Addressing the Relationship Between Emotion Management and Social Capital: An Analysis of Federal Employees

Emotion management describes the ability of an employee to regulate the outward manifestation of emotions during personal interactions. Prior studies examining emotions in organizations have often emphasized the production of emotional labor during service delivery transactions with clients. Such a focus has overlooked the emotion management that takes place between employees. Using ordinary leastsquares (OLS) regression, this study analyzes the 2016 Merit Principles Survey of federal employees to examine the relationship between emotion management and the development of social capital. Findings from the study show a positive relationship between emotion management and social capital. These findings provide evidence that employees capable of properly regulating their emotional states during interactions with their colleagues perceive themselves as having more social capital. The findings, however, also indicate that the positive relationship between emotion management and social capital does not hold for employees of color to the same extent that it does for white employees. This could mean that emotion management by employees of color is less likely to lead to increases in social capital as it is for their white counterparts. These findings have important implications for our understanding of emotion management and its impact on how employees navigate their organizations.

While these studies have provided important insights regarding the emotional experiences of organizational members when interacting with clients, they have generally lacked a discussion of emotional experiences when employees interact with one another as colleagues. Thus, this study seeks to address this gap in the literature by examining the emotion management carried out by employees when interacting with their coworkers.
Building on existing literature, this study examines the relationship between emotion management and the development of social capital. Established through a network of relationships, social capital is a resource that provides a mutual benefit for the actors involved (Putnam, 1995). To draw an understanding of the dynamic between emotion management and social capital, this study addresses two questions. First, is there a relationship between participating in emotion management and the development of social capital? Second, how do gender and race have an impact on the relationship between emotion management and social capital?
These questions contribute to public management scholarship in several ways. First, they provide the field with a deeper understanding of emotion management as it takes place between members of an organization. Emotion management between employees is an essential component of personnel interactions that creates harmony in organizations (Kiel & Watson, 2009). It, however, remains an area lacking extensive research (Puglesi, 1999). Additionally, these questions recognize the potential influence of gender and race, answering recent calls for research that acknowledges and incorporates multiple identities (Bearfield, 2009;Gooden, 2015). Last, these questions connect two concepts-emotion management and social capital-that have often remained separate in the literature but, together, help scholars gain greater insight into individual behavior and organizational outcomes.
Although emotions have provided a nuanced way to understand organizational effectiveness, emotional exchanges remain an undertheorized area of public management scholarship (Meier, Mastracci, & Wilson, 2006). In addition, research on emotions has often emphasized emotional labor on behalf of clients, rather than emotion management on behalf of colleagues. This study, therefore, addresses these gaps in the literature while also incorporating the concept of social capital. This allows for the examination of a unique relationship that can enhance the field's understanding of both emotion management and social capital in the context of public organizations.
The article proceeds as follows: first, a brief review of the literature covering emotions in organizational settings is provided. This is followed by a review of social capital scholarship. Next, the hypotheses used to analyze the relationship between emotion management and social capital are outlined, followed by a presentation of the empirical findings. Last, the conclusion connects the findings to the literature and discusses implications for theory and practice.

Emotions in Organizations: From Emotional Labor to Emotion Management
Interest in the dynamic between emotions and organizations reached new heights in the 1980s with the work of Arlie Hochschild. She first coined the term emotional labor in her book, The managed heart (Kaur & Malodia, 2017). Defined as "the management of feeling to create publicly observable facial and bodily display" (Hochschild, 1983, p. 7), emotional labor finally 7 gave a name to a commonly overlooked form of labor in the workplace. Indeed, in an effort to evoke a desired emotional response from another person (Hochschild, 1983;Newman, Guy, & Mastracci, 2009), employees often balance "their own internal psychological states with organizationally sanctioned ones" (Dehart-Davis, 2017, p. 15). This results in emotional labor.
When completing an interaction, an employee "must intuit the other's state of mind and make split-second adjustments in words, tone, or body language" (Guy, Newman, & Mastracci, 2008, p. 187). In short, emotional labor is a role performed by organizational members requiring the management of emotions (Jin & Guy, 2009). Ultimately, emotional labor is tied to an organization's ability to complete its core service functions.
At its genesis, emotional labor research emphasized historically female jobs (e.g., flight attendants) because the range of emotions studied were considered more feminine in nature (Steinberg & Figart, 1999). Since this time, scholars have recognized that many historically male-dominated professions (e.g., police and correctional officers) also require emotional labor. As such, emotional labor research has expanded to examine these areas (Martin, 1999;Schaible & Gecas, 2010;Stenross & Kleinman, 1989).
Regardless of the profession under examination, however, a primary area of interest among scholars has been the influence of emotional labor on employee well-being (Steinberg & Figart, 1999). Employee well-being has often been measured by job satisfaction and employee burnout. While prior research has found that emotional labor is positively related to social capital (e.g., Ashforth & Humphrey, 1993;Guy, Newman, & Mastracci, 2008;Yang & Guy, 2015), its relationship to burnout is more complex.
Indeed, the relationship between emotional labor and burnout is contingent on the type of emotional work expressed. Employees participating in surface acting, or presenting an emotion they are not actually feeling (Grandey, 2000;Hochschild, 1983), are more likely to experience burnout. However, employees participating in deep acting, or attempting to modify their feelings to align with socially expected outward displays (Grandey, 2000;Hochschild, 1983), experience less burnout (Guy et al., 2008). Overall, prior research has shown that producing emotional labor has a significant impact on the well-being of an organization's members.
Although other disciplines have long explored the many facets of emotions in organizational settings, discussion of emotions did not become commonplace in public administration until the 21st century. Guy and Newman (2004) are credited with introducing the field of public administration to emotional labor with their article, Women's jobs, men's jobs: Sex segregation and emotional labor. Prior to this, there was a notable absence of emotions from the field's scholarship-perhaps due to the fact that, during the industrial era, emotions were perceived as a threat to effective work by organizational leaders (Jin & Guy, 2009).
During this time, emotions were viewed as a barrier against rationality and neutrality (Ashforth & Humphrey, 1995). This led to the exclusion of emotional labor as a valuable contribution from organizational members. Thus, the production of emotional labor was undercompensated, if compensated at all, within organizational settings (Guy & Newman, 2004). However, scholars soon began to acknowledge the utility in using emotions as a framework to understand pay inequities and performance appraisals among public sector employees (Guy & Newman, 2004;Mastracci, Newman, & Guy, 2006;Steinberg, 1999).
Scholars also begun to recognize that employees possessing skills relating to emotional labor had a greater ability to ensure that citizens had positive service encounters with their government (Lu & Guy, 2014). Public organizations establish an environment that teaches employees which emotions are appropriate to convey (Rafaeli & Sutton, 1987). Employees are then expected to meet the emotional expectations of their organization and clients.

8
The increasing importance of emotional labor within public sector institutions is clear by recent growth in scholarship examining emotional labor's relationship with organizational performance (Meier et al., 2006), employee motivation (Hsieh, Yang, & Fu, 2012), job satisfaction (Jin & Guy, 2009;Kaur & Malodia, 2017), and leadership (Kiel & Watson, 2009;Lu & Guy, 2014;Newman et al., 2009). While these empirical developments have been essential to our understanding of emotions within public organizations, scholars have overlooked the potential effects of emotion management and the implications that this can have on coworker relationships and organizational effectiveness.

How Is Emotion Management Unique?
Although closely related, emotion management is distinct from emotional labor. Indeed, emotion management, which still involves regulating outward manifestations of emotions during personal interactions (Pugliesi, 1999) similar to emotional labor, does not have exchange value like emotional labor since it is not sold to a client (Hochschild, 1983). Stated another way, emotion management is not a required work task exchanged for a price (Harlow, 2003). Rather, emotion management is a helpful contribution granted to another individual within the organization (Bolton, 2005). It is not a wage-based organizationally required task like emotional labor (Hochschild, 1990).
As a result, emotion management is considered a form of organizational citizenship behavior (OCB), which can be defined as "helpful behaviors that support the social fabric of the organization that are outside of the core job tasks" (Fox, Spector, Goh, Bruursema, & Kessler, 2012, p. 201). Although OCB does not directly improve task performance or the completion of technical functions, it does enhance the environment in which employees must complete their core job duties (Organ, 1997).
In general, there are five types of organizational citizenship behaviors: altruism, conscientiousness, sportsmanship, courtesy, and civic virtue (Organ, 1988). Although emotion management does not align directly with any of these behaviors, it is strongly related to altruism and courtesy. Collectively, altruism and courtesy represent discretionary employee behaviors that attempt to minimize work-related issues and provide support to other employees when resolving organizationally relevant problems (Podsakoff, MacKenzie, Moorman, & Fetter, 1990). These two types of OCBs form the basis for interpersonal relationships among employees.
Emotion management is similar in that it also emphasizes the significance of employees' interpersonal relationships, but it does so while paying specific attention to the regulation of emotions. Since emotional regulation is not contractual or considered a required task, it tends to align more with the concept of emotion management than emotional labor (Niven, Totterdell, & Holman, 2009).
Employees participating in emotion management are essential to maintaining cordial relationships in the workplace because of their ability to connect with other members of the organization (Kiel & Watson, 2009). However, there has been virtually no research examining emotional exchanges between organizational members. Most studies have focused solely on the role of emotions during service delivery transactions with clients. Although at least one study has examined emotional dynamics between colleagues (Puglesi, 1999), the outcome of interest in this study was still employee well-being (i.e., job satisfaction and stress). Thus, the current study provides further exploration on the emotional dynamics that take place between colleagues and incorporates a new outcome of interest, social capital.

Social Capital in the Workplace
Scholars began systematically analyzing social capital in the mid-1980s (Portes, 1998). However, Putnam (1995) is credited with popularizing the term and providing a definition of social capital that scholars in various disciplines, including public management, have relied on to guide their work (Andrews, 2010;Choi, 2018;Compton & Meier, 2016). This article follows Putnam's (1995) definition of social capital as the "features of social organization such as networks, norms and social trust that facilitate coordination and cooperation for a mutual benefit" (p. 67).
Since becoming mainstream at the close of the 20th century, social capital has developed from a fringe concept into its own field of research. This research connects the work of scholars representing various disciplines (Kwon & Adler, 2014). In all of these disciplines, scholars have examined social capital as a resource that can be transferred at the individual and group levels.
At the group level, social capital is available to an entire unit and can be used by those who establish it as well as by other group members (Kostova & Roth, 2003). Studies analyzing social capital at the individual level have been interested in how the development of social capital leads to personal benefits, such as promotion within organizational settings (Inkpen & Tsang, 2005, p. 150). Although scholars vary in their approaches to studying social capital, the primary theoretical assumption underlying this work has remained consistent. That is, that the goodwill exchanged between individuals and groups is a valuable resource (Adler & Kwon, 2002).
The present study uses the federal agency workplace as a setting to study social capital at the individual level. In general, the workplace presents an ideal setting to study this concept because "many people form rewarding friendships at work, feel a sense of community among co-workers, and enjoy norms of mutual help and reciprocity on the job" (Putnam, 2000, p. 87). It serves to reason, then, that organizations should have a vested interest in social capital because it serves as an asset allowing for knowledge development and transfer among employees (Adler & Kwon, 2002;Inkpen & Tsang, 2005;Maurer, Bartsch, & Ebers, 2011;Nahapiet & Ghoshal, 1998). Indeed, when employees have relationships with one another, they are able to share specialized information that allows for the continued operation of their organization.
In prior studies, scholars have been particularly interested in examining the relationship between social capital and organizational performance (Andrews, 2010;Compton & Meier, 2016;Leana & Pil, 2006;Wichowsky & Moynihan, 2008) as well as between social capital and employee well-being (Kouvonen et al., 2008;Requena, 2003;Suzuki et al., 2010). The findings from studies of social capital and performance, however, have been mixed (Compton & Meier, 2016;Maurer, Bartsch, & Ebers, 2011). Indeed, while some scholars have found a positive relationship between social capital and performance (Leana & Pil, 2006;Batjargal, 2003), others have found no relationship between the two (Lee, Lee, & Pennings, 2001). Additionally, scholars have highlighted the complicated relationship between social capital and performance, specifically pointing out that social capital is composed of different dimensions and that each dimension may have a unique relationship with performance (Andrews, 2010;2011).
Research examining the relationship between social capital and employee well-being has been more conclusive: social capital improves employee well-being. Individuals who perceive themselves as having social capital have lower levels of depression (Kouvonen et al., 2008), better health outcomes (Suzuki et al., 2010), and higher levels of satisfaction with their quality of life at work (Requena, 2003). In sum, social capital can influence both individual and organizational outcomes.

Emotion Management and Social Capital
Emotion management and social capital are both relational concepts. That is, they both emphasize the significance of interpersonal relationships and the implications that follow. Social capital has often been identified as an asset and a resource. However, employees must invest energy in the development and maintenance of relationships with their colleagues in order to reap the benefits of social capital (Maurer et al., 2011).
An employee's investment of energy is represented by emotion management. Employees expend effort managing their psychological state during personal interactions with colleagues in order to build connections that can influence the operation of their organization. The literature indicates that "emotion can be a resource through which organizational relationships are created, interpreted, and altered" (Waldron, 2000, p. 65). In short, emotions are what guide and coordinate interactions that lead to the development of social bonds (Lord & Kanfer, 2002). Because emotions provide the foundations for social ties, in the present study it is assumed that employees who expend effort to participate in emotion management will develop social capital.
Findings from prior research help to establish three testable hypotheses. First, studies have shown that building social capital through strong ties requires an emotional investment (James, 2000). Through personal interactions that evoke the desired emotional state from another person (Newman et al., 2009), individuals can build connections that lead to the establishment of networks and social trust (Putnam, 1995). Recognizing this dynamic, the first hypothesis is: Hypothesis 1: Participating in emotion management is positively related to social capital.
Although this hypothesis frames the relationship between emotion management and social capital as positive, it is also necessary to recognize the potential influence of gender and race. Within any given community, an individual's ability to establish social capital varies (Lin, 2000). Studies have often shown that women and people of color have lower levels of social capital than their white, male counterparts (Burt, 1998;James, 2000). This means that, although women and people of color participate in emotion management within their organizations, this participation may not lead to strong social connections and generate social capital to the same extent as it does for their white, male counterparts. Therefore, the following hypotheses are proposed: Hypothesis2: Women participating in emotion management are less likely to develop social capital than men participating in emotion management.

Hypothesis3: Employees of color participating in emotion management are less likely to develop social capital than white employees participating in emotion management.
Recognizing the lack of scholarship emphasizing emotion management between employees, the current study examines the relationship between the production of emotion management among federal employees and the development of social capital. This study emphasizes employee interactions in an attempt to further the field's understanding of emotion management in public organizations and the implications that follow.

Data and Variables
The data for this study comes from Path 2 of the 2016 Merit Principles Survey which is produced by the U.S. Merit Systems Protection Board. The Protection Board stratifies its sample by federal agency and supervisor status (i.e., non-supervisor, supervisor, executive). The survey is administered electronically via e-mail to 37,397 federal employees representing 24 federal agencies. With a response rate of 38.7%, there were 14,473 employees that responded to the 2016 survey. After restricting the sample to only non-teleworking and nonmilitary federal employees, more than 3,000 observations were retained for this study. See Table 1 for a complete list of all variables along with their means, standard deviations, and minimum and maximum values.

Dependent Variable
Social capital is the dependent variable. This study operationalizes social capital through responses to six survey questions, each possessing a five-point Likert-type scale with values ranging from strongly disagree to strongly agree. Participants responded to the following prompts: 1) I like the quality of relationships I have with my coworkers, 2) my judgement is trusted and relied on at work, 3) I feel personally cared for at work, 4) I feel fully appreciated at work, 5) I am treated with respect at work, 6) my opinion counts at work.
These questions connect to the relational characteristics of Putnam's (1995) definition of social capital, emphasizing the strength of networks, trust, and reciprocity. Since social capital is considered a latent construct (Andrews, 2011), iterated component factor analysis with promax rotation was used in this analysis (Costello & Osborne, 2005). The factor analysis results indicate that all six survey questions load onto a single factor. Results for this analysis are presented in Table 3.

Independent Variables
The independent variable of interest is emotion management. This concept is operationalized through three survey questions, each having a five-point Likert-type scale ranging from strongly disagree to strongly agree. These survey questions follow those used by Puglesi (1999) in the development of an emotion index. Specifically, respondents provided responses to the following prompts: 1) I help coworkers feel better about themselves, 2) I attempt to keep the peace by calming clashes between coworkers, and 3) I help coworkers deal with stresses and difficulties at work. These questions align with definitions of emotion management 13 emphasizing personal exchanges between employees (Meier et al., 2006). Again, using iterated component factor analysis, each question loaded onto a single factor (see Table 3).

Control Variables
Various factors that could influence individual perceptions of social capital. In this study, a standard set of control variables was used to account for these factors. Specifically, included are separate controls for surface acting and deep acting. Previous research has shown that these forms of emotion regulation can have a significant impact on job satisfaction and burnout (Guy et al., 2008). Surface acting was measured using the prompt "My job requires that I hide my true feelings about a situation." Participants were asked to select a response along a five-point Likert-type scale with values ranging from strongly disagree to strongly agree. Deep acting was measured using the survey prompt "I try to actually feel the emotions I must display." For this prompt, respondents were also provided with a five-point Likert-type scale with values ranging from strongly disagree to strongly agree. The responses to both prompts were collapsed into three categories: agree, neither agree nor disagree, and disagree.
To account for basic demographic variations, respondents were asked if they identified as male (coded as "0") or female (coded as "1"). Respondents were also asked to self-identify as white (coded as "0") or a racial minority (coded as "1"). Age was controlled by having respondents disclose if they were under 40 years old (coded as "0") or over 40 years old (coded as "1"). Education was also taken into consideration with respondents disclosing whether they had less than an associate's degree, an associate's degree or bachelor's degree, or a graduate degree. These responses were ultimately collapsed into dichotomous categories, i.e., has not received a college degree (coded as "0") or has received a college degree ("1").
Previous research has shown that the gender composition of an organization can influence gender stereotypes (Ely, 1995) as well as the networking ability of female employees (Choi, 2018). As such, in order to control for these influences, respondents were asked if their organization was mostly composed of male employees, an even mix of male and female employees, or mostly female employees. Supervisor status was also taken into consideration by asking if respondents identified as a non-supervisor, a manager, or an executive within their organization. "I like the quality of relationships I have with my coworkers." "My judgment is trusted and relied on at work." "I feel personally cared for at work." "I feel fully appreciated at work." "I am treated with respect at work." "My opinion counts at work." Eigenvalue 0.70 0.76 0.84 0.85 0.83 0.87 4.24 Emotion Management Social Control "I help coworkers feel better about themselves." "I attempt to keep the peace by calming clashes between coworkers." "I help coworkers deal with stresses and difficulties at work." Eigenvalue "Spread excitement about work to others." "Look for ways to help other with work." "Take initiative to collaborate with others on work." "Foster work related discussions among my colleagues. Social capital can also be impacted by an individual's duration at their current agency. To control for this, respondents were asked if they had been at their current agency three years or less (coded as "0") or four years or more (coded as "1"). This study also takes the primary work location of employees into consideration, with respondents being asked whether they spend most of their time in the field (coded as "0") or at the organization's headquarters (coded "1"). Additionally, respondents were asked if they were nonunion members (coded as "0") or union members (coded as "1"). Membership should allow employees to make additional connections with their colleagues.
Lastly, this study incorporates a social control variable through an index of the following prompts indicating how much each respondent participated in the following activities: 1) spread excitement about work to others, 2) look for ways to help others with work, 3) take initiative to collaborate with others on work, and 4) foster work related discussions among my colleagues. Each prompt was on a five-point Likert-type scale with values ranging from strongly disagree to strongly agree. Following the same method used to generate social capital and emotion management, iterated component factor analysis was used; and, each question loaded onto a single factor. Incorporating this social control was necessary in order to account for employees who are naturally more social and, thus, may build social capital more easily. See Table 2 for the correlation matrix depicting relationships between all variables included in the analysis.

Results
The models are estimated using ordinary least-squares (OLS) regression. The first model addresses whether there is a relationship between the production of employees' emotion management and social capital. The second model addresses whether gender and race have an impact on the relationship between emotion management and social capital. This model also includes interactions between emotion management and both gender and race. The models are presented in Table 4.
Model 1 in Table 4 shows that emotion management has a statistically significant (β=0.07, p <0.001) and positive relationship with social capital. This provides evidence in support of the first hypothesis that participation in emotion management is associated with the attainment of social capital. This result suggests that employees capable of emotion management are more likely to reap the benefits of social capital.
Model 2 shows that the relationship between emotion management and social capital remains positive and statistically significant (β=0.13, p<0.001). The coefficient for the first interaction term (i.e., Female × Emotion Management), however, is negative and not statistically significant. Thus, the second hypothesis, that women participating in emotion management are less likely to develop social capital than men participating in emotion management, cannot be confirmed.
The interaction between race and emotion management also has a negative relationship with social capital; however, it is statistically significant (β=-0.11, p<0.05). This finding provides support for the third hypothesis that employees of color participating in emotion management do not develop social capital at the same level as white employees participating in emotion management. This effect is depicted in Figure 1. This finding aligns with previous studies that have shown that black employees, in particular, struggle to develop strong ties and build social capital within their organization to the same extent as white employees (Ibarra, 1995;James, 2000). Prior research has also shown that when organizational members form relationships with individuals of the same race, they perceive those relationships as providing more psychosocial support than relationships with those of a different race (Thomas, 1990). Stated another way, when individuals develop relationships with someone of the same race, they are more likely to perceive that relationship as having social utility. In the context of this analysis, this could indicate that while employees of color may engage in emotion management, the relationships they build may not feel as socially beneficial (if their organization is predominately white) as the relationships developed between two employees who identify as white.

Figure 1. Interaction Between Emotion Management and Race
While not directly related to the interests of this study, another interesting finding is the negative and significant relationship between surface acting and social capital in both models (β=-0.62, p<0.001 in model 1; and, β=-0.61, p<0.001 in model 2). These findings indicate that employees participating in surface acting perceive themselves as having less social capital than those that are not participating in surface acting. Other research has found surface acting is positively related to burnout (Guy et al., 2008). This study provides further evidence that surface acting has a negative influence on the experiences of organization members.

Discussion and Conclusion
Understanding of emotion management between employees remains limited in the field of public administration. Scholarship in this field has often emphasized the emotional labor that takes place between employees and clients and the impacts it has on job satisfaction (Jin & Guy, 2009;Kaur & Malodia, 2017;Yang & Chang, 2008), burnout (Guy et al., 2008;Jin & Guy, 2009), and performance (Meier et al., 2006). This study addresses an important gap in the literature by examining the association between emotion management between colleagues and social capital using a sample of federal employees. The results provide evidence that emotion management is positively associated with social capital, suggesting that employees willing to invest in interpersonal relationships with their colleagues may reap the benefits of social capital. However, the results also suggest that emotion management by employees of color is less likely to generate social capital than it is for their white counterparts.
These findings have implications for theory and practice. In terms of theoretical relevance, this study presents an initial understanding of the relationship between emotion management and social capital. While both concepts are relational, previous research has not extensively examined the dynamics between the two. Indeed, foundational work in public administration often suggested that for organizations and their members to behave as rational actors, they must to be devoid of emotions (Guy & Newman, 2004). The current study, however, provides evidence that emotions are not only present in public organizations but can help employees connect with their colleagues and develop social capital.
In terms of practical implications, this study confirms the findings of previous research showing that employees of color often struggle to develop social capital at the same level as their white counterparts (Hero, 2003). Since an inability to access social capital can restrict career advancement (Choi, 2018), managers of public organizations should be cognizant of how emotion management can influence this dynamic. These managers should also seek to find ways to ensure that all employees are able to make connections with their colleagues.
While this study provides useful contributions to the field of public administration, it has several limitations that should be acknowledged. First, the study lacks a detailed measurement to disaggregate employees of color. Rather than obtaining each respondent's race, the survey uses an aggregate measure that encompasses all respondents identifying as a racial minority. Future research should focus on disaggregating employees of color in order to understand how the relationship between emotion management and social capital might change based on race.
Second, the study lacks a robust measurement of emotion management. Surveys often do not include a substantial number of questions specifically examining emotion management between employees due to scholars being interested in the emotional labor produced when interacting with clients. Future surveys, however, should consider additional measures of emotion management.
Finally, this study lacks a control variable identifying the racial composition of the organization. With gender composition influencing gender stereotypes (Ely, 1995) and the ability of women to make connections within their organizations (Choi, 2018), racial composition of an organization may have similar effects. Despite each of these limitations, this study presents a deeper understanding of emotion management and its influence on social capital, particularly in public organizations.

Disclosure Statement
The author declares that there are no conflicts of interest that relate to the research, authorship, or publication of this article.