Article body

Leadership is typically defined as the “ability of an individual to influence, motivate, and enable others to contribute toward the effectiveness and success of the organizations in which they are members” (House et al., 2004). There is growing interest in examining leadership in cross-cultural contexts (Robinson & Harvey, 2008; Engelen et al., 2014; Karakitapolu-Aygün et al., 2021; Miao et al., 2018; Smith & Peterson, 2017). The literature has examined a) leader traits that are desirable across cultures (e.g., Miao et al., 2018), b) cross-cultural effectiveness of leadership styles such as transformational (e.g., Engelen et al., 2014), or authoritarian leadership (e.g., Karakitapolu-Aygün et al., 2021), and c) upper echelon leadership (e.g., Smith & Peterson, 2017), among others. We now know that leaders who display attributes or use culturally congruent styles are more likely to do well as leaders (e.g., Miao et al., 2018). Individuals in different but specific cultural settings hold belief systems about what constitutes “good leadership”, which are culturally shared (House et al., 2004). However, the literature has not extensively examined critical leadership behavior: accurate attributions for subordinate performance (Lakshman, 2013). International Business (IB) research has long suggested that isomorphic attributions (the extent to which a person makes accurate attributions about the behavior of a person from another culture) is imperative for cross-national leadership effectiveness (Lakshman et al., 2021; Triandis, 1975). But we still lack an adequate understanding of issues related to the accuracy of attributions that leaders are likely to make for subordinate behavior in cross-national contexts (Dean & Koenig, 2019).

On the other hand, the literature on attribution theories of leadership specifies attribution as the key mediating process through which leaders and subordinates evaluate each other’s behaviors (e.g., Martinko & Mackey, 2019). It suggests that leaders tend to make biased attributions for subordinate behaviors (e.g., van Houwelingen et al., 2021). Thus, while there exists a vast literature on the biased attributions supervisors make for their subordinates’ behaviors (e.g., Martinko & Mackey, 2019), the literature on the accurate attributions of these leaders is non-existent (Forsterling & Morgenstern, 2002), with very few exceptions (e.g., Lakshman, 2013; van Houwelingen et al., 2021). These exceptions have suggested that cognitive complexity or attributional complexity on the part of leaders can increase the accuracy of attributions (e.g., van Houwelingen et al., 2021). However, there has been a lack of answers to the call that the success of leader-subordinate interactions, and thus of leader effectiveness, depends greatly on accurate attributions by leaders (Chen & Van Velsor, 1996).

Based on these theoretical insights, the purpose of our study is to contribute to the literature on leadership in cross-cultural contexts and the literature on attribution theories of leadership. We do so by conducting an empirical examination in five countries to examine the relationship between attributional accuracy and leadership effectiveness. Our conceptual background is the implicit leadership theory approach used by GLOBE researchers (e.g., House et al., 2004). The notion of Culturally endorsed implicit Leadership Theories (CLTs) in the GLOBE study (House et al., 2004) builds on implicit leadership theory, which suggests that people have assumptions and theories about what leader traits and behaviors are likely to be effective. Implicit leadership theories that are culturally endorsed provide a legitimation mechanism in which certain leader traits and behaviors become effective over others (Green, 2017; Watts, Steele, & Den Hartog, 2020).

We examine two characteristics of accurate attributions – the degree of accuracy and the degree to which they are free of racial/cultural bias – which managers make for subordinate behavior in the performance domain. We assess whether or not these characteristics are isomorphic with cultural expectations and thereby constitute desirable leadership. We examine the propositions of our hypothetical model in five countries viz., USA, France, India, Turkey, and Vietnam, each belonging to a different societal cluster, thereby representing diverse cultural contexts. GLOBE researchers classified the USA into the Anglo cluster, France into the Latin Europe cluster, India into the South Asia cluster, and Turkey into the Middle East cluster (House et al., 2004). Vietnam belongs to the Confucian Asia cluster because of a very long period of Chinese influence (Hoang, 2008).

The rest of the paper is organized as follows. We first provide a brief review of the literature on culturally desirable leader traits and highlight the discussion in the IB literature about leadership effectiveness founded on attributional accuracy. We contrast this literature with the dominant approach to leadership theorizing focused on biased attributions of managers. We then provide the implicit leadership theoretical framework for both the attributional process and CLTs (Green, 2017; Robinson & Harvey, 2008) and develop hypotheses of our theoretical model. Subsequently, we describe the method, the results of the analysis, and the contributions of our work in the context of the literature. We conclude with the limitations and practical implications of our research.

Culturally Desirable Leader Traits

The literature on cross-cultural leadership has focused on discussing different leader traits that are culturally desirable (House et al., 2004). GLOBE researchers identified 22 attributes that are universally regarded as positive in terms of their contribution to outstanding leadership (House et al., 2004). Attributes such as trustworthy, excellence oriented, dynamic, encouraging, motivating, communicator, and informed were among those identified as universally desirable for effective leadership. This project also identified 8 attributes that are universally undesirable for effective leadership (e.g., irritable, dictatorial, self-protective, and malevolent). However, the literature has not paid significant attention to the issues related to the accurate attributions of subordinate behaviors by leaders in cross-national contexts. Also, it focuses on identifying culturally desirable leader traits but does not go as far as examining how exactly leaders in any culture actually demonstrate these traits or how exactly the subordinates may perceive them. For instance, what leader attribution demonstrates that (s)he is excellence oriented or malevolent? or what are the consequences in terms of perception by subordinates? Researchers have only started to consider these issues in their study of culturally desirable leadership traits (e.g., Lakshman & Estay, 2016).

On the other hand, the IB literature highlights the importance of accurate attributions. IB researchers appreciate the concept of isomorphic attributions (e.g., Landis & Wasilewski, 1999). It indicates the degree to which attributions made by individuals about others from another culture are accurate (Lakshman et al., 2021). It emphasizes the role of mutual expectations and reciprocal relationships in cross-cultural interactions and therefore is useful in pointing to the importance of attributions and attributional patterns of individuals across cultures (Dean & Koenig, 2019). IB scholars have used this notion to design the cultural assimilator, a training tool aimed at improving attributional accuracy (Lakshman et al., 2021). They consider attributional accuracy as fundamental for having cross-cultural competence and cultural knowledge (Johnson et al., 2006).

Attributional Theories of Leadership

The dominant attributional approach to leadership (Martinko & Mackey, 2019) focuses on biased attributions. Biased attributions can be individual-based, such as gender biases (Dobbins et al., 1983) and biased responses to correct performance problems directed at the employee (Martinko et al., 2007). They can be personalized based, such as self-serving biases (Campbell & Swift, 2006) and biased internal attributions (Martinko et al., 2007), or organizationally based, such as organizationally induced helplessness (Martinko et al., 2007). Because these biases have or result in multiple adverse consequences, these may lead to poor performance spirals (Lindsley et al., 1995), capability traps (Repenning & Sterman, 2002), learned helplessness, leader-subordinate disagreement and conflict (Bagci et al., 2018), loss of trust of workgroup, loss of managers’ credibility, and subordinate dissatisfaction and turnover, all of which makes for very ineffective management of poor performance of employees (Amagoh, 2009), their units, and their organizations, in addition to seriously affecting leadership effectiveness (Lakshman, 2013). This may result in leaders’ eventual loss of credibility, respect, and they may ultimately lose their managerial positions.

A strength of this attributional model is that it views leaders as scientists who seek informational cues explaining causal relationships affecting various aspects of work (Lakshman, 2008). However, it focuses on naïve attributional processes using cognitively simple schemata (Kelley, 1972). For example, Martinko et al. (2007) put forth a two-stage attributional model, in which leaders look for information cues to understand the poor performance of subordinates and then identify appropriate actions to improve performance. Martinko and colleagues (2007) suggested that leaders attribute poor performance of subordinates using three categories of information: distinctiveness, which refers to the uniqueness of the behavior related to the poorly performed task; consistency, which indicates the pattern of behavior over time or across situations; and consensus, which represents the behavior of others and its similarity to the subordinate in question. Based on these attributions, leaders then determine the appropriate course of action, such as punishment or closer supervision (e.g., van Houwelingen et al., 2021).

Recently, researchers have suggested the need to focus on the attribution accuracy of leaders (Grimshaw et al., 2006; Lakshman, 2013), which stems from information processing using complex schemata (Kelley, 1972). Many scholars have noted the importance of accurate attributions by leaders in impacting the actions and behavior of subordinates (e.g., Grimshaw et al., 2006; MacNab & Worthley, 2012; van Hoiwelingen et al., 2020). As leadership is viewed as a result of positive perceptions (Lord & Maher, 1991), effective leaders are more accurate in their analysis of subordinate behaviors (Martinko et al., 2007). Chen and Van Velsor (1996) explicitly argued that leader effectiveness depends primarily on their ability to exercise isomorphic attribution (Lakshman et al., 2021). Importantly, scholars argue for the possibility of accurate attributions of leaders through complex schemata (van Houwelingen et al., 2021). For example, Lakshman (2008) suggested that the foundation of accurate attributions is complex schematic processing such as the use of augmenting and discounting causal schema, interdependence in leader-subordinate contexts, and leader involvement. Sun and Anderson (2012) noted that transformational leaders, who are more complex in their social judgments and attribute complex causes to others’ behaviors, tend not to be biased in their attributions as generally suggested.

In short, it is important to examine the processes related to accurate attributions of leaders, which can be crucial in getting acceptance and commitment to their decisions and achieving performance objectives (Burns, 1978; Estay et al., 2020).

Theoretical Framework and Hypotheses

The CLTs in the GLOBE study suggested that implicit notions of what constitutes good leadership are culturally shared as there is a significant degree of within-culture agreement on effective leadership (e.g., Green, 2017; Watts et al., 2020). We follow this line of thinking by suggesting that attributions made by managers in performance contexts may also form part of culturally shared belief systems about what constitutes effective leadership. We do this based on suggestions in the IB literature that attributions are critical mediating processes through which managers and subordinates interpret and evaluate each other’s behaviors (Chen & Van Velsor, 1996). Accurate attributions of subordinate behavior, based on cognitively complex schematic processing (Kelley, 1972; van Houwelingen et al., 2021), result in subsequent leader interactive behaviors and strategies used to correct performance deficiencies (Lakshman, 2013). Managers making attributions, which are culturally congruent (accurate/unbiased), are more likely to be successful as leaders than those making attributions, which are incongruent (Engelen et al., 2014; Green, 2017). Based on these insights, we develop our theoretical model, as presented in figure 1 below.

Figure 1

Attributional model of Leadership Effectiveness

Attributional model of Leadership Effectiveness

-> See the list of figures

In our theoretical model, there are two exogenous constructs: accuracy of attributions and unbiased attributions. Each of these constructs is independently related to leader interactive behaviors and leadership strategies for correcting performance deficiencies. These two leader behaviors, in turn, are independently related to leadership effectiveness. Additionally, power distance moderates two of the above relationships. Despite our belief in the convergence across these five cultures of the desirability of accurate/unbiased attributions and the resulting behaviors noted above, we argue that power distance may cause variations in how unbiased attributions are viewed in each cultural context (e.g., Engelen et al., 2014).

Accuracy of Attributions

As the dominant approaches in attribution theories of leadership focus on biased attributions, it relies on simple cognitive schemata (Kelley, 1972). By contrast, attributional accuracy, which is defined as the level of congruence with the antecedent discounting and augmenting causal schemata (Lakshman, 2013), is based on Kelley’s (1972) complex cognitive schemata. In suggesting complex schemata, Kelley (1972) explicitly recognizes that individuals interested in making more accurate attributions may want to process information using complex schemata, including the analysis of multiple causes that are consistent with behavior (augmenting schema) and those that are inconsistent with that same behavior (discounting schema). In less complex situations, few causes are consistent with the observed outcome, and thus any single cause possesses augmenting schema. In more complex situations, multiple causes are equally consistent with the outcome. In such a situation, leaders wanting to make more accurate attributions are faced with discounting causes with less import for the observed outcome, thereby focusing on the cause with the most augmenting schema (or least discounting schema).

We draw from implicit leadership theory to suggest that people are likely to consider attributional accuracy as a critical component of good leadership. Accurate attributions relate positively to managerial effectiveness, including aspects such as building trust, building credibility, and enabling subordinates to perform and contribute to overall organizational performance. Accurate attributions are more likely to trigger better resource allocation decisions by managers, more likely to be followed up with functional interactive behaviors, resulting in higher performance (Forsterling & Morgenstern, 2002). Theory suggests that accurate attributions lead to positive interactive behaviors, positive interpretation of subordinate behaviors (Chen & Van Velsor, 1996), and, most importantly, the formulation and implementation of integrative strategies to correct performance deficiencies (Lakshman, 2013). Thus, accurate attributions are likely to be viewed positively by subordinates across all cultures, consistent with their implicit leadership theory. We put forth the following hypotheses.

Hypothesis 1: Across cultures, higher attributional accuracy of managers is positively related to levels of interactive behaviors vis-à-vis subordinates.

Hypothesis 2: Across cultures, higher attributional accuracy of managers is positively related to performance correction strategies used by these managers.

Unbiased Attributions

Attributional accuracy pertains to specific behaviors and performance episodes of subordinates. In contrast, we use the term biased or unbiased attributions to refer to more generally biased or discriminatory (or stereotypical) attributions based on gender, race, or cultural background or lack thereof. Patterns of biased attributions may trigger emotional responses that are not connected to specific interactions (Lakshman, 2013). Subordinates in different cultural contexts are affected both by specific attributions and by general patterns of attributions of their managers (Dean & Koenig, 2019).

The literature suggests that biased attributions are likely to lead to heightened interpersonal conflict (e.g., Bagci et al., 2018) and emotionally disrupt the communication between managers and their subordinates affected by biases (e.g., Lakshman, 2013). Thus, effective managers need to avoid biases stemming from gender, regional, and cultural domains. Biases may trigger attribution-conflict spirals, whereby biased attributions can lead to more severe conflict, which can, in turn, affect attributions made in conflict situations and the subsequent choice of interpretation and communication strategies. Most importantly, gender and cultural biases can lead to a loss of trust, dissatisfaction, and turnover (e.g., Lakshman et al., 2021). Loss of trust could result in the erosion of leadership in the worst case or severely negative leadership effectiveness in the best-case scenario. Moreover, biased and inaccurate attributions may also trigger emotional reactions in subordinates, leading to downward performance-efficacy spirals (Lindsley et al., 1995). These insights lead to the following hypotheses.

Hypothesis 3: Across Cultures, the degree to which attributions are unbiased (gender, racial, cultural) is positively related to interactive behaviors used by managers.

Hypothesis 4: Across cultures, the degree to which attributions are unbiased (gender, racial, cultural) is positively related to performance correction strategies used by managers.

Power Distance

Although several cultural value dimensions may differentially affect the desirability of different leader traits/behaviors (e.g., Li et al., 2021; Miao et al., 2018; Watts et al., 2020), there are many aspects of leadership that are a) universally desirable, and b) more influenced by power distance than other cultural value dimensions (e.g., Karakitapoglu-Aygun et al., 2021; Lakshman et al., 2019). In the context of our study constructs, there may be differences across cultures in terms of the desirability of interactive behaviors and performance correction strategies, stemming from differences in power distance (House et al., 2004). These two constructs –interactive behaviors and performance correction strategies –are part of the typical leader-subordinate communication experience, which has been found to operate similarly across different countries (e.g., Karakitapoglu-Aygun et al., 2021). However, some aspects of these constructs may function universally, while other aspects may be seen or perceived differently as a result of power distance between the leader and the subordinate. In this regard, results from GLOBE suggest that being communicative, trustworthy, encouraging, positive, and a confidence builder are all universally endorsed attributes (Den Hartog et al., 1999), although they may be “seen” in culturally specific ways. Collectivism/Individualism may be more important for differential impact on outcomes such as commitment, and engagement (e.g., Li et al., 2021), and uncertainty avoidance may be more important for outcomes with a long-term horizon such as innovation (e.g., Watts et al., 2020). However, the choice of which cultural dimension to examine would depend on the specific theoretical and work domain, given the lack of a broad theory on how culture affects leadership and the complexity of the leadership construct. We believe that power distance is the most important in this context of attributional processes and related behaviors (e.g., Lakshman et al., 2014). The relationship between unbiased attributions and interactive behaviors may be stronger in low power distance cultures than in high power distance ones. Additionally, the relationship between unbiased attributions and performance correction strategies may be stronger in high power distance cultures than in low, simply because performance correction strategies may be more desirable in high power distance cultures. Thus,

Hypothesis 3a: The relationship between unbiased attributions and interactive behaviors is stronger for those with low power distance orientation than high.

Hypothesis 4a: The relationship between unbiased attributions and performance correction strategies is stronger for those with high power distance orientation than low.

Leader Behaviors and Leadership Effectiveness

Despite differences across cultures, there are many aspects and attributes of leadership that are universal (House et al., 2004; Karakitakoglu-Aygun et al., 2021). In this study context, managers who make more accurate attributions are more communicative and engage in more interactive behaviors (Lakshman, 2013). Additionally, they provide feedback immediately and without undue delay. These are likely to help the “self-esteem” enhancing tendencies of subordinates and thus contribute positively to self-efficacy, satisfaction, and motivation across cultures. This is likely to result in more leadership effectiveness when managers engage in more interactive behaviors than otherwise. Specifically, leader development of integrative strategies to correct performance deficiencies is likely to be seen positively by subordinates, because they contribute to uncertainty reduction and enhancing self-esteem (Lakshman, 2013). Positive subordinate attitudes are likely to fuel virtuous performance-efficacy spirals (Lindsley et al., 1995), thus enhancing subordinate, unit performance, and leadership effectiveness. The underlying basis for the positive relationship between interactive behaviors and performance correction strategies with leadership effectiveness is uncertainty absorption and esteem/satisfaction enhancement.

Hypothesis 5: Across cultures, managerial interactive behaviors are positively related to leadership effectiveness.

Hypothesis 6: Across cultures, managerial performance correction strategies are positively related to leadership effectiveness.

Method

We conducted our empirical study in five countries: France, Vietnam, Turkey, the USA, and India. We used a questionnaire that was translated from English to the appropriate local language (for France, Vietnam, and Turkey) and then back-translated to verify equivalence. We used the English versions in the USA and India. We obtained responses from employees in a variety of business organizations. The respondents reported on the nature of their manager’s attributions and behaviors and their own perceptions of leadership effectiveness. In all five countries, participants in executive education programs were solicited to participate.

We obtained 521 responses on relationships with their managers at their respective places of work. Our overall sample consists of 92 respondents from the USA, 68 from France, 117 from India, 116 from Vietnam, and 128 from Turkey. Our sample represented a balance between male (56.6%) and female respondents, who had an average relationship tenure of 2.6 years with managers, and on average, had 8.19 years of experience. Respondents represented different business sectors, with 12.6% from retail, 12.6% from telecommunication services, 26.8% from financial services, 31% from IT services, 10.4% from manufacturing, and 6.6% from non-profit organizations. We provide details of the samples in each of the countries in Table 1.

Table 1

Details of Country Samples

Details of Country Samples

-> See the list of tables

Respondents were asked to think about a recent manager’s handling of subordinate performance while answering questions. Attributional Accuracy (AA) was measured with a 14-item scale (α=.88). All responses were on 5-point Likert scales ranging from strongly disagree to strongly agree. Seven items pertained to instances of good performance and seven to instances of poor performance. Sample items are – “my supervisor is usually able to identify the cause of my good performance”; and “my supervisor is usually not able to identify the situational factors that led to my poor performance.”

Unbiased Attributions (UA) was measured with six items (α=.87) such as, “in assigning credit or blame, my supervisor is free of gender biases”; and “my supervisor credits employees of similar ethnicity and discredits employees of dissimilar ethnicity (reverse scored).” Higher scores on this scale indicate lower levels of bias vis-à-vis gender, ethnicity, cultural, language, or regional affiliation.

Interactive behaviors were measured with three items (α=.78), assessing the degree to which supervisors interacted positively with subordinates under poor performance contexts. The scale uses items such as, “when I perform poorly, my supervisor interacts with me positively”; and “when I perform poorly, my supervisor does nothing to ease my apprehension.” Managers’ performance correction strategies were measured with three items (α=.74) to assess the degree to which they developed strategies and communicated these to subordinates. It contains items such as, “my supervisor provides me with helpful tips to perform better”; and “oftentimes, my supervisor does not develop plans to help me improve my performance.”

Leadership effectiveness was measured with items from existing scales (e.g., Phillips & Lord, 1981). The six-item scale (α=.94) assessed the degree to which subordinates thought their supervisors were providing leadership. The scale contains items such as, “My supervisor provides good leadership to his/her organization”; and “my supervisor’s behaviors provide leadership to the organization.” Power distance orientation (individual-level) was measured using three items (α=.67) from existing scales (Brockner et al., 2001).

Results

We first conducted a Confirmatory Factor Analysis (CFA). Results suggest that the measurement model fits the data well (Chi-Sq=883.54, p<.01; df=168; CFI=.92; IFI=.92; NFI=.91; RMSEA=.08) and shows evidence of construct- and discriminant- validity (Fornell and Larcker, 1981). We examined the correspondence of the constructs with the corresponding items and found that the error-free variance of the set of items related to constructs (i.e., Average Variance Extracted-AVE) are all above the threshold of 0.5 (Fornell & Larcker, 1981).

We also compared the square root of the AVE of the constructs with the correlations between constructs and found that it is larger than the inter-construct correlations in all cases, providing evidence of discriminant validity (Fornell & Larcker, 1981). In all, results suggest that measurement of the constructs provide evidence of validity and indicate common method variance is not a problem (see Table 2).

Table 2

Means, Standard deviations, correlations, and square root of the AVEa

Means, Standard deviations, correlations, and square root of the AVEa

a : Square root of the AVE along diagonal

All correlations of magnitude larger than .10 are significant at p<.05, N=520.

-> See the list of tables

We tested hypotheses using hierarchical regressions (see Table 3, 4, & 5) and the PROCESS MACRO in SPSS (Preacher & Hayes, 2004). We created dummy variables for each country and entered the four-country dummies in the first step, followed by subordinate experience, gender, and relationship tenure. As subordinate gender and relationship tenure were not significant in any of the regressions, we removed them from the final analyses reported in tables 3, 4, & 5.

Table 3

Regression on Interactive Behaviors

Regression on Interactive Behaviors

* p <0.05, ** p < 0.01, *** p < 0.001 for all tables.

-> See the list of tables

Table 4

Regression on Performance Correction Strategies

Regression on Performance Correction Strategies

* p <0.05, ** p < 0.01, *** p < 0.001 for all tables.

-> See the list of tables

Table 5

Regression on leadership effectiveness

Regression on leadership effectiveness

* p <0.05, ** p < 0.01, *** p < 0.001 for all tables.

-> See the list of tables

As shown in model 1 (table 3), the country dummies and subordinate experience do not explain much variance in interactive behaviors. Model 2 results (table 3) support hypothesis 1 relating the accuracy of attributions and interactive behaviors (β=.73, p<.001). Model 3 results support hypothesis 3 by showing that unbiased attributions are positively related to managerial interactive behaviors (β=.19, p<.001). Model 4 results (table 3) reveal that there is evidence for convergence across the five countries examined in that the relationships between accurate attributions (β=.68, p<.001) and unbiased attributions (β=.58, p<.001) respectively, with interactive behaviors, holds across these settings. Still, there is some divergence across cultures, supporting hypothesis 3a, showing that power distance moderates the relationship between unbiased attributions and interactive behaviors (β=-.43, p<.05).

Following Aiken and West (1991) we plotted the interaction in Fig 2a, which shows a positive relationship between unbiased attributions and interactive behaviors for both low and high-power distance orientation. However, this relationship is stronger in the low power distance orientation case than the opposite. While supporting hypothesis 3a, this evidence is also consistent with our broader arguments for the cultural desirability of unbiased attributions and accurate attributions, respectively, across the five cultures.

Figure 2a

Interaction between Power distance and Unbiased Attributions

Interaction between Power distance and Unbiased Attributions

-> See the list of figures

Results for model 2 (table 4) show that accurate attributions are positively related to performance correction strategies (β=.64, p<.001), in support of hypothesis 2. The results for model 3 (table 4) show that unbiased attributions are not significantly related to performance correction strategies (β=-.02, ns). This is not consistent with hypothesis 4. However, results in model 4 (table 4) show that unbiased attributions are significantly related to performance correction strategies (β=.75, p<.001) while controlling for power distance and its interaction with unbiased attributions, which shows a significant relationship to performance correction strategies (β=-.93, p<.001). Thus, there seems to be support for hypothesis 4 in model 4 results. Additionally, the significant coefficient for the interaction is in support of hypothesis 4a.

We plotted the interaction in Fig. 2b, which shows that the slope of the relationship between unbiased attributions and performance correction strategies is stronger in the case of high-power distance orientation than low. Support for the positive relationship between accurate attributions (β=.82, p<.001), and unbiased attributions (β=.75, p<.001) respectively, with performance correction strategies in Model 4, which controls for the moderating effect of power distance, is in support of convergence across the five cultures.

Figure 2b

Interaction between Power distance and Unbiased Attributions

Interaction between Power distance and Unbiased Attributions

-> See the list of figures

Results in model 2 (table 5) show that each of our attributional constructs – accurate attributions (β=.54, p<.001) and unbiased attributions (β=.23, p<.001) – are directly related to leadership effectiveness. As seen in model 3 (table 5), interactive behaviors (β=.29, p<.001), and performance correction strategies (β=.14, p<.01) are each related positively to leadership effectiveness, in support of hypotheses 5 and 6. Additionally, after the entry of these constructs, the size of the respective coefficients for accurate attributions (β=.26, p<.001) and unbiased attributions (β=.17, p<.001) have decreased from their level in model 2 (table 5). The incremental variance explained by interactive behaviors and performance correction strategies (∆R2=.07, p<.001) indicates that these managerial behaviors mediate the effects of accurate/unbiased attributions on leadership effectiveness. In what follows, we present the results of the PROCESS MACRO to compare the direct and indirect effects more rigorously.

We ran two process regressions (Preacher & Hayes, 2004). The results of the first suggest that the direct effect of accurate attributions on leadership effectiveness is positive and significant (.47, p<0.001), with the 95% bias-corrected confidence interval estimated to be between .33 and .62. The two indirect effects, through interactive behaviors (.38; 95% C.I.: .27 to .50) and performance correction strategies (.11; 95% C.I.: .01 to .21) were both significant as well.

We ran a second process regression with unbiased attributions as the independent variable. Again, we found support for a statistically significant direct effect in addition to two mediating effects, each of which was also significant. First, the direct effect of unbiased attributions on leadership effectiveness is positive and significant (.25, p<.001), with the 95% confidence interval estimated to be between .17 and .34. The two indirect effects, through interactive behaviors (.22; 95% C.I.: .16 to .29) and performance correction strategies (.07; 95% C.I.: .04 to .12) were both significant as well. Thus, the results provide overall support to the model shown in Figure 1.

Discussion

Despite the growing interest in examining leadership in cross-cultural contexts (e.g., Engelen, et al., 2014; Karakitapoğlu-Aygün et al., 2021; Miao et al., 2018; Smith & Peterson, 2017), one key leadership process – attributions – has not been examined. Some of this cross-cultural literature has examined the effectiveness of leader traits (e.g., Miao et al., 2018), while others have examined leadership styles (e.g., Engelen et al., 2014; Karakitapoğlu-Aygün et al., 2021). Leaders are more able to influence, motivate, and enable subordinates to contribute to the success of their organizations when they display traits or styles that are congruent with the cultural context. Although these studies answer the question of “what” type of leaders or leadership is likely to be effective across cultures, this literature does not provide any answers on the nature of leadership processes (i.e., “how”) that are effective across cultures. On the other hand, the attribution theories of leadership tend to focus primarily on biased attributions rather than accurate ones, except for some very recent studies (Lakshman, 2013; van Houwelingen et al., 2021). Thus, we know what happens if leaders are wrong in their attribution but do not know the benefits of their correct attributions. We contribute to these two literatures by providing an answer in the context of one crucial leadership process (i.e., attributional process, Chen & Van Velsor, 1996; Lakshman, 2013) by examining it in five different cultures emanating from different cultural clusters (see House et al., 2004).

Contribution to the literature of cross-cultural leadership

Given that attribution theories of leadership are very important for cross-cultural leadership (Lakshman, 2013), we argued that the lack of research on cultural desirability of managerial attributions of subordinate performance is a critical area, worthy of examination (e.g., Dean & Koenig, 2019). We contribute to this literature by examining the role of two attributional characteristics – being accurate and being unbiased (gender, race, culture) – in the processes leading to leadership effectiveness in five cultures, each of which is from a different societal cluster (House et al., 2004).

Consistent with the IB literature (Lakshman et al., 2021), we find that accurate attributions, and unbiased attributions are strongly related to leadership effectiveness directly and through a behavioral process. This finding holds well across five countries from different cultural clusters. Thus, accurate and unbiased attributions are culturally endorsed in these five cultures. These managerial attributions are potentially more universal as well, which remains to be examined in future research in other country samples. Following the pattern of GLOBE study findings that certain leader traits are universally desirable across cultures, we provide evidence of certain attributional processes of a specific nature that are likely to be universally desired, at least in the context of the five cultures examined. Recent research shows that the quality of the communication experience, in broad terms, is a significant mediating mechanism across three different cultures (Karakitapoglu-Aygun et al., 2021). Our findings in the more specific context of attributional processes and interactive communication between leaders and subordinates are consistent with these findings. However, our findings are more specific and provide details of the communication experience in a manner hitherto unexamined. One clear implication emerging from our findings is that the expectation of trustworthy, credible, motivating, and communicative interactions between leaders and subordinates are likely to be seen across cultures. We find that when such interaction is based on culturally unbiased (or gender, race etc.) attributions and on the accuracy of attributions for work performance, they are more likely to be effective across cultures.

Despite this convergence in the cultural desirability of managerial attributions, we also find some divergence by way of power distance orientation of individuals. While accurate/unbiased attributions are important and desirable across the five cultures examined, they are seen and valued somewhat differently across cultures. Specifically, while interactive behaviors following managerial attributions are relatively more important in low power distance cultures, performance correction strategies are relatively more important in high power distance cultures. These differences point to the nuances of communication differences between low – and high – power distance cultures such as the USA and Vietnam, for instance, in our study. Although the five countries also vary on other cultural dimensions (e.g., collectivism), we theorized that power distance would be more important in this context of leader-subordinate interaction with the focus on the process of leading to higher organizational effectiveness (e.g., Lakshman et al., 2014, 2019). Our findings show that our theorizing is valid as far as these five cultures are concerned, in a manner similar to previous findings from a broader communication context (Karakitapoglu-Aygun et al., 2021).

Our study begins to provide a picture of manager-subordinate relationships and the constituent attributional mechanisms, which are likely to result in cross-cultural leadership effectiveness (Chen & Van Velsor, 1996). Subordinates value accurate attributions of performance, which are also cultural bias-free (e.g., stereotypes; Lakshman, 2013). Unbiased attributions, in particular, are more strongly endorsed in low power distance cultures than in high, although they are important in all. Managers who make more accurate attributions are more likely to engage in interactive behaviors strongly related to leadership effectiveness. Additionally, managers who make more accurate attributions are also more likely to develop and communicate performance correction strategies, critical for leadership effectiveness.

As noted above, we contribute to the literature by adding an answer to the “how” question pertaining to cross-cultural leadership effectiveness in a context where most of the studies examine the “what” question (e.g., Miao et al., 2018). We add to the list of traits (what) identified as universally desirable for leadership effectiveness by identifying one mechanism – leader attribution accuracy – by which (i.e., how) leaders may translate traits such as communicator and excellence oriented. But our study goes beyond a simple list of traits devoid of context and contributes by demonstrating attributions and behaviors that represent traits in performance contexts, which are culturally endorsed for leadership. While studies have argued that specific traits are culturally endorsed (House et al., 2004) and important for leadership effectiveness, these have not been examined in the context of specific leadership processes involving managers and subordinates, such as our study.

Contribution to the literature of attribution theories of leadership

Research on leader attributions has only recently started to focus on accuracy and the resulting behavioral process resulting in effective leadership across cultures (e.g., Lakshman et al., 2021). Despite the importance of attributional processes and accurate attributions, this literature has focused on attributional biases that are likely more reflective of self-protective leadership, at best or reflective of universally undesirable leadership. We contribute to the literature on attribution theories of leadership (e.g., Martinko & Mackey, 2019) by theorizing and finding that accurate attributions made by leaders are universally desirable across the five cultures examined here. Our finding that accurate attributions are related to interactive behaviors and performance correction strategies across five cultures suggests that subordinates may infer traits such as “communicator” and “excellence oriented,” which have been identified as universally desirable. Recent research has found that accuracy of attributions made by leaders may be founded on complex cognitive processing (e.g., van Houwelingen et al., 2021) and/or more complex attributional processing (e.g., Lakshman & Estay, 2016; Lakshman et al., 2021) on the part of managers. Future research needs to examine the cross-cultural generalizability of the relationship between such higher-order cognitive processing and subsequent attributional processes in a broader number of cultures.

Our study also contributes to the literature by identifying unbiased (gender, ethnic, cultural) attributions as strong predictors of interactive behaviors, performance correction strategies, and leadership effectiveness across the five cultures examined. This identifies one way in which (i.e., how) leaders can demonstrate universally desirable traits, at least in the five clusters examined here. These may show to subordinates that leaders are trustworthy and just, in addition to the perception of being communicator and excellence oriented, which are all universally desirable traits of effective leadership. Despite relative differences across low- and high-power distance cultures, these findings hold across the five cultures examined and provide a foundation for an attribution theory of leadership that is more universal, in contrast to extant literature.

Managerial Implications

Our study findings have several key managerial implications. First, we highlight the differences between the “what” and the “how” issues pertaining to leadership effectiveness across cultures. Thus, managers need to possess not only the appropriate traits and use the appropriate styles of leadership, but they also need to pay attention to the specific processes (e.g., attributions) that are critical in this regard. A vast literature on cross-cultural similarities and differences in attributions has noted that this is a natural process in work interactions, some aspects of which are automatic but need to be controlled for improving effectiveness (e.g., Dean & Koenig, 2019). Managers need to be aware of the accuracy of their attributions for subordinate performance across cultural contexts. More importantly, they need to be free of cultural, gender, racial, and ethnic biases in making these performance attributions. Additionally, they need to be aware of the cultural nuances vis-a-vis power distance, among others, in the ensuing interactions and performance management approaches with subordinates. Organizations can design training programs to help managers improve the quality and accuracy of their attributions to ensure high-quality leader-subordinate exchange relationships. Thus, our study findings can help managers and human resource professionals in improving leadership effectiveness.

Our study is limited in its coverage to five of the ten societal clusters identified by GLOBE (House et al., 2004). Our study is also cross-sectional and based only on subordinate responses. Yet, we provide support for our theoretical model and its generalizability, which in itself is a critical contribution.

Conclusion

Our paper started with the premise that the cross-cultural leadership literature has not addressed the issue of accurate attributions of leaders, and the attributional theories of leadership literature have focused primarily on biased attributions. We contribute to these two literatures by relying on CLTs to examine the accurate attributional process domain in five cultures. We hope our study is just the beginning of a broader examination of ties to higher-order cognitive processes (e.g., Lakshman et al., 2021; van Houwelingen et al., 2021) and leadership dimensions and their universal/contingent applicability across cultures. One potential area of future research is the question of what relationship, if any, exists between universally endorsed transformational leaders (e.g., Den Hartog et al., 1999) and the attributions they make in cross-cultural contexts? We think that it is important to discover behavioral processes (Karakitapoglu-Aygun et al., 2021; Robinson & Harvey, 2008) that are reflective of underlying traits, which are likely endorsed differently from one culture to another. This will serve as a critical advancement in cross-cultural leadership literature. In addition, the cultural desirability of accurate attributions in five culture clusters may extend to the other clusters and thus be truly universally desirable. This remains to be examined in future research, but we believe this is a strong possibility.