The role of status and power inequalities in leader-member exchange

Demographic status and LMX quality

I first discuss demographic status, that is, status in the context of an informal social hierarchy (Figure 1). More specifically, I assess the impact of the leader’s demographic status on the LMX quality assessed by the member, and the impact of the member’s demographic status on the LMX quality assessed by the leader. As status involves perceptions about the other person, that is, perceptions about the LMX partner, the analysis of the effects of status is at the interpersonal level.

To theorise about the impact of demographic status on LMX quality, I draw upon social dominance theory (Sidanius and Pratto, 1999). Social dominance theory explains that demographic characteristics, such as race, gender and ethnicity, are hierarchically organised such that higher status is attributed to groups with privilege, power and wealth (Levin et al., 1998; Sidanius and Pratto, 1999). Although age- and gender-related hierarchies tend to exist in all societies, social hierarchies of demographic characteristics are often society-specific (Sidanius and Pratto, 1999). For example, in the United States, it is widely believed that men and whites are of higher, while women and African Americans are of lower status (Sidanius and Pratto, 1999). While these examples present status as a binary construct, social hierarchies tend to be non-binary. Taking race in the United States as an example, white and African Americans are typically placed on the two ends of social hierarchy, while other races, such as Asian Americans, are placed between these two ends (Sidanius and Pratto, 1999; Sy et al., 2010; Zou and Cheryan, 2017). Furthermore, status is relative, that is, an individual or a group has higher or lower status compared to other individuals or groups. Therefore, individuals on the high end of society’s hierarchy are typically considered to have higher status and individuals on the low end of society’s hierarchy are typically considered to have lower status. The status of individuals whose demographic characteristics place them in the middle of society’s hierarchy depends on the comparison with others. As such, for example, whites are generally considered of higher status, and African Americans are generally considered of lower status. On the other hand, Asian Americans are likely to be considered of higher status by African Americans, but of lower status by whites. Therefore, I use the term ‘higher status’ to signify either status at the high end of the social hierarchy or status which is higher compared to the status of the other, and the term ‘lower status’ to signify either status at the low end of the social hierarchy or status which is lower compared to the status of the other.

Social dominance theory asserts that people tend to support the maintenance of social hierarchies linked to demographic characteristics, and as such, demonstrate greater liking toward individuals of higher status in society than toward those of lower status (Sidanius and Pratto, 1999). For example, people tend to be more positively predisposed toward whites than toward African Americans because whites are considered of higher status (Sidanius and Pratto, 1999). Interestingly, evidence in neuroscience points in the same direction; Zink et al. (2008) found that responses of the brain to higher status differ from responses of the brain to lower status. When the target was considered of higher status compared to the participant, the researchers noticed that activity in the participant’s brain areas associated with perceptual/attentional processing was significantly greater than when the target was perceived to be of lower status (Zink et al., 2008). Therefore, higher demographic status is related to liking and attention, while lower demographic status is related to discriminatory attitudes toward lower-status individuals.

As liking and attention are strongly associated with LMX quality (Graen and Scandura, 1987; Martinko et al., 2018), demographic status is likely to affect LMX quality. Considering that we treat status not as a property of the actor, but rather as attributed to a person by others, the member’s status is likely to affect LMX quality assessed by the leader, and the leader’s status is likely to affect LMX quality assessed by the member. In other words, a leader or a member assessing LMX quality is influenced by the status of their LMX partner.

Proposition 1

There is a positive relationship between the LMX partner’s demographic status and LMX quality, such that LMX quality is higher when the LMX partner’s status is higher.

Proposition 1 does not consider the role of the LMX quality assessor’s status. In other words, LMX assessors of higher status (e.g. a man) might be likely to be more positively predisposed toward an LMX partner also of higher status (e.g. also a man) compared to an LMX partner of lower status (e.g. a woman). However, what happens when the LMX assessor is of lower status (e.g. a woman)? More nuance is needed to answer this question.

To add nuance, I draw upon the concept of out-group favouritism and social identity theory (Tajfel and Turner, 1979). Out-group favouritism refers to lower-status individuals and is a ‘preference for members of a group to which one does not belong’ (Jost et al., 2004: 891). Sidanius and Pratto’s (1999) examination of prior empirical data revealed that members of lower-status groups might not favour members of their own group but prefer members of higher-status groups. A famous example of out-group favouritism is the experiment by Clark and Clark (1947). The researchers demonstrated that most African American children, given a choice between a black doll and a white doll, not only preferred to play with the white doll but also thought that the white doll was better than the black doll (Clark and Clark, 1947). The empirical outcomes reviewed by Sidanius and Pratto (1999) confirm the assertion of Tajfel and Turner (1979) that lower-status individuals may derogate members of their own demographic group and show positive attitudes toward individuals of higher demographic status.

Social dominance theory scholars explain that the reason for out-group favouritism is that both higher- and lower-status demographic groups tend to legitimise status differences through ideologies, such as merit-based mobility, which favour higher-status individuals and ignore systemic biases against lower-status individuals (Major et al., 2002; Sidanius and Pratto, 1999). Indeed, it is often the consensus on those ideologies by both higher- and lower-status demographic groups that help maintain inequalities (Sidanius and Pratto, 1999). The explanation of legitimising ideologies is also in agreement with the notion of system justification ideologies, according to which ‘the ideas of the dominant tend to become the ideas of the dominated’ (Jost and Banaji, 1994: 10), reproducing negative stereotypes of lower-status individuals also within their own demographic groups. Consequently, lower-status individuals may believe that their own demographic group bears the responsibility for their lower status and related outcomes (Major et al., 2002). In sum, while, according to similarity-attraction theory, similarity among lower-status individuals would encourage favouritism toward other individuals of lower status, according to social dominance theory, lower-status individuals may favour higher-status individuals over individuals in their own demographic group (Sidanius and Pratto, 1999; Umphress et al., 2007).

To explore when individuals of lower demographic status favour other lower-status individuals and when they favour higher-status individuals, I draw upon social identity theory (Tajfel and Turner, 1979). Social identity theory posits that lower-status individuals employ social identity strategies to improve their self-esteem (Tajfel and Turner, 1979). These social identity strategies are the following: (1) individual mobility, an individual’s attempt to dissociate with their own minority group and join a higher-status group, (2) positive distinctiveness, an attempt to improve the status of their lower-status group either by creatively indicating the group’s positive attributes (social creativity) or by directly antagonising the higher-status group (social competition) (Tajfel and Turner, 1979). Therefore, higher-status individuals in the workplace are likely to identify with their own demographic group, while lower-status individuals identify with their own demographic group when their social identity strategy is social distinctiveness but identify with higher-status groups when their social identity strategy is individual mobility (Chattopadhyay et al., 2004).

Proposition 2

When the LMX quality assessor has lower demographic status, the effect of the LMX partner’s demographic status on LMX quality depends on the LMX quality assessor’s social identity strategy.

To illustrate the difference between predictions based on similarity-attraction theory and predictions based on Propositions 1 and 2, I provide two examples. The first example concerns effects of the leader’s race on member-assessed LMX quality. The second example concerns effects of the member’s gender on leader-assessed LMX quality.

Example 1

Effects of race on member-assessed LMX quality. Table 2 shows predictions of the effects of race on LMX quality, as assessed by the member. In the case of similarity, when both the leader and the member are white, that is, of higher demographic status, there is no difference in the predicted outcomes between similarity-attraction and Propositions 1 and 2. However, when the leader and the member are both African Americans, while similarity-attraction theory predicts a positive effect due to similarity, Propositions 1 and 2 predict that the effect depends on the member’s social identity strategy. When the member’s social identity strategy is individual mobility, the effect is likely to be negative. When the member’s social identity strategy is positive distinctiveness, the effect is likely to be positive. In the case of dissimilarity, when the leader is white and the member is African American, the predictions between similarity-attraction and Propositions 1 and 2 differ. While similarity-attraction theory predicts a negative effect due to dissimilarity, Propositions 1 and 2 predict that the effect depends on the member’s social identity strategy. When the member’s social identity strategy is individual mobility, the effect is likely to be positive. When the member’s social identity strategy is positive distinctiveness, the effect is likely to be negative. However, when the leader is African American and the member is white, both similarity-attraction theory and Propositions 1 and 2 predict a negative effect; similarity-attraction theory predicts a negative effect due to dissimilarity, and Propositions 1 and 2 predict a negative effect due to the leader’s lower demographic status, in combination with the member’s higher demographic status.

OPEN IN VIEWER

Table 2.

Comparison of predictions between similarity-attraction theory and Propositions 1 and 2. Example 1: Effects of race on member-assessed LMX quality.

Race		Effects on member-assessed leader-member exchange (LMX) quality
Leader (LMX partner)	Member (LMX assessor)	Similarity-attraction theory	Propositions 1 and 2
White	White	Positive	Positive
African American	African American	Positive	Negative when the member seeks individual mobility
			Positive when the member seeks positive distinctiveness
White	African American	Negative	Positive when the member seeks individual mobility
			Negative when the member seeks positive distinctiveness
African American	White	Negative	Negative

Example 2

Effects of gender on leader-assessed LMX quality. Table 3 shows predictions of the effects of gender on LMX quality as assessed by the leader. In the case of similarity, when both the leader and the member are male, that is, of higher status, there is no difference in the predicted outcomes between similarity-attraction theory and Propositions 1 and 2. However, when both the leader and the member are female, while similarity-attraction theory predicts a positive effect due to similarity, Propositions 1 and 2 predict that the effect depends on the leader’s social identity strategy. When the leader’s social identity strategy is individual mobility, the effect is likely to be negative. When the leader’s social identity strategy is positive distinctiveness, the effect is likely to be positive. In the case of dissimilarity, when the leader is male and the member is female, both similarity-attraction and Propositions 1 and 2 predict a negative effect; similarity-attraction theory predicts a negative effect due to dissimilarity, and Propositions 1 and 2 predict a negative effect due to the member’s lower demographic status, in combination with the leader’s higher demographic status. However, when the leader is female and the member is male, while similarity-attraction theory predicts a negative effect due to dissimilarity, Propositions 1 and 2 predict that the effect depends on the leader’s social identity strategy. When the leader’s social identity strategy is individual mobility, the effect is likely to be positive. When the leader’s social identity strategy is positive distinctiveness, the effect is likely to be negative.

In sum, Propositions 1 and 2 suggest that a member’s demographic status impacts leader-assessed LMX quality. My predictions agree with those of similarity-attraction theory only when the member and the leader are both of higher status in society’s hierarchy and when the member has lower status than the leader. Conversely, a leader’s demographic status impacts member-assessed LMX quality. My predictions agree with those of similarity-attraction theory only when the leader and the member are both of higher status in society’s hierarchy and when the leader has lower status than the member.

Findings of several relational demography studies support my propositions rather than predictions based on similarity-attraction theory. For example, Schaffer and Riordan (2013) found that member-assessed LMX quality when both the leader and the member were white (higher status) was higher than when both the leader and the member were African American (lower status). Similarity-attraction theory cannot explain why members’ perceptions of LMX quality differ in these two cases of demographic similarity. However, my propositions predict, in agreement with these findings, that LMX quality will be higher when the leader and the member have higher status, while when they have lower status, LMX quality will depend on the LMX quality assessor’s social identity strategy. Dwertmann and Boehm (2016) found that member-assessed LMX quality was higher when the member had a disability and the leader did not, in comparison to when the leader had a disability, and the member did not. While similarity-attraction theory cannot explain why two cases of demographic dissimilarity produce different results, my propositions predict that, as assessed by the member, LMX quality is likely to be lower when the leader is of lower status (has a disability). In the reverse situation, LMX quality will depend on the member’s social identity strategy. From the leader’s perspective, Tsui and O’Reilly (1989) found that among different leader-member race combinations, while white leaders indicated the highest level of affect (liking) toward white members, African American leaders reported the lowest level of affect (liking) toward African American members. These findings also align with my propositions but only partially align with similarity-attraction theory.

OPEN IN VIEWER

Table 3.

Comparison of predictions between similarity-attraction theory and Propositions 1 and 2. Example 2: Effects of gender on leader-assessed LMX quality.

Gender		Effects on leader-assessed leader-member exchange (LMX) quality
Leader (LMX assessor)	Member (LMX partner)	Similarity-attraction theory	Propositions 1 and 2
Male	Male	Positive	Positive
Female	Female	Positive	Negative when the leader seeks individual mobility
			Positive when the leader seeks positive distinctiveness
Male	Female	Negative	Negative
Female	Male	Negative	Positive when the leader seeks individual mobility
			Negative when the leader seeks positive distinctiveness

Note: Different predictions between similarity-attraction and Propositions 1 and 2 are highlighted in grey. LMX: leader-member exchange.

Positional status and LMX quality

In this section, I discuss the effects of positional status, that is, status in the context of formal social hierarchies (Figure 1). Formal hierarchies are characterised by differential status, such that leaders are ascribed higher status than their members (Blom and Lundgren, 2020; Tourish, 2014). Romanticising and celebrating leaders (Bligh and Schyns, 2007; Blom and Alvesson, 2015; Collinson et al., 2018), but not members, is an indication of the higher status attributed to leaders. Therefore, in demographically diverse leader-member dyads, the status of the LMX partner is derived not only from the LMX partner’s demographic characteristics but also from the LMX partner’s position in the formal organisational hierarchy. Following the logic of the preceding section, that is, higher status is associated with greater attention and liking (Sidanius and Pratto, 1999; Zink, et al., 2008), and attention and liking underlie high LMX quality (Graen and Scandura, 1987; Martinko et al., 2018), I expect LMX quality to be higher when the member assesses LMX with their leader (higher positional status) than when the leader assesses LMX with their member (lower positional status).

Proposition 3

There is a positive relationship between the LMX partner’s positional status and LMX quality, such that LMX quality is assessed as higher when the LMX partner is the leader.

The literature on leadership and diversity suggests that leader demographic and positional status interact to affect perceptions about the leader and related outcomes (Dwertmann and Boehm, 2016; Rudman and Phelan, 2008; Rudman et al., 2012). Extending social dominance theory’s assertion that lower-status individuals are discriminated against by defenders of the established social hierarchy (Sidanius and Pratto, 1999), Rudman et al. (2012) assert that female leaders are status incongruent (positional status is higher but demographic status is lower) and discredit the established social hierarchy, thus, eliciting a strong backlash from its defenders. This backlash might extend to LMX quality assessments. In other words, the relationship between demographic status and LMX quality is likely to be steeper when the LMX partner is the leader than when the LMX partner is the member.

Proposition 4

The degree to which the LMX partner’s demographic status affects LMX quality depends on the LMX partner’s positional status.

Positional power and LMX quality

In the preceding sections, I discussed effects of the LMX partner’s status, both demographic and positional. In this and the next section, I discuss effects of the LMX assessor’s power. It is important to remember that power is treated as a property of the actor. As such, the discussion is at the intrapersonal level of analysis.

Starting with positional power, that is, power in the formal social hierarchy (Figure 1) and drawing upon the power literature, I theorise about the effects of the LMX assessor’s positional power on LMX quality. Power scholars have demonstrated that elevated power increases the automaticity of social cognition, while reduced power is related to controlled social cognition (Fiske, 1993; Keltner et al., 2003; Smith and Trope, 2006). Automatic social cognition is judgement based on rapid, effortless evaluations and use of cognitive heuristics, while controlled social cognition is judgement based on deliberate and effortful evaluations which consider a wide range of information (Keltner et al., 2003). Notably, the powerful are likely to stereotype to form judgements not only by design (form opinions based on stereotypes) but also by default (dismiss information inconsistent with stereotypes) (Goodwin et al., 2000).

In contrast, the powerless recognising that the powerful control the outcomes, tend to make the effort to seek individuating information about the powerful, to gain a better understanding, and ideally, predict the behaviour of the powerful (Fiske, 1993; Fiske and Berdahl, 2007). Stevens and Fiske (2000) conducted experiments to understand how the powerless form impressions of the powerful and found that (1) when the powerless depend on the powerful for task-related outcomes, they pay attention to information disconfirming initial expectations and (2) when the powerless depend on the powerful for their evaluation, the powerless tend to pay more attention to positive and less attention to negative information about the powerholder. In sum, to form judgements, the powerful tend to rely on stereotyping and the powerless tend to rely on individuating information.

Applying the logic of the preceding paragraphs to leader-member dyads, I propose that leaders are more likely to be influenced by the LMX partner’s demographic status, compared to members. This is because leaders are more powerful. Indeed, leadership and power go ‘hand in hand’ (Asad and Sadler-Smith, 2020: 49) because the design of formal hierarchies ensures that leaders have more power than members. Independently of their personal attributes or characteristics, leaders have positional power (Fiske and Berdahl, 2007), that is, legitimate (formal power to make task-related demands and expect compliance), rewards (power to administer rewards, including monetary rewards), and coercive (power to punish lack of compliance) powers because of their leadership position in the organisational hierarchy (Bass, 1960; French and Raven, 1959; Reiley and Jacobs, 2016). As such, leaders pay more attention to stereotyping information about their members, while members pay more attention to individuating information about their leaders (Guinote and Phillips, 2010). Therefore, I expect that when leaders assess LMX quality, their assessment is likely to be influenced by stereotyping information based on the member’s demographic status. In contrast, when members assess LMX quality, their assessment is likely to be influenced less by stereotyping information based on the leader’s demographic status and more by individuating information about the leader. In other words, when the leader assesses LMX quality, the relationship between the member’s demographic status and LMX quality is stronger because the leader’s power is more likely to cause automatic cognition, and consequently reliance on the member’s demographic status for the LMX quality assessment. In contrast, when the member assesses LMX quality, the relationship between the leader’s demographic status and LMX quality is weaker because the member is likely to rely more on individuating information (rather than stereotyping information based on demographic status) to assess the quality of LMX with their leader.

Proposition 5

The degree to which the LMX partner’s demographic status affects LMX quality depends on the LMX quality assessor’s positional power.

Demographic power and LMX quality

In the preceding section, I discussed effects of positional power, that is, power in the context of the formal social hierarchy, on the relationship between the LMX partner’s demographic status and LMX quality. In this section, I discuss effects of demographic power, that is, power in the context of an informal social hierarchy (Figure 1).

In addition to positional, power may be personal (e.g. derived from an individual’s expertise) or political (e.g. associated with control over social resources) (Fiske and Berdahl, 2007; French and Raven, 1959; Keltner et al., 2003). Demographic characteristics are associated with both personal and political power. More specifically, demographic characteristics such as white and male (compared to African American and female) are afforded greater personal power because individuals with these characteristics are attributed competence and other stereotypical personal qualities, such as the ability to lead (Berger et al., 1966; Eagly and Karau, 1991; Fiske and Berdahl, 2007). Because of these attributions, individuals with demographic characteristics associated with higher status are also afforded personal power (Fiske and Berdahl, 2007). Furthermore, individuals with demographic characteristics such as white and male are likely to be afforded greater political power because of their instrumental access to broader and more powerful networks, compared to African Americans and women (Fiske and Berdahl, 2007; Ibarra, 1992). In leader-member dyads, a white male member may capitalise on this access to networks, using political power over the leader. In sum, in demographically diverse leader-member dyads, although positional power is always distributed in the leader’s favour, the distribution of personal and political power may vary. For example, an African American female leader has more positional but may have less personal and less political power over a white male member.

Following the logic of the preceding section about how positional power is likely to cause automatic cognition and reliance on stereotyping, rather than individuating, information (Fiske, 1993; Goodwin et al., 2000; Keltner et al., 2003; Smith and Trope, 2006), demographic power is expected to have similar effects. This expectation is in agreement with social dominance theory which posits that individuals with higher demographic power are more likely to defend the established social hierarchies and discriminate more against individuals with lower demographic status (Pratto et al., 1994; Sidanius et al., 1994). Therefore, we can expect that the higher the demographic power of the LMX assessor, the stronger the effect of the LMX partner’s demographic status on LMX quality.

Proposition 6

The degree to which the LMX partner’s demographic status affects LMX quality depends on the LMX quality assessor’s demographic power.

Taken together, my propositions form the model illustrated in Figure 2. In the following section, I discuss the proposed model’s theoretical and practical implications and suggest promising future research opportunities.

OPEN IN VIEWER

Figure 2.

Proposed model.

Theoretical contributions

The proposed model advances theory about demographically diverse leader-member dyads. First, it contributes to the LMX literature by explaining the underlying mechanisms of LMX quality assessment for demographically diverse leader-member dyads. More specifically, it shows how LMX quality is affected by a complex of inequalities at the interpersonal (the LMX partner’s properties affect how the LMX assessor perceives LMX quality) and at the intrapersonal (the LMX assessor’s properties affect how the LMX assessor perceives LMX quality) level.

Second, the proposed model contributes to the status literature by integrating social dominance and social identity theories to offer a nuanced explanation of how effects of status vary depending on whether the assessor is of higher or lower status, and on the assessor’s social identity strategy. Although several studies have considered the role of the assessor’s status on the relationship between demographic characteristics and workplace outcomes (e.g. Chattopadhyay, 1999; Umphress et al., 2007), very few have considered the role of social identity strategy (e.g. Chattopadhyay et al., 2004). By considering the assessor’s social identity strategy, this study sheds light on the conditions under which an assessor of lower demographic status is likely to be positively predisposed toward individuals of similar status, or toward individuals of higher status.

Third, the proposed model contributes to the relational demography literature by specifying how status and power derived from demographic characteristics can interact to affect work-related outcomes. The proposed model offers a more nuanced approach to examine demographic differences than models based solely on similarity-attraction theory. Although relational demography scholars urge researchers to consider status when examining relationships between demographic (dis)similarity and relational outcomes (Chattopadhyay et al., 2010; Dwertmann and Boehm, 2016; Schaffer and Riordan, 2013), to my knowledge, this study is the first to consider joint effects of formal and informal status and power in relational demography. Another contribution to the relational demography literature is showing when similarity-attraction theory is expected to predict outcomes reliably. This is done by comparing LMX quality predictions based on similarity-attraction theory and LMX quality predictions based on social dominance and social identity theories.

Finally, the proposed model contributes to the status and power literature by applying Magee and Galinsky’s (2008) framework on a specific organisational unit, the leader-member dyad.

Practical implications

In addition to its theoretical contributions, the proposed model has practical implications. If validated through empirical research, this model may provide a foundation for practice-based solutions. To guide practitioners and advance research, scholars have identified moderators of the relationship between demographic (dis)similarity and LMX quality. For example, Scandura and Lankau (1996) explain how social psychological processes, such as self-knowledge, interpersonal skills, communication competence and cultural competence, influence the development of demographically diverse leader-member dyads. Brouer et al. (2009) discuss how political skill can moderate the negative effects of demographic dissimilarity. However, until research identifies the main effects of demographic characteristics on LMX quality with more consistency, moderator propositions may be insufficient to offer solutions for practitioners. In other words, practitioners need to know in which cases these solutions may be applicable. By showing when demographic (dis)similarity predicts LMX quality reliably, this study helps us understand when various solutions are applicable. Finally, I hope that my propositions will result in more conclusive empirical results, allowing for the development of more solutions.

Future research opportunities

The proposed model is generalisable across a wide range of demographic characteristics that may be associated with status and power and can consequently encourage a more accurate examination of more demographic characteristics. This model can be applied to demographic characteristics in a binary or non-binary manner. Demographic status has been so far studied mostly in a binary manner, for example, men versus women, white versus non-white and with versus without disability (e.g. Dwertmann and Boehm, 2016; Geddes and Konrad, 2003; Paustian-Underdahl et al., 2017 – for an exception, see Paunova, 2017). My propositions allow the examination of status in a binary or non-binary manner, and as such, they create future research opportunities for demographic characteristics that have not been so far sufficiently studied or can be studied with greater accuracy. Examples of demographic characteristics underexplored with respect to status and power are sexual orientation and disability (Tatli and Özbilgin, 2012). An example of a demographic characteristic that can be examined from a status perspective with greater accuracy is race. Race is often treated in a binary manner, for example, whites versus non-whites (e.g. Paustian-Underdahl et al., 2017). However, race in the United States is not binary. Races such as Asian and Latino Americans are hierarchically organised between white and African Americans (Sidanius and Pratto, 1999; Zou and Cheryan, 2017). As such, if future research treats race-related status as a non-binary variable with various races hierarchically organised based on societal beliefs, research results are likely to be more accurate due to the input of more fine-grained data.

Although my propositions can be tested with binary and non-binary variables, they are developed based on single demographic characteristics rather than groups of demographic characteristics. Research, in general, tends to examine demographic status and power focusing on single demographic characteristics, such as gender and race (Tatli and Özbilgin, 2012). To obtain more generalisable results, an alternative approach might be to consider different types of demographic characteristics. To this end, social dominance theory argues that societies have ‘three qualitatively distinct systems of group-based hierarchy’, ‘an age system’, ‘a gender system’ and ‘an arbitrary-set system’ (Pratto et al., 2006: 273). These systems differ from each other as arbitrary-set system characteristics, such as race and religion, are ‘distinguished by a very high degree of plasticity, both in terms of which group distinctions become socially significant and in the permeability of the group boundaries’ (Pratto et al., 2006: 274). These differences in hierarchies between various demographic characteristic types can be explored as to how they translate in terms of demographic status and power. Therefore, future research can theorise and test how demographic status and power may vary between demographic characteristic types and the implications of this variability on workplace outcomes, including LMX quality.

Another issue that requires the attention of future research is the operationalisation of demographic status based on multiple demographic characteristics. Although demographic characteristics are used as proxies of status and carry independent status value (Ridgeway, 1991), intersectionality theory asserts that demographic characteristics should not be studied in isolation because only by examining the intersection of gender, race and sexual orientation, among others, we can understand the full impact of demographic characteristics, particularly in the cases of multiple subordination (Cole, 2009; Collins, 2019; Shields, 2008; Tatli and Özbilgin, 2012). However, intersecting characteristics create a complex structure of parameters that may differ in both value and direction (Blau, 1977). For example, the status of a disabled African American man results from the intersection of three characteristics (disability, race and gender). One characteristic (gender) is associated with higher status, while the other two (disability and race) are associated with lower status. Not only is the direction of these characteristics different, but their relative demographic status value is unknown. That is, is disability, gender or race more important in determining demographic status? Qualitative studies have often been employed to examine intersectionality (Shields, 2008). However, quantitative studies have not yet examined intersecting demographic categories in a reliable and generalisable way (Hall et al., 2019; Shields, 2008). Thus, the operationalisation of demographic status derived from multiple demographic characteristics constitutes an important future research opportunity.

It should also be noted that, in addition to demographic and positional inequality, other forms of inequality, such as wealth, class, caste and cultural capital (Tilly, 1998), may influence leader-member relationships. The degree to which an inequality is relevant in the workplace depends on the context. For example, this study focuses on intra-societal rather than global inequality. However, the occurrence of globalisation and the subsequent increase in multinational working relationships call for a better understanding of the effects of global inequality (Metcalfe and Woodhams, 2012; Riaz, 2015). To this end, Levy and Reiche (2018) highlight the importance of cultural capital in forming social hierarchies and networks within multinational organisations. Like demographic characteristics, cultural capital can serve as an instrument of status and power (Levy and Reiche, 2018). However, compared to demographic inequality, different factors may be involved in predicting the impact of cultural capital inequality on leader-member relationships. More specifically, the hierarchy of cultural capital may be global, but it can be altered depending on the country of the company’s headquarters (Levy and Reiche, 2018). Therefore, to improve our understanding of global cultural capital inequalities, future research can expand the proposed model by adding contextual factors such as headquarter location. More generally, exploring the effects of cultural capital inequality, as well as other forms of inequality, on leader-member relationships may lead to interesting future research opportunities.

Finally, I hope that my proposed model of demographically diverse leader-member dyads and LMX quality will help revitalise research on this important topic. As discussed in the preceding sections, research interest in demographically diverse leader-member dyads and LMX quality seems to have decreased following the mixed findings of studies based on similarity-attraction theory. This study offers an alternative approach to explore how inequality based on demographic characteristics and inequality based on the formal organisational hierarchy affect LMX quality. This new approach can generate interest among researchers in this important topic and help produce more consistent empirical results.