leadership and its follower outcomes.
Tosse Bontje
10871640
MSc. Business Administration – Leadership & Management Track
Thesis supervisors:
1st: Dr. A.H.B. de Hoogh 2nd: Prof. Dr. D. N. den Hartog
Submission date: Tuesday June 21, 2016 Master thesis: final version
Institution: Amsterdam Business School/University of Amsterdam
Statement of originality
This document is written by Tosse Bontje, 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.
Table of Contents
1. Abstract...5
2. Introduction ...6
3. Theory and Hypotheses...9
3.1 ETHICAL LEADERSHIP AND FOLLOWER ETHICAL BEHAVIOR ... 9
3.2 THE MEDIATING ROLE OF LEADER-‐MEMBER EXCHANGE ... 12
3.3 THE MODERATING ROLE OF LEADER MORAL IDENTITY INTERNALIZATION ... 15
3.4 ANOTHER ROLE FOR LEADER MORAL IDENTITY INTERNALIZATION ... 19
4. Methods... 21
4.1 PROCEDURE AND SAMPLE... 21
4.2 MEASURES... 22
5. Results ... 23
5.1 PRELIMINARY ANALYSES AND DATA SCREENING... 23
5.1.1 Missing values and outliers... 23
5.1.2 Reliability tests... 24
5.1.3 Normality tests... 24
5.1.4 Assumptions tests... 26
5.2 DESCRIPTIVE STATISTICS... 27
5.3 REGRESSION ANALYSES... 28
5.3.1 A direct relationship and LMX as mediator... 29
5.3.2 Moderated mediation... 31
5.3.3 Leader moral identity internalization as predictor... 31
6. Discussion ... 33
6.1 THEORETICAL IMPLICATIONS... 33
6.2 STRENGTHS, LIMITATIONS AND FUTURE RESEARCH... 36
6.2.1 Strengths... 36
6.2.2 Limitations ... 36
6.2.3 Future research suggestions ... 37
6.3 PRACTICAL IMPLICATIONS... 39
7. Conclusion... 41
8. References... 42
APPENDIX B: SURVEY SCALES... 50
APPENDIX C: HISTOGRAMS AND Q-‐Q PLOTS... 52
APPENDIX D: SCATTERPLOTS... 56
APPENDIX E: SCATTERPLOTS OF RESIDUAL TERMS... 57
List of figures Figure 1 – Conceptual model
... 9
List of tables Table 1 – Values, standard errors and Z-‐scores for skewness and kurtosis
... 25
Table 2 – Descriptive statistics
... 27
Table 3 – Results mediation analysis and direct relationship
... 29
Table 4 – Results of moderated mediation analysis
...30/31
Table 5 – Results mediation analysis... 32
Acknowledgements
...41
1. Abstract
This paper examines the influence of true morality on ethical leadership and its follower outcomes. Leader-‐member exchange is proposed as a mediating mechanism in the relationship between ethical leadership and follower ethical behavior. Furthermore, leader moral identity internalization is suggested to have a moderating effect on these relationships. The data was gathered by means of a dyadic (leader-‐follower) survey study (n = 99). Results show that ethical leadership is directly and positively related to follower ethical behavior. So the more ethical a leader, the more ethical the behavior of his or her followers will be. Moreover, the results indicate that ethical leadership leads to better leader-‐member exchange which in turn leads to more ethical follower behavior. Leader moral identity internalization is not found to be a moderator on either of these relationships. Theoretical implications for the existing literature, practical implications for directors, management and organizations in general as well as strengths, limitations and suggestions for future research are discussed.
Keywords: Ethical leadership • Leader-‐member exchange • Follower ethical behavior • Leader
morality
2. Introduction
Volkswagen and its deceiving software: certainly the most striking and eye-‐catching, but definitely not the only, ethical business scandal in recent history. Volkswagen recently admitted to having installed deceiving software to cheat on emission tests on approximately 11 million cars. “For business leaders, the choice is: Do they want to leave behind a legacy of success and integrity? Or, do they want a legacy that built the bottom line but sacrificed core values?” (Fielkow, 2015, p.1). In a world where (un)ethical behavior is increasingly brought to our attention, the role of leaders is inevitably becoming more salient too (for a recent ethical leadership review see: Den Hartog, 2015). Following a widely adopted definition by Brown, Treviño and Harrison (2005, p.120), ethical leadership (EL) can be defined as “the
demonstration of normatively appropriate conduct to followers through personal actions and interpersonal relationship and the promotion of such conduct to employees through two-‐way communication, reinforcement, and decision-‐making”. Not only do the leaders set the limits for what behavior is rewarded or punished, they are also likely to be a role model through the relationship they have with their followers (leader-‐member exchange).
As the above-‐mentioned definition points out, part of EL is about an interaction between a leader and its followers. Recent research, based on social exchange theory, suggests that followers reciprocate EL behavior (Hansen, Brown, Jackson & Dunford, 2013). In other words, the behavior demonstrated by ethical leaders is likely to be copied by that leader’s followers. In line, ethical leaders affect their followers by being a role model. This social learning perspective contends that followers emulate the conveyed identity of an ethical leader and integrate
important elements thereof into their own identity (e.g. trust, shared values, integrity) (Brown et al. 2005).
Building on this identification-‐based process and considering the potential of a leader, this paper proposes an influence of true leader morality on follower outcomes. It is not hard to imagine a situation in which a leader is subject to numerous external factors strongly
themselves as being ethical even though this is not in line with their true personality and
convictions. Den Hartog (2015) describes the gap in the literature on this topic by noting that we are better informed about ethical leaders as moral managers (as perceived by their followers) than as true moral persons.
In this study, I try to address this issue by researching the authenticity of leaders through moral identity internalization (LMII). In other words, the degree to which a person’s morality is truly embedded in ones self-‐concept (Aquino & Reed, 2002). For a leader to be perceived as truly ethical one needs to be either a very good actor or a true moral person. Since followers are likely to detect inauthentic signals (e.g. Zapf, 2002), higher LMII is expected to yield a more positive influence of EL on the relationship one has with its follower. In other words, the
relationship between ethical leadership (follower-‐rated) and leader-‐member exchange (LMX) as perceived by followers is likely to be affected by the true morality of a leader (leader-‐rated). Similarly, due to an increased moral judgment and sense of responsibility, also the direct effect of EL on follower ethical behavior is likely to be stronger for high levels of leader morality than for lower levels.
Recently, researchers have tested and found evidence for the moderating influence of this true leader morality (operationalized as leader authenticity and leader Machiavellianism) on the relationship between EL and follower outcomes (e.g. psychological empowerment, personal initiative, counterproductive behavior) (Zhu, May & Aviolo, 2004, Den Hartog & Belschak, 2012). Nevertheless none of these studies used LMII and none of these studies used follower ethical behavior as an outcome. Using LMII allows for the comparison of leader and follower rated ethicality and morality of the leader; therefore this paper contributes to the existing literature by distinguishing itself in several important ways.
First, the main contribution of this study is linking leader moral identity internalization as a moderator to the relationship between EL and follower ethical behavior. Even though it is clear that leader traits are important influencers of follower behavior (e.g. Kirkpatrick & Locke, 1991), LMII has not been researched yet in this role linking it to ethical follower outcomes.
Leader authenticity operationalized as LMII has been established as an antecedent of EL (Mayer, Aquino, Greenbaum, Kuenzi, 2012) and as an independent variable influencing moral behavior (Reynolds & Ceranic, 2007). However, to the best of my knowledge no studies to date have adopted LMII as a possible moderator in the processes surrounding leader and follower ethicality.
Second, most studies have used follower assessments of leader moral identity whereas this paper will use a leader self-‐measurement (notable exception: Den Hartog & Belschak, 2012). By doing so, this study could reveal possible discrepancies between leader intentions and
follower perceptions and thereby be of added value to inform both existing literature and practice.
Third, this paper is the first to study the mediating role of LMX in the relationship between ethical leadership and follower ethical behavior. Using LMX could indicate the importance for a leader of establishing a good relationship with followers. Finally, this study contributes to the examination of the relationship between EL and ethical outcomes since few studies have done this to date (Mayer et al. 2012). This, in turn, allows for a better general understanding of ethical leadership.
The research questions related hereto are as follows: “How does ethical leadership relate to follower ethical behavior and what is the mediating role of leader-‐member exchange in this relationship?” and “What is the role of leader moral identity internalization in relation to ethical leadership, leader-‐member exchange and follower ethical behavior?”
In sum, the main purpose of this paper is to further understand the influence of true morality on ethical leadership and its follower outcomes. True morality, in this case, would be indicated by the extent to which moral identity is internalized in ones believes and convictions. This will be realized by elaborating on the direct relationship between EL and follower ethical behavior, the mediating role of LMX and the moderating influence of leader moral identity as visualized in the conceptual model as presented in Figure 1. An ethical leader’s positive influence on employees is based on elements of trust, role modeling and honesty, but it all
depends on the true sincerity and morality of the leader. This paper does, thereby, reply to the call by Brown and Mitchell (2010, p.599), who pose that “research is needed to see if moral identity, particularly the internalization dimension, can serve to benefit ethics in organizations generally”.
Figure 1 - Conceptual Model
3. Theory and Hypotheses
3.1 Ethical Leadership and Follower Ethical Behavior
Bass (1985) was one of the first to devote attention to EL by noting that leaders could use their power for the sake of moral and immoral purposes. Building on this moral aspect of EL, Treviño, Brown and Hartman (2000, 2003) developed two constructs: the moral person and the moral manager. The first refers to the character, personality traits and altruistic motivation of an ethical manager both at work and in their personal lives. Typical personal qualities they referred to are honesty, fairness, concern for others and trustworthiness. The second, the moral manager, is focused on how leaders use their influence on followers to promote and encourage ethical follower behavior. They do so by leading by example as an ethical role model, by frequently communicating ethics, setting ethical standards and by using rewards/discipline whenever
Especially the moral manager dimension is relevant in explaining the direct relationship between EL and follower ethical behavior. The moral manager is one who proactively influences the ethical and unethical behavior of followers. From a social learning perspective, followers copy the behavior of ethical leaders because they provide credible and attractive role models (Brown & Treviño, 2006). They are perceived as such through their power and status (Bandura, 1986). They show what behavior is to be seen as ethical and communicate this through the messages they send. Moreover, ethical leaders are said to “walk the ethical talk”. Since people are most likely to seek for ethical guidance from others than themselves (Kohlberg, 1969), leader’s attitudes and behaviors are likely to be mimicked by individuals through social learning (Bandura, 1986). Additionally, from a social contagion perspective (Luthans & Norman, 2006), followers are likely to pick up on positive states and behaviors like keeping your promises, allowing input in decision-‐making and treating others with respect. Based on the
aforementioned, the followers of these moral managers are able to observe ethical behavior, which allows them to develop their own sense of ethical thinking and acting most probably in line with their ethical counterparts.
De Hoogh and Den Hartog (2009) take a social influence perspective on the impact that EL has on followers. From this perspective ethical leaders are driven by moral beliefs and caring values. Here the role of the leader-‐follower relationship (LMX) is an interesting construct, which will be elaborated on in the next section. Prior research shows that moral reasoning of an individual can have an influence on the moral reasoning of groups (Dukerich, Nichols, Elm & Vollrath, 1990). Based on the above and considering the fact that ethical leaders score high on moral reasoning, I propose that the moral reasoning of ethical leaders will result in more ethical follower behavior.
In addition, Brown and Treviño (2006) highlighted the importance of reinforcement on the impact that role models have. Followers are shown what is viewed as ethical and unethical standards and what decisions are rewarded and disciplined, respectively. Hence, followers of
ethical leaders will be familiar with the implications of their choices and subsequently make more ethical decisions.
Over the past decades scholars have developed a broad body of knowledge surrounding ethical leadership. Take for example the antecedents on individual and contextual level as clearly summarized by Den Hartog (2015). Individual antecedents of ethical leadership are some of the big five personality traits (e.g. conscientiousness and agreeableness, Kalshoven, Den Hartog & De Hoogh, 2011a), moral identity (Mayer et al. 2012) and cognitive moral development (Jordan, Brown, Treviño & Finkelstein, 2013) and some contextual level antecedents are culture (House, Hanges, Javidan, Dorfman & Gupta, 2004) and personal characteristics of the followers (Resick et al. 2011).
In the same review Den Hartog (2015) identifies four categories of mediators in the relationship between ethical leadership and its outcomes. Some examples of such mediators are moral judgment, engagement, responsibility and LMX (e.g. Resick, Hargis, Shao & Dust, 2013, Den Hartog & Belschak, 2012, Kalshoven, Den Hartog, De Hoogh, 2013, Walumbwa et al. 2011). The latter will be discussed extensively in the next chapter. In addition, a recent meta-‐analysis of EL found that followers’ trust in the leader mediates the relationship between ethical leadership and performance and job attitudes (Ng & Feldman, 2015).
Furthermore, previous research on EL and its outcomes revealed a negative relationship between EL and cynicism and deviance and a positive relationship with followers showing desired work behaviors, reciprocating role models, satisfaction with the leader, LMX, willingness to report problems to management, effectiveness, performance and effort (e.g. Brown et al. 2005, Den Hartog & De Hoogh 2009, Hansen et al. 2013, Piccolo, Greenbaum, Hartog & Folger 2010, Kalshoven & Boon 2012, Walumbwa et al. 2011). In line, in their meta-‐analysis Ng and Feldman (2015) showed EL to predict task performance, citizenship behavior and
counterproductive work behavior.
Thus, this study adds to this already well-‐informed body of knowledge by examining the direct relationship between EL and follower ethical behavior. When leaders show ethical
behavior, provide ethical guidance, display moral reasoning and use reinforcement to steer ethical behavior, norms for acceptable behavior are established and followers will be likely to engage in ethical behavior. Therefore, I hypothesize:
Hypothesis 1 Ethical Leadership is positively related to Follower Ethical Behavior
3.2 The mediating role of Leader-Member Exchange
Leader-‐member exchange (LMX) refers to the exchange of resources and to the extent to which emotional support is experienced by a leader and its followers (Liden, Sparrowe &
Wayne, 1997). This social exchange can be seen as mutual and as developing over time. Based on social exchange theory, high-‐quality LMX is developed by employees based on three
contingencies (with whom they interact, how this interaction takes place and the experiences resulting from this interaction). In their meta-‐analysis, Ilies, Nahrgang and Morgeson (2007) showed the importance of developing a high-‐quality LMX. They revealed that LMX is an important predictor of individual behaviors (e.g. follower ethical behavior). Especially when a high-‐quality LMX is in place, employees develop the tendency to reciprocate (ethical) behavior, even if they will have to go beyond their required in-‐role-‐behavior (Ilies et al. 2007).
In the general discussion of their findings, Hansen and colleagues call for future research that “links ethical leadership to other outcomes, especially those that are focused on ethics instead of employee job attitudes and job performance” (Hansen et al. 2013, p.445). Therefore, it is suggested here to research the role of LMX in linking EL to an ethics related outcome, namely follower ethical behavior. Besides researching the direct relationship between EL and follower ethical behavior, this paper therefore proposes a mediating role of LMX in that relationship.
In general, ethical leaders bring about positive employee outcomes due to their
trustworthiness, their care for their employees and their fairness (Neubert et al. 2009, Brown et al. 2005). High-‐quality LMX is determined by the emotional support given to one another and the exchange of valued resources between leader and follower. EL is expected to link positively
to high-‐quality LMX for a variety of reasons. First of all, because ethical leaders are considered to be honorable, trustworthy, fair, to have concern for their employees and to be people-‐oriented (Kalshoven, Den Hartog & De Hoogh, 2011b, Resick et al. 2006). Due to these positive influences followers are expected to perceive their relationship with their leader as being of high quality.
Second, building upon the ‘norm of reciprocity’ of social exchange theory, it is likely a high-‐quality LMX develops (Liden et al. 1997). This norm entails that individuals that are treated well by others feel the obligation to return the favor in a positive manner and since high-‐quality LMX emerges through a series of mutual exchanges this will most likely be the case. Third, ethical leaders are committed to power sharing by giving followers a say in decision-‐making and by carefully considering their ideas (De Hoogh & Den Hartog, 2009). The sense of appreciation that employees derive from this is likely to result in a high-‐quality LMX. Finally, following the logic as proposed by Walumbwa et al. (2011), ethical leaders are expected to build relationships with their followers that go beyond the economic exchanges by encouraging voice and using reinforcement. As a result, ethical leaders facilitate high-‐quality LMX due to the way they connect with their followers.
From a social exchange point of view, several authors recently have proven that EL is positively related to LMX. For example, Walumbwa et al. (2011) found EL to be positively related to LMX and, moreover, they found LMX to mediate the relationship between EL and employee performance. Similarly, Hansen et al. (2013) found EL to relate positively to LMX and found a significant mediating influence of LMX on the relationship between EL and employee
commitment. They used social exchange theory to explain how the underlying constructs of social relationships function with regard to EL and its influence on followers. Based upon the aforementioned, I expect a positive relationship between EL and LMX. Hence, the following is hypothesized:
Hypothesis 2a Ethical Leadership is positively related to ‘Leader-‐Member Exchange’
As soon as this high-‐quality relationship is in place, followers are expected to reciprocate by returning the favor for their leader (the so-‐called norm of reciprocity) (Blau, 1964, Gouldner, 1960, Wayne & Green, 1993). As Hansen et al. (2013) describe it, once a person feels that an organization or a person (e.g. ethical leader) has treated him or her positively, one tends to reciprocate. Similarly, Den Hartog (2015) mentions in her review, that followers emulate ethical leaders by focusing on trust, shared values and integrity. These aspects (leading to a high-‐quality LMX), in turn, are expected to exert their influence on follower ethical behavior.
Furthermore, high exchange relationships result in follower loyalty towards leaders (Mahsud, Yukl & Prussia, 2010), which in turn is expected to lead to follower behavior that is consistent with that of the leader. In addition, Ilies, Morgeson and Nahrgang (2005), in their study on authentic leadership, showed that having a high-‐quality relationship not only fosters reciprocity, but also leads to value congruence between leader and follower. Subsequently followers show behavior that is consistent with the leader’s values. So in the case of EL,
followers who have a high-‐quality LMX with their leader are also expected to behave in the same ethical way their leaders do.
Moreover, the basis of decision-‐making lies in the development of mental models serving as a framework in a response to ethical events. These mental models serve as a means to scan the environment for information, evaluating it and determining the appropriate course of action (Hogarth & Makridakis, 1981). If a high-‐quality LMX, rather than a low-‐quality LMX, is in place, ethical leaders are most likely very positive influencers in this process of creating mental models for decision-‐making. Those very same ethical leaders are, from a moral manager dimension, proactively influencing followers’ ethical decision-‐making through their good relationship. Thus, it is realistic to assume that EL results in a high-‐quality LMX and that follower ethical behavior (e.g. decision-‐making) is positively affected by this high-‐quality LMX.
Research to date does not contain any studies directly researching the link between LMX and follower ethical behavior. The only studies that did research ethical leadership and used
LMX as a possible mediator, used employee performance and commitment as its outcomes (Walumbwa et al. 2011, Hansen et al. 2013).
This study therefore examines the mediating role of LMX in the relationship between EL and follower ethical behavior. As the unique character of the LMX construct is its dyadic
relationship as the level of analysis (Gerstner & Day, 1997), this study fits perfectly to measure it, as will become clear in the next chapters. To the best of my knowledge, this is the first attempt to test social exchange as a mediator in this relationship, linking EL to an ethical follower
outcome. EL is expected to positively relate to high-‐quality LMX, which in turn is expected to relate positively to follower ethical behavior. Due to other processes like social learning and social identity (e.g. Walumbwa et al. 2011) and possible personal/contextual influences, only partial, rather than full mediation of LMX is expected. Thus, I hypothesize:
Hypothesis 2b ‘Leader-‐Member Exchange’ is positively related to Follower Ethical Behavior
Hypothesis 2c ‘Leader-‐Member Exchange’ partially mediates the relationship between Ethical
Leadership and Follower Ethical Behavior
3.3 The moderating role of Leader Moral Identity Internalization
Now be the above as it may, this part of the paper argues that both the direct as well as the mediated relationship as described are conditional, depending on whether a leader is being real or not. The behavior of people, leaders alike, can be either authentic or not. For example, a leader could show ethical behavior as a means of ‘window-‐dressing’ rather than pursuing a true personal ideology. In the case of window-‐dressing, it is expected a follower will eventually see a leader’s true colors if he or she is just pretending to be ethical. On the contrary, if leaders are truly convinced of certain moral standards one would expect them to be more genuine and sincere and their actions to have a stronger positive impact. This reasoning is in line with Mayer et al. (2012); these authors refer in their article to the concept of moral identity and its two
dimensions. One of the dimensions is moral identity symbolization, which represents the publicly showed aspect of moral identity (this one will not be used in this paper). The other one is the leader moral identity internalization (LMII) dimension, stating that an individual scoring high on this dimension represents morality that is truly embedded in that person’s self-‐concept (Aquino & Reed, 2002). Such individual is expected to behave in a moral way and encourage, reward and discipline others that behave (un)ethically. Furthermore, persons rating high on moral identity internalization view success from a deontological perspective and derive their personal sense of authenticity from the correctness of their own actions.
To date LMII has been researched in several ways. Olsen et al. (2006) showed a positive relationship between LMII and ratings of transformational leadership. Reynolds and Ceranic (2007) showed how moral identity and moral judgments interact in predicting ethical behavior and Aquino et al. (2008) showed that people scoring lower on moral identity as compared to higher tend to lie more during negotiations. Additionally, Mayer et al. (2008) showed that leader’s scores on both aspects of moral identity (internalization and symbolization) were positively related to ethical leadership (follower-‐rated).
In a similar vein, researchers established antecedents of LMII. The antecedents
discovered to date can be best categorized in the following categories: (1) institutional contexts, (2) participation in moral actions and (3) social interactions. First, Nasir and Kirshner (2003) found an important role for both institutional context as well as cultural practices in the development of ones moral identity. Second, a study by Colby and Damon (1992) showed that the dedication of a person to moral actions increased the moral development of that person. Third, social interactions are proven to shape the moral development of a person. For example, Colby and Damon (1992) showed that especially interactions with peers result in the
development of a person’s moral identity.
In addition, DeCelles, DeRue, Margolis and Ceranic (2012) found that persons scoring high on LMII are less inclined to show self-‐serving behavior. In other words, a person scoring high on LMII is assumed to behave in ways that are supportive of and beneficial to others. This is
likely to lead to more positive employee outcomes of EL (e.g. high-‐quality LMX, follower ethical behavior).
In line with the previous paragraphs, I expect the authenticity of a leader
(operationalized as moral identity internalization) to facilitate a condition under which the positive relationship between EL and LMX is more likely to be observed. The higher a leader scores on LMII, the more that leader’s demonstrated moral behavior is genuine and the higher the effect of EL on LMX will be. The latter reasoning is similar to Zhu et al. (2004), who propose that employees’ perception of authenticity functions as a moderator in the relationship between EL behavior and employee outcomes (e.g. trust in leaders).
To be authentic (scoring high on LMII), one needs to demonstrate behavior consistent with his or her personal held values, inner feelings and moral reasoning. As May, Chan, Hodges and Aviolo (2003) proposed, such leaders are transparent about their intentions and moral evaluations that make up their behavior. In other words, an authentic leader scoring high on LMII reveals his or her true intentions (based on moral reasoning and values) and shows
behavior aligned with these intentions. Zhu et al. (2004), pose that employees want to be treated with respect, authentically and fairly. Following their logic, the alignment of EL with truly held personal values and moral reasoning and considering the fact that employees are likely to pick up on inauthentic signals (Zapf, 2002), will result in employees responding more positively to EL when it’s genuine. This holds true both for the moderating influence of LMII as described above as well as LMII as a moderator on the direct relationship between EL and follower ethical behavior, as described next.
In addition, using Kohlberg’s (1969) theory of cognitive moral development (CMD), one could argue that leaders scoring high on LMII are in the highest stages of their moral
development. These persons determine what is right by looking at their internally held values (without considering a possibly diverging opinion of others) or by using deontological principles of what is right and wrong. Such self-‐determined thinking and decision-‐making is likely to yield high levels of respect and thereby strengthening the influence of EL on LMX.
As argued preceding Hypothesis 2, EL is expected to relate positively to LMX. Thus, the higher the leader scores on LMII, the stronger the positive relationship between EL and LMX will be. On the contrary, I also expect it to hold true that when a leader scores lower on LMII, the relationship between EL and LMX will also weaken.
Furthermore, as already briefly touched upon, I also propose a moderating role for LMII on the direct relationship between EL and follower ethical behavior. I believe this is the case for several reasons. First, as argued for Hypothesis 1, leaders provide credible role models for their followers through the exhibition of ethical behavior. However, even if leaders were to display EL, without true moral identity internalization they would be unable to motivate people to behave in similar ways as them (Morrison, 2001). Thereby weakening the direct effect EL has on follower ethical behavior.
Second, EL is expected to directly effect follower ethical behavior due to reinforcement used by ethical leaders. However, if ethical leaders do not “walk the ethical talk” themselves, their EL is most likely going to exert less influence on follower ethical behavior than it otherwise would. As Aquino and Reed (2002) posed, individuals high on LMII encourage, reward and discipline others that behave (un)ethically. In line, building upon the social identification process as previously mentioned (Ilies et al. 2005), the value congruence would be less strong if leaders show EL based on values that are not truly in line with ones convictions. Such individual is expected to be less supportive of his or her own actions. Since reinforcement applied by leaders is based upon their values and moral reasoning the direct effects of EL on follower ethical behavior are presumed to be stronger for high levels of LMII than for lower levels. A similar logic is applicable to the moderating influence of LMII on the EL-‐LMX relationship. Since high-‐quality LMX develops through a series of mutual exchanges, the result of these exchanges is probably weakened if an ethical leader is not entirely supportive of his or her own course of action.
Third, closely related to the previous paragraph, I argue that the reinforcement used by ethical leaders is more effective in general when leaders are high on moral identity. That is to
say, one of the ways in which EL influences follower outcomes is through reinforcement. Leaders scoring high on LMII pay more attention to and punish unethical follower behavior (Mayer et al., 2012). Therefore the effect that EL has on follower outcomes, such as ethical behavior, is
expected to be stronger when a leader high on LMII is in place.
Thus, adding to the existing research and based on the foregoing, I believe that followers will respond more positively in their judgment of EL if a leader’s behavior is genuine, sincere and authentic as well as in line with ones personal held values and convictions. Consequently, LMII is believed to both moderate the relationship between EL and LMX and the direct
relationship between EL and follower ethical behavior. Considering the above, this results in the following hypotheses:
Hypothesis 3 The mediated relationship of Ethical Leadership with Follower Ethical Behavior
will be moderated by ‘Leader Moral Identity Internalization’. Specifically, the relationship between Ethical Leadership and Leader-‐Member Exchange will be stronger for leaders scoring high on moral identity internalization than for leaders who score low.
Hypothesis 4 ‘Leader Moral Identity Internalization’ moderates the direct relationship between
Ethical Leadership and Follower Ethical Behavior, such that the relationship between Ethical Leadership and Follower Ethical Behavior is stronger when Leader Moral Identity
Internalization is high than when it is low.
3.4 Another role for Leader Moral Identity Internalization
As already acknowledged, previous research has examined LMII in another role than as a moderator. Mayer and colleagues (2012), in their study on the role of identity on group
processes, proved a positive relationship between leader moral identity internalization (leader-‐ rated) and ethical leadership (subordinate-‐rated), establishing LMII as an antecedent of EL. Following their logic, it is likely that leaders who score high on LMII convey to their followers,
through their behavior and decision-‐making, that they are true moral persons. As a result, followers are more likely to perceive those leaders as being ethical leaders, leading to positive follower outcomes (e.g. ethical behavior) as argued in Hypothesis 1.
Even though it is not the main focus of this paper, an additional analysis is conducted to see whether the findings as reported by Mayer et al. (2012) hold true on an individual level as well. More specifically, it is tested whether LMII proves to be an antecedent of EL as well, putting EL in the role of a mediator in the relationship between LMII and follower ethical behavior. There are several arguments supporting this idea, now the most important ones are briefly discussed. First of all, as described in the review by Brown and Treviño (2006), leaders
reasoning at higher moral levels are assumed to behave more ethically and thereby demonstrate more ethical leadership. Second, in their research Jordan et al. (2013) found a relationship between moral reasoning and follower perceptions of EL. This would suggest that leaders scoring high on LMII would be perceived as being ethical leaders. Third, following the logic as proposed by Mayer et al. (2012), one could easily argue that the aggregate of a true moral person scoring high on LMII will result in perceived EL and subsequent follower outcomes. Especially regarding the fact that a leader high on LMII demonstrates behavior that avoids being immoral, pays attention to and corrects unethical behaviors (e.g. of followers) and not only measures success by the results, but specifically pays attention to the ways in which the results are obtained.
Taken together, LMII is expected to possibly occupy a role other than moderator in its relation to EL. Due to other mechanisms (e.g. leader-‐follower value congruence, other behavioral leader components) that could influence the relationship between LMII and follower ethical behavior, a partial, rather than a full mediation is expected. Thus, based on the foregoing, the following is hypothesized:
Hypothesis 5a ‘Leader moral identity internalization’ is positively related to ethical leadership
Hypothesis 5b Ethical leadership partially mediates the relationship between ‘leader moral
identity internalization’ and follower ethical behavior
4. Methods
4.1 Procedure and sample
This study is part of a research project at the University of Amsterdam on ethical-‐ and charismatic leadership. Six students, under the supervision of two professors, collected the data for this dyadic study. For this dyadic study to be possible, two self-‐administered questionnaires were created and distributed online to managers and their employees. Only completely filled out surveys that were possible to link (using matching codes) to the right counterpart (leader-‐ follower) were included in the analysis.
Questionnaires were made available in Dutch and English, contingent on the preference of the respondents. All scales were adopted from English sources, so whenever a Dutch
translation was not yet available the scales were directly translated to Dutch. The accuracy and correctness of the translation was checked by all students and subsequently by the supervisors.
Respondents received an email containing a cover letter (see Appendix A) explaining the purpose of the study, what was expected from them and additional information regarding ethical and procedural aspects of the questionnaire. The respondents were asked to fill out items related to both charismatic and ethical leadership, follower outcomes and related constructs. At the end of the questionnaire some questions regarding demographics were asked.
Data were requested from 174 leader-‐follower dyads. The participation of all
respondents was on an anonymous, voluntary basis and no incentive whatsoever was provided to the respondents in return for their participation. In total, 113 leader-‐follower dyads were returned (so 226 questionnaires), corresponding with a response rate of 64.9%. The 113 returned leader-‐follower were not necessarily from matching dyads. To match them, matching codes were used. In some cases the same matching codes were used by multiple dyads, this
being the case those surveys had to be removed from the dataset since it was impossible to link them back to their unique counterpart. In other cases, either the leader or the follower only filled out the survey, so those had to be deleted from the dataset as well since they were not part of a complete dyad. After matching the surveys, the dataset consisted of 102 matched leader-‐ follower dyads.
Regarding demographic characteristics, 55 percent of the employee respondents are female, and the average age of the employees was 36 years (SD = 12.38). Employees, coming from the Netherlands, Spain, Greece and Belgium, worked in a wide variety of jobs including retail, banking, government, and health care. Their organizational tenure was 11.2 years (SD = 12.2) and 83.7% had successfully finished a bachelor or masters degree. Of all managers, 33 percent is female, and the managers had an average age of 45 years (SD = 10.91). Their
organizational tenure was 16.4 (SD = 13.45), and 85.9% had successfully finished a bachelor or masters degree.
4.2 Measures.
As mentioned, we translated the scales that had not yet been translated from English to Dutch. All scales relevant for this study are adopted from scales as operationalized in previous studies (a full overview can be found in Appendix B). Furthermore, they were all anchored on a seven-‐point Likert scale ranging from (1) “strongly disagree” to (7) “strongly agree”.
Ethical leadership is measured using the 10-‐item Ethical Leadership Scale (ELS)
developed and validated by Brown et al. (2005). Sample items include “My supervisor listens to what I have to say” and “My supervisor can be trusted” (∝ = 0.91).
Leader-member exchange is measured with an 11-‐item scale adopted from Liden, Wayne
and Stilwell (1992). Respondents indicate the extent to which they agree with a certain statement. Samples items include: “I like my supervisor very much as a person” and “I admire my supervisor’s professional skills” (∝ = 0.88).
Follower ethical behavior is measured adopting 10 items from the unethical behavior
measure by Mayer et al. (2012). Since all items were counter indicative for ethical behavior all items were reversed. Employees rated the extent to which they agree with the statements, sample items include: to what extent would you “use company service for personal use” and “claim credit for someone else’s work” (∝ = 0.77).
Leader moral identity internalization is measured using the 5 items of the moral identity
internalization scale as developed by Aquino and Reed (2002). Since two of the items were counter indicative those items were reversed. Supervisors were asked to what degree they agreed with the proposed statements linking them to certain characteristics. Sample statements include “It would make me feel good to be a person who has these characteristics” and “I
strongly desire to have these characteristics” (∝ = 0.83).
Control variables. To control for other effects than the ones being researched, leader and
employee organizational tenure were controlled for. One could expect a person’s organizational tenure to influence his/her behavior. The longer someone works for an organization, the bigger this impact could be. This reasoning is in line with Mayer and colleagues (2009), who also controlled for tenure as it could influence one’s behavior.
5. Results
5.1 Preliminary analyses and data screening
5.1.1 Missing values and outliers. Before starting the actual analyses the data was
checked for missing values and potential outliers. Any missing values were replaced using hot deck imputation in SPSS. In the case of surveys that were missing multiple values (sometimes up to 15), the surveys were deleted from the dataset. A total dataset of N=99 remained.
To check for potential outliers, Z-‐scores of the variables ethical leadership, leader-‐ member exchange, leader moral identity internalization and follower ethical behavior were computed. With a probability of less than 0.1 percent, scores outside the range of -‐3.29 to 3.29
are considered significant outliers (Field, 2013). Both ethical leadership and leader-‐member exchange reported outliers on the lower end, two in total. To discover the impact the outliers had on the normality of the distribution, the Kolmogorov-‐Smirnov and the Shapiro-‐Wilk tests were compared for both the dataset with and without the outliers. The output of the
Kolmogorov-‐Smirnov test did not show any differences between the dataset with and without outliers. On the contrary, the Shapiro-‐Wilk test showed that with the outliers the distribution of ethical leadership and leader-‐member exchange was not normal and without the outliers it did show a normal distribution for both variables. However, one cannot drop outliers from a dataset just because they are outliers. The outliers are not an obvious result of incorrectly entered data, the outliers do not cause significant associations that otherwise would not be there, there is no reason to think that the outliers come from another than the intended population and the outliers also did not affect the assumptions of regression analysis. Therefore, the outliers were not removed from the dataset and the final sample that is used for analysis, is N = 99.
5.1.2 Reliability tests. To ensure that the measures consistently reflect the construct it
is measuring, the Cronbach’s Alpha was computed for all constructs. A generally accepted threshold of ∝ = .7 is used (Field, 2013). Ethical leadership reported a reliability of ∝ = .91, leader-‐member exchange ∝ = .88, follower ethical behavior ∝ = .77 and leader moral identity internalization ∝ = .83. In other words, all variables reported reliability above the threshold and can therefore be considered reliable (see Table 2). Furthermore, no improvements could be made that would justify the loss of data by deleting a certain item. Therefore all items were retained in the dataset and used for analysis.
5.1.3 Normality tests. Subsequently normality tests were conducted for all constructs.
Table 1 outlines the values and outcomes related to the skewness and kurtosis. These reflect the lack of symmetry and how ‘pointy’ a distribution is, respectively. The Z-‐scores mentioned in the table can be compared to the values that you would expect to get for skewness and kurtosis for a normal distribution. These scores are calculated by dividing skewness and kurtosis by their respective standard errors. Ethical leadership, leader-‐member exchange and leader moral