Introduction

Ever since the huge corporate scandals that impacted the world and lead to the global crisis, researchers have been reviewing concepts and theories to attempt to identify the shortcomings in the ethical behaviour of managers and top executives in Multinational Corporations (MNCs) (Brown & Treviño, 2006). There was a time when CEOs and senior managers were admired and respected and were considered to be models of success. However, nowadays even the legitimacy of leadership is questioned and the public’s trust in corporate governance has fallen dramatically. Not surprisingly, empirical leadership research has swiftly changed its focus to analyse the morality, integrity, behaviour, and authenticity of leaders (Ko et al., 2017).

The magnitude of this problem cannot be understated. Several studies have recently demonstrated concern regarding this topic, having analysed the following: 260 corporate scandals that took place from 1996 to 2013 in US-listed companies which originated from countries’ weak legal institutions (Chen, 2016); 76 Malaysian firms which committed financial reporting fraud in 1996–2016, demonstrating a direct association between fraud and the regulator’s auditor (Ghafoor et al., 2019); the domino effect of the corporate scandals of 2008 that led to the financial crisis, which in turn exposed the failure of the functioning of the financial system regulator. Wedell-Wedellsborg (2019, p. 5) stated that “the human mind is skilled at justifying minor incursions when there is a tangible reward at stake—and when the risk of getting caught is low”. Data supports this sentence, with the ACFE report for 2020 stating that “41% of corruption perpetrators were employees and 35% were managers”.

There is consensus among scholars that unscrupulous and dishonest actions disrupt interpersonal relationships, distress individuals, constitute a bad reputation for organizations, and inflict suffering on society (Barsky, 2011; Brenkert, 2009; Gino et al, 2011; Welsh & Ordóñez, 2014). In parallel, scholars also agree that corporate leaders are responsible for flying the flag of ethical environment and behaviour and establish themselves as role models for their followers (Alshammari et al., 2015; Bavik et al., 2018; Sarwar et al., 2020). Nevertheless, there is little research on how effective ethical leadership is in disabling unethical behaviour, by acting on the psychological processes of morality (Hsieh et al., 2020; Seriki et al., 2020; Shaw et al., 2020).

Moral conflict arises when an individual’s attitude or personal interest clashes with the group’s responsibility and harms the organization’s values (Goodarzi et al., 2018). Accordingly, the objective of our research is to contribute to the literature by carrying out an empirical analysis that expands the horizons of ethical leadership and its inestimable contribution to MNCs in promoting and influencing organizational ethical environment. This study explores the whole mediation process, ranging from ethical leadership through to unethical behaviour, as well as the possible interaction that the level of corruption in distinct countries plays in each path of the mediation process.

The remainder of this study is organized as follows. First, we conduct a review of the relevant literature to explain and support each one of the variables and their relationships, namely: ethical leadership, instrumental ethical climate, displacement of responsibility, unethical behaviour, and a corrupt environment. Second, a set of hypotheses and the conceptual model are presented, based on the extant literature on the topics in question. Third, the data collection and analysis procedures are presented and fourth, the results obtained are described and explained. Finally, the theoretical and practical implications are discussed, as well as suggestions for future research.

Literature Review and Hypotheses

Social cognition models are designed to both predict behaviour (O’Connor & Armitage, 2003) and establish roots in theories with decades of development. For instance: the Theory of Reasoned Action was born in 1975, which was then upgraded by the Theory of Planned Behaviour in 1991 (Ajzen & Kruglanski, 2019); the Protection Motivation Theory from 1975 (Good & Hyman, 2020); and the Attribution Theory from 1958 (Martinko & Mackey, 2019). The main concern of these theories is to explain causal nexus of social interactions, and to demonstrate how it influences individuals to form a causal judgement regarding a particular event.

The subject of social cognition remains highly relevant, and all these classic theories are frequently re-visited, having generated important behavioural theories, such as: the theory of planned behaviour (Rahaman et al., 2019; Roos & Hahn, 2019); the protection-motivation theory (Good & Hyman, 2020); and the leadership behaviour and self-leadership theory (Behrendt et al., 2017; Stewart et al., 2019). All these authors differ in their research approach, yet all similarly conclude that different mechanisms are used to justify misbehaviour that damages organizations and the subsequent search for an explanation and subsequent correction sustained on these theories has become a significant tool for social science.

Departing from the research gap that motivated this study, we reviewed the literature on the main constructs that are aligned with the topic under research, approaching it from the Social Cognitive Theory (Bandura, 1986) point of view. This theory is widely tested over time and is suitable for application in different fields, such as education, health, organizations, and business (Schunk & DiBenedetto, 2020), and is also well suited as a supportive theory for our hypotheses.

Unethical Behaviour

Ethical behaviour is defined as being a way of acting that is morally accepted as being “good” and “right” in each situation, as opposed to being “bad” or “wrong” (Sims, 1992, p. 506). It is both legally and morally acceptable to the larger community, according to how individuals handle ethical dilemmas in uncertain situations (Treviño, 1986). The “ethic way” presupposes thinking and acting under commonly accepted laws, norms, and rules in a given society, be they written or unwritten. Conversely, unethical behaviour refers to “decisions and actions either illegally or morally unacceptable to a larger community” (Jones, 1991, p. 367). This well-accepted definition by many scholars is complemented by Gino et al., (2011, p. 191) as being “the acts that have harmful effect upon others”. For example, unethical action can be carried out through the use of a deceptive message (i.e. lying) in communication (Barsky, 2011), or can be excused as being legitimate in the drive for higher performance (Welsh & Ordóñez, 2014). In sum, socially condemned attitudes such as lying, stealing, bullying, and harming the powerless or the fragile, are all examples of bad behaviour which are commonly found in organizational environments (Barsky, 2008; Knoll et al., 2015).

When behaving unethically, individuals tend to experience discomfort from becoming misaligned with their community through cognitive or affective disagreement (Mesdaghinia et al., 2019). While experiencing such discomfort after carrying out unethical actions, individuals tend to try and reduce their dissatisfaction by changing attitudes and behaviour within their organization. However, when the self-regulatory mechanism fails to cope with such disagreement, individuals often disengage morally (Stephens, 2017).

Displacement of Responsibility

Morality guides individuals to act correctly and, in effect, acts as a mechanism that regulates conduct, beliefs, attitudes, and behaviour (O’Keefe et al., 2019). The lack of this mechanism, or the impossibility to act according to “the right thing to do”, creates internal conflicts and moral distress that blurs the perception of one’s moral behaviour (Tigard, 2019). The disengagement from this moral guidance leads to acting immorally, or in extreme situations, to amoral action (Bandura, 1990). Social Cognitive Theory explains that there is a reciprocal interaction between behaviour and cognitive process, where belief in one’s own self-efficacy drives individuals’ ability to control themselves (Bandura, 2002a) and conversely, self-regulatory capabilities restrain individuals from carrying out transgressive behaviour (Moore et al., 2012) through the process of self-sanctioning (auto-condemnation, auto-evaluation, auto-reflection).

Moore et al. (2012) posit that when there is a fracture in the link between self-sanctioning and transgressive behaviour, moral disengagement mechanisms are activated which cause a disruption in moral behaviour and provide manoeuvring reasons to justify the unethical actions in question (Barsky, 2008). The moral disengagement theory, which was primarily developed by Bandura in the 1980s as an extension of the social cognitive theory, describes a cognitive process whereby the following ethical principles do not apply to themselves in all contexts: (1) recognition of the moral issue, (2) “thinking on it”, judgement, (3) establishing moral intentions based on one’s behaviour, (4) taking action, (Bandura, 1990). This “Moral Awareness” (Moore, 2008) posits that it is an individual’s choice to misbehave, evoking individuals” “free will” to make a decision.

The Moral Disengagement Theory (Bandura, 1986, 2002b) establishes eight interrelated mechanisms that individuals use to transform immoral or socially reprehensible actions into acceptable or worthy behaviour in a particular setting. Three of these mechanisms are used to transform unethical acts into justified actions, namely: Moral Justification, Euphemistic Labelling, and Advantage Comparison. A further two mechanisms deflect the responsibility of bad behaviour by obscuring or minimizing it, namely: Displacement of Responsibility and Diffusion of Responsibility and the last three mechanisms are based on finding a victim who serves to reduce or even eliminate the distress of the bad behaviour: Distortion of Consequences, Dehumanisation, and Attribution of Blame (Bandura, 2002b; Moore et al., 2012; Zsolnai, 2016).

Our focus is placed on displacement of responsibility, which is defined by Harris and He (2019) as being the causal attribution of one’s actions to social pressures or other factors, rather than just to oneself. This is a suitable mechanism for explaining a leader–follower relationship, because it twists the relationship perception of the team and distorts the consequences of unethical action by blaming others as being responsible (Bonner et al., 2016; Harris & He, 2019). As this mechanism is prevalent at the individual level (Johnson & Buckley, 2015), the lack of self-regulatory processes results in individuals justifying their actions by transferring the responsibility to a higher level of responsivity (leader), or conversely to a lower level, from the leader to the subordinates.

Scholars have been analysing the relationship between unethical behaviour and moral disengagement practices over the years, both in theoretical and empirical studies. For instance, Barsky (2008) theorized about how moral disengagement is related to unethical behaviour engagement and empirically tested this hypothesis later on, finding a strong positive relationship between the displacement of responsibility and unethical behaviour (Barsky, 2011). In turn, Gino et al. (2011) supported the hypothesis that the depletion of self-regulatory resources increases unethical behaviour. This relationship was also supported in the study of Welsh and Ordóñez (2014), which went further by establishing a mediation of the self-regulatory depletion of the relationship between organizational goals and unethical behaviour. A year later, Knoll et al. (2015) found the same relationship, adding that the relationship is sensitive to the context.

Considering the background explained above, it can be deduced that moral disengagement activated through displacement of responsibility has a direct and positive relationship with unethical behaviour, with the tendency to act in non-ethical ways. Consequently, we hypothesize that

Hypothesis 1

Displacement of responsibility is positively associated with unethical behaviour.

As stated above, ethics is not an individual product, but rather a social product. This implies that groups—identified as being teams within an organizational setting—must play a critical role in the activation of moral disengagement mechanisms for individuals. The literature identifies a failure in the settlement of ethical standards among big corporations, especially at the organizational team level (Moore, 2008). A reasonable team effect may be conditioned by its ethical climate.

Instrumental Ethical Climate

Ethical climate refers to the moral atmosphere at work and the level of ethics practised within an organization, which in turn reflects the culture of an organization and is characterized by a conductive thread that makes members of the organization reflect on questions of what is right and what is wrong (Otaye-Ebede et al., 2020; Tanner et al., 2015). An ethical work climate is defined as being “the shared perception of what is an ethically correct behaviour and how ethical issues should be managed” (Victor & Cullen, 1988, p. 101). What constitutes rightful behaviour depends on the organization’s standards, as governed by codes, norms, and policies that influence employees sufficiently for them to be able to perceive those standards and act accordingly in the decision-making processes (Gronlund et al., 2019; Martin & Cullen, 2006).

The development of research on this subject theoretically identified nine types of ethical climate that have been transformed into five common empirical dimensions, namely: instrumental, caring, independence, rules, law and code (Martin & Cullen, 2006; Newman et al., 2017; Victor & Cullen, 1988). By its nature, an instrumental ethical climate (IEC) can be taken to be intrinsically unethical, and is, thus, of utmost interest for this research, as it is focused on unethical behaviour.

The rational egoism theory, which has progressively become more prominent in philosophy and organizational ethical thought (e.g. Harviainen et al., 2020; Peikoff, 1991; Rand, 1964), is based on self-interest and company profit behaviour. It is the foundation of the IEC. The common attitudes consist of individuals’ decisions being made in order to satisfy their interests, with little or no regard for the consequences of their actions (Ambrose et al., 2008; Cullen et al., 2003; Zhang & Yao, 2019). This environment blurs the judgement of employees facing inconsistencies between the IEC and their level of morality, which is the least-preferred ethical climate for organizations (Martin & Cullen, 2006; Tsai & Huang, 2008).

Victor and Cullen (1988) discovered that it is uncommon for organizations to share the different dimensions of ethical climate in situ, whereby, in general, organizations are characterized by a dominant climate (Tsai & Huang, 2008; Zhang & Yao, 2019). According to this background, it can be deduced that an instrumental ethical climate has a direct positive relationship with displacement of responsibility. Consequently, we hypothesize that

Hypothesis 2

Instrumental ethical climate is positively associated with moral disengagement.

By being in alignment with the literature on unethical behaviour, ethical climate, and moral disengagement, Moore (2015) reasoned that as a psychological process variable, moral disengagement is a promising mediator of other relationships. Indeed, moral disengagement mechanisms help to explain unethical attitudes (Newman et al., 2019). For if moral disengagement disconnects the moral self-regulation and self-condemnation process (Bandura, 2002b), then it is expected to spill over to the corporate ethical climate, and consequently cause individuals to practice unethical behaviour (Shaw et al., 2020).

There is plenty of literature that empirically supports the mediation role of moral disengagement in several relationships, such as the job insecurity effect on interpersonal and organizational deviance (Huang et al., 2016), the perception of cyberbullying on cybervictimization (Cuadrado-Gordillo & Fernández-Antelo, 2019), and the effect of authoritarian-benevolent leadership on unethical pro-organizational behaviour (Shaw et al., 2020). Newman et al. (2019) found that 30 or more studies support the mediation role of moral disengagement over the last two decades, all of which report displacement of responsibility as being a mechanism that promotes unethical work behaviour (Barsky, 2011).

In this case, if indeed displacement of responsibility predicts unethical behaviour, and in turn, if instrumental ethical climate predicts moral disengagement mechanisms, then we hypothesize that

Hypothesis 3

Displacement of responsibility mediates the positive association between instrumental ethical climate and unethical behaviour.

Organizational behaviour research has long recognized that teams are an explanatory factor for individual behaviour within an organization. This is evidenced in theories such as social identity, which explains how social interactions and the identification of team members can influence employees’ behaviour (Lasrado & Arora, 2017), albeit within the phenomenon of rule-bending, which explains certain attitudes that bypass obligations established by the organization (Borry, 2017). However, leadership deserves no less attention, due to its critical power position in modulating organizational culture and the influence that it exerts on the behaviour and attitudes of employees (Hiekkataipale & Lamsa, 2019). The leader’s role includes many important tasks, such as improving not only the performance of employees, but also managing workgroup conflicts (Zhao et al., 2018). With regards to ethics, leadership is a critical factor that should not pass unnoticed, as it can be a source of and a target of judgement about right or wrongdoing. A leader who behaves ethically transmits these values and, thus, creates an environment that encourages group ethical behaviour (Mayer et al., 2012). For this reason, ethical leadership has emerged as an important construct in leadership studies (Treviño et al., 2000).

Ethical Leadership

decades ago, Treviño et al. (2000) visualized the term “Ethical Leadership” as being the responsibility of the managers and executives to lead the ethical behaviour of the company and act as a beacon that shines as a reference for ethical attitudes. The same study established the conditions that managers and executives are important to mature their reputation, especially by behaving as a “Moral Person” and a “Moral Manager”, and goes on to elaborate how leaders should communicate ethical standards to their subordinates (Treviño et al., 2000). These two-base behaviours are the genesis of Ethical Leadership recognition, in other words, “what you say is the way you act”. This proposal gained the status of a theory with further developments being added to organizational leadership studies based on ethical dimensions (Banks et al., 2021).

Modern authors agree that the conceptualization of ethical leadership as a “definition” is that developed by Brown et al., (2005, p. 120), which defines Ethical Leadership as being “the demonstration of normatively appropriate conduct through personal action and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making”. Nevertheless, despite its wide acceptance, this concept is still subject to continuing scrutiny. For instance, some authors call attention to the vagueness of the ethical leadership frame in its influence beyond direct relationship (Wang et al., 2017), while others stress the lack of guidelines for “being an ethical leader” (Malik et al., 2016), or the reduced scope of the application of ethical leadership in different organizational settings, such as Sales, for instance (Badrinarayanan et al., 2019). Despite such claims, the characteristics that an ethical leader must possess remains unchanged among scholars, namely: altruism, caring, fairness, honesty, and being principled (Badrinarayanan et al., 2019; Presbitero & Teng-Calleja, 2019).

There is also consensus among scholars regarding what subordinates and followers expect from the ethical leader, namely attitudes such as: balanced decisions, communicating ethical standards, conducting an ethical personal life, emphasizing moral outcomes, and punishment and rewards accordingly to moral settings, among others (Bonner et al., 2016; Malik et al., 2016). Shakeel et al., (2019, p. 615) state in their definition of ethical leadership is guided by “…principles that advocate learning motivation, healthy optimism and clarity of purpose to uphold the values of empowerment, service to others, concern for human rights, change for betterment and fulfilling duties towards society, future generations, environment, and its sustainability”. Bad governance and poor ethical practices translate into a poor ethical environment, whereby a leader must embrace the responsibility to motivate employees to follow the norms if they want to sustain high standards.

The presence of ethical leadership sustains the organizational ethical climate, and vice versa, as an ethical climate supports a dependent relationship with ethical leadership (Aryati et al., 2018). This interaction promotes ethical values in the organization and the followers maintain these values in constant feedback, creating a circuit that creates the ethical climate (Schminke et al., 2005). Plenty of literature links the role of a leader as an ethical model to team outcomes, i.e. the importance of their attitudes and behaviour, especially concerning integrity, respect, fairness, commitment toward employees to foster the team’s connection, and cohesion (Al Halbusi et al., 2020; Aryati et al., 2018; Bonner et al., 2016).

If leader’s ethical behaviour provides the model for how employees should act in the organization and leads to the development of ethical standards, then in turn the lack of these ethical attitudes fosters a poor and conflictive environment (Wimbush & Shepard, 1994). Thus, we hypothesize that

Hypothesis 4

Ethical leadership is negatively associated with an instrumental ethical climate.

An ethical climate is related with high moral standards which can be perceived as being right or wrong, and it is affected by what management stands for (Al Halbusi et al., 2020). Aryati et al. (2018) suggested that ethical climate is a promising mediator of other relationships, e.g. between ethical leadership and deviant behaviour. IEC was noticed as being a mediator between relational leadership and unethical pro-organizational behaviour (Zhang & Yao, 2019). In turn, when reviewing Zhang and Yao’s paper, a recent work by Almeida et al. (2021) found temporal effects, where IEC fully mediates the relationship between ethical leadership and moral disengagement. The above-described theory supports the inference that if IEC mediates the relationship between ethical leadership and employee’s ethical behaviour, then the latter occurs when individuals switch off the psychological process of morality (Shaw et al., 2020), and thus, we can deduct that IEC mediates the relationship between ethical leadership and moral disengagement. Consequently, we hypothesize that

Hypothesis 5

Instrumental ethical climate mediates the negative association between ethical leadership and displacement of responsibility.

As described above, all the theorized relationships steer us to the key question of our research, namely: “is ethical leadership effective in discouraging teams’ unethical behaviour?” The literature has certainly supported this hypothesis over the past decades and this hypothesis is sustained by theories that posit that ethical leadership fosters commitment and loyalty among employees and prevents bad behaviour (Treviño et al., 2000), while acknowledging that the leader is an influencer of ethical conduct in organizations (Jurkiewicz & Brown, 2000). Ethical leadership has been found to influence employees’ behaviour and performance (Malik et al., 2016), as well as teams’ moral efficacy (Peng et al., 2017), and it provides protection from unethical purchasing behaviour (Ko et al., 2019). The evidence mentioned in the literature does not diminish the fact that ethical leaders may be incapable of changing an individual’s behaviour without accomplishing certain previous steps, such as promoting a good ethical organizational climate. As a chain reaction, a good ethical climate can deter the psychological process underlying moral disengagement, leading ultimately to the prevention of unethical behaviour.

The literature reviewed suggests that ethical climate and moral disengagement operate as sequential mediators. However, several studies have analysed the mediation process by using these variables separately, for instance: ethical climate as a mediator of the positive relationship between ethical leadership and team’s moral efficacy (Peng et al., 2017), and ethical climate as a mediator of the negative relationship between ethical leadership and deviant behaviour in the workplace (Aryati et al., 2018). Moral disengagement activation has also been identified as a mediator for employees feeling shame or guilt due to their own unethical behaviour under an ethical leadership (Liu et al., 2012), as well as a mediator for the negative relationship between ethical leadership and employees’ unethical pro-organizational behaviour (Hsieh et al., 2020). Multiple mediations, especially for psychological variables, offer an accurate assessment and the strongest analysis of the data. Accordingly, by integrating the partial mediation effects found in the literature, we hypothesize that

Hypothesis 6

Ethical leadership is negatively associated with unethical behaviour through a sequential mediation by instrumental ethical climate and displacement of responsibility.

Corruption Environment

Corruption is a worldwide disease, which is much evident in less-developed countries (Scholl & Schermuly, 2020), albeit it is also present in highly developed countries, as highlighted in the recent ACFE report (ACFE, 2020). Corruption is of extreme relevance, and has been mentioned as being an important concern for the top world organizations, such as INTERPOL (“The effects of corruption are far-reaching, as it can undermine political, social and economic stability, and ultimately threaten the safety and security of society as a whole”), the UN (“Corruption undermines democratic institutions, slows economic development and contributes to governmental instability”), or the WBG (“Corruption has a disproportionate impact on the poor and most vulnerable, increasing costs and reducing access to services, including health, education and justice”). Shleifer and Vishny (1993, p. 599) wrote that “corruption is both pervasive and significant around the world”, and almost three decades later, Scholl and Schermuly (2020, p. 171) stated that “corruption is a plague in the world”.

Corruption is exceedingly difficult to define systematically. Davis and Ruhe (2003) tried to do so, but the concept remains blurred among different proposals. Bernardi et al. (2009), associated bribery with corruption, stating that it acts as a trigger for unethical behaviour. Cuervo-Cazurra (2016) went further, associating corruption with bribing a public agent, the abuse of power of public offices, and the selling of government property as if it is one’s own, among others. Feldman (2018) argued that corruption is a failure of public organizations to follow and uphold laws and procedures. However, what is common to all these scholars is that corruption is the selfish behaviour of individuals in key positions, such as public agents who seek personal gain. Based on the definition of Cuervo-Cazurra (2016, p. 16), “corruption is the abuse of entrusted power for private gain”, and it is not perceived in the same way by different cultures.

There has been an increase in the number of studies on corruption in international business over the present decade, with an average of 10.22 publications per year from 2010 to 2018, notoriously surpassing the average of 3.5 publications per year during the previous decade (Bahoo et al., 2020). The explanation for this notable increase could be the lack of consciousness from leaders or a competitive environment promoted by organizations that foster a less-ethical environment by increasing corruption levels and impelling team members to disengage their morality (Scholl & Schermuly, 2020). Moore et al. (2012) mentioned that a bad ethical climate is a driver of moral disengagement and consequently teams’ unethical behaviour. For instance, an organization’s reward structure exerts an influence on the ethical climate, generating greed and the desire to achieve financial goals through any means. Social norms that facilitate corruption are internalized by individuals and are, thus, adopted within the organization (Barr & Serra, 2010).

The facilitating effect that corruption exerts in the organization, and consequently on individuals, influences behaviours. Thus, we hypothesize that

Hypothesis 7

A corrupt environment moderates the relationship of ethical leadership with unethical behaviour through the instrumental ethical climate and the displacement of responsibility, in such a way that when the level of corruption is high, the relationship is weaker.

The conceptual model graphically depicted in Fig. 1 integrates all the above-stated hypotheses.

Fig. 1
figure 1

Conceptual model and hypotheses

Full size image

Method

Procedure and Sample

This study was designed to comprehend two-wave data collection targeting both leaders and their respective teams in dyadic design in order to mitigate common method bias (Podsakoff et al., 2003). A pre-test was carried out with a few individuals in selected organizations who helped to test the answering time for the survey and the understandability of the scales. The pre-test stage brought to light the need to expand the scales into another two languages (Spanish and Portuguese) to facilitate the acquisition of answers from non-English speakers, thus, collecting data from more participants.

Two sets of questionnaires comprising the measures for variables in the conceptual model were designed online, using Qualtrics software, and were distributed in the three selected languages: English, Portuguese, and Spanish. Concurrently, another set of variables was included in two more questionnaires that exclusively targeted the leader (which was also made available in the three languages). Respondents were asked to answer on a 5-point Likert-type scale (ranging from “strongly disagree” to “strongly agree”).

The waves were launched with a 4-week time lag. A code was attributed to each participant to match the data from both waves and to match teams with their respective leader. Sociodemographic questions were included in the first wave for all participants, both for description and control purposes.

The target sample for this study comprised individuals working in teams with a typical three to four members size, who all voluntarily agreed to participate in the study. The primary condition to participate in the study is belonging to work teams in private companies that actively carry out international business, namely in the logistics and transportation industry, more precisely as exporters, freight forwarders, importers, services multinationals, and shipping lines. The second criterion targets distinct regions that could be contrasted by the level of corruption rated by the CPI, namely for higher-level corruption (CPI > 50, Bolivia, Brazil, China, El Salvador, Panama, Pakistan, Peru, and Sri Lanka), and lower-level corruption (CPI < 50, Australia, Denmark, Portugal, United States, and Uruguay), because the study adopts this grouping as the contextual moderator variable. The sample comprises a total of 13 countries, with an initial sample of 226 participants distributed by 48 teams, in the first wave. After scrutinizing the data to identify and exclude missing, and non-matched cases in the second wave, the final valid convenience sample comprises 175 participants (134 dyads), distributed by 41 teams, with 40% of participants falling in the Low CPI group, and the remaining 60% in the High CPI group. Table 1 details the sample.

Table 1 Sample descriptors
Full size table

The team members sample (n = 134) comprises 54.5% male employees, distributed by an age range of mainly up to 50 years old (18–29 y.o. = 21.6%; 30–39 y.o. = 29.1%, 40–49 y.o. = 35.1%, 50 + y.o. = 14.1%), with a college-level education (83.5%). In turn, the team leaders’ sample (n = 41) is older than the team members sample (18–29 y.o. = 7.5%; 30–39 y.o. = 30.6%, 40–49 y.o. = 35.1%, 50 + y.o. = 27.1%), is mostly male (64.9%), and largely with a college-level education (91%).

Data Analysis Strategy

The data analysis followed a twofold strategy, starting with the test of the psychometric quality of the measures, followed by hypotheses testing. Psychometric quality comprises both validity and reliability. The validity, i.e. the capacity of measuring what is intended, is gauged by means of construct validity, which is indicated via factor analysis. Exploratory factor analysis quality is judged based on the Keiser-Meyer-Olkin (KMO) measure of sampling adequacy, as well as Bartlett’s test of sphericity. The criteria take 0.500 as the minimum KMO and a significant Bartlett chi-square statistic (p < 0.001) to infer whether there is sufficient shared variance between items to allow for the extraction of factors. Additionally, the analysis took into consideration Measure Sample Adequacy (MSA) statistics for each item, which should also attain 0.500. The extracted solutions included only items with communalities of at least 0.500, and the factors were extracted based on the Kaiser criterion (eigenvalues equal to or greater than 1). Extracted solutions must account for at least 60% of variance after rotation (Varimax), with this rotation being chosen based on both the theoretical criteria, as this enables clearer solutions and prevents variance inflation in the case of simultaneous use in subsequent analyses. Convergent validity is judged based on Fornell and Larcker’s (1981) criteria, where average extracted variance (AVE) should attain a value of at least 0.500. Reliability, i.e. the consistent pattern of behaviour of the items included in the same factor, was judged using a Cronbach Alpha that should attain at least 0.700, as well as on Composite Reliability (Joreskog r), which adopts the same criterion for acceptance. Should a given factor-solution fail to meet any of the previously stated criteria, then the harmful item is removed, and the analyses are repeated. Process macro was used to test the direct, indirect, and interaction effects, based on a bootstrapping procedure with 5000 repetitions (Hayes, 2018).

Measures

Ethical Leadership (Wave 1 “teams”) was measured with Brown et al.’s (2005) 10-item scale (e.g. your immediate supervisor / manager, “…listens to what employees have to say”, “…disciplines employees who violate ethical standards”), comprising a single factor. The exploratory factor analysis showed a single factor with acceptable indices, but with two items showing substandard commonalities, and after the removal of these items, the solution showed good validity (KMO = 0.911, 0.886 < MSA < 0.957, Bartlett X2 = 846.191, 28 df, p < 0.001), with a single factor explaining 69.7% variance, and showing good reliability (alpha = 0.932, CR = 0.938). The solution also has convergent validity (AVE = 0.656).

Instrumental ethical climate (Wave 1 “teams” and “leaders”) was measured with Victor and Cullen (1988) subscale and answered both by Team members and Team leaders, comprising a four-item scale (e.g. “in this company people protect their own interest above other considerations”, “in this company people are mostly out for themselves”). The exploratory factor analysis showed one case of substandard commonality. After the removal of that item, a single valid factor emerged (KMO = 0.721, 0.704 < MSA < 0.757, Bartlett X2 = 199.976, 3 df, p < 0.001) explaining 75.0% variance, and showing good reliability (alpha = 0.833, CR = 0.826) and convergent validity (AVE = 0.749).

Displacement of responsibility (Wave 2 “teams”) was measured with Bandura et al.’s (1996) four-item scale (e.g. “people cannot be blamed for misbehaving if they are being pressured to do it”, or “nobody should be blamed for using an unpolite language if others also do it”). The exploratory factor analysis showed a valid single factor solution (KMO = 0.729, 0.683 < MSA < 0.806, Bartlett X2 = 190.604, 6 df, p < 0.001), explaining 63.6% variance and showing good reliability (alpha = 0.808, CR = 0.811). The solution also has convergent validity (AVE = 0.635).

Unethical behaviour (Wave 2 “teams”) was measured by Mayer et al. (2012), comprising a 10-items scale (e.g. in this country workers tend to, “…use company services for personal use”, or “…pass blame for errors to an innocent co-worker”). The exploratory factor analysis showed a case of substandard commonality. After the removal of that item, a single valid factor emerged with good validity (KMO = 0.900, 0.843 < MSA < 0.939, Bartlett X2 = 1008.138, 36 df, p < 0.001), explaining 68% variance and showing good reliability (alpha = 0.940, CR = 0.940), as well as convergent validity (AVE = 0.639).

Corruption Environment was measured using the three-year average for CPI Corruption Perception Index in its 2019–2021 edition. This study is carried out by Transparency International, a non-profit global organization which is present in more than 100 countries, with the mission to stop corruption and promote transparency worldwide. The CPI ranks 180 countries and territories by their level of perceived corruption in the public sector and in the opinion of experts and businesspeople from the private sector. The CPI classifies countries on a scale of 0 to 100, with 0 being “highly corrupt”, and 100 “very clean”. For this research, we classify the country’s team that participates in the study into two levels: HCE High Tendency to Corruption Environment, with a scale CPI < 50, and LCE Low Tendency to Corruption Environment, with a scale CPI > 50. A detailed description can be found in Table 2: CPI, rank, and corruption level. As this is a formative construct, it is not subjected to factorial validity testing.

Table 2 Corruption environment level (CPI)
Full size table

The control variables comprehended age (1 = 18–29; 2 = 30–39; 3 = 40–49; 4 = 50–59; 5 = 60–69; 6 = 70 or older), gender (1 = Female; 2 = Male), and education (1 = Basic; 2 = High school; 3 = University graduate; 4 = University post-graduate), as these have been reported as being related to the perceptions of ethical climate and ethical behaviour (e.g. Peterson et al., 2001; Wang & Calvano, 2015).

Results

Considering the complexity of our model, the data were analysed using a regression model based on Conditional Process Analysis—Process Version 4 for SPSS, following the distinct relationship models that were suitable for each of the hypotheses we wanted to test (Hayes, 2018). The two-time data collection comprised the following variables/waves: ethical leadership (by teams in Wave 1); instrumental ethical climate (by teams and leaders in Wave 1) combined by the means of both (“average IEC”), displacement of responsibility (by teams in Wave 2), and unethical behaviour (by teams in Wave 2) variables. CPI is the moderator variable used to test H7.

The descriptive statistics, correlations and Cronbach’s α are shown in Table 3.

Table 3 Descriptive statistics, correlations and Cronbach’s α
Full size table

The analyses were run several times, using two models in Process, in order to test the robustness and stability of the data in different scenarios. It is important to highlight that the outcomes were stable over the different scenarios.

We used Process Model 4 to test our first sequence of Hypotheses 1–3. The analysis showed that displacement of responsibility is positively associated with unethical behaviour (b = 0.208, t = 2.344, p = 0.021, 95% CI [0.033; 0.384]), thus, supporting H1. Similarly, instrumental ethical climate is positively associated with displacement of responsibility (b = 0.358, t = 2.937, p = 0.004, 95% CI [0.117; 0.599]), which, thus, supports H2. Lastly, instrumental ethical climate is associated with unethical behaviour through displacement of responsibility, as shown by the significant positive indirect effect (effect = 0.082, 95% CI [0.005; 0.217), thus, supporting H3.

Process Model 4 was also used to test the second sequence of Hypotheses 4 and 5. The analysis showed that ethical leadership is negatively associated with instrumental ethical climate (b = − 0.254, t = − 3.313, p = 0.001, 95% CI [− 0.406; − 0.102]), thus, supporting H4. Lastly, ethical leadership is associated with displacement of responsibility through instrumental ethical climate, as shown by the significant negative indirect effect (effect = − 0.090, 95% CI [− 0.192; − 0.021]), which supports H5.

To test Hypotheses 6 and 7 we used Process Model 92 for the sequential mediation and respective interaction effects. Hypotheses 6 proposes a sequential mediation from ethical leadership (Tw1) to unethical behaviour (Tw2) via instrumental ethical climate (average w1) and displacement of responsibility (Tw2). Findings show a significant negative indirect effect (effect = − 0.020, 95% CI [− 0.056; − 0.007]), which, thus, supports H6. It is important to emphasize that Process Model 92 also runs direct effect tests simultaneously and provided the same results for the previous hypotheses tests conducted with Process Model 4, namely: (H1 supported, b = 0.194, t = 2.189, p = 0.031, 95% CI [0.018; 0.368]); (H2 supported b = 0.380, t = 2.930, p = 0.004, 95% CI [0.124; 0.638]), (H4 supported, b = − 0.268, t = − 3.525, p = 0.006, 95% CI [− 0.419; − 0.118]), which corroborates the consistency of the analyses.

Finally, we used Process Model 92 to test the effect of a corrupt environment as the moderator on the relationship of ethical leadership with unethical behaviour through instrumental ethical climate and displacement of responsibility. The expected effect hypothesized that when corruption environment is high (lower value on the scale), the relationship is weaker, thus, denoting that countries’ corruption level affects the ethical environment of the organizations. The results showed no interaction among variables, neither on the direct effect, nor on the indirect effects. H7 is, thus, not empirically supported. Table 4 presents a summary of the above-mentioned results.

Table 4 Direct, indirect and interaction effects
Full size table

Discussion

Unethical and illegal behaviours undermine corporate legitimacy, credibility, and trust (Sale, 2019). This perception is commonly shared across countries, transcending economic, political, or legal climates established by different cultures (Payne et al., 1997). Ethics must prevail in any social environment, including the ones pursuing economic goals (Martini & Spataro, 2018). Nevertheless, the ethical factor is all too often neglected by organizations and top executives alike. It seems that the pursuit of objectives without consideration for the means is still overriding organizational values (Sarwar et al., 2020; Wedell-Wedellsborg, 2019). This impelled us to suggest that ethical leadership plays a central role in establishing these organizational values in order to counter unethical behaviours.

Our research results suggest that ethical leadership can be the agent that deters unethical behaviour, by negatively influencing the psychological process of individuals’ moral disengagement. This achievement at the individual level simultaneously causes an effect at the organization’s social level by limiting a poor ethical climate. The process that has been hypothesized to explain this phenomenon follows three logical arguments, namely: (1) if an individual can morally disengage, they will be easily tempted to act unethically; (2) moral disengagement, namely, displacement of responsibility will occur more easily in ethically poor environments, since these can be pointed out as the real (external) reason for behaving unethically; and, consequently; (3) unethical environments generate the displacement of responsibility that fosters unethical behaviour. Our findings supporting the first three hypotheses are in line with the mediation effects reported by Newman et al. (2019). The role that leadership plays in organizations is widely acknowledged as being a key driver of the ethical climate (Mulki et al., 2009). Adding to the previous sequence, we proposed that counteracting unethical behaviours is within the reach of ethical leaders, based on its sequential influence on the corporate ethical climate and moral disengagement. Accordingly, it is reasonable to expect a sequential mediation that incorporates indirect effects that are already documented in the literature (e.g. Zhang & Yao, 2019), and thus, our findings received full empirical support and integrate both the lines of research that focus on ethical climate, either as an antecedent or as a process.

Following a contextual-based analysis of the ethical phenomena, the conceptual model assumed that the organizations are sensitive to their external ethical environment, which in turn depends on the level of corruption established in the culture (Barr & Serra, 2010; Fisman & Miguel, 2007). The use of country-level corruption as a moderator was intended to demonstrate that individuals and organizations in a highly corrupt environment are influenced by their own society, demonstrating low ethical standards and a poor ethical climate (Bahoo et al., 2020). Surprisingly, our study did not support this hypothesis, thus, opening a possible path for a new understanding regarding environmental corruption at the organizational level.

Contributions to the Literature

The findings of our research offer three contributions to the literature on ethics, leadership, and corporate responsibility. Firstly, the literature on processes that mediate the leaders’ ability to disable unethical behaviour is still scarce (Hsieh et al., 2020; Seriki et al., 2020; Shaw et al., 2020) and only a few works have empirically addressed this subject with this level of combination and complexity of the variables (e.g. Barsky, 2011; De Cremer and Vandekerckhove, 2016; Hiekkataipale & Lamsa, 2019; Kia et al., 2019). Our study offers a contribution to the extant literature by suggesting a sequential psychological process that has stronger theoretical integrative power. Additionally, the complexity of the model is matched by its integrative role in bridging the relevant theories, namely the Moral Disengagement Theory (Bandura, 2002b) as an extension of the Social Cognitive Theory (Bandura, 2002a), which offers a strong basis to explain unethical behaviour based on self-regulatory dysfunctions and the Rational Egoism Theory (Rand, 1964), which complements the former by adding IEC as an antecedent of moral disengagement. Ethical leadership theory (Treviño et al., 2000) is brought to the model to offer a view on how ethical climates can be produced or prevented by leaders through influence processes (Aryati et al., 2018).

Secondly, closely related to the first contribution is the overarching nature of Social Cognitive Theory which is expressed in our model from a situated point of view, where ethical agency is set against its macro societal context. CPI becomes an important contextual element that is in line with testing the dualism of personal agency versus social structure (Bandura, 2002a), more popularly known as “bad apples” and “bad barrels” (Kish-Gephart et al., 2010), which are represented in our model as (un)ethical leadership and CPI respectively. CPI has gained such status in business ethics research (e.g. Tang et al., 2018) and the inclusion of boundary conditions has always been considered to be a theoretic advancement in organization science (e.g. in organizational theory, e.g. Burns & Stalker, 1961; in leadership research, e.g. Fiedler, 1964, or in HRM, e.g. Martin-Alcazar et al., 2005, just to name a few), albeit it is an open door to relativism (Demuijnck, 2015). The absence of the moderation effects is not in line with contingency theory, but it highlights the primacy of ethical leaders over the organizational context, thus, suggesting that good apples may not become rotten in bad barrels.

Thirdly, the inclusion of the corruption variable in an international context in this research. Undoubtedly, corruption has proven to be a critical issue, which causes great upset among the general public (Hail et al., 2018), however research in this area is still in its infancy, with scarce literature on international business (Bahoo et al., 2020; Di Guardo et al., 2016; Tang et al., 2018). Our findings offer useful insights into business ethics and the international business literature, as described below.

Managerial Implications

Lastly, our findings have practical implications. The process linking ethical leadership to unethical behaviour signal the importance of the ethical leader in organizations. In times when business ethics are under intense scrutiny, our findings show that ethical leaders can be a key factor to promote a protective organizational ethical climate among employees. The enhancement of the individual’s self-regulatory process prevents employees from engaging in unethical actions. Additionally, another practical implication concerns those stakeholders and business owners who should recognize that the ethical leader is an asset to the organization, rather than a liability due to opportunity costs. Being an ethical leader implies, for instance, caring for environmental protection, supporting employees’ performance, rewarding good attitudes and punishing bad ones, and being fair, trustable, and balanced, with all these actions resulting in a positive return in financial performance and social impact. Finally, the last practical implication of our findings concerns how ethical climate should be conceived by decision makers. More than being just a mere dimension of organizational analysis and academic research, ethical climate should be utilized as a strategic KPI that is reported by each department or organizational unit, which is used to hold leaders accountable.

Limitations and Future Directions

Our study suffers certain limitations. Although the research design followed guidelines intended to overcome common method issues by temporally separating the data collection and using independent data sources (Podsakoff et al., 2003), the sequential model effect does not allow for causality inferences. Additionally, some options for the measurement of country corruption may have to be given a second thought, especially regarding CPI, which is a product of business people’s perception of each country. However, institutional corruption is related to the theory and practices of those organizations that are susceptible to create their own kind of corruption, which tend to become more deceptive, illusory, and unscrupulous (Thompson, 2018). Future research should analyse the moderation of institutional corruption using the conceptual model of our research, in substitution of CPI, which could offer new insights.

Additionally, the sampling is non-random and the sample size is quite modest, although the model has a relatively low number of estimates. Although the use of bootstrapping may not be helpful in countering representativeness bias (Hoyle, 1999), it does add more stability with high repetitions (Hayes, 2018). Future research may extend the conceptual model and test it with a larger sample from geographies that have not been included, but which may still represent extreme values in CPI.

Lastly, the value of ethics is universal by definition, although ethical values can vary according to culture at the national level. Because our data collection considered teams from different countries and different industries, the cultural aspects may have a relevant influence on behaviour, as was found by Eisenbeiß and Brodbeck (2014). Therefore, another limitation in our study is related with the non-consideration of cultural differences (both within-national and within-organizational levels). This limitation is somewhat mitigated by the fact that shared perceptions of corruption, as depicted in CPI, have the potential to be seen as an expression of the culture itself, thus, justifying the use of cultures of corruption (Schneider & Bose, 2017).

Conclusions

Improving an organization’s profit and attaining goals nowadays are not sufficient achievements to be considered as an effective business leader. The way that these targets are achieved and the way that teams perform make all the difference in the success or failure of modern organizations. Any attempt to attain organization performance and recognition is now subject to the scrutiny and approval of the society in which the organization is embedded—where not only the ends matter, but also the means employed to achieve them.

This study suggests that country corruption is not a synonym for organizational corruption. The sample was purposely chosen within business organizations, as CPI is built upon the perceptions of corruption as stated by those businesspeople that can affect commercial activity. Nevertheless, no relationship exists between the country’s level of corruption and the company’s unethical behaviour associated with corruption actions. This may create a precedent for those CEOs and business owners who are planning to expand their participation abroad. For if the ethical standards of new business partners are solid, this may be sufficient to risk finalizing new business deals with them, independently of which country they are based in. More importantly, our findings suggest that the influence of an ethical leader within the corporate boundaries is more powerful than the spillover effects from outside the corporate boundaries.

Overall, ethical leadership is an effective driver of teams’ ethical behaviour, by operating in accordance with the ethical climate that engages individuals morally. Furthermore, ethical leadership seemingly acts equivalently across all the countries surveyed, suggesting a universal theory for its effectiveness. Similarly, the contextual influence of the rotten barrel over the influence of rotten apples seems to be overstated, as our study found that a corruptive context was not effective in annulling the chain effects adopted in the model that originated in the ethical stance of leaders.