Abusive Supervision, Psychological Capital, and Turnover Intention: Evidence from Factory Workers in China / Supervisión abusiva, capital psicológico e intención de dejar el empleo: estudio de la situación de trabajadores de fábrica en China

Seo, Yumi; Chung, Sun Wook

doi:https://doi.org/10.7202/1062088ar

Introduction

The dark side of organizational behaviour has received significant scholarly attention in the past decades, along with increasing occurrence of a variety of bad behaviour in the workplace (Aryee et al., 2008; Burton and Hoobler, 2006). Numerous terms and constructs have been employed to analyze this phenomenon, such as emotional abuse (Keashly, 1997), workplace aggression (Neuman and Baron, 1998), incivility (Andersson and Pearson, 1999), bullying (Einarsen, 2000), abusive supervision (Tepper, 2000), and social undermining (Duffy et al., 2002). In particular, abusive supervision (AS), which is defined as “subordinates’ perceptions of the extent to which supervisors engage in the sustained display of hostile verbal and nonverbal behaviors, excluding physical contact” (Tepper, 2000: 178), has been widely examined in recent years (Harris et al., 2007; Tepper et al., 2017; Zhang and Liu, 2018).

Scholars have shown that abusive supervision has negative consequences on subordinates, organizations and society. Above all, abusive supervision is closely associated with the psychological, physical and family well-being of subordinates (Tepper et al., 2006; Aryee et al., 2008; Rafferty et al., 2010; Hoobler and Brass, 2006). Extant research has also demonstrated that abusive supervision exerts negative effects on work-related attitudes, behaviours and performance (Zellars et al., 2002; Harris et al., 2011; Mitchell and Ambrose, 2012). At the societal level, the negative results arising from supervisory abuses, for example, absenteeism, medical costs, and lower productivity, are salient (Tepper et al., 2006). In addition, scholars have identified a wide range of antecedents of abusive behaviour, not only individual traits of the targeted and the aggressor (e.g., narcissism, neuroticism, or leadership style) but, also, organizational and familial factors (Tepper et al., 2006; Brees et al., 2016; Mawritz et al., 2014; Courtright et al., 2016). Furthermore, a growing number of studies has sought to explore the underlying mechanism that explains how abusive supervision causes detrimental outcomes. Scholars have found that various factors (e.g., social exchange, ego depletion, affective events, and cultural differences) mediated or moderated the relationship between abusive supervision and its consequences (e.g., Tepper et al., 2008; Thau and Mitchell, 2010; Lian et al., 2012; Wheeler et al., 2013).

Building on these recent developments in AS research, the present study seeks to answer the research question on how AS affects individual employees psychologically so that they can show voluntary turnover intention. This study broadens the understanding of abusive supervision in three major ways. First, the present study enriches the extant research on AS by exploring the underlying psychological mechanism between AS and turnover intention, specifically by focusing on the mediational role of psychological capital (PsyCap) (Luthans et al., 2007) based on the conservation of resources theory (Hobfoll, 2002). We examine the individual-level mediating mechanism that explains how AS works as a resource draining behaviour that may reduce employees’ personal resources represented by PsyCap. According to the conservation of resources (COR) theory (Hobfoll, 2002), individuals are motivated to protect their current resources and acquire new resources when clear resource loss is experienced. By explaining the process of how AS can deplete individuals’ resources, which leads to protective behaviour and attitudes, we attempt to integrate COR theory into the existing AS literature.

Second, we also examine the moderating role of broader organizational contexts represented by organizational justice perception in the relationship between AS and turnover intention, including its mediating mechanism. The construct of AS was originally delineated in organizational justice literature (Tepper, 2000). In this literature it was argued that perceptions of procedural and distributive justice moderate detrimental effect of AS on PsyCap and turnover intention. Therefore, we examine the boundary conditions where the effects of AS can be amplified with regard to overall organizational justice perception.

The third important contribution of this paper is to introduce the unique Chinese context and test our model in it. Our study, which is based on young factory workers in China’s electronics industry, deserves close attention in two respects. There has been a growing interest in the changing demography of the workforce in China, which is well known as a world factory. The second-generation workers, who were born during the 1980s and 1990s, form more than 60 percent of the Chinese workforce, in contrast with previous generations (Chang, 2009; Harney, 2008; Chan and Pun, 2009). These young workers, who are mostly the only child in the family due to the government’s child policy, are better educated and more autonomous than their parents, and focus on their own lives. In the workplace, they have higher expectations, express their opinions willingly, and are ready to move to other factories, which is far different from their parents’ generation who obeyed managers and endured all kinds of hardships for subsistence wages (Pun and Lu, 2010; Cooke et al., 2016; Lyddon et al., 2015). In this regard, it would be interesting to examine how this new generation of workers interprets abusive supervisory behaviour in the workplace.

The Chinese context deserves further attention from a macro standpoint. Abusive supervision differs depending on culture (Vogel et al., 2015; Jiang and Gu, 2016). While abusive treatment by supervisors is largely considered unfair and inappropriate in a Western context, high status differences and a hierarchical societal structure may make supervisory abuse more acceptable in a Confucian Asian culture (Farh et al., 2007; Wu et al., 2009). In reality, abusive management tends to be considered normal in a Confucian Asian culture, and employees are expected to endure abusive supervisory behaviour (Vogel et al., 2015). Interestingly, while deeply-rooted traditional values continue to be in play in contemporary China, more and more people consider individual values and Western lifestyles more seriously. These changes are more salient in younger generations, and thus the Chinese context is an interesting setting to examine how young workers understand abusive treatment by supervisors.

In addition, most studies on abusive supervision consider the Western context, drawing data particularly from U.S. samples (Tepper, 2007; Martinko et al., 2013). The findings from a Western context may not be generalizable to other contexts, such as Asian Confucian ones. Despite the fact that China is a world factory, with young female workers forming the majority of the workforce in labour-intensive manufacturing industries, very few studies have considered this context (see recent studies: Lam et al. (2017), Lin, Wang and Chen (2013), and Wang et al. (2012)). Furthermore, our focal point—abusive treatment and its effect on turnover intention—needs more research, given that high turnover is one of the major concerns for most companies in China. The employee turnover rate is reportedly over 40 percent, especially in the eastern coastal area, which is known as a global hub of electronic products (The Economist, 2005, 2013). High turnover can cause various negative effects to companies, such as increased hiring and training costs, lower productivity, and lower quality management (Silloway et al., 2011; Schmidt, 2011). High turnover has become more problematic as it worsens the chronic labour shortages, especially of blue-collar workers, in labour-intensive industries in the eastern coastal areas (Freeman, 2015; Li and Lu, 2014; The Economist, 2013). Given this, this study can inform employers of theoretical and practical guidelines to reduce the high turnover rate by addressing abusive supervisory treatment.

In sum, the present study aims to enrich the extant research on abusive supervision by introducing new mediating and moderating mechanisms with psychological capital and organizational justice perception, and, also, by introducing the unique Chinese context and testing within it. The authors believe that the findings from young factory workers in the Chinese context are a much-needed complement to the extant discussion of abusive supervision. This study proceeds as follows. First, we provide five relevant hypotheses and explain the data collection. Then, we discuss the analysis results. The study ends with a summary of our findings as well as scholarly and practical implications.

Methodology

Abusive supervision and turnover intention

Previous studies have argued that turnover intentions or actual turnover can be one of the most common responses to abusive supervision (Bowling and Beehr, 2006), and that such responses increase organizational costs due to lost productivity (Tepper, 2000). A meta-analytic review of the consequences of abusive supervision showed that the association between abusive supervision and turnover intention was r = .30 (Zhang and Liao, 2015).

As abusive supervisors always badger subordinates about their past mistakes and failures or belittle subordinates in front of others (Tepper, 2000), employees accumulate negative affective events and may perceive their abusive supervisor as a source of fear according to the emotional process theory of abusive supervision (Oh and Farh, 2017). Then, their dominant response would be increased turnover intention or actual turnover, because human beings prefer to get out of harm’s way and escape harm (Frijda et al., 1989) if possible. In addition, employees can withdraw their organizational commitment and reduce their organizational identification and think of turnover as a trade-off for supervisory abuse (Burris et al., 2008). Research on the multifoci approach to the study of organizational justice (Rupp et al., 2014) has argued that an employee’s perception of justice about a source might affect their attitudes and behaviours toward sources other than the justice perception referent. This cross-foci justice effect is shown in the case of supervisory-focused justice affecting organization-directed outcomes due to the centrality of one’s supervisor (Rupp et al., 2014). Following this, we also expect that when subordinates experience abusiveness from their supervisors, their turnover intention can increase as a case of this cross-foci justice effect. From the previous research on abusive supervision and turnover intention, we state our first hypothesis as follows.

Hypothesis 1: Abusive supervision will have positive effects on turnover intention.

Abusive supervision and psychological capital

Although previous studies have examined several societal- and organizational-level outcomes caused by abusive supervision (Tepper et al., 2017; Zhang and Liao, 2015), a gap exists in the study of the psychological mechanisms underlying the relationship between abusive supervision and its outcomes, including turnover intention. While abusive supervision is defined as “subordinates’ perceptions of the extent to which their supervisors engage in the sustained display of hostile verbal and nonverbal behaviours, excluding physical contact” (Tepper, 2000: 178), how these perceptions actually lead to individual-level work outcomes is relatively less studied. To examine the psychological mechanism linking abusive supervision and turnover intention, we assume that the individual employees’ PsyCap can mediate this relationship. PsyCap, which refers to “one’s positive appraisal of circumstances and probability for success based on motivated effort and perseverance” (Luthans et al., 2007: 550), represents an individual’s capacity for work motivation, self-development, cognitive processing, and striving for success that results in performance (Peterson et al., 2011). As a multidimensional construct, PsyCap consists of four positive psychological attributes: self-efficacy, optimism, hope, and resilience (Luthans et al., 2006). PsyCap has attracted much attention from both practitioners and researchers because of its application to workplace-related measurement concerns including job satisfaction, performance, happiness, organizational citizenship behavior (OCB), commitment, stress and anxiety, and turnover intention (Luthans et al., 2007; Avey et al., 2011; Combs et al., 2012).

We focus on the role of AS in decreasing PsyCap because a ‘state-like’ PsyCap can change depending on working contexts, including leadership style and organizational climate (Luthans et al., 2008; Norman et al., 2010). In a dyadic relationship between a supervisor and employees, abusive supervisors can negatively affect the development of PsyCap. Psychological capital is an important resource that employees are motivated to acquire, foster, and maintain to achieve successful outcomes in the future (Peterson et al., 2011) and individuals try to protect, retain, and build their valuable resources in order to guard against potential or actual resource losses stemming from demands in their work environment (Hobfoll, 2002). Although AS has been shown to have negative associations with other employees’ psychological well-being variables (Zhang end Liao, 2015), the direct effect of AS on PsyCap has not been fully investigated. It has been argued that leaders play an important role in the transformation processes underlying PsyCap development (Luthans and Youseff-Morgan, 2017), especially when they exhibit positive leadership; they can facilitate the climate and resources necessary for individuals to drive their own PsyCap development. As much as there is a need for positive leadership, abusive leadership has been identified as a major hurdle in organizations (Luthans and Avolio, 2009). While genuine, authentic belief in the value of people, and the motivation and desire to build employees’ strengths and psychological resources are necessary for PsyCap development (Luthans and Youseff-Morgan, 2017), supervisors’ abusive behaviours work in the opposite direction, which may reduce the PsyCap of employees. The depletion of subordinates’ resources renders them increasingly more likely to engage in deviant behaviours (Mackey et al., 2017) and form negative attitudes such as turnover intention. For example, the self-efficacy of employees could be undermined by the repeated abusive comments and behaviours of their own line managers. A supervisor’s sustained under-evaluation of employee capabilities and negative feedback may decrease the efficacy of employees (Duffy et al., 2002). Self-efficacy beliefs can be shaped from prior performance (Earley and Kanfer, 1985), verbal persuasion (Earley and Kanfer, 1985), and psychological and affective arousal (Bandura, 1997). When supervisors repeatedly remind employees of previous mistakes and failures, withhold credit for positive performance, express beliefs about the incompetence of employees, and publicly blame employees for failure (Tepper, 2000), employees will negatively assess their own ability to succeed, and thus self-efficacy beliefs are decreased (Bandura, 2001).

Hope could also be adversely affected by abusive supervision, because hope refers to the individuals’ willpower and strategic plans for achieving their goals (Luthans et al., 2008). Moreover, because abusive supervision can have detrimental effects on employees’ self-confidence (Priesemuth et al., 2014), the hopes of employees would be diminished from decreased motivation. While optimism is one of the PsyCap attributes that can be developed through modelling (Peterson et al., 2011), an abusive supervisor does not offer positive modelling effects for employees, who may therefore view the environment around them more negatively. Resilience can be developed from the positive belief of people about reality and life (Contu, 2002), but abusive supervision does not offer social support that may help build up the employee’s own resilience. Rather, abusive supervisors are likely to take advantage of the employees’ mistakes or difficulties by ridiculing or humiliating them in public, which may make it more difficult for them to bounce back from failure and adversity.

While very few studies have directly examined the relationship between AS and psychological capital, one study found that abusive supervision was negatively related to the PsyCap of students in a Chinese university setting (Liao and Liu, 2015). We expect that in the manufacturing workplace settings in China, the relationship between AS and PsyCap is more stable and stronger than that found in the school setting, as the interactions between supervisors and employees are intense and occur on a more regular basis than those in school settings. Based on the existing research that finds that abuse by supervisors decreases personal resources, such as PsyCap, we propose that abusive supervision will be negatively related to employees’ psychological capital.

Hypothesis 2: Abusive supervision will be negatively related to employees’ psychological capital.

In addition, PsyCap has been shown to be an important mediator in the relationship between an organization’s contextual factors and its members’ attitudes and behaviours (e.g., Gooty et al., 2009; Luthans et al., 2008; Rego et al., 2012). A decreased PsyCap level can increase the employees’ turnover intention to escape from the current abusive supervisor and find another work environment. According to conservation of resources theory (Hobfoll, 2002), people are motivated to invest resources in order to protest against resource loss, to recover from loss and to gain resources. Turnover and getting another job could be seen as the investment of resources to protect one’s remaining psychological capital, and to avoid the cause of resource loss, which is the abusive supervision in this study. It has been argued that psychological capital is a key factor in understanding the variation in perceived symptoms of stress, as well as intentions to quit and job search behaviours (Avey et al., 2009). Therefore, we argue that the employees’ PsyCap will mediate the previously discussed relationship between abusive supervision and turnover intention.

Hypothesis 3: Psychological capital will mediate the relationship between abusive supervision and turnover intention.

As a relatively more distal contextual variable, organizational justice, the degree of perceived fairness in an organization (Colquitt, 2001), can affect abusive supervision’s influence on turnover intention. While employees experience abusive supervision from their direct bosses, their perception of the larger organizational context may vary from their perception of their supervisors. For example, employees may perceive their work-related outcomes as consistent with the implicit norms of allocation, such as internal equity, namely distributive justice (Adams, 1965; Leventhal, 1976). In addition, even with direct abusive supervisor, employees can perceive that their organizations still have fair decision making processes in terms of consistency, lack of bias, correctability, representation, accuracy and ethicality, which can be represented as procedural justice (Thibaut and Walker, 1975; Leventhal et al., 1980). However, abusive supervision is closely related to the perception of interactional justice (Bies, 2001), as the construct by itself reflects the nature of interpersonal treatment between employees and a supervisor and the information exchange between them. Therefore, in the current study, we focus on the perception of procedural and distributive justice, which are mainly about organizational systems and policies, and relatively more distant from the perception of the dyadic interaction with supervisors.

Perceived organizational justice can affect the relationship between abusive supervision and turnover intention. The employees’ perception of justice is found to predict many workplace outcomes such as performance, OCB, and turnover (Colquitt, 2001). Hackman (1992) referred to high-order contextual factors as ambient stimuli affecting the individual because they do not target a specific subject, but rather form part of the environment to which all the employees of an organization are equally exposed. The perception of organizational justice, especially procedural or distributive justice, could be one of those representative ambient stimuli, because it can widely alter employee attitudes and behaviours. For example, with high levels of procedural justice, employees can be vocal through official organizational mechanisms (e.g., grievance) to rectify the wrongs they face from their direct bosses in the case of abusive treatment. Otherwise, they can just tolerate their boss’s behaviour and hope that their final outcomes (reward) will be fairly distributed regardless of the abusive behaviours with high levels of distributive justice. In either case, the effect of abusive behaviour on turnover intention will be weakened. However, with low levels of procedural or distributive justice, the effect of abusive behaviour on turnover intention will be amplified by increasing the motivation to leave the current workplace. Therefore, we argue that the perception of procedural or distributive justice in the current workplace will weaken the effect of abusive supervision on employee turnover. Conversely, the perception of unfair procedural or distributive justice will amplify the effect of abusive supervision on employee turnover.

Hypothesis 4: The perception of organizational justice (procedural and distributive justice) will moderate the relationship between abusive supervision and turnover intention.

In addition, the employees’ PsyCap may be influenced by both factors, abusive supervision, which comes from close interaction with their own bosses, and the perception of organizational justice, which is the overall perception and expectation of the larger organization. PsyCap has been shown to work as a mediator influencing the effect of organizational justice on attitudinal behaviours (Totawar and Nambudiri, 2014). According to the “broaden and build” theory (Fredrickson, 2001), positive emotional experiences can lead to positive cognition, and individuals can build personal resources such as PsyCap. In addition to the dyadic relationship with their direct supervisor, employees form an exchange relationship with their own organization. Since organizational fairness is one of the fundamental elements of the workplace, it triggers positive emotions that result in the building up of PsyCap. Thus, high levels of organizational justice can weaken the effect of abusive supervision on PsyCap.

Conversely, perceived unfairness or injustice has been shown to be an important dimension in emotion-antecedent appraisal (Mikula et al., 1998). While perceived injustice may increase the likelihood of an emotional reaction, the nature of the reaction is determined by the combined outcome of the appraisal in several dimensions, such as pertinence of consequences, perceived causal agent, and estimated coping ability. While the perceived causal agent of abusive supervision is the direct supervisor, the fairness or unfairness of procedural and distributive justice could be an organizational-level system or policies, and therefore employees may experience complex emotional states leading to the build-up of PsyCap. In addition, procedural justice may enable employees to escape from their current abusive supervisors by challenging their decisions or behaviours through an official channel in the organization, thus helping them to reserve or even develop their own PsyCap, especially in terms of hope and optimism. Therefore, we argue for a moderated mediation mechanism where the perception of procedural or distributive justice can buffer or amplify the mediation effect of PsyCap in the relationship between abusive supervision and turnover intention.

Hypothesis 5: The perception of organizational justice (procedural and distributive justice) will moderate the mediation of psychological capital on turnover intention.

Data collection

We gathered our survey data from employees working in three different electro-mechanical parts assembly factories in the Shandong area in China. The final products included cellphones, printers, and other domestic electric home appliances. Pilot interviews were conducted in February 2016 for developing the survey items, and the main survey was conducted over two time periods in May 2016, with a one-week interval to avoid common methods variances. The time 1 survey included items pertaining to abusive supervision, turnover intention, perception of organizational justice, and PsyCap, and the time 2 survey measured the turnover intention that was used as a dependent variable in the model. The survey items that were initially in English were translated into Chinese using Brinslin’s (1986) back-translation procedure.

We distributed 400 surveys at both time points, and the response rate was 40.25 percent for matching data. Our final sample consisted of 161 female assembly line workers representing the characteristics of workers in electro-mechanical parts assembly factories. Women far outnumber men in labour-intensive manufacturing industries (e.g., garments, toys, and electronics) in China and globally (The Economist, 2013; BSR, 2013). In China’s electronic factories, young women dominate assembly lines, while men work in management or do maintenance tasks (Chang, 2009; China Labor Watch, 2012). It is known that young female workers are welcome due to their nimble fingers, obedience, and lower wages (Chang, 2009). Likewise, in the sampled factories, all assembly jobs were occupied by female workers.

Control variables

With regard to turnover intention, individual characteristics such as age, marital status, and tenure have shown meaningful effects. Therefore, these variables are included in the model.

Abusive supervision

Abusive supervision was measured using 15 items from Tepper (2000). With regard to behavioural description, employees responded to items including; (my boss) “ridicules me” and “expresses anger at me when he/she is mad for another reason” on a 5-point Likert scale as follows: 1 = I cannot remember him/her ever using this behaviour with me; 2 = He/she very seldom uses this behaviour with me; 3 = He/she occasionally uses this behaviour with me; 4 = He/she uses this behaviour rather often with me; 5 = He/she uses this behaviour very often with me. The Cronbach’s α for the total 15 items was .96.

Turnover intention

Turnover intention refers to the employees’ psychological state of wanting to leave the current organization. Employees are thus evaluating alternatives in terms of compensation and working conditions of other organizations in comparison with the current one (Steel, 2002). Following the above definition, we measured the turnover intention for both surveys using four items developed by Steel (2002) with a Cronbach’s α of .85 for the time 1 survey response and .88 for the time 2 survey response. The turnover intention for the time 1 survey was used as a control variable in the model.

Psychological capital

We measured positive PsyCap using the original scale of 24 items developed by Luthans et al. (2007). The employees showed their current psychological state by answering questions related to “How do you feel about yourself right now?” on a 5-point Likert scale, where 1 = strongly disagree, 2 = disagree, 3 = neither disagree nor agree, 4 = somewhat agree, and 5 = strongly agree. The four subscales of self-efficacy, optimism, hope, and resilience contained six items each, and the Cronbach’s α for self-efficacy = .71, for optimism = .70, for hope = .86, for resilience = .78, and for total PCQ = .89. The sample items included, “I feel confident in analyzing a long-term problem to find a solution (self-efficacy),” “I’m optimistic about what will happen to me in the future as it pertains to work (optimism),” “I feel that I am energetic enough to accomplish the work goal (hope),” and “I usually manage difficulties at work one way or another (resilience).”

Organizational justice perception

We used the justice measure developed by Colquitt (2001) where employees answered to what extent they agreed with statements such as “In the current organization, the procedures used to arrive at your outcome have been applied consistently (procedural justice),” “In the current organization, your outcome is appropriate for the work you have completed (distributive justice),” on a 5-point Likert scale, where 1 = to a small extent and 5 = to a large extent. Although our sample consisted of 161 individuals, there were only five departments in three factories, showing a large difference in terms of size, and the workplaces were structured according to three different types of products. Because of the small number of teams, we could not use the justice perception measure as a team-level variable, but used it as an individual-level measure reflecting the employees’ perception of organizational-level justice instead. The Cronbach’s α for procedural justice = .78, and for distributive justice = .82.

Data analysis and results

We conducted regression analysis with SPSS 22.0 software. We tested our hypotheses on the mediating effect of PsyCap and examined the moderated mediation model using the bootstrapping procedure proposed by Hayes (2013). We first analyzed the relationship between abusive supervision and turnover intention measured at time 2 and then examined the mediation mechanism with PsyCap. We next tested the moderated mediation model using organizational justice perception as a moderator, which is expected to moderate the effects of PsyCap on turnover intention. To test the mediation mechanism of PsyCap and the moderation mechanism of justice perception, we used the bootstrap method, which is considered more powerful for estimating the indirect effects in mediation models (Preacher and Hayes, 2004) than Baron and Kenny’s (1986) method and the Sobel test (Sobel, 1982). In this study, we repeated the bootstrap process for the recommended minimum of 1,000 times.

Table 1 shows the means and standard deviations of and correlations among the study variables. Internal consistency reliabilities at the individual level are reported along the diagonal. Abusive supervision is positively related to turnover intention (r = 0.32, p < .01), while it is negatively related to optimism (r = -.18, p < .05). There were several outliers in T1 turnover intention, which led to only a moderate level of correlation r=. 56 with T2 turnover intention. Although we did not exclude those responses, there were somewhat different patterns with regard to the correlation with marriage and age and T1 and T2 turnover intention.

Table 1

**Means, Standard Deviations, and Correlations among Study Variables**

As shown in Table 2, abusive supervision is positively related to turnover intention at time 2, while turnover intention at time 1 is controlled for with other control variables. Thus, Hypothesis 1 is supported. Table 3 shows that abusive supervision is negatively related to PsyCap, but this is only with optimism among the four attributes, thus partially supporting Hypothesis 2. We tested mediation with only optimism in the relationship between abusive supervision and turnover intention, to obtain significant results, as shown in Table 4. While the direct effect of abusive supervision on turnover intention is still significant with the mediator optimism, the significant indirect effect explains the partial mediation effect of optimism in the relationship between the two, partially supporting Hypothesis 3.

Table 2

**Direct Effect of Abusive Supervision on Turnover Intention (time 2)**

Table 3

**Direct Effect of Abusive Supervision on Psychological Capital (Optimism)**

Table 4

**Results of Mediation Test (Mediator: Psychological Capital-Optimism)**

The results of a simple moderation test of procedural and distributive justice in the relationship between abusive supervision and turnover intention are shown in Table 5. Moderation was significant at all three values (mean, -1SD, +1SD) of procedural justice perception. However, regarding distributive justice perception, the conditional effect was not significant for the higher value (+1SD) of moderator. In addition, the direction of effect was somewhat different from expectation. As shown in Figures 2 and 3, justice perception amplified the effect of abusive supervision when people perceived their organization as fair in terms of procedural or distributive justice. These results are different from our expectation that high justice perception will buffer the effect of abusive supervision on turnover intention.

Table 5

**Simple Moderation of Procedural and Distributive Justice Perception**

Figure 2

**Moderation of Distributive Justice on the Effect of Abusive Supervision on Turnover Intention**

Figure 3

**Moderation of Procedural Justice on the Effect of Abusive Supervision on Turnover Intention**

For the test of moderated mediation proposed in Hypothesis 5, we used the bootstrapping method with the process macro developed by Hayes (2013). The results are presented in Table 6. We found that procedural justice perception, but not distributive justice perception, moderated the mediation of optimism, partially supporting Hypothesis 5.

Discussion

The present study showed that abusive supervision is positively associated with turnover intention and that PsyCap, especially optimism, mediates this effect. In addition, the perception of procedural and distributive organizational justice moderated the relationship between abusive supervision and turnover intention, while that of procedural justice moderated the mediation of optimism. As one of the factors affecting employee turnover in China, abusive supervision was associated with a high level of turnover intention, and this was partially explained by the mediation of lowered optimism caused by abusive supervision.

Our findings contain meaningful implications for AS literature in several ways. First, it confirmed the detrimental effect of abusive supervision on turnover intention in the non-Western and manufacturing workplace contexts. Second, it explored underlying mechanisms and boundary conditions between abusive supervision and turnover intention through the mediation of PsyCap and the moderation of organizational justice perception. Third, it shed light on abusive supervision as a negative antecedent of PsyCap. Fourth, it showed that the perception of organizational justice can affect the relationship between abusive supervision and its detrimental results.

Our study contributes to the literature on abusive supervision by showing the psychological mechanism linking abusive supervision and turnover intention. As a resource drainer, abusive supervision reduces the psychological capital of employees, which leads to an increased level of turnover intention. Considering the important role of leaders in establishing employee attitudes towards the organization, the present study provides theoretical and practical implications regarding the detrimental effects of abusive supervision. Previous research on the factors that cause abusive supervision includes the affective and cognitive distress that leaders experience (Byrne et al., 2014). While supervisors experiencing resource depletion tend to show abusive behaviours towards employees, they are at the same time draining employees’ resources, represented by psychological capital according to our results. From the perspective of the organization, multiple layers of interventions to reduce the effects of abusive supervision can be considered. In particular, regarding the organizational context, there should be ways to preserve leaders’ cognitive resources so that they do not experience resource depletion that may lead to the loss of self-regulation associated with abusive supervision.

The reason why optimism, among the four factors of PsyCap, was significantly affected by abusive supervision and mediated the effect of abusive supervision on turnover intention should be considered. Optimism is an explanatory style that attributes personal, permanent, and pervasive causes to positive events and interprets negative events in terms of external, temporary, and situation-specific factors (Luthans et al., 2007: 90). Abusive supervision may affect this optimistic attribution style more directly than other PsyCap because employees under abusive supervision are continuously exposed to the opposite attributional style. The measurement of abusive supervision (Tepper, 2000) includes behaviours such as “reminding me of my past mistakes and failures,” “blaming me to save himself/herself embarrassment,” “not giving me credit for jobs requiring a lot of effort,” “putting me down in front of others,” “making negative comments about me to others,” and “telling me I’m incompetent.” As abusive supervision attributes negative events to personal causes, contrary to an optimistic attribution style, employee optimism is more adversely affected.

The present study also contributes to the literature on abusive supervision by highlighting the moderating role of organizational justice in the relationship between abusive supervision and turnover intention. However, the direction of moderation was different from the hypothesized direction, such that with high organizational justice perception, the effect of abusive supervision on turnover intention was amplified rather than weakened, which can have more meaningful theoretical and practical implications. This amplified effect with a high level of organizational justice could be explained by the contrast effect, as people have different expectations of what should be considered as norms through their own experiences in social exchanges and by observing what happens to others (Van den Bos and Lind, 2001; Folger and Cropanzano, 2001). Fairness theory (Folger and Cropanzano, 2001) argues that the counterfactual contrast between a referent standard and an experienced event establishes the degree of aversion to a negative event. Therefore, with high levels of organizational justice where clear and consistent expectations of fair supervision are the normative rules, employees will have stronger feelings about perceived injustices such as abusive supervision, resulting in more negative affective and behavioural reactions than employees with low levels of organizational justice. As a result, with high levels of organizational justice, the positive effect of abusive supervision on turnover intention is amplified. In addition, with a high level of organizational justice, trustworthy behaviours can be explained by situational forces such as fair organizational policies, but negative events are viewed as more intentional and malicious (Bies, 2001). Therefore, the target of blame, in this case an abusive supervisor, is relatively clear since the experience of unfairness is an unusual violation of norms. When individuals can externalize blame, they feel outward-focused emotions such as anger, and this attribution of blame may lead to individual behaviour to challenge such violations, showing an increased level of turnover intention. Conversely, with low levels of organizational justice where consistent fair practices are not guaranteed by the organization, it is more difficult to make attributions of blame when facing abusive supervision. Therefore, it is assumed that the relationship between abusive supervision and turnover intention is not as strong as it is with high levels of organizational justice.

The practical implications of this moderating effect should be interpreted with caution. As the impact of abusive supervision on turnover intention was stronger with high levels of organizational justice, it might seem adequate to keep fairness standards in the workplace lower rather than higher to reduce the impact of abusive supervision. However, low levels of organizational justice will cause distress for supervisors as well as subordinates. When leaders are psychologically distressed, their personal resources tend to be depleted so that they lack the ability to maintain appropriate social interactions and instead exhibit abusive supervision behaviours (Li et al., 2016). In turn, these abusive behaviours may drain subordinates’ resources. Therefore, companies should adopt micro-level interventions that may help to reduce the psychological distress of supervisors and develop employees’ psychological capital in order to reduce the level and detrimental effects of abusive supervision. In the same vein, as a way of reducing turnover in China, it is necessary to provide relevant and timely education to line managers about the effects of abusive supervision so that abusive supervision by managers does not occur in the workplace.

Despite the importance of our findings, several limitations should be noted. First, we collected data from northern China, and thus we cannot generalize our findings to all factories in China, and further in Asia. By a similar logic, because we collected data from the manufacturing industry, it may be difficult to make a generalization to other industries such as service, finance, or public organizations. Future studies can benefit from collecting data from various industries in multiple countries. Second, many of our variables were self-reported by employees. Future studies should use multiple sources to faithfully measure the perceptions of employees and those of supervisors. In this study, we did not measure the demographic variables of supervisors, but a previous study found that the effect of AS on employee reluctance to voice legitimate concerns about their workplace was moderated by gender dissimilarity such that the mediation effect of psychological distress was weaker for subordinates with gender-dissimilar supervisors (Park et al., 2018). Future studies can consider the effect of gender dissimilarity in examining the direct and indirect effects of AS and turnover. Finally, although our study sheds light on psychological capital by identifying AS as an antecedent that decreases individual PsyCap, it only showed that optimism out of the four sub-dimensions of psychological capital relates to abusive supervision and turnover intention. Future studies need to pay more attention to how the sub-dimensions of psychological capital work in this relationship.

Abstract

Résumé

Resumen

Introduction

Methodology

Abusive supervision and turnover intention