Saving the moral self: unethical leadership and employee behaviors – a moral cleansing perspective

Purpose – The purpose of this study is to investigate the concept of vicarious moral cleansing and scrutinize whether unethical behavior of leaders initiate moral cleansing in subordinates or not. This study also highlights a boundary condition under which employees are motivated to cleanse their moral self-image through increased organizational citizenship behaviors and decreased counterproductive work behaviors. Design/methodology/approach – The study is quantitative based on hypothesis testing. By adopting convenience sampling technique, employees working at all managerial levels of service sector organizations were asked to fill out the questionnaires. Being a time-lagged study, data for independent variable (unethical leadership) and moderator (relational self-construal) were collected at T1, data for mediator (moral self-image) were collected at T2 and data for outcomes (OCBs, CWBs) were collected at T3 from same respondents. To rule out the possibility of common method bias and social desirability bias, a multi-wave design was adopted and respondents were asked to provide unique keys/IDs instead of their names. Findings – This study investigated the impact that unethical leaders impose on employee self-concept. Moreover, this study also explored the motivational tendencies of moral self-image. Findings suggest that employees ’ desirable or undesirable behaviors against leader are dependent upon the perceptions related with their own role, self-image and perception of leader ’ s integrity and intentions. Leader ’ s unethicality is perceived threatening for their own moral self-image and they deal with it constructively. Thisstudyhaslaidthefoundationforpresenceofvicariousmoralcleansinginorganizationalsetup,andit isadvisedthatresearchersmustinvestigatethisphenomenonindifferentsettingstoprovideusefulinsights. Research limitations/implications – Due to lack of resources, employing a pure longitudinal research design was not feasible, and therefore a time-lagged research design was used to gather data from only two cities of Pakistan. However, authors believe that a longitudinal research design, with data collection from a larger sample, will provide more fine-grained results. Secondly, use of perceived leader ’ s integrity scale to measure unethical leadership is another limitation. Although the authors tried to address this issue by conducting an EFAand adoptingonlysuitable items,yet anew scale whichisableto measurethe true essence of unethical leadership ought to be developed. Originality/value – Useofmoralself-imageasanindicatorofmoralcleansingisanadditionalcontributionof this study, as previous studies used levels of guilt as driving force behind moral cleansing and compensatory cleansing. Most of the studies on unethical leadership as well as moral cleansing took place in the Western context and scholars ’ stress that culture can substantially influence outcomes of these constructs. Thus, this study extendsthe literature on moral cleansingand moral self-regulation by developingand testing a model in cultural settings of Pakistan.


Introduction
Current era has witnessed many communal scandals linked with unethical and immoral practices of leaders, which urged researchers to investigate when and how unethical leaders are affecting their subordinates and how the subordinates respond toward them (Schyns and Schilllings, 2013;Shaw et al., 2020).Unethical leadership, as defined by Brown and Mitchell (2010), is a leadership style where leaders themselves act unethically and also impose practices and procedures which permit and promote unethicality of their followers.These leaders indulge in immoral and unethical practices, set up an environment conducive for unethical practices and impose great harm for the organization, yet research on this important leadership style is lagging both empirically and theoretically (Brown and Mitchell, 2010;Eisenbeiß and Brodbeck, 2014).
Mainstream research comprehends that leader's unethical conduct results in low productivity, poor performance, decreased well-being, emotional dissonance and psychological disorders (Ferris et al., 2009;Mesdaghinia et al., 2019;Vriend et al., 2020).Moreover, subordinates of these leaders respond with hostility, deviance, supervisor-directed aggression and organizational sabotage (Chi and Liang, 2013;Burton and Hoobler, 2011;Thau and Mitchell, 2010).Yet, there is some recent evidence which suggests that employees differ in their response toward leaders and different employees deal with them differently.Employees play an agentic role, and instead of passively absorbing the harm caused by unethical leaders, they behave mindfully and consciously try to mitigate the effects of these leaders (Milosevic et al., 2020), that is, some employees respond toward unethical leaders with deviance, while others indulge in impression management, thus handling this adverse experience positively.The literature investigating the positive responses to dark leadership is sparse.Yet, scholars and practitioners are interested in knowing that why and when employees indulge in prosocial actions, exercise self-control and regulate their behavior despite the negative experiences with the leader (Vogel and Mitchell, 2017;Wee et al., 2017a, b).Lin et al. (2016) investigated the ethical and unethical behavior of individuals through the moral licensing lens and provided evidence that ethical behavior of an individual (leader in their case) could transform into abusive behavior via the mechanism of moral licensing and stressed to test the opposite mechanism, that is, that reverse of this could also be possible such that present unethical act could force an individual to indulge in ethical or altruistic actions to restore his/her threatened moral self-image.
This research comprehends that these varying responses to unethical leadership could be explained through the phenomenon of moral cleansing.Moral cleansing phenomenon is based on the premise that all individuals desire for a positive self-regard and whenever they feel that they have acted immorally and/or their identity of being a moral individual is under scrutiny, they undergo compensatory actions to restore their moral balance (Sachdeva et al., 2009;West and Zhong, 2015).Moreover, moral cleansing not only comes into action by one's own action but it also gets initiated by morally dubious behaviors of close others.Although initial investigation of Monin and Miller (2001) posited that unethical practices of close others could initiate moral cleansing actions, yet there is no empirical investigation of this to date.Specifically, moral cleansing is highly prevalent at workplace; yet the theoretical and empirical investigations for exploring and understanding causes and effects of moral cleansing are sparse.This research is thus aimed at investigating whether or not unethical leadership instigates moral cleansing of their subordinates.Particularly, it is argued that moral self-image (MSI), that is, one's assessment of being a moral individual at any given instant, could explain the impact of unethical leadership on outcomes.It is argued that unethical leaders would threaten the selfconcept of their subordinates, and to uphold their self-concept, they will undergo citizenship behaviors and refrain from CWBs.OCBs are discretionary actions which benefit the organizations and their members, whereas CWBs are deliberate actions to harm the organization or peers.The conceptual framework is presented in Figure 1.

Unethical leadership and employee behaviors
This research is aimed at making following contributions to existing literature.First of all, it is augmenting the literature on unethical leadership by investigating it in organizational settings and unleashing its effects on subordinates.By doing this, it answers calls of researchers to study how unethical leaders affect their subordinates.Secondly, this research explains the moral self-regulation of subordinates of unethical leaders.By doing this, it aims to address calls of researchers to investigate that how and when moral self-regulation comes into action under unethical leaders.Thirdly, this research studies the self-serving motivations of individuals behind pro-social actions, that is, this research comprehends that OCBs are not always enacted for the organization or its members; rather, sometimes they serve as a reparative role in building one's self-image.Fourthly, this research also adds into domain of moral cleansing by investigating moral cleansing in work settings, and provides support for the fact that not only action of oneself, but also activities of close others (vicarious moral cleansing), such as unethical leaders, enforce cleansing action of their subordinates.This article is among the few studies which are testing the concept of vicarious moral cleansing, specifically in organizational settings.Above all, this study is developing and testing in new cultural settings of Pakistan.

Definition of unethical leadership
Before proceeding further, it is worthwhile to define unethical leadership and differentiate it from other comparable constructs.The first definition for unethical leadership came from Brown and Mitchell (2010) who defined unethical leadership style as " behaviors conducted and decisions made by organizational leaders that are illegal and/or violate moral standards, and those that impose processes and structures that promote unethical conduct by followers" (p.588).Later, € Unal et al. (2012) added into their work and stated that leaders who are unjust, egotistic, immoral and who violate the rights of others lie in frame of unethical leadership.Afterward, Eisenbeiß and Brodbeck (2014) conducted a cross-cultural research to highlight the commonly held perceptions about unethical leaders and summarized that immoral foundations, egoism, urge to exploit and manipulate others and indifference toward feelings of subordinates are frequently associated with unethical leadership.La s akov a and Remi sov a (2017) described unethical leadership as a style where intentional or unintentional behaviors of leaders are harmful for their peers, subordinates and/or for the organization.Hence, unethical leadership style can be summed up as a leadership style where leaders act unethically themselves, possess self-serving intentions and to achieve their self-interest they exploit their subordinates and organizational resources.
Though many conceptually overlapping constructs exist in extant literature such as abusive supervision, tyrannical leadership and despotic leadership, yet unethical leadership is distinct.Specifically, not a single construct as mentioned above has taken exploitation and violation of interest of subordinates and organization together.For example, abusive supervision as well as tyrannical leadership exploits their followers and mistreat them, but their behaviors toward organizations are not a part of these constructs.Similarly, despotic and pseudo-transformational leaders exploit followers for their self-interest, yet none of these leaders force them to act unethically.However, unethical leaders not only act immorally themselves but also force their subordinates to act so.Hence, we can conclude that unethical leadership is a broader concept which might be conceptually overlapping with some dark leadership styles, yet it is distinct in its scope.

Moral cleansing theory
The concept of moral cleansing is a part of moral licensing theory (Monin and Miller, 2001) which is supported by the moral balance model (Nisan and Horenczyk, 1990).Moral licensing theory put forth the concept of moral licensing and moral cleansing, and argues that one's self-worth is derived from one's actions.Every morally dubious action diminishes one's moral self-worth, and every good action enhances one's moral self-image (Sachdeva et al., 2009).In other words, people tend to maintain a desired level of moral standing and constantly monitor their actions (West and Zhong, 2015).Individuals who perceive that their moral standing exceeds the desired level tend to feel licensed to engage in immoral behavior (moral licensing), and they are motivated to engage in moral behavior when they feel that their moral standing is below a desired level (moral cleansing).
According to moral cleansing theory, every immoral/unethical action has repercussions on one's moral self-image, which then motivates one to act morally.Moral actions are motivated by the desire to retain a preferred moral self-worth.Further, this theory contends that not only one's own actions but the actions of close others have the same effect on one's self-image, that is, individuals tend to act morally in response to their own morally dubious actions as well others' morally degenerate actions.This is termed as vicarious moral cleansing.
Existing research has validated the notion of moral cleansing.The research shows that individuals often experience negative emotions, such as guilt and shame, which leads them to believe they are morally deficient (Ding et al., 2016;Wang et al., 2021;West and Zhong, 2015), prompting them to act in a morally acclaimed manner.Drawing support from the basic tenants of moral cleansing theory, this study argues that unethical leaders' behavior will negatively affect employees' moral self-image.Considering that unethical leaders not only act unethically but also encourage and compel their subordinates to do the same, it is likely that the moral self-image of employees will also be threatened under these leaders.In order to retain their moral self-worth, employees will then engage in more OCBs and less CWBs.
Unethical leadership and moral self-image Jordan et al. (2015) stressed that individuals wish to retain a positive self-image and constantly evaluate their moral standing by asking themselves, "How moral I am?"The answer to this question changes with every favorable or unfavorable doing.Monin and Jordan (2009) argued that moral self-image is not static and, while verifying the malleable nature of self-image, stressed that moral self-image varies from one instant to another depending upon the situation.They highlighted that situational influence has substantial impact on one's self-image, which then shapes subsequent behaviors.
Followers emulate their leaders by internalizing their moral/immoral values, and these values, standards and behavioral patterns are made a part of one's self-image (Bandura, 1991;Gardner et al., 2005).Unethical leaders are self-serving who misbehave with others, abuse and mistreat them, violate organizational norms, exploit organizational resources and are dishonest (Brown and Mitchell, 2010;€ Unal et al., 2012); thus, followers, while perceiving these Unethical leadership and employee behaviors moral transgressions as a reflection of their own behavior, would begin to perceive a diminished moral self-image.Moreover, moral self-image is also shaped by one's own actions (Monin and Miller, 2001;Sachdeva et al., 2009); unethical leaders encourage or sometimes force their followers to act unethically.They promote unethical conduct of their followers by making unethical requests or by making them follow the dishonorable rules and regulations (Sparks, 2012).Such leaders implant disputes at workplace, turn employees against each other, force their subordinates to lie to cover for their own misdeeds and fuel clashes among employees for their self-interest (Craig and Gustafson, 1998).Thus, it is assumed that individuals who are working under unethical leaders would be, intentionally or unintentionally, acting unethically.All such behaviors are found to be detrimental for one's moral self-image.
Moral cleansing theory posits that one's own actions as well as the actions of close others have considerable impact on one's self-image.This research argues that manipulative, deceptive, aggressive, immoral and abusive conduct of unethical leaders (Brown and Mitchell, 2010;€ Unal et al., 2012) lies in the span of morally dubious behaviors which would affect moral self-image of their followers.Similarly, while working under these leaders, individuals are forced to act unethically, and previous research verified that unethical acts discount one's current moral self-image.Thus, depending upon this argument, it is proposed that H1.Unethical leadership is negatively related to moral self-image.
Moral self-image and outcomes (OCBs, CWBs) Individuals are substantially concerned about portraying an image of a moral being (Monin and Jordan, 2009).They want to appear moral to others as well as to themselves (Adler, 2006).This motivates them to undergo morally praiseworthy behaviors to present them as moral beings.All individuals possess an ideal image of character, which they want to achieve and strive to reach the ideal levels of morality in their daily endeavors (Barkan et al., 2015;Ploner and Regner, 2013).When individuals perceive that their expected and actual moral selfimages are not aligned, then they regulate their behaviors to reduce the discrepancy (Higgins, 1987;Jordan et al., 2015).
Individuals always strive to maintain a positive self-image, and when they perceive that their self-image is threatened or their morality is questioned, then they strive to rebuild it by undergoing actions that could help them restore the morality in their own eyes as well as in eyes of others (Ploner and Regner, 2013).For instance, in an experiment when participants recalled their immoral behaviors, they tried to balance their declining self-image by either involving themselves more in moral activities, conveying pro-social intentions or by reducing later deceptive/cheating conduct (Jordan et al., 2011).This serves as base argument for proposing that a diminished moral self-image could motivate individuals for prosocial actions at workplace such as OCBs and less CWBs.
Extant research has shown that OCBs are also performed for managing one's image such that helping others at workplace, staying late for work, taking extra responsibilities (Emami et al., 2012) and involving in other relevant activities helped individuals in maintaining a positive image (Bolino et al., 2008).Similarly, individuals often indulge in unethical practices for the sake of self-interest, but if these actions pose a threat to their moral self-image, then they withhold such behavior (Mazar et al., 2008).Similarly, Bandura (2004) stressed that when people know the penalties of their actions, then they are more likely to exercise self-control and avoid immoral behavior.
This study argues that motivation behind OCBs and refraining from CWBs is to boost one's threatened moral self-image.Individuals who perceive low moral self-image would try to distance themselves from any action that would bring more harm to their self-concept; in fact, they would indulge in more pro-social actions toward their peers and organization to enhance their moral self-worth.This proposition is in line with moral cleansing concept, which argues that when individuals identify a threat to their moral self or perceive that their moral self has been discounted due to any reason, then they actively try to heighten their image.Thus, it is proposed that H2a.Moral self-image is negatively related to OCBs.
H2b. Moral self-image is positively related to CWBs.

Mediating role of moral self-image between unethical leadership and outcomes
Moral cleansing theory posits that one's own morally vile actions as well as those of close others can make individuals suspect their own standing of being a moral individual (Monin and Miller, 2001).Such actions can have emotional and cognitive ramifications, which then lead to reparative actions that mitigate the effects of these actions (Liao et al., 2018).
Existing research has provided theoretical and empirical evidence that leader's behavior has substantial impact on follower's sense of self and they can threaten or even change the way individuals envisage themselves (Krylova et al., 2017;Vogel and Mitchell, 2017).For example, these followers of ethical leaders begin to induce morality in their working selfconcept and develop a self-image of being a moral person.Similarly, it is plausible to expect that followers of unethical leaders will perceive themselves to be unethical and thus will perceive a low moral self-image.Moreover, unethical leaders promote unethical behavior of their followers and contrive an environment where unethicality is promoted (Brown and Mitchell, 2010).Therefore, it is expected that in such a toxic environment, individuals will be bound to act in ways that do not go along with their moral values which will have implications for their moral self-image.Thus, this study posits that when individuals will perceive a threat to their morality and consider their moral self-image under scrutiny, then they would be encouraged to undergo reparative actions in the form of increased OCBs as well as refrain from acting unethically through decreased CWBs.
In line with existing evidence and in accordance with the concept of vicarious moral cleansing, it is posited that individuals, while seeing the deleterious and unethical conduct of their unethical leaders as a demonstration of their own character, will perceive that their moral self-image has deteriorated.The pressure of such leaders on subordinates to act unethically will have negative consequences for their moral self-image.Thus, to restore their moral self-image, they would then undergo compensatory actions in the form of increased OCBs and decreased CWBs.Moreover, they will also develop whistleblowing intentions which are a suitable predictor of accrual whistleblowing.Thus, it is proposed that H3a.Moral self-image will mediate the relationship between unethical leadership and OCBs.
H3b. Moral self-image will mediate the relationship between unethical leadership and CWBs.
Relational self-construal as moderator in the relationship between unethical leadership and moral self-image Among many associations found at workplace, leader-subordinate relationship holds a prime position and individuals define themselves in term of their leaders/supervisors (Brewer and Chen, 2007).Relational self-construal is defined as an individual's tendency to construe oneself in terms of relations with others including family member, supervisor and friend (Cross et al., 2003).Individuals differ in the way in which they incorporate their valued associations into their self-view (Heintzelman and Bacon, 2015) such that for individuals with Unethical leadership and employee behaviors higher levels of relational self-construal, close relationships hold a significant position in their self-concept (Cross et al., 2003).Individuals high in relational self-construal, aiming at healthy relationship with significant others (i.e.leaders), do not stand against their unethical demands but comply with them and act unethically or unfairly (Van Houwelingenn et al., 2017).This research argues that such followers will align their behavior with that of their leader's unethical behavior, and as a result their current moral self-image will be jeopardized.
The concept of vicarious moral cleansing asserts that actions of not only oneself but also close others such as friends, family member and colleagues are capable of instigating cleansing actions.Van Knippenberg et al. (2005) verified the notion that leaders have substantial influence on follower's self-concept and different leaders' behaviors could elicit different aspects of self.They expected that since different leaders' behaviors influence different facets of self-concept, the follower's self-construalthat is, the degree to which individuals include others in their self-definitionmight interact with leader's behaviors and influence other facets of self-conception.Since unethical leaders frequently misbehave, lie, abuse and exploit others (Brown and Mitchell, 2010;Eisenbeiß and Brodbeck, 2014); therefore, those who are close to their leaders will perceive that their immoral actions are spilling on to their own self-image.Thus, in line with moral cleansing theory and arguments presented above, it is plausible to expect that relational self-construal will moderate the relationship between unethical leadership and moral-self-image.
Depending upon this argument, it is proposed that H4.Relational self-construal will moderate the negative relationship between unethical leadership and moral self-image such that this relationship will be stronger at higher levels of relational self-construal as compared to low

Methodology
This study is a quantitative study based on hypothesis testing.To gather data, a survey was conducted among service sector employees of Pakistan.By adopting convenience sampling technique, employees working at all managerial levels of service sector organizations of Rawalpindi and Islamabad were contacted and asked to fill out the questionnaires.Being a time-lagged study, data for IV (unethical leadership) and moderator (relational self-construal) were collected at T1, data for mediator (moral self-image) were collected at T2 and data for outcomes (OCBs, CWBs) were collected at T3 from same respondents.Previous researchers in same domain adopted the same methodology (Naseer et al., 2016;Rasool et al., 2018).The participation in the study was voluntary, and respondents were informed about the scope of the study before filling out the questionnaire.
To rule out the possibility of common method bias and social desirability bias, a multi-wave design was adopted and respondents were asked to provide unique keys/IDs instead of their names.At T1, 600 questionnaires were distributed, whereas 557 were retrieved back; at T2, these 557 respondents were accessed and requested to provide data, but only 445 were retrieved back.At T3, of 445, only 390 responses were retrieved.Later, incomplete and unengaged responses were discarded, and a total of 362 usable responses were generated.
The response rate for this study is 60.3%

Instruments
This study adopted the existing questionnaires.The questions were in English language, as English is the official language of Pakistan and also the medium of education.Moreover, no respondent complained about not understanding the questions, and thus authors did not feel any need to translate the questions into native language, that is, Urdu.Following questions were used.

EJMBE 32,3
Unethical leadership.Previous studies utilized 31 items that perceived leader integrity scale developed by Craig and Gustafson (1998) to tap into unethical leadership (e.g.Spark, 2012).They argued that items of this scale particularly refer to unethical behavior enacted by leaders.Still this scale has some items similar to that of abusive supervision scale.Therefore, exploratory factor analysis (EFA) was conducted to derive the items that measure unethical leadership, and later analysis was performed by utilizing these items.EFA yielded 26 items which truly represent unethical leadership; thus, these 26 items were used for further analysis (detailed results are presented in the next section).
Moral self-image.Data for moral self-image were collected through a 9-item scale developed by Jordan et al. (2015).Responses on 7-point Likert scale (1 5 much less than what I want to be to 7 5 much more than what I want to be) were obtained.
OCB.A 13-item scale developed by William and Anderson (1991) was used.It is divided into two sub-scales measuring organizational citizenship behavior individual (OCBI) and organizational citizenship behavior organization (OCBO).The items represent OCBs directed at peers and organization.Sample items include "takes time to listen to co-workers' problems and worries" and "gives notice when unable to come to work."Responses were obtained on a 7-point Likert scale ranging from "1-Never" to "7-Always." CWB.A 13-item scale developed by Aquino et al. (1999) was used.The items represent CWBs directed at peers and organization.Responses on a 7-point Likert scale (1 5 Never, 7 5 Always) were obtained.Sample items include "called in sick when he/she was not really ill" and "refused to talk to a coworker." Relational self-construal.A 11-item scale of relational-interdependent self-construal scale by Cross et al. (2000) was used.Responses on a 7-point Likert scale (1 5 strongly disagree to 7 5 strongly agree) were obtained.The statements were adapted to add leader as the focal person (e.g."my relationship with the leader is an important part of my self-image").

Analysis tools
SPSS and AMOS software were used for obtaining results.For preliminary analysis, SPSS v21 was used.Factor analysis was performed by using AMOS.PROCESS macro by Hayes was used for regression analysis.Since this study proposed mediation and moderation hypothesis, model 4 and model 1 were used, respectively, to test these hypotheses.

Exploratory factor analysis
As discussed above, there is no validated scale of unethical leadership, and previously scholars have used perceived leader's integrity scale to measure unethical leadership; therefore, an EFA was performed to assess the suitability of the items of this scale to measure unethical leadership.
The first step in the EFA is to check the suitability of data for factor analysis (Williams et al., 2012).For this purpose, two tests are applied which include Kaiser-Meyer-Olkin (KMO) test to check for the adequacy of the sample (Kaiser, 1970(Kaiser, , 1974) and Bartlett's test for sphericity (Bartlett, 1950).The results of KMO and Bartlett's tests are reported below.KMO values for individual questions were greater than 0.645, which is well above the acceptable limit of 0.5 (Field, 2009).Bartlett's test of sphericity χ 2 5 27603.555,p 5 0.000, indicated that correlations between questions were sufficiently large for principal component analysis (PCA) (Table 1).
As the data were found suitable for factor analysis, we proceeded for the next stepthe calculation of factor loadings.A PCA with varimax rotation was conducted on the items composing the unethical leadership instrument used in this study.The results obtained are shown in the Table 2.The factor loadings for item 21, item 22, item 26, item 30 and item 31 Unethical leadership and employee behaviors were 0.29, 0.19, 0.33, 0.18 and respectively.These factor loadings are below acceptable limit (<0.4), and hence these five items were removed from further analysis.So EFA results highlight that among 31 items, 26 items are true representative of unethical leadership construct; hence, these 26 items were used further.The factor loadings and Eigen values along with the variance explained are shown in Appendix.
Confirmatory factor analysis (CFA) After EFA, CFA was performed to assess the discriminant and convergent validity of the study constructs.For CFAs, a five-factor model was run, and corresponding model fit indices were obtained.The following data represent the goodness of fit.We used the extracted 26item scale for unethical leadership for conducting CFA.Factor loadings for all the variables were in acceptable limit, that is, greater than 0.3, whereas other indicators of fitness including IFI-0.90, CFI 5 0.90, TLI 5 0.89, BFI 5 0.80 are also in acceptable range, that is, >0.8.Also, the RMSEA for five-factor model is 0.48 and χ 2 /Df is 2.78.All these values are within the standard limit put forth by Marsh and Hocevar (1985).

Control variables
When collecting data, respondents provided information about their demographics, such as their age, gender and qualifications.SPSS was used to run an analysis of variance test to determine whether any of the demographics affected the mediating or dependent variable.However, none of the demographic variables had a significant effect on the mediating variable (MSI) as well as the dependent variables (OCBs, CWBs)., Thus, none of the demographic variables were controlled in the analysis.(r 5 0.24, p < 0.05) and OCBs (r 5 0.38, p < 0.05), and negative correlation with moral self-image (r 5 À0.12, p < 0.01) and CWBs (r 5 À0.36, p < 0.05).Similarly, moral selfimage shows a positive correlation with relational self-construal (r 5 0.45, p < 0.05), CWB (r 5 0.42, p < 0.05) and negative correlation with OCBs (r 5 À0.19, p < 0.05).

Descriptive statistics
Proposed direct effect hypothesis, mediation hypothesis and moderation hypothesis were tested using PROCESS macro by Preacher and Hayes.Detailed results of direct and mediation effects are depicted in Table 3a and b.H1 proposed a negative association between unethical leadership and moral self-image.Results fully support H1, that is, unethical leadership is negatively and significantly associated with moral self-image (B 5 À0.3, t 5 0.02, p < 0.005).
After the main effect and mediation effect hypothesis, the moderation hypothesis was also tested by conducting hierarchical moderated regression analysis using PROCESS macro.The moderation analysis results provided impact of high and low values of moderator (relational self-construal) on UL-MSI relationship.An interaction plot at ±1 SD of moderator is shown in Figure 2. H4 proposed that the negative relationship between unethical leadership

Unethical leadership and employee behaviors
and moral self-image would be fortified individuals having higher levels of relational selfconstrual.Table 4 represents the results of the moderated regression analysis.H4 received full support as the interaction term UL x RSC was significant for moral self-image (B 5 À0.12, SE 5 0.01, p < 0.05).Furthermore, the bootstrap results for impact of different values of moderator on unethical leadership and moral self-image relationship also support H4, such that for high values of relational self-construal, the relationship between unethical leadership and moral self-image is the strongest (Table 4).Same can be seen in the interaction plot formulated at ±1 SD (Figure 2).As shown in Figure 2, the relationship between unethical leadership and moral self-image is moderated by relational self-construal.Figure 2 highlights that the negative relationship between unethical leadership and moral self-image is significant and stronger at higher levels of relational self-construal (B 5 À0.12, p < 0.001), whereas the unethical leadership and moral self-image relationship is insignificant at low levels of relational self-construal, thus lending support to H4.

Discussion
The aim of this study was to investigate how and when moral self-regulation in the form of moral cleansing is instigated in employees of unethical leaders.Research suggests that moral cleansing is highly prevalent in our day-to-day life; however, only few studies have investigated it in organizational settings (Liao et al., 2018).Specifically, while discussing moral cleansing, research has neglected the concept of vicarious moral cleansing.This research has shed light on the phenomenon of vicarious moral cleansing and integrated the literature of dark leadership and identification to explain the process of moral cleansing in response to transgressions of others (leader).The findings of this study have fully supported the proposed hypothesis.Though the proposed associations have not been developed and tested before, yet a bourgeoning line of research highlights that instead of always behaving destructively, individuals act thoughtfully or practically in response to detrimental supervisory/leadership behaviors.For instance, followers of abusive leaders sometime increase their job performance with an aim to avoid future abuse (Shao et al., 2018) or they indulge in ingratiation or façade creation to appear likeable in leader's eyes and restore their self-esteem (Vogel and Mitchell, 2017).Even individuals also increase their citizenship behaviors when they fear that they be negatively evaluated (Syed et al., 2018).Research in this domain is at nascent stage.This study is adding into this domain by comprehending the role of moral cleansing in compensatory altruistic behaviors under unethical leaders.
This study is unique as it is relating unethical leadership with outcomes via mechanism of moral self-image and increase in OCBs and decrease in CWBs under the tenets of moral cleansing effect.This study has specifically focused unethical leadership, which is an important yet theoretically and empirically overlooked construct (Eisenbeiß and Brodbeck, 2014).To the best of researcher's knowledge, no prior study has specifically investigated the moral selfregulation in the form of moral cleansing under a dark leader such as unethical leader.This research surpasses the mainstream literature and provides insight into the fact that followers of unethical leaders who identify with them have implications for their moral self-image.This diminished self-image forces them to undergo compensatory behaviors, and through these they try to revert to their moral self-image.Moreover, use of moral self-image as an indicator of moral cleansing is also an additional contribution of this study, as previous studies used levels of guilt (Liao et al., 2018) as driving force behind moral cleansing and compensatory cleansing.
Most of the studies on unethical leadership as well as moral cleansing took place in the Western context, and scholars' stress that culture can substantially influence outcomes of In fact, the findings suggest that despite the negative influence of unethical leaders, employees tend to react mindfully, that is, instead of retaliating they try to neutralize the harm by using their discretionary altruistic behavior.Also, they actively try to evade themselves from any questionable action so that their moral self-image is not further harmed.

Practical implications
The findings of current study support the notion that increase in OCBs and decrease in CWBs are not necessarily enacted in the best interest of organization or peers; rather, employees may have personal reasons for such behavior, such as boosting their own moral self-image (Organ, 1997).Since these behaviors and intentions are thought to be a means of boosting one's self-esteem, it is quite possible that once employees feel they have maintained a positive image, they would cease to perform them.Thus, by understanding employees' behavior and the motivating factors behind them, managers and organizations will be able to work with them more effectively.
Although the findings indicate that unethical leadership promotes positive outcomes, it does not mean that it should be encouraged.Despite short-term benefits, in the long run, unethical leadership is to be discouraged.Employees with a diminished moral self-image become depressed, exhausted and burned-out over time.Therefore, a workplace with unethical leaders is emotionally stressful and morally demanding for their subordinates.Organizations should thus pay attention to the emotional state of their employees and keep a check on the conduct of the managers or supervisors.It is equally important to provide a safe channel for employees to speak up and highlight the challenges they are facing.Doing so will prevent long-term losses for the organizations.

Limitations and future research directions
This study has some limitations which must not be overlooked.First of all, due to scarcity of time and resources, employing a pure longitudinal research design was not feasible, and therefore a time-lagged research design was used to gather data from only two cities of Pakistan.However, authors believe that a longitudinal research design, with data collection from a larger sample, will provide more fine-grained results.Secondly, use of perceived leader's integrity scale to measure unethical leadership is another limitation.Previous researchers also utilized the same scale, yet it is critiqued that this scale has similarities with abusive supervision scale.Although we tried to address this issue by conducting an EFA and adopting only suitable items, a new scale which is able to measure the true essence of unethical leadership ought to be developed in future.Moreover, unethical leadership is not a new construct; however, theoretical and empirical work on this important construct is sparse.Thus, we encourage scholars to investigate the effects of such leaders by taking individuallevel, team-level and organizational-level outcomes into account.Specifically, the mechanisms and boundary conditions through which these leaders exercise their influence are fruitful areas for investigation.

Conclusion
This study investigated the impact that unethical leaders impose on employee self-concept.Moreover, this study also explored the motivational tendencies of moral self-image.Based on the premise that every individual strives to maintain a moral self-image, thus study argued EJMBE 32,3 individual's desirable or undesirable actions are motivated by the perceptions of their own moral standing at any instant.The findings suggest that employees' desirable or undesirable behaviors against leader are dependent upon the perceptions related with their own role, self-image and perception of leader's integrity and intentions.Leader's unethicality is perceived threatening for their own moral self-image, and they deal with it constructively.In addition to extending the research on discretionary behaviors, the findings revealed that altruistic behaviors are not only enacted in the best interest of organization, but that they might have self-serving motivates (increasing moral self-image) behind them.Therefore, one may not always refrain from CWBs just to maintain the moral self-image at a desired level.As once the moral self-image is balanced, employees may stop OCBs and engage in CWBs.Moreover, unethical leaders harm the moral self-image and emotional well-being of their subordinates, which adversely affects the organization.We can therefore deduce that unethical leadership must be discouraged, and effective ways must be devised to keep moral dilemmas in check in the workplace.This study has laid the foundation for the presence of vicarious moral cleansing in organizational setup, and it is advised that researchers must investigate this phenomenon in different settings and provide useful insights.
Figure 1.The conceptual framework Figure 2. Graph depicting relationship between unethical leadership and moral self-image on high, medium and low levels of moderator

Table 1 .
KMO and Barlett's test N 5 362; UL 5 unethical leadership; MSI 5 moral self-image; bootstrap sample size 5 5,000.LL 5 lower limit, CI 5 confidence interval, UL 5 upper limit.*p < 0.05, **p < 0.01, ***p < 0.001 Thus, this study extends literature on moral cleansing and moral selfregulation by developing and testing a model in cultural settings of Pakistan.Since Pakistan has a distinct culture where people possess the collectivistic orientation and have high power distance, in such an environment aggressing against the leader or organization is least likely.