The Effect of Ethical Leadership on Work Engagement and Workaholism: Examining Self-Efﬁcacy as a Moderator

: This study aims to investigate how ethical leadership can inﬂuence work engagement and workaholism through the potential moderating effect of self-efﬁcacy. There have been debates on the similarities, their negative correlation, and differences between these two work outcomes. To show one new aspect of evidence regarding the debate, we chose ethical leadership as the common antecedent of the outcomes and analyzed the relationships while considering a boundary condition, self-efﬁcacy. For this purpose, using an online questionnaire, we collected primary data from 80 graduate students from a university in Indonesia. An experimental research design was applied, and we used t-test and hierarchical regression analysis to conﬁrm the relationship mentioned above. Results indicate that ethical leadership has a positive effect on work engagement, while it has an insigniﬁcant effect on workaholism. Moreover, self-efﬁcacy did not moderate the relationships between ethical leadership and work engagement, or ethical leadership and workaholism. One novelty of the present study is the ﬁnding of different consequences of the two “similar” work outcomes from ethical leadership. Implications, limitations, and direction for future research are also discussed.


Introduction
The current working world has changed significantly due to global competitiveness and rapid innovation, both of which lead employees to work harder than they ever have before (van Beek et al. 2012). In addition to these working conditions, there are two types of hard work, also known as heavy work investment: work engagement and workaholism (e.g., Bakker et al. 2014;Shimazu et al. 2015;van Beek et al. 2012). In general, work engagement is associated with positive work outcomes, while workaholism is associated with negative outcomes (e.g., Di Stefano and Gaudiino 2019; Shimazu et al. 2015;van Beek et al. 2012). However, the meanings of the two terms also overlap (Di Stefano and Gaudiino 2019).
Prior studies have discussed the similarities, differences, and relations between work engagement and workaholism (Di Stefano and Gaudiino 2019). However, our understanding of both concepts is currently limited; hence, they need to be investigated further. Snir and Harpaz (2012) argued that, from the perspective of heavy work investment, work engagement and workaholism are two sides of the same coin. Therefore, based on prior research, this study investigates these two concepts by considering the similarities between them.
Examining the factors that support work engagement is essential for organizational sustainability (Bush and Balven 2018;Den Hartog and Belschak 2012;Engelbrecht et al. 2017;Tims et al. 2011). Although previous studies have investigated the antecedents of work engagement and workaholism, they have mainly focused on their relationship to working conditions; more specifically, these studies have relied on the job demands and resources (JD-R) theory (Molino et al. 2016;Saks and Gruman 2014). Therefore, the present study aims to investigate the determinants of work engagement and workaholism by expanding the JD-R theory to include leadership and personal resources (Macey and Schneider 2008;Saks and Gruman 2014). The idea to expand the theory was based on the indication that behavior, attitudes, and well-being are influenced by a combination of individual characteristics and the environment (Guglielmi et al. 2012), and the fact that recent research has empirically identified leadership and personal resources as the determinants of engagement (Saks and Gruman 2014). As work engagement and workaholism are similar in terms of heavy work investment, we expect that the antecedents of these two concepts might be similar, even though they might operate in different directions due to the potential overlap between the concepts. Furthermore, recent trends in the scientific debates about work engagement and workaholism have recommended investigating the similarities and differences between these two concepts, including their determinants.
Both academics and practitioners have examined how leadership affects management (Dadhich and Bhal 2008). This is mainly because leadership plays a significant role in the organization's success or failure, while engagement plays a significant role in the organization's success and competitiveness (Bennis 2007). Among others, this study examines ethical leadership, which has recently received more attention from scholars due to the strong effect ethical leadership has on its followers (Den Hartog and Belschak 2012). According to Brown et al. (2005, p. 120), ethical leadership is "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." Measurements of ethical leaders' actions through surveys reveal that ethical leadership positively influences employees (Den Hartog and Belschak 2012;Engelbrecht et al. 2017;Naeem et al. 2020). Brown and Treviño (2006) introduced numerous scandals in many types of unethical organizations including Enron, government, sports, and even religious organizations and emphasized the importance of ethical dimensions of leadership. They also argued that ethical leadership motivates and influences employees through ethical behavior and attitudes. More specifically, Nassif et al. (2020) argued that transparency of reward and punishment and communication of ethical standards and expectations are the source of positive influence of ethical leadership. Nejati et al. (2019) stressed that ethical leaders communicate that ethical and responsible behavior is essential and they lead by example to affect their followers positively. The theory of self-determination argues that even though work engagement and workaholism are different constructs, they have overlapping features that scholars have yet to clearly recognize (Di Stefano and Gaudiino 2019). Therefore, it is fruitful to investigate ethical leadership as a possible determinant of work engagement and workaholism.
Furthermore, personal resources, defined as "aspects of the self that are generally linked to resiliency and individuals' sense of their ability to control and impact upon their environment successfully" (Xanthopoulou et al. 2007, p. 124), have recently been included in the JD-R model (Janssen et al. 2020;Saks and Gruman 2014) as an expansion of the theory. Molino et al. (2016) suggested that future research into workaholism should explore the role of personal resources as its antecedent. However, the potential positioning of personal resources varies. For example, Saks and Gruman (2014) proposed that personal resources mediate the relationship between leadership styles and engagement, whereas Salmela-Aro and Upadyaya (2018) proposed that personal resources have a positive and direct influence on work engagement and a negative influence on job burnout (which is related to workaholism). In another study, Mérida-López et al. (2020) argued that self-efficacy, as a type of personal resource, may act as a moderator. Therefore, this study uses self-efficacy as a moderator because we expect that individuals' internal factors have both positive and negative moderating influences on the external factor's (leadership) effect on both work engagement and workaholism.
By examining the influence of ethical leadership on both work engagement and workaholism, and by examining the moderating effect of personal resources on those relationships, we expect this study will contribute to the literature. To date, there have been debates on the similarities, their negative correlation, and the differences between these two work outcomes. We expect to be able to show one new aspect of evidence regarding the debate by choosing ethical leadership as the common antecedent of the outcomes and analyzing the relationships while considering the boundary condition, self-efficacy. Shimazu and Schaufeli (2009) argued that it is important to investigate work engagement and workaholism empirically because in addition to the differences, there also exist similarities between the two concepts. Shimazu et al. (2012) also stated that there is a positive relationship between these two concepts, even though it is weak.

Work Engagement and Workaholism
Related to the definition, Schaufeli et al. (2002) developed the most frequently used and widely validated definition of work engagement (e.g., Demerouti et al. 2001;Salanova et al. 2005). They defined engagement at work as "a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption" (p. 172). Here, the element of vigor refers to the experience of high levels of energy and mental resilience while working, the willingness to invest effort in work, and persistence even in the face of difficulties. Dedication refers to the experience of a sense of significance, meaning, inspiration, pride, enthusiasm, and challenge. Finally, absorption is the experience of feeling a sense of concentration and engrossment in one's work (e.g., Bakker et al. 2007).
Oates (1971) defined workaholism as a phenomenon of working beyond reasonable expectations, which leads to an addiction to work, while Schaufeli et al. (2009, p. 322) defined it as "the tendency to work excessively hard and being obsessed with work, which manifests itself in working compulsively." In another study, Scottl et al. (1997) argued that there are three characteristics of workaholism. First, when presented with the opportunity, workaholics will spend many hours working. Second, outside working hours, workaholics continue to think about their work; thus, they are unwilling to disengage from their work. Third, workaholics work far beyond the reasonable requirements of their organization. Schaufeli et al. (2006b), however, argued that there are two components of workaholism: first, working excessively (i.e., working beyond the limits of time and energy), and second, working compulsively (working with a high and irresistible inner drive).
Comparing work engagement and workaholism as types of heavy work investment, referring to the study of Schaufeli et al. (2006b), Di Stefano and Gaudiino (2019) argued that there is a potential correlation between working excessively (work with exceeding time and energy) and absorption (concentrating and engrossed in one's work) and between working compulsively (working with a high and irresistible inner drive) and absorption.

The Effect of Ethical Leadership
Researchers have paid less attention to how leaders foster work engagement . Some researchers have predicted that leaders may influence their followers' engagement by spreading their positive attitudes or working behaviors. Christian et al. (2011) argued that some types of leadership, such as transformational and transactional (leader-member exchange), have been found to be positively related to engagement. Furthermore, past literature shows that ethical leadership improves employee well-being (Chughtai et al. 2015;Donaldson-Feilder et al. 2013;Kuoppala et al. 2008).
Some studies have suggested that ethical leadership is a combination of transformational and transactional leadership styles (Dadhich and Bhal 2008). According to Treviño et al. (2003), ethical leadership uses both transformational and transactional leadership styles to influence standard-setting processes, performance appraisals, and rewards and punishments to hold followers accountable to ethical conduct. Therefore, understanding the importance of ethical leadership's effect on individual well-being, we expect that ethical leadership may foster work engagement or even reduce workaholism in the daily working environment. This is because ethical leaders use their authority and responsibility to empower and provide their followers with more freedom to initiate work (Den Hartog and Belschak 2012;Engelbrecht et al. 2017;Macey et al. 2011).
Thus, we suggest that (at least part of) ethical leaders' impact on followers comes from their enhancing this positive motivational state of engagement (especially vigor and dedication). Mauno et al. (2007) noted that vigor has conceptual similarity with work motivation, whereas dedication is more closely related to job involvement. These elements of work engagement and workaholism seem especially relevant to the ethical leadership process. Based on these arguments, we expect that ethical leadership might influence work engagement and workaholism in opposite ways. Therefore, our hypotheses are as follows: Hypotheses 1 (H1). Ethical leadership has a positive effect on work engagement.

Self-Efficacy as a Personal Resource
According to Hobfoll et al. (2003), personal resources affect individuals' ability to control their environment. Some studies describe personal resources as mainly consisting of four elements: self-efficacy, optimism, hope, and resiliency. These four types of personal resources are also known as psychological capital (PsyCap). For the reasons explained below, this study measured only self-efficacy, even though the integrated PsyCap of all four elements has a greater impact than the sum of the effects of individual factors (Luthans et al. 2007).
This study examines the moderating effect of only self-efficacy for the following reasons. Xanthopoulou et al. (2009) found that three types of personal resources (selfefficacy, organization-based self-esteem, and optimism) are predictors of engagement and exhaustion. According to Guglielmi et al. (2012), self-efficacy is the most important personal resource as it acts as a self-motivating mechanism and supports motivation by influencing the challenges people pursue. Individuals with high self-efficacy can face a demanding working situation by building more resources (Hobfoll 2001). Furthermore, according to Salanova et al. (2005), the relationship between self-efficacy and work engagement has received less research attention. Therefore, to increase survey engagement, we decided to reduce the number of questions by including only self-efficacy as the measured individual resource. Furthermore, we included only self-efficacy in our research design due to its significance among other types of personal resources and its empirically demonstrated relationship with both work engagement and workaholism.
We aim to further investigate the potential moderation effect of self-efficacy on ethical leadership's respective relationships with work engagement and workaholism, both categorized as heavy work investment, as personal resources affect individuals' ability to control their environment. When individuals have more control over their working environment, we expect that it will affect their attitude toward daily working life; that is, they will either engage in their work or become workaholics, as both concepts are internally driven. Other studies argued that self-efficacy influences the improvement of individuals' perceptions of work in the social context, which is closely related to engagement at the workplace (Consiglio et al. 2016;Zuo et al. 2021). The positive effect of ethical leadership on work engagement might be stronger when individuals have higher self-efficacy because self-efficacy acts as a self-motivation mechanism. We expect that when ethical leaders give their followers adequate motivational power and inspiration, employees are more engaged, and rates of workaholism are lower; we also believe this applies to both those with high self-efficacy and those with low self-efficacy. As self-efficacy is defined as "beliefs in one's capabilities to organize and execute the course of action required to produce given attainments" (Bandura 1997, p. 3), ethical leaders should support their followers by providing them with an example of how to behave, especially in terms of ethical behavior. Therefore, we expect that self-efficacy might strengthen or weaken ethical leadership's effect on both work engagement and workaholism. Hence, we hypothesize the following: Hypotheses 3 (H3). Self-efficacy has a positive moderating effect on the relationship between ethical leadership and work engagement.
Hypotheses 4 (H4). Self-efficacy has a positive moderating effect on the relationship between ethical leadership and workaholism. Figure 1 presents our research framework, based on the hypotheses developed above. We investigate the effect of ethical leadership on the two similar but distinct outcomes, work engagement and workaholism, and the moderating effect of self-efficacy as an individual resource on the main relationship.
Adm. Sci. 2021, 10, 1 5 of 12 given attainments" (Bandura 1997, p. 3), ethical leaders should support their followers by providing them with an example of how to behave, especially in terms of ethical behavior. Therefore, we expect that self-efficacy might strengthen or weaken ethical leadership's effect on both work engagement and workaholism. Hence, we hypothesize the following:

Hypotheses 3 (H3). Self-efficacy has a positive moderating effect on the relationship between ethical leadership and work engagement.
Hypotheses 4 (H4). Self-efficacy has a positive moderating effect on the relationship between ethical leadership and workaholism. Figure 1 presents our research framework, based on the hypotheses developed above. We investigate the effect of ethical leadership on the two similar but distinct outcomes, work engagement and workaholism, and the moderating effect of self-efficacy as an individual resource on the main relationship.

Samples and Procedures
This study applied an experimental research design. We distributed an online questionnaire to graduate students at the Faculty of Psychology, the University of Indonesia. We expected graduate students to have better survey engagement because they have had and/or will have similar experiences of conducting research. Participants were required to have previous work experience and participate voluntarily. We distributed the questionnaire to the graduate students using the snowballing technique after procuring research ethics approval from the faculty. The survey questionnaire consisted of a hypothetical vignette for manipulating the respondent's ethical leadership experience, the measurement items for work engagement, workaholism, and self-efficacy, and demographic characteristics. We developed scenarios using two vignettes; the first vignette was related to ethical leadership, while the second was related to non-ethical leadership (please refer to Appendix A). We assigned vignettes to the participants randomly, using the last two digits of their student IDs; therefore, each participant read only one scenario. Self-efficacy questions based on respondents' individual conditions and demographic characteristics were asked before the vignettes; ethical leadership questions for the manipulation check and questions about work engagement and workaholism were asked based on respondents' understanding of the vignette presented to them. After 1 week of data collection in mid-July 2020, 82 respondents had completed the survey, exceeding our data analysis requirements. After data cleaning, 80 questionnaires were retained. Finally, we used SPSS 26 to conduct the data analysis.

Measurements
This study used measurements that have been well validated in prior studies related to ethical leadership, work engagement, workaholism, and personal resources. Ethical

Samples and Procedures
This study applied an experimental research design. We distributed an online questionnaire to graduate students at the Faculty of Psychology, the University of Indonesia. We expected graduate students to have better survey engagement because they have had and/or will have similar experiences of conducting research. Participants were required to have previous work experience and participate voluntarily. We distributed the questionnaire to the graduate students using the snowballing technique after procuring research ethics approval from the faculty. The survey questionnaire consisted of a hypothetical vignette for manipulating the respondent's ethical leadership experience, the measurement items for work engagement, workaholism, and self-efficacy, and demographic characteristics. We developed scenarios using two vignettes; the first vignette was related to ethical leadership, while the second was related to non-ethical leadership (please refer to Appendix A). We assigned vignettes to the participants randomly, using the last two digits of their student IDs; therefore, each participant read only one scenario. Self-efficacy questions based on respondents' individual conditions and demographic characteristics were asked before the vignettes; ethical leadership questions for the manipulation check and questions about work engagement and workaholism were asked based on respondents' understanding of the vignette presented to them. After 1 week of data collection in mid-July 2020, 82 respondents had completed the survey, exceeding our data analysis requirements. After data cleaning, 80 questionnaires were retained. Finally, we used SPSS 26 to conduct the data analysis.

Measurements
This study used measurements that have been well validated in prior studies related to ethical leadership, work engagement, workaholism, and personal resources. Ethical leadership (used for manipulation check) was measured on a 5-point Likert-type scale, consisting of 10 questions developed by Brown et al. (2005); a sample measurement item is as follows: "disciplines employees who violate ethical standards." Work engagement was measured (α = 0.923) on a 7-point Likert-type scale adopted from the Utrecht Work Engagement Scales-short version (Schaufeli et al. 2006a), which consists of nine questions.
This measurement tool covers all three subscales of employee engagement: vigor, dedication, and absorption. A sample question is as follows: "at my work, I feel bursting with energy." A 5-point Likert-type scale consisting of 10 questions, adopted from Schaufeli et al. (2008), also known as the Dutch Workaholism Scale, was used to measure workaholism (α = 0.757). A sample item is, "it is hard for me to relax when I am not working." To measure self-efficacy (α = 0.850), a 5-point Likert-type scale that consists of eight questions developed by Chen et al. (2001), was utilized. A sample item is, "I will be able to achieve most of the goals that I have set for myself."

Demographic Data
We cleaned the data to eliminate outliers and arrived at 80 samples. This exceeded the minimum data required for the next analysis. Table 1 shows respondents' demographic data.

Testing the Experiment Application
We conducted a manipulation check using a t-test to make sure that our experiment succeeded. Table 2 shows the mean (M) and standard deviation (SD) scores of perceived ethical leadership. There are significant differences in perceived ethical leadership between those who read the ethical leadership scenario (M = 40.383; SD = 6.572) and those who read the non-ethical leadership scenario (M = 27.061; SD = 12.604). The significant difference between the two groups (F = 40.570 and p < 0.05) indicates that our intervention was successful. Table 3 shows the correlation information of the items measured.

Testing the Hypotheses
We implemented a hierarchical regression analysis to test the hypotheses. As can be seen from Tables 4 and 5, the results suggest that ethical leadership has a positive effect on work engagement (p < 0.05). However, ethical leadership has an insignificant effect on workaholism. Furthermore, this study found that self-efficacy does not have a moderating effect on the relationship between ethical leadership and work engagement or ethical leadership and workaholism. Table 4. Result of hierarchical regression analysis for the effect of ethical leadership on work engagement with moderation effect of self-efficacy.
Step 1 Step 2 Step  Table 5. Result of hierarchical regression analysis for the effect of ethical leadership on workaholism with moderation effect of self-efficacy.
Step 1 Step 2 Step Because of the small sample size, we were concerned about whether random assignment was appropriately implemented. Therefore, we compared the demographic characteristics of the two groups. Three variables, including gender, marital status, and workplace experience, show insignificant results (t-test for gender and χ 2 tests for the others; all p > 0.05). As age was significantly different (t-test; p < 0.05), we analyzed the hypothesized relationships with controlling for age. The results were similar to the original analysis without controlling for age.

Discussion
This study examined the respective influence of ethical leadership on work engagement and workaholism and investigated whether self-efficacy moderates these relationships. Based on our analysis, ethical leadership has a positive significant effect on work engagement but has an insignificant effect on workaholism. This result implies that ethical leaders may create engaged employees but may not influence workaholics. We believe there are at least two reasons ethical leadership has an insignificant effect on workaholism. First, considering that workaholism consists of two features-working excessively and working compulsively (Schaufeli et al. 2006b)-ethical leaders supporting ethical conduct in their daily working environments might not ask their employees to work less excessively or compulsively; instead, they may simply promote ethical behavior. Another possibility is related to the cultural dimensions of Indonesia, the country in which this study was conducted. Here, we consider a concept argued by Hofstede et al. (2005)-collectivism versus individualism. A collectivist culture is a society that prefers to work as a team or collectively. In such a society, individual performance is not pursued above the group or team performance. Therefore, individuals in a collectivist society like Indonesia may feel no need to work beyond reasonable expectations to reach higher performance as an individual because working as a team may be more meaningful than individual achievement. However, in individualistic societies, we can apply the discussion that was made for developing the hypotheses, due to the lack of conditions found in the collectivist countries.
Here, we would like to argue that ethical managers are likely to act as role models for accomplishing tasks based on ethical standard applied in their organizations, by explaining or showing, for example, how to accomplish work or tasks on an effective time schedule and based on a standard operating procedure. Similarly, Lo Presti et al. (2020) studied workaholism in an individualistic country, Canada, and suggested removing any work aspects that could potentially foster addictive work behavior, by constantly monitoring within organizations. Here, we expect that these tasks may be potentially accomplished by ethical managers who are expected to prevent their subordinates from working compulsively and excessively.
Furthermore, this study finds that self-efficacy does not moderate the relationship between ethical leadership and work engagement, or ethical leadership and workaholism. This result supports the theory that work engagement and workaholism are relatively stable internal factors that cannot be easily influenced by individual resources. Therefore, in addition to finding support for the similarities suggested by prior research, this study also finds that these two concepts are different from the point of view of antecedents. Another reason for the insignificant moderation effect might be related to the single personal resource used in our analysis (i.e., self-efficacy). Therefore, incorporating other personal resources (e.g., optimism, hope, and/or resiliency) might enhance the discussion.
Based on the results of this study, as a practical implication, company managers need to act more ethically to further engage their employees. Ethical leadership implies that leaders need to create, support, and become role models of ethical behavior in the daily working environment. The results also suggest that managers who promote ethical behavior have no influence on the workaholic behavior of their employees. The implication is that the managers do not have to be very concerned about the consequences of workaholism. However, they should explore reasons other than ethical leadership, when an unacceptable level of workaholism is obseved in the workplace. We propose that when promoting ethical behavior to create engaged employees, leaders need to consider the cultural dimensions of their workplace and adjust their plan accordingly. With regard to the insignificant moderation of self-efficacy, the results imply that ethical leaders may be able to perform their jobs without focusing on subordinates' self-efficacy, at least for the purpose of improving or not damaging their work engagement and workaholism.
This study's main limitations are twofold. First, it utilized a small sample, which might limit the discussion even though the model's explanatory and out-of-sample predictive powers are at acceptable levels. Second, this study examined the effect of only one personal resource-self-efficacy-to reduce the total number of questions and increase the engagement of survey respondents. Thus, future research should explore all types of individual resources categorized as psychological capital. The results of this study suggest that there is need for further research into the antecedents of workaholism in collectivistic countries like Indonesia. Another suggestion for further research is to possibly utilize structured equation modeling for investigating the mediated-moderation analysis of personal resources including self-efficacy, as there has been empirical evidence on the mediation of personal resources between leadership and work engagement (Shamshad and Khan 2020).

Conclusions
The results of this study show that ethical leadership helps support the work engagement of employees, as expected, but, contrary to our expectations, is not associated with workaholism. Managers should behave ethically to improve the work engagement. Their leadership style may not relate to the consequences of workaholism, but they should explore other possible reasons when an unacceptable level of workaholism is obseved in the workplace.
Furthermore, self-efficacy alone has no moderating effect on ethical leadership's respective relationships with work engagement and workaholism. The results imply that ethical leaders may be able to perform their duties without focusing much on the selfefficacy of subordinates, at least for the purpose of improving or not damaging their work engagement and workaholism.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are openly available in Mendeley at http://dx.doi.org/10.17632/8f9jctrjpd.1.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
The following vignette/scenario was used for the experiment. We assigned the participants randomly by using the last two digits of their student ID number; therefore, each participant only read one vignette/scenario. Vignette 1 (Unethical leadership) Imagine you are working at a company, under a boss or supervisor. Your boss or supervisor never considers your voice and he or she is not trustworthy. When he/she makes decisions, he/she seldom asks you "what is the right thing to do," and makes unfair and unbalanced decisions. Neither does he/she discuss business ethics or values with you. To work ethically, he/she does not set an example of how to do things the right way because he/she does not conduct his/her personal life in an ethical manner. Consequently, he/she is not capable of disciplining employees who violate work ethics. Furthermore, he/she defines success simply by results, not the way they are obtained. Therefore, he/she does not have the best interests of the employees in mind.
Vignette 2 (Ethical leadership) Imagine you are working at a company, under a boss or supervisor.
Your boss or supervisor considers your voice and he or she is trustworthy. When he/she makes decisions, he/she usually asks you "what is the right thing to do," and makes his/her decisions in a fair and balanced way. He/she also discusses business ethics or values with you. To work ethically, he/she sets an example of how to do things the right way and he/she proves it by conducting his/her personal life in an ethical manner. Consequently, he/she disciplines employees who violate work ethics. Furthermore, he/she defines success not just by results, but also by the way they are obtained. Therefore, he/she has the best interests of the employees in mind.