Leader-follower fit : the curvilinear relationship of ethical leadership and OCB and the moderating effects of subordinate’s moral identity and coworkers’ moral behavior

(1)

Leader-Follower Fit: The Curvilinear Relationship of Ethical Leadership and OCB and the Moderating Effects of Subordinate’s Moral Identity and Coworkers’ Moral Behavior

Birte Staden, 11148934 University of Amsterdam

Amsterdam Business School Date: 24th_{of June 2016} Final Master Thesis

MSc. in Business Administration – Leadership & Management Track Supervisor: Dr. Annebel De Hoogh

(2)

Statement of originality

This document is written by Student Birte Staden who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

The current study investigates the curvilinear relationship between ethical leadership and organizational citizenship behavior (OCB) and tests whether subordinates’ moral identity and coworkers’ moral behavior have impact on this relationship. The main focus within this study lies on the moderating role of these two factors. Additionally, the mediating role of coworkers’ moral behavior is examined. The data was collected through a quantitative questionnaire study resulting in 99 leader-follower dyads. A curvilinear relationship is found, however, contrary to expectations this relationship is positive instead of negative. Neither subordinates’ moral identity or coworkers’ moral behavior is found to influence the strength of this specific relationship. Thus, they are no moderating factors of the relationship between ethical leadership and OCB. In addition, coworkers’ moral behavior is not found to mediate this curvilinear relationship. These findings contribute to the debate about the influence of ethical leadership on employees work behavior and provide a basis for further research in this intriguing field.

Keywords: ethical leadership, organizational citizenship behavior, moral identity, coworkers’

(6)

Introduction

Have you ever experienced a work situation in which you felt discomfort because the decision your manager took did not match your believes about the right thing to do in this situation? Or have you ever experienced the perfect fit between you and your manager? Did you think: “Wow! My manager acts how I would want to act as well”? These are important aspects that can lead to satisfaction, commitment and a positive or negative attitude at work (Wang, Law, Hackett, Wang & Chen, 2005).

In today’s society, due to the fast changing economy effective leadership is more necessary than ever. After several scandals of large corporations such as Enron (Den Hartog, 2015; Mayer, Nurmohamed, Treviño, Shapiro & Schminke, 2013; Brown & Treviño, 2006), employees need leaders that provide trust and confidence in these uncertain economic times. These cases show that due to the unethical behavior of their leaders whole organizations had to suffer, for example through the result of bankruptcy and dismissal of employees. Therefore, especially ethical leadership behavior is nowadays highly appreciated and gains a lot of attention. This is so, as ethical leaders possess highly valued characteristics such as fairness, honesty and moral decision-making through which they show behaviors that distinguish between right and wrong conduct (Brown, Treviño, & Harrison, 2005; Brown & Treviño, 2006; Den Hartog, 2015).

However, there are always two sides of a story and therefore the focus should not only be on the leader but on the follower as well. Important is that not only the style of leaders and their consistent behavior have influence on their effectiveness, but also the employees values and personality traits are of great importance (Detert, Treviño, Burris & Andiappan, 2007). Employees with different values react differently to leader behaviors. Even though the leader might follow his/her moral beliefs when taking an action out of good intentions, the influence this action might have in regard to the employees’ work behavior seems to be influenced

(7)

through the employee’s moral values (Turnipseed, 2002). Thus, leaders do not only have to focus on one specific leadership style that they represent, but have to take the personality and values of their subordinates into account when taking decisions, being role models and reinforcing their subordinates. Furthermore, not only the employees’ moral tendency is important, but also the context and thereby the close social work environment they are surrounded with (Avolio, Sosik, Jung & Berson, 2003), such as their coworkers.

Even though the demand for ethical leaders is large and empirical progress has been made in this field of research, theoretical knowledge is still limited (Den Hartog & De Hoogh, 2009; Den Hartog, 2015). Current research on ethical leadership has found links to “attitudinal, motivational, well-being, and performance-related outcomes; ethical norms and decisions; and the (ethical) behaviors stemming from these” (Den Hartog, 2015, p. 417). Especially the focus of unintended negative effects of ethical leadership behavior is scarce and only recently the focus of attention. To my knowledge there are only two studies that focus on these negative effects (see Stouten, van Dijke, Mayer, De Cremer & Euwema, 2013; Miao, Newman, Yu & Xu, 2013). However, to fully understand the concept of ethical leadership behavior it is important to consider both, the positive and negative side.

The focus of this research will be to explore these two sides in more detail by investigating the curvilinear relationship of ethical leadership with organizational citizenship behavior (OCB). Analyzing a curvilinear relationship can identify when an initially positive relationship changes into a negative relationship or the other way round. This can create an inverted or a normal U-shape graph, respectively. In this study, as ethical leadership continually increases, OCB is expected to change from a positive to a negative relationship with ethical leadership. Thus, the positive influence that ethical leadership behavior can have might change into a negative influence. Importantly, by analyzing the curvilinear relationship not only positive or negative relationships can be taken into account but both at the same time.

(8)

OCB is a form of pro-social behavior (Brown & Treviño, 2006) and is highly valued by organizations, as it is any behavior that supports organizational functioning and effectiveness (Kalshoven, Den Hartog, De Hoogh, 2011). Furthermore, the context in which ethical leadership is executed plays an important role and needs further investigation (Den Hartog, 2015; Mayer et al., 2013). Also, earlier studies call for further research in ethical leadership in combination with follower types and their characteristics (see Detert et al., 2007; Avolio, 2007). Therefore, by analyzing this curvilinear relationship and by linking it with individual personality types and the social environment as moderating factors, this research will extent the prevailing literature and add to the current understanding of this relationship. Specifically, this research focuses on the moderating role of subordinates’ moral identity and coworkers’ moral behavior. In addition, I will investigate coworkers’ moral behavior as a mediating variable, because I expect that coworkers orient themselves at the behavior of the manager and thereby indirectly reflect the ethical attitude of the manager and influence other employees’ behavior.

This topic is not just of theoretical relevance, but also of practical relevance, as indicated by Den Hartog (2015, p. 428) “ethical leadership matters for organizations”. Therefore, managers need to be aware of the fact that by interacting with different employees they have to trigger different qualities of each individual employee. For example, reinforcement or communication techniques that work for one employee do not necessarily work in the same way for a different employee. In this research, the focus will be specifically on subordinates’ moral identity as values that have to be considered by the leader when interacting with their subordinates. In the past, the concept of moral identity has been widely used to explain non-organizational phenomena, and only recently has been focused on in organizational and management literature (Mayer, Aquino, Greenbaum & Kuenzi, 2012). Moreover, the emphasis has been mainly on the leaders’ moral identity (Aquino & Reed,

(9)

2002) and not the followers. By using this concept as characteristics of the subordinates’ moral behavior, this research extents the current line of research on this topic.

In sum, by linking ethical leadership to follower OCB and investigating the moderating effects of subordinates’ moral identity and the moderating and mediating effects of coworkers’ moral behavior this research intends to contribute to the current literature and helps to understand these processes in more detail. Thereby the current study adds to the existing literature by looking at the positive and negative sides of ethical leadership behavior and by integrating subordinates’ moral identity and the coworkers’ moral behavior as contextual influences. Furthermore, this research also adds to the practical knowledge of leadership techniques that should be used depending on the subordinate and the surrounding social work environment.

The questions that will be addressed in this research are: To what extent has ethical leadership unintended negative relationships with the subordinates’ work behavior? Especially, how is ethical leadership behavior related to subordinates OCB? And to what extent does the subordinates’ moral identity and perceived coworkers’ ethical behavior moderate this relationship? The theoretical model for this study is summarized in Figure 1.

(10)

Figure 1. Theoretical model.

This paper first provides an overview of the current literature about ethical leadership, OCB, subordinates’ moral identity, and coworkers’ moral behavior and provides hypotheses, which help to answer the research questions. Secondly, the methodology is described. Thirdly, the results are provided and are followed by a discussion including theoretical and practical implications, limitations, and suggestions for further research.

Theoretical background

Given this model, this paper examines the relationship of ethical leadership behavior and employees’ OCB and the moderating roles of the subordinates’ moral identity and the moderating and mediating role of coworkers’ moral behavior. Now this model will be explained in more detail by addressing the individual factors and by reasoning why these relationships are important.

(11)

Ethical leadership

In general, leadership can be defined as “the ability to influence a group of individuals towards the achievement of shared goals” (Aryee, Sun, Chen & Debra, 2007, p. 191). In the past, leadership has been analyzed to a great extent (Den Hartog & Koopman, 2001). The center of analysis has been on “leadership traits, styles and the contingency approach”, however, since several decades the focus is more on leader behavior (Den Hartog & Koopman, 2001, p. 167). These new leadership theories include behaviors that show charisma, inspiration and values. Through these leadership approaches, leaders try to influence the behavior of their subordinates to increase their motivation, commitment, respect and trust. Due to recent corporate scandals, trust in the moral values of leaders seems to be reduced. Therefore, companies’ interest in ethical leaders has been valued more than a decade ago. However, not only the interest for ethical leadership increased in practice, but also academically (Mayer et al. 2013).

In the current literature there is no consensus about one clear definition of ethical leadership. Earlier research has seen ethical leadership as part of charismatic or transformational leadership (Brown et al., 2005; Ng & Feldman, 2015). Even though clear links between the ethical dimension of leadership behavior can be linked to charismatic and transformational leadership, it actually can be seen as a leadership style on its own (Brown et al., 2005; Brown & Treviño, 2006). Brown and his colleagues (2005) identified that ethical leader behavior only partially overlap with transformational leader behavior. But that ethical leaders often use transformational leadership approaches when interacting with the subordinates. A main focus within the recent literature seems to be on the perceived leadership behavior, as the perception of a leader’s behavior is found to be crucial to his/her success (Treviño, Hartman & Brown, 2000). In line with the definition of Brown et al. (2005, p. 120), in this research ethical leadership can be defined as “the demonstration of

(12)

normatively appropriate conduct through personal actions and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making”.

Treviño et al. (2000) illustrate that the perception of ethical leaders consists of two moral aspects: the moral person and the moral manager. The moral person is identified to show favorable personal characteristics, such as honesty, fairness and trustworthiness. Additionally, by being a moral person, the leader acts as an ethical role model. The moral manager focuses on the management of morality in the way that ethical behaviors are supported whereas unethical behaviors are disciplined (Mayer et al., 2012). This is based on the ethical standards that were set by the leader and transmitted to the subordinates (Den Hartog, 2015). According to Treviño et al. (2000) in being perceived as a moral person and a moral manager, the leader is seen to be an ethical leader. This is not the case if the leader is either a moral person or a moral manager. Thus, taking all together, ethical leaders can be seen as fair, honest, trustworthy, and acting as role models that show the follower how they should morally behave and steer followers into behaving in a moral way. Furthermore, they are said to be caring for other people and are self-disciplined (Den Hartog, 2015; Ng & Feldman, 2015).

All of these traits reflect the positive influence that ethical leaders have on their subordinates in favor for their organization. Thus, the focus has been mainly on positive effects of this emerging leadership style. The current debate regarding ethical leadership focuses more on contrasting the positive effects of ethical leadership versus the negative effects of unethical leadership (see Brown & Mitchell, 2010; Brown & Treviño, 2006). However, as identified by Brown and Treviño (2006), a high level of unethical leadership is not the same as low level of ethical leadership. The unintended negative effects that ethical leadership behavior possibly can cause have not been found much of attention yet. These

(13)

possible unintended negative effects relate to the undermining of positive behavior outcomes (Le, Oh, Robbins, Ilies, Holland & Westrick, 2011). I believe it is important to not only consider the possible increase in positive behavior, but also to investigate under what conditions these positive behaviors possibly can be undermined. Therefore, this research will analyze possible unintended negative effects of ethical leadership behavior on follower behavior that might follow an initial positive relationship.

To explain the positive follower outcomes of ethical leader behavior, a frequent focus of research lies on social learning theory and social exchange theory (Den Hartog, 2015). According to Bandura’s social learning theory (1977), individuals learn how to behave in certain situations by observing, experiencing and imitating the behavior of other people in their environment. Especially important for individuals to take on the same behavior and values, in case of ethical leadership, is that the leader is seen to be an “attractive and credible role model” (Brown & Treviño, 2006, p. 597). Therefore, as leaders reflect ethical behavior by being a role model, the followers learn unconsciously to behave in the same way as the leader. This is so, because they experience and observe ethical actions taken by the leader. Thus, leaders are seen to have a large impact on their subordinates’ moral behavior. This is in line with social learning theory, which proposes that the experience of ethical behavior mainly takes place through reward and punishment that is carried out by the leader (Bandura, 1977; Brown et al., 2005).

Together with the social learning approach, the social exchange theory assumes that if a leader reflects moral values and behaves in an ethical way towards the subordinate, the subordinate will reciprocate such a behavior towards the leader. Thus, if the subordinate experiences fair and honest treatments by the leader, to return the favor he/she will reciprocate the behavior by also acting fair and honest (Hansen, Alge, Brown, Jackson & Dunford, 2013).

(14)

Also for the current hypothesis development, social learning theory and social exchange will be used.

Research in the area of ethical leadership has focused on different outcome variables. These are especially subordinate attitudes, behaviors or effort and performance outcomes. Much of attention has been on positive effects of ethical leadership (see Den Hartog, 2015 and Ng & Feldman, 2015). Positive influences have been found of ethical leadership on commitment, ethical behavior, and OCB (Den Hartog, 2015; Ng & Feldman, 2015). Furthermore, it has been found to increase the subordinates well-being and optimism in regard to the future (Den Hartog, 2015; De Hoogh & Den Hartog, 2008). As OCB is a kind of pro-social behavior that is initiated out of intrinsic rather than extrinsic motivation and involves a strong ethical element, the citizenship behavior is highly valued by organizations (Brown & Treviño, 2006; Stouten et al. 2013). Furthermore, it is an important behavioral outcome, as it can be influenced by the leader’s management style (Den Hartog, De Hoogh and Keegan, 2007). However, the leader might undermine OCB under certain conditions as well. Therefore, this paper concentrates on the behavioral outcome of OCB and analyzes to what extent ethical leadership can have a positive and negative relationship with OCB. The next section will introduce OCB in more detail and link it with ethical leadership behavior.

Organizational Citizenship Behavior

OCBs can be defined as any behavior that supports organizational functioning and effectiveness. “Such behaviors include helping, taking the consequences for others of one's own actions into account, being loyal, and keeping up with internal organizational information” (Kalshoven, Den Hartog, De Hoogh, 2011, p. 59). However, there is not one composition of OCB, but different dimensions have been identified within the literature. For example, MacKenzie, Podsakoff and Fetter (1991) focus in their study specifically on the

(15)

dimensions of “altruism, civic virtue, courtesy, and sportsmanship” (p. 123). In line with this, Den Hartog et al. (2007) describe OCBs as voluntary behaviors that are performed to benefit the common, so the organization or the department they are working for, and not necessarily for the individuals benefit as well. Furthermore, two important aspects of OCB are identified to be “interpersonal helping and compliance with the rules and procedures of the collective” (p. 1132). Thus, through voluntarily helping other team members and following not only formal rules and procedures but also informal ones, ethical components of acting in the right way can be identified as part of OCB. Therefore, OCB can be seen to include some ethical components (see Turnipseed, 2002). These behaviors are proactive, as they motivate and encourage things to happen within the organization and are therefore crucial for the long-term effectiveness of the organization (Van Dyne & Le Pine, 1998; Stouten et al., 2013). They enable employees to adapt to changes that the organization faces in current economic uncertainties and to act according to organizational standards that might not be clarified yet (Stouten et al., 2013). Therefore, it is important for managers to understand the meaningfulness of this construct and how to reinforce followers to behave accordingly. Furthermore, managers should be aware of the fact, that some interactions might cause OCB to be undermined.

For this paper, the focus will be specifically on the dimension of helping, proactivity and keeping up with internal organizational information. These behaviors include commitment towards the organization and therefore are important positive aspects that can result from the right leadership style. Additionally, helping can be linked to behaving in an honest and caring way and therefore can be connected to ethical traits. Also, taking a proactive role in following voluntarily formal and informal rules or procedures includes an ethical behavior to act in the right way. Furthermore, to keep up with internal organizational

(16)

information is an important aspect that supports this line of thinking further (Turnipseed, 2002).

Earlier research has focused on the relationship between ethical leadership and OCB (e.g. Brown & Treviño, 2006; Stouten et al., 2013). Even though empirical evidence investigating this relationship is scarce (Kalshoven et al., 2011), different research has shown that ethical leadership behavior positively influences organizational citizenship behaviors and by this showing a positive linear relationship of these two constructs (Mayer, Kuenzi, Greenbaum, Bardes, & Salvador, 2009; Mayer et al. 2012; Walumbwa & Schaubroeck, 2009). For example, by providing trust towards the subordinate, ethical leaders teach them through role modeling how to behave in a trustful way and influence their willingness to reciprocate the same behavior. Furthermore, in representing and communicating the organizations ethical standards, leaders encourage their subordinates to behave in the same way (Mayer et al., 2009).

However, it has only recently been proposed that this relationship might be curvilinear instead of linear (Stouten et al., 2013). This curvilinear relationship would mean that at some point the initially positive relationship between ethical leadership and OCB would turn into a negative relationship. By investigating the relationship between ethical leadership and OCB, Stouten et al. (2013) found indeed a negative curvilinear relationship. They tested their assumptions in three field studies in organizational settings by using surveys and in one experimental study to find support for causality. The three field studies revealed a curvilinear relationship between ethical leadership and OCB. In total, through the four studies, they found that moderate ethical leaders motivate OCB more compared to leaders with a low or high level of ethical leadership behavior. Different to the current study, which focuses on factors that strengthen this curvilinear relationship, they used perceptions of moral reproach as a mediating factor to explain the curvilinear relationship. So, they argue that high ethical

(17)

leaders are perceived to look down on their employees and by that demotivating them. As the study indicates that a curvilinear relationship exists, it is important to find additional support for this claim and increase the understanding of this relationship. This is exactly the purpose of the current research. By using a survey design only, the current research uses a slightly different approach of investigating this relationship, but adds to the question of generalizability of their study.

In line with this, I propose that ethical leadership behavior positively and negatively influences the subordinates OCB depending on the leader’s level of ethical behavior. Specifically, this relationship is assumed to be curvilinear instead of linear, by first showing a positive relationship and then turning to a negative relationship. Thus, as ethical leaders are said to be fair, honest, trustworthy, and are caring for other people (Den Hartog, 2015), they act as role models and by that are positively influencing the motivation of the employees to contribute to the organizational success. This is in line with Bandura’s social learning theory (1977) and can be supported by Brown and Treviño (2006) who argue that OCB is stimulated through the leader by being an appealing role model “who focus follower's attention on their ethical standards and their normatively appropriate behavior” (p. 607). However, I believe that this motivation can be negatively influenced when the ethical expectations of the leaders on the subordinate are perceived to be too high or too low. In case of too high-perceived ethical expectations (too high ethical leadership), the subordinate may experience a great deal of pressure to live up to the standards given by the leader. By not reaching these standards the subordinate might feel that he/she is continuously failing. Over time this reduces the motivation of the subordinate to do his/her best to reach these high standards and thereby the contribution towards the organizational effectiveness and functioning is reduced. In case of too low perceived ethical expectations (too low ethical leadership), the leader might not be seen as a role model that represents trust, honesty and fairness. In that case, the subordinate

(18)

does not identify with the leader in such a way that OCB would be increased, as the motivation to do so is reduced or even not present. Thus, the subordinate would contribute also in this case less to the organizational effectiveness. Therefore, it can be hypothesized:

Hypothesis 1: Ethical leadership behavior has a negative curvilinear relationship (inverted

U-shape) with subordinates OCB.

Subordinates’ moral identity

Due to the fact that individuals are different in regard to their personality, moral values and attitudes, how the leader influences the subordinate seems to be highly influenced through the individual moral attitudes of the subordinate. Thus, to take personal differences into account, it is suggested that the relationship between ethical leadership behavior and OCB is moderated by the degree of the subordinate’s moral identity.

Moral identity can be defined “as a self-schema organized around a set of moral trait associations (e.g. honest, caring, compassionate)” (Mayer et al., 2012, p. 152). Furthermore, “moral identity has been described as one kind of self-regulatory mechanism that motivates moral action” (as cited in Aquino & Reed, 2002, p. 1423). For individuals high in moral identity, moral decision-making is central to their behavior and they are inclined to act consistently. Furthermore, commitment and morally appropriate behavior is highly valued and self-control is high. Therefore, when facing a specific situation or decision, people high in moral identity act self-regulated in the way they value the most. By doing so, in a moral dilemma, they are able to weight what is right and wrong before making the decision of how to act. However, as the ethical values are internalized, this decision making mostly happens automated. So, when facing the possibility to behave unethical, such as cheating, people high in moral identity do not value the gain that can be achieved through the unethical behavior over retention of their self-concept. And thus automatically behave honest in a situation where

(19)

cheating would be possible (Gino, Schweitzer, Mead & Ariely, 2011). In their study, Gino et al. (2011) analyzed the relationship between self-control depletion and unethical behavior and the impact moral identity has on this relationship. They found that when comparing participants of high and low moral identity with a depleted self-control, the former were less likely to behave in an unethical way. This highlights the fact that with high moral identity people internalize ethical behavior and act automatically in the morally right direction.

In line with this, Aquino and Reed (2002) divide the concept of moral identity into two dimensions, the ‘moral identity symbolization’ and ‘moral identity internalization’. The former contains the social perspective of the individuals’ moral identity, in which the individual shows the traits that are owned by acting in a moral way. On the other hand, the latter captures the personal perspective, in which the self-concept includes the individual moral traits (Mayer et al., 2012). Mayer et al. (2012) describe people high in internalization to be more likely to prevent immoral behaviors in order to sustain the self-concept. Furthermore, they are more inclined to focus and correct on other people’s unethical behavior.

For this paper, moral identity internalization is especially relevant and will be focused on, as this includes the individual values that one stands for without necessarily obviously showing the traits. Furthermore, internalization takes the person’s feelings into account and relates to the individual’s self. I believe that the concept of moral identity is especially important to consider, as it involves self-schemas associated with moral traits (Mayer et al., 2012), and by this comprises a more in depth moral tendency than to consider individual characteristics only.

In line with the stated literature, it can be suggested that the moral identity of the subordinate influences his/her perception of the ethical behavior of the leader. An individual with a high moral identity can be assumed to react to and perceive the ethical behavior to a different degree than an individual with low moral identity. On the one hand, individuals with

(20)

high moral identity are less likely to identify with a leader that represents low ethical leadership behavior and thereby showing less OCB. This is so, as the moral standards of the individual would be higher than the leader’s. However, in the case of a leader with high ethical leadership behavior, the leader might be perceived to be too rigid (Stouten et al., 2013) and that the moral standard cannot be reachable. Thus, the motivation to follow the leader would be reduced. This reasoning indicates a curvilinear relationship, which is first of positive nature and then turns towards a negative relationship. On the other hand, it can be expected that individuals with low moral identity might reach the point where it negatively turns already earlier, as the high ethical standards are earlier perceived to be not reachable. The curvilinear curve would start at a higher point, as low ethical leadership behavior would be more likely to be in line with the low moral identity of the subordinate. Therefore, the subordinate would see the leader as an example and OCB would be still low, but higher compared to a subordinate with high moral identity. In general, this curve is expected to stay on a lower level of OCB and begins to decrease at a point, earlier than the subordinate with high moral identity. This argumentation stresses that the right fit between the leader and the subordinate is crucial for an organization to be successful (Wang et al., 2005).

Next to the moderation effect, it can be assumed that the moral identity of the individual by itself has a direct effect on OCB. This is important to keep in mind when analyzing this relationship, however, the main focus of this research will be on the moderation effect of ethical leadership on OCB.Therefore, the following hypothesis is developed:

Hypothesis 2: The negative curvilinear relationship between ethical leadership and OCB will

(21)

Coworkers’ moral behavior

Next to the moral identity of the subordinate, the actors within the social environment are expected to have a large effect on the proposed relationship as well. Thus, the presence of coworkers and their personal attitudes can shape the commitment of an individual towards the organization one works for (Mayer et al., 2013). This is especially important, because differently to a mentor relationship which is intentional, a coworker as an influence of an employee’s behavior is informal and cannot be controlled or planned (Weaver, Treviño & Agle, 2005).

Employees work every day closely together with their coworkers and experience their behavior, perceive their moral attitudes, and receive help to understand the leadership messages (Mayer et al., 2013). By this they can be assumed to gain ideas of how to behave in certain situations. This is in line with the social information processing theory (Salancik & Pfeffer, 1978), which states that individuals observe their social environment to decide on how to best behave in specific situations. Especially when a situation leaves individuals in uncertainty or doubt they are said to be more likely to follow the behavior of the social environment. As the work environment is complex, many uncertain situations can arise. Therefore, the employee will look for someone who shows appropriate behavior. Leaders and coworkers are usually the most present individuals in the close working environment for an employee and therefore they will focus, orientate and base their behavior on those coworkers, next to the leader.

Mayer and colleagues (2013) proposed that for an employee to decide whether to report internal unethical conduct, they most likely orientate at their social environment. By this in their study they focused on the direct supervisor and the employee’s coworkers. The research findings support this proposition through a survey study, a field study and an experimental study; thereby they investigated not only on behavioral intentions through the

(22)

survey study, but also actual coworkers’ behavior through the field and experimental study. Thus, the findings provide support that supervisors and coworkers have a large effect on the disclosure of unethical actions. Thus, especially in cases when ethical behaviors have to be performed, mostly the leader can be seen to be the first reference person. However, as the coworkers are a point of orientation as well, to which degree they confirm with the leader’s opinion likely plays a great role as well. Therefore, due to the presence of coworkers in the work environment they influence the relationship between the ethical leader and the subordinates OCB.

Weaver et al. (2005) supported this by conducting in-depth interviews with employees and asking them about their experience with moral role models. The interviewees mainly named their coworkers and their supervisors as the individuals in their close environment to be their role models. Even though this seems obvious, little research has focused on the coworkers’ influence on work related experiences of employees (Mayer et al. 2013). Thus, by empirically testing the influence that coworkers’ moral behavior can have on the relationship of ethical leadership on OCB, this study extents this current line of thought.

In addition, taking the social learning theory (Bandura, 1977) into consideration, by experiencing and observing the behavior of other’s, the individual learns to act in the same way, even though that might be unconscious. To follow the ethical decisions of the leader or coworker and behave in the same way, however, might not always be in line with the actual behavior that the employee would choose based on his/her own standards. This stresses the importance to consider both personality and environmental factors when investigating the relationship between ethical leadership and OCB.

As mentioned above, each employee is individually different in regard to their moral attitude and from that their resulting ethical behavior. Thus, also the degree to which the coworker behaves in line with the ethical leader is crucial for each subordinate to engage in

(23)

OCB. Therefore, it can be assumed that the strength of the curvilinear relationship between ethical leadership and OCB will be affected through the coworkers’ moral behavior as a moderator. I expect that the positive and negative curvilinear relationship between ethical leadership and subordinates OCB will be stronger when the coworkers behave in the same moral way as the leader does, as this supports the leader behavior. However, when the coworkers behave opposed to the leader’s ethical behavior, then the positive and negative curvilinear relationship will be expected to be less strong. Thus, this makes clear that for a leader to influence all the employees into the same direction is really important for the leader to be effective and to motivate the employee to contribute to the organizational functioning.

Other experimental studies have found that “peers influence the unethical behavioral intentions and behaviors” (Mayer et al., 2013, p. 90). This indicates that coworkers’ behavior does have a direct effect on the behavior of the subordinate. Furthermore, due to the fact that ethical leaders act as role models who motivate their employees to behave ethically in the same way as the leader (Brown & Treviño, 2006), leaders can be expected to influence the moral behavior of all their subordinates. Thus, by creating an ethical work environment in which morally appropriate behavior is provided as an example, leaders positively influence the subordinates to participate in pro-social behaviors, such as OCB. This is important, as this means that through the moral attitude of the leader the coworkers’ moral behavior will be influenced, which directly influences the pro-social OCB of the subordinate. Thereby, ethical leadership influences OCB also indirectly through the coworkers’ moral behavior.

In this research the focus will be mainly on the coworkers’ moral behavior as a moderating factor and it can be hypothesized:

Hypothesis 3: The negative curvilinear relationship between ethical leadership and OCB will

be stronger for subordinates with high coworkers’ moral behavior than for those with low coworkers’ moral behavior.

(24)

In addition, it can be expected:

Hypothesis 4: The negative curvilinear relationship between ethical leadership and OCB is

mediated by coworkers’ moral behavior.

Research Method

Sample and Procedure

The data was collected jointly by a team of six international students following the Master Business Administration with the track of leadership & management at the University of Amsterdam. Specifically, both leader and follower data was gathered, for which matched questionnaires of leader-follower dyads were used. To do so, first the necessary measure items for these six students were compiled and combined to design one questionnaire for the leader and for the follower. Moreover, the questions were either provided in Dutch or English, depending on the preference of the participant. To prevent possible bias of the questionnaires, it was made sure that in both versions the questions have the same meaning as well as the chronological order is given. The questionnaires were made online by using Qualtrics.

Important is that the dyad questionnaires can be matched, so that it is easy to see which questionnaires belong together. Therefore, code words were used which both the leader and the follower filled in at the beginning of the questionnaire. Each dyad came up with an individual code word. To prevent for possible duplication, this code word was instructed to be in the form of a word of at least six letters linked with two numbers. A box for the code word was provided on the first page of the questionnaire, so that nobody could fail to notice the box.

The six students distributed the online questionnaires to 174 employees and their direct supervisor in various industries in the Netherland, Greece, Spain, and Germany. An email including a link to the questionnaire was sent to the possible participants. These emails

(25)

announced the study and asked for participation, stressed the anonymity of the questionnaire and assured confidential treatment of the collected data. Furthermore, contact information was provided to make sure that the respondents could directly contact one of the team members in case of questions.

To make sure that the response rate could be provided, an excel sheet was used to keep track of all the employees and supervisors that were asked to participate. Two and four weeks after the distribution of the emails, a reminder was sent to all the requested participants. This is so, because there was no possibility of linking the responses to the people who were requested to participate. After five weeks the data collection was closed. Only complete sets of employees and their direct supervisor were included in the dataset.

In total, 146 employee questionnaires and 126 manager questionnaires were completed. Ultimately 122 matched questionnaires of employee-manager dyads were obtained. However, due to missing data and double entry 23 dyads had to be excluded. Therefore, the final sample exists out of N = 99 dyads. Subsequently, the response rate was 57%. Of the respondent employees 44% were male and 56% were female, with a mean age of 35 (SD = 12.42). Of the supervisor respondents 66% were male and 34% female, with a mean age of 45 (SD = 10.99).

Measures

In the present research, the following variables are used: ethical leadership, OCB, subordinates’ moral identity and coworkers’ moral behavior. All responses were given on a 7-point Likert scale ranging from 1 (totally disagree) to 7 (totally agree). The subordinates rated ethical leadership, moral identity and coworkers’ moral behavior, whereas the managers rated OCB (see Appendix B).

(26)

Ethical leadership was measured based on Brown et al.’s (2005) 10-item Ethical

Leadership Scale. All items are indicative. An example item is “My manager makes fair and balanced decisions” or “My manager disciplines employees who violate ethical standards”. Cronbach’s α was 0.91.

Organizational citizenship behavior was measured on a 13-item scale (based on Van

Dyne & Le Pine, 1998; MacKenzie et al., 1991), with Cronbach’s α of 0.87. All items are indicative and examples read “This particular co-worker communicates his/her opinions about work issues to others in this group even if his/her opinion is different and others in the group disagree with him/her” and “This particular co-worker considers the impact of his/her actions on others”.

Subordinate moral identity was measured based on Mayer et al.’s (2012) 5-item

‘Moral Identity Internalization’ scale in which the questions were related to a list of some characteristics of a particular type of person: caring, compassionate, fair, friendly, generous, helpful, hardworking, honest and kind. An example question includes, “It would make me feel good to be a person who has these characteristics”. Two of the five items are counter-indicative and an example reads, “I would be ashamed to be a person who had these characteristics”. Cronbach’s α was 0.81.

Coworkers’ moral behavior was measured through 3 items (based on Mayer et al.

2013), with Cronbach’s α of 0.83. All items are indicative. One item reads, “Overall, my coworkers set a good example of ethical business behavior” or “My coworkers support me in following my company’s standards of ethical behavior”.

Control variables. Additionally to the variables described, control variables were

included in the questionnaires. The control variables that were used in the questionnaire of the leader were gender, age, country of employment, organizational tenure, period of time being manager of the employee (supervisor tenure) and contact frequency. The control variables that

(27)

were used in the employee questionnaire were gender, age, education, country of employment, organizational tenure and industry of the organization. For this specific analysis, especially the control variables of gender, age, organizational tenure and supervisor tenure were considered in line with the analysis of Stouten et al. (2013), as they also analyzed the curvilinear relationship between ethical leadership and OCB. Additionally, I controlled for supervisor tenure because it relates to ethical leadership (r = -0.28, p = 0.01). Employee gender and age were also included as controls, as gender relates to moral identity (r = 0.37, p = 0.00) and age to OCB (r = 0.25, p = 0.01). Manager age, gender, and tenure were not controlled for because they have not been found to relate to ethical leadership or OCB in this sample.

Data Analysis and Results

This section will first present the steps that are taken to perform the data analysis based on the collected data and will then present the results of the performed analyses in detail. Finally, an exploratory analysis is performed and presented.

Data analysis

The data that was collected through the different questionnaires was analyzed with IBM’s Statistical Package for Social Sciences (SPSS) version 23. First, the English and Dutch employee data sets were merged into one overall employee data file. This was also done with the leader data sets. Additionally, the resulting two data files were merged based on the code words that were entered at the beginning of the questionnaire and invented by the matching dyads. This resulted in one final data set, which included both the leader and the follower data. In total, 23 employee and leader responses had to be excluded due to missing code words, double entry, or missing of the majority of the responses. As the sample only

(28)

contained a few missing values for the data analysis, the missing data was listwise deleted, so that only cases without any missing data in one variable was analyzed.

Second, to prepare the data for the analysis, frequency checks were used for each variable to identify errors in the data and counter-indicative items were recoded. Subsequently, scale reliabilities were computed. All of the variable scales show a high reliability. The Cronbach’s Alpha for ethical leadership is 0.91, for OCB is 0.87, for subordinates’ moral identity is 0.81, and for coworkers’ moral behavior is 0.83. Also the corrected item-total correlation indicates a good correlation of the individual items with the overall score of the scale (all above 0.30). Also for all of the variables, reliability would not substantially be affected if one of the items would be deleted and therefore no item needed to be excluded from the scale. Next, scale means of the main variables were computed by averaging all of the scale items of each variable scale.

To test the hypotheses a correlation analysis and regression analyses were performed. The correlation analysis was used to measure the strength and the meaning of the relationship between different variables. For the first hypothesis a linear regression analysis was used, including a squared independent variable for ethical leadership behavior to test the proposed curvilinear relationship between ethical leadership and OCB. The proposed moderation and mediation effect on the curvilinear relationship was analyzed using PROCESS (Hayes, 2015). The correlation and regression analysis included each the following control variables: employee gender, age and tenure and the supervisor tenure.

Results

Descriptive Statistics

Means, standard deviations and correlations of the used variables are presented in Table 1. Table 1 indicates that the leaders in the sample show on average a high level of

(29)

ethical leadership behavior (M = 5.42; SD = 0.90). Additionally, employees are on average seen to have a high level of OCB (M = 5.47; SD = 0.62) and a high level of moral identity (M = 5.32; SD = 1.01). The coworkers’ moral behavior is perceived to be on average high as well (M = 4.84; SD = 1.11).

Skewness, kurtosis and normality tests were performed to see how accurate the data is. These tests indicate that all four main variables are moderate negatively skewed. Thus, the majority of the respondents reported relatively high levels of OCB, ethical leadership, moral identity and coworkers’ moral behavior and therefore. 68% of the values for OCB within the dataset lie between 4.85 and 6.09, which is one standard deviation below and above the mean. 95% of the values for OCB lie between 4.23 and 6.71, which is two standard deviations below and above the mean. Thus, 67 and 94 supervisors reported high level of OCB of their employees, respectively. For ethical leadership, 68% of the values lie between 4.52 and 6.32 and 95% lie between 3.62 and 7.22. Thus, 67 and 94 employees reported their leaders to be high in ethical leader behavior, respectively. 68% of the values for subordinate moral identity lie between 4.31 and 6.33 and 95% lie between 3.30 and 7.34. Therefore, 67 and 94 employees rated their moral identity to be rather high, respectively. For coworkers’ moral behavior, 68% of the values lie between 3.73 and 5.95 and 95% lie between 2.62 and 7.06. Thus, 67 and 94 employees rated their coworkers to be relatively high in moral behavior, respectively. This negative skewness can be explained by the fact that the dyads that participated in the study seem to have a good relationship with each other. Thus, it could be the case that mostly participants that had a good relationship with each other were open to participate in the study or were open to approach either his/her leader or employee for participation.

The test to indicate that the data met the assumption of collinearity indicated that multicollinearity was not a concern. For all variables, the tolerance level is larger than 0.1 and

(30)

the Variance inflation factor (VIF) is below 10 (Ethical leadership, Tolerance = 0.89, VIF = 1.11; Subordinate Moral Identity, Tolerance = 0.99, VIF = 1.00; Coworkers’ moral behavior, Tolerance = 0.89, VIF = 1.11).

Consistent with expectations, the correlation between coworkers’ moral behavior and ethical leadership is positively and significant, r (97) = 0.32, p = 0.001. Contradicting expectations, the correlation between coworkers’ moral behavior and OCB is positive but not significant. Also, the correlation between subordinate moral identity and OCB and the correlation between ethical leadership and OCB is positive but not significantly. This is not in line with expectations and earlier findings.

Table 1. Means, Standard Deviations, Correlations.

Note. N = 99. The 7-point scales range from 1 (totally disagree) to 7 (totally agree). Reliabilities are reported along the diagonal.

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

Table 2 shows the regression results to test for the first hypothesis that ethical leadership behavior has a negative curvilinear relationship (inverted U-shape) with subordinates OCB. In the first step of the regression analysis, the predictors of employees’ gender, age, tenure, supervisor tenure and ethical leadership were included in the model. This model was statistically significant F (5, 90) = 2.92; p = 0.02 and explained 14% of variance in OCB. In the second step EL2 has been included in the model of the analysis and does show

(31)

statistically significant results F (6, 89) = 3.37; p < 0.01. The whole model explained a total variance of 19%. By including squared ethical leadership into the model, additional 5% of variance in OCB is explained after controlling for employees’ gender, age, tenure and supervisor tenure (R2 change = 0.05; F (1, 89) = 4.97, p = 0.03). This indicates that there is a curvilinear relationship between EL and OCB. However, different than expected is the Beta value of EL2 significantly positive (b = 1.51, p = 0.03), which implies that the trend in the quadratic effect is going up. This would mean that the slope is going upwards instead of downwards as predicted (see Figure 2). Specifically, this would indicate that at lower levels of ethical leadership there would be a negative relation with OCB, however, at high levels of ethical leadership OCB would increase. Thus, Hypothesis 1 cannot be supported.

Table 2. Regression results for ethical leadership.

(32)

Figure 2. The curvilinear relationship of ethical leadership and OCB.

Moral Identity

To test whether moral identity is a moderator of the curvilinear relationship between ethical leadership and OCB (see Hypothesis 2), a regression analysis was performed using PROCESS. The results for this moderation analysis are presented in Table 3. In line with the procedure suggested by Hayes (2015) to perform the moderation analysis for a curvilinear relationship by using model 1 and tricking PROCESS, first, OCB was inserted as the dependent variable, ethical leadership squared as the independent variable and moral identity as the moderation variable. Second, ethical leadership and the interaction of ethical leadership and moral identity as covariates were inserted. Additionally, the same control variables as for the analysis of the first hypothesis were included in the analysis: employee gender, age, tenure and supervisor tenure.

If the amount of variance explained in the dependent variable significantly increases through including the interaction term, then moral identity can be identified to be a moderator

(33)

of the curvilinear relationship between ethical leadership behavior and OCB. However, the regression coefficient for X2_{M (Ethical leadership squared x Subordinate moral identity) is} not statistically different from zero, t (86) = 0.03, p = 0.57. Furthermore, by including this interaction term, the amount of variance explained does not significantly increase (R2 change = 0.00; p = 0.57). Thus, moral identity cannot be identified to be a moderator of the curvilinear relationship of ethical leadership and OCB and therefore, Hypothesis 2 cannot be supported.

Table 3. Moderation moral identity.

Note. N = 99. Control variables: employee gender, age, tenure and supervisor tenure.

Coworkers’ moral behavior Moderation

A regression was performed using PROCESS to test whether coworkers’ moral behavior is a moderator of the curvilinear relationship between ethical leadership and OCB (see Hypothesis 3). The results for this moderation analysis are presented in Table 4. The same steps were performed as for the moderation of moral identity: OCB was inserted as the dependent variable, ethical leadership squared as the independent variable and coworkers’ moral behavior as the moderation variable. Furthermore, ethical leadership and the interaction of ethical leadership and coworkers’ moral behavior were inserted as covariates. Additionally, the same control variables were included in this analysis: employee gender, age, tenure and supervisor tenure.

(34)

If the amount of variance explained in the dependent variable significantly increases through including the interaction term, then coworkers’ moral behavior can be identified to be a moderator of the curvilinear relationship between ethical leadership behavior and OCB. However, the regression coefficient for X2CMB (Ethical leadership squared x Coworkers’ moral behavior) is not statistically different from zero, t (86) = -0.16, p = 0.87. Furthermore, by including this interaction term, the amount of variance explained does not significantly increase (R2_{change = 0.002; p = 0.87). Therefore, also coworkers’ moral behavior cannot be} identified to be a moderator of the curvilinear relationship of ethical leadership and OCB and Hypothesis 3 cannot be supported.

Table 4. Moderation coworkers’ moral behavior.

Note. N = 99. Control variables: employee gender, age, tenure and supervisor tenure.

Mediation

To test whether coworkers’ moral behavior is a mediator of the curvilinear relationship between ethical leadership and OCB (see Hypothesis 4), a hierarchical regression analysis was performed. The results for the regression analysis are reported in Table 5. For this analysis again PROCESS was used, however, to analyze the mediation model 4 was used. First, OCB was inserted as the dependent variable, ethical leadership squared as the independent variable and coworkers’ moral behavior as the mediation variable. Second, ethical leadership and the interaction of ethical leadership and coworkers’ moral behavior were inserted as covariates. Additionally, the same control variables as for the moderation

(35)

analyses were included in this analysis: employee gender, age, tenure and supervisor tenure. The outcomes for the curvilinear mediation analysis were tested on accuracy using MEDCURVE. The outcomes of the analysis using PROCESS and MEDCURVE show exactly the same results.

The total effect of ethical leadership on OCB is c = 0.09, meaning that two leaders who differ one unit in ethical leadership are estimated to occupy employees who differ by 0.09 units in their OCB. This total effect is significantly different from zero, b = 0.09, t = 2.04, p = 0.04 (path c, Figure 3). The positive sign indicates that the person perceiving a greater amount of ethical leadership behavior of the leader is perceived to show a higher amount on OCB. Also, ethical leadership does explain a significant proportion of variance in OCB scores, R2 = 0.19, F (7,88) = 2.89, p < 0.01.

Figure 3. Model of the total effect of ethical leadership on OCB.

The direct effect of ethical leadership on coworkers’ moral behavior, a = 0.02, would indicate that ethical leadership predicts coworkers’ moral behavior. However, this direct effect is not statistically significant, b = 0.02, t = 1.84, p = 0.07 (path a, Figure 4), and therefore cannot be interpreted. The direct effect of coworkers’ moral behavior on OCB, b = 0.52, would indicate that coworkers’ moral behavior would predict the employees OCB. However, also this effect is not statistically significant, b = 0.52, t = 1.18, p = 0.24 (path b, Figure 4). The direct effect of ethical leadership on OCB, c’ = 0.08, is the estimated difference in OCB between two employees experiencing the same level of coworkers’ moral behavior but whom supervisors differ by one unit in their ethical leadership behavior. This

(36)

would indicate that employees with an ethical leader are more inclined to have higher OCB, keeping the coworkers’ moral behavior constant. However, this direct effect is not statistically significant, b = 0.08, t = 1.77, p = 0.08 (path c’, Figure 4).

Figure 4. Model of ethical leadership as a predictor of OCB, mediated by coworkers’ moral behavior.

Finally, the indirect effect of 0.01 is not statistically different from zero when looking at the 95% BC bootstrap confidence interval. When the bootstrap interval fails to include 0, then the p-value is said to be lower or equal to 0.05 and the effect is said to be significant. Given this, ethical leadership does not have a significant indirect effect on OCB through coworkers’ moral behavior, b = 0.01, BC CI [-0.0123, 0.0795]. Thus, even though the total effect is significant, the in depth analysis shows that coworkers’ moral behavior is not mediating the curvilinear relationship between ethical leadership and OCB and therefore, Hypothesis 4 cannot be supported.

(37)

Table 5. Consequent of the regression analysis.

Note. N = 99. Control variables: employee gender, age, tenure and supervisor tenure. * p < .05.

** p < .01.

Exploratory analysis

Additionally to the performed analyses with the sample of 99 dyads, this sample was reduced to N = 95 by excluding the four outliers, who showed high OCB when working with a low ethical leader. These four outliers could be easily identified by plotting a scatter graph and exclude the dots with their corresponding participants from the sample. This was done to test whether the relationship indeed shows a curvilinear relationship or a linear relationship instead.

In the first step of the regression analysis, the predictors of employee gender, age, tenure, supervisor tenure and ethical leadership were included in the model. This model was statistically significant F (5, 86) = 3.92; p = 0.00 and explained 18.6% of variance in OCB. In the second step EL2 has been included in the model of the analysis and does show statistically significant results F (6, 85) = 3.29; p = 0.01. The whole model explained the total variance of 18.9%. By including squared ethical leadership into the model, no significant additional variance in OCB is explained after controlling for employee gender, age, tenure and supervisor tenure (R2_{change = 0.003; F (1, 85) = 0.32; p = 0.57). This indicates that there is} no curvilinear relationship between EL and OCB. Specifically, this would indicate that the relationship between ethical leadership and OCB is linear instead of curvilinear. Indeed, when analyzing the linear relationship of ethical leadership and OCB by performing a linear regression analysis with EL2 as the independent variable and OCB as the dependent variable

(38)

and covariates of employee gender, age, tenure and supervisor tenure, this analysis does show a significant linear relationship (b = 0.31; p = 0.00).

Table 6. Regression results for ethical leadership.

Note. N = 95. ** p < .01.

To test whether moral identity or coworkers’ moral behavior moderates this linear relationship of ethical leadership and OCB, a regression analysis was performed using PROCESS. This time, as no curvilinear relationship is analyzed, PROCESS does not need to be tricked. However, still model 1 was used. For the first moderation analysis, ethical leadership was inserted as the independent variable, OCB as the dependent variable and moral identity as the moderator variable. Additionally, the employees’ gender, age, tenure and supervisor tenure was included as control variables. However, the interaction term for XM (Ethical leadership x Subordinate moral identity) is not statistically different from zero, t (84) = -0.54, p = 0.59. Furthermore, by including this interaction term, the amount of variance explained does not significantly increase (R2 change = 0.002; p = 0.59). Thus, moral identity cannot be identified to be a moderator of the linear relationship of ethical leadership and OCB. For the second moderation analysis, the same steps were performed, however, coworkers’ moral behavior was inserted as the moderation variable. Also this interaction term for XM (Ethical leadership x Coworkers’ moral behavior) is not statistically different from zero, t (84) = -0.32, p = 0.75. Furthermore, by including this interaction term, the amount of

(39)

variance explained does not significantly increase (R2 change = 0.001; p = 0.75). Thus, coworkers’ moral behavior also cannot be identified to be a moderator of the linear relationship of ethical leadership and OCB.

Next, a mediation analysis was performed to see whether coworkers’ moral behavior does influence the linear relationship of ethical leadership and OCB indirectly. This was done by using PROCESS model 4 and by inserting ethical leadership as the independent variable, OCB as the dependent variable and coworkers’ moral behavior as the mediation variable. The same control variables as for the other analyses were included in the analysis: employee gender, age, tenure and supervisor tenure. Even though significant direct effects of ethical leadership on coworkers’ moral behavior (b = 0.68; t = 4.72; p = 0.00) and OCB (b = 0.24; t = 2.63; p = 0.01) and a total effect of EL on OCB (b = 0.25; t = 3.10; p = 0.00) could be found, an indirect effect could not be identified (b = 0.01, BC CI [-0.0640, 0.0766]). Thus, it can be concluded that coworkers’ moral behavior is also not a mediator of the linear relationship of ethical leadership and OCB.

Discussion

This research was designed to test unintended negative consequences that ethical leadership might have on a positive outcome of OCB of this increasingly important leadership style. Through the current study it was intended to answer the following research questions: To what extent has ethical leadership unintended negative relationships with the subordinates’ work behavior? Especially, how is ethical leadership behavior related to subordinates OCB? And to what extent does the subordinates’ moral identity and perceived coworkers’ ethical behavior moderate this relationship? The aim of this paper was first to investigate whether ethical leadership behavior had a curvilinear relationship with OCB instead of linear, as found in one previous study (Stouten et al., 2013). Thereby, the positive and negative sides of ethical

(40)

leadership are researched in more detail to contribute to the current debate about ethical leadership. Second, it was meant to explore whether moral identity serves as a moderator and coworkers’ moral behavior as a moderator and mediator of this specific relationship.

In line with expectations, a curvilinear relationship between ethical leadership and OCB was found. However, this curvilinear relationship was of a positive nature instead of negative as predicted and found in an earlier study (Stouten et al., 2013). Contrary to the predictions, moral identity and coworkers’ moral behavior did not moderate the curvilinear relationship between ethical leadership and OCB. Also, inconsistent with the hypothesis that ethical leadership influences OCB indirectly through coworkers’ moral behavior, coworkers’ moral behavior is not found to mediate this curvilinear relationship.

Even though the hypotheses cannot be supported, the findings of this study have important theoretical and practical implications, which will be discussed in greater detail in the following section. Additionally, future research possibilities and potential limitations of this study will be addressed.

Theoretical implications and future research directions

All in all, the current study increases the empirical knowledge about ethical leadership behavior, which is increasingly a topic of interest in organizational settings. Furthermore, this research adds to the understanding of the current debate about positive, but also unintended negative roles that ethical leadership can play in regard to generally positive outcomes, such as OCB. Specifically, this research questions the findings that high ethical leadership has unintended effects on employees’ OCB (Stouten et al., 2013). As positive and negative sides of ethical leadership are taken into account, a wider concept of ethical leadership is considered than most previous research did.

Leader-follower fit : the curvilinear relationship of ethical leadership and OCB and the moderating effects of subordinate’s moral identity and coworkers’ moral behavior

Statement of originality

Table of contents