS two separate processes or they complement each other in predicting important organizational outcomes? Answering these significant questions will assist managers in overcoming the challenge of leading individuals and the team, all at one time and provide direction on how to refine their leadership behaviors accordingly.

In Figure 1. we outline a multilevel model of the dual LPS (i.e., both the individual and the group-focused LPS). First, we propose that the individual-focused LPS predicts the individual performance at the individual-level. Secondly, we argue that the group-focused LPS predicts team performance at the team level. Thirdly, we examine one cross-level effect from the group-focused LPS to the followers’ in role performance (i.e., the effect of team-level on the individual-level).

Figure 1. A multilevel model of the dual effects of Leader Political Skill and performance.

2. Literature review

Literature shows that there has been an increase of scholarly interest in social effectiveness constructs such as political skill that quantifies some characteristics of social efficacy (Ferris, Perrewe, & Douglas, 2002). Political skill has been described as “the ability to effectively understand others at work and to use such knowledge to influence others to act in ways that enhance one’s personal and/or organizational objectives” (Ferris et al., 2005, p. 127). Numerous studies have established and reported the political skill construct validity and have also investigated its connection with other social effectiveness constructs (e.g., Ferris et al., 2008, 2002; Semadar, Robins, & Ferris, 2006). Literature reports that it is theoretically distinct from political savvy, emotional intelligence, leadership self-efficacy, general mental ability and self-monitoring abilities (Ferris et al., 2005). Moreover, previous scholars have associated political skill with numerous-desired consequences such as job performance (Blickle et al., 2008; Wihler, Blickle, Ellen, Hochwarter, & Ferris, 2017), career success (Todd, Harris, Harris, & Wheeler, 2009), leadership effectiveness (Perrewé et al., 2004) and the attenuation of various workplace stressors (Cullen, Gerbasi, & Chrobot-Mason, 2018; Harvey, Harris, Harris, & Wheeler, 2007). Furthermore, it was theorized that political skill encompasses four distinct but related dimensions, i.e. social astuteness, interpersonal influence, networking ability and apparent sincerity (Ferris et al., 2005).

Political skill symbolizes an expression of self-assurance and confidence in social interactions (Kapoutsis, 2016). These leaders use their social astuteness to adapt their behaviors accordingly and to influence their followers effectively. They can quickly attune to contextual demands in such a way that it earns them the trust and respect of their followers (Ferris et al., 2005). Such leaders appear to be genuine and sincere, thus, inspiring the faith and support of the people around them. Their personalities reflect an aura of self-confidence and personal assurance that not only attracts others but also creates a feeling of comfort and security for others. Such leaders can maintain such a fine balance of positive self-image that it is never seen as an unfavorable aspect of their personalities, rather is always appreciated and respected (Ferris, Perrewé, Anthony, & Gilmore, 2003).

2.1. Political theory of leadership

Pfeffer (1981) and Mintzberg (1983) introduced political perspectives on organizational behavior more than three decades ago through their independent works. Both of them emphasized on political skill as a critical characteristic for successfully navigating through organizational settings. Consequently, in response to the call for a political theory of leadership by House and Aditya (1997), Ammeter, Douglas, Gardner, Hochwarter, and Ferris (2002) developed the political theory of leadership and presented political skill as one of the critical concepts of the theory. They explained that leadership is best defined as a social influence mechanism. They proposed a conceptual model and elaborated the antecedents and the consequences of leader political skill. They also argued that political theory of leadership shares its roots with the charismatic and transformational leadership attributes, i.e. politically skilled leaders display a certain charisma that is perceived favorably by their followers (Ferris et al., 2017). In fact, a study has shown that transformational leadership mediates the relation of leader political skill and the effectiveness of the leaders (Ewen et al., 2013). Furthermore, many scholars have investigated the impacts of politically skilled leadership on leadership effectiveness and have found a significant positive relation (e.g., Brouer, Douglas, Treadway, & Ferris, 2013; Buch, Thompson, & Kuvaas, 2016; Ewen et al., 2013; Semadar et al., 2006; Treadway et al., 2004).

2.2. Individual-focused leader political skill and follower performance

We define individual-focused LPS as, “the ability of the leaders to effectively understand individual followers at work, and to use such knowledge to influence those followers to act in ways that enhance leaders’, their individual followers’ and/or organizational objectives.” Literature reveals that the emphasis on abilities essential to guarantee leader performance and consequently follower performance, in recent years has shifted more towards the social aspects of leadership like social astuteness and interpersonal influence (George, 2000). More than ever before, effective leaders now in modern organizations are required to pay attention towards interpersonal abilities such as mentoring, coordination, and coaching. Whereas politically skilled leaders can effectively fulfill their roles of being mentors or coaches and Individual-focused LSP helps these individuals to demonstrate to their followers that their individual needs are being contemplated and will be taken care of. They display the right approaches to inspire individual followers in taking up challenging assignments and to encourage them for doing a better job than they are expected to do (Ammeter et al., 2002).

Politically skilled leaders are very proficient in dealing with their organizational experiences and coping with the high demands and ambiguities of the organizational arena, which makes them well-positioned to influence the opinions and experiences of their followers. Many empirical studies validate the notion that when leaders are able to take care of followers’ individual needs, they can positively influence their followers’ performance. Thus, it helps these leaders to elicit the desired reactions from their followers as well as obtain higher-levels of follower performance and productivity (Brouer et al., 2013). Treadway et al. (2004) showed that leader political skill is positively related to the followers’ perceived organizational support and trust. They also showed that these leaders could enhance followers’ affective organizational commitment by means of enhancing followers’ perceptions of organizational support. Moreover, such leaders can also enhance perceptions of fairness and justice by enhancing followers’ perceptions of organizational support (Gavin, Green, & Fairhurst, 1995). Whereas it is important to note that these perceptions of organizational support and fairness are also positively linked to follower performance (Janssen, 2001).

One of the other ways through which politically skilled leaders enhance their followers’ performance is by increasing their followers’ access to organizational capital. Such leaders have the ability to build strong ties and large organizational capital (Ferris et al., 2003), which provides timely access to the resources needed by the followers. Thus, it not only enhances the perceptions of satisfaction and support for the followers but also serves as a means of enhancing their performance and effectiveness. Hence, the behaviors of individual-focused LPS are more likely to be seen as honest and sincere and thus are bound to enhance not only follower satisfaction (Ewen et al., 2013) but also follower performance (Brouer, 2007). Thus, on the basis of the above literature, we propose the following hypothesis.

H1. Individual-focused LPS is positively related to followers’ in-role performance

2.3. Group-focused leader political skill and team performance

We define group-focused LPS as, “the ability of leaders to effectively understand their teams at work, and to use such knowledge to influence their teams to act in ways that enhance leaders’, teams’ and/or organizational objectives.” Presuming that one important task of the leaders is removing any hurdles that may obstruct team performance, being politically skilled to some extent would be a precious resource for leaders. This phenomenon is particularly important in team-based structures because research shows that the group-focused social effectiveness is one of the most significant requirements of effective leader behavior with one of the most promising potentials to positively impact team performance (Kozlowski, Gully, McHugh, Salas, & Cannon-Bowers, 1996). Political skill, as investigated by many scholars is one of the most important predictors of leader performance (Semadar et al., 2006). Therefore, more and more studies have established the significance of the leader’s social effectiveness on team performance (Stewart & Manz, 1997). Batt and Appelbaum (1995) observed that teams in which followers have considerable participation and autonomy (i.e., leaders give up control to a certain extent and rely on their followers) outperform teams where leaders fail to relinquish any control, thus limiting the followers’ autonomy and involvement in the teams. Leaders are usually hesitant to let followers participate substantially because it results in lesser control of the team and arises a new risk for the leader. However, on the contrary, politically skilled leaders are proficient at effectively restraining autonomous followers by using political skill as a subtle way of acquiring the team control. This is especially important because the environment that a leader fosters for follower participation has a considerable effect on team performance (Barker, 1993). Thus, politically skilled leaders with a subtle yet strong influence can manage their teams without being coercive or hostile (Ahearn et al., 2004).

In addition, the fact that transformational leadership mediates political skill and leader effectiveness illustrates the ways in which politically skilled leaders positively influence their team’s effectiveness. They can motivate their team members towards collective goals, and it allows these leaders to align their followers’ attitudes and values with the team or organizational goals and objectives (Ewen et al., 2013). Group-focused LPS helps leaders in building a shared vision and thus aligning the individual capabilities and energies of their followers towards the common goals. It results in an interdependence among the team members and a sense of motivation towards the accomplishment of collective goals. When such leaders put team interests above their own interest, team members are willing to put their individual interests aside to work for the collective good.

Furthermore, research has found that work groups and teams overseen by politically skilled leaders demonstrated higher performance than those overseen by non-politically skilled leaders (Ahearn et al., 2004; Douglas & Ammeter, 2004). This role requires the necessary political skills to be successful in working with and through others to positively influence team performance (Liu, Wang & Cao, 2011; Ewen et al., 2013). Thus, on the basis of the above literature, we propose the following hypothesis.

H2. Group-focused LPS is positively related to team performance.

2.4. Cross-level effect

Individual-focused LPS and group-focused LPS processes, although separate from each other but are not entirely independent of each other. Instead, they are linked to each other because of the cross-level effects. A cross-level effect is the influence of a team-level variable on an individual-level variable. We anticipate that group-focused LPS behavior will result in the enhancement of the individual-level outcomes because individual members can relate to their team leader’s behaviors. Even though group-focused LPS is focused towards accomplishing team tasks and is directed towards team members as a whole, this behavior may also strengthen ties between the leaders and their followers at the individual-level as a result of a trickle-down effect. Thus, leaders are consequently followed because of the consistency in followers’ beliefs and the leader’s actions.

When a leader focuses on shared beliefs and shared values in order to articulate a shared vision, followers can relate and identify with leader’s vision and beliefs in order to accomplish team goals (Lord, Brown, & Freiberg, 1999). Politically skilled leaders have a confident aura and can act as strong role models, thus enhancing their followers’ belief in the team as well as the individual efforts of the employees. Such group-focused LPS will not only result in the articulation of a shared vision, but will also act as a source of motivation, and inspiration for the followers (Wihler et al., 2016). Politically skilled leaders are good at building invaluable and effective connections and associations with their followers (Ferris et al., 2003). Thus, followers will feel determined to accomplish their individual as well as organizational goals. Group-focused LPS while influencing their followers’ beliefs, values and vision will also positively influences their individual performance. Thus, on the basis of the above literature, we propose the following hypothesis.

H3. Group-focused LPS is positively related to followers’ in-role performance.

3. Methods

3.1. Data collection and sample

The data were collected from teams working in multiple banks. Survey method was used, and the employees along with their leaders were fully informed about the aims and scope of this research. They were also assured that their answers would be kept confidential. The data were collected from 61 teams (i.e., 305 followers and 61 leaders). Volunteer employees were requested to rate the perceptions of their leader’s political skill and 61 leaders rated the performance of their followers at the individual as well as the team-level. In order to get matched reports of the leaders and followers (i.e., followers reporting about their leaders and leaders reporting about their followers) the survey packet included six questionnaires, one for the leader and remaining five for the followers. Colored questionnaires were used in order to get matched surveys of each team. The color of each follower’s survey matched with the color of the survey of the leader for that particular employee. After gaining the consent of the leaders, we contacted their teams, and the volunteer followers were handed over their questionnaires with envelops to seal their responses in order to maintain the confidentiality of their responses.

3.2. Individual-focused leader political skill

We measured individual-focused leader political skill by adapting the 6-item political skill scale developed by Ferris et al. (1999). To capture the followers’ perceptions of individual-focused leader political skill, each item was allocated an individual-focused referent. Sample items are “My leader is able to make me feel comfortable and at ease around themselves” and “My leader understands me well.” The scale showed good reliability and was rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree).

3.3. Group-focused leader political skill

Group-focused leader political skill was also assessed by adapting the 6-item political skill scale developed by Ferris et al. (1999). In an effort to capture the followers’ perceptions of group-focused leader political skill each item was allocated a team-focused referent. Sample items are “Our team leader finds it easy to envision themselves in the position of others in the team” and “Our team leader usually tries to find common ground with the team.” The scale showed good reliability and was rated on a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). A composite score of group-focused leader political skill for each team was calculated by averaging team members’ ratings. These composite scores were supported by adequate within-team interrater agreement and interrater reliability indices.

3.4. Follower performance

We measured followers’ in-role performance by the 7-item scale developed by Williams and Anderson (1991). The measure was rated by their corresponding leaders. Example items include “This employee adequately completes assigned duties” and “This employee meets formal performance requirements of the job.” The scale showed good reliability and was rated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

3.5. Team performance

Team Performance was assessed by a 3-item scale used by Schaubroeck, Lam, and Cha (2007). The sample items are “This team is very competent,” and “This team gets its work done very effectively.” The items were rated by the corresponding team’s leader on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) and the scale showed good reliability.

3.6. Control variables

We measured followers’ gender, age, and experience in order to control their impact on followers’ in-role performance. As according to the previous studies certain individual characteristics such as age, gender, and work experience can positively affect in-role performance because accumulated knowledge results in a better performance (Ng & Feldman, 2009; Strober, 1990). Similarly, for the group level, following previous studies (Chen, Kirkman, Kanfer, Allen, & Rosen, 2007) we measured team size, in order to control its documented impact on team performance. But, following Becker’s (2005) recommendations that a control variable should only be used when it significantly correlates with the dependent variable. We computed correlation coefficients between control variables and the dependent variable at their corresponding levels. As it can be seen in Table 5 that at the individual level, we found that all of the above three individual level control variables (i.e., followers’ age, gender, and experience) were unrelated to the followers’ in-role performance and thus were not included in the subsequent analysis. Whereas Table 6 shows that team size at the group level significantly correlated with the team performance and thus was used in the final analysis.

Table 1. Exploratory factor loadings of individual-level variables, KMO = 0.90

Scales	Factor loadings
	1	2
Individual-focused leader political skill
My leader finds it easy to envision his/herself in my position.	.779	.107
My leader is able to make me feel comfortable and at ease around him/herself.	.806	.015
It is easy for my leader to develop good rapport with me.	.849	−.003
My leader understands me well.	.842	−.052
My leader is good at getting me to respond positively to him/her.	.811	−.040
My leader usually tries to find common ground with me.	.800	−.016
In-role performance
This employee completes his/her assigned duties properly.	.015	.779
This employee fulfills responsibilities mentioned in his/her job description.	−.039	.820
This employee performs tasks that are expected of him/her.	−.035	.832
This employee meets formal performance requirements of the job.	.035	.787
This employee fails to perform essential duties.	−.014	.859
This employee engages in activities that will directly affect his/her performance evaluation.	.009	.810
This employee neglects aspects of the job that he/she is obligated to perform.	.031	.786

Table 2. Exploratory factor loadings of group-level variables, KMO = 0.84

Scales	Factor loadings
	1	2
Group-focused leader political skill
Our team leader finds it easy to envision his/herself in the position of others in the team.	.786	.014
Our team leader is able to make most people in the team feel comfortable and at ease around him/herself.	.829	−.033
It is easy for our team leader to develop good rapport with the team.	.839	.006
Our team leader understands team well.	.818	.011
Our team leader is good at getting the team to respond positively to him/herself.	.838	.027
Our team leader usually tries to find common ground with the team.	.835	−.029
Team performance
The team under my supervision is very competent.	.031	.908
The team under my supervision works effectively.	−.058	.849
The team under my supervision performs its job well.	.025	.855

Table 3. CFA Loadings & Cronbach’s Alpha (α)

Variables	Loading	Cronbach’s α	CR	AVE
Individual-level
Individual-focused leader political skill	0.74–0.82	0.89	0.89	0.59
In-role performance	0.73–0.83	0.91	0.91	0.60
Team-level
Group-focused leader political skill	0.73–0.81	0.90	0.90	0.61
Team performance	0.71–0.93	0.83	0.84	0.65

Note CR represents Composite Reliability, and AVE represents Average Variance Extracted.

Table 4. Measurement models for main constructs

	Models	Factors	χ^2	df	χ^2/df	RMSEA	SRMR	GFI	NFI	TLI	CFI	Δχ^2
M1	Hypothesized two-factor model (individual-level variables)	Individual-focused LPS; In-role Performance;	153.65**	64	2.40	0.06	0.04	0.92	0.93	0.95	0.96	718.75**
M2	Hypothesized two-factor model (team-level variables)	Group-focused LPS; team performance	76.06**	26	2.92	0.08	0.03	.94	.94	.95	.96	796.34**
M3	Two-factor model	Scales rated by the leader; Scales rated by the followers	872.4**	208	4.87	.10	.07	.78	.79	.81	.83

*p < .05 **p < .01

Table 5. Mean, standard deviation and correlations of individual-level variables

	Variables	Mean	SD	1	2	3	4	5
1	Age	31.42	6.21
2	Gender	.314	.46	−.179**
3	Experience	5.01	4.23	.668**	−.075
4	Individual-Focused Leader Political Skill	5.65	.96	.091	.195**	.104	0.77
5	In-role performance	4.16	.62	−.066	.067	.015	.184**	0.77

Diagonal represents the square root of average variance extracted (AVE); whereas estimated correlations are shown below the diagonal.

NOTE: N = 305. *p < 0.05. **p < 0.01.

Table 6. Mean, standard deviation and correlations of team-level variables

	Variables	Mean	SD	1	2	3
1	Team size	8.70	3.35
2	Group-focused leader political skill	5.51	.67	.041	0.78
3	Team performance	4.11	.55	.175**	.154**	0.80

Diagonal represents the square root of average variance extracted (AVE), whereas estimated correlations are shown below the diagonal.

NOTE: N = 305. *p < 0.05. **p< 0.01.

4. Results

Two exploratory factor analyses were conducted using principle component extraction method along with Promax rotation to assess the unidimensionality of the constructs. First factor analysis was performed on the lower-level variables which showed a meritorious KMO (0.90), a significant Bartlett’s test of sphericity and the two-factor solution explained more than 66% variance. Table 1 shows all the factor loadings of the items on their corresponding factors. Further, no cross-loadings were of sufficient magnitude to reject unidimensionality also establishes the unidimensionality of the constructs.

As a pragmatic approach, Table 1 reveals that both the convergent and discriminant validity was established, as all the loadings were higher than 0.6 individually and their averages are above 0.7 (Fornell & Larcker, 1981). Moreover, the factor correlation matrix revealed that the correlations were less than 0.18 confirming discriminant validity.

Second factor analysis was performed on the team-level variables which also showed a meritorious KMO (0.84), a significant Bartlett’s test of sphericity and the two-factor solution explained more than 65% variance. Table 2 shows all the factor loadings of the items on their corresponding factors. Further, no cross-loadings of sufficient magnitude were found that can reject unidimensionality. The table also reveals that both the convergent and discriminant validity was established, as all the loadings were higher than 0.6 individually and their averages are above 0.7 (Fornell & Larcker, 1981). Moreover, the factor correlation matrix revealed that the correlation was 0.11 confirming discriminant validity.

4.1. Confirmatory factor analysis

Further, we conducted two sets of confirmatory factor analyses, one for the lower-level variables and the second for the team-level variables. Table 3 provides the loadings, composite reliability (CR) and AVE of all the variables. Internal consistency, reliability, and validity of the constructs were established by measuring both the Cronbach’s alpha and the composite reliability. Although the Cronbach’s alpha values greater than 0.70 indicate a good reliability (Nunnally, 1978), but it has certain limitations; therefore, composite reliability was also used for the assessment of internal consistency reliability. The table shows that the alpha reliabilities (.83–.91) and composite reliabilities (.84–.91) were above .70, and average variances extracted of all variables were comfortably above .50 which establishes the convergent validity of the variables (Hair, Black, Babin, Anderson, & Tatham, 2010).

Table 4 shows the values of the two-factor confirmatory analysis of lower-level variables, the team-level variables and another two-factor model which combines the scales rated by the leader on one factor and the scales rated by the followers on the other factor. This two-factor model was included in order to show that if these results were a product of common method bias then M₃ (i.e. scales rated by the leader on one factor and scales rated by the followers on another factor) should show a better model fit than other models (i.e., each of the other models have one scale rated by the leader on one factor and the scale rated by the followers on the other factor). However, the table shows that the model fit for M₁ and M₂ is significantly better than the model fit for M₃.

Next, we assessed the discriminant validity of the constructs. We used Fornell and Larcker (1981) criterion as a more conservative measure to establish discriminant validity. According to this criterion, the square root of AVE of each construct must be greater than all the correlations of that construct with other constructs. These values are reported on the diagonals in the correlation matrix (see Table 5), and a careful examination of these values indicates that square root value of each constructs’ AVE is comfortably above its correlations with all the other constructs. Thus, discriminant validity among the study constructs is established.

4.2. Descriptive statistics and correlations

Table 5 and Table 6 show the means, standard deviations, and correlations of all the lower and team-level variables and correlation of all the variables along with the square root of AVE for all the constructs in parenthesis on diagonals. All the lower-level variables and the team-level variables showed significant positive correlations, thus, lending preliminary support to our hypotheses.

The variable group-focused leader political skill was assessed at the individual-level and was analyzed at the team-level of analysis; thus we had to aggregate this variable to the team-level for further analyses. Hence in order to assess the agreement within teams, we calculated a number of multilevel statistics. First of all, we calculated rWG(J), intraclass correlation (ICC1) and intraclass correlation (ICC2) as a measure of the reliability of team means (Bliese, Klein, & Kozlowski, 2000). Lastly, we ran F-tests to measure if average scores differed significantly across teams or not. In order to calculate rWG(J), we applied uniform distributions, and the application of a uniform distribution yielded a mean of 0.92 along with positive values (.64≤ rWG(J)≤.99) and only one negative estimate. So, in this case, the value was reset to 0 because it is estimated that with a larger sample of raters the values will conform to the common range from 0 to 1. The ICC1 was .34, and ICC2 was .72, F = 3.58, p < .001.

The suggested cut-off values for the interpretation of inter-rater agreement are between .60 and .70 whereas, the rWG(J) values between .51 and .70 also indicate a moderate agreement (LeBreton & Senter, 2008). Thus, based on these criteria, consensus model, the underlying theoretical null distribution, as well as the number of raters we inferred that the rWG(J) values in this study are adequate to signify a moderate agreement within teams for the aggregation of the group-focused leader political skill.

Table 7 shows the multilevel statistics of this study. The ICC1 value for this study was .34 (for group-focused leader political skill) indicating large effect, which means that the team membership explained substantial variance in individual’s perceptions of group-focused leader political skill. Furthermore, the value of ICC2 for this study also showed good levels of reliability, i.e. .72. To sum it up, we deduced that these results support the aggregation of the lower-level measure of group-focused leader political skill for further analyses.

Table 8 shows the hypotheses testing results. Hypothesis 1 posited that individual-focused leader political skill is positively related to followers’ performance. The results revealed that individual-focused leader political skill significantly predicts their followers’ performance, F (1, 303) = 10.568, p < .001, with an R² of .034. Hypothesis 2 proposes that group-focused leader political skill is positively related to their followers’ team performance. Results revealed that group-focused leader political skill significantly predicts their followers’ team performance, F (1, 302) = 8.297, p < .001, with an R² of .052. Hypothesis 3 proposes that group-focused leader political skill is positively related to their followers’ performance at the individual-level. Results revealed that group-focused leader political skill significantly predicts their followers’ performance at the lower-level (F (1, 303) = 11.872, p < .001), with an R² of .038.

Table 7. Multilevel statistics

Variables	One way ANOVA	ICC1	ICC2	Mean	SD	F-Ratio	Range
Group-focused leader political skill	Sig	0.34	0.72	0.92	0.14	4.28	0.64–0.99

Table 8. Hypotheses test results

				BC 95% CI^1
	Path	Path coefficient	SE	Lower	Upper
H1	Individual-Focused LPS—> In-Role Performance	0.11**	0.03	0.04	0.19
H2	Group-Focused LPS—> Team Performance	0.12**	0.04	0.03	0.21
H3	Group-Focused LPS—> In-Role Performance	0.18**	0.05	0.07	0.28

*p < .05, **p < .01

5. Discussion

Previous scholars have examined the direct effects of leader political skill on followers’ performance as well as their attitudes, but those studies have not considered the impact of the dual effects of leader political skill on followers’ outcomes at different levels. Similarly, no attempt has been made to measure the cross-level effect of group-focused LPS on individual performance of the followers. Therefore, employing the theoretical notions, we hypothesized that politically skilled leaders would not only have dual political skill effects (i.e., at both the individual and the team-levels in parallel), but the group-focused LPS will also have a trickle-down effect on the individual outcomes. We hypothesized that the group-focused LPS would predict team outcomes like team performance, and the individual-focused LPS will predict individual outcomes like individual performance in parallel. Moreover, group-focused LPS will also have a trickle-down effect and will influence the followers’ performance at the individual-level. When a leader displays group-focused LPS behaviors, followers may not only connect with the leader’s behaviors towards the team but also may feel more motivated towards individual outcomes. That is, followers may view their leader as more reliable and dependable figures and put more effort towards the individual goals as well as the shared vision articulated by the team’s leader. Consequently, the group-focused LPS itself then becomes an integral part of the leader’s aura in how followers perceive their leaders which ultimately develops strong associations between the leader and individual followers.

5.1. Theoretical implications

Our study offers empirical proof of the integrative multilevel model of the dual effects of LPS and follower outcomes at both the individual and the team-levels. Its contribution lies in answering the calls for multilevel research on leader political skill (Kozlowski & Bell, 2003). Our research advances the leader political skill theory by displaying that dual effects of leader political skill (i.e., individual-focused LPS and group-focused LPS) can predict different outcomes at their corresponding levels. Our study explicitly addresses the concerns raised by various researchers regarding the proper levels of analysis while building theory and testing hypotheses of multilevel models (Hox, Moerbeek, & Van de Schoot, 2002; Yammarino, Dionne, Chun, & Dansereau, 2005).

The second contribution of our study lies in testing the cross-level effect of group-focused LPS to the followers’ performance at the individual-level. Although directed towards the team members as a whole, group-focused LPS still influences followers at the individual-level. This cross-level effect is the result of the trickle-down influence of group-focused LPS on the individual outcomes of the followers.

5.2. Managerial implications

This study offers two important implications for managers. Foremost, given that team-work is becoming more and more valuable in this era’s modern organizations, leaders are now expected to motivate their followers towards both the individual and the team goals simultaneously (Kozlowski & Bell, 2003). Thus, this research signifies that the dual effects of leader political skill can drive both the individual and the team performance as a whole. Similarly, it is very crucial for the leaders to realize which sets of behavior are expected from them in order to accomplish their own and organizational goals with and through their followers. For instance, when the leader is held responsible for the performance of their teams, but behaviors being exhibited by the team lack team cohesion and collective efforts, now is the time for the leader to display group-focused LPS to improve team performance. On the contrary, when dealing with individual followers, leaders must be adept at displaying individual-focused LPS tailored towards the needs of individual followers in order to drive their individual performance. In essence, leading both the individuals and the team at the same time is a challenging job for the leaders. Whereas the key to pulling this task successfully is to display the corresponding LPS behavior aligned with the corresponding levels of the outcomes.

Furthermore, this study has been carried out in a unique cultural context because most of the political skill studies have been done in western cultures that emphasize on individualistic aspects of their cultures; however, this study elucidates that the dual effects of LPS can predict favorable outcomes at two levels in organizations working in this culture.

5.3. Limitations and future implications

One of the limitations of our study is the cross-sectional nature of the data; thus, care should be taken in drawing any causal inferences (i.e., one may argue that the positive experiences of leader effectiveness lead to a rise in political skill of leaders), however, theory would predict that that our hypotheses accurately place leader political skill as an antecedent to follower outcomes. Moreover, in order to reduce any other biases, the constructs under observation were assessed by two sources. The performance of the teams was assessed by their corresponding team leaders, and the individual and group-focused LPS was evaluated by the followers’ perceptions of their leader’s political skill aggregated at the team-level. Although we had different raters for the constructs used in this study, but the fact that leaders rated their own team’s performance might have resulted in the self-serving bias, that is, leaders might have elevated the ratings of their own teams to come across as an effective leader. Thus, future research can benefit from using more sources of rating performance, for instance, asking the supervisor of the leader to rate team performance instead of relying only on the leader’s ratings. Nevertheless, on the other hand, to reduce the selection bias, the followers’ participation was voluntary (i.e., they were neither nominated by their leaders nor were they forced by their leaders, instead they were contacted by the researchers themselves) and were fully informed about the purpose of the study. Therefore, in contrast to the leader-nominated followers, such sample has lesser selection bias and are expected to give more accurate responses.

Moreover, future research can also assess the impact of differentiated LPS on various followers in a team. A politically skilled leader may have a better connection with some followers than with others, and this differentiated LPS might affect the dynamics of the team on the whole. Similarly, other contextual moderators may also affect the team dynamics and their performance; thus their inclusion can also open more avenues for future research. Scholars can also test multilevel mediation models to explore intermediate linkages. Such intermediate linkages will provide more insights into how these leader characteristics influence follower outcomes. Therefore, our study can serve as a basis of reference for these further studies on multilevel LPS behaviors.

6. Conclusion

Since the introduction of Political theory of leadership, researchers have explored the favorable outcomes of leader political skill at both the individual and the team-levels separately. But these studies have neither integrated the individual and team levels nor have drawn a clear distinction between leader-follower and leader-team interactions. Whereas our study investigates the leader political skill forms a multilevel perspective by drawing a distinction between the group-focused LPS and individual-focused LPS. This integration has allowed us to investigate the trickle-down effect of group-focused LPS on followers’ performance at the individual level. Furthermore, the distinction between the dual effects of LPS will also allow future researchers to test multilevel models and investigate further cross-level effects.

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Beenish Qamar

Beenish Qamar is a Ph.D. scholar at Lyallpur Business School, Government College University Faisalabad, Pakistan. Her research interests include political skill at work, self-concept, and leadership.

Sharjeel Saleem

Sharjeel Saleem is an assistant professor at Lyallpur Business School, GCUF, Pakistan and has obtained his Ph.D. from University of Vienna, Austria. His research interests include expatriate adjustment, organizational behavior, and leadership. He has published in renowned international journals and has presented at international conferences including the Academy of Management Conference and AHRD Conference. He is currently reviewing for international journal of human resource management and personnel review.

Mohsin Bashir

Mohsin Bashir is an assistant professor at Lyallpur Business School, GCUF, Pakistan and has obtained his Ph.D. from Huazhong University of Science and Technology, China. He has published his work in well-known international journals and has a vast experience in industry and academia. His research interests include culture, organizational behavior and organizational psychology.