Corporate social responsibility (CSR) refers to “context-specific organisational actions and policies that take into account stakeholders’ expectations and the triple bottom line of economic, social, and environmental performance” (Aguinis 2011, p. 855). During the past few decades, CSR has emerged as an inescapable priority for business leaders in most countries and an important source of organizational competitive advantage and long-term sustainability (Porter and Kramer 2011). Although CSR often has no immediate financial benefits (Orlitzky et al. 2003), it affects the sustainability of contemporary firms due to its effects on multiple stakeholders, such as employees, customers, and communities (Pfeffer 2010). It is suggested that employees are central stakeholders of CSR (Wood 1991; Wood and Jones 1995) and that individuals are actually responsible for the execution of CSR initiatives and bear most of CSR’s consequences (Aguinis and Glavas 2012). Thus, organizations cannot achieve their goals via rules and control measures alone—they also require their employees’ acceptance and support of those goals (Mossholder et al. 2011). Employees “carry the main burden of responsibility for implementing ethical corporate behaviour in the daily working life of the company, [thus] the achievement of those outcomes will depend largely on employee willingness to collaborate” (Collier and Esteban 2007, p. 20). Therefore, employee support for CSR initiatives is critical to their successful implementation, as well as to employee job satisfaction and more general work attitudes and behaviors.

Empirical research has indicated that employees may have different attitudes toward CSR, such as being committed, indifferent, or dissident (Rodrigo and Arenas 2008). However, past research (e.g., Rodrigo and Arenas 2008) did not differentiate between employees’ general attitudes toward CSR and employees’ attitudes toward organizations’ specific CSR initiatives. An individual’s attitude toward CSR in general (also called “individual CSR values”) is determined mainly by his or her upbringing and working and living environments (Rodrigo and Arenas 2008). In contrast, an individual’s attitude toward an organization’s specific CSR initiatives can be influenced by factors beyond general attitudes, such as perceived CSR fit, perceived motive, or the attributes of the organization’s CSR initiatives (Elliot 2011; Ellis 2009; Vlachos et al. 2013; Wagner et al. 2009), and how the organization’s CSR initiatives affect employees’ interests and well-being (Zappalá 2004). Hence, it is possible that, even when employees generally support CSR as a concept, they may not necessarily support their organization’s specific CSR initiatives.

Employees are often required to enact espoused organizational values in their work behaviors (for example, as part of their professional roles), even when they do not fully support these values (Gruys et al. 2008; Hewlin 2003). Under these circumstances, facades of conformity occur. Supplies–values fit theory posits that supplies–values misfit is likely to have negative employee consequences, such as compromised employee work performance, job dissatisfaction, and intention to quit or actual turnover (Edwards 2007). Hence, it is important for organizations to gain employee support for organizational policies and practices that demonstrate organizational values. Orlitzky and Swanson (2006) suggested that one way organizations can increase employee support for external CSR (CSR initiatives targeting external stakeholders) is to formulate and implement socially responsible human resource management (SRHRM) practices. SRHRM is a set of human resource management (HRM) practices adopted by organizations to affect employee attitudes and behaviors in order to facilitate the implementation of external CSR initiatives (Shen and Benson 2016). According to Shen and Benson (2016), SRHRM practices may include recruiting socially responsible employees; providing CSR training; and considering employees’ social contributions during promotion, performance appraisal, reward, and compensation. However, it remains unknown whether organizations are able to achieve the goal of increasing employee support for external CSR through the implementation of SRHRM practices. To fill this important literature gap, this research explores the relationship between SRHRM and employee support for organizational external CSR and the underlying mechanisms.

Theoretical Background

In this study, we define employee support for organizational external CSR as the extent to which employees perceive the organization’s external CSR initiatives to be appropriate and are willing to contribute to their implementation. External CSR initiatives mainly address issues such as poverty reduction (Jenkins 2005), climate change (Le Menestrel et al. 2002), environmental sustainability (Basu and Palazzo 2008), and community involvement—such as providing financial and in-kind assistance and contributions of time and expertise to the community (Zappalá 2004). Unlike internal CSR that is directed toward employees, external CSR may not always align with employees’ interests (Royle 2005). The HRM behavioral literature suggests that the effects of HRM practices on employee attitudes and behaviors vary with the attributes of HRM practices (Bowen and Ostroff 2004; Lengnick-Hall et al. 2009; Nishii et al. 2008). As noted by Nishii et al. (2008), “in order for HR practices to exert their desired effect on employee attitudes and behaviors, they first have to be perceived and interpreted subjectively by employees in ways that will engender such attitudinal and behavioral reactions” (p. 504). In line with this literature, we propose that SRHRM will increase employee support for organizational external CSR initiatives.

Building on the organizational climate literature (Reichers and Schneider 1990), we argue that SRHRM is related to employee support for organizational external CSR through the social and psychological process of organizational CSR climate. Organizational CSR climate is employees’ shared perceptions of how various stakeholders’ interests are addressed within and by the organization. Moreover, we extend the stakeholder theory (Freeman 1984) to explore the nuanced relationships of SRHRM and organizational CSR climate with employee support for organizational external CSR. In particular, the stakeholder theory postulates that a firm has multiple stakeholders who may have different or even conflicting interests in its activities. Thus, stakeholders’ reactions to the firm’s activities are likely to be subject to how their respective interests are met. With this logic, we argue that the effects of SRHRM and organizational CSR climate on employee attitudes toward organizational external CSR are contingent on employee perceptions of how they are treated by the organization.

The current research contributes to the literature in several ways. First, CSR is a unique topic that enjoys the growing attention of both researchers and practitioners. Research has already explored the effect of CSR on employee outcomes—for example, Brammer et al. (2007) and Hofman and Newman (2014) investigated the effect of CSR on organizational commitment, while Kim et al. (2010) studied the relationship between CSR and employee organizational identification. Further, Newman et al. (2015) examined the effect of CSR on employee work performance and organizational citizenship behavior. However, because CSR research is rooted in several macro-level disciplines, including strategy and corporate governance, the vast majority of research has been conducted at the organizational and institutional levels of analysis, such as strategy and corporate governance. Yet the most important knowledge gaps in the CSR literature relate to a lack of understanding of individual-level issues (Aguinis and Glavas 2012; Devinney 2013; Morgeson et al. 2013). The micro-foundations of CSR (employees’ roles in and consequences of CSR) remain largely unexplored. By adopting a cross-level approach to consider the effects of both individual-level predictors and organizational contexts on employee support for external CSR, this research will further the micro-foundations of CSR research.

Second, this research answers the call to extend CSR research into the HRM domain (Morgeson et al. 2013). SRHRM is a newly developed concept contributed by scholars including Orlitzky and Swanson (2006) and Shen and his colleagues (Shen 2011; Shen and Benson 2016). SRHRM has thus far received only limited empirical research attention, with the exception of Shen and Zhu (2011) and Shen and Benson (2016). Shen and Zhu (2011) examined the relationship between SRHRM and employee organizational commitment, while Shen and Benson (2016) considered the effect of SRHRM on employee work performance and extra-role helping behavior. However, overall, as stated by Morgeson et al. (2013), CSR research lacks consideration of HRM issues. According to the HRM attribution literature (Nishii et al. 2008), SRHRM should improve employee attitudes and behaviors toward external CSR. However, this employee consequence of SRHRM has not been previously studied. This study adopts a behavioral HRM perspective (Bowen and Ostroff 2004; Lengnick-Hall et al. 2009) to build its theoretical model upon the HRM attribution literature (Nishii et al. 2008) and explore the effect of SRHRM on employee support for organizational external CSR initiatives. We develop our conceptual framework based on the organizational climate literature, which suggests that organizational climates are important psychological processes of HRM (Gelade and Ivery 2003; Mossholder et al. 2011; Wright and Boswell 2002). The conceptual model is used to test the mediation of organizational CSR climate, which has not been previously studied in HRM research. Thus, this study offers a better understanding of employee consequences of SRHRM, including the social and psychological process involved.

Third, despite the increasing academic attention devoted to CSR, little research has empirically explored the effect of organizational policies on employees’ support for external CSR. Aguinis and Glavas (2012) recently became aware of this deficiency and called for more studies to develop “a better understanding of the predictors that influence individuals to carry out CSR activities” (p. 953). This research helps bridge this gap by: (1) differentiating between employees’ general attitudes toward CSR and employees’ attitudes toward organizations’ specific CSR initiatives, (2) exploring the effect of organization-level practices (SRHRM practices) on employee support for organizational external CSR, and (3) recognizing potential conflicting interests of different stakeholders and considering the role of perceived CSR directed toward employees. By differentiating external CSR from CSR directed toward employees, our study sheds some light on the nuanced employee consequences of CSR. In short, we adopt the HRM behavioral perspective in order to better understand why and when SRHRM influences employees’ support for their firms’ CSR initiatives targeting external stakeholders.

We test our theoretical model in the context of the People’s Republic of China, which has successfully developed a socialist market economy through a few decades of economic reform. At the early stage of the economic reform (between the late 1970s and late 1990s), the rapid economic development was achieved often at the cost of environmental failure and employee rights violation. Under President Xi Jingping’s administration, China has been investing great effort to transition to a green economy and become an environment advocate. As a result, regardless of their country of origin, firms are under scrutiny for environment violation (Tatlow 2016). In order to ensure employees engage in CSR initiatives, firms in China are likely to implement SRHRM practices. Hence, China has become an appropriate test field for conducting research on SRHRM. The findings of this research will have practical implications for both Chinese indigenous firms and foreign multinational enterprises.

As noted earlier, our model relies on a behavioral HRM literature framework, suggesting that employee perceptions regarding human resource interventions serve as input for subsequent employee attitudes and behaviors. The behavioral HRM theoretical framework allows us to conceptualize individual perceptions within a broader organizational context, which has been noted as an important factor in furthering the CSR literature (Aguinis and Glavas 2012; Morgeson et al. 2013). Our theoretical model is multilevel in nature and includes a cross-level mediation relationship (a cross-level direct relationship mediated by an organization-level construct) and two cross-level interactive relationships (a cross-level direct relationship and a cross-level indirect relationship moderated by a lower-level construct). In the next section, we offer the theoretical background and rationale for each of our hypotheses. Figure 1 is a graphic representation of our conceptual model.

Fig. 1
figure 1

Conceptual model and hypotheses

Full size image

Hypotheses Development

Socially Responsible HRM and Employee Support for Organizational External CSR

The behavioral perspective on HRM suggests that employee work attitudes and behaviors are in part a reaction to their interpretations and appraisals of organizational HRM practices (Guest 2011; Mossholder et al. 2011; Wright and Boswell 2002). Organizations adopt a range of HRM practices for different purposes. Attribution theory explains why people behavior in certain ways. More specifically, people tend to analyze the causes and effects of others’ behaviors; a person’s perception of the purpose of other people’s behaviors influences that person’s future behavior (Heider 1958; Weiner 1986). Researchers such as Bowen and Ostroff (2004) and Nishii et al. (2008) have argued that employees tend to personalize organizations. The perceptions of how employees are treated by the organization largely influence their attitudinal and behavioral responses to the organization. Based on attribution theory, it is argued that a firm’s HRM to a great extent determines employees’ perceptions of the organization, and therefore, affects their work attitudes and behaviors (Nishii et al. 2008).

SRHRM is a comprehensive concept that consists of various SRHRM practices, aiming to facilitate the implementation of organizational external CSR policies. Specifically, considering individual CSR values during employee selection increases the chance of recruiting employees who have positive general attitudes toward CSR. CSR training helps to distill the organization’s CSR values to employees. Moreover, Ellis (2009) argued that a lack of awareness or misperception of CSR can have negative effects on employees. Thus, CSR training enhances employees’ awareness of CSR and improves their ability to engage in CSR initiatives. Recognition of employees’ social performance—such as green behavior, voluntary work, community involvement, and donations to nonprofit organizations or people in need—during appraisals and pay sets reward orientation (Gelade and Ivery 2003) and enhance employees’ external CSR cognition (Orlitzky and Swanson 2006). Overall, using a set of SRHRM practices enables organizations to increase employee support for external CSR initiatives. Thus, we offer the following hypothesis:

Hypothesis 1

SRHRM practices will be positively related to employee support for the organization’s external CSR initiatives.

Mediating Role of Organizational CSR Climate

The HRM literature has advanced considerably during the last few decades. One notable theoretical advancement is the wide recognition among HRM researchers that HRM practices may not directly affect employees’ work attitudes and behaviors; rather, they transmit effects to employees through certain social and psychological processes (Guest 2011; Jiang et al. 2012). It is suggested that organizational climates are important underlying mechanisms linking HRM practices and employees’ work attitudes and behaviors (Mossholder et al. 2011). Reichers and Schneider (1990) defined organizational climate as “the shared perception of the way things are done around here” (p. 22). Organizational climate is a process of quantifying organizational culture, resulting largely from the adoption of certain organizational policies. Therefore, organizational climate changes over time, such as when organizational policies change (Schneider et al. 2013). HRM practices are major factors shaping organizational climates (Gelade and Ivery 2003; Mossholder et al. 2011). Employees make sense of their work environment in terms of what is accepted and the consequences of certain behaviors, and take action accordingly (Burke et al. 1992). A successful organization cares about external stakeholders beyond the maximization of economic profits and embraces a culture that values external CSR commitment. The adoption of SRHRM practices enables the organization to communicate its CSR values to its employees by integrating this espoused value into its organizational systems and encouraging employees to enact it in their daily behaviors. By so doing, the organization is likely to develop an organizational CSR climate, which is an employees’ shared perception about the work environment committed to addressing the interests of various stakeholders.

The organizational climate literature suggests that organizational climate influences the cognitive and affective states of employees, causes employees to behave in certain ways, and motivates employees to exhibit the behavior required and accepted in the workplace (Reichers and Schneider 1990). This aligns with social learning theory (Bandura 1977), which suggests that individual behavior is influenced by environmental factors and stimuli, and not by psychological factors alone. Organizational climates influence employees’ work attitudes and behaviors because member behaviors are, to a certain extent, shaped by the group (Hackman 1992). A number of studies (e.g., Mossholder et al. 2011) have provided empirical evidence for the mediating role of organizational climates in the HRM—employee work attitudes and behaviors relationship. Similarly, an organizational CSR climate will serve as “social stimuli” to influence employee cognitions and foster employee attitudes toward external CSR, consistent with the organization’s espoused CSR values. Hence, SRHRM leads to an organizational CSR climate that increases employee support for the organization’s external CSR initiatives. This leads to the following mediation hypothesis:

Hypothesis 2

Organizational CSR climate will mediate the relationship between SRHRM practices and employee support for organizational external CSR initiatives.

Moderating Role of CSR Directed Toward Employees

Employees may place pressure on companies to engage in CSR (Aguilera et al. 2007; Brammer et al. 2011). One reason for this is that employees expect firms that are actively engaged in external CSR to behave toward their employees also in a socially responsible manner (Royle 2005). However, there is evidence that while “some corporations are keen to take on the rhetoric of CSR; they may be less keen to act in a socially responsible manner [to their employees]” (Royle 2005, p. 51). It is possible that companies will attempt to offset the costs of their external social goals by compromising employees’ welfare, development, and even working conditions (Klein 2001). For example, some organizations require employees to sacrifice part of their salaries to support external CSR policies and actions (Zappalá 2004).

It is important to note that, although external CSR is directed toward external stakeholders, an organization’s engagement in external CSR—regardless of its motives—will have a range of organizational benefits, such as increased appeal to customers (Larson et al. 2008; Luo and Bhattacharya 2006) and job seekers (Gully et al. 2013; Rupp et al. 2013). Although these benefits can eventually result in financial returns, they may not be immediately effective or may be disproportionate to the amount of investment in CSR (Orlitzky et al. 2003). As a result, an organization’s engagement in external CSR may not benefit employees, or at least not in the short term.

According to stakeholder theory (Freeman 1984), the interests of stakeholders are likely to be different. For example, an employee’s value orientation is often self-concerned (Meglino and Korsgaard 2007); thus, in some cases, external CSR may not benefit, or may even contradict, the welfare of incumbent employees—especially those who do not have long-term employment relationships with the organization. Hence, while we generally expect SRHRM and organizational CSR climate to be positive driving forces in increasing employee support for an organization’s external CSR initiatives, their positive effects are likely to be contingent on the extent to which the organization is also engaged in CSR that effectively addresses employees’ interests. As Pfeffer (2010) suggested, a number of employee-oriented CSR initiatives—such as providing life insurance, avoiding layoffs, enabling a work-life balance, providing job autonomy, and maintaining equitable remuneration—are important for maintaining the health and well-being of employees, and have positive effects on employees.

Based on the discussions above, we propose that the effects of SRHRM and organizational CSR climate on employee support for the organization’s external CSR initiatives may depend on the level of CSR directed toward employees. This level determines employees’ perceptions of how they are valued by their organization as legitimate members (Edwards and Peccei 2010). When employees perceive that they are not socially validated as legitimate organizational members, they tend to care less about how their organization treats external stakeholders and how external stakeholders perceive their organization (Dutton et al. 1994; Edwards and Peccei 2010). Under these circumstances, employees are likely to view their organizational external CSR initiatives as a threat to their interests and welfare, and as inauthentic—simply window dressing or a marketing ploy (Becker-Olsen et al. 2006). This undermines the effect of SRHRM practices and organizational CSR climate on employee support for the organization’s external CSR initiatives. Conversely, a high level of CSR directed toward employees enhances their sense of being valued as legitimate organizational members. This intensifies the hypothesized positive effect of SRHRM practices and organizational CSR climate on employee support for the organization’s external CSR initiatives. Hence, we predict that CSR directed toward employees will moderate the effect of SRHRM on employee support for external CSR. Moreover, as CSR mediates the relationship between SRHRM and employee support for external CSR, we predict that CSR directed toward employees also moderates the indirect effect of SRHRM. Therefore, we hypothesize the following:

Hypothesis 3

CSR directed toward employees will moderate the relationship between SRHRM practices and employee support for the organization’s external CSR initiatives, such that the relationship will be stronger when the levels of CSR directed toward employees are high.

Hypothesis 4

CSR directed toward employees will moderate the indirect effect of SRHRM on employee support for the organization’s external CSR initiatives via organizational CSR climate, such that the indirect effect will be stronger when the levels of CSR directed toward employees are high.

Method

Data Collection Procedures and Participants

We first administered questionnaires that included measures for SRHRM practices, organizational CSR climate, and CSR directed toward employees (the predictors in our conceptual model). Two months later, we conducted a second round of data collection by measuring the support for organizational external CSR initiatives (the dependent variable in our conceptual model). The participants were given the same codes when completing the two-wave surveys, and we paired the surveys by matching the codes. Collecting data for predictor and outcome variables on separate occasions helped reduce the possibility of common method variance (CMV) (Podsakoff et al. 2003) and causal ambiguity (Finkel 1995). To facilitate respondents’ understanding of the meanings of the measuring statements, we provided a glossary of terms (for example, for “external CSR” and “social performance”) and examples of external CSR (such as engagement in environmental protection, donation to charities, and community engagement activities), CSR directed toward employees, and SRHRM practices.

We collected data from employees in 32 firms in the People’s Republic of China. Four of the 32 firms were joint ventures, two firms were from the USA, one firm was Japanese, and one firm was a South Korean multinational enterprise. The other companies were all domestic and privately owned. To reduce response bias, we sent letters via each organization’s human resources department to 50 randomly selected employees, explaining the research purpose, the voluntary nature of participation in the study, the data collection procedures, and assurance of confidentiality. The questionnaires were distributed via internal mail and collected in the workplace using sealed collection boxes. The response rates varied from 32 to 84% across firms, with 812 employees returning both surveys, and 776 surveys matched and useable. Of the 776 participants, 69% were male and 18.4% had managerial positions. The mean age was 43 years and, in terms of educational level, 42.1% of participants had completed junior high school.

Measures

We used five-point Likert-type scales ranging from 1 = strongly disagree to 5 = strongly agree for all study variables. The questionnaire was developed in English, translated into Mandarin, and back-translated by two bilingual researchers.

SRHRM Practices

SRHRM was measured using Shen and Benson’s (2016) six-factor scale that was based on the Western literature (Orlitzky and Swanson 2006) and statistically validated in the Chinese context. The six statements were: (1) My company considers candidates’ general attitudes toward CSR in selection, (2) My company uses training to promote CSR as an espoused organizational value, (3) My company provides CSR training to develop employees’ skills in stakeholder engagement and communication, (4) My company considers employee social performance in promotions, (5) My company considers employee social performance in performance appraisals, and (6) My company relates employee social performance to rewards and compensation. The Cronbach’s alpha for this measure was 0.81.

Organizational CSR Climate

A scale was purposefully developed to measure organizational CSR climate based on the CSR literature (e.g., Ellis 2009; Maignan and Ferrell 2004) and organizational climate literature (Reichers and Schneider 1990). We generated a range of literature-based statements and asked 35 Chinese Master of Business Administration students to rate the relevance of the statements to the Chinese context on a five-point Likert scale (1 = not relevant at all, 3 = neutral, and 5 = very relevant). The following six statements scored an average of 4 or 5: (1) In this company, employees are expected to do what is right to support social causes; (2) This company recognizes and rewards employees’ contributions to social causes; (3) In this company, indifference to social causes is regarded as irresponsible; (4) In this company, people protect their own interests above supporting social causes (R); (5) Successful people in this company support social causes; and (6) In this company, each person is expected to work efficiently, even if it does not benefit social causes (R). Item 2 was virtually identical to SRHRM item 6; thus, it was removed from further analysis. Using principal component analysis, the exploratory factor analysis extracted one factor (explaining 71.95% of variance, with individual factor loading ranging from 0.80 to 0.88). The confirmatory factor analysis (CFA) supported the one-factor structure: χ2(5) = 14.35, p < 0.001, comparative fit index (CFI) = 0.97, incremental fit index (IFI) = 0.97, and root mean square error of approximation (RMSEA) = 0.06. The Cronbach’s alpha was 0.90.

Perceived CSR Directed Toward Employees

We measured perceived CSR directed toward employees using the five-item employee responsibilities dimension of CSR developed by Maignan and Ferrell (2004). The five statements were: (1) My company treats all employees fairly and respectfully, regardless of gender or ethnical background; (2) My company provides all employees with salaries that properly and fairly reward their contributions; (3) My company supports all employees who want to pursue further education and develop careers; (4) My company helps all employees coordinate their private and professional lives; and (5) My company incorporates the interests of all employees into business decisions. Cronbach’s alpha was 0.84.

Employee Support for CSR Initiatives Targeting External Stakeholders

We measured employee support for external CSR using three statements: (1) Overall, I support my company’s CSR initiatives that address the interests of external stakeholders; (2) I am willing to contribute to my company’s external CSR programs; and (3) I praise my company for its external CSR initiatives. The normality test revealed that the 5% trimmed mean of 3.087 was close to the original mean of 3.077, indicating data normality. Cronbach’s alpha was 0.79.

Control Variables

Our research design controlled for gender, age, education, and hierarchical position in the organization because these factors may influence employee attitudes toward CSR (Rodrigo and Arenas 2008). We also controlled for employees’ general attitudes toward CSR, as they are likely to affect support for the organization’s specific CSR initiatives. Employee general attitudes toward CSR were measured using two statements from Turker (2009): (1) “Being socially responsible is the most important thing a firm can do” and (2) “All business has a social responsibility beyond making profits.” The Cronbach’s alpha was 0.76. As expected, the correlation between employees’ general attitudes toward CSR and support for external CSR initiatives was positive and significant (r = 0.32, p < 0.01). We ran CFAs to check whether these two variables were distinctive. The results showed that the two-factor model fitted the data well: χ2(4) = 9.74, p < 0.001, CFI = 0.97, IFI = 0.97, and RMSEA = 0.06, and fitted better than the one-factor model: χ2(5) = 18.35, p < 0.001, CFI = 0.93, IFI = 0.93, RMSEA = 0.15, Δχ2(1) = 8.61, and p < 0.01. These two measures shared 23% of variance, suggesting that, although related as expected, they were distinguishable constructs.

Finally, we controlled for employee perceptions of CSR fit and motive, which is believed to influence employee reactions to CSR (Elliot 2011; Ellis 2009). This was measured using a shortened scale from Speed and Thompson (2000). The four items were: (1) My organization is engaged in external CSR because it cares about the society, (2) My organization has genuine interests in the social causes that it supports, (3) There is a logical connection between my organization and the external CSR initiatives in which it is engaged, and (4) My organization and the social initiatives in which it is engaged fit well. Cronbach’s alpha was 0.87.

Data-Analytic Approach

Employees who are nested within the same organization are exposed to the same HRM practices; hence, their attitudinal and behavioral responses to HRM may not be independent of organizational contextual effects (Bliese and Hanges 2004). Kenny and Judd (1996) noted that “observations may be dependent, for instance, because they share some common features, come from some common sources, are affected by social interactions, or are arranged spatially or sequentially in time” (p. 138). Therefore, it is appropriate to conceptualize and measure SRHRM at the organizational level. According to Rousseau (1985), when data collected from individuals are aggregated to higher levels, the following two preconditions should be met. First, there must be systematic between-group variations, which can be tested by performing one-way analysis of variance (ANOVA) and calculating intra-class correlation (ICC1) and reliability of the mean (ICC2). ICC1 indicates the proportion of variance due to team variability. ICC1 values different from zero and close to 0.20 are desirable (Bliese 2000). ICC2 differentiates groups in terms of individuals’ ratings, with ICC2 values greater than 0.60 desirable (Bliese 2000). Second, there is a high level of inter-rater agreement for Level-2 variables within the Level-2 units. rwg has been developed to assess the level of inter-rater agreement (Le Breton et al. 2005). According to James et al. (1984), an rwg value greater than 0.70 is acceptable.

We implemented these recommended procedures to confirm the appropriateness of aggregating SRHRM practices scores at the firm level. First, the results of an ANOVA suggested statistically significant between-firm variations in SRHRM practices across the 32 organizations: F[31, 776] = 3.79, p < 0.001. Second, we assessed rwg, which had an average score of 86. Third, we computed ICC1 and ICC2, which were 0.10 and 0.74, respectively (Bliese 2000). These results provided evidence in support of aggregating SRHRM practices scores at the firm level of analysis.

Conceptually, an organizational climate is a shared perception of organizational members. Hence, we also undertook the same procedures to explore whether it was appropriate to aggregate the data for organizational CSR climate to the organizational level. A one-way ANOVA revealed significant between-firm variations in organizational CSR climate: F[31, 776] = 3.18, p < 0.001. The average rwg was 0.90. ICC1 and ICC2 were 0.15 and 0.81, respectively. These findings justified the aggregation of organizational CSR climate to the organizational level.

Due to the fact that all employees in the same organizations are exposed to the same CSR directed toward employees, their perceptions may be similar within the same organization and different between organizations. We also explored whether it might be appropriate to aggregate this variable to the organizational level. A one-way ANOVA revealed significant between-firm variations in organizational CSR climate: F[31, 776] = 2.32, p < 0.001. However, the average rwg was 0.62, which was below the cutoff point of 0.70 (James et al. 1984). Similarly, ICC1 and ICC2 were 0.05 and 0.57, respectively, which were below the recommended cutoff points of 0.06 and 0.70, respectively (Bliese 2000; Glick 1985). These findings did not support aggregating perceived CSR directed toward employees to the organizational level.

Other study variables were conceptually individual-level constructs. As a result, our research design was multilevel in nature. The hierarchal data structure leads to gross errors of prediction if a researcher uses statistical approaches (such as ordinary least squares regression) that are not designed to model data structures that include dependence due to clustering of entities (Bliese and Hanges 2004; Hox 2010; Snijders and Bosker 2012). In addition, traditional multilevel modeling is inappropriate for testing cross-level mediation because it violates the assumption of independence of observations, resulting in downwardly biased standard errors and conflation of indirect effects (Preacher et al. 2010). To overcome these problems, we performed multilevel structural equation modeling using Mplus 7 to simultaneously estimate the between- and within-level variation, thereby imposing an equality constraint on the fixed component of the between and within slopes, and separating within effects from between effects (Preacher et al. 2010). The significance of the cross-level indirect effect was estimated by exploring confidence intervals using the Monte Carlo method recommended by Preacher et al. (2010). Finally, we followed the one-step procedure suggested by Croon and Van Veldhoven (2007) to estimate direct and indirect paths simultaneously. We fitted models by comparing the partial mediation models with the full mediation model and adopted the procedures described by Edwards and Lambert (2007) for moderated mediation.

Results

Validity Evidence for the Measurement Model

Before proceeding with the substantive analyses, we conducted a series of CFAs to gather additional validity evidence regarding the measurement model by comparing the fit of the hypothesized four-factor model with other nested models. Our proposed model—including SRHRM practices, organizational CSR climate, CSR directed toward employees, and support for external CSR initiatives—indicated good fit to the data: χ2(146) = 400.59, p < 0.001, CFI = 0.97, IFI = 0.97, and RMSEA = 0.06. The four-factor model was a better fit than:

  • the three-factor model collapsing SRHRM and organizational CSR climate: χ2(149) = 445.69, p < 0.001, CFI = 0.89, IFI = 0.89, and RMSEA = 0.093, Δχ2(3) = 45.10, and p < 0.001

  • the two-factor model collapsing SRHRM, organizational CSR climate, and CSR directed toward employees: χ2(151) = 514.91, p < 0.001, CFI = 0.82, IFI = 0.82, and RMSEA = 0.11, Δχ2(5) = 114.31, and p < 0.001

  • the one-factor model: χ2(152) = 541.12, p < 0.001, CFI = 0.79, IFI = 0.79, and RMSEA = 0.13, Δχ2(6) = 140.53, and p < 0.001.

In addition, all loadings were statistically significant and exceeded the minimum recommended value of 0.40 (Holtman et al. 2002). Hence, the four-factor model had the best fit to the data.

Although we collected the data on two separate occasions, we used self-reported measures, and all three variables were collected from the same source at time 1, meaning that the potential for CMV could still remain. We followed the procedure suggested by Podsakoff et al. (2003) to compare the fit of the three-factor model to a model allowing all items to load on the method factor. The method model had a better fit (χ2[85] = 209.95, p < 0.001, CFI = 0.96, IFI = 0.96, and RMSEA = 0.06) than the three-factor model (χ2[101] = 253.51, p < 0.001, CFI = 0.95, IFI = 0.95, and RMSEA = 0.07, Δχ2[16] = 43.56, and p < 0.01). However, the CFI change was only 0.01, which was less than the 0.05 rule of thumb (Bagozzi and Yi 1990). The method factor accounted for just 7% of the total variance, which was lower than the 25% rule of thumb (Williams et al. 1989). Hence, CMV was not a serious problem with the data. In summary, these results provided evidence in support of the satisfactory psychometric properties of the measurement model.

Descriptive Statistics

Table 1 presents the descriptive statistics, including the correlations and internal consistency estimates. The results indicated that SRHRM was positively related to organizational CSR climate, which was also positively related to employee support for organizational external CSR. Moreover, CSR directed toward employees was positively related to employee support for organizational external CSR.

Table 1 Means, SDs, correlations between variables, and internal consistency reliability
Full size table

Hypothesis Tests

Before testing our substantive hypotheses, we estimated null models to examine the extent to which variance in the individual-level criterion variable (employee support for external CSR) resides at the organizational level of analysis. The results provided evidence regarding between-firm variance in support for external CSR initiatives: χ2(31) = 79.78, p = 0.000, and ICC1 = 0.14. In other words, 14% of variance in individual-level support for external CSR scores resided between firms. Hence, it was appropriate to implement the multilevel data-analytic strategy to test our conceptual model (Snijders and Bosker 2012).

We then proceeded to test our substantive hypotheses. First, we assessed the fit of the partial mediation model shown in Fig. 1. Second, we assessed a competing full mediation model by removing the direct relationship between SRHRM practices and support for organizational external CSR. As noted by Preacher et al. (2010), “any mediation effect in a model at least one of X, M, or Y is assessed at Level-2 must occur strictly at the between-group level” (p. 210). Thus, we fitted the competing models at the between-group level. The results showed that the partial mediation model fitted the data significantly better (χ2[74] = 178.34, p < 0.001, CFI = 0.98, IFI = 0.98, and RMSEA = 0.06) than did the full mediation model (χ2[75] = 219.75, p < 0.001, CFI = 0.90, IFI = 0.81, and RMSEA = 0.10, Δχ2[1] = 41.41, p < 0.001). Hence, we retained the partial mediation model as the final model to assess our individual hypotheses.

Hypothesis 1 refers to a cross-level direct relationship between SRHRM practices and employee support for external CSR. An examination of the cross-level standardized path coefficient for this relationship provided support for this hypothesis: γ = 0.34, p < 0.01. Hypothesis 2 refers to the cross-level mediating effect of organizational CSR climate in the relationship between SRHRM practices and support for external CSR initiatives. To test this mediating relationship, we followed the approach suggested by MacKinnon (2008) by examining the significance of the direct effect. We examined the path coefficient between SRHRM practices and organizational CSR climate (γ = 0.35, p < 0.001) and the path between organizational CSR climate and employee support for external CSR initiatives (γ = 0.24, p < 0.01) at the cross-level. The results with 20,000 Monte Carlo replications revealed that the indirect relationship between SRHRM and employee support for external CSR via organizational CSR climate was significant: indirect effect = 0.08, p < 0.01 with 95% bias-corrected bootstrap confidence interval [0.041, 0.160]. Hence, Hypothesis 2 was supported.

We adopted the following two steps to assess Hypotheses 3 and 4, referring to moderated direct and indirect relationships. Both the relationships of SRHRM and organizational CSR climate with employee support for external CSR were at the cross-level; thus, we estimated the moderating effects of CSR directed toward employees at the cross-level only. The results indicated that the product term SRHRM practices × CSR directed toward employees and the product term organizational CSR climate × CSR directed toward employees were positively related to employee support for external CSR (γ = 0.08, p < 0.01; γ = 0.06, p < 0.01).

Next, we tested Hypothesis 4 regarding the moderated mediation relationship by following Bauer et al.’s (2006) recommendation to estimate the indirect effects of SRHRM at high [1 standard deviation (SD) above the mean) and low (1 SD below the mean) levels of CSR directed toward employees. The results showed that the indirect effect of SRHRM was stronger (γ = 0.13**) when the level of CSR directed toward employees was high, compared to when the level was low (γ = 0.04*, Δγ = 0.09, p < 0.01). These results demonstrated that the indirect effect of SRHRM on employee support for the organization’s external initiatives via CSR climate was moderated by CSR directed toward employees. Table 2 provides the results of path analysis for the first- and second-stage mediations, and the direct, indirect, and total effects at both high and low levels of CSR directed toward employees.

Table 2 Results of multilevel mediation and moderation analyses
Full size table

We used the hierarchical linear-regression model two-way interaction graphing tool described by Preacher et al. (2006) to create graphs illustrating the nature of the cross-level interaction effects. As shown in Fig. 2, each interaction effect was in the predicted direction, such that all hypothesized relationships increased as the values of CSR directed toward employees increased. Together, these results offered support for Hypotheses 3 and 4.

Fig. 2
figure 2

Graphic representation of interaction effects. Note Top panel CSR directed toward employees as a moderator of the relationship between SRHRM practices and employee support for external CSR (Hypothesis 3). Bottom panel CSR directed toward employees as a moderator of the relationship between organizational CSR climate and employee support for external CSR (Hypothesis 4). High, one SD above the mean; low, one SD below the mean

Full size image

Discussion

The topic of CSR is of great importance for organizations and society in general. However, owing to a lack of research, we know little about the micro-foundations of CSR. Accordingly, the goal of this study was to investigate this “black box” in CSR research—specifically, how SRHRM practices influence employee support for a firm’s CSR initiatives targeting external stakeholders, and what is the nature of the social and psychological process of this relationship. Further, we examined under what conditions this relationship becomes stronger or weaker. Consistent with the behavioral HRM literature, our results suggested that micro-level variables provide important insights to why and when employees are likely to support organizational external CSR initiatives, when employee general attitudes toward CSR are controlled for. Implementing SRHRM practices developed organizational CSR climate, which increased employee support for external CSR initiatives. Our results also suggested that the effects of SRHRM and organizational CSR climate on employee support for external CSR initiatives were contingent on the implementation of employee-focused CSR—that is, CSR directed toward employees. Specifically, the direct and indirect effects of SRHRM through the mediation of organizational CSR climate on employee support for external CSR became stronger (more positive) when the levels of CSR directed toward employees were higher. In contrast, when there was a lack of CSR addressing employees’ interests, they might regard external CSR as merely a marketing ploy, and the direct and indirect effects of SRHRM on employee support for external CSR consequently became weaker (less positive).

Our study contributes significantly to the literature. First, it helps fill the micro–macro gap in the literature. Organizational behavior (OB), HRM, and industrial and organizational (I–O) psychology focus mostly on individual-level and (to some extent) team-level phenomena. An examination of I–O psychology textbooks (e.g., Cascio and Aguinis 2011) and compendia (e.g., Rogelberg 2007; Zedeck 2011) indicates that major topic headings include personnel selection, training, performance appraisal and management, individual differences, and job analysis and design. The vast majority of these topics address individual-level phenomena. For the most part, organization-, industry-, and society-level phenomena are not discussed in detail and do not play a major role. Therefore, it seems that the time is right for OB, HRM, and I–O psychology to adopt a broader perspective that includes higher levels of analysis (Aguinis 2011). Conversely, CSR research has been conducted mainly at the organizational level. Our research adopts a cross-level approach by considering the effects of both organization- and individual-level variables on an individual-level outcome. Thus, our study can serve as a springboard for future research on CSR, as well as other substantive domains that consider both micro- and macro-level issues, thereby helping bridge the micro–macro gap.

Second, the extant literature suggests that employees’ attitudes toward CSR are determined mainly by employees’ CSR values, which are influenced largely by their upbringings and work environments (Rodrigo and Arenas 2008). Our research distinguishes employee attitudes toward CSR in general and employee attitudes toward their organization’s specific CSR initiatives. This distinction is important because it enables a better understanding of employees’ nuanced attitudes toward CSR and the associated factors.

Third, our research answers the recent call by Aguinis and Glavas (2012) and Morgeson et al. (2013) to extend CSR research into the HRM realm. The findings of this study provide empirical evidence in support of the behavioral HRM literature in relation to the attributes (Nishii et al. 2008) and social and psychological processes (such as organizational climate) (Gelade and Ivery 2003; Jiang et al. 2012; Mossholder et al. 2011) of HRM. It should be noted that HRM practices can influence employee behaviors through different mediating paths, such as human capital, motivation, and opportunity enhancement (Jiang et al. 2012). The current research considered only one transactional path—organizational social behavior climate. Thus, we call for further research exploring other possible mediating paths.

Further, using stakeholder theory allowed us to confirm that levels of CSR directed toward employees are the contingent factor for whether employee support for external CSR is likely to be strengthened or weakened. This finding is significant and can be extended to a broader context. Individuals, organizations, and the broader community may only support socially responsible initiatives on the condition that their interests are not threatened. Therefore, there is a need for further research to apply stakeholder theory to explore the socially responsible behaviors of organizations’ multiple social players. In short, our study suggests that future CSR research should adopt a multilevel conceptual framework, consider appropriate mediating paths, and consider the interests of multiple stakeholders.

Research Limitations and Further Research Directions

This study has several limitations that need to be addressed. First, the requirements and standards for CSR, SRHRM practices, and people’s interests in CSR vary between firms, economies, and cultures. This study was conducted in China, where the economic and cultural environment may differ significantly from other national contexts. Hence, the strengths of the relationship tested in this study may be not the same elsewhere. We suggest that future research use cross-cultural data in order to increase the generalizability of the research findings. Second, because this study used cross-sectional data, it was difficult to establish causal relationships. To address this limitation, future research should collect data at multiple points in time in order to explore the changes in employee attitudes toward CSR as a result of the implementation of SRHRM practices.

Third, our conceptual model considered only one mediator and one moderator, yet it is suggested that HRM practices may influence employee outcomes through multiple social and psychological processes (Jiang et al. 2012). In addition, SRHRM is different from general HRM, as the former is also part of CSR. Hence, the social and psychological processes through which SRHRM is linked to employee outcomes may be different from those for general HRM practices. It would be interesting for future research to explore the alternative mediators and moderators. Finally, we only considered one employee outcome in this study. As SRHRM has received very limited research attention, future research should endeavor to investigate other employee outcomes, as well as organizational outcomes. Such research would greatly enrich the HRM and CSR literature.

Implications for Practice

The increased importance of CSR for organizations and society means that our study has significant implications for organizations striving to successfully implement CSR initiatives. Based on the HRM behavioral literature on the relationship between HRM practices and the resulting employee attitudes and behavior (e.g., Bowen and Ostroff 2004; Lengnick-Hall et al. 2009), our study provides a concrete, actionable recommendation for how to undertake this. First, organizations should implement effective SRHRM practices to develop organizational CSR climate and obtain employee support for external CSR. More positive employee attitudes toward HRM practices can be attained by considering job candidates’ CSR values in recruitment and selection; offering CSR training opportunities; and considering employees’ social performance in promotion decisions, performance appraisals, and rewards and compensation systems. Such practices are also likely to communicate organizational CSR values and change employees’ cognition of CSR, which enhances employees’ support for the implementation of organizational external CSR initiatives.

Second, organizations should strive to address the interests of employees, as well as those of other stakeholders. Employee support for external CSR initiatives is not unconditional. Although employees may support CSR in principle, only employees who perceive that their organization cares about their interests and well-being are likely to regard its external CSR as legitimate, and to support and be committed to external CSR. Hence, in order to effectively engage in external CSR, a firm needs to provide adequate organizational support and justice to employees—not just its external CSR-promotion policies—to meet the conditions under which employees are more likely to support external CSR initiatives. In a broader context, a firm has multiple stakeholders who may have different or even conflicting interests and needs regarding CSR. The way these stakeholders react to the organization’s CSR initiatives is largely dependent on how their respective interests and needs are addressed. Thus, it is critical for an organization to achieve a balance in meeting the needs of employees and various stakeholders.

Concluding Remarks

CSR research has been conducted primarily at the macro-level of analysis and in fields such as strategy and corporate governance. We hope that our study will serve as a catalyst for future empirical research on the employee consequences of CSR. We believe that this research has great potential to enhance knowledge regarding why and when CSR policies and actions are likely to lead to positive outcomes for individuals, organizations, and society. Moreover, we believe that this research has great potential to help narrow the much-lamented micro–macro gaps in OB, HRM, and I–O psychology and related fields.