Affective and Normative Commitment as Intervening Variables of the Links between Ethical Leadership, Religiosity, and Fraud

Article history: Received date: 14 June 2020 Received in revised form: 21 August 2020 Accepted: 1 September 2020 Available online: 25 September 2020


Introduction
Fraud mitigation has become a crucial issue both in private and public sector organizations. Governments have increased their efforts to mitigate fraud risks due to the rise of fraud occurrence, public demands for better internal control, and citizens' dissatisfaction with the scale of loss caused by frauds (Savovska, Vatyan, Danescu, & Crnkovic, 2017).
In Indonesia, Komisi Pemberantasan Korupsi/ KPK (or Indonesian'a Corruption Eradication Commission) has recently issued guidance of corruption prevention for companies due to the increasing number of fraud in the public sector that involves private companies (Komisi Pemberantasan Korupsi, 2019). For the last five years (2013 -2018), 370 individuals in public sector committed corruption. Around 60% of the total individuals who perpetrated white-collar crime were exposed by KPK (see Table 1). The fraud perpetrators were not only low-level employees but also top-level ones. For example, in July and September 2019, KPK arrested a mayor and a governor for receiving bribery from private companies (Rachman, 2019;Irawan, 2019). Two members of the Indonesian House of Representatives (DPR) were also arrested for committing the same crime in March and August 2019 (Rachman, 2019;.  I/II/III  24  43  10  7  2  86  Judge  -5  3  1  3  12  Prosecutor  1  -1  3  -5  Total  161  82  49  40  38 370 Source: (Komisi Pemberantasan Korupsi, 2019) Among factors that may mitigate fraud namely, control environment, has been extensively studied, particularly on ethical leaderships, religiosity, and commitment. One of the reasons why a control environment is important in mitigating fraud is that leaders have an important role in the control environment due to their power to enforce integrity and ethical values and to monitor various levels of the organization (COSO, 2013). However, prior studies have several limitations. First, these studies tend to focus only on private companies (Posey, Roberts, & Lowry, 2015;Mohdali & Pope, 2014;Gorjidooz & Greenman, 2014;Philipp & Lopez, 2013;Bhal & Dadhich, 2011). Thus, this study investigates fraud in the context of government institutions.
Second, although prior studies demonstrated the importance of ethical leadership and religiosity in influencing employees' ethical behavior (Pope, 2014;Walker, Smither, & DeBode, 2012) and in reducing financial misreporting in certain circumstances (Chen, 2010), the studies unrevealed how organizational commitment, specifically affective and normative, may also mediate the effect of ethical leadership and religiosity to ethical behavior.
Meanwhile, several prior studies have provided empirical evidence that employees with a higher commitment to their organization tend to have more intention to fully contribute to their working place and act ethically (Posey et al., 2015;Lu & Lin, 2013;Caillier, 2013). Some studies have found that the employees' commitment is significantly affected by how they perceive the ethical behavior of their leaders (Blau, 1964as cited in Hansen, Alge, Brown, Jackson, & Dunford, 2012 and their level of religiosity (Farrukh, Ying, & Ahmed, 2016). However, in some cases, normative commitments may reduce employees' ethical behavior. This is because employees with strong normative commitment may resist to act differently with their groups (Becker, 1960). Consequently, if fraud is deeply internalized into an organization where fraud is perceived as normal acts in various levels, employees with strong normative commitments may conform and do the fraudulent acts.
Third, past studies about fraud mostly examined the determinants of fraud (Murphy & Free, 2016;Yusof & Lai, 2014;Cohen, Yuan, Lesage, & Stolowy, 2010). Studies that examined factors that hinder employee's legal fraud engagement in the public sector are still scanty. Legal fraud is different from ethical fraud. Legal fraud is as a fraud that breaches the existed laws, but ethical fraud is an action that is not in line with existing ethical norms (Anand, Dacin, & Murphy, 2015).
This study, thus, extends prior research by focusing more on the impact of ethical leaderships, religiosity, and commitment to legal fraud mitigation in government institutions. This study aims to evaluate the mediating effect of organizational commitments, particularly affective and normative commitments to the relationship between ethical leadership, intrinsic religiosity, and government employees' legal fraud engagement.
This research contributes to fraud literature and practices. It evaluates not only the mitigation effect of ethical leadership, religiosity, affective, and normative commitments on fraud in government institutions, but also the mediating effect of affective and normative commitments. The research findings can be used as a reference for the government to improve its public sector internal control, especially the control environment primarily because the increasing number of local leaders who committed corruption raises concern about the effectiveness of the control environment in government institutions. If leaders of government institutions committed fraud, the internal control of the institution is jeopardized, which may increase the fraud risk.
This paper is divided into 5 sections. Section II defines and discusses the concept of the variables, results of prior studies, and finally propose hypotheses. Section III describes in detail the research design adopted in this study. Section 4 comprises the finding interpretation and discussion. Lastly, section 5 provides conclusions and recommendations for further researches.

2.
Literature review and hypotheses development Ethical leadership and employee's engagement on legal fraud Ethical leadership is defined as "the demonstration of normatively appropriate conduct through personal actions and interpersonal relationships, and the promotion of such conduct to followers through two-way communication, reinforcement, and decision-making" (Brown, Trevino, & Harrison, 2005, p. 120). The belief that ethical leadership may have an impact on employee behavior and performance relies on social learning theory.
Social learning theory postulates that when individuals perceive certain behaviors as modeled behaviors, they tend to observe and to emulate those behaviors (Bandura, 1977). Due to their positions in organizations, a leader is expected to act normatively (e.g. fairness, integrity, reliable) to become a role model for normative behaviors. The leader also communicates the ethical conduct to their employees, reinforcing it through rewards and punishments, and ethical decision-making (Neubert, Carlson, Kacmar, Roberts, & Chonko, 2009). Their subordinates will then perceive their leader as an honest, fair, responsible, and ethically acting in their personal and professional lives (Mayer, Aquino, Greenbaum, & Kuenzi, 2012). As a result, employees will learn what kind of behavior or outcome is expected for them.
Employees in private and public sectors are also more willing to report wrongdoings when they perceive that their leaders act ethically (Wen & Chen, 2016;Zhang & Liao, 2016;Hassan, Wright, & Yukl, 2014;Bhal & Dadhich, 2011). The finding of Said et al. (2018) may differ from many other studies possibly because, to some extent, fraud occurrence in companies is more affected by factors other than individuals' own willingness to act unethically. For example, Chen (2010) suggests that leaders may act unethically not because they want to, but because of the pressure from their shareholders' and other parties' expectations. CEOs would keep acting fraudulently only if they feel they cannot remove themselves from being responsible for the frauds they have made. Hence, drawing on the above discussion, we expect that ethical leadership may reduce employees' fraud. H1: Ethical leadership negatively affects public sector employees' engagement with legal fraud.

Intrinsic religiosity and employee fraud engagement
Religiosity or religious commitment can be defined as the level to which an individual follows their religious feeling and belief, and implements them in their daily life (Worthington, 1988as cited in Worthington Jr et al., 2003. It is a person's faith in God accompanied by the person's commitment to implement principles set by God (McDaniel & Burnett, 1990). Religion helps individuals to decide about what is right or wrong (Modarres & Rafiee, 2011). According to religious values theory, people with high religious values may set certain values and beliefs to evaluate their world (Worthington, 1988as cited in Worthington Jr et al., 2003. This theory is supported by various studies. For example, Putrevu & Swimberghek (2013) find religiosity contributes positively to a person's moral or ethical judgments. It increases people's ethical awareness about ethical issues (Ali, Kuzay, Gungormus, & Alas, 2015). Religiosity encourages individuals to voluntarily comply with tax regulation (Pope, 2014). In the field of fraud, religiosity also plays an important role to reduce fraud occurrence. Individuals with higher religious commitment also tend to commit less fraud (Said et al., 2018).
Religiosity comprises both intrinsic and extrinsic orientation (Allport & Ross, 1967). Intrinsic religiosity means that individuals' beliefs or personal religious internalized into their daily life. Meanwhile, extrinsic religiosity means that individuals tend to practice the doctrine taught by their religious affiliations for social or their benefits, for example, security, self-justification, social status, and so forth. Individuals are usually between these two continuums.
This study particularly evaluates the impact of intrinsic religiosity to individual fraud because individuals with high intrinsic religiosity are more willing to use their beliefs as guide guidance to make day-to-day decision making and behavior (King & Crowther, 2004). In other words, religion will positively affect individuals' ethical behavior only when individuals live their religious guidelines.
Result findings generally indicate that intrinsic-religiousness individuals tend to evaluate circumstances around them by using religious criteria and act based on the criteria (Worthington Jr et al., 2003). Walker et al., (2012) also demonstrate that intrinsic-religiousness individuals are more reluctant to accept ethically questionable scenarios, while extrinsic-religiosity individuals are more willing to accept the ethically questionable scenarios. This is consistent with the study of Arli & Tjiptono (2013) in Indonesia that suggests individuals with stronger intrinsicreligiosity tend to disagree more with consumer unethical behavior compared with individuals with stronger extrinsic-religiosity. Thus, concerning the religiosity, we expect the following: H2: Employee's intrinsic religiosity negatively affects public sector employees' engagement to legal fraud.

Ethical leadership, affective commitment, and employee's engagement to legal fraud
Affective commitment, which was introduced by Meyer & Allen (1984, 1991, is rooted in attitudinal commitment theory (Mercurio, 2015). The theory argues that when employees are attached emotionally to organizations they work at, the employees are likely to commit to contributing to their organizations (Kanter, 1968). Previous studies suggest that affective commitment reduces employee turnover and absenteeism (Garland, Hogan, Kelley, Kim, & Lambert, 2013;Somers, 2009 How ethical leadership may have an impact to employee's affective commitment is based on the social exchange theory. The theory suggests that ethical leadership is likely to influence employees' commitment to the organizations (Blau, 1964 as cited in Hansen et al, 2012). At the early stage, relationships between a leader and their subordinates are more characterized by rules and lower trust between each other. Since ethical leaders tend to have high integrity and act fairly, they are likely to develop better rapport with their subordinates. Over time, the better relationship improves subordinates' commitment to the leaders or organization represented by the leaders (Neuber et al., 2009;Brown et al., 2005). As a role model, ethical leadership also seems to develop a committed environment (Ruiz, Ruiz, & Martinez, 2011). Ko & Hur (2013) find that manager trustworthiness negatively affects employees' intention in the public sector to leave their current organizations. Similarly, Hassan, Wright, & Yukl (2014) suggest that ethical leadership plays a prominent role in employees' organizational commitment in public agencies in the United States. Absenteeism tends to be lower when employees perceive that their leaders are ethical.
When employees' affective commitment is strong, employees may very likely become emotionally attached to their profession or to the organizations they work at (Meyer, Allen, & Smith, 1993;Meyer, Allen, & Gellatly, 1990). In government institutions, the motivation of public sector officers with strong affective commitment is likely guided by certain professional ideals, for example, by their commitment to work for public interest rather than personal benefit (Shafer, Simmons, & Yip, 2016). Devece et al., (2015) provide empirical evidence that the relationship of affective commitment and employees' organizationalcitizenship behavior does not change in a low or high unemployment environment. Kim (2012) found that affective commitment mediates the effect of transformational leadership on organizational citizenship behavior in the public sector. Transformational leadership may only improve organizational citizenship behavior through increasing of employees' affective commitment to organizations.
Taken together, this suggests that ethical leadership may increase employees' affective commitment, causing employees to avoid fraudulently acts. This study thus proposes the following hypotheses: H3: Ethical leadership positively affects public sector employee's affective commitment. H4: Public sector employee's ethical commitment negatively affects public sector employee's engagement to legal fraud.

Intrinsic religiosity, normative commitment, and employee's engagement to legal fraud
Normative commitment refers to the extent to which employees feel they need to remain at their current organization because they feel obligated to do so (Meyer et al., 1993). Employees' sense of obligation to remain in the company is influenced by the internalization of normative pressures experienced by the employees, for instance, loyalty and duty, family expectation, or advance incentives given by their organization (Allen & Meyer, 1996;Wiener, 1982). In other words, normative commitment suggests that employees' commitment to the organization is based on the belief that they have a moral obligation to perform values of loyalty and sense of duty. This is consistent with the values of religiosity.
Religiosity also encourages individuals to highly value religious values, including loyalty and responsibility, which is likely to cause individuals to be more committed to what they do including the commitment to the organizations or professions (Farrukh et al., 2016). Several empirical evidence reveal the impact of religiosity on normative commitment. For example, teachers who work in Catholic Schools commit to their profession because they perceive their profession as a vocational call from God (Mancuso, 2002). Religiosity also improves educators to normative commitment in Pakistan (Farrukh et al., 2016).
Individuals with strong normative commitment are likely to work based on normative values, e.g. profession's ideals (Shafer et al., 2016). Their belief that they must be loyal and responsible may indicate that they "are willing to make a personal sacrifice for the sake of the organization" (Wiener, 1982). They believe that it is the right thing to do.
People with strong normative commitment may also believe that it would be better for them to leave their organization when they perceive that the organization does not act ethically anymore (Domino, Wingreen, & Blanton, 2014). A study conducted by Meyer, Stanley, Herscovitch, & Topolnytsky (2002) suggests that there is a correlation between normative commitment and employees' organizational citizenship behavior. Employees with higher professional commitment also tend to be more reluctant to engage in tax fraud (Shafer et al., 2016).
To recap, this suggests that intrinsic religiosity is likely to improve employees' normative commitments, which leads employees to avoid acting fraudulently. This study thus proposes the following hypotheses. Figure 1 depicts the theoretical framework of this study. H5: Employee's intrinsic religiosity positively affects employee's normative commitment. H6: Employee's normative commitment negatively affects employee's engagement to legal fraud.

Research method
This study takes several actions to minimize common method biases (Podsakoff, MacKenzie, & Podsakoff, 2012). For example, the data collection procedure was designed to increase participants' motivation to provide accurate answers, while it also has to decrease participants' intention to answer the question that they may not understand or experience. The questionnaire consists of four sections that focus on ethical leadership, religiosity, affective and normative commitments, and legal fraud. Measurement items in each section of the questionnaire have been used and validated in previous studies. Each section applies the 5-Likert scale as summarized in Table  2.
Section A of the questionnaire requires the respondents to evaluate the ethical leadership at the organization where they work by utilizing ten measurement items developed by Brown et al. (2005) and also used in Lu & Lin (2014), and Demirtas & Akdogan (2015). Section B asks the respondents about their degree of religiosity. It is adapted from the Religious Commitment Inventory (RCI-10) developed by Worthington Jr et al. (2003) and has been extensively used in many studies with various samples (Farrukh et al., 2016;Mohdali et. al., 2016;Martin & Bateman, 2014;Putrevu & Swimberghek, 2013). The RCI-10 measures two types of commitment (motivational and behavioral) to a religious value system regardless of the beliefs or faiths in the value system (Mokhlis, 2009). The scales comprise six statements about intrapersonal religiosity (cognitive) and four statements expressing interpersonal religiosity (behavioral). This study only measures the intrapersonal (intrinsic) religiosity of respondents.
Section C measures the participants' affective and normative commitments by the measurement developed by Meyer, Allen, & Smith (1993), Gellatly (1990), andMeyer (1996). The scale has been used in many studies (Farrukh et al., 2016;Caillier, 2013;Kim H, 2012). Respondents are required to answer six questions for each affective and normative commitment. In Section D, respondents are required to inform the level of their engagement in various misconduct activities, particularly asset misappropriation, corruption, claiming false reimbursement, and misuse of data privacy for the last 3 years. The items adopt only the legal fraud used by Said et al. (2018) with modification. The scale is from never happen (scale 1) to very high (scale 5).  Six items about intra-personal religiosity RCI-10 developed by Worthington Jr et al.,

5-Likert
Normative Commitment (NC) Normative Commitment is a feeling of responsibility of an employee towards his organization to stay afloat.

5-Likert
Legal Fraud Engagement (LF) Legal fraud is an individual's involvement in fraud that breach law in the last 3 years of their working experience.
Four items were adopted from Sa id et al. (2018) with modification.

5-Likert
As this study used a survey to collect data, the respondents are government institution employees, specifically those who have worked more than 3 years and more than 25 years old. The reason is that one of the questions in the questionnaire asked the participants about their level of engagement to fraud in the last three years of their working experience.
The age criterion was required as it is less likely to have three years of working experience as full-time government employees at a younger age. Due to the specific target respondent, the study does not have any information about the population characteristics. Hence, with that targeted respondent, the cluster sampling method was selected. Cluster sampling is a sampling selection strategy with providing specific criteria at the available random population (Hair, Celsi, Money, Samouel, & Page, 2011). Two hundred and fifty (250) questionnaires were sent to four government institutions in the period January -March 2020 through each of the institutions' public relations department. The number of returned questionnaires was 177.
A total of 168 respondents met the sample criteria. However, 65 surveys were eliminated because of incompleteness. As a result, there were 111 usable responses or a 44.4% response rate.

Results and discussion Respondents' profiles
To ascertain whether the participants understand about ethical leadership in the institution they work at, they need to indicate how long they have worked in their current institution. Table 2 shows that almost half of the participants (43.2%) have worked in their current institution for more than 15 years, while 15.3% and 16.2% of participants have 11-15 years and 6-10 years of working experience, respectively. Only 25.2% of participants who have 3-5 years of working experience as a public officer.
Further, demographic information including age, gender, and work experience was also collected. Most respondents (62.2%) are males and more than half of participants (54%) are under 41 years old. The minimum respondents' education level is a diploma degree (31.5%). More than half of the participants (68.5%) have obtained at least an undergraduate degree.

Validity and reliability tests of the measurement model
The study used SMART PLS 3 version 3.3.2 for data analysis. SMART PLS 3 is a nonparametric statistical tool to analyze relatively small sample sizes and complex models, and makes no distributional data assumption (Hair, Hult, Ringle, Sarstedt, 2017). The convergent and discriminant validities were analyzed by using factor analysis and Average Variance Extracted (AVE) method and cross loading analysis. Convergent validity examines the representation of the factor parameter to their latent variable (Hair et al., 2017). Convergent validity (Table 4) shows the results of factor loading analysis and AVE. Factor analysis summarizes a large number of criteria into fewer dimensions. Only variables with a factor loading 0.7 or above were considered as valid (Hair et al., 2017). All items, except NC6 (factor loading = 0.562) are valid. AVE result also reveals that each construct's validity result was more than 0.5. AVE value of 0.50 or higher indicates that the construct is able to explain most of the variance of its indicators.
Discriminant validity is the extent to which a construct is unique compared with other constructs.
Good discriminant validity indicates that the observed items share more common variance with their respective construct than any variance that the construct shares with other constructs (Koufteros, 1999;Fornell & Larcker, 1981). Table 5 presents that discriminant validity for each variable was upheld. For example, the loading value of EL1 until EL10 was larger on ethical leadership compared to their cross-loading values on other constructs. Similarly, loading values of AC1 until AC6 were larger on affective commitment compared to their cross-loading values on ethical leadership, normative commitment, and other constructs.
The Cronbach's Alpha and composite reliability are used to measure the reliability of observed indicator variables. Both of them are common internal consistency measurement. The reliability test result (Table 4) shows that the values of Cronbach Alpha and composite reliability for all variables were above 0.7. Hair et al. (2017) indicate that Cronbach's Alpha and composite reliability value of 0.7 or higher can be considered as satisfactory. This result provided evidence for reliable latent variables.

Results of hypotheses tests
To test the hypotheses, this study used partial least square structural equation modeling (PLS-SEM). The research model in Figures 2 and 3 displays that there are six direct and two indirect paths to examine. The direct paths are ethical leadership and legal fraud (EL  LF), ethical leadership and affective commitment (EL  AC), affective commitment and legal fraud (AC  LF), intrinsic religiosity and legal fraud (IR  LF), intrinsic religiosity and normative commitment (IR  NC), and normative commitment and legal fraud (NC  LF).
The indirect paths are ethical leadership, affective commitment, and legal fraud (EL  AC  LF), and intrinsic religiosity, normative commitment, and legal fraud (IR  NC  LF). The value of adjusted R-Square presented by Figure  3 presents that the coefficient determination of legal fraud is 90.1%. It suggests that overall, 90.1% of the variations in legal fraud value can be explained by predictors examined in this model. Only 9.9% of the variations in legal fraud value are affected by other factors outside the model. Ethical leadership may also explain 29.5% of variation in affective commitment value, while intrinsic religiosity explains 28.5% of the variations are in normative commitment value.  Table 6 provides a more detailed explanation about the result of hypotheses testing. The interpretation is straightforward. The first hypothesis states that ethical leadership negatively affects public sector employees' engagement with legal fraud. The hypothesis testing shows that ethical leadership had a direct negative effect on employee's legal fraud engagement (EL  LF; β = -0.275; p = 0.000). This suggests that the first stated hypothesis is accepted.
Except for the sixth hypothesis, the other hypotheses are also accepted. Intrinsic religiosity directly and negatively affects employee's legal fraud engagement, (IR  LF; β = -0.359; p = 0.000), indicating that the second hypothesis is accepted. Ethical leadership also positively affects affective commitment (EL  AC; β = 0.549; p = 0.000), while affective commitment negatively affects legal fraud (AC  LF; β = -0.586; p = 0.000). Thus, the third and fourth hypotheses are accepted. The positive effect of intrinsic religiosity to normative commitment (IR  NC; β = 0.540; p = 0.000) also suggests that the fifth hypothesis is accepted. Hypothesis six is rejected because normative commitment does not affect legal fraud significantly (NC  LF; β = 0.087; p = 0.075). The result of path analysis for mediating effect can be seen in Table 7. The path analysis results suggest that the total effect of ethical leadership to legal fraud is -0,596. This effect is the combination of the direct effect (EL  LF; β = -0.275) and the indirect effect (EL  AC  LF; β = -0.321). It indicates that that affective commitment mediates the effect of ethical leadership to employee's legal fraud engagement. On the other hand, normative commitment does not mediate the relationship between intrinsic religiosity and legal fraud because normative commitment does not have any significant effect to legal fraud. The data analysis highlights the importance of ethical leadership and intrinsic religiosity in mitigating public sector employees' engagement on legal fraud. Ethical leaders would benefit their organizations because when employees perceive that their leaders are ethical, they tend to avoid acting fraudulently. This provides insights that those leaders in government institutions hold a strategic role in anti-fraud working behavior in government institutions. For example, to increase employees' ethical behavior, the leaders should become examples for their employees and communicate their values clearly. An ethical leader who acts fairly and cares for their subordinates should make their subordinates feel obligated to their organization. The obligatory feeling may lead to employees' beneficial behavior to the organization. This finding extends previous studies about the positive effect of ethical leadership to organization citizenship behavior (Yang & Wei, 2018;Wang & Sung, 2016), employee's ethical behavior (Lu & Lin, 2013), and employees' intention to report fraud (Zhang & Liao, 2016;Wen & Chen, 2016).
The mediating effect of affective commitment also supports the importance of ethical leadership to reduce fraud occurrence in government institutions. The results suggest that ethical leaders affect on the level of their employees' fraud engagement, in part, because they improve the commitment of their employees. Thus, it can be inferred that by improving their employees' affective commitment, the employees are even more motivated to act ethically and to avoid involving in any act of legal fraud. Their strong emotional attachment to their government institutions motivates them not to involve in any activities that may harm the institutions. Thus, leaders need to share their organizations commitment values to their subordinates. By sharing their values, ethical leaders seem to build a committed working environment. This finding provides further evidence that affective commitment does not only improve public sector employees' willingness to report any wrongdoing (Kim, 2012), but it will also influence employees to avoid conducting fraud.
Another factor that may mitigate public sector employees' engagement to fraud is employees' intrinsic religiosity, which is consistent with previous studies, e.g. Said et al. (2018). The higher is the value of religiosity internalized in a government employee, the lower the likelihood they will involve in misuse of organization assets and confidential information and corruption. The employees are likely to work ethically and follow procedures and standards. This result is not surprising because religiosity has been known to have an influence on human behavior and ethical beliefs (Vitell, Paolillo, & Singh, 2006;Weaver & Agle, 2002). Besides, religiosity plays prominent role in the formation of beliefs, values, and culture in Indonesian society. Indonesian people tend uphold and preserve religious values. Even several local regulations in Indonesia are based on religious values, for example, the implementation of Islamic law in several local governments in Indonesia.
Interestingly, normative commitment does not mediate the effect of religious commitment to employees' involvement in the fraud. The connection between intrinsic religiosity and normative commitment is not sufficient to mitigate legal fraud occurrence. Although this result contradicted the study done by Shafer et al, (2016), there is a plausible explanation. Public sector employees' normative commitment in Indonesia may have a strong correlation with their continuance commitment. When employees' normative commitment is highly correlated with their continuance commitment, they may regard that it is wrong to deviate from their social group or society's values or practice (Becker, 1960). This kind of relationship between normative commitment and continuance commitment is often found in countries with high value of collectivism and power distance (Fischer & Mansell, 2009), including Indonesia (Hofstede Insights, 2020). The working environment of public sector in Indonesia seems to depend on hierarchy, centralized power, and employees may conform their beliefs to society or the organization they work at.
The potential problem with having employees who are willing to conform their ideal into the beliefs of the organization they work at is when fraud practice has been deeply internalized to the organization. Public sector employees may not have any other options but following the fraud acts. An empirical study suggests that legal fraud may have been institutionalized in public sector in Indonesia (Prabowo & Cooper, 2016). When the corruption acts are integrated in the operations of government institutions, corruption practice is systematically developed and becomes normal acts in the organizations. As a result, in a collectivism and power distance culture, public sector employees may find a rationalization to act fraudulently. It will be incorrect for them not to obey their leaders and not to join their colleagues. Only people with strong principles, such as having high religious values may confront fraud acts. If their confrontation is not successful, they may leave the government institution for good.

Conclusions
The study found that only affective commitment mediates the effect of ethical leadership to legal fraud engagement. Employees' affective commitment improves the effect of their ethical leadership in reducing legal fraud occurrence in government institutions. This result implies that ethical leaders who highly value fairness, integrity, trustworthiness, care about the well-being of their subordinates, and are more likely to improve their relationships with their followers. In return, the subordinates are likely to show loyalty and commitment to their leaders and organizations. Government employees with strong affective commitment are very likely to attach to their organization and avoid acting fraudulently. They are more likely to work for the best interest of their organization instead of their personal benefit.
Meanwhile, the normative commitment does not mediate the effect of intrinsic religiosity to public sector employees' legal fraud engagement. One possible reason is that the public sector employees may tend to perceive that it is incorrect not to follow common practice performed by their colleagues and leaders. This is because Indonesia's cultural dimension tends to be collectivism and high-power distance. Thus, when corruption practice is internalized into the operations of government organizations, public sector employees may perceive that it is incorrect not to act fraudulently as shown by their working environment. Having said that, this reason still needs further evidence. It opens possibilities for future research and discussion.
This study contributes to academic literature and public sector organizations. It extends prior literature about the role of affective commitment and ethical leadership in mitigating fraud engagement in Indonesia's government institutions. It also contributes to government institutions by providing empirical evidence about how important it is for public institutions led by leaders who highly value ethics, since they may mitigate fraud occurrence in public directly or through the improvement of their employees' commitment.
Leaders in public sector organizations must act carefully by considering ethical and moral issues. Their action will affect their subordinates' commitment to organizations. When fraud acts are already internalized in government organizations, appointing ethical leaders is likely to be a crucial step to recreate an ethical working environment. Ethical leaders will, directly and indirectly, influence their subordinates to be more reluctant to engage in legal fraud. In doing so, leaders need to maintain communication with their subordinates all the time. Government institutions may also improve their programs or training to produce leaders with high ethical values and good interpersonal skills.
This study has several limitations. First, the sample is homogeneous since it consists of only employees who work in government institutions. Generalizing this finding to private and for-profit sectors must be done carefully. Second, the number of respondents in this study is relatively small and they are only distributed to a few institutions. Larger respondents and more various institutions are likely to provide better insights into the model applied in this study. Future studies may also consider cultural context and organizational settings to provide more insights on how ethical leaders may affect legal fraud in the public sector.