INTRODUCTION
Research shows that effective leadership behaviors are different and not easily transferable across different cultural contexts (Agarwal, DeCarlo, & Vyas, Reference Agarwal, DeCarlo and Vyas1999; Aycan, Schyns, Sun, Felfe, & Saher, Reference Aycan, Schyns, Sun, Felfe and Saher2013; Pillai, Scandura, & Williams Reference Pillai, Scandura and Williams1999), however, the reasons behind this remain little understood. An example of transformational leadership (TL) offers a case in point (Avolio, Walumbwa, & Weber, Reference Avolio, Walumbwa and Weber2009; Judge & Piccolo, Reference Judge and Piccolo2004; van Knippenberg & Sitkin, Reference van Knippenberg and Sitkin2013). Whereas it is generally presumed that TL exerts a number of positive effects on employees, we do not understand well the processes and mechanisms through which it exerts these effects nor the reasons behind their cross-cultural variation (Kark, Shamir, & Chen, Reference Kark, Shamir and Chen2003; Walumbwa, Avolio, & Zhu, Reference Walumbwa, Avolio and Zhu2008; Cho & Dansereau, Reference Cho and Dansereau2010). In this article, we shed light on these issues by examining the cross-cultural variation of the effects of idealized influence and individualized consideration – two TL behaviors – on followers’ organizational identification in the two culturally distinct contexts of Finland and Russia.
Our study builds on three recent advancements in our understanding of TL in particular and leadership in general. First, there is a growing acknowledgement that in the past leadership literature has largely been leader-centric (Hollander, Reference Hollander1992; Uhl-Bien, Riggio, Lowe, & Carsten, Reference Uhl-Bien, Riggio, Lowe and Carsten2014), focusing on the role of leaders in motivating or directing followers in action and goal achievement. In contrast, followers have been conceived as recipients of leaders’ influence who comply without resistance or initiative to leaders’ orders and directives (e.g., Shamir, Reference Shamir, Shamir, Pillai, Bligh and Uhl-Bien2007). However, recently, research began to advocate a relational view of leadership that acknowledges the role of followers in the co-creation of leadership together with leaders (Fairhurst & Uhl-Bien, Reference Fairhurst and Uhl-Bien2012; Shamir, Reference Shamir, Uhl-Bien and Ospina2012). This literature attributes a greater role for followers’ traits and values in defining leadership and its behavioral outcomes, thus moving away from treating leaders’ style as the sole antecedent to organizational outcomes (Shamir, Reference Shamir, Shamir, Pillai, Bligh and Uhl-Bien2007; Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014). It argues that to study leadership, research needs to account for followers since without their recognition and granting legitimacy to leaders’ influence attempts there is no leadership (DeRue & Ashford, Reference DeRue and Ashford2010).
Second, there is a growing recognition that leadership is a social practice framed by social norms, socio-emotional traits, and cognitive values pertinent to the cultural context in which it takes place (see Dickson, Den Hartog, & Mitchelson, Reference Dickson, Den Hartog and Mitchelson2003; Gelfand, Bhawuk, Nishi, & Bechtold, Reference Gelfand, Bhawuk, Nishi, Bechtold, House, Hanges, Javidan, Dorfman and Gupta2004). In line with this, scholars began to examine contextual/cultural factors that may affect how, for example, TL exerts its influence on followers (Jackson, Meyer, & Wang, Reference Jackson, Meyer and Wang2013; Jung, Yammarino, & Lee, Reference Jung, Yammarino and Lee2009; Walumbwa, Lawler, & Avolio, Reference Walumbwa, Lawler and Avolio2007). Contrary to the claim that TL transcends organizational and national borders (e.g., Bass, Reference Bass1997), there is growing evidence (although mainly in Western contexts) indicating that the effects and mechanisms of leadership behaviors are often culture-contingent (see House, Hanges, Javidan, Dorfman, & Gupta, Reference House, Hanges, Javidan, Dorfman and Gupta2004; Paris, Howell, Dorfman, & Trafimow, 2009; Pillai et al., Reference Pillai, Scandura and Williams1999). In other words, to be accepted and effective, leadership behaviors have to be in congruence with the norms of the culture in which the leader operates (Wendt, Euwema, & Van Emmerik, Reference Wendt, Euwema and Van Emmerik2009).
Finally, given that the focus of this article is on TL behaviors, several scholars have initiated a discussion concerning the need to scrutinize the construct of TL in more detail than previously (van Knippenberg & Sitkin, Reference van Knippenberg and Sitkin2013; Wang & Howell, Reference Wang and Howell2010; Wu, Tsui, & Kinicki, Reference Wu, Tsui and Kinicki2010). They inquire whether by lumping different leadership behaviors under the conceptual umbrella term of ‘TL’ scholars disregard the important nuances of TL effects on employees. Preserving and paying attention to these nuances may yield new discoveries and improve our understanding of TL, its constitutive behaviors and their effects. For instance, van Knippenberg and Sitkin (Reference van Knippenberg and Sitkin2013: 46) contend that the concept of TL needs to be disentangled into its constitutive dimensions and the effects of these dimensions have to be analyzed separately, ‘unrestrained by the conceptual baggage and poor measurement associated by [their] inclusion as …element[s] of… transformational leadership’. As a way forward, van Knippenberg and Sitkin (Reference van Knippenberg and Sitkin2013: 3) suggested that ‘theory and measurement [should] concentrate on conceptualizing and operationalizing more precise and distinct elements and effects of leadership without the handicap of the higher-order label of charismatic-transformational leadership’. Instead of lumping different dimensions of TL together to form the higher-order construct – as it has been the dominant practice in prior research – they advocated examining the effects of individual TL behaviors on followers.
Hence, applying the above to the case of TL behaviors, we can expect the following. First, followers play an important role in defining which behaviors or influence attempts qualify as TL. Second, considering the natural diversity among followers due to their individual and cultural differences, different followers are likely to perceive TL behaviors differently. Finally, the different constitutive behaviors of TL are likely to have (or perceived to have) different effects on different followers.
Given the above, we shed light on these currently little understood issues in two ways. First, we theorize and examine the cross-cultural variation of the effects of TL behaviors. We do so by analyzing the effects of two specific constitutive behavioral dimensions of TL, namely idealized influence and individualized consideration leadership behaviors, on followers’ organizational identification in the two different cultural contexts of Russia and Finland. We focus on these particular behaviors because research found them to exert the most salient influences on followers (e.g. Judge & Piccolo, Reference Judge and Piccolo2004; Lowe, Kroeck, & Sivasubramaniam, Reference Lowe, Kroeck and Sivasubramaniam1996). Idealized influence leadership entails role-modeling behavior that aims to motivate followers to internalize the leader's vision, values, and mission, whereas individualized consideration leadership aims to attend to followers’ emotional needs, to act as a support, and to listen to followers’ concerns (Piccolo & Colquitt, Reference Piccolo and Colquitt2006). Context-wise, we argue that Russia and Finland provide a suitable context for testing the cross-cultural variation of the effects of TL behaviors on followers because the two countries possess very distinct cultural profiles, especially so in terms of individualism and power distance (see Hofstede, Reference Hofstede2001). We test our theoretical ideas by analyzing the perceptions of white-collar employees in four Finland-based multinational corporations (n = 295) and their subsidiaries in Russia (n = 104) concerning their proximal team leaders, who are middle-level managers.
Second, to illuminate the mechanisms through which the two TL behaviors operate, we examine the role of role ambiguity, which reflects both the followers’ role and self-concept, as a mediator in the relationship between TL behaviors and followers’ organizational identification. We define role ambiguity as a lack of clarity about expectations, tasks and goals and an unpredictability in the consequences of one's role performance (Ilgen & Hollenbeck, Reference Ilgen and Hollenbeck1991). Whilst research found that role ambiguity has important organizational implications that influence employees’ job satisfaction, turnover intentions and performance (O'Driscoll & Beehr, Reference O'Driscoll and Beehr1994; Tubre & Collins, Reference Tubre and Collins2000), it is also crucial in influencing followers’ perceptions of leadership in an organization, which is in line with the follower-centric approach to leadership. Followers’ perceptions of a particular leadership behavior hinge on their follower role, which refers to followers’ perceptions of their position in relation to leaders (see Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014). A range of individual, relational and work unit factors can influence how the follower role is defined (Shamir, Reference Shamir, Shamir, Pillai, Bligh and Uhl-Bien2007). Given that role perceptions directly influence role behaviors (see Katz & Kahn, Reference Katz and Kahn1978), these factors and characteristics are important drivers of how followers perceive leadership and what outcomes they achieve.
The article contributes to the leadership literature in three ways. First, it increases our understanding of the followers’ perspective on leadership in general and TL behaviors in particular (Fairhurst & Uhl-Bien, Reference Fairhurst and Uhl-Bien2012; Hollander, Reference Hollander1992; Shamir, Reference Shamir, Uhl-Bien and Ospina2012; Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014) by illustrating how followers with different cultural backgrounds perceive and react to TL behaviors differently. Second, it sheds light on the effects of two specific TL behaviors (van Knippenberg & Sitkin, Reference van Knippenberg and Sitkin2013; Wang & Howell, Reference Wang and Howell2010; Wu et al., Reference Wu, Tsui and Kinicki2010) by explicating how they exert their differential influences (as perceived by followers) on organizational identification in two culturally distinct contexts. Finally, it elucidates the role of a self-concept-based mechanism in transforming leadership behaviors into crucial organizational outcomes (Kark & Shamir, Reference Kark and Shamir2002; Shamir, House, & Arthur, Reference Shamir, House and Arthur1993; Shamir, Zakay, Breinin, & Popper, Reference Shamir, Zakay, Breinin and Popper1998, Reference Shamir, Zakay, Breinin and Popper2000) by examining how followers’ role ambiguity – as a manifestation of followers’ perceived roles in their relationship with proximal leaders – mediates the relationship between leadership behaviors and organizational identification.
The article unfolds as follows. First, we introduce the self-concept-based theory of leadership, which we later use to theorize the differential effects of TL behaviors in Finland and Russia, and to discuss the cultural contingency of TL effects and the distinct effects of TL behaviors. Second, we develop a set of hypotheses. Third, we test them using linear mixed modeling. Finally, we present and discuss our findings, and then conclude.
THEORETICAL BACKGROUND AND HYPOTHESES
The Self-Concept-Based Motivational Theory of Leadership
The theory postulates that the explanation for TL effects lies in understanding how TL behaviors influence followers’ self-concepts (Shamir et al., Reference Shamir, House and Arthur1993). It suggests that the TL motivating effects occur because TL behaviors influence followers’ self-concepts in three key ways: by increasing follower self-efficacy, by facilitating followers’ social identification with their group, and by linking work values to follower values.
To realize these effects, a transformational leader him/herself has to engage in role modeling, whereby the leader provides an ideal, a point of reference and a role model for followers, and frame alignment, which refers to the linkage of individual and leader interpretative orientations, such that some set of followers’ interests, values and beliefs and the leader's activities, goals and ideology become congruent and complementary. When done properly, these are likely to result in several important outcomes among the followers, namely a higher salience of collective identity in the followers’ self-concepts, a stronger identification with the leader, a higher level of self-esteem and self-worth, and a higher sense of self-efficacy.
However, Shamir et al. (Reference Shamir, House and Arthur1993) acknowledge that the implied effects are not likely to be similar on all followers (e.g., Bono & Judge, Reference Bono and Judge2003; Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998). Another crucial point that the authors make is that for the TL effects in realizing TL behaviors have to be ‘congruent with the existing values and identities held by potential followers’ (Shamir et al., Reference Shamir, House and Arthur1993: 588). We may add that the effects are also likely to form themselves differently depending on the existing values and ideologies held by potential followers. It can be that the described self-implicated effects of TL on followers’ self-concepts may appear to take place and result in very similar individual outcomes, e.g., decreased turnover intentions and/or increased organizational identification, yet, the mechanisms through which these outcomes come into being will differ depending on the followers’ cultural values and orientations. Thus, we argue that, depending on the context, different types of TL behaviors can yield the motivating effects and outcomes described above.
The Cultural Contingency of Transformational Leadership Effects
Research has shown that countries differ in terms of cultural values, such as individualism–collectivism, power distance, uncertainty avoidance, femininity–masculinity (Hofstede, Reference Hofstede2001), and different cultural values have a different predictive power for a broad range of important individual attitudes and behaviors in the workplace, including leadership style preferences and outcomes (see Taras, Kirkman, & Steel, Reference Taras, Kirkman and Steel2010; Tsui, Nifadkar, & Ou, Reference Tsui, Nifadkar and Ou2007). For instance, Taras et al. (Reference Taras, Kirkman and Steel2010) found that at the individual level power distance is strongly associated with a preference for directive leadership whereas uncertainty avoidance with avoiding participative leadership.
The same appears to be true for the specific case of TL. Contrary to the claims that TL behaviors have universal effects on followers (Den Hartog, House, Hanges, Ruiz-Quintanilla, & Dorfman, Reference Den Hartog, House, Hanges, Ruiz-Quintanilla and Dorfman1999; Dorfman, Howell, Hibino, Lee, Tate, & Bautista, Reference Dorfman, Howell, Hibino, Lee, Tate and Bautista1997), research increasingly shows that TL effects are contingent on cultural factors. Studies found TL behaviors to be more effective among followers exhibiting stronger group and collectivistic orientations (Bass, Reference Bass1997; Jung et al., Reference Jung, Yammarino and Lee2009; Pillai & Meindl, Reference Pillai and Meindl1998). In collectivistic cultures, TL is a stronger motivator for followers to pursue longer-term, organizational goals instead of immediate self-interests (Jung, Bass, & Sosik, Reference Jung, Bass and Sosik1995), to engage in teamwork and embrace a collective mission (Pillai & Meindl, Reference Pillai and Meindl1998), and to internalize collective identity as part of their own self-concept (Dvir & Shamir, Reference Dvir and Shamir2003; Shamir et al. Reference Shamir, House and Arthur1993). In contrast, Jung et al. (Reference Jung, Yammarino and Lee2009) speculated that organizational cultures emphasizing individualistic values might limit the potential effects of TL on followers’ effectiveness and firm performance. Testing these claims in different cultural settings, Jung and Avolio (Reference Jung and Avolio1999) found that TL has stronger effects on performance among Asian American versus Caucasian American followers, and Jung et al. (Reference Jung, Yammarino and Lee2009) revealed a stronger relationship between TL and leader effectiveness in Korea than in the US.
Other studies found power distance to influence TL effects across cultural contexts. For instance, Kirkman, Chen, Farh, Chen, and Lowe (Reference Kirkman, Chen, Farh, Chen and Lowe2009) proposed that individual-level power distance orientation might influence the effects of TL on followers’ procedural justice perceptions. Followers with higher power distance orientations are likely to behave submissively around their leaders, to avoid disagreements by obeying their leaders’ instructions, and to perceive their leaders as respectful, superior and knowledgeable. Therefore, lower power distance orientations enhance the effects of TL on followers’ procedural justice perceptions. The authors verified this claim empirically by comparing the contexts of the US and China. In support, House et al. (Reference House, Hanges, Javidan, Dorfman and Gupta2004) also found TL effectiveness to relate negatively to power distance at the country level of analysis.
Explaining the cultural contingency of TL effects, scholars pointed out that followers in different cultures carry different implicit leadership theories or ideas about what leadership is (see Brodbeck et al., Reference Brodbeck, Frese, Akerblom, Audia, Bakacsi, Bendova and Castel2000; Tsui et al., Reference Tsui, Nifadkar and Ou2007). For instance, in a relatively rare study of this type, Ensari and Murphy (Reference Ensari and Murphy2003) studied cross-cultural differences in the attribution of charisma and found that in the individualistic culture of the US, a leader's prototypical characteristics were more effective in the formation of leadership impression, whereas company performance was more effective in leadership attributions in the collectivistic culture of Turkey. Hence, this by no means exhaustive overview points towards the cross-cultural variation of TL effects and the importance of explicating how these effects come into being.
The Distinct Effects of Transformational Leadership Behaviors
Over the years, a substantial amount of research has accumulated, indicating that individual TL behaviors, namely idealized influence, inspirational motivation, intellectual stimulation, and individualized consideration, might have different effects on followers (see Cho & Dansereau, Reference Cho and Dansereau2010; Wang & Howell, Reference Wang and Howell2010; Wu et al., Reference Wu, Tsui and Kinicki2010). For instance, Cho and Dansereau (Reference Cho and Dansereau2010) showed that in the Korean context individualized consideration and idealized influence have different effects on followers. Whereas the former enhanced leader-directed organizational citizenship behavior via interpersonal justice, the latter affected group-directed organizational citizenship behavior via procedural justice climate. Further, Wu et al. (Reference Wu, Tsui and Kinicki2010) found that individual-focused leadership, comprising individualized consideration and intellectual stimulation, facilitated leader identification and self-efficacy, whereas group-focused leadership, consisting of idealized influence and inspirational motivation, positively influenced group identification and collective efficacy among followers.
More generally, Lowe et al. (Reference Lowe, Kroeck and Sivasubramaniam1996) meta-analyzed studies that used the Multifactor Leadership Questionnaire (MLQ) measure of TL and found idealized influence and individualized consideration to have the most generalizable validities among all the four dimensions of TL. Conger and Kanungo (Reference Conger and Kanungo1998) also showed the idealized influence dimension to exhibit the strongest relationships with outcome variables as compared to other TL dimensions. Finally, in the most recent and comprehensive meta-analysis of TL research, Judge and Piccolo (Reference Judge and Piccolo2004) established the validity of idealized influence to be the highest among all TL dimensions and the most generalizable in terms of its effects on followers’ job satisfaction, leader satisfaction and motivation as well as leaders’ job performance and group or organization performance.
Thus, two conclusions stem from these studies: (1) the effects of TL constitutive dimensions on followers can be dissimilar from each other, and (2) idealized influence and individualized consideration appear the most widely studied but also the most influential TL behavioral dimensions in terms of their effects on followers.
Considering this evidence, van Knippenberg and Sitkin (Reference van Knippenberg and Sitkin2013) have noted that to advance the TL research it would be timely and relevant to take the concept of TL ‘back to the drawing board’ to disentangle both its conceptual and methodological problems that it has amassed over the years. The authors note that by treating TL as an aggregate construct research may fail ‘to specify how each [constitutive TL] dimension has a distinct influence on mediating processes and outcomes’ (2). Prior to that, others have noted the same issue as being problematic also. Bono and Judge (Reference Bono and Judge2003: 554–555) lamented that ‘so little is known about the processes by which transformational… leaders have their effects on followers” and attributed it to the lack of conceptual clarity in “the specification of leader behaviors’. Whereas Bass (Reference Bass1999: 24) argued that ‘much more explanation is needed about the inner workings of transformational leadership’. We may add that we also know little about the cultural variation of the effects of TL constitutive dimensions.”
Thus, given the above, in the following section, we employ the self-concept-based theory of leadership (Kark & Shamir, Reference Kark and Shamir2002; Shamir et al., Reference Shamir, House and Arthur1993; Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998, Reference Shamir, Zakay, Breinin and Popper2000) to hypothesize and test the effects of idealized influence and individualized consideration on followers’ organizational identification in multinational corporations operating in the culturally distinct contexts of Russia and Finland.
Hypotheses’ Development
Transformational leadership and followers’ organizational identification
Research shows that TL positively affects followers’ organizational identification (Kark & Shamir, Reference Kark and Shamir2002; Kark et al., Reference Kark, Shamir and Chen2003; Epitropaki, Reference Epitropaki2013; Epitropaki & Martin, Reference Epitropaki and Martin2005; Kark & Shamir, Reference Kark and Shamir2002; Kark et al., Reference Kark, Shamir and Chen2003; Shamir et al., Reference Shamir, House and Arthur1993), defined as the process whereby followers’ beliefs and values in relation to their organization become self-referential and self-defining (Pratt, Reference Pratt, Whetten and Godfrey1998). Identification implies that followers base their self-concept and self-esteem at least partly on their belonging to the organization, experiencing successes and failures of the organization as personal successes and failures (Kark & Shamir, Reference Kark and Shamir2002; Mael & Ashforth, Reference Mael and Ashforth1992; Shamir et al., Reference Shamir, House and Arthur1993).
However, the existing literature does not usually differentiate between the dimensions of TL when examining the relationship between TL and followers’ organizational identification, thus not explicating what TL dimensions prime the organizational identification of different followers. The usual argument is that TL – as an amalgamation of different behaviors and practices – primes organizational/collective identification (Epitropaki, Reference Epitropaki2013; Epitropaki & Martin, Reference Epitropaki and Martin2005; Walumbwa et al., Reference Walumbwa, Avolio and Zhu2008) or leader identification (Kark et al., Reference Kark, Shamir and Chen2003; Liu, Zhu, & Lang, Reference Liu, Zhu and Yang2010) among followers.
Moreover, the rare studies, which differentiate between TL dimensions’ effects on organizational identification, are either of conceptual (Kark & Shamir, Reference Kark and Shamir2002) or experimental (Paul, Costley, Howell, Dorfman, & Trafimow, Reference Paul, Costley, Howell, Dorfman and Trafimow2001) nature. Alternatively, they focus solely on charismatic leadership and its constitutive behaviors and use new measurement instruments developed specifically for the research setting in question, such as military units (see Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998, Reference Shamir, Zakay, Breinin and Popper2000).
As a result, the nature of TL dimensions’ effects on followers’ organizational identification in general remains poorly understood and empirically verified. Moreover, research showing how these effects differ across different cultural contexts is, to the best of our knowledge, non-existent. Yet, we foresee that idealized influence and individualized consideration are likely to interact differently with followers’ organizational identification in Russia and Finland.
Idealized influence/individualized consideration and organizational identification in Finland
We expect that in Finland, a Western context with high individualism, low power distance and masculinity, and average uncertainty avoidance scores (see Table 1), idealized influence will be effective in influencing followers’ self-concepts for priming their organizational identification. Given Finland's cultural scores, to facilitate followers’ organizational identification Finnish leaders will be expected to provide a generally compelling and transparent description and image of the organization, detailing what are its objectives, goals, values and the overall mission, thus pinpointing its unique and distinct features (Ashforth & Mael, Reference Ashforth and Mael1989; Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998).
Table 1. Comparison of Russia and Finland on Hofstede's cultural dimensions (Hofstede, Reference Hofstede2001)
Furthermore, they are likely to do well by stimulating a friendly and collaborative workplace atmosphere, which is conducive to team work and knowledge sharing amongst followers, and providing a personal example, i.e., role modelling, for their followers, demonstrating commitment and dedication to the organization and its values. By doing so, leaders will be able to map the general boundaries, i.e. frame alignment, within which employees could then perform their work tasks more or less independently but in alignment with organizational goals and values (cf. Kark & Shamir, Reference Kark and Shamir2002).
Thus, by (i) supplying the followers with clear and compelling organizational values, vision and objectives, (ii) stimulating efficient teamwork, and (iii) providing an example of organizational commitment and identification, Finnish leaders are likely to enhance their followers’ self-concepts and motivate them to perceive both their organizational and own goals and values as being congruent. This is then likely to increase the followers’ organizational identification.
At the same time, the low power distance and highly individualistic orientations in particular are likely to encourage followers, in Finland, to question freely their leaders’ perceiving them as constituents of their work teams. In Finland, a leader's opinion about an operational or organizational issue is but one of the opinions, although of a somewhat higher weight, to be considered. The perception that one can question the leader's opinion and the realization that one is relatively independent from one's direct leader's supervision and close attention to act within the framework of organizational values and objectives is likely to increase the follower's sense of empowerment and control over what and how needs to be done (Randolph & Sashkin, Reference Randolph and Sashkin2002).
On the contrary, subjected to individualized consideration, followers in Finland might regard leaders’ personal attention and interest in their feelings as unnecessary, a violation of privacy, and an indirect attempt to control and monitor the ways they work (cf. Pellegrini & Scandura, Reference Pellegrini and Scandura2006). Perceiving their workplace as infused with such a type of subtle monitoring and control is neither likely to boost Finnish employees’ self-concepts nor make them identify with their organizations. Therefore, in contrast to idealized influence, we anticipate that individualized consideration, in Finland, is not likely to be very effective in facilitating followers’ organizational identification.
Hence, in Finland, leaders are more likely to prime followers’ organizational identification by engaging in more impersonal and (socially) distant types of leadership (see Shamir, Reference Shamir1995), thus providing followers with an empowering environment to facilitate their self-realization of being part of a larger collective (i.e., organization) with compelling and inspiring values and objectives. By providing a positive vision of the future, motivating employees to internalize an organizational mission, values and objectives, leaders are likely to motivate followers to transcend their own self-interests and to start perceiving their job-related activities as contributing to a larger common good and thus be more meaningful and rewarding (see Jung et al., Reference Jung, Yammarino and Lee2009; Shamir et al., Reference Shamir, House and Arthur1993). This will then be conducive to followers’ self-concept enhancement. The described behaviors are the core elements of idealized influence but not individualized consideration. Therefore, we hypothesize the following:
Hypothesis 1:
In Finland, idealized influence but not individualized consideration leadership behavior will be positively associated with followers’ organizational identification.
Idealized influence/individualized consideration and organizational identification in Russia
We foresee that in Russia TL will also positively affect followers’ organizational identification by effectively enhancing their self-concepts but in contrast to Finland mainly through individualized consideration. In contemporary Russia, employees can be expected to be skeptical about those aspects of TL that attempt to prime followers’ group membership by using various slogans and symbols (e.g., logos, labels, flags) or rituals and ceremonies (e.g., singing company songs), which, as elements of idealized influence, were proposed to enhance followers’ self-concepts (e.g., Kark & Shamir, Reference Kark and Shamir2002; Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998). The skepticism is due to the legacy of the Soviet Union where such slogans and rituals were widespread and used extensively for propaganda purposes. As a result, Russian people today generally feel disenchanted about such symbolic tools.
Interestingly, studying leadership in military units, Shamir et al. (Reference Shamir, Zakay, Breinin and Popper2000) showed that in certain situations followers simply ‘do not buy into’ idealized influence behavior. The authors found that soldiers in military units were not at all receptive (in terms of their social identification) towards leadership behavior from army staff that put an emphasis on shared values. Although in a different context, it shows that sometimes leadership behaviors alienate followers if perceived as being instrumental, hypocritical or untrustworthy (see also Greenbaum, Mawritz, & Piccolo, Reference Greenbaum, Mawritz and Piccolo2015). In Russia, followers may react similarly to the aspects of idealized influence that resemble the ideological tools of the Soviet period. Therefore, we expect that in Russia idealized influence behavior will not be effective in enhancing followers’ self-concepts and in priming their organizational identification.
In contrast, we anticipate that due to Russia's high score in power distance and the traditional preference of Russian people for paternalistic and (socially) proximal leaders (Kets de Vries, Reference Kets de Vries2001) individualized consideration behavior will have, in Russia, a positive influence on followers’ self-concepts and organizational identification. Russia has a long tradition of paternalistic relations between leaders and followers in all possible domains (Kets de Vries, Reference Kets de Vries2001; Puffer, Reference Puffer1994). As such, a paternalistic relationship presupposes that leaders provide guidance, protection and care to followers in exchange for their trust and loyalty (Aycan et al., Reference Aycan, Schyns, Sun, Felfe and Saher2013). Until recently, a high degree of paternalism characterized Russian organizations, meaning that subordinates treated their superiors as more than just work-related superiors, but as someone with whom they identify closely and who knows a lot about them because of his/her active involvement in their private lives (Michailova, Reference Michailova2000). Further, Russian employees’ high power distance and uncertainty avoidance orientations (see Table 1 above) complement such paternalism (Elenkov, Reference Elenkov1997). The combination of these factors suggests that Russian followers are likely to treat individual attention from a higher standing and a more powerful leader as a sign of trust and privilege. They are also likely to perceive such leaders as articulating less ambiguous and more clearly defined work-related goals and instructions, and providing a source of continuous and sought-after approval and legitimacy for their work-related activities (Fey, Reference Fey2005). In this way, such TL behavior can be more effective in Russia than in less paternalistic, power inequality and uncertainty avoidant countries, for example, Finland, in enhancing followers’ self-concepts and in facilitating organizational identification.
Interestingly, research shows that especially in environments characterized by high power distance, a follower's identification with the leader, to which, in Russia, the strongly paternalistic nature of the leader-follower relationship is conducive, is likely to transform into his/her identification with the organization. For instance, Shamir et al. (Reference Shamir, Zakay, Breinin and Popper2000) showed that military personnel's identification with their immediate leaders relates positively to their identification with their military units. In such environments, a follower is more likely to equate his/her paternalistic leader (and the follower's identification with him/her) with the entire organization (and the follower's identification with it), pairing them cognitively, affectively and behaviorally (see Sluss & Ashforth, Reference Sluss and Ashforth2008). We expect the same process to occur in Russian organizations whereby followers’ identifications with their proximal leaders, facilitated by the leaders’ individualized consideration behavior, can translate into a higher-level identification with followers’ organizations. Therefore, based on all of the above, we hypothesize the following:
Hypothesis 2:
In Russia, individualized consideration but not idealized influence leadership behavior will be positively associated with followers’ organizational identification.
Mediating effects of role ambiguity in Finland and Russia
As mentioned above, leadership research has gradually acknowledged its leader-centric bias and advocated considering the role of followers in defining what behaviors or influence attempts qualify as leadership (Hollander, Reference Hollander1992; Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014). Because of natural diversity among followers due to individual and cultural differences, different followers are likely to perceive leadership differently. The literature assigns a central role defining these perceptions to the follower's role, which refers to the followers’ perception of their position and responsibilities in relation to leaders (see Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014). Given that role perceptions directly influence role behaviors (see Katz & Kahn, Reference Katz and Kahn1978), these factors and characteristics are important drivers of how followers perceive leadership and what outcomes this achieves. Moreover, how followers position themselves in relation to leaders, their role perceptions, is likely to depend on their self-concept. Thus, in this paper, we focus on the role of role ambiguity, as a manifestation of followers’ role perception, in mediating the relationships between different TL behaviors and followers’ organizational identification.
We put forward the following arguments in support of the proposed mediation. First, Kahn et al. (Reference Kahn, Wolfe, Quinn, Snoek and Rosenthal1964) has theorized that the experience of role ambiguity arises from expectations and communications as they are dispatched from a role sender. Although in organizations role goals and expectations can emanate also from other constituents, for example co-workers, the role of immediate, proximal leaders in this process is likely to be central (Griffin, Reference Griffin1981; Griffin, Bateman, Wayne, & Head, Reference Griffin, Bateman, Wayne and Head1987; O'Driscoll & Beehr, Reference O'Driscoll and Beehr1994). The degree of role ambiguity is likely to depend on how followers perceive their role to be in relation to their leaders.
Second, prior research suggests that transformational leaders are able to reframe stressful job tasks as opportunities for growth rather than mere sources of stress (Sosik & Godshalk, Reference Sosik and Godshalk2000). They also enhance their followers’ task-related self-efficacy beliefs and social support perceptions (Shamir et al., Reference Shamir, House and Arthur1993; Lyons & Schneider, Reference Lyons and Schneider2009), create low stress and emotional exhaustion but a high-performance workplace environment (Avolio, Zhu, Koh, & Puja, Reference Avolio, Zhu, Koh and Puja2004), and generate positive emotions and lower threat appraisals among their followers (Bono, Foldes, Vinson, & Muros, Reference Bono, Foldes, Vinson and Muros2007; Lyons & Schneider, Reference Lyons and Schneider2009). Thus, it seems plausible to suggest that in this way transformational leaders are likely to decrease their followers’ role ambiguity.
Finally, prior research showed that role ambiguity as an important role characteristic predicts organizational identification (Greene, Reference Greene1978; Wan-Huggins, Riordan, & Griffeth, Reference Wan-Huggins, Riordan and Griffeth1998). When an employee cannot fulfill the required role of a job because the information concerning the leader's expectations of the role is lacking or ambiguous, he or she is likely to experience increased tension, anxiety and stress (House & Rizzo, Reference House and Rizzo1972; Rizzo, House, & Lirtzman, Reference Rizzo, House and Lirtzman1970). High tension and stress may be perceived by the employee as a negative organizational experience and thus may have a detrimental effect on his/her identification with this particular organization (see also Kreiner & Ashforth, Reference Kreiner and Ashforth2004; Thoresen, Kaplan, Barsky, & de Chermont, Reference Thoresen, Kaplan, Barsky and de Chermont2003).
Therefore, we expect the followers’ role ambiguity to mediate the relationships between TL behaviors and followers’ organizational identification. However, we also foresee that the mediation is likely to be different in Russia and Finland. In what follows, we hypothesize these mediation effects.
Mediating effect of role ambiguity in Finland
We argue that, in Finland, role ambiguity will mediate the relationship between idealized influence, but not individualized consideration, and followers’ organizational identification. On the one hand, followers in Finnish organizations are more likely to be able to deduce what their leaders expect from them from organizational culture, goals, values and objectives, as well as from their colleagues’ behavior. Leaders’ exemplary behavior as an element of their idealized influence behavior can further clarify these expectations by outlining the boundaries within which the followers are to perform their tasks and duties. Agarwal et al. (Reference Agarwal, DeCarlo and Vyas1999) found that such behaviors decreased role stress and role ambiguity in the US with its high individualism and low power distance, but not in India with its high collectivistic and power distance orientations.
On the other hand, individualized consideration is not likely to decrease followers’ role ambiguity in Finland because in the context of low power distance and high individualism proximal leaders are likely to be perceived as but one of the constituents of the followers’ work and task environments. Moreover, as Peterson et al. (Reference Peterson, Smith, Akande, Ayestaran, Bochner, Callan and Viedge1995) suggest, in cultures with low power distance and high individualism, employees perceive potential problems with role ambiguity as emanating from situations or events that their direct leaders cannot understand adequately themselves. To resolve these problems, employees in such cultures engage in self-reliant actions and turn to other sources (e.g., prior experience, corporate culture, norms and values) and parties (e.g., coworkers, team members or subordinates) for consideration (Peterson, Smith, Bond, & Misumi, Reference Peterson, Smith, Bond and Misumi1990; ibid.). Such self-reliant actions provide employees with a possibility to clarify their role expectations by consulting multiple sources of meaning. Hence, for Finland, we hypothesize the following:
Hypothesis 3:
In Finland, followers’ role ambiguity will mediate the relationship between idealized influence, but not individualized consideration leadership behavior, and followers’ organizational identification.
Mediating effect of role ambiguity in Russia
In Russia, we foresee that role ambiguity will mediate the relationship between individualized consideration, but not idealized influence, and followers’ organizational identification. On the one hand, building on our prior arguments, idealized influence is likely to be less effective in Russia than in Finland. The idealized influence practices, such as communicating corporate vision, inculcating corporate values, motivating employees towards achieving corporate goals and objectives, have been developed and promoted in the West. The applicability and transferability of these practices (at least ‘as it is’) into Russian organizations have been questioned (Elenkov, Reference Elenkov1998; Luthans, Peterson, & Ibrayeva, Reference Luthans, Peterson and Ibrayeva1998; May, Puffer, & McCarthy, Reference May, Puffer and McCarthy2005). Furthermore, contemporary Russian organizations still seem to be oscillating between traditional, locally bred and Western-originated managerial practices, values and approaches (see May et al., Reference May, Puffer and McCarthy2005; McCarthy, Puffer, May, Ledgerwood, & Stewart Jr., Reference McCarthy, Puffer, May, Ledgerwood and Stewart2008). For instance, empowering and engaging Russian employees using Western-originated practices was found to be difficult (Elenkov, Reference Elenkov1998; Michailova, Reference Michailova2002). These results hint at the potential non-susceptibility of Russian employees to Western-origin managerial ideas that form the backbone of idealized influence leadership behavior. It may be hard for Russian employees to deduce clear demands and expectations for their role in the organization from organizational culture, values, objectives, HRM and managerial practices. They are more accustomed to their proximal leaders directly communicating these expectations to them.
On the other hand, to feel less uncertain and more secure within an organization, Russian employees need to feel support, consideration and trust from their immediate leaders, which individualized consideration leadership might help to achieve. Russian employees have traditionally shown preference for a more contact-intensive, (socially) proximal and controlling type of leadership over a more (socially) distant and delegating one (Fey, Adaeva, & Vitkovskaia, Reference Fey, Adaeva and Vitkovskaia2001; Kets de Vries, Reference Kets de Vries2001). It concurs with prior studies that found employees with high power distance orientations preferring direct and close supervision and being less likely to argue with managerial decisions, with which they do not agree (Bochner & Hesketh, Reference Bochner and Hesketh1994). Peterson et al. (Reference Peterson, Smith, Akande, Ayestaran, Bochner, Callan and Viedge1995) argued that to manage effectively their work roles, employees in high power distance cultures need recourse to an unambiguous source of power. A clearly specified authority decreases role ambiguities in such cultures. In their study across 21 countries, the authors found that power distance is negatively associated with role ambiguity. Another study showed that when leaders’ protection and presence are felt and leaders show their consideration for followers, in high power distance and low individualism contexts (e.g., Russia) role ambiguity decreases (e.g., Agarwal et al., Reference Agarwal, DeCarlo and Vyas1999). Feeling protected and being loyal to leaders decrease the sense of role ambiguity in this type of context. Thus, in Russia, individualized consideration but not idealized influence is likely to be effective in decreasing followers’ role ambiguity, so that role ambiguity will mediate the relationship between individualized consideration and followers’ organizational identification. Hence, we hypothesize the following:
Hypothesis 4:
In Russia, followers’ role ambiguity mediates the relationship between individualized consideration, but not idealized influence leadership behavior, and followers’ organizational identification.
Figure 1 below depicts the overall theoretical model examined in this study.
Figure 1. Theoretical model
METHODS
Participants
The data for the study come from a large-scale comparative project on the influence of leadership and HRM practices in Russia and Finland. Four Finland-based multinational corporations participated in the project. We surveyed white-collar employees in (a) the corporate headquarters in Finland (in construction, metal, power engineering, and food producing industries) and (b) the Russian subsidiaries of three of these companies (in construction, power engineering, and food producing industries). Unfortunately, we were not able to gain access to the Russian subsidiary of the metal industry corporation. The data collection took place in 2013 using an online questionnaire. Altogether, we targeted 483 employees in Finland and 185 employees in Russia. The obtained responses were as follows: 295 employees in 72 teams in Finland (response rate 61%, the average number of employees per team 4.1) and 104 employees in 28 teams in Russia (response rate 56%, the average number of employees per team 3.7). Thus, the total number of respondents included in this study is 399 employees in 100 teams. The average age of respondents was 44 in Finland (s.d. = 10.2) and 36 in Russia (s.d. = 9.5). 66% of Finnish and 47% of Russian respondents were male.
Measures
To measure idealized influence and individualized consideration leadership behaviors, we asked the respondents to evaluate their immediate, proximal leaders, i.e. their team leaders, who represented the organizations’ middle-level management. We used the TL construct developed by Podsakoff and colleagues (Podsakoff, Mackenzie, Moorman, & Fetter, Reference Podsakoff, MacKenzie, Moorman and Fetter1990; Podsakoff, Mackenzie, & Bommer, Reference Podsakoff, MacKenzie and Bommer1996). The items were adopted from a shortened version used in previous research (e.g., Kirkman et al., Reference Kirkman, Chen, Farh, Chen and Lowe2009; MacKenzie, Podsakoff, & Rich, Reference MacKenzie, Podsakoff and Rich2001): three items for idealized influence (Cronbach's alphas: Finland = 0.85 and Russia = 0.87) and two items for individualized consideration (Cronbach's alphas: Finland = 0.81 and Russia = 0.79). For all the items in the study (listed in Appendix I), we used a five-point Likert scale ranging from ‘1' = ‘Strongly disagree’ to ‘5' = ‘Strongly agree’.
We used three best loading items from Rizzo, House, and Lirztman (Reference Rizzo, House and Lirtzman1970) to measure role ambiguity. The items were reverse coded in the analyses (Cronbach's alphas: Russia = 0.84; Finland = 0.87). Organizational identification was measured using three best loading items from Reade (Reference Reade2001) (Cronbach's alphas: Russia = 0.84; Finland = 0.83).
Controls
Prior research has identified followers’ age, gender, and tenure as important variables that can influence followers’ attitudes and perceived leader effectiveness (Riordan, Griffith, & Weatherly, Reference Riordan, Griffith and Weatherly2003; Walumbwa, Wang, Lawler, & Shi, Reference Walumbwa, Wang, Lawler and Shi2004). Moreover, role ambiguity and organizational identification are likely to depend on followers’ hierarchical positions and average working hours. We therefore included followers’ age, gender, tenure (in current organization, in current position, and in having the same supervisor), hierarchical position and average working hours as controls in all our analyses.
Model tests
To check for multicollinearity, we examined VIF values. They ranged from 1.234 to 2.222 in Finland and from 1.087 to 2.066 in Russia thus suggesting no multicollinearity issues. Further, to examine the distinctiveness of our measures for the four constructs (idealized influence, individualized consideration, role ambiguity, and organizational identification) across the two samples, we conducted a CFA using Mplus 7. Since our data are nested within teams, we used hierarchical CFA with “team number” as a clustering variable. We tested the fit to the data of the expected four-factor model and compared it with two competing models (a one factor model where all constructs loaded onto one common factor and a three factor model where the two leadership behaviors were combined into one construct). Table 2 shows that the hypothesized four-factor model provided the best fit to the data in both samples.
Table 2. CFA results (with team number as a clustering variable)
Cross-cultural measurement invariance
The questionnaire was originally developed in English. Later, it was translated and back-translated into Finnish and Russian in line with the established cross-cultural translation procedures (Brislin, Reference Brislin, Attman, Rapoport and Wohlwill1980). To determine measurement invariance across the two samples, we followed several steps as outlined in Liao, Sun, and Thomas (Reference Liao, Sun, Thomas, Sanders, Cogin and Bainbridge2014). Because the power and precision of chi-square statistics are sensitive to sample size (Meade & Bauer, Reference Meade and Bauer2007), we adjusted the two samples to be relatively equal in size. To do that, for the analysis with Mplus 7, we randomly selected the same number of teams from the Finnish sample as we had in the Russian sample (i.e., 28 teams from each sample were included in these analyses: 112 employees in Finland and 104 in Russia).
We first tested a configural invariance model to verify that the same factor structure is applicable in both samples and that the used items are adequate measures of each latent variable in both samples. The model fitted the data well (see Williams, Vandenberg, & Edwards, Reference Williams, Vandenberg and Edwards2009): (χ 2(98) = 142.204, p = 0.002; CFI = 0.967; TLI = 0.955; RMSEA = 0.065; SRMR = 0.070). Next, we tested a metric invariance model where the factor loadings were constrained to be equal in both samples. The model yielded an acceptable model fit: (χ 2(90) = 149.824, p = 0.000; CFI = 0.955; TLI = 0.945; RMSEA = 0.078; SRMR = 0.102). The χ 2 difference test, comparing the metric invariance model with the configural invariance model, was insignificant (p = 0.471). This confirms the cross-cultural validity of our constructs (Liao et al., Reference Liao, Sun, Thomas, Sanders, Cogin and Bainbridge2014; Williams et al., Reference Williams, Vandenberg and Edwards2009).
Common method variance bias
Because we collected our data from a single source, we used several measures to ensure that common method variance (CMV) bias is not a serious concern for our analysis. First, several measures were undertaken ex ante. To begin with, we assured all our respondents of the anonymity and confidentiality of the survey. Further, we piloted all three versions of the instrument, i.e., the original one in English and the two translated ones in Finnish and Russian, on at least five respondents each to identify any ambiguous or unfamiliar expressions and ensure that the items are formulated concisely and understandably (Podsakoff, Mackenzie, Lee, & Podsakoff, Reference Podsakoff, MacKenzie, Lee and Podsakoff2003). In addition, following the recommendation in Chang, van Witteloostuijn, and Eden (Reference Chang, van Witteloostuijn and Eden2010), we scattered the measures used in the study throughout the questionnaire, which included many other questions and took on average around 20 minutes for the respondents to fill in. We also randomized the order of items in each construct to make it difficult for the respondents to create cognitively the correlation needed to produce a CMV-biased pattern of responses.
Second, we used several ex post statistical analyses. First, the one-factor models in Table 2 are equivalent to Harman's single-factor tests. As shown in Table 2, the models yielded bad fits to the data. However, because Harman's single-factor test has been criticized (Chang et al., Reference Chang, van Witteloostuijn and Eden2010), we also followed the single-method-factor procedure (see Podsakoff et al., Reference Podsakoff, MacKenzie, Lee and Podsakoff2003) suitable for situations when the precise source of CMV bias cannot be identified. Thus, for each distinct sample we examined and compared two models: (1) the measurement model (Model 1) and (2) the measurement model with an additional common method factor (Model 2). The χ 2 difference test between the two was insignificant in both samples (p = 0.459 in the Finnish sample and p = 0.345 in the Russian sample). These results indicate that, although in both samples Model 2 had a slightly better fit than Model 1, the differences were not significant. Hence, we conclude that CMV bias is not a serious problem in our analyses.
RESULTS
Table 3 below presents the correlation matrix and Cronbach's alphas of the constructs included in the analyses. All Cronbach's alphas exceed the commonly used threshold of 0.7. To ensure convergent validity, we also examined the item-to-item correlation table, which indicated that in both samples the items correlated highest with other items from the same constructs (see Shook, Ketchen, Hult, & Kacmar, Reference Shook, Ketchen, Hult and Kacmar2004).
Table 3. Correlation table
Notes: Russia N = 104, Finland N = 295; Gender: male = 1, female = 2; Job role/function: supervisor = 1, non-supervisor = 2
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
To test our hypotheses, we used linear mixed modeling (LMM) in SPSS. The choice was determined by the nature of our data, which is nested in teams. Hence, we used ‘team number’ as a blocking variable in our analyses. Using LMM, we were able to account for the dependence of our observations on team membership and, in this way, improve the precision of our estimates, ensuring that our results of the fixed effects of the variables in focus are generalizable across all teams both in the Russian and Finnish samples (see LeBreton & Senter, Reference LeBreton and Senter2008). We examined the intra-class correlations (ICC) for the four constructs. In the Finnish sample, the ICC1 values were between 0.33 and 0.42, and the ICC2 values between 0.67 and 0.75. In the Russian sample, the ICC1 values were between 0.37 and 0.47, and the ICC2 values between 0.68 and 0.77. These values indicate significant effects of group membership in the case of teams and a high reliability of within-team ratings (ibid.).
We used hierarchical LMM to test Hypotheses 1 and 2 (see Table 4 below). In Step 1, we entered the control variables. Then, in Step 2, we added the two leadership behaviors: idealized influence and individualized consideration. The results supported Hypothesis 1, so that in Finland only idealized influence had a positive association with followers’ organizational identification (idealized influence: t = 3.70, p < 0.001; individualized consideration: t = 0.44, n.s.). The two leadership behaviors explained 5% of the variance in followers’ organizational identification in Finland. Hypothesis 2, which posited that in Russia only individualized consideration will have a positive association with followers’ organizational identification, was partially supported. Both leadership behaviors turned out to be significantly associated with the dependent variable (idealized influence: t = 3.52, p < 0.01; individualized consideration: t = 2.55, p < 0.05). The two explained 13% of the variance in followers’ organizational identification in Russia. In both countries, the leadership behaviors explained a relatively modest variance in followers’ organizational identification, which might indicate that there are other important factors, such as HR practices, corporate strategy, relationship with peers, etc. influencing identification. Expectedly, the leadership behaviors play a more marginal role in facilitating followers’ identification in Finland where the importance of proximal leadership is likely to be lower as compared to Russia.
Table 4. Results of LMM analyses (fixed effects; ‘team number’ is used as a blocking variable)
Notes: t values are reported; p values in parentheses; Russia N = 104, Finland N = 295.
***p < 0.001, **p < 0.01, *p < 0.05
To test Hypotheses 3 and 4, we followed the procedure recommended by Baron and Kenny (Reference Baron and Kenny1986). It posits that three regressions are needed to test for a mediating effect. First, a dependent variable should be regressed on an independent variable. Second, a potential mediator should be regressed on an independent variable. Finally, a dependent variable should be regressed on both an independent variable and a potential mediator. A mediation effect exists if the first two regressions are significant and in the third regression the path between a mediator and a dependent variable is significant and the path between an independent and a dependent variable is not significant (i.e., full mediation) or is weaker than in the first regression (i.e., partial mediation).
Table 4 shows that in Finland only idealized influence is significantly related to followers’ organizational identification (idealized influence: t = 3.70, p < 0.001; individualized consideration: t = 0.44; n.s.). It also shows that idealized influence is significantly related to role ambiguity (t = −4.15, p < 0.001). Although in Finland individualized consideration is also significantly related to role ambiguity (t = −2.12, p < 0.05), it does not fulfill the first condition as stipulated by Baron and Kenny (Reference Baron and Kenny1986). Therefore, we tested the Finnish sample only for a mediation effect of role ambiguity in the relationship between idealized influence and followers’ organizational identification. Substituting unstandardized beta coefficients into the Sobel (Reference Sobel and Leinhart1982) test calculator (available at http://quantpsy.org/sobel/sobel.htm), we found the mediation effect of role ambiguity to be significant (z = 3.37, p < 0.001). Because the relationship between idealized influence and followers’ organizational identification remained significant after role ambiguity had been introduced as a mediator (t = 2.43, p < 0.05), we conclude that role ambiguity partially mediates the relationship.
In Russia, although both leadership behaviors were significantly related to followers’ organizational identification (see Table 4; idealized influence: t = 3.52, p < 0.01; individualized consideration: t = 2.55, p < 0.05), only individualized consideration was significantly related to role ambiguity (see Table 4; t = −2.39, p < 0.05). Hence, in the Russian sample, we tested the relationship between individualized consideration and followers’ organizational identification for a possible mediation effect. Again, using the same online calculator, we found the mediation effect of role ambiguity to be significant (z = 2.10, p < 0.05). Because after the inclusion of role ambiguity the relationship between individualized consideration and followers’ organizational identification changed into an insignificant one (see Table 4; t = 1.70, n.s), we conclude that role ambiguity fully mediates the relationship. Therefore, our results supported both Hypothesis 3 and Hypothesis 4.
DISCUSSION
Theoretical Advances
The article makes three contributions to leadership research. The first one reflects the statement by Yukl (Reference Yukl1998: 328) that ‘a variety of different influence processes may be involved in transformational leadership, and different transformational behaviors may involve different influence processes… [and] research on these processes is needed to gain better understanding of transformational leadership’. In this article, we concurred with van Knippenberg and Sitkin (Reference van Knippenberg and Sitkin2013) on the need to disaggregate TL leadership behaviors and examine their individual effects. We also argued in favor of adopting a more follower-centric perspective on leadership (Fairhurst & Uhl-Bien, Reference Fairhurst and Uhl-Bien2012; Hollander, Reference Hollander1992; Shamir, Reference Shamir, Uhl-Bien and Ospina2012; Uhl-Bien et al., Reference Uhl-Bien, Riggio, Lowe and Carsten2014). Consequently, we examined the effects of two TL's leadership behaviors (van Knippenberg & Sitkin, Reference van Knippenberg and Sitkin2013; Wang & Howell, Reference Wang and Howell2010; Wu et al., Reference Wu, Tsui and Kinicki2010) and explicated how these behaviors exert their differential influences (as perceived by followers) on followers’ organizational identification. We find that indeed the two behaviors influence followers differently, so that the influence of idealized influence is different from that of individualized consideration. Moreover, the influence of the individual behaviors differs across different cultural contexts. As our study shows, followers with different cultural backgrounds perceive the influence of individualized consideration differently. This discussion then links to our second contribution.
Our second contribution relates to what Spreitzer, Perttula, and Xin (Reference Spreitzer, Perttula and Xin2005: 207) formulated as follows: ‘We know little about the extent to which transformational leadership behaviors are effective across those with different cultural values’. To add to our understanding of the nature of TL influences across different cultural contexts (see also Kirkman et al., Reference Kirkman, Chen, Farh, Chen and Lowe2009), we examined the cross-cultural variation in TL behaviors’ effects on followers’ organizational identification across the two distinct cultural contexts of Russia and Finland. Our analysis confirms that different leadership behaviors have different effects in different cultural contexts (e.g., House et al., Reference House, Hanges, Javidan, Dorfman and Gupta2004; Paris et al., Reference Paris, Howell, Dorfman and Hanges2009). We find that whereas in Russia both idealized influence and individualized consideration leadership behaviors facilitate followers’ organizational identification, in Finland, only the idealized influence leadership behavior does. However, it is to be acknowledged that although statistically significant, the substantive impacts (so called effect sizes measured with Cohen's f 2, see Table 4) of the leadership behaviors was rather modest both in Russia and even more so in Finland. It suggests that there are other factors, e.g., HR practices or organizational climate, which affect followers’ organizational identification and need to be considered in future research.
With this in mind, our findings seem to provide some counterintuitive evidence (at least in relation to prior research) to the claim that to be effective leaders need to match the values and identities they stress and promote with the cultural-cognitive structures held by their followers (Lord, Brown, & Freiberg, Reference Lord, Brown and Freiberg1999; Shamir et al., Reference Shamir, House and Arthur1993). Based on this claim, prior research has consistently purported the idea that followers with more collectivistic orientations are more likely than their counterparts with more individualistic orientations to respond to group-oriented leadership behaviors that presumably prime social (organizational) identification (Bass, Reference Bass1997; Jung et al., Reference Jung, Yammarino and Lee2009; Pillai & Meindl, Reference Pillai and Meindl1998). Our analysis indicates that followers from a cultural group with more collectivistic orientations, i.e., Russia, can develop their organizational identification based on person-oriented and socially proximal leadership behaviors, such as individualized consideration, too. Whereas Finnish followers, who according to Hofstede (Reference Hofstede2001) are less collectivistic-oriented, proved to be more susceptible to group-oriented and socially distant leadership behaviors, i.e., idealized influence. In this way, our analysis provides a more nuanced understanding of the two TL behaviors in two distinct cultural contexts, which allows us to illuminate the cross-cultural variation of these behaviors’ effects on followers’ organizational identification. Arguably, that's something that is difficult or almost impossible to do by employing the higher-order construct of TL, as per most of the prior research (cf. van Knippenberg & Sitkin, Reference van Knippenberg and Sitkin2013; Wang & Howell, Reference Wang and Howell2010; Wu et al., Reference Wu, Tsui and Kinicki2010). As such, aggregating all TL behaviors together leaves few possibilities to pinpoint which specific behaviors are more effective in what cultural contexts and in what ways.
Our third contribution lies in elucidating the role of a particular self-concept-based mechanism in transforming leadership behaviors into crucial organizational outcomes (Kark & Shamir, Reference Kark and Shamir2002; Shamir et al., Reference Shamir, House and Arthur1993; Shamir et al., Reference Shamir, Zakay, Breinin and Popper1998, Reference Shamir, Zakay, Breinin and Popper2000). To do that we examined the role of followers’ role ambiguity – as a manifestation of the followers’ perceived role in the relationship with their proximal leaders – in mediating the relationship between leadership behaviors and organizational identification in Russia and Finland. Our analysis shows that in different cultures different leadership approaches effectively strengthen followers’ self-concepts and hence reduce their role ambiguity. It appears that in Finland followers tend to rely on organizational structures, processes, systems, goals and objectives to decipher and infer their role expectations. Whereas in Russia followers’ perceived role ambiguity tends to decrease when these followers feel secure, supported and cared for by their proximal leaders.
Moreover, our analysis also illuminates some notable differences in how role ambiguity mediates the relationships between TL behaviors and followers’ organizational identification in Russia and Finland. In Russia, role ambiguity fully mediates the relationship between individualized consideration and followers’ organizational identification, whereas in Finland it partially mediates the relationship between idealized influence and followers’ organizational identification. Our analysis thus indicates that in Russia the identification of followers with an organization seems to center on the figure of a direct leader. It suggests that the widely propagated, in the West, means of facilitating employees’ organizational identification such as communicating an exciting vision of the future, inculcating corporate values in employees’ psyche, and offering inspiring goals and objectives might not be the most effective ones in Russia. Yet, other, more person-oriented and socially proximal behaviors, such as individualized consideration leadership, if they come from a respected and hierarchy-based powerful leader, can be efficient and effective. In this case, followers’ organizational identification seems to be primed precisely through leaders’ personalities and personal attributes.
Moreover, when addressing the question of why and how employees identify with their organizations, it seems that there are different motives as well as means for doing so in different cultures. Whereas generally employees tend to identify because they strive to satisfy a number of individual needs, such as safety, affiliation and uncertainty reduction, and to create a sense of order and meaning in the world around them (see Hogg & Terry, Reference Hogg and Terry2000; Pratt, Reference Pratt, Whetten and Godfrey1998), these actively sought-after elements seem to have different referents in different cultures. We find that in Finland these elements tend to be associated more with group-level organizational structures, meanings, values, missions, and so on. Yet, in the more paternalistic and less egalitarian culture of Russia, employees perceive safety, uncertainty reduction, sense of meaning and order, and so on as embedded more strongly in good relationships with their direct leaders. These relationships then facilitate the employees’ organizational identification. In short, our analysis shows that the widely circulating adage that ‘people quit or stay loyal to bosses, not organizations’ appears to be more applicable to Russia than Finland.
Limitations and Future Research Recommendations
The analysis has several limitations. First, the study uses single-source data. To minimize potential risks of CMV bias, we have undertaken several procedural and statistical measures. Based on the latter, we can state that CMV bias does not seem to affect the quality of our results. Second, we have adopted shortened versions for some of our measures and, most importantly, for our measures of leadership behaviors. We did so for two reasons. The first is that the shortened measures of TL behaviors have been already used in the literature (e.g., Kirkman et al., Reference Kirkman, Chen, Farh, Chen and Lowe2009). The second reason has to do with the difficulty of collecting data in Russian organizations, as noted in the literature (see Michailova & Liuhto, Reference Michailova and Liuhto2001). To ensure the respondents’ participation in the survey we had to optimize the length of the measurement instrument. However, our measures might be one of the reasons why the substantive impacts, i.e. the effect sizes, of our models was rather modest both in Russia and Finland. If possible, future research could verify our results using the full-length measures of TL behaviors. Third and relatedly, another limitation of our paper is the large amount of unexplained variance (the effect size of the overall model varied from 0.18 in Finland to 0.24 in Russia). It points towards the need to go beyond static cultural explanations, as we did in this paper where we argued for the differences in the impacts of the two leadership behaviors largely based on Hofstede's cultural dimensions, and incorporate other variables and explanations that may be relevant for predicting followers’ role ambiguity and organizational identification in Russia and Finland. Future research could examine among other factors, for instance, HR practices, organizational climate, psychological safety and/or the quality of relationships with peers.
Fourth, because of our interest in TL behaviors within multinational companies operating across different cultural contexts, we deliberately focused on comparing the effects of TL behaviors between the companies’ headquarters in Finland and their subsidiaries in Russia. It could be that the effects found in the Russian subsidiaries differ from those in domestic Russian organizations. Future research needs to identify these possible differences. Fifth, our construct of ‘organizational identification’ may have been somewhat ambiguous in the case of the Russian respondents as to whether reference was made to the local subsidiary organization or perhaps also more generally to the MNC in question. Indeed, extant research suggest that identification with the subsidiary may be nested in the identification with the MNC (Smale et al., Reference Smale, Björkman, Ehrnrooth, John, Mäkelä and Sumelius2015), in particular when the latter is more salient than the local subsidiary. This may well have been the case for some of our Russian employees. While our results should be interpreted with this in mind, we also note that our main interest was to examine the relationship between TL behaviors and organizational identification regardless of the specific foci of the latter. Finally, as in any cross-cultural research, we cannot fully exclude the possibility that our respondents in Finland and Russia comprehended the questions in the survey differently. To minimize this possibility, we have performed a number of statistical tests to establish the measurement equivalence between the two samples.
CONCLUSION
The question concerning the transferability of leadership behaviors and practices across different cultural contexts continues to puzzle the minds of researchers and practitioners alike. In this paper, we examined the cross-cultural variation of the effects of two transformational leadership behaviors, namely idealized influence and individualized consideration, on followers’ organizational identification in the two culturally distinct contexts of Finland and Russia. Our analysis strongly suggests that due to the countries’ cultural and socioeconomic differences, in order to prime followers’ organizational identification, leaders are better off employing very different leadership behaviors in the two countries. It points toward the importance of taking the follower-centric perspective on leadership seriously as well as the cultural contingency of leadership and its effectiveness in general.
APPENDIX I
The List of Items Used in the Study
Role ambiguity (adopted from Rizzo et al., Reference Rizzo, House and Lirtzman1970)
• I know exactly what is expected of me in my job
• I know what my job responsibilities are
• I have clear, planned goals and objectives for my job
Identification with organization (adopted from Reade, Reference Reade2001)
• My values and the values of this organization are the same
• I share the goals of this organization
• What this organization stands for is important to me
Transformational leadership behaviors (based on Podsakoff et al., Reference Podsakoff, MacKenzie, Moorman and Fetter1990, Reference Podsakoff, MacKenzie and Bommer1996; MacKanzie et al., Reference MacKenzie, Podsakoff and Rich2001; Kirkman et al., Reference Kirkman, Chen, Farh, Chen and Lowe2009)
Idealized influence leadership behavior
• (My team leader) inspires others with his/her plans for the future
• (My team leader) provides an appropriate role model to follow
• (My team leader) develops a team attitude and spirit among employees
Individualized consideration leadership behavior
• (My team leader) shows respect for my personal feelings
• (My team leader) considers my personal feelings before acting
