Introduction

Leaders often face the conundrum of how to deal with employee misconduct (Ong, 2021). Some leaders may insist on appropriate punishment following organizational rules and regulations, but some may enact lenient behavior (Strelan, 2007; Zipay et al., 2021). Leniency is “the act of lessening or removing a negative consequence for misconduct” (Zipay et al., 2021, p.354). Leaders who display lenient behavior tend to reduce the blame for the misconduct or even wipe the slate clean (Strelan, 2007), avoiding the negative consequences for wrongdoers (Butterfield et al., 1996). Therefore, leader leniency, as means of addressing workplace misconduct, could repair relationships and restore the employee’s motivation in the workplace (Bertels et al., 2014; Wenzel & Okimoto, 2014), thereby reducing the harm that misconduct can cause to organizations (Robinson, 2008).

Although previous research has highlighted that the lenient reaction to misconduct generally brings benefits (Bertels et al., 2014; Wenzel & Okimoto, 2014), scholars have recently challenged this positive view of leniency by noting the negative outcomes of leniency (e.g., Treviño, 1992; Zipay et al., 2021). For example, these scholars held that leniency might not stop the misconduct because lenient leaders neglected standards of behavior (Treviño, 1992; Zipay et al., 2021), thus sacrificing justice. Therefore, these inconsistent findings suggest that leader leniency may have potential costs while previous findings may only see the positive sides of leader leniency. Moreover, existing research has not clearly understood what exactly leader leniency leads to and why and when leader leniency produces bright and dark side effects for subordinates. To enhance understanding of the consequences of leader leniency, we explore the potential double-edged sword effect of leader leniency and try to understand why leader leniency helps some subordinates become “good soldiers” and others become “bad apples.” Theoretically, in so doing, we could create consensus between divergent research streams and demonstrate that leader leniency is not always beneficial or detrimental. Practically, knowledge of the nomological net of leader leniency is critical because organizations can leverage it to encourage effective leader leniency and take steps to mitigate negative consequences.

We propose that attributional perspective may address these issues and help us recognize when positive and negative consequences of leader leniency tend to occur. Attribution theory proposes that individuals have an innate tendency to explain the occurrence of events in their surroundings (Weiner, 1985). Thus, attribution theory is particularly relevant to our model because it argues that subordinates’ attribution of leader behavior affects the effect of leader behavior on subordinates’ behaviors (Qin et al., 2020). Accordingly, we explore when and why leader leniency brings organizational citizenship behavior (OCB, one discretionary behavior which benefits organizations but is not related to in-role job performance; Yam et al., 2017) and workplace deviance (another discretionary behavior that violates organizational norms and undermines the well-being of other members and the organization; Bennett & Robinson, 2000) drawing upon attribution theory. We choose these two outcomes because, theoretically, OCB and workplace deviance are the representative outcomes in the subordinates’ perceptions of leader behaviors based on the attribution theory (Qin et al., 2020; Sun et al., 2019). Practically, these two discretionary behaviors are expected and crucial for organizations (Bies et al., 2016) and are likely to arise in the context of leaders’ handling of subordinate misconduct.

Then, we suggest that leader leniency triggers both positive and negative behavior via different mechanisms. As the attribution literature noted, the way individuals attribute others’ actions determines how these actions influence their psychological states (Qin et al., 2020). Specifically, we propose that leader leniency affects OCB and workplace deviance via two mechanisms—one for the bright side (subordinate gratitude, i.e., “a state affect that occurs when a person experiences a benefit and assigns positive attribution to that experience”; Spence et al., 2014) and one for the dark side (psychological entitlement, i.e., a momentary sense that one deserves more than others; Campbell et al., 2004). We choose gratitude and psychological entitlement as mediators because these two variables are generally considered as individuals’ psychological consequences of attributing others’ behavior (Harvey & Dasborough, 2015; Qin et al., 2020; Spence et al., 2014). In addition to being well-aligned with the attribution theory, our focus on gratitude and psychological entitlement as drivers of employee discretionary behaviors (cf., Qin et al., 2020; Sun et al., 2019). Significantly, subordinate relational attributions of leader leniency contribute to reconciling these two seemingly incompatible effects of leader leniency on subordinates.

Moreover, we propose that relational attribution is a critical boundary condition for the process through which enacting leader leniency impacts subordinates. Relational attribution refers to “those explanations made by a focal individual that locate the cause of an event within the relationship the individual has with another person” (Eberly et al., 2011, p.732). In this respect, the explanations that identify the cause of leader leniency made by subordinates might be related to the relationship between leaders and subordinates because these relationships are the essential basis for the evaluation of subordinates on the leader’s behavior (Sun et al., 2019). Moreover, research shows that attributions could either strengthen or weaken how the event triggers both cognitive and affective reactions and behaviors (Qin et al., 2020; Sun et al., 2019). Such that, leader leniency may induce distinct outcomes when subordinates contribute a high or low level of relational attributions to leader leniency. Precisely, we propose that subordinates with low relational attributions of leader leniency are likely to believe leaders treat them leniently because of pure kindness but not their relationships, then spur gratitude and further OCB. Conversely, when subordinates have high relational attributions of leader leniency, they are prone to interpret leader leniency as a special relationship between them and the leader, which will create the feeling that they are unique compared to other colleagues, namely, a sense of psychological entitlement, and further lead to workplace deviance. In summary, we suggest that leader leniency has a double-edged sword effect on subordinate behavior through two paths contingent on the distinct subordinate’s relational attributions of leader leniency.

Our study reevaluates the concept of leader leniency, presenting a balanced view that recognizes both its potential benefits and drawbacks. This perspective invites a rethinking of conventional wisdom that views leader leniency solely as beneficial. Second, we extend our investigation to the realm of gratitude within the workplace, identifying key factors that trigger gratitude among employees. While traditional views have linked gratitude to positive acts like leader leniency, our approach suggests that the emergence of gratitude is more complex and contingent on how such acts are perceived by recipients. Finally, our research contributes to relational attribution theory how leader leniency can lead to varied outcomes through the lens of relational attributions, thus broadening its applicability to a wider range of workplace scenarios.

Theory and hypotheses

Attribution theory

Attribution theory suggests that people are inherently driven to comprehend others’ behaviors in terms of their causes, as a means to make sense of their environment (Weiner, 1985). In the leadership literature, attribution theory frequently serves to elucidate how leader behavior and subordinates’ attribution of leader behavior interact with subordinates’ attitudes and behaviors (Bharanitharan et al., 2021; Qin et al., 2020; Sun et al., 2019). In this study, we select relational attribution in attribution theory because it emphasizes attributional explanations for the inquiry of “why something happened between my leader and me” (Sun et al., 2019, p 530), which takes the nature of the human relationship as the cause of the event and aligns with our research context. Thus, relational attribution provides an approach to explicating interpersonal interactions in a dyadic process (e.g., leaders and subordinates; Eberly et al., 2011). In addition, theoretical and empirical evidence contended that leader behavior and subordinate relational attribution of leader behavior interact with subordinates’ attitudinal and behavioral outcomes (Sun et al., 2019). Therefore, according to attribution theory, we explore the interactive effect of leader leniency and subordinate relational attribution of leader leniency on downstream subordinate attitude and behavior.

The interactive effect of leader leniency and subordinate relational attribution of leader leniency on gratitude

We propose that whether subordinates have gratitude towards leader leniency depends on the extent to which they make relational attributions of leader leniency. Gratitude is typically elicited as a positive response when individuals discern a favor to be the result of the benefactor’s sincere intentions (Bartlett & DeSteno, 2006). In a similar vein, followers might also experience gratitude in response to the leader leniency because of leaders’ benevolent reactions to their missteps (Butterfield et al., 1996). Furthermore, studies have shown that merely receiving a favor is insufficient to elicit gratitude. Instead, it is the recipients’ perceptions of the authenticity behind the action that inform their responses (Belmi & Pfefer, 2015; Fehr et al., 2017; Watkins et al., 2006). This suggests that the elicitation of employee gratitude may hinge on the interplay between the event (e.g., leader leniency) and the employee’s interpretation of the leader’s leniency, particularly in terms of relational attributions. Relational attributions refer to a set of explanatory approaches where individuals attribute the reasons for their experiences to the relationships they maintain with others (Eberly et al., 2011). Research indicates that such attributions can modulate the extent to which employees value certain leader behaviors (Sun et al., 2019). Building on this logic, we argue that relational attributions may serve as a pivotal factor in determining the impact of leader leniency. Specifically, if subordinates perceive their leaders’ leniency primarily as a tactic to maintain their relationship, they may not feel grateful because this perception undermines the unconditional nature of the support (Sun et al., 2019), stripping leader leniency of its altruism. Consequently, subordinates might overlook the added value of leader leniency, diminishing their likelihood of experiencing gratitude.

Conversely, when subordinates attribute leader leniency to factors beyond their relationship, this perspective may foster gratitude. Specifically, a low relational attribution implies that subordinates do not primarily view their relationship with the leader as the catalyst for lenient behavior. Instead, they are inclined to perceive such leniency as an act of genuine kindness and generosity, which in turn triggers a grateful response (Watkins et al., 2006). Based on these insights, we articulate the following hypothesis:

Hypothesis 1

Leader leniency and subordinate relational attribution of leader leniency will interact to influence subordinate gratitude, such that the relationship will be positive when subordinate relational attribution is low and will not exist when subordinate relational attribution is high.

Subordinate gratitude and organizational citizenship behavior

We further propose that subordinate gratitude inspired by leader leniency and low relational attribution positively relates to OCB. Gratitude is not only a product of the supportive and helpful treatment from others but also a motivational factor that can drive positive mental attitudes and benefit others (McCullough et al., 2008). Accordingly, gratitude elicits a desire to reciprocate (Bartlett & DeSteno, 2006), leads employees to notice the positive qualities of others and be more aware of and attentive to others’ needs (Chen et al., 2022), thus engaging in prosocial behavior (Ma et al., 2017). Therefore, when subordinates make low relational attributions of leader leniency, these subordinates may feel that leader leniency is selfless, leading to gratitude subsequently. In response to leaders’ benevolence, subordinates are prone to meet the demands of leaders and organizations and even engage in activities beyond the call of duty, such as volunteering for extra work, namely, OCB (Organ et al., 2006). Thus, combined with Hypothesis 1 which proposes the interactive effect of leader leniency and relational attributions on subordinate gratitude, we propose:

Hypothesis 2

The interaction of leader leniency and subordinate relational attributions has an indirect effect on OCB via gratitude, such that the indirect effect will be positive when subordinate relational attribution is low and will not exist when subordinate relational attribution is high.

The interactive effect of leader leniency and subordinate relational attribution on psychological entitlement

We propose that when subordinate relational attribution of leader leniency is high, leader leniency may lead to subordinate psychological entitlement. Specifically, if subordinates make relational attributions of leader leniency, they may experience inflated perceptions of themselves (Harvey & Martinko, 2009) because they have good relationships with leaders. Even more, when leaders treat subordinates leniently, subordinates who make high relational attributions tend to interpret leader leniency as their sense of superiority (e.g., leaders must tolerate their misconduct to maintain their relationship, Sun et al., 2019). Thus, drawing upon attribution theory, when subordinates make high relational attributions of leader leniency, they are more likely to believe they deserve to be granted such leniency and experience an inflated sense of self-worth, which is called psychological entitlement (Graffin et al., 2013).

On the contrary, when subordinates make low relational attributions of leader leniency, they are less likely to attribute their leader leniency to their deservedness and superiority. In such conditions, subordinates are less likely to inflate their egos and are unlikely to experience psychological entitlement (Qin et al., 2020) when leaders treat them leniently. Therefore, we propose the following hypothesis:

Hypothesis 3

Leader leniency and subordinate relational attribution of leader leniency will interact to influence subordinate psychological entitlement, such that the relationship will be positive when subordinate relational attribution is high and will not exist when subordinate relational attribution is low.

Subordinate psychological entitlement and workplace deviance

We further argue that subordinate psychological entitlement trigged by leader leniency is positively associated with subordinate workplace deviance. As we noted before, psychological entitlement reflects individuals’ inflated self-opinion and expectations without adhering to reciprocity rules (Wu et al., 2022). In this vein, employees with a sense of psychological entitlement may believe that they have more rights than others, thereby undermining the principle of reciprocity and engaging in behaviors that violate organizational norms and undermine the well-being of other members and the organization (Bennett & Robinson, 2000), i.e., workplace deviance. Indeed, extant research has found that psychological entitlement is generally associated with deviance (Qin et al., 2020). Therefore, when subordinates who receive leader leniency experience psychological entitlement under the high relational attributions, they may feel that they should have gained more than they have, supplementing unmet needs by engaging in deviant behavior (Yam et al., 2017). Thus, combined with Hypothesis 3, we propose:

Hypothesis 4

The interaction of leader leniency and subordinate relational attributions has an indirect effect on workplace deviance via psychological entitlement, such that the indirect effect will be positive when subordinate relational attribution is high and will not exist when subordinate relational attribution is low.

Study overview

This study adopted two studies to test our research model. In Study 1, we investigated the interactive effect of leader leniency and subordinates’ relational attribution on subordinate gratitude (Hypothesis 1) and psychological entitlement (Hypothesis 3) using a recall experimental study as in previous research (Zipay et al., 2021). In Study 2, we conducted a multi-source and time-lagged study to validate the full model (Hypotheses 1 to 4). Our research design establishes internal and external validity and provides robust findings (Qin et al., 2020).

Study 1

Samples Participants

We recruited 240 participants who were randomly chosen from Chinese various industries’ full-time employees. In the experiment, eleven samples were dropped because they failed the attention test or material reality check. The final sample included 229 participants (48% male and 52% female, 88.2% with a bachelor’s degree or above, Mage = 27.35 years old, SD = 6.994; Mtenure = 2.09 years, SD = 1.471; Mdyadic tenure =2.10 years, SD = 1.343). All participants were ensured to take part in our experiment voluntarily and we engaged participants in experiments by sending an online link. All participants received 25 RMB (3.68 USD) at the end of this study.

Procedure and experimental design

We manipulated leader leniency and subordinate relational attribution to form a 2 (leader leniency: high vs. low) × 2 (subordinate relational attribution: high vs. low) experimental study. We randomly assigned participants to one of four between-subjects conditions: high leader leniency -high subordinate relational attribution, high leader leniency -low subordinate relational attribution, low leader leniency -high subordinate relational attribution, and low leader leniency -low subordinate relational attribution. To manipulate leader leniency, we asked participants to recall and write down their most recent interactions with leaders. Then, we presented subordinate relational attribution of leader leniency manipulation. Following these manipulations, participants completed measures of gratitude, psychological entitlement, manipulation checks, and demographic variables.

Experimental materials

Leader leniency manipulation. We first asked participants to read the following definition and some examples of leader leniency so that they could understand the leader leniency more accurately:

Leniency, which is a less punitive or severe response to misconduct, often manifests in actions to mitigate or eliminate the negative consequences of misconduct. For example, the leader should deduct Li Hua’s late work, but he/she did not. Moreover, the leader should have deducted all the bonuses for Li Hua’s misconduct toward the company. However, he/she only deducted a small portion of his year-end bonus after verbal criticism.

Then, we asked participants to recall and write down the way their leaders responded to their misbehavior. In the high leader leniency condition, we asked participants to recall and write down the most recent event experienced at work that was handled with leniency by their leaders. In the control condition, we asked participants to recall and write down the most recent event experienced at work that was not handled with leniency by their leaders.

Subordinate relational attribution of leader leniency manipulation. After participants completed the recall task, we instructed them to read research reports designed to manipulate subordinate relational attribution of leader leniency as in previous research (Qin et al., 2020). The statements are based on the conceptualization of relational attribution and prior experimental study on attribution (Qin et al., 2020; Zipay et al., 2021). Specifically, in the high relational attribution condition, participants read the following:

Research has shown that leaders may sometimes respond leniently to their subordinates when facing their misconduct, such as giving them a lighter punishment or even directly exempting them from punishment. There has been a long debate about why leaders may respond leniently to their subordinates. However, a new study involving thousands of US and China samples showed that leaders might show lenient responses to subordinates’ misconduct because of their relationships. Namely, leaders will exhibit lenient attitudes because they have a good relationship with their subordinates.

Considering the research results, when your leader is lenient with you, you may attribute this to your private relationship with your leaders (e.g., having established a good relationship).

In the low relational attribution condition, participants read the following:

Research has shown that leaders may sometimes respond leniently to their subordinates when facing their misconduct, such as giving them a lighter punishment or even directly exempting them from punishment. There has been a long debate about why leaders may respond leniently to their subordinates. However, a new study involving thousands of US and China samples found that leaders may respond leniently to subordinates’ misconduct not because of their relationships. Namely, leaders will exhibit lenient attitudes not because they have a good relationship with their subordinates.

Considering the research results, when your leader is lenient, you may not attribute this to your private relationship with your leaders (e.g., having established a good relationship).

Measures

In this study, we used the established scales to ensure high reliability. In Study 1 and Study 2, the English scales were translated and back-translated using Brislin’s (1980) standard method to ensure measurement equivalence. Unless otherwise noted, all scales in the two studies were measured by a 5-point Likert (1 = strongly disagree; 5 = strongly agree).

Gratitude. We measured subordinate gratitude adapted from Spence et al. (2014) five-item scale of gratitude. To capture the current state, we asked participants to rate how they felt about their directed leader after they recalled the event of leader leniency and read the research report about the relational attribution of leader leniency. A sample item is “I feel grateful to the leader” (α = 0.923).

Psychological entitlement. We measured subordinate psychological entitlement with a four-item scale from Yam et al. (2017). To capture the current state, we asked participants to rate the extent to which they felt this way after they recalled the event of leader leniency and read the research report about the relational attribution of leader leniency. A sample item is " I honestly feel I am just more deserving than others “(α = 0.752).

Manipulation checks. Finally, participants rated their perceptions of leader behavior in the recall exercise using the three-item leader leniency scale developed by Zipay et al. (2021) (α = 0.874). In addition, we measured subordinate relational attribution of leader leniency using a three-item scale developed by Burton et al. (2014) (α = 0.874).

Results

We first tested the experimental manipulation. t-test results showed that the mean score of the experimental condition of leader leniency (M high = 4.222) was significantly higher than that of the control condition (M control = 2.745), t = 17.622, p <.001. Also, the mean score of the experimental condition of subordinate relational attribution (M high = 3.456) was higher than that of the control condition (M control = 2.255), t = 10.637, p <.001. Therefore, the experimental manipulations were successful.

Table 1 shows the means, standard deviations, and correlations among Study 1 variables.

Table 1 Descriptive statistics and correlations in Study 1
Full size table

We conducted an ANOVA analysis to test our hypotheses. First, the results showed that the interaction effect between leader leniency and subordinate relational attribution was related to subordinate gratitude (F (1, 225) = 10.050, p <.01). As Fig. 1 showed, when subordinate relational attribution was low, participants in the leader leniency condition (M = 4.324, SD = 0.515) reported significantly higher gratitude than those in the control condition (M = 3.512, SD = 0.873; F (1, 225) = 32.050, p <.001); When subordinate relational attribution was high, participants in the leader leniency condition (M = 3.857, SD = 0.822) did not report significantly higher gratitude than those in the control condition (M = 3.690, SD = 0.818; F (1, 225) = 1.352, p >.05). Thus, Hypothesis 1 was supported.

Fig. 1
figure 1

The interactive effect of leader leniency and subordinate relational attribution on gratitude in Study 1

Full size image

Second, the results also found that the interaction effect between leader leniency and subordinate relational attribution was related to subordinate psychological entitlement (F (1, 225) = 4.278, p <.05). As Fig. 2 showed, when subordinate relational attribution was high, participants in the leader leniency condition (M = 3.205, SD = 0.929) reported significantly higher psychological entitlement than those in the control condition (M = 2.845, SD = 0.711; F (1, 225) = 6.909, p <.01); When subordinate relational attribution was low, participants in the leader leniency condition (M = 2.767, SD = 0.566) did not report significantly different psychological entitlement than those in the control condition (M = 2.807, SD = 0.681; F (1, 225) = 0.085, p >.05). Thus, Hypothesis 3 was supported.

Fig. 2
figure 2

The interactive effect of leader leniency and subordinate relational attribution on psychological entitlement in Study 1

Full size image

Study 2

Samples and procedures

Study 2 was a multi-sources and multi-wave study in China, with each interval of one month. Before conducting the survey, we contacted the subordinates or leaders of the enterprises in advance. After explaining the purpose of the study and the research process to the subjects, the employees or leaders agreed to help us complete the survey. Specifically, when the participants we connected agreed to engage in the survey, we connected their HR department with their help and asked HR to provide us with a list of one leader and their direct subordinates. Then we randomly selected one subordinate to form a leader-subordinate dyad with the leader. Only when they have agreed to participate in this survey can they start filling out the questionnaire. To ensure the matching of leaders and subordinates, we coded the leaders and subordinates with numbers in the questionnaire for matching, and we collected questionnaires online. At Time 1, subordinates rated the leader leniency, relational attribution of leader leniency, internal and external attributions of leader leniency, liking for leaders, and demographic information. 302 subordinates participated at Time (1) At Time 2, subordinates who participated in T1 then evaluated their gratitude and psychological entitlement. 18 subordinates did not participate in this survey for reasons such as leaving their jobs. Therefore, 284 subordinates’ questionnaires were received at Time (2) At Time 3, direct leaders of these subordinates participated in the T2 survey to evaluate subordinates’ OCB and rated their own demographic information; subordinates rated their own workplace deviance. Finally, 268 leaders and 268 subordinates were obtained after eliminating invalid questionnaires, with a valid response rate of 88.74%. All participants received 50 RMB (7.46 USD) at the end of this study.

Among the subordinates, 42.5% were male, and the average age of subordinates was 29.25 years old (SD = 7.897); in terms of education level, the largest proportion was 50.4% with a bachelor’s degree, and the average number of years of work with their direct leader was 2.345 years (SD = 1.427). Among the leaders, the average age of the leader was 34 years old, 50% of women and 50% of men. Moreover, 73.5% of leaders held a bachelor’s degree or above. 74.3% of leaders have been working for more than five years. Participants held various jobs (e.g., sales, engineers, and IT professionals) in various industries (e.g., retail, manufacturing, IT, and education).

Measures

Leader leniency (T1). We measured leader leniency with three three-item scale developed by Zipay et al. (2021). A sample item is " My leader has given me at work a lighter punishment for my misconduct than he/she could have.” (α = 0.783).

Subordinate relational attribution of leader leniency (T1). We rated subordinate relational attribution using a three-item scale developed by Burton et al. (2014). A sample item is " The cause of my supervisor’s lenient behavior is a result of the relationship we have.” (α = 0.847).

Gratitude (T2). We used five items adapted from Spence et al. (2014) to rate subordinate gratitude. A sample item is " I feel grateful to the leader.” (α = 0.914).

Psychological Entitlement (T2). We used four items from Yam et al. (2017) to evaluate subordinate psychological entitlement. An example is, " I honestly feel I’m just more deserving than others " (α = 0.918).

Organizational citizenship behavior (T3). We measured organizational citizenship behavior with a four-item scale developed by Cardona et al. (2004). A sample item is “The subordinate takes the trouble to mentor new members of his/her department, even when not asked to do so” (α = 0.861).

Workplace deviance (T3). We measured workplace deviance with a fourteen-item scale developed by Aquino and Bradfield (1999). A sample item is " Intentionally arrives late for work” (α = 0.943).

Control variables. Based on previous studies, we controlled for subordinates’ gender, age, education, and the time spent with leaders (Berry et al., 2007). In addition, research has shown that subordinates’ internal and external attributions of leader behavior may also influence subordinates’ behavior (Sun et al., 2019), so we controlled the internal and external attributions of leader leniency. Internal attributions (α = 0.829) and external attributions (α = 0.807) were measured using the four-item scale and the three-item, respectively, developed by Burton et al. (2014). Furthermore, feelings about the leader may influence subordinates’ perceptions and judgments towards the leader (Hall & Lord, 1995); we also controlled the subordinates’ liking for leaders as in previous research (Bharanitharan et al., 2021). subordinates’ liking for leaders was measured using a four-item scale developed by Brown and Keeping (2005) (α = 0.924).

Analytic strategy

We used SPSS 25.0 and Mplus 7.4 to test our preliminary analysis and hypotheses. We performed path analysis using Mplus 7.4 to test hypotheses. Particularly, leader leniency and subordinate relational attribution of leader leniency were grand-mean centered prior to analysis. Finally, we bootstrapped the sampling distribution to test indirect effects (Preacher & Hayes, 2008). Specifically, we conducted bias-corrected bootstrapping with 5,000 iterations.

Results

Common method bias and confirmatory factor analysis

According to Harman’s one-factor test results, the first unrotated factor explained 19.75% of the total variance (< 50%). Therefore, the common method bias of this study was effectively controlled. Moreover, Table 2 showed the results of CFA. The six-factor model (χ2/df = 2.105, CFI = 0.912, TLI = 0.902, SRMR = 0.058, RMSEA = 0.064) had the best fitting effect compared with other competitive models, indicating that the core variables have good discriminant validity in this study.

Table 2 Results of confirmatory factor analysis in Study 2
Full size table

Hypotheses testing

Table 3 showed the descriptive statistics, correlations, and Cronbach’s α.

Table 3 Descriptive statistics and correlations in Study 2
Full size table

Then, we conducted a Bootstrapping analysis to examine our hypotheses with Mplus 7.4, Fig. 3; Table 4 showed the results.

Fig. 3
figure 3

Path coefficients of the full model in Study 2.  *p  < .05; **p  < .01, ***p  < .001. Unstandardized regression coefficients are reported

Full size image
Table 4 Results of moderated mediation analysis (full model) in Study 2
Full size table

After accounting for the control variables, the interactive effect of leader leniency and subordinate relational attribution on subordinate gratitude was negative and significant (β = − 0.143, p <.01). When subordinate relational attribution was low, the effect was 0.180, 95% CI= [0.026, 0.323]; When subordinate relational attribution was high, the effect was − 0.089, 95% CI= [− 0.246, 0.062], indicating the effect was not significant; the difference between these effects was significant (estimate = − 0.269, 95% CI= [− 0.474, − 0.079]). Therefore, Hypothesis 1 was supported. We further conducted a simple test to verify the interactive effect of leader leniency and subordinate relational attribution on subordinate gratitude (see Fig. 4).

Fig. 4
figure 4

The interactive effect of leader leniency and subordinate relational attribution on gratitude in Study 2

Full size image

Subordinate gratitude was significantly related to the subordinate OCB (β = 0.201, p <.01). Furthermore, the indirect effect of leader leniency on subordinate OCB via subordinate gratitude was significant when relational attribution was low (estimate = 0.036, 95% CI= [0.008, 0.081]), but was not significant when relational attribution was high (estimate = − 0.018, 95% CI= [− 0.063, 0.008]). The difference between these indirect effects was significant (estimate = − 0.054, 95% CI= [− 0.122, − 0.016]). Therefore, Hypothesis 2 was supported.

Similarly, the interactive effect of leader leniency and subordinate relational attribution on subordinate psychological entitlement was positive and significant (β = 0.185, p <.01). When subordinate relational attribution was high, the effect was 0.235, 95% CI= [0.028, 0.434]; When subordinate relational attribution was low, the effect was − 0.112, 95% CI= [− 0.313, 0.065], indicating the effect was not significant; the difference between these effects was significant (estimate = 0.348, 95% CI= [0.089, 0.613]). Thereby, Hypothesis 3 was supported. We further conducted a simple test to verify the interactive effect of leader leniency and subordinate relational attribution on subordinate psychological entitlement (see Fig. 5).

Fig. 5
figure 5

The interactive effect of leader leniency and subordinate relational attribution on psychological entitlement in Study 2

Full size image

Subordinate psychological entitlement was significantly related to subordinate workplace deviance (β = 0.097, p <.01). In addition, the indirect effect of leader leniency on subordinate workplace via subordinate psychological entitlement was significant when relational attribution was high (estimate = 0.023, 95% CI= [0.004, 0.060]) but was not significant when relational attribution was low (estimate = − 0.011, 95% CI= [− 0.043, 0.004]). The difference between these indirect effects was significant (estimate = 0.034, 95% CI= [0.006, 0.085]). Thus, Hypothesis 4 was supported.

General discussion

Although being lenient to subordinate misconduct is a common phenomenon in the workplace, few studies have focused on the impact of leader leniency on subordinates. Drawing upon the attribution theory, we clarified that leader leniency has a double-sword effect on subordinates. Especially, across an experiment study (Study 1) and multi-source and multi-wave survey (Study 2), we explored that when subordinate relational attribution of leader leniency was low, leader leniency led to subordinate gratitude, in turn, promoted OCB. Conversely, when subordinate relational attribution of leader leniency was high, leader leniency induced subordinate psychological entitlement, thus resulting in workplace deviance.

Theoretical implications

Our research makes several theoretical contributions. First, our core contribution is to clarify the effects of leader leniency and to offer consensus for the research by providing a more comprehensive picture of both the positive and negative effects of leader leniency. Previous research has mainly contended that the lenient reaction to misconduct generally brings benefits, such as it could repair relationships and restore the employee’s motivation in the workplace (Bertels et al., 2014; Wenzel & Okimoto, 2014). Nevertheless, we question the prevailing agreement that leader leniency that leader leniency is universally beneficial and propose that it could bring benefits and costs. Examining both the positive and negative effects of leader leniency offers a more dialectical perspective than previously assumed, underscoring the necessity to acknowledge the potential drawbacks of leader leniency. Thus, our work provides consensus to the leader leniency literature by explaining how leader leniency brings positive and negative outcomes.

Secondly, our study enhances our understanding of the factors that lead to gratitude, thereby deepening our insights into how gratitude arises within the workplace. Previous studies have established that the primary source of gratitude is positive favor factors. For instance, research has identified various positive factors that trigger feelings of gratitude, such as interactional justice, perceived organizational support, the reactive helping behavior of others, generous favor-giving by others, and a team’s error management climate (Chen et al., 2022; Ford et al., 2018; Kim et al., 2018; Lee et al., 2019; Ouyang et al., 2018). However, our findings challenge the conventional perspective. Specifically, we have shown that receiving favors (e.g., leader leniency) from others does not always induce gratitude, and this response depends on how subordinates interpret these actions. In essence, our research confirms that not every act of favor-giving results in the development of gratitude (Lee et al., 2019) and delineates its limiting conditions. Thus, our study expands the field of gratitude research and prompts further investigations into the mechanisms that promote the emergence of gratitude in the workplace.

Finally, our research enriches attribution theory in two significant ways. Firstly, we expand upon attribution theory by presenting a more nuanced understanding of relational attributions in various contexts. Prior research has predominantly concentrated on the application of relational attribution to negative events, such as those involving abusive supervision or adverse performance feedback (e.g., Burton et al., 2014; Eberly et al., 2017). However, through an experimental study and a field study, our findings demonstrate how leader leniency, indicative of positive leader behavior, interacts with subordinates’ relational attributions to shape their responses to such leniency. By extending the range of events that provoke relational attributions beyond negative occurrences to include positive events as well, we redirect the focus from exclusively negative contexts to encompass positive ones. In doing so, we significantly broaden the applicability of relational attribution theory.

Secondly, our study enhances the relational attribution theory by refining our understanding of the outcomes associated with relational attributions. Initial studies highlighted that relational attributions often lead to adaptive behaviors aimed at improving interpersonal relations (Eberly et al., 2011). Consequently, much of the existing research on relational attributions has been devoted to investigating their positive outcomes, such as the expression of remedial voice, engagement in interpersonal citizenship behaviors, and efforts toward relational enhancement (Eberly et al., 2011, 2017). However, it has been noted that not every instance of relational attribution results in efforts to strengthen relationships (Puranik et al., 2019), and this can vary based on the perceived significance of the situation and the context in which relational attributions are made. Our findings add a new dimension to this discussion by illustrating that relational attributions can also lead to negative behaviors, such as workplace deviance, particularly in scenarios involving leader leniency. Through this, our research broadens the attribution theory literature by pinpointing specific conditions that elicit relational attributions and highlighting the diverse outcomes they may engender.

Practical implications

Our research also brings practical implications for organizations and leaders. First, leader leniency is generally deemed beneficial for wrongdoers and organizations (Zipay et al., 2021); our results indicated that leader leniency is a double-edged sword for subordinate behavior. Our study found that leader leniency may also trigger subordinate psychological entitlement and workplace deviance, besides its positive outcomes. Therefore, we recommend that leaders consider the potential costs of leader leniency. For instance, organizations could build a moral climate, develop ethical standards, monitor employee behavior, and enforce ethical behavior (Yam et al., 2017) so employees do not feel entitled because of leader leniency.

Second, we found that subordinates’ judgments about why leaders enact lenient behavior affect the effects of leader leniency. Thus, leaders should consider the role of subordinate attributions about leader leniency when they enact lenient behavior. Particularly, leaders should be aware that subordinates are more likely to engage in workplace deviance when subordinates attribute leader leniency to the relationship with leaders. Therefore, to reduce the costs of relational attribution, leaders should take steps to reduce subordinate relational attribution of leader leniency. For example, leaders should better conduct a uniform standard of leniency for each subordinate’s misconduct instead of showing leniency to specific subordinates, especially those who have a good relationship with them.

Third, we found that when subordinates’ relational attribution is high, leader leniency will induce subordinates’ psychological entitlement, leading to workplace deviance. Our results showed that workplace deviance could be caused not only by the employee’s personality, and employees’ situational perceptions (Mackey et al., 2021) but also by subordinates’ unconscious self-inflation. Accordingly, leaders should deeply analyze the causes of workplace deviance and take targeted measures to reduce and mitigate workplace deviance from a broader perspective (e.g., avoid granting excessive leniency only to subordinates with whom they have a good relationship).

Limitations and future directions

This study has provided a theoretical framework for understanding the effects of leader leniency. Meanwhile, as with most studies, this study is not without limitations. First, we conducted a multi-source and time-lagged survey to test our model in Study 2. We asked leaders to rate subordinate OCB to reduce common method bias. However, we rated subordinate workplace deviance using self-reported data because of the following reasons. First, studies have pointed out that self-reported data is particularly suitable for gauging workplace deviance due to its subjective nature because that the individual involved in the deviant behavior is the most accurate source of information about their actions (Ferris et al., 2016; Aquino & Douglas, 2003). Second, empirical reviews conducted by Berry et al. (2012) indicate that other reports of workplace deviance add little beyond self-reports. Although a viable approach, the self-report rate is susceptible to social desirability that may distort ratings (Podsakoff et al., 2003); we encourage future research could attempt other data collection methods (e.g., objective assessments or multisource measures) to avoid these possible problems. Second, research suggested that gratitude is multilevel, including event, individual, and organizational levels (Fehr et al., 2017). Different levels of gratitude have different effects. In Study 2, gratitude was only measured at a specific time, which may not reflect the individual level effectively. Future research may consider using repeated gratitude measurements over time to better reflect individual-level gratitude. Finally, our study explored the effects of leader leniency on subordinate behavior in a relatively short time. However, over time, leaders may adjust their follow-up strategies for using lenient behaviors after realizing that employees have reactions to their leniency (Zipay et al., 2021). Accordingly, we encourage scholars can capture these temporal effects by conducting appropriate designs, such as longitudinal studies.

Conclusion

Drawing upon the attribution theory, we found that leader leniency is a double-edged sword and its effects depend on subordinates’ relational attribution. Our findings reveal that leader leniency leads to OCB and workplace deviance through subordinate gratitude and subordinate psychological entitlement. We hope this study sparks future research on leader leniency.