Cross-Cultural Variation in Political Leadership Styles

Initial research in the area of behavioural leadership theory aimed to group behaviours under a common denominator (Likert 1961; Stogdill & Shartle, 1955). This led to the emergence of two major groups that characterised two distinct styles. An array of names has been used to label these groups, but there is currently common agreement that one of the styles contains task-oriented properties while the other contains person-oriented properties.

Research explored which of the two styles is superior (i.e. research by Bales, 1950; Lewin, Lippitt, & White, 1939; Likert, 1961; Stogdill & Shartle, 1955), with results proving largely fruitless (Gastil, 1994), as most scholars agree that a combination of the two provides the most robust explanation of leadership (Blake & Mouton, 1978; Sorrentino & Field, 1986). Burns (1978), however, refused to agree that a superior style is absent and looked to describe leaders who have emerged and ensured effectiveness in times of crisis. Such leaders produced positive results by instigating change and they were seen as transformational. Moreover, transformational leaders have been described as trustworthy, honest, motivational and encouraging (Den Hartog, House, Hanges, et al., 1999). Avolio and Bass (1988) and Bass and Avolio (1990) presented a more elaborate transformational leadership theory—the Full Range Model—wherein transformational leadership style is contrasted with transactional and passive/avoidant styles. Like Burns (1978), Avolio and Bass (1988) described transformational leaders as: (a) those who raise subordinates’ awareness of the importance of achieving commonly set goals; (b) those who encourage subordinates to disregard what interests them for the sake of the common goal; and (c) those who develop their subordinates’ needs to a higher level, wherein they feel autonomous yet affiliated with the group. Bass (1985) viewed the compared transactional leadership as descriptive of those leaders who recognise what subordinates want to get from their job, help subordinates get what they want (if their performance allows) and exchange rewards for effort. Additionally, in this model the ‘passive/avoidant’ style signified a ‘negative’ leadership style, which was used as a baseline anchor point (Avolio, 1999).

In general, transformational theory is considered to have had a large impact upon leadership as a scientific domain. According to Antonakis (2012), it helped leadership researchers by providing them with direction at a time when the field suffered much pessimism. It presented researchers with a set of behaviours that were considered superior for inducing effectiveness, compared to the set of behaviours measured in the so-called two-factor theories (i.e. task- and person-oriented behaviours). Moreover, the Full Range Model benefits from a specially designed test (the Multifactor Leadership Questionnaire), currently used to measure the frequency of behaviours associated with the three styles.

Bass (1997) insisted on the universal relevance of transformational leadership and studies have cited that it is the most desired (Den Hartog et al., 1999) and the most effective of styles (Barling, Weber, & Kelloway, 1996; Derue, Nahrgang, Wellman, & Humphrey, 2011; Geyer & Steyrer, 1998; Judge & Bono, 2000; Judge & Piccolo, 2004; Lowe, Kroeck, & Sivasubramaniam, 1996; Rowold & Heinitz, 2007). Despite support for the stability of transformational leadership across settings, some research suggests that in some respects it varies as a function of the environment. Even though contextual factors were not seen to affect the effectiveness and the desirability of the style they were seen to affect its pervasiveness. According to findings the occurrence of the style is contingent on organisational and cultural factors (Bruch & Walter, 2007; Jung, Bass, & Sosik, 1995; Kuchinke, 1999; Leong & Fischer, 2011; Ling, Simsek, Lubatkin, & Veiga, 2008; Lowe et al., 1996; Ozorovskaja, Voordijk, & Wilderom, 2007; Pawar & Eastman, 1997). Organisational structure, type, hierarchy level and size have been thought to affect the frequency of its use. Transformational leaders have been seen to emerge mainly in the upper levels (Bruch & Walter, 2007) of the hierarchy in organic (Bass, 1985) and public (Lowe et al., 1996) organisations of small sizes (Ling et al., 2008). Similarly, cultures experiencing crisis (Conger, 1999) are reported as providing better contexts for the emergence of transformational leaders. Additional culture-contingent variables, such as political structure (Alas, Tafel, & Tuulik, 2007; Ozorovskaja et al., 2007) and differential values (Jung et al., 1995; Leong & Fischer, 2011), have also played a role in determining pervasiveness. For instance, transitional democracies and collectivist/mastery-oriented cultures (Jung et al., 1995; Leong & Fischer, 2011), show more transformational behaviours in their leaders.

Nonetheless, despite the widespread exploration of influences on transformational leadership practices, gaps in the literature are present. There is a notable absence of studies of political leaders even though transformational leadership theory is rooted in Burns’s (1978) study of political leaders. The difficulty in approaching politicians could explain this trend. However, it does not fully justify the apparent near total absence of studies quantitatively exploring behaviours displayed by political leaders. Other gaps in the literature are also present. They concern the lack of studies considering all three styles within the Full Range Model. Trends relating to the pervasiveness of transactional leadership and the effects of context on it are not as clear, because, often, studies exploring transformational leadership have not contrasted its effects with those of transactional and passive/avoidant styles. The exploration of culture as a context variable also seems to have been limited and when investigated an interest in Western cultures is predominant. While helpful, only few studies explore transformational leadership (Den Hartog et al., 1999) in developing and transitional democracies, and their regional clusters lack full diversity. For instance, within their Eastern European post-communist cluster, Den Hartog et al. (1999) included only Central European cultures such as those of Poland, Hungary and the Czech Republic. These countries share some historical similarity with other post-communist cultures—like those of Bulgaria or Lithuania—which allows for some generalisations. However, one must generalise with caution because Brodbeck et al. (2000) noted differences between historically similar cultures, which did not always share close geographical proximity.

Due to the aforementioned gaps, the transformational theory literature could benefit from studies exploring the presence and dynamics of the Full Range Model styles (i.e. transformational, transactional and passive/avoidant; Bass & Avolio, 1995) in the area of political leadership. The inclusion of political leader MLQ (Multifactor Leadership Questionnaire) self-ratings from direct leader assessment, the assessment of the transformational, transactional and passive avoidant leadership frequency data and the examination of the effects of variables such as culture might allow for more complete knowledge of political leadership style. The research question therefore explores the variance in the pervasiveness of the Full Range leadership style behaviours across cultures. A comparison between UK (developed Western democracy) and Bulgaria (developing post-communist South-Eastern European democracy) is made. In general, the comparison of Bulgaria to an established Western democracy could inform the hypothesis formation of future research by underlining the difference in leadership variables in Western and South-Eastern Europe. The comparison could also reveal much about the differences and similarities between cultures with varying historical, political and cultural profiles. Any knowledge of the differences and similarities between Bulgarian and British leaders could possibly be welcomed by structures such as the European Union, where understanding and allowances in communication could aid collaborative work in areas such as immigration. Furthermore, the findings could inform transformational theory by discussing variables which might have influence upon it and by noting the extent to which the Full Range model styles vary across cultural contexts.

The results of past research (Jung et al., 1995) describe leaders in collectivist cultures and cultures in critical conditions (Conger, 1999)—such as Bulgaria—as more transformational in nature. Ozorovskaja et al. (2007) asserted that post-communist leaders—such as those in Bulgaria—practice more transactional leadership behaviours, which as cited, is tied to high power distance (inequality) and authoritarianism. Based on such findings the expected outcome is for Bulgarian political leaders to rate themselves as displaying higher levels of both transformational and transactional leadership behaviours, compared to British leaders. In relation to passive/avoidant leadership we could expect cross-cultural equality, as both cultures are presumed equally likely to avoid any association with salient indicators of ‘bad leadership’, however due to the lack of findings it could be wise to form a question--as opposed to hypothesis—asking whether there are cross-cultural differences in passive/avoidant leadership style behaviours displayed by political leaders.

Method

Analyses and Results

In order to explore the likely variations, political leader frequency scores on the MLQ scales (i.e. five transformational leadership scales, two transactional leadership scales and two passive/avoidant leadership scales) within the two cultural samples were compared. The preferred test for these comparisons was MANCOVA allowing for the exploration of multiple dependent variables and therefore reducing the likelihood of type I error. MANCOVA also addresses the issue of covariate effects, often leading to an increase in within-group error. Aspects such as gender (Eagly, Johannesen-Schmidt, & van Engen, 2003; Eagly & Johnson, 1990) and age (Kobacoff & Stoffey, 2001; Oshagbemi, 2008) have been previously shown to affect the frequency of leadership behaviours displayed by leaders; since these aspects were measured here, controlling for their effects helped when inferences about group variation were drawn. Further variables, such as ‘political ideology’ could also bring on confounding effects. Unfortunately, the large presence of missing data (see Table 1), associated with this variable prevented us from including it in the analysis. In general the inclusion of unreliable variables in MANCOVA is discouraged, as degrees of freedom and therefore power is lost with each addition of a variable. Political ideology should nevertheless be considered in future work.

While exploring assumptions for carrying out MANCOVA, some of the requested analyses showed sizeable correlations between scales belonging to each of the measured styles. The finding posed problems similar to those experienced by others (e.g., Hetland & Sandal, 2003). In order to solve the possible presence of redundant dependent variables, variable integration was undertaken. Combining the frequency scores of the five transformational scales into a single rating variable called the ‘transformational leadership frequency score’ and combining the scores of the transactional and passive/avoidant scales into two more single score variables (i.e. the ‘transactional leadership frequency score’ and the ‘passive/avoidant leadership frequency score’) was something other scholars in the area had done (Hetland & Sandal, 2003). The combination of variables reduced the original nine leadership style frequency variables into three new dependent variables: transformational, transactional and passive/avoidant leadership frequency scores. Before this combination, exploratory factor analysis with Varimax orthogonal rotation was carried out to investigate (independently of prior expectations) the higher order factor structure of the MLQ, in order to confirm the five transformational, the two transactional and the two passive/avoidant scales loaded onto three separate factors. The need for this was further prompted by evidence (Ardichvili & Kuchinke, 2002; Avolio & Bass, 2004; Geyer & Steyrer, 1998; Hetland & Sandal, 2003) suggesting that the transactional leadership scale contingent reward is highly correlated with all transformational leadership scales. Exploring the factor loadings was therefore necessary before variable integration took place.

The emerged factor structure following the exploratory factor analysis is illustrated in Table 2.

As presented in Table 2, the results note that, while the five transformational scales and two passive/avoidant scales loaded onto their respective factors, the transactional leadership scale contingent reward loaded quite highly onto both the transformational leadership factor and the factor containing the remaining transactional leadership scale management by exception–active.

It can, therefore, be argued that contingent reward should be included in the combined transformational leadership frequency score variable, rather than in the combined transactional leadership variable, where it is thought to belong. However, past research (Hetland & Sandal, 2003) has noted otherwise and, despite its high loading onto the transformational leadership factor, contingent reward was still included in the transactional factor during variable combination. This common practice has been supported by the test publishers and by other researchers (Avolio & Bass, 2004; Hetland & Sandal, 2003), who insist contingent reward is applicable to both types of leadership and note its high loading onto two of the three leadership behaviour factors/styles. Based on this, contingent reward was included in the transactional leadership frequency score variable.

Moreover, in terms of the remaining assumptions the data were suitable for a MANCOVA. Outliers were absent and all three dependent variables appeared as normally distributed for each level of the IVs (i.e. all skewness and kurtosis values were lower than the more stringent cut-off value of +/-1) (see Table 3 and 4Table 4 for descriptive statistics and skewness and kurtosis values). Missing values associated with the dependent variables (i.e. the three leadership styles) were not present and as described in Table 1 such were only evident with regard to the demographic variables. Heterogeneity was also absent, with all Levene’s tests suggesting insignificance (see Table 5).

The results of the multivariate analysis of covariance revealed a significant main effect of culture, F(3, 134) = 26.94, p < .001.

As hypothesised the follow up univariate analyses confirmed that Bulgarian and British political leaders varied with regard to their transactional, F(1, 136) = 59.82, p < .001, and unexpectedly with regard to their passive/avoidant style scores, F(1, 136) = 15.18, p < .001. A Bonferroni correction to the alpha level was employed to prevent the type I error usually associated with multiple follow-up analyses. On both occasions, Bulgarians appear to have scored higher (M = 3.13, SD = 0.49; M = 1.28, SD = 0.58, respectively), compared to UK nationals (M = 2.47, SD = 0.57; M = 0.87, SD = 0.6, respectively) (see Table 3 for all means and standard deviations). No significant differences were found with regard to the transformational style where Bulgarian and UK political leaders appeared to score similarly.

In addition, the effect sizes can be considered too. Where significance was reported, effect sizes (signified by eta squared [η²]) were medium to large (Cohen, 1988) (see Table 6 for η²).

Discussion

The results of the analysis of culture differences in leadership styles partially supported the hypotheses. They suggest that, compared to British leaders, Bulgarian leaders scored significantly higher only in terms of transactional and passive/avoidant leadership style behaviours. While the transactional style differences were predicted, the lack of transformational style differences was unexpected. Moreover, the higher instances of passive/avoidant behaviours in Bulgarian leaders is an interesting and noteworthy finding, which tentatively answers the question posed in the introduction.

In order to explain the transactional leadership difference, one could explore the cultural value differences across the cultures studied. Bulgaria, for example, has been found to score highly in terms of power distance (Minkov, 2011), promoting authoritarianism, steep hierarchy, obedience towards those at the top, centralised power and reduced concern for employees in work settings. Some of these aforementioned concepts—such as authoritarianism—are associated with task-oriented behaviours--also related to, and present in, transactional leaders. In this way, one can associate power distance with the elevated levels of transactional behaviours in Bulgaria found in this research.

The provision of explanations for the passive/avoidant style variation is more challenging. It was expected that behaviours detrimental to the completion of a task are equally and universally absent in the leadership arena. Nevertheless, while passive/avoidant behaviours were still negatively associated with leadership in Bulgaria, they were certainly more frequently enacted by Bulgarian, than by British, leaders. No studies have looked at this in Bulgaria, but studies exploring countries with similar historical challenges—like Russia—have generated results similar to those obtained here. Ardichvili and Kuchinke’s (2002) findings presented low scores on the passive/avoidant leadership style scales laissez-faire and management by exception in all of the tested cultures. However, the scores were substantially higher for leaders from Russia, Georgia, Kazakhstan and Kyrgyzstan, compared to leaders from Germany and the USA. Similarly, Puffer (1996)—who studied the leadership styles of Russian leaders—suggested that these leaders display a collectivist attitude characterised by diffusion of responsibility (Mazar & Aggarwal, 2011), and a tendency to delegate decision making due to the wish to avoid the responsibility associated with unforeseen circumstances. A similarity between collectivism-based diffusion of responsibility and passive/avoidant leadership is evident, as both concepts are associated with decision making avoidance, delayed response and failure to interfere when needed. It is therefore likely that a high score on one (e.g. collectivism) could lead to a high score on the other (i.e. passive/avoidant leadership). Bulgaria’s more collectivist nature (Minkov, 2011) could explain its higher experience of inadequate leadership practices. Moreover, after looking at denial and avoidance of threat-related information, Metselaar (2012) described governments with authoritarian experiences—such as governments in Bulgaria—as more likely to tolerate denial and avoidance, compared to stable democratic systems.

Interestingly, even though a difference in the enactment of transformational leadership behaviours in Bulgaria and the UK was expected, such a difference was absent. The findings oppose those who characterise cultures in more critical environments (Conger, 1999), cultures with collectivist values (Jung et al., 1995) and cultures with authoritarian experiences (Eisenstadt, 1968), as displaying higher instances of transformational leadership. In general, charisma—a crucial aspect of transformational leadership—is often treated as an anti-democratic force and its heightened presence in cultures with history of totalitarian ruling was therefore expected. One possible answer to why leaders in Bulgaria and the UK scored similarly in terms of the practiced transformational leadership behaviours and therefore opposed what was hypothesised can be derived from Weber’s (1978) work. According to Weber, charisma can be equally developed in both dictatorial/autocratic and democratic situations. Weber (1978) described charismatic appeal as highly interactive with bureaucratic administration, but also as a concept which had democratic ramifications. Similarly, according to Gerth and Mills (1946), charisma can be ‘the vehicle of man’s freedom in history’; these researchers also suggested that, depending on its routinisation, charisma can exist equally in both democratic and undemocratic settings. Moreover, Bass (1997) agreed that transformational leadership can be both autocratic and directive, and democratic and participative. As noted earlier, he also advocated for the universality of transformational behaviours. Propositions as such could explain why transformational leadership appeared equally in Bulgaria and the UK.