Organizational identification (OI), defined as the perception of oneness or belongingness with the organization (Ashforth & Mael, 1989), has attracted burgeoning interest in management research (Edwards, 2005). As a core psychological state, OI reflects the underlying bond between the employee and the organization (Mael & Ashforth, 1992). Mainstream OI research adopts a “positive outlook” perspective, whereby OI is potentially capable of explaining favorable changes in the employee’s attitudes and behaviors, such as increased job satisfaction, job performance, commitment, work engagement, motivation, and organizational citizenship behaviors, alongside reduced turnover intention (Bartel, 2001; Blader & Tyler, 2009; Liu, Loi, & Lam, 2011). These positive results have led to OI being perceived as a “magic bullet” for enhancing desirable work attitudes and behaviors (Ashforth & Mael, 1989; Lee, Park, & Koo, 2015; Liu et al., 2011; Mael & Ashforth, 1992).

Notwithstanding this potential value of OI, some scholars have warned against an overly positive and potentially biased literature, noting that OI may have undesirable consequences as an invisible force shaping unethical work behaviors (Ashforth & Anand, 2003; Dukerich, Kramer, & McLean Parks, 1998). The malevolent side of OI is a timely and relevant topic given organizational scholars’ recent emphasis on identifying which contextual and individual-level factors drive unethical organizational behavior (Gino, Schweitzer, Mead, & Ariely, 2011; Moore, Detert, Klebe Treviño, Baker, & Mayer, 2012). Burgeoning interest in the malevolent side of OI is also highlighted in a recent review describing the potentially deleterious consequences of high OI (Conroy, Henle, Shore, & Stelman, 2017).

Although research on the positive consequences of OI has made significant inroads, only a few studies have provided empirical support for the negative perspective. For instance, Umphress, Bingham, and Mitchell (2010) found high levels of OI to promote unethical pro-organizational behaviors. Similarly, Chen, Chen, and Sheldon (2016) showed that the process of moral disengagement explained the positive relationship between organizational identification and unethical pro-organizational behaviors.

Given the untapped nature of this research field, this study is primarily motivated by the need to explore the processes and conditions through which OI is associated with unethical organizational behaviors. In establishing the underlying mechanisms and boundary conditions associated with OI and unethical behaviors, we focus on three types of unethical behaviors: (a) attempting to harm the organization (i.e., counterproductive work behaviors “CWBs”; Fox, Spector, & Miles, 2001); (b) helping oneself at others’ expense (i.e., careerism; Katz & Sala, 1996); (c) and engaging in unethical acts with intent to benefit the organization and/or its members (i.e., unethical pro-organizational behaviors; Umphress & Bingham, 2011).

In this study, we mainly build on the principles of social identity theory (SIT; Tajfel & Turner, 1985) to explain the dynamics between our core variables. More specifically, one of our core assumptions is that strong organizational attachment can blind employees to ethicality, generating an expanded self-view from their enormous loyalty to the organization, which yields feelings of increased psychological entitlement.

Heeding calls to integrate more individual-level factors as key drivers of unethical work behaviors (Andreoli & Lefkowitz, 2009; Kish-Gephart, Harrison, & Trevino, 2010), we also propose that individuals with a high “manipulative personality” (often referred to as the “dark triad”)Footnote 1 (Jonason & Webster, 2010) as reflected in callous, self-serving, and devious tendencies, are more likely to cultivate high entitlement perceptions, especially when identifying strongly with their organization.

In contrast, less manipulative individuals are sensitive, honest, modest, and empathetic to others’ needs (Jones & Paulhus, 2010). When identifying with their organization, such individuals might not feel compelled to acquire more from the organization, thus mitigating the positive relationship between OI and psychological entitlement. In turn, increased entitlement perceptions make highly manipulative individuals cross the line by increasingly engaging in both pro-organizational and self-interested unethical work behaviors (see Fig. 1 for our theoretical model).

Fig. 1
figure 1

Research model. A moderated mediation model: manipulative personality moderates the indirect effects of organizational identification on unethical work behaviors (unethical pro-organizational behaviors, careerism, and counterproductive work behaviors) through psychological entitlement

Full size image

Our research makes multiple contributions. First, we introduce psychological entitlement as a mechanism to explain the dynamics between OI and unethical work behaviors. We thereby address researchers’ recent calls to unravel the processes underlying the undesirable consequences of OI (Blader, Patil, & Packer, 2017; Lee, Park et al., 2015). Second, by treating psychological entitlement as an intervening mechanism between OI and unethical work behaviors, we extend the construct’s nomological network. Third, we contribute to the OI literature by examining manipulative personality as an important dispositional trait between OI and psychological entitlement relationship. Finally, we investigate whether the “neutralization” mechanism of psychological entitlement and the moderating role of manipulative personality hold for two categories of unethical work behaviors: (a) self-interested unethical behaviors that may harm an organization and its members (i.e., CWBs and careerism), and (b) unethical pro-organizational behaviors. Considering both categories in the same study allows meaningful comparisons across these sets of work behaviors.

Theory and hypotheses development

Organizational identification and unethical work behaviors

The central theoretical framework on which we draw to explain the influence of OI on pro-organizational and self-interested unethical work behaviors is SIT, one of the prominent approaches in organizational identification research (Tajfel & Turner, 1985). In general, SIT posits social identity as the part of an individual’s self-derived from belonging to a social group, and the worth and emotional attachment one feels through inclusion therein. A social identity is formed through three intra-individual processes of group-based identification: (a) social categorization, (b) social comparison, and (c) social identification. In the first stage, individuals mentally evaluate social information to classify themselves into social groups, according to gender, social class, religion, ethnic background, organizational membership, etc. After thus categorizing themselves, individuals engage in social comparisons, assessing their own group against appropriate others. Having done so, individuals will finally form an emotional bond with their own group, which results in social identification (Tajfel, 1982; Tajfel & Turner, 1985).

Given these underlying dynamics of SIT, people with high OI may respond differently in the type of unethical work behavior they exhibit. Because unethical work behavior is complex and may manifest in different forms, we differentiate three forms of unethical behavior: unethical pro-organizational behaviors, CWBs, and careerism. Unethical pro-organizational behaviors encompass unlawful and ethically questionable actions that are performed to help or advantage the organization (Brief & Motowidlo, 1986). CWBs are intentional, self-serving, and unethical acts, oriented towards harming the organization and its members (Fox et al., 2001). Finally, careerism refers to the individual belief that career progress can only be achieved through unethical and illicit tactics (Feldman & Weitz, 1991).

In line with SIT, we assert that individuals who strongly identify with their organization (Tajfel, 1982) may participate in behavior benefiting the organization at the expense of important moral values. Their strong sense of association may make them immune to immoral behavior and indifferent to ethical standards in their endeavors to help the organization, thus fostering high levels of unethical pro-organizational behaviors. By contrast, since highly identified employees take pride in their organizational membership and attach emotional significance to the organization through their sense of belonging, they are less likely to harm the organization or violate organizational norms, which decreases their likelihood of engaging in CWBs. Additionally, consistent with SIT, we assert that employees who exhibit high OI define themselves in terms of the organization’s identity and act in ways that promote the organization’s best interests (Ashforth & Mael, 1989). Their strong focus on the organization’s vantage point makes highly identified individuals likely to work harder and engage less in careerism. Thus, we hypothesize the following:

  • H1: OI is positively related to unethical pro-organizational behaviors and negatively related to CWBs and careerism.

Mediating effect of psychological entitlement on the organizational identification–unethical work behaviors relationship

Building on the malevolent perspective of identification and SIT, we also assume that high OI promotes destructive outcomes through activating psychological entitlement, here defined as an individual’s conviction that he/she deserves more and is authorized to gain higher privileges, rewards, and status (Campbell, Bonacci, Shelton, Exline, & Bushman, 2004). Specifically, we argue that psychological entitlement neutralizes the moral and ethical imperatives associated with unethical work behaviors. Individuals who feel entitled are more likely to develop disproportionate self-centeredness, perceiving themselves as special, unique, and worthy of preferential treatment (Westerlaken, Jordan, & Ramsay, 2017). These self-serving thoughts contribute to the belief that they transcend important ethical standards.

Research indicates that individuals who deeply identify with their organization are so overly immersed that they not only view the organization as an important part of themselves, but also believe they are indispensable to its functioning (Galvin, Lange, & Ashforth, 2015). Such highly identified employees might think they know what is best for the organization, leading to high psychological entitlement (Emmons, 1984). Additionally, strong identification makes individuals’ identity intensely subsumed so that they gain more control and influence in the organization, of which they feel they are a critical component. Such highly identified individuals might develop a sense of importance in the organization, potentially making them feel that it is actually “all about them” (Chatterjee & Hambrick, 2007).

According to SIT, high identification with an esteemed organization gives individuals an immediate basis for affirmative self-view and favorable social comparisons between themselves and others (Tajfel & Turner, 1979). This self-inflated sense that they deserve more than their actual input may drive psychologically entitled employees to act unethically, violate norms when they do not get what they want (Harvey & Harris, 2010; Lee, Schwarz, Newman, & Legood, 2015), and gain rewards without effort. Based on the aforementioned arguments, we hypothesize the following:

  • H2: Psychological entitlement mediates the relationship between OI and unethical work behaviors (i.e., unethical pro-organizational behaviors, CWBs, and careerism).

Moderating effect of manipulative personality

Manipulative personality often called “the dark triad” (Jonason, Li, Webster, & Schmitt, 2009; Jonason & Webster, 2010) encompasses three socially aversive personality traits with shared and overlapping tendencies: Machiavellianism (deceitful), narcissism (self-loving), and psychopathy (cold-hearted and impulsive) (Paulhus & Williams, 2002). Researchers continue to argue whether the traits comprising manipulative personality, i.e., “the triad” should be treated as separate constructs (Muris, Merckelbach, Otgaar, & Meijer, 2017) or as a single unified construct (Jonason et al., 2009). Advocates of the latter perspective cite empirical evidence of using shortened scales, which can only measure shared callousness and manipulation components (Jonason et al., 2009).

In general, “the triad” remains entangled with the combined study of the three distinct constructs (Jonason et al., 2009; Jonason & Webster, 2010). Unfortunately, studies concerning the “core of manipulative personality” are limited (Jonason et al., 2009), and more research is required to advance understanding of the shared variance and overlap. Accordingly, we chose to adopt the unitarian view of the triad treating it as a single construct. Also, psychometrically sound instruments such as the Dirty Dozen scale (Jonason & Webster, 2010) focus more on the common characteristics of selfishness, manipulation, and callousness, which tend to be the shared or common components of these triad personalities (Muris et al., 2017).

Given the socially aversive nature of the triad personality, it has unsurprisingly been found to relate strongly to dysfunctional work behaviors (O’Boyle, Forsyth, Banks, & McDaniel, 2012; Spain, Harms, & Lebreton, 2013). Highly manipulative individuals tend to be attention seekers, excessively proud, and extremely egoistic (Jones & Paulhus, 2010; Paulhus & Williams, 2002), which makes them self-absorbed. When such individuals strongly identify with their organization, they are more likely to assume that the organization’s identity is subsumed within them, and thus consider their own identity as central and their organization’s identity as secondary. Research has also demonstrated that highly manipulative individuals have a constant need for power, experience a continuous struggle for higher status/prestige, and have an insatiable appetite for being in the spotlight (Christie & Geis, 1970; Rauthmann, 2012). Being overly materialistic, they firmly believe that when they strongly identify with their organization, their loyalty to and bond with it should be reciprocated through special favors and compensation, which feeds into higher psychological entitlement. Finally, their sense of jealousy, lethargy, and greed (Veselka, Giammarco, & Vernon, 2014) helps foster a strong belief that they can manipulate rules to get ahead of others without investing their due share of effort, again invoking higher psychological entitlement. Hence, drawing from our core theoretical framework (i.e., SIT), we propose that high manipulative personality might aggravate the effects of OI on psychological entitlement.

The above argumentation establishes manipulative personality as an important boundary condition that moderates the relationship between OI and psychological entitlement. Accordingly, it is also conceivable that manipulative personality conditionally influences the strength of the indirect relationship between OI and unethical work behaviors via psychological entitlement, indicating a pattern of moderated mediation between the study’s variables (see Fig. 1). Hence, we expect a strong (weak) relationship between OI when manipulative personality is high (low). In summary, we hypothesize the following:

  • H3: Manipulative personality moderates the positive relationship between OI and psychological entitlement, such that it is more pronounced for individuals with high manipulative personality.

  • H4: Manipulative personality moderates the indirect effects of OI on unethical work behaviors through psychological entitlement. Specifically, psychological entitlement mediates the indirect effects when manipulative personality is high but not when it is low.

Methods

Sample and data collection

We employed a three-wave research design, allowing temporal segregation of the measurement of our predictors (T1: OI), mediators and moderators (T2: psychological entitlement and manipulative personality), and outcome variables (T3: unethical pro-organizational behaviors, CWBs, and careerism). The time lag between each measurement point was 3 weeks. Adopting such a design helps to curtail potential issues arising from solely self-reported and single-source data collection methods. To further mitigate the concern of self-report bias, data on CWBs were collected from co-workers; data on all other variables were collected through self-report.

Peer reporting to assess workplace behaviors, such as CWBs, has been especially effective in collectivistic cultures, where people are more aware of others’ behaviors (Barclay & Aquino, 2010; Fox, Spector, Goh, & Bruursema, 2007; Penney & Spector, 2005; Raja & Johns, 2010). Although some studies have employed self-rated measures of CWBs (Palmer, Komarraju, Carter, & Karau, 2017; Smith, Wallace, & Jordan, 2016), this literature is not much appreciated. That is, such negative behaviors usually go unreported when tapped with self-reports (Fox et al., 2007; Penney & Spector, 2005). Also, a major limitation of this approach is the influence of social desirability (Fida, Paciello, Tramontano, Barbaranelli, & Farnese, 2015; Palmer et al., 2017). Furthermore, peer ratings are not only valid and reliable sources of employee data (Gardner, Scogin, Vipperman, & Varela, 1998; Greguras & Robie, 1998) but also provide important information due to frequent interactions and the greater ability of peers to observe their colleagues’ work behaviors (Harris & Schaubroeck, 1988).

An advantage of peer reports over supervisor reports is that they enable unique dyadic pairing, thereby eliminating the data nesting issues associated with having one supervisor for multiple respondents. More specifically, data nesting was eliminated by ensuring that one peer only reported for one respondent. Cases where one peer reported for two or more respondents were deleted prior to our analyses.

A field survey was designed and disseminated to employees working for six different organizations in the service sector in Islamabad (Pakistan): two universities, one telecom company, one private bank, one private hospital, and one district court. Respondents were mostly entry- and mid-level managers. In Pakistan, a master’s degree is usually a prerequisite for being appointed to a managerial role.

Because our study’s respondents are in white-collar jobs ranging from entry- to mid-level management, most were highly educated. Several managerial positions are represented in the sample, such as “Assistant Manager Finance,” “Senior Advocate,” “Assistant Professor,” “Senior Lab Engineer,” and “Branch Manager.”

Before consenting to participate, participants received a cover letter explaining the study’s objectives and assuring them of the strictest confidentiality. The letter also stated that participation was voluntary and included the principal investigator’s contact information in case participants had questions about the study or wanted to share feedback. Participants were asked to identify peers they had known for at least 6 months to provide peer-reported data on CWBs. To avoid nesting issues in dyadic data, we ensured that every employee had a unique peer assessor for CWBs.

At T1, we distributed 650 questionnaires; 510 were completed, resulting in a response rate of 78%. Three weeks after T1, we asked the participants to complete another questionnaire to measure psychological entitlement (mediator) and manipulative personality (moderator). A total of 400 useable surveys were completed at T2, yielding a response rate of 62%. Finally, 3 weeks after T2, we again contacted the same participants and asked them to complete a third survey measuring the study’s outcomes. Also, at T3 we asked the co-workers of our respondents to complete a questionnaire measuring CWBs. The three time waves yielded a total of 345 matched self-report questionnaires. After 39 of the peer-report questionnaires were discarded due to missing data, the final sample comprised 306 useable responses, giving an overall response rate of 47%.

Prior to our analyses, we checked the data for missing values and inattentive response patterns (e.g., straight lining). In addition to visually inspecting for response patterns, we calculated within-subject standard deviations for all variables combined in one section of the questionnaire involving the concurrent measurement of more than one construct; for example, manipulative personality and psychological entitlement (T2), and unethical work behaviors and careerism (T3) (Leiner, 2017). Respondents who gave the same answer to all questions in one section would have a standard deviation of zero over these questions, thus indicating straight-lining behavior. Individual standard deviations ranged between 0.36 and 3.01 (Averagesd = 1.32, SDsd = 0.62) for the combination of manipulative personality and psychological entitlement, and between 0.50 and 3.03 (Averagesd = 1.47, SDsd = 0.53) for the combination of unethical pro-organizational behaviors and careerism. Overall, these findings suggest that straight lining is very unlikely to have affected the reliabilities of this study’s scales.

Participants in our final sample were mainly male (78%) and married (61.4%), and the majority held a master’s degree (57.5%) and worked in HR/administrative departments (74.2%). Additionally, a considerable proportion of participants held mid-level managerial positions (37.3%). Finally, on average, our participants were 31.2 years old (SD = 6.8), had spent 6.6 years working at their present company (SD = 6.1), and had 8.3 years of total working experience (SD = 6.8).

Measures

The survey was administered in English, being the official correspondence language in all offices and the instruction language in all high schools and universities in Pakistan. Previous studies in similar fields have used English questionnaires in Pakistan without encountering any language issues (e.g., Abbas, Raja, Darr, & Bouckenooghe, 2014; Naseer, Raja, Syed, Donia, & Darr, 2016; Raja & Johns, 2010). Further, due to the time-lagged nature of data collection, we had to brief respondents about the process for completing the questionnaires and answer any questions about items that were unclear or poorly understood. For all these reasons, we decided to keep the original English versions of all the measures, rather than translating into Urdu. Unless otherwise mentioned, all scales were measured on a seven-point Likert scale, ranging from 1 = “strongly disagree” to 7 = “strongly agree.”

Organizational identification

OI was measured using a six-item scale developed by Mael and Ashforth (1992). Sample items include “When someone criticizes my organization, it feels like a personal insult”; “When I talk about my organization, I usually say ‘we’ rather than ‘they.’” Cronbach’s alpha reliability for this scale was very good (α = 0.92).

Psychological entitlement

A nine-item scale developed by Campbell et al. (2004) was used to assess psychological entitlement. Sample items include “I honestly feel I’m more deserving than others”; “I demand the best because I’m worth it.” Cronbach’s alpha reliability for this scale was very good (α = 0.95).

Manipulative personality

Manipulative personality was tapped using the typical dark triad measure of 12-item scale developed by Jonason and Webster (2010). Sample items include “I tend to manipulate others to get my way”; “I tend to want others to pay attention to me”; “I tend to be callous or insensitive.” The scale’s internal consistency was very good (α = 0.98).

Unethical pro-organizational behaviors

These were measured through self-reports using a six-item scale developed by Umphress et al. (2010). Sample items include “If it would help my organization, I would exaggerate the truth about my company’s products or services to customers and clients”; “If it would help my organization, I would misrepresent the truth to make my organization look good.” Cronbach’s alpha reliability for this scale was good (α = 0.88).

Careerism

We used a 23-item scale designed by Feldman and Weitz (1991) to measure careerism. Sample items include “It’s hard to get ahead in an organization on sheer merit alone”; “Who you know is more important than what you know.” Cronbach’s alpha reliability for this measure was very good (α = 0.94).

Counterproductive work behaviors

CWBs were measured through participants’ peers with a 14-item scale developed by Aquino, Lewis, and Bradfield (1999). The scale comprises two dimensions: CWBs directed towards the organization as a whole (CWBO) and CWBs targeted towards individuals (CWBI). Sample items include “Intentionally arrived late for work” (CWBO) and “Purposely ignored his/her supervisor’s instructions” (CWBI). Because data on both dimensions of CWBs were collected from co-workers, we performed confirmatory factor analysis (CFA) to establish discriminant validity. Our analyses indicated that the single-factor aggregate CWB model demonstrated a better fit (χ2 = 176.69, df = 48, p < 0.001; CFI = 0.98, NFI = 0.98, GFI = 0.93, RMSEA = 0.09) than the two-factor model (χ2 = 591.78, df = 67, p < 0.001; CFI = 0.93, NFI = 0.92, GFI = 0.75, RMSEA = 0.16). Hence, for our analyses, we used the aggregate CWB scale, which yielded excellent internal consistency (α = 0.98).

Control variables

One-way ANOVAs revealed variations in our dependent variables in relation to the organization, organization type, department, designation, gender, age, education, and work experience (with the current organization and overall). Four dummy-coded variables were created to control for the effects of organization, organization type, department, and designation. They were a total of six organizations participating in this inquiry. Responses collected from organization 5 (i.e., district court) were significantly different for our dependent variables in comparison to all other organizations resulting in the creation of a dummy code for this organization (1 = Organization 5 and 0 = all other organizations). Similarly, there were three categories of organization type (i.e., 1 = Government, 2 = Semi-Government and 3 = Private). Based on the results of one-way analysis of variance, we created a dummy code for government versus others. As for the control variable departments (i.e., 1 = HR/Administration, 2 = Finance/Audit, 3 = IT/Telecom, 4 = Operations/Logistics/Procurement, 5 = Marketing/Sales), HR/Administration showed significant variation so we a created a single dummy-coded variable for department (1 = HR/Administration and 0 = all other departments). Lastly, regarding designation for employees (1 = Technical/Frontline, 2 = Entry level management, 3 = Middle level management, 4 = Top level management), we created a dummy code variable for Technical/Frontline in comparison to all other designations t (1 = Technical/Frontline and 0 = all other designations). These four dummy-coded variables were included as covariates in our analyses. Because both gender (0 = male, 1 = female) and education (0 = bachelor’s and below, 1 = master’s and above) had only two categories, we could directly control for them. Finally, the continuous variables of age, experience with the present organization, and total work experience were directly controlled for in all subsequent analyses.

Confirmatory factor analysis

We ran a CFA to establish the discriminant validity of variables measured in the same time wave using the same source. Following Anderson and Gerbing (1988), we conducted CFA by pairing a two-factor model with a single-factor model. Because psychological entitlement and manipulative personality were measured at T2 through self-reports, we estimated a two-factor model and compared it with a single-factor model. The CFA findings indicated that the two-factor model yielded a better fit (χ2 = 472.99, df = 158, p < 0.001; CFI = 0.96, GFI = 0.88, NFI = 0.95, RMESA = 0.08) than the single-factor model (χ2 = 3118.56, df = 169, p < 0.001; CFI = 0.65, GFI = 0.35, NFI = 0.64, RMESA = 0.24). Additionally, unethical pro-organizational behaviors and careerism were measured through self-reports at T3. Thus, we conducted another CFA to ascertain the discriminant validity of unethical pro-organizational behaviors and careerism as a two-factor model in comparison with that of a combined model. The CFA results showed that the two-factor model (χ2 = 1015.28, df = 269, p < 0.001; CFI = 0.90, GFI = 0.82, NFI = 0.87, RMESA = 0.09) had a better fit than the single-factor model (χ2 = 2287.67, df = 313, p < 0.001; CFI = 0.71, GFI = 0.61, NFI = 0.72, RMESA = 0.14). Finally, to demonstrate the acceptability and fit of our hypothesized six-factor measurement model, we ran a series of alternative measurement models. None of these alternative models yielded a better fit (see Table 1).

Table 1 Confirmatory factor analysis for hypothesized model and alternative measurement models
Full size table

Results

Table 2 presents the descriptive statistics, bivariate correlations, and Cronbach’s alpha reliabilities of the key study variables. OI was positively correlated with psychological entitlement (r = 0.37, p < 0.01), unethical pro-organizational behaviors (r = 0.46, p < 0.01) and CWBs (r = 0.16, p < 0.01), and negatively correlated with careerism (r = − 0.37, p < 0.01). Similarly, psychological entitlement was positively related to unethical pro-organizational behaviors (r = 0.63, p < 0.01), careerism (r = 0.16, p < 0.01), and CWBs (r = 0.41, p < 0.01).

Table 2 Means, standard deviations, correlations, and reliabilities for key study variables
Full size table

We utilized Preacher and Hayes’ (2004) bootstrapping macro to test for direct, mediation, moderation, and moderated mediation effects. For the moderation effects, we plotted the significant interactions based on simple slope analyses for mean ± 1 SD. Table 3 (a, b, and c) shows the results for both the direct and mediation effects hypotheses (H1 and H2). OI was significantly and positively related to unethical pro-organizational behaviors (B = 0.29, t = 5.57, p < 0.001) and negatively related to careerism (B = − 0.40, t = − 8.38, p < 0.001). No significant relationship was observed with CWBs (B = − 0.08, t = − 1.04, p = ns) hence partially supporting H1. Supporting H2, OI had an indirect effect through psychological entitlement on all three variables measuring unethical work behaviors: unethical pro-organizational behaviors (indirect effect = 0.16, p < 0.001); careerism (indirect effect = 0.06, p < 0.001); and CWBs (indirect effect = 0.11, p < 0.01). Further supporting H2, the bootstrapped 95% confidence intervals around the indirect effects did not include zero for unethical pro-organizational behaviors (0.10, 0.24), CWBs (0.04, 0.19), or careerism (0.02, 0.11).

Table 3 Mediated regression analysis results
Full size table

H3 predicted that the positive relationship between OI and psychological entitlement would be contingent on manipulative personality. The findings in Table 4 underline that the OI × manipulative personality interaction was significant for psychological entitlement (B = 0.16, SE = 0.03, p < 0.001). The bootstrap findings corroborate that OI had conditional direct effects on psychological entitlement at differing levels of the moderator, specifically, in the case of high manipulative personality (see Table 4). Next, we formulated the interaction for mean ± 1 SD.

Table 4 Hierarchical moderated regression analysis
Full size table

Supporting H3, Fig. 2 illustrates that the positive relationship between OI and psychological entitlement was more pronounced (and positive) when manipulative personality was high (β = 0.67, t = 11.10, p < 0.001), whereas it became nonsignificant for low manipulative personality (β = 0.08, t = 0.79, p = ns).

Fig. 2
figure 2

Interaction effects of organizational identification and manipulative personality on psychological entitlement. N = 306. MP manipulative personality. Slope for high manipulative personality (β = 0.67, t = 11.10, p < 0.001). Slope for low manipulative personality (β = 0.08, t = 0.79, p = ns)

Full size image

Finally, Table 5 presents the findings for moderated mediation effects (i.e., H4), whereby manipulative personality moderates the indirect effects of OI on unethical work behaviors through psychological entitlement. We investigated the conditional indirect effects of OI on unethical pro-organizational behaviors, CWBs, and careerism at three different values of manipulative personality (mean ± 1 SD; see Table 5). The mediated effects of OI through psychological entitlement on unethical pro-organizational behaviors (indirect effect = 0.28, p < 0.001), CWBs (indirect effect = 0.19, p < 0.001), and careerism (indirect effect = 0.10, p < 0.001) were conditional upon mean + 1 SD of manipulative personality. Further supporting H4, the bootstrapped 95% CIs did not contain zero for unethical pro-organizational behaviors (0.19, 0.37), CWBs (0.07, 0.32), or careerism (0.03, 0.18). No such effects were found for the condition of mean − 1 SD of manipulative personality.

Table 5 Moderated mediation analysis
Full size table

Discussion

Since its inception, OI has frequently been emphasized as a fundamental construct providing crucial insight into how people behave and develop different attitudes and behaviors (Ashforth & Mael, 1989; Hogg & Terry, 2000). Mainstream literature emphasizes the positive outlook of OI and continues to expand rapidly. However, in comparison to this voluminous research, significantly less focus and energy has been devoted to investigating the unfavorable effects or malevolent side of OI (Blader et al., 2017; Conroy et al., 2017). In addition, the OI literature has failed to explain the processes and boundary conditions for OI leading to positive or negative consequences (Blader et al., 2017; Brown, 2017; Conroy et al., 2017).

This study answered the call for more empirical research on the malevolent side of OI by investigating how unethical work behaviors emerge from OI through the intervening mechanism of psychological entitlement. By considering manipulative personality, we also illustrate that individual-level factors can be an important boundary condition for these mediation effects. In general, we found strong support for our hypothesized model (see Fig. 1).

The study’s findings suggest that the indirect relationship between OI and unethical work behaviors via psychological entitlement was stronger for highly manipulative individuals. These results extend related research by Umphress and Bingham (2011) in at least two ways. First, Umphress and Bingham (2011) are silent on the nature and types of neutralization mechanisms (Bandura, 1999) when examining the malevolent side of OI. Our study supports the process view by highlighting psychological entitlement as an important process mediating the relationship between OI and unethical work behaviors. Second, in theorizing on the negative impact of OI, Umphress and Bingham (2011) only consider unethical pro-organizational behaviors. Our study explores a wider range of unethical work behaviors, including self-centered unethical behaviors.

Apart from these important theoretical contributions, another key strength of our study was its distinctive time-lagged research design, with data collected in three time waves from independent sources (i.e., self-peer dyads). Employing multi-wave and two-source data decreases the risk of mono-method bias, particularly when examining complex moderated mediation models. Additionally, we relied on bootstrapping to test mediation, moderation, and moderated mediation (Preacher & Hayes, 2004). Finding strong support for the direct, indirect, and conditional indirect effects highlights the prominence and statistical power of our approach for testing complex moderated mediation models.

Limitations and future research directions

Although this study makes unique theoretical and empirical contributions, it is not free of limitations. While we used temporally segregated data, the study’s design cannot be considered longitudinal since none of the core variables was measured repeatedly across all three time waves. Temporal segregation between data on the independent, mediating, and criterion variables was intended to reduce the potential for common method bias (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). However, other forms of error could have emerged from segregation for which we did not control. Since a lengthy time lag can mask the theoretical relationships under examination, we decided to use a relatively short time lag of 3 weeks. One possible weakness in our adopted design is transient error, which is produced by longitudinal variations in respondents’ mood or feelings or in the efficiency of the information processing mechanisms they use to answer questions (Becker, 2000; Schmidt, Le, & Ilies, 2003). We could have statistically controlled for this transient error by measuring all the variables in all three time waves. Also, contrary to our hypothesis, the direct effect of OI on CWBs was nonsignificant. However, all other hypotheses were supported, including for CWBs as an outcome variable, which mitigates concern over the lack of support for H1b.

Given these limitations, future research should replicate this study’s results using a purely longitudinal design across different contexts. Future studies should also consider including other mediating and moderating mechanisms to fine-tune our understanding of the processes and conditions for OI becoming deleterious. For example, perceived organizational politics and leader-member exchange can be considered as potential moderators in the OI–psychological entitlement relationship.

Prior studies have investigated the direct effects of important demographic variables such as gender and age on job outcomes. For instance, females are less likely to engage in CWBs (Gonzalez-Mule, DeGeest, Kiersch, & Mount, 2013) and feel less deserving than males (Ciani, Summers, & Easter, 2008; Major, McFarlin, & Gagnon, 1984). Moreover, younger employees feel more entitled, are more likely to engage in CWBs, and demonstrate more unethical pro-organizational behaviors than their older peers (Gruys & Sackett, 2003; Mano-Negrin & Kirschenbaum, 1999; Xie & Johns, 1995). The main effects of these demographic variables on the dependent variables were outside the scope of this study’s objective, although we controlled for them in our analyses. It would also be interesting for future studies to examine the role of gender and age as boundary conditions in the OI–entitlement and OI–unethical behaviors relationships (considering CWBs, careerism, and unethical pro-organizational behaviors). We also created interaction terms of gender and age with OI and ran moderating analysis to test these two demographic variables as moderators in OI–entitlement, OI-CWBs, OI–unethical pro-organizational behaviors, and OI–careerism. None of these interactions were statistically significant. Nonetheless, future studies should investigate the pertinent role of these variables in explaining when and for whom OI shows more deleterious effects.

Another key question is whether our results would hold across cultures. The collectivistic nature of Pakistani culture (Hofstede, 1980) may impact the generalizability of our results in two important ways. First, our respondents may be expected to report high OI scores as individuals from collectivistic cultures tend to report more group/organizational memberships when asked to respond to “I am...” statements, compared to individuals from more individualistic cultures (Trafimow, Triandis, & Goto, 1991; Triandis, McCusker, & Hui, 1990). This may have led to an inflated correlation between OI and outcomes for our study.

Second, in a recent meta-analytic review, individualism was found to have significant negative effects on the relationship between OI and outcomes (B = − 0.20, p < 0.05), suggesting that the OI–outcomes relationship is stronger in collectivistic than in individualistic cultures (Lee, Park et al., 2015). These findings demonstrate that the size of the OI effect may be weaker in individualistic cultures. However, we believe that our findings are generalizable to other collectivistic countries, such as United Arab Emirates, India, China, and Japan. Future studies should replicate our approach in other sectors and cultures. In addition to unethical work behaviors, it would also be interesting to investigate other undesirable attitudinal and behavioral outcomes of OI, such as turnover intention, psychological detachment from work during off-job time, and other well-being outcomes.

Finally, we treated manipulative personality as a unified construct that shares the tendencies of underlying traits, such as manipulation and callousness. Also, given the pivotal role of psychopathy, narcissism, and Machiavellianism as distinct traits, short measures such as the Dirty Dozen scale fail to capture their uniqueness, focusing instead on assessing the shared callousness and manipulation underlying all three traits. In short, the literature on manipulative personality has largely ignored the possibility of the separate manipulative traits being multi-dimensional, representing heterogeneous sets of characteristics (Monaghan, Bizumic, & Sellbom, 2016). Researchers should consider using measures that comprehensively assess all the different aspects of these three manipulative traits.

Practical implications

Our study’s findings have crucial implications for managers and organizations. First, our findings suggest that high identification creates unethical work behaviors through instigating feelings of psychological entitlement. Managers must be aware that overidentification and extreme forms of identification should be promptly and appropriately addressed, since those employees may otherwise feel entitled to engage in unethical work behaviors. We propose open and timely interaction to inform overidentified workers that the organization rewards employees based on performance, rather than for illegitimate and nongenuine reasons.

Second, our findings underscore that manipulative personality acts as a moderator in the relationship between OI and psychological entitlement, such that it becomes aggravated for highly manipulative individuals. Hence, companies should be especially cautious in hiring individuals with manipulative tendencies: as our study illustrated, these self-absorbed and callous individuals will likely use their identification as an excuse to exploit their organization. Additionally, we propose the development of a code of ethics prescribing what behaviors are acceptable and which are not beneficial to organizations. Along the same lines, policies and systems should be implemented that hold all organizational members accountable for their actions, and no employee should be exempted, including those with strong OI. For example, employee rewards and compensation should be determined through a system of fairness and merit, rather than granting special favors or preferential treatment to a few “entitled” individuals. Finally, promotions and other resource distributions should be based on performance, competence, and experience, rather than personal and professional connections.

Conclusion

Our research offers unique insight by examining how and for whom OI has deleterious consequences. In a world where the usefulness of positive constructs such as OI has been overemphasized, we contend that high OI might also harbor an often ignored but particularly gloomy reality. This study aimed to provide empirical support for the existence of the negative effects of OI; in doing so, we created foundations for future research to further untangle the malevolent side of identification.