Procedural justice enactment as an instrument of position protection: The three-way interaction between leaders' power position stability, followers' warmth, and followers' competence

Studies have started to examine factors that explain when and why leaders enact procedural justice. However, these studies have not considered the idea that justice enactment can be a self-serving instrument for leaders. In this paper, we propose a threat-based tripartite model of procedural justice enactment . Specifically, we examine how leaders in unstable (vs. stable) power positions combine information from the two fundamental dimensions of person perception — that is, their perceptions of a follower's competence and warmth — to shape the level of procedural justice they enact toward the follower. In support of our model, the results of a multisource organizational field study and a laboratory experiment show that leaders in unstable power positions enact procedural justice, particularly toward followers whom they perceive as highly competent but low in warmth. We discuss our findings in light of their implications for the justice and leadership literatures.


| INTRODUCTION
Procedural justice, that is, the extent to which a leader upholds fair procedural rules such as voice, accuracy, and timeliness in the decision-making process (Leventhal, 1980;Thibaut & Walker, 1975;Tyler, 1988), is highly important for employees and organizations.
Employees value procedural justice because they view it as a moral imperative (Folger, 2001), it serves their instrumental purpose of achieving long-term personal goals (Leventhal, 1980;Thibaut & Walker, 1975), and it signals that employees are respected members of the collective (De Cremer & Tyler, 2005;Tyler & Lind, 1992).As a result, procedural justice enhances employee wellbeing, as indexed by increased job satisfaction, positive affect, and trust in the supervisor (Colquitt, Conlon, Wesson, Porter, & Ng, 2001).
Procedural justice is important to organizations, among other reasons because it stimulates organizational commitment, in-role performance, organizational citizenship behavior, and a fair organizational climate (Brebels, De Cremer, van Dijke, & Van Hiel, 2011;see Colquitt et al., 2013, for a meta-analysis).
Given that procedural justice entails beneficial consequences for employees and organizations, justice scholars suggest that it is important to understand factors that explain when and why leaders enact procedural justice toward their employees (Brockner, Wiesenfeld, Siegel, Bobocel, & Liu, 2015;Graso, Camps, Strah, & Brebels, 2020).However, to date, our understanding of such factors is limited at best.For this reason, researchers have begun to shift their focus from the perspective of the justice recipient to that of the justice-enacting authority, that is, the leader (Brockner et al., 2015;Graso et al., 2020).Consistent with the notion that a leader's central task is to serve collective goals (Day, 2001;Kaiser, Hogan, & Craig, 2008;Yukl, 2002), these studies have largely taken an "other-serving approach."For instance, studies have shown that leaders with characteristics that reflect or enhance other-serving motives such as agreeableness, conscientiousness (Mayer, Nishii, Schneider, & Goldstein, 2007), moral identity (Brebels et al., 2011), status (Blader & Chen, 2012), and empathy (Cornelis, van Hiel, De Cremer, & Mayer, 2013) are relatively likely to serve the needs of their followers by enacting procedural justice.Other studies suggest that leaders enact procedural justice if they perceive their followers to have needs for control and belongingness (Cornelis, van Hiel, & De Cremer, 2012;Cornelis et al., 2013;Hoogervorst, De Cremer, & van Dijke, 2013).Finally, research shows that leaders enact procedural justice to facilitate organizational effectiveness because it stimulates employee compliance (Scott, Garza, Conlon, & Kim, 2014;Zhao, Chen, & Brockner, 2015).
Although leaders' other-serving motives can facilitate procedural justice, this approach fails to consider that fair procedures may also be enacted instrumentally to serve the leaders' own needs (De Cremer & van Dijke, 2009;Graso et al., 2020;Lerner & Clayton, 2011;Qin, Ren, Zhang, & Johnson, 2018).Indeed, an abundance of studies show that leaders often act in self-interested ways (De Cremer & van Dijk, 2005;Maner & Mead, 2010;Stouten, De Cremer, & van Dijk, 2005;Rus, van Knippenberg, & Wisse, 2012; see Williams, 2014, for an overview).Therefore, a complete picture of procedural justice enactment requires studying how the enactment of procedural justice can serve the self-interests of leaders.
To address this research gap, we integrate the instrumental approach of justice enactment (De Cremer & van Dijke, 2009;Graso et al., 2020;Lerner & Clayton, 2011;Qin et al., 2018) with the person perception literature (Cuddy, Glick, & Beninger, 2011;Fiske, Cuddy, & Glick, 2007) and propose a threat-based tripartite model of procedural justice enactment.This model considers leaders' power position stability and their perceptions of followers' warmth and competence as three threat-related antecedents that are likely to trigger leaders' self-serving motives and procedural justice enactment.The variable that is perhaps most likely to make self-serving motives salient among leaders is the (in-)stability of their power position (Mead & Maner, 2012;Mooijman, van Dijk, van Dijk, & Ellemers, 2019;Williams, 2014;Wisse, Rus, Keller, & Sleebos, 2019).We argue that when leaders' power position is unstable, they will use procedural justice enactment as an instrument to protect their power position.Furthermore, the more threatening followers are perceived to be to the leaders' power position, the more leaders will exhibit procedural justice toward these followers.Leaders assess the positional threat posed by followers based on two fundamental dimensions of person perceptions: followers' competence and warmth (Tai, Narayanan, & McAllister, 2012;Yu, Duffy, & Tepper, 2018).Thus, we argue that followers' competence and warmth interact with leaders' power position stability to influence leaders' procedural justice enactment.Specifically, we propose that leaders whose power position is unstable are most likely to enact procedural justice toward followers who they perceive as competent but cold.Our proposed conceptual model is presented in Figure 1.
Thus, in line with the quote from Evan Simpson (1976)  Second, research on self-serving leadership has focused on how leaders' self-interests facilitate leaders' behaviors that are detrimental to followers (Wade, O'Reilly, & Pollock, 2006).By examining whether leaders' self-interests can foster procedural justice, which is seemingly beneficial to followers, our research broadens the way leadership scholars think about behavioral consequences of leader self-interests.A prominent antecedent of self-interested behavior, which may also facilitate the instrumental enactment of procedural justice by leaders, is position stability (Williams, 2014).The superior position over employees that leaders occupy confers to them valuable resources such as power, status, higher pay, and autonomy, thereby creating a pleasurable state for leaders (Anderson, Kraus, Galinsky, & Keltner, 2012;Magee & Galinsky, 2008;van Dijke & Poppe, 2006).However, leadership positions are not always stable (Mead & Maner, 2012;Tajfel, 1984;van Vugt, Hogan, & Kaiser, 2008).Given that people are motivated to maintain pleasurable states (Ryan & Deci, 2001;Young, 1952) and that a superior position creates such a state, leaders in unstable power positions are motivated to maintain their positions (Williams, 2014).Power position instability therefore shifts leaders' attention toward the protection of their positions and away from organizational goals (Jordan, Sivanathan, & Galinsky, 2011;Mead & Maner, 2012;van Vugt et al., 2008).
Leaders will be particularly sensitive to their position stability when followers are threatening.Threatening followers are those who are perceived as being highly competent yet lacking good intentions.
Research has identified two fundamental dimensions of person perception that people rely on to assess whether others are threatening: whether the other person has good intentions (i.e., warmth) and the ability to act upon these intentions (i.e., competence) (Cuddy, Glick, & Beninger, 2011;Fiske et al., 2007;Tai et al., 2012).Warmth refers to perceived characteristics, such as cordiality, helpfulness, and kindness.
Competence refers to a person's perceived skills, abilities, and knowledge to carry out these intentions (Cuddy et al., 2011;Fiske et al., 2007).Leaders arguably will see competent but cold followers as threats because these followers have the ability to carry out their bad intentions to challenge the leader's position.Indeed, previous studies have shown that people are more likely to see out-group individuals as threats when they perceive them as being competent and cold (Awale, Chan, & Ho, 2019).More directly relevant to our research, Tai et al. (2012) suggested that employees with high competence and low warmth are more likely to trigger leaders' envy toward employees.Similarly, Yu et al. (2018) found that when leaders perceive subordinates as being competent and cold, the self-esteem threat that leaders experience in the workplace is heightened.Thus, leaders interpret the level of threat that followers can pose to leaders' position by assessing whether competent followers actually have the intention to challenge them (i.e., based on the impression that they are cold).
Procedural justice is commonly regarded as a position-affirming instrument because it legitimizes the leader's position in the eyes of followers (Brockner et al., 2015;De Cremer & van Dijke, 2009;Tyler, 2006;van Knippenberg, van Knippenberg, & De Cremer, 2007).Specifically, by being consistent, accurate, unbiased, and timely in the decision-making process, the leader shows to subordinates that he/she is the appropriate person to occupy the superior position.This legitimacy fosters followers' voluntary compliance and cooperation with the leader (see Tyler, 2006, for a review).For instance, research shows that the enactment of procedural justice makes people regard legal authorities as more legitimate, and consequently, they are more willing to comply with these authorities and with the law in general (Sunshine & Tyler, 2003).Relatedly, procedural justice has been found to enhance tax authorities' legitimacy and consequently facilitate citizens' voluntary tax paying (Kirchler, Hoelzl, & Wahl, 2008).In the workplace, leaders' procedural justice legitimizes their positions and invites compliance from employees, especially from organizational rule breakers (Tyler, 2006;Zhao et al., 2015).Based on these findings, we argue that when power positions are unstable, leaders may be motivated to enact procedures in a fair way out of instrumental considerations to affirm and thus protect their positions.Specifically, we argue that leaders are most likely to instrumentally enact procedural justice toward followers who are perceived as a threat to their unstable power positionsthose who are perceived as competent and cold.
Specifically, we hypothesize the following: Hypothesis 1.An unstable (vs.stable) power position will lead to more procedural justice enactment, particularly toward a follower who is perceived as highly (vs.lowly) competent and at the same time low (vs.high) in warmth.
As argued earlier, our reasoning for the three-way interactive effect on procedural justice enactment is that leaders view procedural justice as an instrument that can stabilize their positions (Brockner et al., 2015;De Cremer & van Dijke, 2009;Tyler, 2006;van Knippenberg et al., 2007).Specifically, based on our threat-based tripartite model of procedural justice enactment, an unstable power position combined with a follower who is low in warmth and high in competence present leaders with the most severe positional threat.
Given that procedural justice is a position-affirming instrument that legitimizes leaders' positions in the eyes of followers, leaders are most likely to be motivated to enact procedures in a fair way out of this instrumental motive to affirm and protect their positions.Therefore, we make this process explicit and test for the mediating role of perceived instrumentality of procedural justice enactment in Hypothesis 2 as follows: Hypothesis 2. The three-way interactive effect of the stability of leaders' power position, their perceptions of a follower's competence, and their perceptions of this follower's warmth on procedural justice enactment is mediated by leaders' perceived instrumentality of procedural justice.
We tested our hypotheses in two studies.Study 1 was a multisource field study conducted in China in which we collected responses from leader-follower dyads.To bolster our confidence in the internal validity of our findings, Study 2 was a controlled laboratory study using an established managerial in-basket task (Treviño, 1992).We conducted this study in the Netherlands.et al., 2013;Judge, Scott, & Ilies, 2006).We restricted participation to individuals who, at the time of the survey, held a leadership position at work (i.e., supervising at least one follower).The research agency invited 318 leaders.These leaders completed a survey and were asked to indicate the name and the email address of the last subordinate they interacted with before working on the survey.In so doing, we ensured that choosing the focal subordinate would be random (Chun, Yammarino, Dionne, Sosik, & Moon, 2009).The research agency then sent the follower an email that directed the follower to an online link to the survey questions.
We received complete responses from 160 leader-follower dyads.

| Measures
Unless noted differently, we used 5-point response scales (1 = not at all, to 5 = completely).To ensure translation equivalence, two bilingual researchers separately translated the items from English to Chinese and back to English.Comparisons showed no discrepancies in the meaning of the items.

| Leader power position stability
We measured the leaders' power position stability with the three-item power-loss concern scale developed by Mooijman et al. (2019).The

| Perceived follower competence
To assess the leaders' perceptions of follower competence, we used a five-item scale developed by Fiske et al. (2002).This scale was developed to assess the competence dimension of social perception (Fiske, Cuddy, Glick, & Xu, 2002).It has been widely used to assess the perception of others' competence in previous studies (Cuddy, Fiske, & Glick, 2008).Sample items are "This employee is competent" and "This employee is intelligent" (Cronbach's α = .60).

| Perceived follower warmth
We assessed perceived follower warmth with eight items based on Fiske et al. (2002).Sample items are "This employee is agreeable" and "This employee is cooperative" (Cronbach's α = .66).

| Procedural justice enactment
The followers rated their leaders' procedural justice enactment using the five-item procedural justice scale developed by Brebels et al. (2011).The participants were requested to refer to "the procedures your immediate supervisor uses to make decisions about pay, rewards, evaluations, promotions, assignments, etc." Sample items are "Would your supervisor grant voice to you when making such decisions?"and "Would your supervisor take time to listen to your opinions when important decisions have to be made about you?" (Cronbach's α = .70).

| Control variables
We controlled for the leaders' other-serving motives.Specifically, the followers indicated to what extent their leaders make decisions about pay, rewards, evaluations, promotions, and assignments by considering the following statements.They rated four items developed for this study: "He/she truly cares about my well-being," "He/she takes my needs into consideration," "He/she cares about my satisfaction," and "He/she does things that are beneficial to me" (Cronbach's α = .65).

| Results
Table 1 shows the descriptive statistics and correlations between the study variables.We conducted a hierarchical regression analysis (Aiken & West, 1991).The interaction terms were based on the standardized scores of the independent variables.The results are shown in Table 2. Consistent with Hypothesis 1, we found a significant three-way interaction between the stability of the leader's power position, perceived follower competence, and perceived follower warmth (b = .08,p = .008). 1 Figure 2 visualizes this interaction.We estimated the simple interaction between the leaders' position stability and perceived follower competence at low (1 SD below the mean) and high (1 SD above the mean) levels of perceived follower warmth.As expected, power stability and perceived follower competence interacted when perceived follower warmth was low (b = −.14, p = .006)but not when perceived follower warmth was high (b = .02,p = .691).
Further simple slope analyses supported Hypothesis 1; when perceived follower warmth was low (1 SD below the mean), high (vs.low) power stability predicted lower procedural justice enactment when perceived follower competence was high (1 SD above the mean; b = −.23,p = .011)but not when perceived follower competence was low (1 SD below the mean; b = .04,p = .481).When perceived follower warmth was high (1 SD above the mean), power stability did not predict procedural justice enactment, regardless of whether perceived follower competence was high (1 SD above the mean; b = .08,p = .080)or low (1 SD below the mean; b = .04,p = .634).
In sum, based on responses from leaders and followers, Study 1 provides initial support for the threat-based tripartite model of procedural justice enactment which proposes that leader power position stability, perceived follower competence and warmth are three threat-related factors that interactively influence leaders' procedural justice enactment.These findings underscore the importance for organizations to be aware of threat-related antecedents that may promote leaders' procedural justice enactment.

| STUDY 2
We conducted Study 2 for two reasons.First, although the findings of Study 1 had high ecological validity, this study does not allow drawing causal conclusions.Study 2 was a laboratory experiment in a leadership context.Second, we tested the underlying process that explains why leaders' power position stability, perceived followers' competence, and perceived followers' warmth interact to influence procedural justice enactment.Following our reasoning, we tested for the mediating role of perceived instrumentality of procedural justice. 1 Without controlling for other-serving motives, the three-way interaction remained significant, b = .07,p = .029,and its shape was similar to the shape reported in the text.

| Procedure
In the laboratory, each participant was seated in a soundproof cubicle, and all the instructions were communicated via a personal computer.
To simulate actual workplace experiences, we adapted a managerial in-basket test, which is a simulation task that is often used to assess specific competencies of job applicants in selection procedures.This test combines high internal validity and some ecological validity as it simulates real workplace situations (Treviño, 1992).After reading the instructions for the in-basket test and the description of the organization where they were supposed to be working, each of the participants was placed in a situation in which they were the leader in charge of the selection process for a hypothetical organization's plant manager.At the beginning of the task, the participants had to first answer several questions on their leadership skills.They were told that these questions were to be used for determining their role (leader or follower) in the described situation.In reality, all the participants were assigned to the leader role.This step was necessary to ensure that the assignment of the leadership role was believable to the participants (Hoogervorst et al., 2013;Stouten & Tripp, 2009).The participants were placed in a company named "Duron Paints."They learned that they had just come back from a vacation and they had to reply to several emails.They then read the information about the plant manager selection plan.
The participants had to go through five emails in total.Three emails served to manipulate leader position stability, follower warmth, and follower competence.To reduce demand characteristics and prevent the participants from becoming suspicious about the actual purpose of the study, we presented these three emails along with two filler emails that contained neutral information.We first manipulated follower warmth (high vs. low) based on the definition of warmth, F I G U R E 2 Significant simple interaction of perceived follower competence and leaders' position stability when perceived follower warmth is low (upper panel) and when perceived follower warmth is high (lower panel) in Study 1 which refers to good intentions toward others (Cuddy et al., 2008).
The participants in the follower with high (/low) warmth conditions read the following: Hi (participant's name).During your holiday, I had a chance to get to know Jim, the assistant you recently hired, better.I must say that I gained a very favorable (/unfavorable) impression of him.During the weeks that I worked with Jim, my impression is that he is (/not) very benevolent to people, and he is (/not) concerned about others' needs and desires.He would not (/would not hesitate to) take hurtful actions against others if his own interests were threatened.I think he can (/cannot) be trusted to serve the interests of the team.Vincent.
Participants then read two potential responses that they could use to reply.To ensure that the responses would not influence the follower warmth manipulation, we used two neutral responses: "Thank you for the information!We will discuss this more in our meeting."And "I have read your email.I will think about this."The participants indicated their choice based on a 7-point scale (1 = not at all; 7 = completely).
We then manipulated leadership position stability (unstable vs. stable) based on similar procedures from previous studies (Case & Maner, 2014;Jordan et al., 2011;Maner & Mead, 2010).Specifically, instead of operationalizing position stability as a general feeling as in Study 1, we manipulated position stability as a threat from the specific follower, that is, Jim.The participants in the unstable power position condition read the following: Dear (participant's name After the participants indicated their preference for one of the two neutral responses on a 7-point scale (1 = not at all, 7 = completely), we manipulated follower competence with another email (Cuddy et al., 2008).Specifically, the participants in the high (/low) competence follower conditions read the following: Dear (participant's name).During your absence, I worked with the assistant you recently hired, Jim.He seems to be (/not be) very competent at work.I must say that I have no (/some) doubts about his level of competence.Peter.

| Measures
Unless noted otherwise, we used 7-point response scales (1 = not at all, to 7 = completely).After they had read each email, we asked the participants to answer two questions evaluating their understanding of the email message.After reading the email that manipulated follower warmth, the participants rated follower warmth with two items: "Jim is a person who can be trusted to care for the team" and "Jim is benevolent" (Cuddy et al., 2008; Cronbach's α = .63).After reading the email that manipulated leader power position stability, the participants rated two items: "It is uncertain whether you will keep your job as plant manager" and "Your role as a plant manager might be taken over by Jim" (reverse coded; Maner & Mead, 2010; Cronbach's α = .77).After reading the email that manipulated follower competence, the participants rated two items: "The assistant named Jim is competent" and "The assistant named Jim is capable" (Cuddy et al., 2008; Cronbach's α = .93).
Before the participants proceeded to the procedural justice mea- We measured procedural justice enactment with a four-item scale based on Colquitt (2001).Previous literature on selection procedural justice suggests that openness, treatment, and two-way communication are the most important dimensions of selection procedural justice (Bauer et al., 2001).Hence, we adopted four items that were directly relevant to these dimensions of procedural justice.The participants were asked to what extent they would exhibit the following behaviors toward Jim: "Would you let Jim express his views and feelings during such decision-making procedures?," "Would you let Jim have influence over the outcome arrived at by such procedures?,""Would you apply such decision-making procedures with respect to Jim free of bias," and "Would you apply such decision-making procedures with respect to Jim consistently?"(Cronbach's α = .65).
As the gender of the described follower was always male (i.e., Jim), we measured the participants' gender as a control variable.
Furthermore, the extent to which students have work experience may influence their decision in the in-basket task (Keys & Wolfe, 1990).To control for work experience, we asked the participants whether they were currently working (1 = no, 2 = yes) and how many months of work experience they had so far.

| Manipulation checks
To check whether the manipulation of follower warmth was successful, we conducted a one-way analysis of variance (ANOVA) with the follower warmth manipulation as the independent variable and the follower warmth manipulation check as the dependent variable.This analysis revealed a significant main effect of follower warmth, F (1,192) = 417.83,p < .001,η2 = .69.The participants in the high follower warmth condition perceived Jim as being warmer (M = 6.11,SD = 1.09) than did the participants in the low follower warmth condition did (M = 2.51, SD = 1.37).We did not include follower competence and power position stability in this analysis as independent variables because these factors were manipulated after we had checked the manipulation of follower warmth.
To check if the manipulation of leader position stability was successful, we conducted a follower warmth (low vs. high) × leader power position stability (unstable vs. stable) ANOVA on the leader power position stability manipulation check.This analysis revealed a significant main effect of leader power position stability, F(1,190) = 97.18,p < .001,η 2 = .34.The participants in the stable power position condition perceived their positions as more stable (M = 4.76, SD = 2.01) than did participants in the unstable power position condition did (M = 2.22, SD = 1.50).The effect of follower warmth, F(1,190) = .41,p = .52,η 2 = .002,and the interaction between follower warmth and leader position stability were not significant, F(1,190) = .007,p = .934,η 2 = .000.We did not include follower competence as a predictor in this analysis because it was manipulated after we had checked the manipulation of follower warmth.

| Hypotheses tests
We conducted a 2 (follower warmth: low vs. high) × 2 (leader power position stability: unstable vs. stable) × 2 (follower competence: low vs. high) analysis of covariance (ANCOVA), with gender and work experience as the control variables and procedural justice enactment as the dependent variable.The results are presented in Table 3.Most importantly, this analysis revealed a significant three-way interaction, F(1,183) = 5.09, p = .025, 2 η 2 = .03. Figure 3   for a similar procedure).A fully mediated moderation effect is supported if four conditions are met.(a) The interaction between the independent variable and the moderators is significantly related to the dependent variable (i.e., procedural justice enactment).(b) The interaction between the independent variable and the moderators (i.e., leader position stability, follower warmth, and follower competence) is significantly related to the mediator (i.e., perceived instrumentality).(c) After controlling for the mediator × moderator terms (i.e., perceived instrumentality × leader position stability, perceived instrumentality × follower warmth, and perceived instrumentality × follower competence) and other predictors, the mediator remains significantly related to the dependent variable.
(c) After controlling for the mediator, the effect of the interaction between the independent variable and the moderators on the dependent variable becomes nonsignificant.
Supporting Hypothesis 2, the results in Table 4 show that (a) procedural justice enactment was significantly predicted by the focal three-way interaction.(b) The focal three-way interaction also had a significant effect on perceived instrumentality, F(1,183) = 6.32, p = .013,η 2 = .03.The results in Table 4 show that (c) after controlling for the interactions among the mediator and moderators and other predictors, perceived instrumentality was positively related to procedural justice enactment (see Model 1), b = .17,p = .026.(d) After including the mediator and controlling for other two-way and threeway interactions and predictors, the three-way interaction effect on procedural justice enactment became nonsignificant (see Model 2), b = .13,p = .07,while perceived instrumentality significantly predicted procedural justice enactment, b = .13,p = .026.
To test the significance of the indirect effect, we used Hayes' PROCESS macro (Model 12; 5,000 bootstrap samples; Preacher & Hayes, 2008;Hayes, 2013).The PROCESS macro generated bias-corrected confidence intervals (CIs) for the indirect effect of the three-way interaction of leader position stability (unstable position = −1 vs. stable position = 1), follower competence (low competence = −1 vs. high competence = 1), and follower warmth (low warmth = −1 vs. high warmth = 1) on procedural justice enactment via perceived instrumentality.The results revealed that power position instability significantly facilitated participants' procedural justice enactment via perceived instrumentality when the follower was perceived to be highly competent but low in warmth: indirect effect = −.05,SE = .07,95% CI [−.16, −.01].There was no indirect effect of power position instability, via perceived instrumentality, on procedural justice enactment in the other three conditions (see Table 5).

| GENERAL DISCUSSION
We found in a multisource field study and an experimental study that leaders' power position stability, leaders' perceptions of followers' competence, and leaders' perceptions of followers' warmth interacted to influence leaders' procedural justice enactment.Specifically, unstable (vs.stable) power positions led to more procedural justice enactment toward followers who were perceived to be simultaneously high in competence and low in warmth.Furthermore, the results of the experimental study revealed that the perceived instrumentality of procedural justice mediated the focal three-way interaction.Importantly, we found that the threat-based tripartite model of procedural justice enactment holds whether leaders experience a general feeling of power position instability (Study 1) or they experience a direct positional threat from the specific follower (Study 2).Below, we discuss the implications and limitations of these findings and offer suggestions for future research.

| Theoretical implications
By proposing and validating the threat-based tripartite model of procedural justice enactment, our research contributes to the procedural justice research in two ways.First, we establish an instrumental model of procedural justice from an actor-focused perspective.Previous justice models have mainly focused on the perspective of justice recipients and suggest that one of the primary motives underlying their need for justice is the instrumental purpose of achieving long-term personal goals (Leventhal, 1980;Thibaut & Walker, 1975).However, to capture a complete picture of procedural justice enactment, it is important to also take an actor-focused perspective to uncover leaders' instrumental motives for procedural justice enactment.This helps justice scholars understand the factors that promote leaders' procedural justice enactment.Our research broadens the understanding of justice scholars about when and why leaders are motivated to enact procedural justice out of instrumental considerations.
Second, our study goes beyond normative models of procedural justice by examining when and why leaders may engage in procedural justice.Influential leadership theory has proposed and tested normative models that prescribe when leaders should act fairly to achieve leadership effectiveness (e.g., when leaders should involve employees in decision making procedure-e.g., depending on employee competence; Vroom & Jago, 1988).However, we know little about whether and when leaders actually engage in procedural justice.Hence, there is a dearth on descriptive models of procedural justice.This is an important limitation, given that the tension between employee involvement and more directive styles is a central struggle for many leaders (Hill, 2003).The threat-based tripartite model of procedural justice enactment provides evidence for the descriptive models of procedural justice.
On a broader note, our study also adds to our understanding of leadership dynamics.First, our study provides insights into a central leadership perspective, the notion of self-serving leadership, which examines leaders' behaviors that place the leader's own well-being and interests above the needs of both followers and the organization (Decoster, Stouten, Camps, & Tripp, 2014).Previous studies have mainly focused on self-serving behaviors that are detrimental to employees and organizations (e.g., misuse of departmental budget or of public funds for personal gain; Wade et al., 2006).In contrast, our findings suggest that the self-serving interests of leaders can also foster seemingly positive and unselfish leadership behaviors, such as procedural justice.Thus, our findings suggest that in order to more fully understand leaders' self-serving actions we need to go beyond phenotypically negative behaviors and consider seemingly positive behaviors and their underlying motives.
Second, the present results are consistent with prior studies showing that leaders' behaviors are influenced by followers' competence, particularly if the leader is in an unstable position (Georgesen & Harris, 2006;Maner & Mead, 2010;Mead & Maner, 2012).However, the present results also qualify this view in an important way.Indeed, they suggest that whether competent followers will actually be seen as a threat to the leader also depends on followers' warmth.This finding is crucial because it suggests that focusing on warmth or competence alone may result in somewhat narrow, incomplete predictions and that we need to examine both fundamental dimensions of person perception together to accurately understand leaders' reactions to their followers (Fiske et al., 2007).

| Practical implications
Existing research on how to promote leaders' procedural justice enactment suggests that organizations can promote leaders' procedural justice by emphasizing followers' needs and highlighting leaders' responsibilities to serve organizational goals (Cornelis et al., 2012;Cornelis et al., 2013;Hoogervorst et al., 2013).However, as shown in our threat-based tripartite model of procedural justice enactment, leaders' procedural justice enactment can also be facilitated by leaders' self-interests such as serving the instrumental purpose of stabilizing the leader's power position.This suggests that organizations could also promote procedural justice enactment among leaders by appealing to leaders' own needs and self-interests.For example, with an increasingly competitive job market, many organizational positions have become less stable in terms of long-term employment (Greenhalgh & Rosenblatt, 2010).As a result, leaders in higher positions will be motivated to protect their positions against those who pose a threat.Organizations could promote procedural justice enactment among these leaders by emphasizing its benefits in making followers, especially threatening followers, perceive them as legitimate and fitting for their managerial position.This may encourage leaders to deal with potential competitors in a more constructive way by enacting procedural justice.

| Limitations and future directions
Despite a number of contributions to the literature, the present set of studies also has limitations.One potential limitation that should be discussed is that, in line with the vast majority of studies on antecedents of justice enactment (see Graso et al., 2020 for an overview), we focused on the enactment of procedural justice.However, recent research has shown that instrumental motives also shape the enactment of distributive justice (Qin et al., 2018).Future research should examine other types of justice, such as interpersonal and informational justice, as an outcome of leaders' power position instability and followers' warmth and competence.
Our findings suggest that leaders are more likely to exhibit procedural justice toward subordinates who they perceive as competent and cold.This is because they see procedural justice as an instrument by which to stabilize their unstable positions.Ironically, social comparison theory suggests that differential treatment from leaders can make subordinates behave in destructive ways (Thau et al., 2007).For example, Thau, Aquino, and Poortvliet (2007) found that employees who strongly compare themselves with others are more likely to exhibit antisocial behaviors when they think they are unfairly treated.
It is thus possible that other subordinates react negatively to leaders' high level of procedural justice enactment toward subordinates who they perceive as competent and cold at the same time.Future research should focus more explicitly on the broader effects that procedural justice enactment targeted at a specific employee has on other employees.
We found that leaders exhibit higher levels of procedural justice toward followers who they perceive as competent and cold.This is because they see exhibiting procedural justice to these followers as being instrumental in stabilizing their positions.However, followers high in warmth may be perceived as "leader material" because they can gain status and respect from other employees (Fragale, Overbeck, & Neale, 2011).It is thus possible that under certain conditions, followers with high competence and warmth are perceived as threats by leaders.For example, in Study 2, the follower as the justice recipient in the experiment was a male.We found that the gender of participants had an effect on their procedural justice enactment.The person perception literature suggests that perceivers favor women who exhibit traits that conform to their traditional roles, such as warmth and kindness, while they are more likely to trigger negative reactions when they exhibit traits that are less traditional, such as competiveness and competence (Cikara & Fiske, 2009).It is

| CONCLUSION
In this paper, we explored why leaders exhibit procedural justice by examining a threat-based tripartite model of procedural justice enactment.This model proposes that the stability of leaders' power position and leaders' perceptions of followers' competence and warmth, are three threat-related factors that interactively influence leaders' use of procedural justice as a position-affirming instrument.We validated the model and found that leaders enact procedural justice in a particularly threatening situation-when they have unstable power positions and perceive followers as being high in competence but low in warmth.
in the beginning of this paper, our model implies that leaders do not pit self-interest against justice-they view justice enactment as a means to benefit themselves.The present research contributes to the justice literature in at least two ways: First, by examining a threat-based tripartite model of procedural justice enactment from an actor-focused perspective, our research provides a novel perspective that takes leaders' selfinterest into account.Our research thus broadens the understanding of justice scholars about why leaders act in procedurally just ways.
leaders were requested to indicate to what extent the statements describe their thoughts about their current position.The items were "I dread the possibility of being a subordinate," "I don't like losing my position," and "I fear losing my power position."The Cronbach's α value for this scale was .86.
visualizes this effect.As in Study 1, we tested the simple interaction between leader power position stability and follower competence at low and high levels of follower warmth.Consistent with Study 1, power position stability and follower competence interacted when follower warmth was low, F(1,183) = 3.91, p = .049,η 2 = .02,but not when follower warmth was high, F(1,183) = 1.37, p = .243,η 2 = .01.In line with our prediction, a series of contrast analyses showed that when the follower was low in warmth and high in competence, participants in the unstable power position condition (M = 4.62, SD = .95)exhibited more procedural justice than participants in the stable power position condition (M = 4.01, SD = 1.06),F(1,183) = 4.82, p = .029,η 2 = .03.However, when the follower was low in warmth and low in competence, the participants in the unstable power position condition (M = 3.92, SD = .65)did not show more procedural justice than the participants in the stable power position condition (M = 4.09, SD = 1.02),F(1,183) = .43,p = .512,η 2 = .002.When the follower was high in warmth and high in competence, the participants in the unstable power position condition (M = 4.51, SD = 1.21) did not exhibit more procedural justice than the participants in the stable power position condition (M = 4.81 SD = .72),F(1,183) = .71,p = .402,η 2 = .004.Similarly, when the follower was high in warmth and low in competence, the participants in the unstable power position condition (M = 4.76, SD = .86)did not show more procedural justice than the participants in the stable power position condition (M = 4.52, SD = 1.19),F(1,183) = .66,p = .417,η 2 = .004.Next, we sought to formally test the full moderated mediation hypothesis.Hypothesis 2 stated that leaders were more likely to enact procedural justice toward followers with high competence and low warmth because they viewed procedural justice as an instrument to stabilize their leader positions.The model was tested based on Muller, Judd, and Yzerbyt's (2005) procedure (see also Zhang & Zhou, 2014 F I G U R E 3 Significant simple interaction of perceived follower competence and leaders' position stability when perceived follower warmth is low (upper panel) and when perceived follower warmth is high (lower panel) in Study 2 possible that female followers who are competent and warm are perceived as threats to one's leadership positions.At the same time, it is also possible that female leaders perceive female followers who are competent and warm as less threatening(McColl-Kennedy & Anderson, 2005).Future research should examine boundary conditions under which followers with different combinations of warmth and competence are more likely to be perceived as threats to leadership positions.In addition, prior research suggests that procedural justice may be more effective in affirming a leader's position if followers perceive it as being sincere(De Cremer & van Dijke, 2009).Therefore, it would be interesting to test whether followers would be able to tell that the underlying purpose of procedural justice is to serve the leader's self-interest and whether they would then exhibit more or less compliance toward the leader.It is possible that perceived instrumentality can backfire and foster destructive follower behaviors.Since leaders enact procedural justice with the hope of stabilizing their positions, it is important to examine followers' perceptions and behaviors afterwards.Last, our research took an instrumental perspective to examine leaders' procedural justice enactment as an outcome of three threat-related variables: leaders' position stability, follower competence, and follower warmth.However, recent research suggests that different justice motives may interactively influence leaders' justice enactment.Qin et al. (2018) suggest that leaders' justice enactment is driven by two general motives: instrumental motives and value expressive motives.Value expressive motives suggest that leaders enact justice to serve a signaling function by communicating core values and beliefs.The authors found that value expressive motives and instrumental motives are orthogonal rather than opposing ends on one continuum.These motives interact to influence leaders' overall justice enactment.This finding suggests that leaders' justice motives may coexist.Although we have shown in Study 1 that the threat-based tripartite model holds after controlling for other-serving motives, future research should examine whether instrumental motives and other-serving motives are two orthogonal or competing mechanisms in facilitating procedural justice enactment.
Can you send me a confirmation that you have read this message and delete it afterwards?I want to keep this issue private.David.
).You are in charge of the selection process for another plant manager.As you know, we originally wanted to have two plant managers in the company, but I think it is fair to tell you upfront that we are not yet certain whether we want to keep two plant managers.We may only need to Dear (participant's name).You are in charge of the selection process of another plant manager.There is a strong candidate within your plant.This candidate is Jim, the assistant you recently hired.I think it is fair to tell you upfront that we want to keep both you and Jim as plant managers.It is impossible that Jim can take over your role.
T A B L E 5 Results of mediation analysis in Study 2Three-way interaction (PS × CO × WA) à perceived instrumentality à procedural justice enactment Note: N = 194.Low CO/WA refers to one standard deviation below the mean of CO/WA; high CO/WA refers to one standard deviation above the mean of CO/WA.Abbreviations: CO, follower competence; P MX , path from three-way interaction to perceived instrumentality; P YM , path from perceived instrumentality to procedural justice enactment; P YX , path from three-way interaction to procedural justice enactment; WA, follower warmth.