Our search focused on the top journals publishing empirical work on leadership and followership. More specifically, we selected journals by impact factor that were listed in the Top 20 of the ‘Management’ category in Web of Science. We eliminated all journals publishing only theoretical or review articles as well as out of scope journals, leaving us with only five journals: Academy of Management Journal, Journal of Applied Psychology, Journal of Management, Journal of Organizational Behavior, and The Leadership Quarterly. We added the journal Leadership because it frequently publishes manuscripts on followership with a more qualitative approach compared to the other five included journals.

We gathered all articles on July 8^th, 2022, using WebOfScience. We searched for manuscripts including combinations of the word “follow” in the text. We focused on articles published between 2017 and 2021 to reflect contemporary practice, resulting in an initial pool of 323 articles. The next step in our search was to include studies that covered any sort of followership constructs (e.g., attitudinal, emotional, cognitive, behavioral, traits). We decided to exclude articles using “followers” as a source to measure leadership perceptions (e.g., a direct report rating a manager leadership style). Both authors manually checked each article for inclusion, and this procedure left us with 137 articles containing 210 distinct followership constructs.¹

We decided to focus on eight descriptive variables based on our initial assessment of the followership literature.

Table 1, which reports all descriptive statistics from our review, shows that a majority of articles included were quantitative (N = 84; 61.3%). The remaining articles were theoretical or review papers (N = 48; 35%), and only a few were qualitative (N = 5; 3.6%). Attitudinal (N = 113 out of 311 coded constructs; 36.3% of all variables), perceptual behaviors (N = 69; 22.2%) and cognitive constructs (N = 53; 17.0%) represented the majority of followership constructs. Actual following behaviors only made up for a small fraction of all constructs (N = 18, 5.8%) along with emotional (N = 20, 6.4%), others (N = 13, 4.2%), and traits (N = 10, 3.2%).

Followership constructs mainly served as outcome-based constructs such as dependent (N = 153 out of 304 coded constructs, 50.3%) or mediator (N = 71, 23.4%) variables. In about one-fourth of the coded constructs, they served as explanatory variables, either as an independent (N = 47, 15.5%) or as a moderator (N = 33, 10.9%) variable. Our review indicates that a vast majority of followership variables were measured using scales (N = 166 out of 198; 83.8%), and the remaining measurements included (among others) experimental manipulations, case analysis, archival records, and behavioral coding. The data source for followership variables is mostly “followers” themselves (N = 163 out of 198 variables; 82.3%) with only a minority being their “leaders” (N = 35; 17.7%).

All “followers” in our reviewed studies referred to direct reports, individuals in low power positions, or individuals labelled as “followers” (i.e., role-based approach). That is, no studies explicitly investigated the emergence of followers from a process approach. Our review identified seven articles (Meinecke et al., 2017; Molenberghs et al., 2017; Stam et al., 2018; Boulu-Reshef et al., 2020; Gilani et al., 2020; Güntner et al., 2020; Van De Mieroop, 2020) that investigated followers who were partly operationalized as representing influenced individuals (i.e., a combination of role-based and process approaches). However, in all seven cases, followers were individuals who were already considered to be “follower” before the study took place and so did not emerge as followers during the social interaction under study. Out of these seven articles, only five included followers as having some sort of common or shared goals with the leader (Meinecke et al., 2017; Molenberghs et al., 2017; Boulu-Reshef et al., 2020; Gilani et al., 2020; Van De Mieroop, 2020).

Our review unequivocally shows that the role-based approach, which equates followers with direct reports or subordinates, dominates the study of followership (91.6% of the reviewed papers). The role-based approach is represented in mainstream followership streams such as implicit followership theories (Sy, 2010), followership role orientations (Carsten et al., 2010) and typologies of followership (Kelley, 1992, 2008; Kellerman, 2008; Chaleff, 2009). Carsten (2017) summarized this rank-based view, arguing that the field of “followership considers the skills, behaviors, and influence that individuals use while interacting with “higher-ups” in an effort to advance the mission of the organization” (p. 144). By extension, this approach considers that leaders are individuals in position of authority or power such as managers, supervisors, or bosses.²

However, many scholars in the leadership and followership literatures have criticized the essence of the role-based approach to study followership and leadership (e.g., Bedeian and Hunt, 2006; Stech, 2008; Derue, 2011; Ashford and Sitkin, 2019; Larsson and Nielsen, 2021; Bastardoz and Day, 2022). Interestingly, almost a century ago, Cowley (1928) already differentiated between a leader (“an individual who is moving in a particular direction and who succeeds in inducing others to follow after him,” p. 145) and a head[wo]man (“an individual who, because of ability or prestige, has attained to a position of headship,” p. 146). Unfortunately, the fields of leadership and followership have never seriously embraced this call to distinguish leaders from individuals having formal power. Instead, the role-based approach is ubiquitous and is further reinforced by mainstream representations of leaders as CEOs, country Presidents, or individuals with large communities of “followers” on social media.

A role-based approach conflicts with the view that leadership and followership consist in an influential process, and that an individual must influence another to emerge as a leader. Individuals in a low position of formal power are not de facto “followers” if they are not influenced (Bedeian and Hunt, 2006) or do not agree on shared goals (Bastardoz and Day, 2022). This misalignment between the conceptualization (i.e., who we say we study) and operationalization (i.e., who we actually study) of followers impedes the scientific progress of the field and limits its validity. Interestingly, a role-based operationalization is also inconsistent with dictionary definitions that consider followers as individuals “in the service of one another,” “that follows the opinions or teaching of another,” or “that imitates another” (https://www.merriam-webster.com/dictionary/follower).³

Our systematic review revealed a second major issue related to how the concept of followership is operationalized. Specifically, most of the variables focus on attitudes, cognitions, or perceptions of behaviors, rather than directly assessing behaviors themselves (see also a recent review of the followership field with similar results by Oc et al., 2023). This is an issue because a broader number of constructs refer to actual behaviors. This finding replicates what Banks et al. (in press) found in the leadership literature, and echoes (Baumeister et al., 2007) criticism of the psychology literature. This situation is inappropriate and threatens the validity of followership research because of three main reasons.

First, variables that are measured with scales such as attitudes, cognitions, and perceptions or evaluations of behaviors (79% of the variables studied in our review) are endogenous and will lead to biased and inconsistent estimates when treated as antecedents in an empirical model (Sajons, 2020; Bastardoz et al., 2023). Given the nature of the human cognition to fill blanks, categorize individuals, and rely on salient perceptual memories (Lord et al., 1978, 1984; Cronshaw and Lord, 1987), idiosyncratic unmeasured factors affect attitudes, cognitions, and perceptions. For example, perceptions of actual behaviors will be biased because each individual has their own sensemaking, implicit theories (i.e., prototype matching), and inferential process that filter their recollections of memory and events (Lord et al., 1978). Furthermore, omitted variables, reverse causality, and simultaneity frequently affect scale measures and thus introduce endogeneity and threats to causal inference (Antonakis et al., 2010, 2016; Wulff et al., 2023). As a case in point, Güntner et al. (2020) show how a subordinate’s behavior affects and is simultaneously affected by a management style, rendering any scale measures of these constructs endogenous and causally invalid. Furthermore, scales can hardly account for the dynamic and flexible nature of followership suggested by the process approach.

Second, the few concrete and actual behaviors uncovered in our review mainly concern behaviors of direct reports that do not fit with a process view of followership. We found some behaviors that do not refer to being influenced such as prohibitive or promotive voice (Kim and Toh, 2019), absenteeism (Nevicka et al., 2018) and free-riding (Boulu-Reshef et al., 2020). Some behaviors also focus on individual goals that inherently differ from shared goals such as perpetrated rudeness (Kluemper et al., 2019), resistance (Güntner et al., 2020), dissent (Uhl-Bien et al., 2014), proactive behaviors (Fuller et al., 2015), and independence (Carsten, 2017). All these behaviors are important and must be studied in their own right (e.g., see the large and blooming literature on employee voice; Morrison, 2014, 2023);⁴ however, these behaviors do not pertain to the followership field because they do not refer to individuals being influenced to reach a shared goal.

Third, major taxonomies of followership behaviors (or styles) (Kelley, 1992, 2008; Kellerman, 2008; Chaleff, 2009), which are in line with a role-based approach because they imply that employees’ behaviors are static and context-free, remain influential in the field. These taxonomies provide critical insights about subordinates’ effectiveness, engagement, and courage. However, these taxonomies are limited because they categorize all individuals in a very limited number of types (four in Chaleff, five in Kelley and Kellerman). Also, direct reports’ behaviors result from a myriad of factors (e.g., the direct report’s traits and motivation but also the supervisor’s characteristics and the organizational context; see Brown and Thornborrow, 1996) rendering them endogenous (Bastardoz and Van Vugt, 2019; Güntner et al., 2020). These taxonomies imply that managers have limited responsibility if they face bystander, alienated, or passive employees because it is the nature of individuals to behave as such; however, such reasoning is probably misleading because a majority of individuals become bystander, alienated, or passive in the presence of situations or individuals rendering them as such (e.g., a bad manager). Overall, we encourage followership researchers to refer more cautiously to these taxonomies.

This section indicates that the followership field mainly fails to study actual followership behaviors. We now discuss a way forward to account for following behaviors in line with a process approach.

Our review highlights a lack of studies investing the emergence of followership from individuals in positions of power. Due to its overreliance on the role-based approach, the field focuses almost exclusively on formal followership, that is, when a direct report follows a supervisor.⁵ However, formal followership and top-down leadership are not always desirable or required. Formal followership works relatively well in presence of a simple task or routine with clear performance standards and stable conditions (Lee and Edmondson, 2017). In such strong situations (Meyer et al., 2010), downward following offers little advantage. For instance, Wellman et al. (2020) found that an inverse pyramid-shaped hierarchy increases organizational performance when task variety is high (but not when task variety is low). When the task is complex and requires decentralized expertise, skills, or information, organizations may benefit from managers who follow their direct reports.

The idea that everyone – and particularly individuals higher-up in organizations – can emerge as a follower is not inherently new. Currently, whenever a process approach is applied to the study of leaders and followers, it generally zooms in on leadership constructs such as leader emergence or shared leadership (Acton et al., 2019; Hanna et al., 2021; Xu et al., 2021). Although many calls have been made to study informal leader emergence (Badura et al., 2021), non-designated leaders (Ashford and Sitkin, 2019) or pure leadership (Bastardoz and Day, 2022), few calls have been made to study its followership counterpart, that is, downward following.

To understand the emergence of downward following, we start by discussing managers’ traits and states, particularly their social constructions regarding individuals in positions of power. Managers high on the honesty-humility dimension (Lee and Ashton, 2004), and low on the dark triad (Paulhus and Williams, 2002) and social dominance (Pratto et al., 1994) are more likely to downward follow. Empirical evidence for instance suggests that humble CEOs are more likely to empower (Ou et al., 2014) and collaborate (Ou et al., 2018) with other individuals.

Because upward leading and downward following are associated with power relations, struggles, and tensions, it matters how managers construe and enact their formal role in an organization. Upward leading attempts may be rejected to protect one own’s turf (Mintzberg, 1985; Maner and Mead, 2010), particularly in climates where perceptions of power are salient (Galinsky et al., 2008). Thus, individuals construing power as a responsibility (rather than an opportunity; De Wit et al., 2017), having a personal sense of power (Sessions et al., 2020), and navigating stable power structures (Jordan et al., 2011; Feenstra et al., 2020; Gray et al., 2023) are less likely to see upward leading as a threat and are more likely to downward follow.

Furthermore, a social construction of employees as industrious individuals (Carsten et al., 2010; Whiteley et al., 2012) and a shared leadership structure schema (Derue and Ashford, 2010; Derue et al., 2015; Wellman et al., 2022) are two important characteristics that predict the emergence of downward following. Finally, we argue that managers who embrace to a certain degree a follower identity (Epitropaki et al., 2017) – even if only temporarily – will be more likely to downward follow because following is congruent with their identity.

When predicting the emergence of downward following from an employee or direct report perspective, we rely on the literatures on informal leader emergence to inform our conceptual model (e.g., Judge et al., 2002; Hanna et al., 2021). We recognize that this discussion is not fundamentally new; however, it is new to consider employees’ characteristics as antecedents of downward following. We distinguish here between employees’ traits, motives, states, and social constructions.

Important employees’ traits predicting the emergence of downward following in their managers include general intelligence (Taggar et al., 1999; Judge et al., 2004), personality dimensions such as extraversion, conscientiousness, or emotional stability (Judge et al., 2002), proactivity (Grant and Ashford, 2008), and self-monitoring (Zaccaro et al., 1991). Motivational factors include employees’ motivation to lead (Chan and Drasgow, 2001; Luria and Berson, 2013), need for achievement and power (Mcclelland, 1975), and need for competence and growth (Ryan and Deci, 2000). Critical employees’ states predicting managers’ downward following refer to direct reports having high self-efficacy (Kwok et al., 2021), high leader identity (Miscenko et al., 2017), being legitimate to take a leading role (Aime et al., 2014), and being prototypical of a leader (Derue et al., 2015). Employees’ social constructions include low leadership risk perceptions (Zhang et al., 2020) and a proactive social construction of the employee role (Carsten et al., 2010), which are antecedents of managers emerging as downward followers.

All these employees’ characteristics increase the likelihood of a direct report’s upward leadership. As indicated earlier, a direct report’s leading attempt – which results from the different traits, motives, states, and social constructions listed above – is a necessary condition for the emergence of downward followership. For instance, Klasmeier et al. (2022) found that managerial empowering behaviors (a proxy of downward following behaviors) result from employees’ influential acts and behaviors, which suggests that managers are unlikely to emerge as downward followers if direct reports do not step up as upward leaders.

We now discuss boundary conditions in the emergence of downward following, and more specifically the optimal conditions that favor its occurrence. Downward following is more likely to occur when employees have valuable resources to lead a specific task, situation, or project. For instance, direct reports who have expertise, skills, or knowledge to perform a specific task should lead this task (Aime et al., 2014). Also, when managers do not have sufficient time to appropriately lead a project, downward following is likely the best strategy. A low psychological distance, which may result from familiarity (Farh and Chen, 2018), interactions over time (Zhang et al., 2020), or similar status (Benson et al., 2016), may be necessary to build the required trust for downward followership to emerge. In conditions of high psychological distance, the situation is riskier for a manager because a direct report who leads may (intentionally or not) have misaligned goals. Also, a high psychological distance may prevent an accurate and timely collaboration between the manager and their direct report. Another dyadic characteristic pertains to the supervisor and direct reports sharing a relational or social identity (Sluss and Ashforth, 2007; Haslam et al., 2010). For instance, Oc et al. (2019) found that powerholders were more receptive to dissenting opinions and more open to ideas when they emerged from subordinates sharing a relational identity. Research on the social identity of leadership and followership (Platow et al., 2015; Steffens et al., 2021) suggests that a shared group or organizational identity is critical for downward following.

The kind of task or group project will also matter for the emergence of downward following. Managers should lead tasks and projects that have a strategic importance or may be subject to political behaviors such as rewards or promotion decisions (Ferris and Treadway, 2012). Similarly, a manager should lead tasks such as strategic work (e.g., translating the organizational strategy from the TMT at the micro level; Antonakis and House, 2014), goal setting (Ordóñez et al., 2009), or performance evaluations. In contrast, we expect more frequent downward following behaviors in situations such as implementing a group project requiring complex solutions, scanning the competitive environment, or performing creative tasks. Similarly, in the presence of long and large-scale projects, we expect the temporary emergence of downward following behaviors to be critical for the effectiveness of teams (Zhu et al., 2011). To summarize, downward followership will likely emerge for tasks or projects that are not too politically sensitive and require building on diverse information, expertise, or skills.

At the group level, we argue that downward following behaviors are more likely to emerge when the climate (i.e., group members’ shared perceptions of their work environment) is safe and allows for risk taking, when the hierarchy is not too salient, and when group members have an accurate transactive memory system. First, employees are more likely to influence managers in climates that are psychologically safe (Edmondson, 1999; Bradley et al., 2012), empowering (Seibert et al., 2004), or encouraging voice behaviors (Morrison et al., 2011). Such trustworthy climates foster risk taking behaviors, collaboration across group members, and shared decision making, which are all critically important for downward following. Second, downward followership and upward leadership will emerge when the hierarchy is not too salient, or at least when groups do not embrace an authority ranking model (Fiske, 1992). For instance, Klein et al. (2006) found that hierarchical environments prevent upward leading attempts; such context seems thus to preclude the emergence of downward following. Wellman (2017) suggests that formal hierarchical differentiation – when authority cues are salient – sends signals to group members that authority and formal power difference matters. We expect to observe a low likelihood of downward following in such environments. Third, group members (and managers in particular) need to have some appropriate meta-knowledge – in the form of an accurate transactive memory system - describing who is knowledgeable, expert, or skilled in a group (Lewis and Herndon, 2011). For instance, Xu et al. (2021) found that transactive memory systems were related to situationally-aligned leadership, which represents the emergence of leaders who fit the requirements of the situation. If managers cannot easily locate where (or whether) resources can be found in a team, we predict that managers will downward follow to a lesser extent.

Drawing on the distinction between omnibus and discrete contextual levels (Johns, 2006; Oc, 2018), we discuss how omnibus contextual factors moderate the emergence of downward following behaviors.⁸ We argue that slow changing business environments, low competitive industries, crises situations, and high-power distance as well as collectivist national cultures all represent factors limiting the emergence of downward following.

Certain business environmental conditions call for different sources of influence (i.e., leaders). For instance, fast changing, uncertain or very competitive business environments require adaptations, flexibility, and a high degree of collaboration to keep up with such environments (Brown and Eisenhardt, 1997). In contrast, slow changing, more stable, or monopolistic business environments offer clearer operational and performance standards that necessitate less downward following. Similarly, a crisis situation (at its onset) will limit the opportunity for downward following. During crisis periods, individuals look up to those in positions of power for certainty, direction, and a vision of the future (Bastardoz et al., 2022). Such tense situations do not call for downward followership, but rather formal leadership. Finally, managers or supervisors originating from national cultures high on the power distance or collectivism dimensions (Hofstede, 2011) are less likely to downward follow because such behaviors are not culturally expected; rather, managers are expected to “lead” and provide directions to the group (Blair and Bligh, 2018). Empirical evidence suggests that low power distance cultures relate positively to voice behaviors (Botero and Van Dyne, 2009) and empowering leadership (c.f., Sharma and Kirkman, 2015).