Leading by Example: Testing a Moderated Mediation Model of Ethical Leadership, Value Congruence, and Followers' Openness to Ethical Influence

In this study, we investigate a key question of ethical leadership theory: Are ethical leaders able to gain followers' openness to their ethical influence? When describing ethical leadership, we distinguish between leaders&#8217; moral person behavior (i.e., behaviors that are normatively appropriate) and moral management (i.e., using position power to reinforce ethical standards). We then draw on mechanisms of social learning and social exchange and test a moderated mediation model. It is predicted that the effect of leaders' moral person behavior on followers&#8217; openness to ethical influence is mediated by leader&#8208;follower ethical value congruence. Moreover, we examine whether moral management amplifies the effects of moral person behavior. For hypothesis testing, existing measures from the literature are integrated to operationalize four underlying normative reference points of leaders' moral person behavior: humane, justice, responsibility/sustainability, and moderation orientation. After establishing the psychometric properties of this integrative measure, the main hypotheses are tested in two separate studies. While results showed a positive effect of leaders' moral person behavior on followers' openness to ethical influence through ethical value congruence, the moderating effect of moral management was, contrary to our expectation, negative. With this, our study contributes to a better understanding of the mechanisms of ethical leadership.


| INTRODUC TI ON
Failures in ethical leadership have consistently been identified as an important account of corporate scandals and misbehavior in organizations (Knights & O'Leary, 2005). Accordingly, ethical leadership is no longer seen as a naive claim of some idealists or a "Sunday school" subject reserved for theologians and philosophers, but has gained widespread public interest and also become a major topic in organizational behavior research (Den Hartog, 2015). One of the most influential conceptualizations of ethical leadership has defined it as "the demonstration of normatively appropriate conduct (…) and the promotion of such conduct to followers" (Brown, Trevino, & Harrison, 2005, p. 120). An important tenet of this approach is that ethical leadership represents a composite of being a moral person and being a moral manager 1 (Trevino, Laura, & Brown, 2000).
While being a moral person refers to visible traits and behaviors of leaders that are normatively appropriate, being a moral manager describes leaders' proactive efforts to urge followers into an ethical direction. By drawing on social learning theory (Bandura, 1986), it is thought that ethical leaders elicit ethical values, attitudes, and behaviors among followers by modeling ethical behavior (i.e., being a moral person) and proactively pushing ethical standards through Despite its prominence in the literature, there are still several open questions about how and when ethical leadership works and how and when it might not work or work less well. First, the question what we mean when we say "ethical leadership works" offers room for further scrutiny. Since leadership is by definition about influence (Northouse, 2013), the sine qua non of ethical leadership is ethical influence. Thus, the best test for ethical leadership is: To what degree is the leader able to gain followers' voluntary openness to his/ her ethical influence? To what degree is he/she seen as an ethical example and to what degree do followers voluntary seek and accept his/her leadership when it comes to ethical issues? These questions, however, have received surprisingly little attention in the pertinent literature. Furthermore, the bulk of empirical research in the field of ethical leadership has, so far, merged the moral person and the moral manager dimension into a composite score. While this reflects the common notion of ethical leadership as a one-dimensional construct, there is emerging evidence that the two facets represent related, yet distinct concepts and that the interplay between them may be more complex. For instance, Rowold, Borgmann, and Heinitz (2009) as well as Páez and Salgado (2016) used Brown et al.'s (2005) seminal ethical leadership scale and found that the moral person and the moral management items loaded on distinct factors, showing differential relationships with related constructs (e.g., transformational leadership) and outcome variables (e.g., job satisfaction, OCB, workplace deviance, and perceived leader performance). Against this background, the question arises as to how these two dimensions interact when it comes to gaining followers' voluntary openness to leaders' ethical influence. Related to this, further scrutiny is also warranted regarding the range of behaviors that we consider as normatively appropriate and through which ethical leaders, as moral persons, may lead by moral example. Brown et al. (2005) have operationalized this part exclusively around trustworthiness and concerns of care and fairness. While other researchers have refined and extended this view by including additional dimensions, such as power sharing and concern for sustainability (Kalshoven, Den Hartog, & De Hoogh, 2011), there is still a dearth of empirical research about the effectiveness of such an integrated perspective on being a moral exemplar in gaining followers' openness to ethical influence.
With these gaps and questions in mind, our main goal in the present research was to empirically examine the extent to which ethical leaders promote followers' openness to their ethical influence. To specify the content of leader behaviors that are normatively appropriate (i.e., moral person behaviors), we use the framework of Eisenbeiss (2012) which introduced four essential normative reference points of ethical leader behaviors (i.e., humane, fairness, moderation, and sustainability and responsibility orientation). On this basis, our theoretical model posits that leaders who show such normatively appropriate behaviors represent significant moral exemplars who are able to transmit ethical values to followers through role modeling and social exchange. This is expected to foster ethical value congruence between leaders and followers, which in turn is seen as an essential enabler of followers' openness to the ethical influence of their leaders.
A second goal of our study was to shed light on the interplay of moral person behaviors and moral management. Specifically, by relating the two dimensions to different mechanisms of social learning, we propose that the indirect effect of leaders' moral person behaviors on followers via ethical value congruence is amplified by the degree to which leaders show moral management.
Taken together, our research has a number of intended contributions. First, we respond to the recent call for more thoroughly investigating the effects of ethical leadership "on more theoretically-specific outcomes" (Lemoine, Hartnell, & Leroy, 2019, p. 155). By focusing on followers' openness to the influence of ethical leaders we examine a key criterion of effective ethical leadership and thus, contribute to the validity of the ethical leadership construct. Related to this, our work also echoes the call of Brown and Mitchell (2010) emphasizing ethical value congruence of leaders and their followers as a potentially critical explanatory variable for the effectiveness of ethical leaders. Second, by drawing on Eisenbeiss' (2012) integrative work on the normative foundations of ethical leadership, we seek to substantiate a more comprehensive perspective on what it means to be a moral person in terms of ethically desirable leadership behaviors. Third, we contribute to the ethical leadership literature by investigating moral management as a potential multiplier of leaders' moral person behaviors. A clearer understanding of this interplay is not only important for the practical concerns of selecting for and developing ethical leadership; such information will also provide a deeper theoretical perspective on how ethical leadership works.

| THEORY AND HYP OTHE S IS DE VELOPMENT
In what follows, we first specify the normative underpinning of ethical leadership and then provide a theoretical rationale for the study hypotheses. Brown and colleagues (2005) have operazionalized ethically desirable behaviors of leaders exclusively around the notion of being trustworthy, fair, and considerate. Other researchers have refined and extended this view by including power sharing, altruism, and concern for sustainability as additional dimensions of ethical leadership (Kalshoven, Hartog, & Hoogh, 2011). A highly useful contribution to integrating and clarifying these perspectives has been provided by Eisenbeiss (2012).

| The normative foundations of ethical leadership
In an attempt to determine the normative underpinning of ethical leadership, she systematically analyzed predominant Western and Eastern moral philosophies as well as ethics principles of the world religions, resulting in four basic "normative reference points" (Eisenbeiss, 2012, p. 792) of ethical leadership: (a) humane orientation, (b) justice orientation, (c) responsibility and sustainability orientation, and (d) moderation orientation. Humane orientation means that leaders treat followers with dignity and respect and show genuine concern for their well-being and development. Justice orientation refers to fair and consistent decision making, whereas moderation orientation involves behaviors that reflect temperance and humility. Finally, responsibility and sustainability orientation means that leader behavior is guided by social responsibility values and that leaders show concern for the welfare of society and the environment. While Eisenbeiss (2012, p. 729) acknowledges that these four ethical orientations "present established leadership attributes in general leadership literature in the social sciences as well," she criticizes that ethical leadership approaches have mostly "concentrated on humane and justice orientation but have neglected both responsibility and sustainability orientation and moderation orientation." On this basis, we assert that these four dimensions are particularly well suited to describe ethical leaders as moral persons. Accordingly, we use the term moral person behavior to describe normatively appropriate leader behavior, reflecting these four dimensions.

| Leaders' moral person behavior and followers' openness to leaders' ethical influence
Leadership is inherently about influence (Northouse, 2013) Brown et al. (2005) as well as more recent research on ethically positive forms of leadership (Lemoine et al., 2019), we primarily draw on social learning (Bandura, 1986) and social exchange theory (Blau, 1964) to explain how ethical leaders, by engaging in moral person behavior, may stimulate followers' openness to their ethical influence.
Social learning theory posits that individuals learn appropriate behaviors observationally through a role-modeling process (Bandura, 1986). It is noteworthy that observing and learning from those who practice moral principles have consistently been confirmed as important developmental pathways in theories of moral development (Grusec, Chaparro, Johnston, & Sherman, 2014). One of the central tenets of social learning theory is that individuals are more likely to pay attention to and learn from attractive and credible role models. Since leadership entails power and status as well as high degrees of visibility, leaders are likely deemed attractive role models by followers. According to Brown and Trevino (2006), this is particularly true for leaders who engage in normatively desirable behaviors.
In addition, we refer to social exchange theory (Blau, 1964) to bolster the proposed relationship between leaders' moral person behavior and followers' openness to their ethical influence. In line with the observation of Lemoine et al. (2019) as well meta-analytical evidence provided by Hoch, Bommer, Dulebohn, and Wu (2018), leaders are likely to establish high-quality relationships with their followers when they are fair, humble, and considerate. Such relationships include a high degree of trust, respect, and mutual obligation and therefore, followers should be more inclined to go along with the leadership and to willingly seek and accept the ethical influence of the leader. This is in line with empirical research in the field of leader categorization, indicating that the degree to which followers show openness to leadership is strongly determined by their implicit representations of an "ideal" leader (Van Quaquebeke, Graf, & Eckloff, 2014). The more leaders match their followers' conception of an ideal leader, the more favorably followers respond toward leaders, most notably in terms of openness to influence. Through engaging in normatively desirable leadership behaviors and developing high-quality exchanges with followers, leaders are likely to match such implicit representations of ideal leadership. With this, it is plausible that followers establish greater faith in their leader's decision making and choices their leader makes when it comes to ethical issues. Hence, the following prediction is made:

Hypothesis 1 Leaders' moral person behavior is positively related to
followers' openness to their leaders' ethical influence.

| The mediating role of ethical value congruence
Although the relationship between leaders' moral person behavior and followers' openness to ethical influence appears to be quite straightforward, we propose that this link is not appropriately understood solely as a direct effect. To our thinking, a more proximal response on the part of followers intervenes between the behaviors they observe and their choice to willingly seek and accept the ethical influence of the leader.
We focus on perceived ethical value congruence as the key mediating process because there is an inherent link between ethical values and ethical leadership and because prior research has consistently emphasized its vital, yet underresearched role for the effects of ethical leadership (Brown & Mitchell, 2010). Following Edwards and Cable (2009), we define ethical value congruence as the similarity between ethical values held by leaders and followers. More specifically, we focus on followers' subjective perceptions of congruence since prior research has consistently shown that perceived fit is a better predictor of attitudinal outcomes than objective fit (Kristof- Brown, Zimmerman, & Johnson, 2005).
To explain the intervening effect of perceived value congruence, we again refer to processes of social exchange and social learning.
Specifically, followers' desire to promote and retain a high-quality relationship with an attractive leader is likely to induce a sense of identification with the leader and the inclination to emulate the leader's attitudes and values. Indirect support for this notion comes from studies indicating that charismatic and transformational leaders convey their vision and values to followers through identification processes and role modeling (Conger, 1999), resulting in a higher level of value congruence (Brown & Trevino, 2006;Hoffman, Bynum, Piccolo, & Sutton, 2011). A similar mechanism can be assumed for leaders as moral persons and in fact, there is initial empirical evidence supporting this notion. A study reported by Tang et al. (2015) found that perceptions of leader-follower value congruence accounted for the effectiveness of ethical leaders with regard to the reduction of turnover intentions among followers. It is, however, important to note that their study treated ethical leadership the conventional way, that is, as a composite of the moral person and the moral manager dimension. Also, it looked at general value congruence and not congruence of ethical values.
In general, value congruence naturally results in more and better communication, while reducing the likelihood of misunderstandings and conflict (Edwards & Cable, 2009

| The moderating role of moral management
Thus far, we reasoned that leaders are moral exemplars by engaging in normatively appropriate behavior. Implicit to this is the assumption that leaders are positive role models by simply living ethical values and not by explicitly describing desired values and instructing employees about them. In fact, role modeling does not necessarily relate to purposeful influence, including the use of explicit guidance or learning techniques. This is in line with the notion of role modeling as an "incidental form of social learning" (Warhurst, 2011, p. 876), whereby leaders may not even be aware that they are being role modeled. That said, in the present research we develop the argument that moral management can fuel the more passive role that leaders play as moral persons. Moral management refers to intentional efforts of leaders to make ethics a salient theme in the organizational context (Trevino et al., 2000). These efforts are more transactional in that a moral manager uses his or her position to communicate clear ethical standards and to reward and discipline followers, depending on whether they follow or violate ethical expectations. As such, moral management corresponds more closely with the notion of verbal instructional modeling within the social learning framework (Bandura, 1986), implying a rather purposeful relationship that provides explicit guidance, support, and advice. Furthermore, as moral managers, leaders more explicitly require followers to "buy into a social exchange system" (Lemoine et al., 2019, p. 170) by consistently stressing and reinforcing the importance of ethics. That said, we posit moral management as a boundary condition that amplifies the impact of moral person behaviors. Followers look to the leaders and clear messages about ethical standards and expectations represent powerful clues that help followers to align their values. By actively putting ethics on the leadership agenda, leaders create more opportunities for followers to observe, remember, and emulate the value-based behaviors of their leaders. Also, moral management promotes what social learning theory has described as incentive and motivational processes (Bandura, 1986). When incentives are available (i.e., rewards or the avoidance of punishment), followers will more easily pay attention to and retain principled behaviors that they observe and, moreover, observation is more quickly internalized. This is in line with Schein's (1992) notion that leaders embed their beliefs and values by signaling what they pay attention to, what they measure and control and, importantly, how they react to critical incidents (such as moral transgressions of followers). Taken together, the following prediction is made:

Hypothesis 3 Moral management operates as a moderator by amplifying the positive relationship between leaders' moral person behavior and perceived leader-follower ethical value congruence.
Finally, although we posit ethical value congruence as the key mechanism through which leaders' moral person behavior exerts its effects on followers, it is plausible that other similar mechanisms are at work too, such as general reciprocity and personal identification.
Whatever exactly links followers' perceptions of their leader as behaving normatively appropriate to their choice to accept and seek the ethical influence of their leader, it is plausible that its strength is a function of the degree to which leaders are able to make their ethics messages salient among followers. Hence, we propose that moral management influences also the direct relationship between leaders' moral person behavior and followers' openness to their leaders' ethical influence:

Hypothesis 4 Moral management operates as a moderator by amplifying the positive relationship between leaders' moral person behavior and followers' openness to their leaders' ethical influence.
Our theoretical model is summarized in Figure 1.

| ME THODOLOGY AND S TUDY OVERVIE W
The hypotheses under investigation were addressed in two studies.
Study 1 tested the links specified in our theoretical model. Study 2 F I G U R E 1 Proposed theoretical model was designed to replicate the results from Study 1 and, in addition, explored whether the proposed effects would hold when controlling for transformational leadership as a competing leadership style. This is important because it permits to address the issue of omitted variable bias and to assess the incremental validity of ethical leadership when predicting our focal outcome. However, prior to these two studies, we tackle the question of how to adequately measure leaders' moral person behavior. Below, we first describe this measurement strategy and how it was developed and then report the two main studies for hypothesis testing.

| Assessing leaders' moral person behaviors
To measure the four normative reference points of moral person behaviors (i.e., humane, fairness, moderation, and sustainability and responsibility orientation), we adopted and integrated existing items from the pertinent literature. Specifically, we selected 16 items from the following instruments 2 : The Ethical Leadership Scale , the Ethical Leadership at Work Questionnaire (Kalshoven et al., 2011), the Authentic Leadership Inventory (Neider & Schriesheim, 2011), the Servant Leadership Instrument (Ehrhart, 2004), the Servant Leadership Survey (Van Dierendonck & Nuijten, 2011), and the Multifactor Leadership Questionnaire (Avolio, Bass, & Jung, 1999). Then, this newly integrated measure was pretested in two separate samples (hereafter: sample 1 and sample 2).

Sample 1 included 121 employees from various organizations in
Germany (average tenure with the leader of 3.5 years). They were instructed to indicate how frequently each item fits their leader, using a 5-point continuum (1 = almost never-5 = almost always).To test the factorial validity of the measure, the fit of three different models was compared via confirmatory factor analysis (CFA), using the lavaan package in R (Rosseel, 2012). For model comparison, we followed the recommendations of Meade, Johnson, and Braddy (2008), according to which a change of .002 in CFI (Comparative Fit Index) indicates that the models are significantly different. The first model adopted a second-order factor structure in which items loaded onto their respective factors (i.e., humane orientation, fairness orientation, moderation orientation, responsibility/sustainability orientation) and the four factors loaded on a second-order latent ethical leadership factor. The second model was a first-order factor model in which items loaded onto their respective factors and the four factors were allowed to correlate. In the last model, all items were allowed to load on one factor. The fit statistics for the three models are reported in the upper section of Table 1. The results show that the single-factor model was inferior to both the second-order factor model and the first-order factor model. The difference between the first-order and the second-order factor models was not statistically significant.
Since we strove for an economical measure, we selected three items per subscale, based on the revealed factor loadings. This resulted in a total of 12 items for the final measure of leaders' moral person behavior. These 12 items were administered to a separate sample (sample 2), consisting of 325 employees from an automotive company in Germany (average tenure with the leader of 9.3 years). The CFA results are reported in the lower section of Table 1 and confirm the pattern found in sample 1. Again, while both the second-order factor model and the first-order factor model were preferable over the single-factor model, the fit of these two models was statistically equivalent. Thus, we conclude that moral person behavior is not a higher level construct that underlies its dimensions because the dimensions of moral person behavior are not different manifestations of the construct. Rather, it is better described by what Law, Wong, and Mobley (1998, p. 745) refer to as the "aggregate model," implying that "the multidimensional construct is formed as an algebraic composite of its dimensions." In other words, leaders' moral person behavior consists of the sum of its dimensions and none of these dimensions alone adequately captures the complexity of global moral person behavior. The 12-item measure derived from the above approach is shown in the Appendix and was used to test the hypotheses under investigation in the two studies that are reported next. Model

| Sample and procedure
Data for Study 1 came from 206 employees recruited from various organizations in Germany. Snowball sampling starting from the networks of two graduate students involved in data collection was used. We followed the recommendations by Demerouti and Rispens (2014) to ensure data quality in student-recruited samples.
Participants were asked to rate their immediate leaders in terms of moral person behaviors and moral management. In addition, they provided self-ratings on the proposed outcome variables (i.e., perceived ethical value congruence and openness to their leaders' ethical influence), sociodemographics, and mood state (i.e., positive and negative affectivity) as a potential source of method variance (Podsakoff, MacKenzie, Jeong-Yeon, & Podsakoff, 2003).
The mean age of the participants was 29.27 (SD = 9.05); 52% were male and 54% had a university degree. The majority of the participants worked in the for-profit sector (74%

| Measures
To measure leaders' moral person behavior, the integrative measure with 12 items described above was used.  Thompson (2007). Respondents were asked to indicate how they have felt during the past 4 weeks on a 5-point rating scale (1 = never to 5 = always).

| Measurement issues
In a first step, we examined the factor structure of the 12-item measure of leaders' moral person behavior in the current data set For all other constructs, two parcels were built using the factorial algorithm (Rogers & Schmitt, 2004)   Controlling for the mediator, the direct effect of moral person behavior on followers' openness to ethical influence was .55 (p < .001).

| Hypothesis tests
These patterns provided full support for Hypotheses 1 and 2. Next, we examined the proposed moderation of the direct and indirect effect through moral management (Hypotheses 3 and 4). The results of the moderated mediation analysis are reported in Table 3.   Table 4). Regarding the overall model, in line with the negative interac-  Abbreviations: CI, confidence interval; LL, lower limit; UL, upper limit.

| S TUDY 2
Study 2 was designed to see whether the patterns found in the previous study can be replicated in a separate sample. Furthermore, we sought to test whether the effects of leaders' moral person behavior would hold when controlling for the effects of transformational leadership as a competing correlated leadership style. With its strong focus on charisma as well as visionary and inspirational communication, transformational leadership is a highly prominent and established approach in the field of positive leadership (Bass, 1999;Hoch et al., 2018). Yet, although transformational leaders are typically portrayed as ethical role models, it has consistently been noted that the transformational leadership model lacks an explicit moral dimension (Hoch et al., 2018). Therefore, testing ethical leadership against transformational leadership is particularly suited to demonstrate its unique contribution.

| Sample and procedure
Data for this study came from employees working in various industries in Germany. We gathered 125 responses via a professional provider of international access panels. In addition, 36 responses came from the network of a graduate student, who was involved in data collection, resulting in a total of 161 valid responses for data analysis. Importantly, there were no statistically significant differences regarding the study variables between the two sample sources.
Instead of including an affect measure, responses were collected at two points in time, separated by approximately two weeks, in order to reduce common method bias (Podsakoff et al., 2003).
The mean age of the participants was 35.59 (SD = 13.07); 49% were male and 28% had a university degree. The majority of the participants worked in the for-profit sector (70%). With regard to the occupational background, 14% worked in the public/government sector, 9% each came from commerce and consumer affairs, education, and building and construction industry, and 8% worked in financial services (51% other sectors). The average tenure with the current leader was 5.55 years (SD = 6.19).

| Measures
The same measures and response formats as in Study 1 were used. In addition, transformational leadership was measured using parts of the instrument developed by Rafferty and Griffin (2004). Specifically, we used three subscales, consisting of three items each, to capture the features of transformational leadership, that is, vision, intellectual stimulation, and inspirational communication. Following Conway and Lance (2010), the two remaining subscales, that is, supportive leadership and individualized consideration, were not included because they show substantially high item overlap with the humane orientation subscale of the moral person behavior measure, which can bias relationships.
At Time 1, respondents were asked to rate their immediate leader in terms of transformational leadership and moral person behavior, and to provide personal information. At Time 2, respondents provided ratings of moral management and self-reports on the focal outcome variables (i.e., perceived ethical value congruence and openness to ethical influence).

| Measurement issues
As in Study 1, we first tested the factorial validity of the 12-item moral person behavior measure via CFA. Again, a single-factor model

F I G U R E 3 Conditional indirect effect (Study 1)
| 323
Next, we conducted a series of CFAs to examine the integrity of the overall measurement model and to see whether the study variables captured distinct constructs versus common source effects.
Five constructs were included: moral person behavior, transformational leadership, moral management, ethical value congruence, and openness to ethical influence. Again, we formed four item parcels for moral person behavior (i.e., justice, moderation, humane, and responsibility/sustainability orientation) and, in addition, three item parcels for transformational leadership (i.e., vision, intellectual stimulation, and inspirational communication). 5 For all other variables, items were used as indicators. The results of this procedure are shown in Table 5, providing solid support for discriminant validity. 6

| Hypothesis testing
The descriptive statistics and correlations among the study variables are presented in Table 6. For hypothesis testing, the same procedures as in Study 1 were used and we again controlled for tenure with the leader in all analyses. Also, variables were not centered and all coefficients reported here are unstandardized.

TA B L E 5 Measurement models (Study 2)
Model Note: N = 161, Δχ 2 and ΔCFI represent the difference in χ 2 and CFI values between the respective model and Model 1 (i.e., the proposed 5-factor model).
*p < .05; ***p < .001. As shown in Table 7 have resulted in less precise estimates of regression coefficients (Cohen, Cohen, West, & Aiken, 2003). Therefore, we followed the procedures suggested by Lorenzo-Seva, Ferrando, and Chico (2010) and conducted relative importance analysis to obtain a more precise understanding of the specific role of each predictor. Results ( Abbreviations: CI, confidence interval; LL, lower limit; UL, upper limit. † p < .10; ***p < .001. and openness to ethical influence, leaders' moral person behavior drove the prediction to a higher extent.

TA B L E 6 Descriptive statistics and correlations (Study 2)
Next, the proposed conditional effects through moral management (Hypotheses 3 and 4) were tested. The results of the moderated mediation analysis are reported in Table 9 Table 10). Also, the index of moderated mediation (Hayes, 2015) was again negative (

| Brief discussion
This study supports the pattern of results found in the previous study, while controlling for the effects of transformational leadership. It is noteworthy that we again found moral management to exert a negative moderating effect on the relationship between leaders' moral person behavior and ethical value congruence.

| G ENER AL D ISCUSS I ON
The purpose of this research was to examine whether and through which mechanisms leaders gain followers' openness to their ethical influence. In two separate studies we found support for our predictions: Leaders' moral person behavior (i.e., having and exhibiting a humane, fairness, moderation as well as responsibility and sustainability orientation) was positively related to followers' perceived ethical value congruence, which in turn predicted followers' openness to ethical influence. In contrast to what we expected, moral management had a negative moderating effect on the relationship between moral person behavior and ethical value congruence.

| Theoretical implications
The general picture that emerges from this research suggests that the conceptual basis of ethical leadership may be more complex than it is often assumed in the pertinent literature. As a rather straightforward implication, our results indicate that it is worthwhile to extend the behavioral foundations of the moral person dimension of ethical leadership by incorporating a broader set of normatively appropriate behaviors. We thereby add to initial empirical evidence that moral person behavior goes beyond principles of care and justice and include leadership behaviors referring to moderation and temperance as well as responsibility and concern for the greater good (Eisenbeiss, Van Knippenberg, & Fahrbach, 2015). With this, we now have a clearer understanding of the behavioral complexity in ethical leadership and what it means to lead by example. The positive relationship between these leader behaviors and followers' increased openness to and acceptance of the leader's ethical influence suggests that the four normative foundations that we took from Eisenbeiss (2012) are indeed valid proxies for the moral goodness of a leader (Chiu & Hackett, 2017). Furthermore, we find that ethical value congruence mediates the relationship between leaders' moral person behaviors and followers' openness to the ethical influence of leaders. This fills a gap in the pertinent literature because it verifies a link that has been implicated in many previous studies as an important explanatory factor of leadership in general and moral forms of leadership in particular (Brown & Mitchell, 2010). The examination of ethical value congruence represents a particularly meaningful asset of our study, since prior research in this field has usually examined general value congruence which, however, does not necessarily entail congruent ethical values (Tang et al., 2015). An important empirical contribution that is related to this is that the effects of moral per- Abbreviations: CI, confidence interval; LL, lower limit; UL, upper limit. the notion that highly correlated leadership constructs can still have distinct effects and in particular, they confirm that ethics-centric leader behaviors are uniquely important when compared with classical, more task-and goal-focused leader behaviors (Lemoine et al., 2019).
An important theoretical implication of the present effort refers to the relationship between moral person behaviors and moral management. Our results substantiate the notion that these two features of ethical leadership represent analytically and empirically distinct types of leader behavior (Páez & Salgado, 2016;Rowold et al., 2009). This is not only reflected in the CFA results; they both also relate differently with our focal outcomes. Specifically, when looking at their effects in our data in isolation, they both predict ethical These moral evaluations had a stronger impact than legitimacy, that is, the acceptance of authority, and, importantly, they were also more effective than expected rewards or punishments on rule-related be- to be perceived as hypocrites (Trevino et al., 2000), or at least that such leaders are not regarded as attractive role models. At the other hand, however, they challenge the assumption that leaders who engage in moral person behaviors without actively putting ethics on the leadership agenda may be perceived as ethically neutral (Trevino et al., 2000). While Trevino et al. (2000)

| Practical implications
The first practical implication is rather simple: Our results suggest that leaders' moral person behavior is instrumental in promoting ethical value congruence, which may then foster leadership effectiveness. We know from prior research that value congruence may foster actual ethical performance, such as reduced organizational deviance (Brown & Trevino, 2006 Another issue refers to the measurement of moral person behaviors, specifically the justice component. While our measure focuses on perceived adherence to justice rules (e.g., the leader analyzes relevant data before making a decision), it also captures global fairness perceptions (i.e., the leader is perceived as generally making fair and balanced decisions). Although perceptions of justice rule adherence are likely to trigger a more general sense of appropriateness (Ambrose & Schminke, 2009), we recognize that this may, to some extent, vary across individuals and contexts. Thus, future studies in this field should measure a broader range of justice rules leaders may adhere to, most notably distributive and procedural justice aspects.
Following the research by Samara and Paul (2019), such studies would also permit to investigate whether and to what degree different contexts (e.g., family businesses or employee-owned companies) produce distinct notions of what is normatively appropriate leader behavior.
Furthermore, also with regard to measurement precision, it would be useful in future research to consider alternative forms of ethical value congruence. Future research could measure more objective forms of ethical value congruence (Kristof-Brown et al., 2005) to determine whether there is more definitive support for the role of ethical value congruence within this context. Also, researchers might use inventories of specific ethical values, such as the Schwartz Value Survey (Schwartz & Boehnke, 2004), to examine whether specific patterns of congruence exert distinct effects on followers' openness to and acceptance of the ethical influence of their leaders.
A final limitation of or research refers to the cultural context.
Ethics scholars have stressed that "real-life perceptions of ethicality are pluralistic and culturally determined" (Chiu & Hackett, 2017, p. 32). In a similar vein, as indicated above, it is possible that also organizations differ in their endorsement of specific ethical orientations (Fehr et al., 2015;Samara & Paul, 2019). Thus, future research should test our predictions not only in more culturally diverse settings but also take the organizational culture into account.  1 We wish to note that, consistent with much prior research in this field, we use the terms ethical and moral synonymously.

| CON CLUS ION
2 Some of these instruments were included in a previous data collection which the authors of this research conducted in a separate organization. These data are not part of the present research. Three experts (i.e., the first author of this study as well as a doctoral student and a graduate assistant in the field of leadership research) assigned the pertaining items to the four ethical orientations of Eisenbeiss (2012). Then, items were selected based on statistical (i.e., factor loadings) and theoretical (i.e., semantics, conceptual fit) grounds. Because this procedure did not result in an item pool that adequately covered all four ethical orientations, the three experts identified and included theoretically suitable items from additional instruments. 3 The same pattern of discriminance was obtained when items are used as indicators for all latent variables. The model fit of the proposed four-factor model (χ 2 = 561.14, df = 269, p < .001, χ 2 /df = 2.08, CFI = 0.92, RMSEA = .07) was preferable over all competing models. 4 It is important to note that in this equation, the direct effect of moral person behavior on both ethical value congruence and openness for ethical influence is conditional (i.e., it restricts the values of the moderator to zero). The same applies to the direct effects of the moderator variable on the outcomes (i.e., it restricts the values of independent variable to zero, see Hayes, 2018). To facilitate interpretation, we estimated the direct effects when the moderator was excluded. This yielded the following results: Ethical value congruence was more strongly predicted by moral person behaviors (b = .90, p < .001) relative to moral management (b = .28, p < .001). Also openness to ethical influence was more strongly predicted by moral person behaviors (b = .86, p < .001) relative to moral management (b = .36, p < .001). 5 A separate CFA for transformational leadership, consisting of three correlated factors yielded an excellent model fit (χ 2 = 30.13, df = 24, p =.18, χ 2 /df = 1.25, CFI = .99, RMSEA = .04). 6 The same pattern of discriminance was obtained when items are used as indicators for all latent variables. The model fit of the proposed 5-factor model (χ 2 = 912.16, df = 517, p < .001, χ 2 /df = 1.76, CFI = .88, RMSEA = .07) was preferable over all competing models. 7 To facilitate interpretation, we again estimated the direct effects when the moderator was excluded. This yielded the following results: Ethical value congruence was more strongly predicted by moral person behaviors (b = .75, p < .001) relative to moral management (b = .25, p < .01). Moreover, openness to ethical influence was more strongly predicted by moral person behaviors (b = .66, p < .001) as compared to moral management (b = .31, p < .001).