Scrooge Posing as Mother Teresa: How Hypocritical Social Responsibility Strategies Hurt Employees and Firms

Abstract

Extant research provides compelling conceptual and empirical arguments that company-external (e.g., philanthropic) as well as company-internal (i.e., employee-directed) CSR efforts positively affect employees, but does so largely in studies assessing effects from the two CSR types independently of each other. In contrast, this paper investigates external–internal CSR jointly, examining the effects of (in)consistent external–internal CSR strategies on employee attitudes, intentions, and behaviors. The research takes a social and moral identification theory view and advances the core hypothesis that inconsistent CSR strategies, defined as favoring external over internal stakeholders, trigger employees’ perceptions of corporate hypocrisy which, in turn, lead to emotional exhaustion and turnover. In Study 1, a cross-industry employee survey (n = 3410) indicates that inconsistent CSR strategies with larger external than internal efforts increase employees’ turnover intentions via perceived corporate hypocrisy and emotional exhaustion. In Study 2, a multi-source secondary dataset (n = 1902) demonstrates that inconsistent CSR strategies increase firms’ actual employee turnover. Combined, the two studies demonstrate the importance of taking into account the interests of both external and internal stakeholders of the firm when researching and managing CSR.

Introduction

Your employees are your company’s real competitive advantage. They’re the ones making the magic happen—so long as their needs are being met. (Richard Branson)

Corporate social responsibility (CSR) has become a mainstream practice, and almost every company includes CSR activities addressing a large variety of causes in its annual planning (Cone 2015). As CSR engagement is approaching obligation, skepticism and cynicism regarding companies’ reported CSR efforts is also on the rise (Leonidou and Skarmeas 2015). One common concern is that companies only invest in symbolic CSR for advertising reasons, for example, to create perceptions of sustainability, commonly referred to as “greenwashing” (e.g., Brammer et al. 2015; Leonidou and Skarmeas 2015; Donia and Sirsly 2016). “Greenwashing” demonstrates that a company maintains two types of ideologies—one for external use and one for internal (Brunsson 1993)—evoking impressions of corporate hypocrisy, which is defined as perceptions of inconsistencies between an organization’s actions and its collective identity (Aqueveque and Encina 2010; Brunsson 1993; Cartwright and Holmes 2006). While research has examined customer and market perceptions of corporate hypocrisy (Wagner et al. 2009), employee perceptions of how an employer’s CSR investments are distributed have not been explored. Our first research question, therefore, asks: Do employees perceive a double standard (i.e., corporate hypocrisy) when their employers invest differentially across CSR causes?

A well-known example of such inconsistent distribution of CSR investment is Walmart. The company spent $1 billion on external CSR initiatives in 2013 (Walmart Stores 2013). Instead of being praised for its massive CSR engagement, Walmart was accused of being “Scrooge posing as Mother Teresa”¹ for paying low wages to workers, thus requiring many to use food stamps and governmental assistance. Public perceptions of Walmart’s hypocrisy led to a large boycott, lost sales, and a damaged reputation. Walmart heavily invested in what we call external CSR activities (ECSR) which are directed at societal or environmental causes and primarily benefit external stakeholders (Brammer et al. 2007). At the same time, Walmart largely neglected internal CSR activities (ICSR) defined as those CSR investments that are directed at enhancing employees’ physical and psychological working environment (e.g., Lindgreen et al. 2009; Turker 2009; Verdeyen et al. 2004). While this example shows the public’s outrage at Walmart’s inconsistent CSR strategy, employee micro-level dynamics in such situations have not been investigated. Therefore, our second research question asks: How are work-related attitudes and behaviors of employees impacted under different conditions of (in)consistent CSR strategies?

Our research questions are important to answer as employees, due to their insider status, are likely to be the first stakeholder group to become aware of CSR inconsistencies. Comparing their everyday work experience with how the company appears in the public media, employees can quickly discover a company’s potential double standards (Brunsson 1993). Ensuing hypocrisy perceptions among employees can be particularly damaging to the firm because employees play a central role in extracting value from a firm’s CSR activities (Korschun et al. 2014; Rupp et al. 2013; Vlachos et al. 2010) as they are considered a highly credible source of CSR information for company-external stakeholders (Dawkins 2004; Morsing et al. 2008).

We seek to answer our research questions by first developing a conceptual model based on social and moral identification theory, and accounting for the multi-motive aspect of employee interpretations of and reactions to CSR. More specifically, our model identifies the micro-level dynamics of employee hypocrisy perceptions resulting from inconsistent CSR investment, driving emotional exhaustion and increasing employees’ intentions to quit and real employee turnover (Fig. 1).

Fig. 1

Overview of empirical studies and hypothesis tests

We test our conceptual model in two studies to reveal effects of inconsistent CSR strategies on employees’ subjective perceptions and behavioral intentions (Study 1) as well as on objectively measured CSR activities and real employee turnover (Study 2). Study 1 examines the psychological mechanisms underlying the employee-level impact of CSR strategies with larger external than internal CSR efforts. We generate data from a cross-industry employee survey (n = 3410) and conduct our serial mediation analyses using both difference scores and polynomial regressions, finding convergent support for our hypotheses. In Study 2, we test whether the demonstrated attitudinal outcomes (i.e., turnover intentions in Study 1) are also observable in behavioral outcomes (i.e., objectively measured employee turnover). We use a multi-source secondary dataset (n = 1902) and a lagged-variable regression model accounting for the nested data structure to demonstrate that employee turnover significantly increases as external CSR (ECSR) outweighs internal CSR (ICSR).

Our research contributes to the literature by explaining and empirically demonstrating, for the first time, how and when differential CSR investments across internal and external causes evoke negative attitudinal and behavioral consequences among employees. With our two studies, we respond to recent urgent calls for insights into why and how CSR affects employees (Glavas and Kelley 2014; Glavas 2016; Wang et al. 2016). Despite the growing body of research on employees’ reactions to and perceptions of CSR, many questions relating to how CSR affects employees remain unanswered (Edwards and Kudret 2017; Glavas and Kelley 2014; Glavas 2016; Vlachos et al. 2017; Wang et al. 2016).

Conceptual Framework and Hypotheses Development

Companies have to take into account the needs of multiple stakeholders when allocating budgets and making investments. As not all stakeholder groups can be served equally, this might lead to necessary trade-offs (Brunsson 1993). In the following, we will argue that CSR strategies with larger levels of ECSR compared to ICSR evoke corporate hypocrisy perceptions and trigger detrimental effects in employees (i.e., emotional exhaustion and intention to quit). We focus on emotional exhaustion and intention to quit as both have been identified as consequences of organizational hypocrisy by previous research (Kouzes and Pozner 1995; Philippe and Koehler 2005).

Companies seeking to engage in CSR must assign budgets, efforts, and policies to activities directed toward either internal (employee) or external stakeholders (customers, community, or larger society), or both. CSR activities can address potentially conflicting stakeholder needs, which often force trade-offs in investments (Raghubir et al. 2010) that will necessarily affect stakeholder perceptions. As a firm’s primary internal stakeholders, employees perceive multiple CSR activities (El Akremi et al. 2015) and react differently to different facets of CSR (Edwards and Kudret 2017; Lee et al. 2012). Consistent with the internal–external stakeholder distinction (Brammer et al. 2007; Turker 2009; Verdeyen et al. 2004), we differentiate between internally and externally oriented CSR initiatives (De Roeck et al. 2014) and propose that CSR strategies can be either consistent or inconsistent between the two types. As previously defined, ECSR benefits societal or environmental causes (Brammer et al. 2007), whereas ICSR encompasses activities that go beyond legal requirements to include good working conditions, employee training, career opportunities, organizational justice, and family-friendly policies (e.g., El Akremi et al. 2015; Lindgreen et al. 2009; Turker 2009). We define a consistent CSR strategy as having congruent levels of ICSR and ECSR activities (i.e., both are high or low). As we take the employee perspective on CSR, we define inconsistent CSR strategies as reflecting incongruent levels of ECSR and ICSR activities with larger company-external than internal CSR investments. We thus conceptualize inconsistent CSR strategies favoring external stakeholders as strategies in which ECSR activities are perceived to be (Study 1) or actually are (Study 2) greater than ICSR activities.

The Impact of Inconsistent CSR Strategies on Employee Perceptions of Corporate Hypocrisy: A Social and Moral Identity Theory Perspective

A substantial number of studies have investigated the impact of CSR on employees, with the overwhelming majority indicating that CSR leads to positive employee-level outcomes (Brammer et al. 2007; Carmeli et al. 2007; De Roeck et al. 2014; El Akremi et al. 2015; Farooq et al. 2014; Hillenbrand et al. 2013; Hofman and Newman 2014; Jones 2010; Korschun et al. 2014; Lindgreen et al. 2009; Rupp et al. 2013; Turker 2009). These positive outcomes in employees are primarily driven by social and moral identity dynamics. May et al. (2015) have expanded the view of the positive impact of CSR on employees that has been predominantly explained by social identity theory (Tajfel 1978; Tajfel and Turner 1985) by integrating the theory of the moral self (e.g., Shao et al. 2008) with social identity theory and thereby introducing the concept of moral identification. Moral identification is defined as the perception of oneness or belongingness associated with an organization that exhibits ethical traits (May et al. 2015). Hence, in social and moral identity theory (May et al. 2015; Tajfel 1978; Tajfel and Turner 1985) an employee derives value from her membership in an organization because she believes her employer to be in good social and moral standing, which (at least in part) is interpreted as reflecting her own social and moral identity and standing.

Employees’ social identity and self-worth are enhanced by their organizational membership when employees perceive their employer as having high prestige and an attractive image (Ashforth and Mael 1989; Dutton et al. 1994; Pratt 1998). CSR enhances perceived external prestige and pride (De Roeck et al. 2016; Jones et al. 2014) and offers employees an avenue to derive value from being associated with a company in good social standing (De Roeck et al. 2014; Jones 2010; Kim et al. 2010; Korschun et al. 2014; Lamm et al. 2015). These mechanisms have been empirically investigated for both ICSR and ECSR (e.g., Brammer et al. 2015; Carmeli et al. 2007; De Roeck and Delobbe 2012; De Roeck et al. 2014, 2016).

Employees’ moral identity can be bolstered through the employer’s CSR activities when employees believe the moral values related to the CSR activities to be aligned with their own value systems. Moral identification is defined as the perception of oneness or belongingness associated with an organization that exhibits ethical traits (May et al. 2015). For example, when employees hold broad pro-social values, they seek these values in the firms they work for (Jones et al. 2014). Indeed, a 2012 Forbes survey² of current and prospective employees underscores the importance of moral identification and reveals that respondents desire “a sense of pride and fulfillment from their work, a purpose, and, importantly, a [company] whose values match their own.”

While positive effects in employees from CSR through social and moral identification are documented in the literature, we posit that this dynamic may turn negative when CSR activities communicate a potential misalignment with employees’ moral and social value systems. So far, the literature has focused on the misalignment between words and deeds of companies. That is, when firms say something different in their CSR policies and communications than they actually implement. This word-deed mismatch then evokes perceptions of hypocrisy (Wagner et al. 2009).

Particularly damaging in terms of social and moral identification dynamics, but not investigated so far, could be a deed–deed mismatch. That is, hypocrisy perceptions among employees may also arise when companies practice what they preach in one CSR dimension, but do so less in others. As positive identity dynamics require unequivocal moral and social standing and unambiguous interpretations, any incongruence among various implementations of CSR can destabilize the value derivation process among employees. This happens when a company demonstrates concern for the wellbeing of individuals outside of organizational boundaries (via ECSR engagement) but not within. If employees perceive inconsistencies between an organization’s actions and its collective identity (part of which is signaled to the public via ECSR activities), they might pause to question: “What is this organization really about?” (Dutton et al. 1994, p. 243). This disruption causes employees to reassess the match between the organization’s values and their own and conclude that the organization is setting double standards for private and public morals (Fein et al. 1990; Monin and Merrit 2010), i.e., that the organization “maintains two ideologies, one for external use and one for internal use” (Brunsson 1993, p. 4). As a result, we propose that inconsistent CSR strategies will lead to employees’ perceptions of corporate hypocrisy:

H ₁

Employees’ perceptions of an inconsistent CSR strategy (i.e., ECSR > ICSR) are positively associated with their perceptions of the employer’s corporate hypocrisy.

Hypocrisy Perceptions and Emotional Exhaustion as Mediators to Intention to Quit

When positive moral and social identity dynamics are disrupted and hypocrisy perceptions evoked, as discussed above, employees are likely to seek a way to disassociate themselves from the organization. As prior research demonstrates, perceiving an organization as hypocritical stresses employees and leads to intention to quit (Kouzes and Posner 1995; Philippe and Koehler 2005). How hypocrisy perceptions evoked by inconsistent CSR investments lead to greater intentions-to-quit can be argued based on social and moral identity theory (Dutton et al. 1994; May et al. 2015; Tajfel 1978) and how this process is mediated by emotional exhaustion is suggested by Rupp et al’s (2013) multi-motive model of employee reactions to CSR. More specifically, three employee motives (relational, heuristic, and deontic/moral) will shape employees’ social and moral identity dynamics in light of their employers’ CSR activities.

First, relational motives encompass employees’ need for self-esteem enhancement and the fostering of a favorable social identity as outlined thus far. Hence, employees have a strong basic need to identify with their organizations and thus are drawn to firms with a positive image or a value system that matches their own (e.g., Dutton et al. 1994), which can be also labeled as moral identification (May et al. 2015). But what is the identity of an organization that displays high moral standards to the outside and has much lower moral standards on the inside? These conflicting cues lead to a disruption of moral identification as employees realize that their organization is hypocritical. This process of disidentification puts an emotional strain on employees, as they expend cognitive and emotional resources to deal with the conflict between their own and the company’s value system. Membership in organizations with perceived negative character traits thus leads to depression and stress (Dutton et al. 1994) as employees struggle to reconcile the perceived hypocritical reality with the desired ideal (Foreman and Whetten 2002). An additional pressure is that a hypocritical company makes false promises to external stakeholders (e.g., a firm positions itself as socially responsible to external stakeholders while being irresponsible internally). Hence, employees might be additionally stressed because the aspect of the organization they partially identify with (high external prestige) is threatened. Evidence suggests that employees who perceive their employer as unable to live up to external promises become emotionally exhausted (Suh et al. 2011).

Second, Rupp et al. (2013) identify heuristic motives that encompass employees’ needs for predictability and certainty and determine their reactions toward a companies’ CSR strategy. CSR actions can have a heuristic value for the interpretation of current and the prediction of future treatment by the employer. Indeed, CSR is often regarded as a positive signal that an organization has high moral standards and can therefore be trusted (Homburg et al. 2013). However, inconsistent CSR strategies are indicative of an organization’s double standard (Brunsson 1993; Philippe and Koehler 2005). Therefore, employees will find it difficult to understand and make predictions about their employers’ morality and, more generally, future behavior. This uncertainty becomes a source of anxiety and potential emotional exhaustion (Greenberger et al. 1989). As Philippe and Koehler (2005) put it: “A decrease in consistency will be mirrored by an increase in the level of stress and may affect an organization member’s desire to remain” (p. 14).

Third, Rupp et al. (2013) argue that individuals have deontic needs as they care about and react to (in)justice because unfair treatment violates moral and ethical norms. As a consequence, employees will react negatively to a firm whose actions are perceived as socially irresponsible. Inconsistent CSR strategies are likely to cause employees to feel treated in an unfair way, violating their conviction that they (as internal stakeholders) should receive at least as much attention from their employer as firm-external stakeholders. Perceptions of unfairness may be caused by their first-party perceptions of distributive injustice (i.e., as individuals they are receiving less attentions from the company than firm-external stakeholders) or by their perceptions of their co-workers being treated in an unfair way. This may even cause an “us versus them” thinking (Bhattacharya and Elsbach 2002), further adding to the disidentification (Kreiner and Ashforth 2004) and the emotional strain caused by this dissonance.

In summary, we propose that inconsistent CSR strategies cause employees to perceive the organization as hypocritical, triggering a process of moral disidentification and, in turn, emotional exhaustion. Taking these arguments into account, we thus propose:

H ₂

Employees’ perceptions of an inconsistent CSR strategy (i.e., ECSR > ICSR) are positively associated with emotional exhaustion, mediated by employees’ perceptions of the employer’s corporate hypocrisy.

A growing body of research suggests that hypocrisy perceptions of leaders and organizations lead to employee turnover or turnover intentions (e.g., Greenbaum et al. 2015; Philippe and Koehler 2005). In the following, we will argue how emotional exhaustion at least partially mediates the path from employees’ perceptions of organizational hypocrisy evoked by inconsistent CSR strategies to employees’ intention to quit. We find support for this serial mediation in research that documents the pivotal role of employees’ moral identification in firms’ ability to retain qualified staff (May et al. 2015). In general, moral identification serves as a motivational impetus for employees to remain with their employing organization. However, based on moral self-theory (e.g., Aquino and Reed 2002) and the attraction-selection-attrition (ASA) framework (Schneider et al. 2000), May et al. (2015) develop boundary conditions for this positive effect of moral identification on employee retention. More specifically, employees with strong moral identification strive for self-consistency, i.e., seeking consistency between their own and the organizations values (Aquino and Reed 2002; Shao et al. 2008). As long as their organization continues to have a sound and consistent ethical profile, employees will want to retain organizational membership (May et al. 2015). These employees are likely to have chosen their employer based on matching values (May et al. 2015). If employees striving for moral identification become aware of their employer’s CSR inconsistencies (and thus perceive the employer to be hypocritical), they will try to reconcile the misfit between their own and the organization’s moral principles (e.g., Dutton et al. 1994; Glavas and Godwin 2013) as they place a high value on working for organizations they can identify with (e.g., Ashforth and Mael 1989; Dutton et al. 1994) and as they strive for self-consistency, i.e., a match between their personal and the organizations’ values (May et al. 2015). When employees cannot rationalize the gap between their values and the organization’s values, they become emotionally exhausted by this process, and leaving the employer will be a logical consequence to exit the strained relationship. In addition to turnover due to emotional exhaustion from unresolved moral identities, May et al. (2015) argue that moral identification has a regulatory effect on employees’ own (un)ethical behavior as the organization provides guidance on the right moral conduct. Employees who perceive their organizations as hypocritical might therefore be confused about how to correctly behave in ethical decision situations. This confusion and the resources employees have to expend to reflect on the ethically appropriate behaviors may well pose an additional emotional strain on them. As a consequence, employees might seek a way out of the organization. Therefore, hypocrisy perceptions evoked by an inconsistent CSR strategy are likely to lead to greater intention to quit due to the emotional exhaustion that the mismatch in values and a lack of moral guidance cause in employees. Furthermore, hypocrisy research underscores that espoused fundamental values (as communicated via external CSR, for example) evoke false expectations and that this results in a conflict that stresses employees and diminishes their capacity to continue their job (Kouzes and Pozner 1995; Philippe and Koehler 2005). Taking these arguments into account, we propose:

H ₃

Employees’ perceptions of an inconsistent CSR strategy (i.e., ECSR > ICSR) are positively associated with intention to quit, mediated by (a) employees’ perceptions of the employer’s corporate hypocrisy and (b) employees’ degree of emotional exhaustion.

We know from the theory of reasoned action and the theory of planned behavior (Ajzen 1991; Madden et al. 1992) that intentions and real behavior may very well deviate from each other due to various additional factors. Meta-analyses reveal that these models explain on average between 40 and 50% of the variance in intention, but only between 19 and 38% of the variance in behavior (Sutton 1998). In the context of our research, deviations between turnover intentions and real quitting may arise from various sources, such as a lack of outside options, risk-avoidance, potential attachment to colleagues, or mere inertia. The first to quit in cases of adverse working conditions are likely high-potential employees who have multiple outside options and are less bound to the employer. However, if adverse conditions (such as the mismatch between the employee’s and the organization’s values and the emotional exhaustion that this mismatch causes in employees) continue, employees will move beyond thinking about quitting and put these intentions into action.

Since much is made of CSR as a tool to retain high-quality employees, it is critical not only to assess our proposed logic on intentions but also on actual behaviors. The central question is: do the proposed detrimental effects of inconsistent CSR strategies only affect turnover intentions, or do these intentions lead to real quitting behavior? Answering this question is especially important as turnover is a major factor in organizational costs. For example, retaining a highly skilled employee who earns $100,000 by a sound CSR strategy translates into a $400,000 benefit (Sprinkle and Maines 2010). Given the importance of assessing our proposed logic at the self-reported intentional as well as objectively measured behavioral level, we conduct Study 2 and test whether inconsistent CSR portfolios lead to actual employee turnover.

Methods and Results

We conduct two studies to test our hypotheses. In Study 1, we focus on the dependent variable of turnover intentions and investigate whether, in line with hypotheses H₁–H₃, inconsistent CSR strategies that favor company-external stakeholders lead to employees’ perceptions of corporate hypocrisy, thereby causing emotional exhaustion and turnover intentions. Study 2 examines objectively measured CSR and turnover data. Observed behavioral dependent variables in management and marketing contexts have been argued to be critical due to the intention-behavior gap (Ajzen et al. 2004, 2009; Chandon et al. 2005; Morales et al. 2017; Podsakoff and Organ 1986; Sutton 1998). In this article, we remedy the problems of self-report survey research by showing that our model not only predicts self-reported intentions (Study 1), but also actual behavior (Study 2).

Study 1: The Mediating Role of Employees’ Perceived Corporate Hypocrisy and Emotional Exhaustion

Data Collection and Sample Composition

We distributed an online survey to a cross-sectional panel of 69,799 individuals living in Germany, asking them to rate their current employers. Of these, 5488 (7.9%) participants completed the survey, of which 2022 respondents (36.8%) did not qualify for inclusion because they were either unemployed, retired, free-lancing, or self-employed. Another 56 cases were excluded due to missing values on dependent measures, resulting in an effective sample of 3410 respondents. The overall response rate of 7.9% is comparable to those from other online surveys in CSR studies (e.g., Berry and Junkus 2013: 6%; Jin et al. 2013: 7%). The sample covers a broad range of different company- (size and industry), participant- (gender, age, educational level, household net income), and work-related characteristics (work experience, tenure, and leadership position). The average participant was 35.3 years old (SD = 10.3 years), and most participants had worked for their employers between 1 and 5 years (44.7%). As 71.3% of respondents were women, we conducted additional robustness checks we report later on (for a complete overview of all demographics, please refer to Table W1 in Web Appendix).

Measures

We utilized or adapted existing multi-item scales with a 7-point Likert scale (1 = “strongly disagree”; 7 = “strongly agree”). All multiple-item measures show adequate reliability (all α ≥ .87), and all constructs display adequate discriminant validity (Fornell and Larcker 1981). We examined the psychometric properties of all multi-item constructs through a confirmatory factor analysis (CFA) in Mplus 7. The analysis indicates satisfactory global fit measures: confirmatory fit index (CFI) = .97, Tucker-Lewis index (TLI) = .97, and root mean square error of approximation (RMSEA) = .05. Additionally, we compared our measurement model (that includes the multi-item constructs of ICSR, ECSR, corporate hypocrisy, and emotional exhaustion) to (a) a three-factor model with a joint measure of ICSR and ECSR (CFI .91, TLI .89, RMSEA .10) and to (b) a two-factor model that includes next to the joint CSR measure also a joint measure of hypocrisy and emotional exhaustion (CFI .67, TLI .62, RMSEA .17). These models display worse fit indices which supports our measurement model. “Appendix 1” lists all items for our core constructs and their psychometric properties. Table 1 summarizes the constructs’ descriptive statistics and correlations.

Table 1 Study 1: Descriptive statistics and correlations

ICSR and ECSR

ICSR perceptions were measured by a 6-item scale (α = .89) from Lindgreen et al. (2009), with respondents rating their employer from 1 = “I do not agree at all” to 7 = “I fully agree” with items such as “My employer incorporates interests of employees in business decisions” or “…provides employees with salaries that properly and fairly reward their work.” ECSR perceptions (ECSR; r = .74) were assessed through a shortened version of Menon and Kahn (2003) with statements such as “My employer donates part of the profits to good causes.”

Inconsistent CSR Strategies

In line with past contributions exploring the effect of a difference between symbolic and substantive CSR actions (Walker and Wan 2012), we computed our measure of inconsistency by subtracting the value of ICSR from the value of ECSR. Greater positive values on the measure indicate larger inconsistencies favoring ECSR over ICSR investments. In turn, negative values indicate inconsistencies of the ICSR > ECSR type, and in the case of consistent strategies the difference score takes the value of 0.

Employee Outcomes

To capture the three employee attitude constructs of interest, we use the perceived corporate hypocrisy scale (α = .92) from Wagner et al. (2009). An example item is “My employer pretends to be someone that he is not.” To measure emotional exhaustion, we use a shortened scale (α = .93) from Maslach and Jackson (1981). An example item is “I feel burned out from my work.” We use an original scale to capture intention to quit (r = .74) by asking respondents how many times they had thought of quitting within the last 12 months, as research has shown that the frequency of thoughts of quitting adequately captures turnover intentions (Hom et al. 1992).

Control Variables

We include a comprehensive set of individual, company-specific, and work-related control variables based on past research. We control for job satisfaction measured by a straightforward item “I feel a great sense of satisfaction from my present work” (e.g., Christen et al. 2006), as it affects intention to quit (e.g., Liu et al. 2012). We control for age, gender, and work experience (Liu et al. 2012) and education (Batt and Colvin 2011), as they affect the general level of job satisfaction, turnover intention, and emotional exhaustion (Batt and Colvin 2011; Liu et al. 2012; Maslach and Jackson 1981). We measured work experience with an ordered, time-based categorical variable. Work-related control factors capture whether the respondent has a leadership position (marked by agreement with “I give out orders to other employees”) and his/her tenure (Taylor et al. 1996). We also control for firm size and industry, as both variables have been shown to affect employee attitudes and behaviors (e.g., Huselid 1995; Shaw et al. 2009).

Robustness Checks

Gender Differences

As men are under-represented in our sample (28.7% of all respondents), we perform a multiple group analysis (Muthén and Muthén 1998–2012) with the respondents’ gender as the grouping factor to check whether our model estimates remain robust. We conduct a Chi-square difference test between a constrained model and an unconstrained model for each path in our model. The model fit does not significantly improve by independently examining effects in two groups, except for one path: male employees react more strongly to an inconsistent CSR strategy and exhibit greater increases in perceptions of hypocrisy (β_male = .23, p < .01; β_female = .13, p < .01). An equal gender distribution in our sample would thus likely have amplified our findings. This finding changes the magnitude of the result, not the valence of the relationship, and we include gender as a control variable.

Non-response Bias

As we have no archival data on non-respondents, we used a technique suggested by Rogelberg and Stanton (2007) to check for non-response bias. We compare early respondents (n = 1981) to late ones, who answered after a reminder message (n = 1429). The two groups show no significant differences on various socio-demographic measures (age, gender, income, education), job-related factors (tenure, leadership position), and key outcome variables, such as corporate hypocrisy perceptions, emotional exhaustion, and intention to quit. Notable exceptions were age (M_early = 34.6, SD = 10.2 vs. M_late = 36.4, SD = 10.6, t (3109) = −4.77, p < .01), education, and tenure. Comparison of the distribution on the education categories (see Web Appendix Table W1) shows that early respondents seem to have a higher education than late respondents (Pearson’s χ² (3, N = 3347) = 9.52, p < .05). Hence, we ran our hypothesized model with education as control variable. However, education did not exhibit any significant effect on our dependent variables corporate hypocrisy, emotional exhaustion, and intention to quit (all β range between −.01 and −.02, at p > .10). Hence, we did not include education as a control variable in the final model. Furthermore, early respondents tend to have a significantly shorter tenure than late respondents (Pearson’s χ² (6, N = 3410) = 32.59, p < .01). Due to the results of comparing early with late respondents, we implemented age and tenure as control variables. Please refer to Web Appendix W2 for a full reporting of all statistics for the non-response bias testing. This finding decreases the likelihood of response bias (Rogelberg and Stanton 2007).

Common Method Bias

To assess bias stemming from common method variance (CMV), we conduct a Harman’s single-factor test. The single factor does not explain more than 50% of variance (38.29%). We also estimate a confirmatory factor analysis (CFA) in which all first-order factors are restricted to load on a single factor. The results indicate a poor model fit (CFI = .53; TLI = .46; RMSEA = .21; SRMR = .15). We also use the Lindell and Whitney (2001) procedure, which is based on the idea that the degree of common method bias present in a dataset can be assessed by determining the correlation between key dependent variables in the framework and a marker variable, i.e., a variable that theoretically should be uncorrelated with the key dependents. This correlation can then be used to correct the correlation matrix for common method bias. Hence, we chose the correlation between the intention to apply at a specific company (a large retailer) and intention to quit, as well as emotional exhaustion and corporate hypocrisy (all r < .02) to correct the correlation matrix for common method bias. The largest correlation can be found between perceived corporate hypocrisy and employment intention at company X (r = .02); hence, we took this as a correction factor. The statistical significance of the correlations does not change indicating the absence of a common method bias. These findings indicate that common method variance is unlikely to inappropriately influence our results.

Hypothesis Testing and Results

We tested our hypothesized model by employing structural equation modeling (SEM) using Mplus 7 (Muthén and Muthén 1998–2012). For indirect effects, we used the bootstrapping approach implemented in Mplus 7 (i.e., MODEL INDIRECT). We also show the independent effects of ICSR and ECSR (Model 2) in addition to the main model capturing inconsistent CSR strategies (Model 1). Table 2 summarizes the standardized path coefficients and significance levels for both models.

Table 2 Study 1: The employee-level impact of inconsistent CSR strategies

Goodness of Fit

The measures of overall fit meet conventional standards and indicate adequate model fit (CFI = .98; TLI = .97; RMSEA = .05; SRMR = .02).

Inconsistent CSR Strategies on Perceived Corporate Hypocrisy

Supporting H₁, an inconsistent CSR strategy with larger external than internal efforts significantly increases employee perceptions of corporate hypocrisy (β = .11, p < .01). Results also underscore that assessing ICSR and ECSR independently can yield misleading results, as the effects from ICSR and ECSR show that both reduce hypocrisy perceptions when assessed separately (see Model 2 in Table 2).

The Mediating Impact of Corporate Hypocrisy and Emotional Exhaustion on Intention to Quit

In line with H₂, an inconsistent CSR strategy increases employees’ emotional exhaustion by enhancing their perceptions of corporate hypocrisy (β_{ECSR>ICSR → Hypo → EE} = .03, p < .01). We also find a direct increasing effect of inconsistent CSR strategies on emotional exhaustion (β = .03, p < .01). Hypocrisy perceptions also mediate the impact of inconsistent CSR strategies on intention to quit (β_{ECSR>ICSR → Hypo → IQ} = .02, p < .01) with emotional exhaustion being an additional mediator (β_{ECSR>ICSR → Hypo → EE→ IQ} = .01, p < .01). These findings support H_3a and H_3b.

Squared Difference Score

Reflecting existing research on greenwashing (Walker and Wan 2012), we use a subtractive difference score (ECSR–ICSR). This method is a relatively simple way to gain insights into the effects of differences between two related constructs such as ICSR and ECSR. When ECSR is high and ICSR is low, the difference score has a high value, signifying a high inconsistency tilted toward external stakeholders. When ECSR and ICSR are equal (i.e., in the case of a balanced strategy), the difference score is zero. When ICSR exceeds ECSR, the value of the difference score turns negative. This score is an important distinguishing characteristic to squared difference scores, which have also been used to capture differences (e.g., Homburg and Jensen 2007). The distinctive feature of squared difference scores is that they exhibit the same positive value in the both directions of the inconsistency (i.e., in cases where ICSR exceeds ECSR, the value of the variable would equal a case in which ECSR exceeds ICSR to the same extent). This makes sense if the direction of the difference is irrelevant. In our case, the direction is important and a difference score is able to capture this. However, the difference score is still a simplified representation of the real effects because it does not control for the main effects of the two variables and does not take into account nonlinearities in the effect of the difference. One easy way to explore whether nonlinearities exist is to integrate a squared difference score in combination with a “direction dummy” and an interaction term between the two. To capture the size of the difference, we subtract the item scores of ECSR from the item scores of ICSR and square that difference (Tsui and O’Reilly 1989). To account for the direction of the difference, we implement a dummy variable (which takes the value of “1” for inconsistent CSR strategies with ECSR > ICSR and “0” for balanced strategies and strategies with ECSR < ICSR). We also created a multiplicative interaction term between the squared difference score and the direction dummy. Results indicate that, indeed, the interaction term has a significantly positive effect on hypocrisy perceptions (β = .12, p < .01). In other words, if the inconsistency takes the form ECSR > ICSR, the effect on hypocrisy perceptions is significantly stronger than in cases of balanced strategies or cases in which ICSR is greater than ECSR. These results confirm that the effect of the difference on corporate hypocrisy perceptions is nonlinear in a way that it is significantly different for both directions of the indifference. Indeed, when we specify a regression that includes not the magnitude of the difference but merely a dummy that takes the value 1 in cases in which ICSR > ECSR, we find that the coefficient of this dummy variable is negative (β = −.15, p < .00). In such cases, it is recommended to also use more elaborate statistical techniques to validate the findings, such as polynomial regressions (Edwards and Parry 1993).

Polynomial Regressions

The structural equation model analysis and the robustness checks indicate that perceived corporate hypocrisy plays a central role for observing the negative effects of an inconsistent CSR strategy with larger ECSR than ICSR efforts. Following the literature examining congruence hypotheses (e.g., Jordan et al. 2013), we employ polynomial regressions (Edwards and Parry 1993) to explore the field of tension between ICSR and ECSR efforts in more detail. We conduct a regression analysis in which we first include the interaction term between ICSR and ECSR and in a second step their nonlinear quadratic effects and observe the changes in R². Whereas the interaction term between ICSR and ECSR does not significantly improve R², the inclusion of quadratic effects enhances the explanatory power of our model (p < .10). These results suggest that a simple multiplicative interaction analysis would likely not reveal the effects of interest.

The polynomial regression analyses entail estimating a quadratic regression model with perceived corporate hypocrisy as the dependent variable (Z) and both ICSR (X) and ECSR (Y) levels as the independent variables. The full polynomial equation is given as:

$$Z \, = b_{0} + b_{1} X \, + b_{2} Y \, + b_{3} X^{2} + b_{4} XY \, + b_{5} Y^{2} + e$$

Results from the polynomial regression (R² = .30) indicate that ICSR has a negative linear (b₁ = −.68, p < .01) but not quadratic (b₃ = −.01, p > .10) effect on corporate hypocrisy, that ECSR does not have a linear (b₂ = −.04, ns) but has a quadratic (b₅ = −.02, p < .05) effect, and that the interaction term is not significant (b₃ = .00, p > .10). The coefficients in this polynomial regression are not interpreted as in regular regression (Shanock et al. 2010). Instead, the slopes and curvatures of congruence and incongruence lines are examined to draw conclusions. Questions regarding the impact of ICSR and ECSR incongruence on perceptions of corporate hypocrisy explore the shape of the surface along various lines of interest. One of these is the line along which perceived levels of ICSR and ECSR are unequal.

Using the polynomial regression coefficients, we find that the slope of the incongruence line is negative (b₁ − b₂ = −.64, p < .01), with the curvature term not significant (b₃ − b₄ + b₅ = −.03, p > .10), indicating additive effects. In line with the previous analyses, the negative slope along the line of incongruence as it relates to perceived corporate hypocrisy indicates that the direction of the inconsistency matters. More specifically, perceived hypocrisy escalates as employee perceptions of discrepancy grow (i.e., ICSR becoming lower than ECSR), while perceived hypocrisy actually diminishes in the other direction of inconsistency (i.e., when perceptions of ICSR are higher than those of ECSR).

Similarly, the slope of the congruence line is negative (b₁ + b₂ = −.72, p < .01), with the curvature not significant (b₃ + b₄ + b₅ = −.02, p > .10), again indicating an additive relationship. The negative slope of this line suggests that perceived hypocrisy decreases as employees perceive more consistency between ICSR and ECSR. Combined, the results indicate that the directionality of the inconsistency determines hypocrisy perceptions and that ICSR is a benchmark category when judging CSR efforts.

Study 2: The Impact of Inconsistent CSR Strategies on Objective Employee Turnover at the Firm Level

Data Collection and Sample Composition

Employing organizational-level secondary data, our second study assesses the proposed effects on real employee turnover. Our data are drawn from several sources, allowing for a comprehensive assessment of independent and joint effects of ICSR and ECSR investments on employee turnover. CSR measures are based on Thomson Reuters’ ASSET4ESG database, which provides detailed information on companies’ CSR actions. It is one of the largest and most comprehensive sources of CSR data. A team of over 120 experienced analysts collects and standardizes information from a variety of publicly available sources such as sustainability/CSR reports, company websites, annual reports, proxy filings, and NGOs as well as news and media information, ensuring accuracy and comparability.³ Detailed information on the ASSET4ESG database can be found on the Thomson Reuters’ website under following link: http://financial.thomsonreuters.com/en/products/data-analytics/company-data/esg-research-data.html.

From ASSET4ESG, we initially drew 37,469 observations from 3746 unique firms based in 81 countries across 10 years (2002–2011). We excluded all cases with missing values for ICSR and/or ECSR and employee turnover, resulting in 1902 observations including 623 unique companies from 50 countries over a period of 10 years. (Tables W3 and W4 in Web Appendix describe the sample composition.) We include relevant control variables commonly regarded as potential explanatory variables in the context of employee turnover, which we extracted from Thomson Reuters’ Datastream database. Table 3 summarizes the descriptive statistics and correlations.

Table 3 Study 2: Descriptive statistics and correlations

Measures

ICSR and ECSR

To adequately capture the constructs of ICSR and ECSR, we reviewed the literature to select appropriate items from the large ASSET4ESG pool. We based our selection on stakeholder-based categorizations of ICSR and ECSR (Maignan and Ferrell 2000; Öberseder et al. 2013) and chose items on employee training, education, and work–life balance; donations; emission reduction; corporate citizenship (e.g., local community support); and product safety. “Appendix 2” provides an overview of the selected items. According to established definitions (Brammer et al. 2007; Turker 2009), only employee training, education, and work–life balance can be classified as ICSR. To validate that this is the case, we asked two research assistants (blind to the study’s objectives and research questions) to rate all selected ASSET4ESG CSR measures as relating to internal or external stakeholders. Each CSR measure was rated (on a scale of 1–7) on two items: “The activity primarily aims at improving the relationship with employees” and “The activity primarily aims at improving the relationship with firm-external stakeholders (e.g., customers, communities, society).” The raters confirmed that only the measures of employee training, education, and work–life balance related to employees, with an overall high degree of agreement between coders (correlation between ratings = .964, p < .001). Thus, the final ECSR measure represents environmental sustainability, corporate citizenship, donations, and product quality and safety and is formed as an equally weighted, standardized index. Similarly, the ICSR measure is an equally weighted, standardized index representing employee training and education, and work–life balance.

Inconsistent CSR Strategies

We take the same approach as in Study 1 to capture the degree of CSR strategy inconsistency.

Employee Turnover

The turnover rate, expressed as a percentage, is calculated by dividing the number of employees who left by the total number of employees at the beginning of the period.

Control Variables

Drawing on previous research, we control for multiple factors to rule out alternative explanations (Schmitt et al. 1991). As studies relating to employee turnover have frequently addressed sales growth, variability in sales, and R&D intensity (e.g., Huselid 1995), we include these measures as controls in our analysis. The level of sales and R&D expenditures are the natural log value in USD (provided by Datastream). Sales growth was calculated as the percentage of change in sales in 1 year as compared to the previous year. We include the variability of sales, since higher sales variability might increase the need for layoffs at least temporarily, translating into a higher turnover rate. Advertising expenditures serve as a proxy for firm visibility. ROA is used as a control for the firm’s overall financial health, which could explain significant variance in employee turnover. Salary, measured by staff costs, may also lower turnover and is included as a control.

Hypothesis Testing and Results

To test for the main and inconsistency effects of ICSR and ECSR initiatives on employee turnover, we specify a regression model that takes into account the nested structure of the data. We use the TYPE = COMPLEX procedure implemented in Mplus 7, which adjusts the standard errors for non-independence of observations. We use the individual firms as a cluster variable. The “TYPE = COMPLEX” command in Mplus is adequate when the interest does not lie in disentangling the effects on the firm level from the effects on the individual level but rather in testing the effect of overall population characteristics on certain outcome variables. Muthén and Satorra (1995) term this type of approach with clustered data aggregate analysis (i.e., a single regression line is estimated, but appropriate standard errors are estimated). In the case of our data, an analysis of variance with the company as independent and employee turnover as dependent variable does reveal a significant effect, indicating that employee turnover significantly varies across firms. Thus, taking into account the non-independence of observations is necessary.

In line with previous research, we lag the CSR-related independent variables by 1 year (Hawn and Ioannou 2014). We further include ROA as an independent variable, also lagged by 1 year. The regression models use employee turnover as the dependent variable. As in past research (Walker and Wan 2012), we calculate separate models for the effects of ICSR and ECSR and the inconsistency model (i.e., ECSR–ICSR).

Inconsistent CSR Strategies on Employee Turnover

Supporting an organizational-level effect of inconsistent CSR strategies, larger ECSR than ICSR efforts significantly increase employee turnover (see Table 4, Model 1; β = .27, p < .01). Additionally, we find that ICSR efforts significantly reduce real employee turnover (see Table 4, Model 2; β = −.30, p < .01), whereas ECSR has no significant impact (β = .14, p > .10). Table 4 summarizes the results of all estimated regression models.

Table 4 Study 2: The impact of CSR inconsistencies on real employee turnover

The results of Study 2 confirm that inconsistent CSR strategies lead to elevated employee turnover in firms and potentially endanger the company’s financial performance (e.g., Glebbeek and Bax 2004; Huselid 1995). Our results demonstrate the potentially misleading conclusions that may result from assessing ICSR and ECSR activities separately and underscore the critical role of ICSR.

General Discussion: Contributions and Implications

Numerous examples from business practice show that when companies cannot sustain an ethical stance internally, they may engage in ECSR activities for quick gains, such as positive publicity and enhanced public perceptions (Barnett and Salomon 2006). Our studies’ results underscore that an inconsistent CSR strategy is short-sighted, as it puts at risk the relationship between a company and its employees. We show for the first time that inconsistent CSR strategies evoke hypocrisy perceptions in employees, triggering detrimental effects such as emotional exhaustion and intention to quit. As a result, inconsistent CSR strategies result in greater organizational turnover rates. By examining the effects of ICSR and ECSR separately as well as holistically, our research is one of the first to investigate potentially negative employee-level effects of CSR strategies.

Theoretical Contributions

Our research responds to scholarly calls to investigate why and how CSR affects employees (Glavas 2016; Glavas and Kelley 2014; Wang et al. 2016). We contribute in three important ways to micro-level employee-related CSR and hypocrisy research.

First, by considering the multifaceted nature of CSR (Edwards and Kudret 2017; Lee et al. 2012; Wang et al. 2016), we achieve a more precise understanding of CSR’s differential effects on employees, adding to the emerging evidence that such differences exist and affect important employee outcomes (e.g., Glavas and Kelley 2014; Hofman and Newman 2014; Wang et al. 2016). By assessing separate and joint effects of ICSR and ECSR, our two empirical studies highlight that observing the impact of ICSR and ECSR independently may lead to erroneous conclusions. When set in relation to each other (i.e., when looking at inconsistencies in the form of a surplus of ECSR relative to ICSR), separately measured positive effects of ICSR and ECSR can turn into negative joint effects. Therefore, our research supports existing theoretical arguments that a company must consider potential payoffs between stakeholder groups when designing a CSR strategy (Bhattacharya et al. 2009). On the basis of our findings, we propose that ICSR might be treated as a benchmark category for all stakeholder groups to determine the credibility and consistency of a company’s overall CSR efforts. For future research, this strongly suggests taking a relative perspective on CSR dynamics, with ICSR representing the reference category. At the very least, our study suggests that any employee-related CSR research minimally includes ICSR as a control variable, in order to realistically investigate the effectiveness and employee-level reactions to certain (external) CSR strategies. Even for CSR research examining effects on customers, ICSR should be included for two reasons. First, customers have been shown to value employee-related CSR activities (Öberseder et al. 2013). Second, with the expanding importance of social media and a growing number of platforms where employees rate their employers (e.g., www.glassdoor.com), ICSR is becoming increasingly visible to external stakeholders.

Second, our research also contributes novel insights into how CSR affects employees. More specifically, we uncover the dynamics that underpin a potential “dark side” of CSR (Skarmeas and Leonidou 2013; Wagner et al. 2009; Yoon et al. 2006), which develops when employees must reconcile inconsistent CSR strategies of their employers. Rather than simply examining the boundary conditions of positive mechanisms, we introduce perceived hypocrisy as a mediator that explains how inconsistent CSR strategies lead to employees seeking to exit. We also theoretically argue how inconsistency can disrupt positive social and moral identification dynamics in employees (e.g., May et al. 2015) by evoking hypocrisy perceptions and therefore leading to emotional exhaustion, culminating into a higher intention to quit. Our research points to the need for additional studies into the micro-level dynamics and psychological mechanisms of CSR inconsistency. For example, a direct test of how hypocrisy perceptions affect employees’ relational, heuristic, and deontic/moral motives (Rupp et al. 2013) would be able to weigh the relative importance of those motives.

Third, we contribute to research on organizational hypocrisy (e.g., Brunsson 1993; Philippe and Koehler 2005) by showing that employee-level hypocrisy perceptions are evoked by inconsistent CSR strategies with larger external than internal investments that might perceived as signs of a double moral. Thereby, we extend CSR-related hypocrisy research that has shown so far that hypocrisy judgments can be evoked among customers by word–deed CSR misalignments (Wagner et al. 2009). Hence, future studies could address whether perceived corporate hypocrisy might be the mediating mechanism that captures perceptions of CSR inconsistencies across various stakeholder groups, such as suppliers and retailers, and across various CSR types (for example, are hypocrisy perceptions also evoked for imbalances on different external CSR strategies, e.g., when companies invest a lot into addressing societal issues, but neglect environmental ones).

Managerial Implications

An incomplete understanding of how CSR affects employees makes it difficult for organizations to design CSR strategies that satisfy both internal and external stakeholder demands. Whereas well-designed CSR strategies satisfy a multitude of relational, heuristic, and deontic needs in employees (Rupp et al. 2013) that also lead to the (moral) identification of employees with their company, only a third of all employees would recommend their company on the basis of its ethical behavior (Dawkins 2004). Hence, companies are struggling to formulate and manage CSR strategies that effectively take into account the needs of their employees. Our results indicate that such perceptions among employees may have serious consequences for the company. By empirically demonstrating the detrimental impact of inconsistent CSR strategies on employees, our studies contribute to a paradigm change in CSR management from externally focused to internally well-balanced CSR strategies. To avoid the equation “CSR minus HR = PR” (see footnote 3), firms should pay particular attention to the following action domains.

Sharing CSR Management Responsibilities

Responsibilities for CSR management are often split. As ICSR relates directly to the working environment of employees, the human resources department is often responsible for designing an ICSR strategy. As ECSR activities focus mostly on external stakeholders, such as enhanced customer awareness and a better reputation, the responsibility for the ECSR strategy often lies with the public relations or marketing department. Setting up a CSR department or creating a CSR committee that consists of members of all functions of the organization as a separate centralized policy unit will facilitate a holistic stakeholder-oriented approach to CSR that balances conflicting interests and ensures employees are fully represented in CSR strategies and tactics.

Setting CSR Objectives and Accounting for CSR Returns

The relative costs and benefits of ICSR and ECSR are hard to compare. Whereas ECSR investments are often motivated by quick returns and lower costs (Brammer and Millington 2008), ICSR investments often are much more substantial and costly, and thus tax a firm’s financial bottom line (Barnett and Salomon 2006). Hence, in marketing-driven organizations, the business case for ECSR might be much more attractive than the case for ICSR. The implementation of a comprehensive tracking system for the totality of ICSR and ECSR efforts that entails both costs and long-term indices could facilitate designing an informed and deliberate CSR strategy. Based on our results, such a tracking system should include psychological as well as objective employee-level indicators, such as perceived external and internal company image (including hypocrisy perceptions), emotional exhaustion and turnover rates. As burnout rates and costs due to staff turnover massively impact a company’s bottom line, including these indicators might turn the business case from ECSR toward ICSR.

Communicating CSR to and by Employees

The majority of companies communicate their ECSR activities to their employees. The reasons for communication reflect dual interests. Whereas 84% of companies want to increase employee satisfaction and commitment, 57% also have in mind the link between internal and external stakeholder outcomes—that is, “to create publicity and a good reputation through word-of-mouth advertising” (Birth et al. 2008, p. 185). Companies should be aware that when they have a weak ICSR strategy, communicating their good external deeds can backfire by fostering employees’ perceptions of double moral standards and hypocrisy. In turn, employees might convey their insider knowledge of the “real moral standards” of a company to their surroundings including customers and the greater public. As employees are considered a highly credible source of CSR information (and as more credible than company-issued messages), their negative word of mouth might render company-driven communication efforts ineffective. To have consistent communication across stakeholders, companies should emphasize their ICSR activities to their employees and take a highly credible “inside-out” communication approach (Morsing et al. 2008). Ultimately, this would mean integrating CSR strategy into human resource management to keep employees engaged.

Eliminating “Necessary” Hypocrisy in Terms of CSR

Brunsson (1993) provocatively proclaimed that, to a certain extent, organizational hypocrisy might be necessary to satisfy diverging stakeholder needs. In terms of CSR, we see the tendency that customers value ICSR and often perceive it as more genuine than ECSR (Öberseder et al. 2013; Schons et al. 2017). Our study highlights the perils of investing too heavily in ECSR from an employee perspective and underscores the relational power of ICSR. Thus, for investing more in ICSR, there seems to be no conflict of values and interests between external and internal stakeholders. Indeed, shifting budgets from an inconsistent toward a well-balanced strategy might be a winning proposition for all sides.

Limitations and Recommendations for Future Research

Like most research, our inquiry into the effects of inconsistent CSR strategies on employees has several limitations, indicating future research opportunities. We discuss three such limitations and potential future research directions.

First, our theoretical arguments suggest that employees’ moral (dis)identification plays a central role in causing detrimental employee-level effects. However, we did not specifically measure moral identification in our empirical study, indicating a clear future research direction. More specifically, based on research on moral identification (May et al. 2015), we expect employees’ individual need for moral identification to be a possible boundary condition. The effects observed in our study could be more pronounced for employees with a high need for moral identification. Furthermore, employees’ current state of moral (dis)identification is a possible additional mediator in our current model. Future research can specifically measure employees’ need for moral identification and current state of moral identification and investigate how employees with high versus low levels of (need for) moral identification react to inconsistent CSR strategies.

Second, our empirical studies lacked a true longitudinal design necessary to capture the full dynamics of inconsistent CSR strategies’ impact on employees. While our secondary data approach in Study 2 provided some temporal accounting by including 10 years and integrating lagged effects, our survey-based Study 1 was cross-sectional in nature and prevented us from observing changes over time. A particularly powerful approach to probing our findings in more detail would be a field experiment tracking employees in organizations that implement CSR strategies with varying levels of inconsistency. In an optimal setting, such a study would capture employees’ perceptions before and after the introduction of a new CSR program or communication. Such a field experiment would allow for investigation of short-, medium-, and long-term effects of inconsistent CSR strategies. We see two venues for future research that might be particularly fruitful. As argued above, inconsistent CSR strategies might affect employees with a high need for moral identification more negatively, potentially prompting faster and more likely exits from an employer (May et al. 2015). In line with the attraction-selection-attrition framework (Schneider et al. 2000), inconsistent CSR strategies might thus lead to an overall erosion of moral values across the organization. Future research can focus on testing this assumption by, for example, measuring moral identification of and ethical behavior by employees over time, carefully tracking the composition of the employee base as a consequence of the ethical stance of the company. As monitoring attitudes and behaviors over time always is a challenge to researchers, as a first step, scenario-based experimental research could monitor how consistent and inconsistent CSR strategies affect the intention to employ of applicants with high or low ethical values (i.e., focusing on the selection phase). Second, a field experiment or field study comparing different organizations might also address how different stages and strengths of employees’ organizational and moral identification might interact to produce greater or weaker perceptions of corporate hypocrisy from inconsistent CSR strategies and consequently, how inconsistent CSR strategies affect employees’ moral identification, emotional exhaustion, and intention to quit over time. While the heterogeneity in our survey-based sample in Study 1 provides some measure of confidence that our findings are robust across organizations and industries, a more in-depth examination of a single organization’s employees differentiated on the basis of moral identification could provide important boundary insights.

A third limitation of our study is that our data (particularly in Study 1) come from a developed economy, Germany, where employees expect and have rights to certain workplace protections and quality assurances. In contrast, in emerging economies, the sensitivity to hypocrisy might be much lower and behavioral outcomes significantly restricted. Future research can compare employees’ reactions to inconsistent CSR strategies in emerging versus developed markets.

Appendices

Appendix 1

See Table 5.

Table 5 Study 1: Overview of measures and scale items

Appendix 2

See Table 6.

Table 6 Study 2: Overview of ICSR and ECSR measures derived from the ASSET4ESG pool