Impact of Abusive Supervision on Work Engagement: Mediation by Organizational Justice and Moderation by Resilience

1. Ph. D Scholar, Air University School of Management, Air University, Islamabad, Pakistan 2. Ph. D Scholar, Riphah School of Leadership, Riphah International University, Rawalpindi, Punjab, Pakistan 3. Assistant Professor, Hailey College of Banking and Finance, University of the Punjab, Lahore, Pakistan PAPER INFO ABSTRACT Received: April 07, 2020 Accepted: June 15, 2020 Online: June 30, 2020 The study investigated the mechanism through which abusive supervision impacts employee work engagement. Organizational justice was tested as mediator and resilience as moderator over the aforementioned relationship. A sample of 396 employees was contacted from 3S (Sales, Spare Parts and Services) and 2S (Spare Parts and Services) dealerships of automobiles sector of Pakistan using convenience sampling, time-lag technique and structured questionnaires and was analyzed by using SPSS and Smart PLS software. Findings support the relationships as hypothesized in the study. Current study contributed to the destructive leadership and organizational justice literature by highlighting the mechanism behind negative outcomes of abusive supervision relating to employees work engagement and organizational justice. Managers, therefore, need to take measures for discouraging abusive supervision at the work place and should find ways to develop resilience in the employees. In the end the limitations of the study and recommendations were also discussed.


Introduction
Employee-organization relationship is subject to a continuous change owing to the modern economic challenges (Baker et al., 2008) based on cut-throat competition and characterized by technology and innovation (Eldor & Harpaz, 2016). This changing relationship has brought the conventional concept of employee's performance into question (Islam et al., 2016). Modern day researchers are shifting their focus of studies from employee's proficiency to employee's level of commitment and engagement (Grefen et al., 2007). Employee performance is multifaceted which comprises of within-role, extra-role and anti-role behaviors (Elaine, Leslie de & Isabel, 2011). Employees performance is affected the behavior and attitude by the manager they are facing. Hogan & Kaiser (2005) studied destructive leadership behaviors (DLB) and found that employees perceive DLB to be a cause of their dissatisfaction at the jobs. Kirrane, Kilroy, & O'Connor(2019) count abusive supervision in the list of destructive leadership behaviors. They found abusive supervision, being a DLB, to be a cause of employees' disengagements. Abusive supervision is hostile behavior by supervisors towards their subordinates. Verbal abuses and emotional aggression by supervisors, excluding physical involvement, is counted among the abusive behaviors (Tepper, 2000). Abusive supervisors mistreat and mislead their subordinates by taking undue advantage of their powers and position (Ashforth & Vikas, 2003). They yell names, pass derogatory remarks, are found taunting, screaming on others and not recognizing the contributions by others (Keashly, 1998).
Work engagement is an important work-attitude that is impacted by abusive supervision (Wanget. al., 2020). It is gaining much of the attention of the researchers studying organizational psychology in the present times (Sonnentag, 2011). It is a concept that defines positive state of mind linked with an enjoyment of the functions performed (Schaufeli and Bakker, 2001). When employees are engaged in their work their mental and physical health also improves and their benefit to the organization also increases in form of improved performance and commitment.
Impact of abusive supervision has also been studied using organizational justice as mediator (Tepper, 2001), on different organizational and individual outcomes. Present study is also using organizational justice as a mediator among the relationship between abusive supervision and work engagement. Organizational justice is an important factor in understanding employees' response to abusive supervision (Morrison, 2014). It is well known that fairness perception enables employees to poise a positive approach and engage profoundly in their work considering themselves to be an important part of the organizational system (Crawshaw, et al., 2013).
Employees when face abusive behaviors instinctively try to find out the cause and the perpetrator of that behavior. According to Mikula, (2003) employees' subsequent response to decide the perpetrator of the abuse is explained by "the attribution-of-blame model of fairness theory for judging injustice". The attribution-ofblame model has three dimensions, "causation, control and intention". When organizations are believed to produce abusive agents the dimensions of intention and control are evoked that produces injustice fairness perceptions among subordinates (Wang & Jiang, 2015). Kim, Lee and Yun (2016) demonstrate that perception of negative effects of abusive supervision will differ individual to individual, depending on their personality traits and motives. They have studied the moderating effect of employees' personal characteristics in relation with abusive supervision and its outcomes. Employees' reaction to supervisor's behavior also depends on the personalities of employees. Subordinates with positive personality traits like agreeableness and conscientiousness are found to be less resistive to the abusive behavior of supervisor, weakening the relationship between abusive supervision and its outcomes (Tepper et al., 2001), whereas negative personality traits are found to be strengthening the aforesaid relationship (Mitchell & Ambrose, 2007).
One of the important personality characteristics is resilience that also gives one the psychological strength to handle stress (Luthans et al., 2007a). The subordinates being more resilient will be able to retain their emotional state than those with lower resilience. The bouncing back after facing an abusive situation makes them more strong, determined and ready for the next situation (Luthans et al., 2007a). It is found to be moderating the relationship between discretionary behaviors, knowledge sharing and abusive supervision. Fairness theory's postulate emphasizing the role of accountability in determining injustice (Folger &Cropanzano, 1998 brings attention to understand the responding employees' attribution styles and personalities characteristics. Abused employees with different motives and different personalities may respond differently to abusive supervision and their sense of organizational justice will also be different determining their work attitudes. One of the personality attributes that is studied here, as a moderator, is resilience.Present study is trying to highlight the impacts of abusive supervision on employees' work engagement through the mediation of organizational justice and moderation of resilience.Theoretical Framework

Abusive Supervision and Work Engagement
Work engagement is a positive work-related mindset that consists of vigor (having high energy and mental ability to bounce back while working), dedication (accepting challenge at work and feel enthusiastic about work) and absorption (mentally engrossed in the work) (Schaufeli, Salanova, Gonzalez-Roma, & Bakker, 2002). Although researchers have studied effects of many variables like personality, job enrichment, leadership styles etc. in determining employees' work engagement (Christian, Garza, & Slaughter, 2011) but there is dearth of studies examining the negative behaviors of leaders like abusive supervision on work engagement. Leaders' behaviors greatly affect employees' work attitudes (Li et al., 2018). Leaders' with positive behaviors influence employees' work attitudes positively and negative behaviors of leaders influence work attitudes of employees negatively (Li et al., 2018) especially work engagement. They further say that it would be valuable research to integrate leadership styles and employees' emotional response as leadership style greatly influences subordinates mental state and potential development (Oldham & Cummings, 1996). The relationship between abusive supervision and work engagement can be described by understanding leadership theory. Manager-employee relationship consists of a mutual transactional process (Bass & Avolio, 1993). Managers try to motivate subordinates to engage in their work by creating a task-oriented environment (Bass & Avolio, 1993).
To build up link among work engagement and abusive supervision Kahn's (1990) model of engagement is very useful. Kahan suggested that if employees feel valued, worthwhile, meaningful, safe, less fearful of negative outcomes and resourceful then they are work engaged. All these feelings will vanish if they perceive their supervisor behavior to be a stressor. There is a little evidence of direct impact of abusive supervision on work engagement. But researchers have studied the relationship between abusive supervision and emotional exhaustion (Breaux et al., 2008;Wu & Hu, 2009) and psychological distress (Restubog, Scott, &Zagenczyk, 2011). Tepper (2000) discussed the outcomes of abusive supervision more thoroughly than the studies carried out before. He found abusive supervision and work commitments to have a strong negative relationship. Work engagement being a very closely related concept with job commitment may have the same negative impacts of abusive supervision as in case of work commitment. When commitment decreases then vigor also decreases, and we also know that commitment to work gets a worker to get involved in the job and it also increases a workers' absorption level in the job (Wefald& Downey, 2009). Tepper's results of his 2000 study show that increase in abusive supervision results in decrease in work commitment of the subordinates. Above discussion brings us to the conclusion that abusive supervision may also decrease work engagement of subordinates. So, it is hypothesized here as:

Organizational Justice as Mediator
In his doctoral thesis Eib et al., (2014) highlighted the need of studying justice by considering the actors of justice (supervisors, managers etc.). He wrote that most of the justice research focused the receiver of the justice i.e. subordinates not the individuals who are responsible for keeping a just environment, the authority figures like managers. They suggested that there can be an inner ethical environment among individuals that make them feel good not only when they receive just treatment but also when they deliver justice. It also that enacting justice becomes difficult in certain situations and for certain individuals. There is a scant of literature available discussing the justice actors than the justice receivers. If actors of justice are included in the study greater insight into the causes and determinants of organizational justice is possible. The mangers who enact fairness are of stable personalities and characteristics such as caring attitude and moral obligation to treat subordinates like they treat themselves (Brebels et al., 2011;Patient & Skarlicki, 2010). Abusive supervisors display attitudes that are not caring at all and that are far away from moral obligations to treat subordinates as they treat themselves. Abusive supervisor can be termed as injustice actor than justice actor. It implies that he enhances injustice perceptions of employees resulting in negative outcomes.
Work engagement is closely related with the concept of commitment. It has been found that organizational justice mediates the relationship between abusive supervision and organizational commitment of subordinates (Aryee et al., 2007;Alisher et al., 2016). Owing to mediation process by organizational justice it can thus be hypothesized that:

H1:
Organizational justice mediates the relationship between work engagement of employees and their perceptions of abusive supervision so that abusive supervision decreases the perception of justice of employees that further decreases employees work engagement.

Resilience as Moderator
Although research on outcomes and antecedents of resilience is not much but there is also a need to study its intervening and moderating impacts. At work-place there can be different constructive or destructive behaviors shown by supervisors and leaders so it would be worthwhile to study these behaviors under the impact of resilience.
Various studies have found that resilient employees maintained their health, performance, happiness etc. even after facing many negative factors like downsizing (Maddi, 1987). Even in the studies of the last decades, Luthans et al. (2005) found a significant relationship between the performance of employees and change/transformation and Larson & Luthans (2006) found resilience having significant relationship with job satisfaction of factory workers. Similarly, Luthans & Youssef (2007b) found that employees' level of resilience is significantly related with employees' commitment, satisfaction and happiness.
The bouncing back from every adversity and stress makes resilient individual to cope adverse situations successfully as well. Smith et al., (2008) calls resilience to recover from stress and every new recovery makes the resilient more immune to the new stress situation. This immunity makes resilient individuals feel less stressed in the adverse situations than the individuals with low resilience (Connor & Davidson, 2003). It is conceptualized as a positive resource capacity builder (Luthans, 2002). Instead of feeling stressed and outcast resilient individuals flourish in the times of adversity that may be caused by feeling of increased responsibility (Christensen &Knardahl 2010).
Work engagement is a motivational state employees' well-being relating to their work (Baker et al., 2008). It is associated with employee performance at work (Halbeselben & wheeler, 2008). This state of engagement varies from person to person and situation to situation especially under the influence of certain situational factors like PsyCap (Xanthopoulou et al., 2009). COR theory presents the understanding to establish the link between PsyCap and work engagement. It is a motivational process that involves positive work behaviors. Workers positive personality traits work as a spur to elevate the levels of work engagements of the employees. Mostly this impact is indirect rather than direct. The individual having higher PsyCap do not let hostile situations demotivate them rather take them as challenges and their personality characteristics refine after passing through every adverse situation. It's a sort of building resources that help employees show positive work behaviors like work engagement (Guido et al., 2018).
In past researches positive outcomes of resilience have been studied in the past decade. Being a positive attribute and quality resilience can shed negative impacts of different factors. Present study focused on negative impact of abusive supervision on employees' work engagement, so it was worthwhile to study this impact under the influence of resilience. It was anticipated also that resilience will also weaken the impact of abusive supervision on employees' injustice perception. So, it can be hypothesized here that:

H2.
Resilience moderates the relationship between work engagement and abusive supervision.

H3.
Resilience moderates the relationship between organizational justice and abusive supervision such that this relationship becomes weak for higher levels of resilience and strong for lower levels of resilience.

Material and Methods
Study setting of present study is field setting for it is non-contrived with minimal researcher interference, as respondents were approached in their natural work settings, without disturbing their routine operations. Respondents (Employees of 2S& 3S automobile dealerships of Pakistan) were asked to spare some of their time to fill the research questionnaires. Every individual employee was asked to fill in the research questionnaire separately. Time distance for the study is cross-sectional, whereas data were collected using the Time Lag technique.
All the scales used are adopted. Present study used Tepper (2000) Data were collected by approaching the dealers in person. The managements of the dealerships were contacted and some 630 questionnaires were distributed in wave I. Some 574 questionnaires were returned back out of which 568 were useable and complete. After a gap of 15 days those questionnaires were returned back in wave II and this time 480 questionnaires were returned back out of which 475 were useable. After another lapse of 15 days those 475 questionnaires were distributed back and only 405 questionnaires were received back out of which only 396 were found complete and useable for further analyses. Thus, providing the response rate of 62.86 %.

Results and Discussion
After collection of data all the questionnaires were thoroughly checked for correction of errors and to ensure highest standards of quality of data. All questionnaires were coded and then entered in the SPSS for further analyses. Statistical Package for Social Sciences (SPSS 21.0) and SmartPLS 3 were used for analyses (CFA, regression, mediation and moderation analyses).
The table 1 below shows CR and DV for the study constructs. The acceptable value range for Cronbach's Alpha is more than 0.70 as accepted by SmartPLS. In the case of present study, the Cronbach's Alpha values for study variables are higher than 0.70 implies the reliability of the instrument. Other values as given in the table below are also in the acceptable ranges.

Discriminant Validity
The Fornell-Larcker criterion is one of the common approaches to assess the discriminant validity as shown in Table 2 diagonal elements (bold). The table 2 below represents the values of discriminant validity of every variable of the study.

Model Testing
Before proceeding with the main model testing, it was thought better to identify the control variables that may produce confounded results that decrease explanatory power of the model (Pedhazur & Schmelkin, 1991).

Control Variables
Present studied tested the demographic characteristics of the respondents as the variables that may confound the results of the model. ANOVA has been used to find out the impacts of demographic characteristics to determine whether to use them as control variables or not. If the impact of demographic characteristics is found to be significant then that variable is used as control variable otherwise not. The results of all the control variables show that they have insignificant impact on dependent variable work engagement. As the impacts of demographic characteristics on dependent variable of the study are not significant so there is no need to control them while testing the hypothesized relationships. Hypotheses Testing

Model: The Relationships between Variables of the Study
The figure below represents the relationships as hypothesized in the study. It represents the direct relationships between employees' discretionary behaviors, work engagement and abusive supervision, the mediation of organizational justice between independent and dependent variables and moderation by resilience over the relationships between independent, dependent and mediating variables.

Mediation by Organizational Justice among the relationship between abusive supervision and employee Work Engagement
The results of mediation show that being an independent variable abusive supervision impacts organizational justice significantly. Whereas, organizational justice has also significant impact on employee work engagement and in the presence of mediator employee work engagement is also impacted significantly by abusive supervision. The indirect effect, which matters here to establish the state of hypothesis being accepted or rejected, issignificantly (p=0.001) and positively (β =0.108) related to employee work engagement. Hence our finding supports mediation of organizational justice among abusive supervision and work engagement. Hypothesis H1 is accepted. The results are given in the below table.

Moderation by Resilience over the Relation among Work Engagement and Abusive Supervision
The figure representing model of the study shows the moderation effect of resilience over the relationship between abusive supervision and Employee work engagement. The result shows significant impact of abusive supervision on organizational justice. Resilience insignificantly impacts organizational justice with coefficient value of 1.288, and 0.396 as p-value. Moderating effect (abusive supervision x resilience) with 0.000 p-value-.0179 coefficient value impacts organizational justice significantly. If moderating effect has significant impact on dependent variable then the moderation by the moderator is established. In the present case all results suggest the moderation by resilience to support hypothesis H2. Resilience moderates the relation among work engagement and abusive supervision. The table above shows effect size of moderator (F 2 ) for moderation by resilience. According to the studies by Hair et al., (2017), Kenny et al., (2016) and Aguinis et al., (2005) the realistic moderation effect size values can be as low as 0.005, medium values can be 0.01 and large values can be 0.025 or above. The moderating effect size value of resilience on employee work engagement here is 0.052 which is a high effect size. It implies that resilience contributes much in explaining the influence of abusive supervision on work engagement.

Moderation by Resilience over the Relation among Organizational Justice and Abusive Supervision
The figure representing the tested model of the study above also shows the moderation effect of resilience between abusive supervision and Organizational Justice. The result shows significant impact of abusive supervision on organizational justice with a coefficient value of 0.450 and p-value equal to 0.000. Resilience also significantly impacts organizational justice with coefficient value of 0.120, and 0.037 as p-value. Moderating effect (abusive supervision x resilience) with 0.000 p-value-.0136 coefficient value impacts organizational justice significantly. If moderating effect has significant impact on dependent variable then the moderation by the moderator is also established. In the present case all results suggest the moderation by resilience to support hypothesis H3. Resilience moderates the relation among organizational justice and abusive supervision. The table above shows effect size of moderator (F 2 ) for moderation by resilience. The moderating effect size value of resilience on organizational justice here is 0.031 which is a high effect size. It implies that resilience contributes much in explaining the influence of abusive supervision on organizational justice.

Discussion
Present study examined the negative influence of abusive supervision on employee work engagement. Mediating mechanism of organizational justice in the relationships between abusive supervision and employee work engagement was assessed. Moreover, moderating roles played by resilience were investigated in the relationships between abusive supervision work engagements, as well as between abusive supervision and organizational justice, which were all tested to be accepted as hypothesized.
First hypothesis of the study (H1) abusive supervision is negatively related to employee work engagement. It was found out that abusive supervision significantly decreases employee work engagement. This finding is consistent with the findings of previous studies. The employees under the influence of abusive supervision feel insecure and humiliated (Tepper, 2006) that disengages them from their work. It develops a sense of helplessness in the subordinates (Ashforth, 2003) and reduces job engagement (Zhang, et al., 2011). Hypothesis H2 states that Organizational justice mediates the relationship between work engagement of employees and their perceptions of abusive supervision so that abusive supervision decreases the perception of justice of employees that further decreases employees work engagement. Work engagement is mostly studied in the context of job demand-resource model (Demerouti et al., 2001) and Organizational positive activities ). There are other factors that may have more influence on employees getting engaged one of them is individual differences (Esakhani, 2016). These are the individual differences that help some workers thrive in their work in the same organization where some other workers get disengaged. Work engagement can be characterized in terms of individual differences rather than organizational motivations and policies (Esakhani, 2016) but Organizational justice is perhaps the only organizational factor that have more effects in making employees engaged. Where the discussion is of external factors like organizational justice and abusive supervision there work engagement also depends on internal factors like personality of the employees as well.
Hypotheses H2 and H3state that the relationship between employee work engagements and abusive supervision and the relationship between organizational justice and abusive supervision are moderated by resilience in such a way that these relationships become weaker for higher levels of resilience and stronger for lower levels of resilience. Results showed that resilience moderated these relationships. These findings are consistent with the findings of previous studies ofKaratepe&Karadas (2015) and Utsey et al., (2008). Abusive supervision is a severe workplace stressor, faced by employees in an organization (Tepper, 2001). So, if employees are resilient the adverse effects of abusive supervision can be averted rather abusive supervision will make them more resilient and maintain a positive attitude. Fredrickson et al., (2008) stated that resilience makes employees more proactive in facing the adversity. It lessens the tensions and decreases the stresses caused by adverse environment. This managing stressful situation is due to utilization of their psychological resources.

Conclusion
Present research work tried to unearth the negative outcomes of leaders' destructive behavior and successfully explained the phenomenon. Negative impacts of abusive supervision on employee work engagement were investigated. Work engagement and abusive supervision were found to be related as were stated in the objectives. Abusive supervision gets the employees disengaged from work and hinders their creative process.
The mediations of organizational justice among the relationships between employees have been established. Abusive supervision, being a stressor, damages perceptions of organizational justice that ultimately affects employee work engagement. Resilience proved to be a significant moderator in the relationships between employee work engagement and abusive supervision. Similarly, the moderation done by resilience over the relationship between organizational justice and abusive supervision is significant that proves moderation by resilience.
Management must pay significant attention to build up a cordial employeemanager relationship, owing to the importance of this relationship in achieving organizational objective. The results suggest that it is the duty of organizations to teach leadership interventions to supervisors so that the incidents of abusive supervision may be minimized. There are always certain limitations in every research work as there can be no perfect research. Present study has also several limitations, mentioning them here will surely increase reader's understanding of the work. Cross-sectional design of the study is another limitation. Owing to little resources and time management issues data were collected by using cross-sectional design but future research can be carried out by using longitudinal design for data collection thus clarifying the casual direction between abusive supervision and its outcomes.