The Effect of Stretch Goals on Destructive Leadership and Counterproductive Work Behavior in Indonesian State-Owned- Enterprises

DOI: 10.22146/jlo.70140 Introduction/Main Objectives: This study aims to examine the effect of stretch goals on destructive leadership with burnout as the mediating variable and then the effect of destructive leadership on counterproductive work behavior of employees with psychological capital as a moderating variable. Background Problems: The phenomenon of irregularities that occur in SOE in Indonesia is interesting to study. Deviations committed by SOE leaders in Indonesia include fraud, gratification, and data manipulation. The increase in the number of irregularities has a negative effect on organizational performance because it causes several counterproductive work behaviors in employees. Novelty: Empirical research on destructive leadership is still rare because previous research has focused only on the conceptual side. Research Methods: The design of this study used a survey with a questionnaire completed by 724 respondents who were leaders, and employees. The hypothesis testing used Structural Equation Modeling (SEM). Finding/Results: The findings of this study show a positive influence of stretch goals on burnout and a positive influence of stretch goals on destructive leadership, but burnout has no mediating role on the effect of stretch goals on destructive leadership. There is no effect of perceived destructive leadership on employees’ counterproductive work behavior, but psychological capital has a moderating role on the effect of perceived destructive leadership on employees’ counterproductive work behavior. Conclusion: The practical implication in this study is that stretch goals that are not balanced with resources can cause individuals to behave destructively even though they are at a managerial level.


Introduction/Main Objectives:
This study aims to examine the effect of stretch goals on destructive leadership with burnout as the mediating variable and then the effect of destructive leadership on counterproductive work behavior of employees with psychological capital as a moderating variable. Background Problems: The phenomenon of irregularities that occur in SOE in Indonesia is interesting to study. Deviations committed by SOE leaders in Indonesia include fraud, gratification, and data manipulation. The increase in the number of irregularities has a negative effect on organizational performance because it causes several counterproductive work behaviors in employees. Novelty: Empirical research on destructive leadership is still rare because previous research has focused only on the conceptual side. Research Methods: The design of this study used a survey with a questionnaire completed by 724 respondents who were leaders, and employees. The hypothesis testing used Structural Equation Modeling (SEM). Finding/Results: The findings of this study show a positive influence of stretch goals on burnout and a positive influence of stretch goals on destructive leadership, but burnout has no mediating role on the effect of stretch goals on destructive leadership. There is no effect of perceived destructive leadership on employees' counterproductive work behavior, but psychological capital has a moderating role on the effect of perceived destructive leadership on employees' counterproductive work behavior. Conclusion: The practical implication in this study is that stretch goals that are not balanced with resources can cause individuals to behave destructively even though they are at a managerial level.

Introduction
A number of studies suggest that stretch goals affect a leader's destructive behavior (Barsky, 2008;Sitkin et al., 2011;Bazerman & Tenbrunsel, 2011;Schweitzer et al., 2004). Stretch goals that are perceived as new and difficult have an unavoidable negative effect (Sitkin et al., 2011;Lemoine et al., 2016;Cunha et al., 2016;Cheng et al., 2007). Stretch goals concern many researchers because previous studies have found that stretch goals can cause destructive behavior. However, this research is still only conceptual. Empirical testing of the effect of stretch goals on destructive leadership has not been carried out much.
Although goal setting can clarify tasks and responsibilities related to a particular job or workgroup  and goal setting can focus on individual goals and strengthen individual contributions and work outcomes, this study shows the negative effect. In some cases, highperformance goals can lead to undesirable results. For example, goals have been associated with an increase in destructive leadership behaviors, especially when one has to meet difficult performance targets (Schweitzer et al., 2004). If the leader has "demands" from the organization to achieve stretch goals, it leads to destructive leadership. When a person feels vulnerable to resource depletion and even loss of resources, he or she feels a certain amount of pressure and tries to maintain and protect those depleted resources (Hobfoll, 1989(Hobfoll, , 1998(Hobfoll, , 2001. Individuals who experience the occurrence of resource depletion then struggle to get back their resources (Hobfoll, 2001;Bandura, 1991).
In addition to the direct relationship between stretch goals and destructive leadership, according to conservation of resources (COR) theory, burnout can be chosen as a mediator. One of the main dimensions of burnout, namely emotional exhaustion, is characterized by reduced energy and employees feeling that their emotional resources are depleted. In accordance with what was conveyed by Urban et al. (2012), emotional exhaustion refers to a feeling that there has been an excessive depleting of one's emotional and physical resources after someone receives an assignment in the form of stretch goals that are felt to be heavy (Urban et al., 2012). According to resource conservation (COR theory (Hobfoll, 1989, 1998, 2001, when people have stretch goals which are ambitiously designed with radical results but beyond the current capacity and output, then they have the potential to experience feelings of frustration and hopelessness. Furthermore, the findings of the study by Schweitzer, Ordonez, and Douma (2004) indicate that someone with unfulfilled goals is more likely to engage in destructive behavior.
Deviant behavior by employees that occurs in business and government organizations is closely related to the leadership style of their leaders (Sari, 2020). The study of  stated that destructive leadership triggers unproductive work behaviors such as decreased performance, increased absenteeism, and other things resulting in organizational goals not being achieved. The character and bad intentions of leaders associated with this destructive leadership can lead to such counterproductive work behavior. This is because these goals are an important component of destructiveness, so various unpleasant actions by superiors can ultimately lead to negative reactions from subordinates which can lead to counterproductive work behavior (Bass & Steidlmeier, 1999). This is evidence that leaders who behave destructively have a negative effect, both on their subordinates (Bies & Tripp, 1998;Tepper, 2000Tepper, , 2007, the organization they work for, or both (Kellerman, 2004;Vredenburgh & Brender, 1998). According to Mullins (2015), almost 33% of employees experience destructive leadership (Aasland et al., 2010). The findings in their study show that 83.7% of employees experienced at least some destructive behavior by their superiors, and 33.5% reported experiencing destructive behavior more frequently (Aasland et al., 2010). Furthermore, the destructive behavior of leaders that has been studied by Tepper et al. (2006) estimates that organizations in the USA spend nearly 24 billion dollars annually as a consequence.
However, this negative effect can be reduced by the ability of the individual to be led. According to Liang Guo et al. (2018), the effect of the leader's character on employees is also influenced by those employees' abilities and/or their psychological capital. When an employee has high psychological capital, they tend to resist the negative effects of a destructive leader. Therefore, in this study, psychological capital is positioned as a moderator of the destructive influence of leadership on counterproductive work behavior. Based on the background presented, the researcher wants to examine the effect of stretch goals on destructive leadership and the mediating role of burnout on the effect of stretch goals on destructive leadership. Furthermore, the researcher has also tested the effect of psychological capital on destructive leadership and the role of psychological capital as a moderator on the effect of destructive leadership on counterproductive work behavior.
This research contributes theoretically in several ways. First, although the theory of goal setting states that, in general, goal setting motivates a person to achieve goals well and, in the end, the organization benefits from achieving goals, the results of this study show that there is a negative effect. If the leader has "demands" from the organization to achieve stretch goals, it leads to destructive leadership. Second, the mediating role in the form of burnout contributes to COR theory. This theory can be used to explain the emergence of destructive leadership better. Burnout and stress are two concepts with similarities; stress occurs when an imbalance exists between demands from the environment and individual resources. Meanwhile, burnout occurs due to a continuous process of adaptation to disturbances arising from longterm imbalances (Cooper et al., 1996). Stress can have both positive and negative effects. Stress can still be modified into a positive impetus for change for the better. Meanwhile, burnout will only have a negative effect (Farber, 1991).
In addition, this study contributes to the literature by establishing the role of individual factors, namely psychological capital, in moderating the relationship between destructive leadership and employee counterproductive behavior. The higher the psychological capital owned by employees, the weaker the influence of destructive leadership on employee counterproductive behavior. Contextually, this research was conducted on state-owned companies, namely government-owned companies that have a combination of characteristics of both public companies and private companies, so this study has a different feel from previous research conducted purely on private companies, which are profit-oriented.

Destructive Leadership
Destructive leadership is defined as systematic and repetitive behavior by leaders, supervisors, or managers that violates the legitimate interests of the organization, by overriding and/or sabotaging the goals, tasks, organizational resources, and effectiveness and/or motivation, well-being, or job satisfaction from their subordinates (Einarsen et al., 2007). Although, empirically, it is very difficult to distinguish different concepts regarding the dark side of leadership (Hershcovis, 2011;Tepper & Henle, 2011), this definition from Einarsen et al., (2007) is more comprehensive.
The form of destructive leadership that is labelled "social undermining" was presented by Duffy et al. (2002). The social construct of undermining is related to deviant behavior and aggressive behavior. This social undermining behavior is never directed at the organization as a whole. However, some social undermining behaviors include acts of aggression, namely working slowly and not protecting the welfare of the target.
According to Hershcovis (2011), there are several forms of mistreatment in the workplace in the form of social undermining, incivility, bullying, abusive supervision, and interpersonal conflict. These five examples of mistreatment are included in the construct of abuse, namely harsh and cruel supervision related to hostile behavior, both verbal and nonverbal that is continuous, but not including physical contact (Tepper, 2000). In their study, Krasikova et al. (2013) say that there are similarities between abusive supervision and petty tyranny constructs in terms of destructive actions, but they do not take any action against organizational goals.
Destructive actions related to the achievement of destructive goals are known as pseudo-transformational leadership and personalized charismatic leadership. Meanwhile, strategic bullying and managerial tyranny are included in destructive actions, but they are related to the achievement of goals constructively.
Destructive leadership combines forms of deviant behavior and organizational leadership. Thus, destructive leadership differs from (a) currently emerging constructive forms (e.g., transformational leadership) which do not involve harmful behavior, (b) counterproductive work behavior, aggression, and similar phenomena that do not involve others, (c) ineffective leadership behavior that is unintentionally harmful, and (d) poor job performance and accidents (for example, due to breaches of safety standards) that do not involve leading others and are unintentionally harmful.

Destructive Leadership
According to Mawritz et al. (2013), there are negative effects that have the potential to arise when there are difficult goals and low levels of acceptance, namely in the form of destructive behavior. Stretch goals are perceived as having a high level of difficulty and appear to be difficult to achieve, leading to greater effort and/or persistence of goals that are quite difficult or abstract (Sitkin et al., 2011;Lemoine et al., 2016;Cunha et al., 2016.;Cheng et al., 2007). According to clinical psychologists, goals that are too high can cause psychological stress (Beck, 1967;Ellis, 1962Ellis, , 1994Ellis, , 2002Hewitt & Flett, 1991, in Hrabluik, 2009) and by feeling psychological pressure, a person will be unable to do his job well. (Burns, 1980;Frost & Marten, 1990;Hollender, 1965in Hrabluik, 2009.
Stretch goals with high novelty certainly require more knowledge and skills, so this causes an increase in uncertainty. Carroll and Tosi (1970) found that goal clarity was correlated with increased effort. If the leader believes that stretch goals are impossible, then the perception and assumptions of possibility and impossibility are important in determining and operationalizing stretch goals. According to Thompson et al. (1997), if achieving a goal is considered a "mission impossible", it is thought to cause the individual to stop making efforts (Locke & Latham, 1984, 1990. Even though the individual is aware that risk-taking behavior can create future problems for the organization, individuals will try to find ways to justify destructive behavior. If individuals are unable to complete tasks and achieve goals, the possible consequences for them include being fired, reprimanded, or demoted. In line with what was stated by Palazzo et al. (2010), individuals tend to exhibit risktaking behavior when they face a number of uncertainties (Takeuchi et al., 2012). In conditions of uncertainty, individuals become unable to see the ethical dimensions and destructively make decisions where important things are at stake. Furthermore, under certain conditions, over time, sensemaking becomes more narrow and "rigid", tensions become less strong, and the problem of ethical values begins to fade (Tenbrunsel & Messick, 2004). According to Lewicki (1983), the relationship between goals and unethical behavior will be stronger when people perceive they have a short time to achieve goals.
The effect of stretch goals on destructive leadership can be explained by deception theory (Levine, 2014). The concept of deception is defined as a message consciously conveyed by the sender to help create false beliefs or conclusions in the recipient (Buller & Burgoon, 1996). Threats that occur to middle-level leaders due to pressures from upper management make them engage in various behaviors, one of which is committing irregularities such as fraud. In the context of this study, the threat of very high targets or goals can cause middle-level leaders to commit fraudulent actions. Individuals who have not succeeded in achieving stretch goals try to present information in the form of reports that are different from the truth. According to Levine (2014), when people have information that they consider too problematic to disclose, they commit fraud. This occurs because the leader perceives the contextual demands of "putting them in place" to reveal the unpleasant truth. In all of these cases, the driving force behind the behavior is a temporary and contextual consideration of the practical good that fits the information at hand. Based on the explanation above, the researcher proposes the following hypothesis: H1: Stretch goals have a positive effect on destructive leadership

The Mediating Role of Burnout on
The Positive Effect of Stretch Goals on

Destructive Leadership
Leaders who are assumed to be ineffective seek to protect resources that are perceived to be limited. They reduce their effort, resulting in lower quality performance outcomes (Halbesleben & Bowler, 2005). On the other hand, a leader who is assumed to be effective has a variety of resources, including cognitive abilities (Barling et al., 2000;Howell & Avolio, 1993;Ross & Offermann, 1997;Walter & Bruch, 2003). In conclusion, the occurrence of threats to the loss of resources can affect an individual's behavior. This is due to the extreme difficulty of stretch goals and it is not yet known how to achieve these goals using current capabilities (i.e., practice, skills, and organizational knowledge) (Hobfoll, 1989(Hobfoll, , 2001. Feelings of depletion or loss of resources can be used to understand the occurrence of stress and strain (Halbesleben & Buckley, 2004, & Hobfoll, 2001). An individual has a psychological response to a situation that exceeds the individual's capacity or resources (Lepine et al., 2004). The imbalance between demands and available resources, with low social support, autonomy, rewards, and boredom, has the potential to cause stress if it lasts for a long time, resulting in burnout. Symptoms of depression and anxiety at work as a form of resource depletion, will be predictors of destructive leadership (Byrne et al., 2014).
A high burnout rate indicates that workers have insufficient resources to handle the demands of their job effectively. Burnout reduces the capacity of employees to control the work environment, which causes disruption to work performance (Taris, 2006). Someone invests other resources to address resource depletion, so aggressive actions are taken to address resource depletion. At this stage, people who experience burnout have increased losses, and each loss can lead to depletion of the resources needed to deal with the next problem with the threat of loss. This shows that people strive to acquire, retain, protect, and regenerate valuable resources and minimize the threat of resource loss, which is consistent with COR theory. Leaders who are emotionally and physically exhausted and do not have the capacity to reinvest time and effort into stretch goals end up taking destructive actions (Sitkin et al., 2011;Lemoine, Blum, Roman, 2016;Cunha et al., 2016;Cheng, Luckett & Maham, 2007). Based on the explanation above, the researcher proposes the following hypothesis: H2: Burnout mediates the positive effect of stretch goals on destructive leadership

Counterproductive Work Behavior
When employees perceive that their leaders are taking destructive actions, it triggers them to engage in CWB (Ones et al., 2017). Counterproductive work behavior carried out by employees is divided into two orientations, namely interpersonal and organizational. This is conveyed by Robinson and Bennett (1995), who state that CWB has two dimensions or orientations, namely interpersonal CWB such as gossiping about co-workers and organizational CWB such as taking sabbaticals that exceed their rights. Workers who exhibit CWB tend to be more prone to stress and choose to resign from their workplace (O'Leary-Kelly, Griffin, & Glew, 1996).
According to Ones et al. (2017), the majority of research on destructive leadership and counterproductive employee work behavior assumes that destructive leadership has a negative effect on the work climate. It certainly encourages undesirable behavior by employees. Destructive leaders have authority, so they try to maintain their power. Employees who are subordinate to a destructive leader feel a lack of support from the organization, which results in CWB, such as anxiety, increased turnover, decreased job satisfaction, low organizational commitment, and organizational identification (Shoss et al., 2013).
This statement is in accordance with social exchange theory, which states that the value of a relationship affects the final outcome or outcome. A person is faced with a choice to continue a relationship or end it (such as leaving work). According to the utilitarian perspective on social exchange theory, one of the basic truths about life is that everyone is trying to avoid suffering and maximize pleasure. Humans will only act if they think they will get a reward. Actors are seen as having goals and choices at the same time. The actions taken by actors are always oriented toward the desire to achieve goals that are in accordance with the level of their choice. Based on the explanation above, the researcher proposes the following hypothesis: H3: Destructive leadership has a positive effect on counterproductive work behavior.

Leadership on Counterproductive Work
Behavior Several studies conducted by researchers in the field of leadership have tried to link psychological capital with leadership (e.g., Avey et al., 2011;Avey et al., 2008;Avolio et al., 2004;Avolio & Luthans, 2006;Clapp-Smith et al., 2009;Jensen & Luthans, 2006;Luthans & Avolio, 2003;Story et al., 2013;Youssef & Luthans, 2012). The study conducted by Bono and Ilies (2006) found that leaders have a positive effect on followers through positive expression. The researcher who researches in the field of psychological capital and leadership is Erkutlu (2014). Erkutlu (2014) examines the moderation of psychological capital on the influence of narcissism, one of the dimensions of destructive leadership, on psychological well-being. The test shows supported results.
The researcher uses self-concept theory to explain the moderating role of the effect of destructive leadership on employee counterproductive work behavior. Selfconcept is a concept about the individual himself which includes how a person perceives, thinks, and evaluates himself so that his actions are in accordance with his concept.
This self-perception is psychological, social, and physical. Selfconcept is a prediction that is prepared for oneself (Calhoun & Acocella, 1995). The researcher uses a proxy from the results of Mitchell's research (2008, in Brown & Mitchell, 2010, which found that the relationship between perceived destructive leader behavior and employee deviation or employee counterproductive work behavior weakened when employees showed a high level of internalization. So, when they have high psychological capital, employees are better prepared to adapt in uncertain circumstances and without the support of the organization H4: Psychological capital moderates the positive influence of destructive leadership on counterproductive work behavior.

Respondents and Procedures
This study uses a quantitative survey given to leaders and employees of 14 SOE companies in Indonesia. Responses were collected online using electronic devices. Leaders filled out 405 questionnaires. Eightythree of them could not be processed for this research because they did not meet the characteristics and criteria for the sample. Thus, the data that could be processed for this research were 322 questionnaires from leaders.
Meanwhile, 501 employee questionnaires were filled out. Ninety-nine of these could not be processed for research because they did not meet the characteristics and criteria of the sample. Thus, the data that could be processed for this research came from 402 employee questionnaires. The demographics of the respondents are presented in a table according to age, length of service, and gender.

Measures
Destructive Leadership (DL) is measured using the destructive leadership scale (DLS) developed by .
This questionnaire measures destructive leadership which includes 12 statement items. Responses to each statement were expressed on a 4-point Likert scale ranging from 1 (never) to 4 (very often or almost always).
Burnout (BO) is measured using the Maslach Burnout Inventory (MBI) from Maslach and Jackson (1986). The measurement of burnout in this study is divided into three parts; 10 statement items to measure emotional exhaustion, 6 statement items to measure depersonalization, and 6 statement items to measure achievement decline. Respondents were asked to measure their perceptions on a 6-point Likert scale from 1 (several times a year or less) to 6 (every day).
Psychological Capital (PC) is measured using 12 statement items developed by Fred Luthans et al. (2007). Responses to the psychological capital variable were expressed on a four-point Likert scale from 1 (never) to 4 (very often).
Counterproductive Work Behavior (CWB) is measured as follows: interpersonal counterproductive work behavior variable (CWBi) is measured using 7 statement items, while organizational (CWBo) is measured using 12 statement items developed by Spector et al. (2010). Responses to the counterproductive work behavior variable were expressed on a seven-point Likert scale.

Result and Discussion
The researcher conducted a validity test for the variables of destructive leadership, psychological capital, and counterproductive work behavior aimed at SOE employees. The loading factor value is greater than 0.5 except for 8 statement items. Statement items with a loading factor value below 0.5 are statement items LD2, LD5, LD9, and LD11 (destructive leadership variables) as described in the previous section. Other statement items are pertain to psychological capital statements, namely PC8 ("Reduce your opportunity to express opinions at meetings by giving less time to speak or by giving a chance to speak at the end") and PC9 ("Be friendly by encouraging you/colleagues to extend the lunch break"). The counterproductive work behavior variables whose loading factor value is below 0.5 are CWB1 ("Making fun of someone at work") and CWB2 ("Saying something hurtful to someone at work").
Meanwhile, the Cronbach alpha value for each research variable is greater than 0.6. Based on these results, it can be concluded that the questionnaire used in the study has met the criteria for reliability and can be used for collecting research samples as a whole.
This study also uses Harman's Single Factors Test to determine the bias in the structural model of the leadership group. By including all indicators in the exploratory factor analysis (EFA), the % variance value of the extraction sum of the squared loading stage is 18.872% (<50%). Thus, it can be concluded that there is no common method bias in the structural model of the leadership group. Meanwhile, the employee structural model obtained the value of percentage variance in the extraction sum of the squared loading stage of 26.460% (<50%). Thus, it can be concluded that there is no common method bias in the structural model of the employee group.

Structural Equation Modeling (SEM)
To test the six hypotheses, researchers used SEM which is good to use for research designed to confirm a research study design rather than exploring or explaining a phenomenon. In addition, SEM is a statistical tool that can be used to solve multilevel research models that cannot be solved by linear regression equations.
The steps of the SEM test carried out by researchers are as follows: a. Confirmatory factor analysis In the SEM analysis, the first step is to measure the items used in the structural model using confirmatory factor analysis. Confirmatory factor analysis evaluation tests convergent validity through standardized regression weight values > 0.5, discriminant validity testing through average variance extracted (AVE) > 0.5 and reliability construct testing with a value determination greater than 0.7. The results of the confirmatory factor analysis evaluation for the structural leadership model are shown in Table 1 as follows:  Table 2 shows that all items used in the employee structural model also have a standardized regression weight value greater than 0.50 so that these indicators have met convergent validity. Meanwhile, the average variance extracted (AVE) and construct reliability values in each variable also obtained values greater than 0.50 and 0.7, so the measurement of all variables in the employee structural model has met discriminant validity and construct reliability.
b. Univariate and multivariate outlier Evaluation of univariate outliers with Z Score and multivariate outliers through the Mahalanobis distance found the number of samples that were outliers to be 83 which were then excluded from the sample of leaders that was analyzed. The Z Score value is obtained from the reduction process, which is already in the range of -3 to +3 and the Mahalanobis distance is 58.970 < chisquare table 65.247, while the employee structural model obtained the number of outliers as many as 99 samples that must be reduced, so the Z Score value is obtained in accordance with the provisions and the Mahalanobis distance 61.741 < chi-square table 66,618.
c. Univariate and multivariate normality The results of the univariate evaluation with reduced data samples in each structural model obtained the value of cr skewness and kurtosis, which were already in the range of -2.58 to +2.58. As for the evaluation of multivariate normality, the multivariate CR value is 2.363 for the structural leadership model and 2.278 for the employee structural model, both of which are already smaller than 2.58. d. Goodness of fit It was initially concluded that the results of the estimation of the structural model of leadership and employees still did not meet the specified criteria, so the structural model had to be given modification indices to improve the goodness of fit. The following is a comparison of the goodness of fit evaluation of the structural research model. Table 3 shows the evaluation of the goodness of fit structural model of leadership and employees, and the initial result is that almost all of the criteria are still not in accordance with the provisions of the goodness of fit index. Furthermore, the results of the goodness of fit evaluation of the two structural models that had been improved with modification indices resulted in better goodness of fit index values where there was only one criterion, namely the pvalue of the chi-square model, which is still smaller than 5% and other goodness of fit index criteria are included in the marginal fit and fit index criteria. Referring to these results, the process of testing the causality of the paths that have been developed in the two structural models uses a structural model that has been carried out by a model modification process. A path is declared to have a significant effect if the critical ratio value is greater than 2 or if it produces a p-value less than 5%. The following are the results of the estimation of the structural leadership model that was carried out using the model modification process. In testing the influence of the control variable, the gender of the leader, on the burnout variable, the emotional exhaustion dimension obtained an effect of -0.193 with a CR value of -3.088 and a probability effect of 0.002 with a probability value of 0.002 < 0.05. Meanwhile, testing the influence of the control variable, the gender of the leader, on the burnout variable, the dimensions of the decline in achievement obtained an effect of -0.191 with a CR value of -3.237 and a probability value of 0.001 with a probability value of 0.001 < 0.05. So it can be concluded that gender has a significant negative effect on burnout dimensions of emotional exhaustion and decreased achievement. The value of standardized regression weight shows the direction of negative influence, which means that if the leader is female, the emotional exhaustion and decrease in achievement will be lower. The results of the two-dimensional test of the burnout variable are different from the test results between the gender control variable and the depersonalization burnout variable. From the standardized regression weight value -0.117, CR value -1.727 > -2.0 and probability value 0.084 > 0.05, it can be concluded that gender has no significant effect on the burnout dimension of depersonalization. In testing the effect of the stretch goals variable on the burnout variable, the emotional fatigue dimension is known to have an effect value of 0.143, a CR value of 2.253 > 2.0, and a probability value of 0.024 < 0.05. Then, the effect of the stretch goals variable on the burnout variable in the depersonalization dimension obtained an effect of 0.182, CR 2.532 > 2.0, and a probability value of 0.011 < 0.05. Finally, the influence of the stretch goals variable on the burnout variable in terms of the achievement decline dimension produces an influence value of 0.117, CR 2.038 > 2.0 and a probability value of 0.047 < 0.05. It can be concluded that stretch goals have a significant positive effect on burnout dimensions of emotional exhaustion, depersonalization, and decreased achievement. The positive standardized regression weight values for the effect of stretch goals on burnout dimensions of emotional exhaustion, depersonalization, and decreased achievement are 0.143, 0.183, and 0.117. This means that if the perceived stretch goals are higher, the emotional exhaustion, depersonalization, and decreased leadership performance will also be higher.
In testing the influence of the stretch goals variable on the destructive leadership variable, the standardized regression weight value is 0.411 with a CR value of 5.348 > 2.0 and a probability value of 0.000 < 0.05, so it can be concluded that stretch goals have a significant positive effect on destructive leadership. With a positive standardized regression weight value of 0.411, if the perceived stretch goals are higher, the destructive leadership of the leader will also be higher. Conversely, if the perceived stretch goals are getting lower, the destructive leadership of the leader will be lower.
Testing the effect of the gender control variable on the destructive leadership variable obtained a value of -0.027 with a CR value of -0.303 > -2.0 and a probability value of 0.762 > 0.05. Meanwhile, testing the effect of the control variable working period on the destructive leadership variable obtained a value of 0.046 with a CR value of 0.693 < 2.0 and a probability value of 0.489 > 0.05. So, it can be concluded that gender and work period (tenure) also have no significant effect on the destructive leadership of the leader. Testing the effect of burnout dimensions of emotional exhaustion on the destructive leadership variables resulted in an influence value of 0.698 with a CR of 1.351 < 2.0 and a probability value of 0.177 > 0.05. In testing the effect of the burnout variable's depersonalization dimension on the destructive leadership variable, the effect was -0.357 with a CR value of -0.452 > -2.0 and a probability value of 0.651 > 0.05. Then, testing the effect of the burnout variable on the dimensions of decreasing achievement on the destructive leadership variable, the effect value is 0.075, the CR value is 0.188 < 2.0, and the probability value is 0.851 > 0.05. So it can be concluded that all the dimensions of the burnout variable have no significant effect on the destructive leadership of the leader.
As for the test results, the age control variable is considered to have a significant negative effect on the burnout dimension of emotional exhaustion. There is a standardized regression weight value of -0.142 with a CR value of -2.169 < -2.0 and a probability value of 0.030 <0.05. The same pattern was also found in testing the effect of age on the burnout variables of the depersonalization dimension and the dimension of decreased achievement. The test results have a standardized regression weight value of -0.199, a CR value of -2.698 < -2.0, and a probability value of 0.007 <0.05, so it can be concluded that age has a significant negative effect on the burnout dimension of depersonalization. Then, in testing the effect of age on the burnout variable, the dimension of decreased achievement obtained an effect value of -0.126, a CR value of -2.045 < -2.0, and a probability value of 0.041 <0.05. So, it can be concluded that with increasing age, emotional exhaustion, depersonalization, and decreased leadership performance will be lower. It can be concluded that age has a significant negative effect on destructive leadership variables. This is because there is an influence value of -0.157 with a CR value of -2.033 < -2.0 and a probability value of 0.042 <0.05, so, with increasing age, the leader's destructive leadership will be lower. In testing the effect of the control variable gender on interpersonal and organizational counterproductive work behavior, the effect value is -0.008 and -0.039 with CR values of -0.150 and -0.779 > -2.0 and the probability value is 0.881 and 0.436 > 0.05. Meanwhile, testing the effect of the working period control variable on interpersonal and organizational counterproductive work behavior obtained an effect value of 0.081 and 0.060 with CR values of 1.326 and 1.076 < 2.0 and probability value of 0.185 and 0.282 > 0.05. So, it can be concluded that gender and work period (tenure) have no significant effect on interpersonal and organizational counterproductive work behavior. In testing the influence of destructive leadership on interpersonal and organizational counterproductive work behavior, the effect value is -0.069 and -0.045 with CR values of -1.231 and -0.874 and the probability value is 0.218 and 0.382. It can be seen that the CR value is -1.231 and -0.874 > -2.0 and the probability value is 0.218 and 0.382 > 0.05, so it can be concluded that destructive leadership has no significant effect on interpersonal and organizational counterproductive work behavior. The standardized regression weight value shows the direction of the negative effect of -0.069 and -0.045.
In testing the effect of the interaction between psychological capital on interpersonal and organizational counterproductive work behavior, the results obtained are 0.139 and 0.137 with CR values of 2.385 and 2.599 and probability value of 0.017 and 0.009. It can be that the CR value is > 2.0 and the probability value is < 0.05, so it can be concluded that the interaction between psychological capital has a significant positive effect on interpersonal and organizational counterproductive work behavior. So when the psychological capital owned by employees is high, the organizational counterproductive work behavior will be high as well, and vice versa. In testing the effect of age on interpersonal and organizational counterproductive work behavior, the effect value is -0.111 and -0.210 with CR values of -2.050 and -3.329 and probability value of 0.047 and 0.000. So, it can be concluded that age has a significant negative effect on interpersonal and organizational counterproductive work behavior. This means that the older the employee, the lower the interpersonal and organizational counterproductive work behavior.
In testing the effect of the interaction between destructive leadership and psychological capital on counterproductive interpersonal work behavior, the effect value is 0.266 with a CR value of 3.538 and a probability value of 0.000. It can be seen that the CR value is 3.3538 > 2.0 and the probability value is 0.000 < 0.05. Meanwhile, the effect of the interaction between destructive leadership and psychological capital on organizational counterproductive work behavior obtained an effect value of 0.200 with a CR value of 3.033 and a probability effect value of 0.002. It can be seen that the CR value is 3.033 > 2.0 and the probability value is 0.002 < 0.05. So, it can be concluded that the interaction between destructive leadership and psychological capital has a significant positive effect on interpersonal and organizational counterproductive work behavior. The significant positive effect of the interaction of destructive leadership with psychological capital means that it strengthens the negative influence of destructive leadership on interpersonal and organizational counterproductive work behavior where with good psychological capital, interpersonal and organizational counterproductive work behavior will be smaller due to destructive leadership, which is assessed by employees on average fairly high average.

Discussion and Future Direction
This study aims to examine the effect of stretch goals and burnout on destructive leadership and the mediation role of burnout on the impact of stretch goals on destructive leadership. Furthermore, this study also examines the effect of destructive leadership on counterproductive work behavior and the moderation role of psychological capital on the impact of destructive leadership on counterproductive work behavior.
The first hypothesis is supported; stretch goals have a positive effect on destructive leadership. This finding is in accordance with the opinion of Latham (1986), Latham and Locke (2006), LePine et al. (2004), Ordonez et al. (2009), Schweitzer et al. (2004, Barsky (2008) andBandura, et al. (1996) regarding the purpose of influencing a person to take destructive actions, either directly or indirectly. Barsky (2008) states that goals are the strongest predictor of destructive behavior; this is because the level of difficulty of the goal is high but with low acceptance (Locke & Bryan, 1967).
The second hypothesis is not supported since there is no mediation of burnout on the impact of stretch goals on destructive leadership. This can be explained using the theory of self-regulation (Bandura, 1986(Bandura, , 1989. When leaders set goals that are perceived as complex and have high novelty, then rationally, those leaders will not appear to be 'maladaptive' or dysfunctional, but, instead, will appear to be still making rational plans. The leader will consider the 'many benefits' and 'lowest costs' and what is best done. Carver and Scheier (1998) add that self-regulation is the ability to maintain the process towards goals and this has been shown to be important for goal attainment.
The third hypothesis regarding the positive influence of destructive leadership on CWB is not supported by the results. So, in this case, it can be interpreted that the CWB engaged in by employees, both interpersonal and organizational, cannot be predicted from the leadership's destructive leadership. In line with the COR theory (Hobfoll, 1998(Hobfoll, , 2001, SOE employees always try to maintain the quality and quantity of their resources and behave productively regardless of what they perceive as destructive leadership. Furthermore, according to Chatman (1989), in the theory of person-organization fit, the practice of corporate values inherent in every employee can be associated with the pattern of company values. This means that the values of SOE employees are in accordance with the pattern of organizational values. These include value congruence, goal congruence, appropriate supply needs, and capability demands (Muchinsky & Monahan, 1987).
The fourth hypothesis regarding the moderating effect of psychological capital on the impact of destructive leadership on counterproductive work behavior is supported by the results. There is a significant positive effect on the interaction, which indicates a role for psychological capital in reducing the level of CWB behavior, both interpersonal CWB and organizational CWB. The moderating role of psychological capital on the impact of destructive leadership on CWB can be explained through the self-concept theory. The positive self-concept of SOE employees makes them carry out positive selfevaluations, positive self-esteem, and positive self-acceptance. With psychological capital, the perception of oneself becomes one of the important factors influencing behavior (Avey et al., 2010). Furthermore, also supported by the COR theory, Hobfoll (2002) defines psychological capital as an entity that is valued centrally within themselves without being influenced by external parties such as leaders.
This research certainly has some limitations. First, the testing of this model has been carried out only in the Indonesian context. Therefore, to apply the results of this study to different cultural contexts, care must be taken. Future research could test the research model in different country contexts. Second, the design of this research uses a survey with a questionnaire, which measures proactive and deviant behavior, conducted independently by the respondents (selfreported), so there may be subjectivity in assessing one's own behavior. However, selfreporting has been widely used in previous research on leadership and employee performance (Schaubroeck et al., 2007;Illies & Reiter-Palmon, 2008;Wu & Yu, 2009;Jaramilo et al., 2011;Amir, 2019). Further research could increase respondents' sources or use dyadic data. Finally, this study only measures stretch goals as antecedents of destructive leadership. Future research could use a research framework from several empirical studies that use different antecedents from this research.

Conclusion and Suggestion
The main purpose of this study is to conduct a confirmation test of the destructive leadership model by linking the stretch goals variable, the role of burnout as a mediator, and the employee's counterproductive work behavior variable positioned as a consequence of destructive leadership and psychological capital which acts as a moderating variable. The test is carried out by integrating various approaches and theories such as goal setting theory, conservation of resource theory, social exchange theory, self-concept theory, interpersonal deception theory and personorganization fit theory.
Broadly speaking, there are two findings. First, there is a positive influence of stretch goals on destructive leadership, but there is no mediating role of burnout on the effect of stretch goals on destructive leadership. Second, there is no effect of perceived destructive leadership on employees' counterproductive work behavior, but there is a moderating role on the effect of perceived destructive leadership on employees' counterproductive work behavior.
Some basic theories that can be used as the theoretical basis for destructive leadership are goal setting theory, conservation of resource theory, social exchange theory, self-concept theory, interpersonal deception theory and personorganization fit theory. Researchers using proxies for these theories, especially goal setting theory, can provide an explanation that individuals carry out destructive leadership behaviors and counterproductive work behaviors.
The results show that destructive leadership can be predicted from the existence of stretch goals. Meanwhile, burnout, which consists of dimensions of emotional exhaustion, depersonalization and lack of achievement, does not act as a mediator of the direct influence of stretch goals on destructive leadership. The results of testing show that the positive direct influence of destructive leadership on counterproductive employee work behavior is not supported, but there is a moderating role of psychological capital on the effect of destructive leadership on counterproductive employee work behavior.