Moderators of the relationship between turnover intentions and performance

Abstract

Purpose

Drawing from the theory of reasoned action, this study investigated the moderators of the relationship between turnover intentions and performance at work.

Design/methodology/approach

This study employed hierarchical multiple regression to test three proposed hypotheses regarding the above relationship. It used 1,011 dyad data from employees and their supervisors from eight professional organizations in Vietnam from employees and their supervisors to reduce research bias.

Findings

Employee attitude toward change and the level of job engagement of the employee affected the nature of the relationship between turnover intentions and job performance. When the attitude toward change was less favorable, the relationship between turnover intentions and job performance was positive. However, when the attitude toward change was more favorable, the relationship between turnover intentions and job performance was non-significant. For the moderating role of job engagement, we found that for employees with a high level of job engagement, the relationship between turnover intentions and job performance was positive. However, for employees with a low level of job engagement, the relationship between turnover intentions and job performance was non-significant.

Practical implications

Unlike the implications from previous research, turnover intentions of employees might not adversely affect their performance. Under two conditions – a high level of job engagement and a less favorable attitude toward change - employees with turnover intentions might actually perform better.

Originality/value

Unlike the vast number of studies that have investigated the relationship between job performance and turnover intentions (as a proxy of turnover), this paper focuses on the relationship between turnover intentions and job performance to show evidence for two important boundary conditions.

Citation

Bui, H.T.M., Pinto, J., Tran Vu, A.V.H., Mai, N.T. and Nguyen, T.Q. (2024), "Moderators of the relationship between turnover intentions and performance", Journal of Trade Science, Vol. 12 No. 2, pp. 82-99. https://doi.org/10.1108/JTS-02-2024-0006

Introduction

The relationship between an employee’s performance at work and their continued employment is a fundamental aspect of work and management (Oh et al., 2024). Conventional wisdom implicitly assumes that dissatisfied employees, including those with turnover intentions, leave while satisfied employees stay (Li et al., 2016). However, the extensive research on this relationship and the two focal variables, namely job performance and employee turnover, reveal a more complex and unclear picture. Some of the contributory factors discussed below include the difficulty in operationalizing these variables (Klotz et al., 2021), the temporal aspects of quitting or employee turnover (Xue et al., 2024; Peltokorpi et al., 2023) and the implications of these temporal aspects or process (cf. exit transition process, Klotz et al., 2021), such as the direction of the relationship between employee performance and quitting. The reverse case, that turnover intention (proxy metric for quitting) impacts job performance is examined in this study. Further, the role of two individual-level moderating variables, employee attitude to change and the level of employee’s job engagement, are investigated to find evidence for two important boundary conditions.

Job performance and employee turnover are two of the most important outcome variables in personnel psychology, industrial-organizational (I-O) psychology and organizational behavior research areas (Jackofsky, 1984; Peltokorpi et al., 2023; Wang et al., 2022). Employee turnover and intention to quit have been among the most widely investigated organizational phenomena (Hur and Abner, 2024; León and García-Saavedra, 2021; Ng et al., 2019) and are among the most crucial employment relations outcomes for employees and organizations (Batt and Colvin, 2011; Nyberg and Ployhart, 2013).

Although accurate and reliable data on job performance can be obtained relatively easily, measuring employee turnover for research purposes is more complicated. First, it is argued that turnover needs to be reconstructed as a useful dependent variable. Scholars have distinguished between involuntary turnover (i.e. poor performers leave) and voluntary turnover, and between voluntary turnover and functional turnover (i.e. good performers leave). Dysfunctional turnover has been of greater interest to both scholars and practitioners (Batt and Colvin, 2011). Second, although individual-level quitting can be observed easily, reliably and objectively, it is difficult for researchers to capture information on employee attitude and other behaviors at the time of quitting, and a longitudinal design is required, which is more difficult to implement (Nyberg and Ployhart, 2013). Hence, scholars drawing on Fishbein and Ajzen’s theory of reasoned action (1975) have developed the “intention to quit” construct (Firth et al., 2004) which has been variously labeled as turnover intentions (Tett and Meyer, 1993), intended turnover (Werbel and Bedeian, 1989), intention to leave (Robinson and O’Leary-Kelly, 1998) and behavioral intentions to leave (Hulin et al., 1985); all these terms are treated as synonyms in this study. Since this construct is measured through a scale (Lapointe et al., 2013), it is a continuous variable and therefore, amenable to a wider range of statistical analyses that are easier to interpret. As a result, largely due to these reasons, scholars have often used turnover intentions or intentions to quit as a surrogate measure of actual turnover (Peltokorpi et al., 2023).

Much individual-level research on the relationship between job performance and turnover has automatically considered the former as an antecedent of the latter (since once an employee quits, there can be no job performance). Jackofsky (1984) developed the classic integration of job performance in the process model of turnover, and since intention to quit is a surrogate or proxy for turnover (Meyer et al., 2002), several studies have treated job performance as an antecedent of turnover intention derived from this conceptual model. However, turnover intention is not the same as actual turnover (Zimmerman and Darnold, 2007), as employees may have an intention to quit but may not necessarily quit (Klotz et al., 2021; Peltokorpi et al., 2023). Further, attitudes about job dissatisfaction typically explain only around 5% of turnover variance, and intentions to quit rarely exceed 10 to 15% of variance (Allen et al., 2005). In some cases, turnover intentions and job performance have automatically been treated as joint outcomes or correlates without their causal relationship being examined (Ozyilmaz et al., 2018).

Conventional academic wisdom implicitly assumes that dissatisfied employees, i.e. those with turnover intentions, leave and satisfied employees stay (Li et al., 2016). However, recently there have been interests in this area, with theoretical developments such as the pre-quitting behavior construct (Gardner et al., 2018) and the proximal withdrawal states theory (Li et al., 2016). This study contributes to this research stream, which underscores the importance of a closer examination of the linkages between turnover intentions and job performance.

Literature review

Understanding turnover is significant to human resources and career management (Verbruggen and van Emmerik, 2020). In this section, the focus is on one specific turnover cognition variable, the antecedent variable, turnover intention (Klotz et al., 2021; Peltokorpi et al., 2023), as it is significant with regard to employee satisfaction (with their job, the organization or both), and may therefore impact job performance among other variables, such as career satisfaction (Verbruggen and van Emmerik, 2020). Further, the turnover intention construct is not as well understood as job performance and unpacking it is necessary because it underpins the hypothesis development. The turnover intentions research stream has its foundations in cognitive models of turnover in which the three primary partial determinants of voluntary turnover are: (1) desirability of movement from the organization; (2) ease of movement to another organization; and (3) intention to quit (Jackofsky, 1984). The last of these, intention to quit, has been theorized as the final cognitive variable immediately preceding voluntary turnover (Mobley et al., 1979), and subsequent empirical research has supported this approach (Carter et al., 2020). Intention to quit is defined as “an individual’s perceived probability of staying in an employing organization or terminating employment” (Werbel and Bedeian, 1989, p. 275), whereas turnover intention is a conscious and deliberate willingness to leave the organization within a specific time interval, e.g. six months, and is the last in a sequence of withdrawal cognitions (Tett and Meyer, 1993). As mentioned earlier, these terms are used interchangeably in this study. The intention to quit construct is drawn from the theory of reasoned action, which explains that a person’s intention to perform a specific behavior is the immediate determinant of the behavior (Fishbein and Azjen, 1975). The intention to quit has been empirically found to be the most powerful predictor of turnover behavior (Peltokorpi et al., 2015).

Recent research extends and elaborates upon this foundational work. For instance, Hom and colleagues (Hom et al., 2012; Li et al., 2016) developed the proximal withdrawal states theory, which distinguishes four types of employees: (1) enthusiastic stayers (those who want to stay and can stay); (2) enthusiastic leavers (those who want to leave and can leave); (3) reluctant stayers (those who want to leave but feel they must stay); and (4) reluctant leavers (those who want to stay but feel they must leave). The proximal withdrawal states theory was developed because prevailing theories focused primarily on enthusiastic stayers and leavers and neglected reluctant stayers and leavers. Reluctant stayers may be constrained in leaving jobs they dislike but reluctant leavers are often pressured to leave a job they want to keep (Li et al., 2016). This potential variety among employees regarding their turnover intentions, their volitional control over leaving or staying and their consequent psychological states (Li et al., 2016) suggests the relationship between turnover intentions and job performance may also be more complex and nuanced than conventional academic wisdom indicates.

In addition, Gardner et al. (2018) developed the pre-quitting behavior construct in which they distilled 58 prototypical pre-quitting behaviors of employees to 12: (1) their work productivity has decreased more than usual; (2) they have acted less like a team player than usual; (3) they have been doing the minimum amount of work more frequently than usual; (4) they have been less interested in pleasing their manager than usual; (5) they have been less willing to commit to long-term timelines than usual; (6) they have exhibited a negative change in attitude; (7) they have exhibited less effort and work motivation than usual; (8) they have exhibited less focus on job-related matters than usual; (9) they have expressed dissatisfaction with their current job more frequently than usual; (10) they have expressed dissatisfaction with their supervisor more frequently than usual; (11) they have left early from work more frequently than usual; and (12) they have lost enthusiasm for the mission of the organization. The variety inherent in this final pre-quitting behavior scale points to the complexity of the relationship between turnover intentions and job performance.

Hypotheses development

Scholars researching the relationship between job performance and turnover or intention to quit have largely treated the former as an antecedent and the latter as a dependent variable. A few have considered the reverse relationship, i.e. the effect of intention to quit on job performance, and this study’s hypotheses are developed based on this direction. The relationship between turnover and performance has been studied at both the organization-level (Arthur, 1994) and the individual-level (Hulin et al., 1985). At the organizational level, Arthur (1994) found that turnover level and job performance were negatively related, and this relationship was stronger in human resource systems that concentrated on commitment maximization than those focusing on cost reduction.

At the individual-level, Hulin et al. (1985) theorize that when employees are dissatisfied with their jobs, it results in one or more of three types of withdrawal behavior intentions: (1) to reduce job inputs (resulting in psychological job withdrawal); (2) to change work situation (e.g. unionization activity, transfer attempts); and (3) to quit. Thus, intention to quit could occur concurrently with psychological withdrawal or increased non-job activity, such as attempts to unionize or engineer a transfer. In either case, it would lead to poorer job performance, and turnover intentions would be negatively associated with job performance. Further, the longer the time period that the individual maintains the intention to quit but cannot actually quit, the greater the chances of developing feelings of helplessness, which could cause a decline in job performance (Hulin et al., 1985). Another causal explanation of this relationship is that job performance could be regarded as an employee investment in an organization that would yield positive returns over the course of their employment. However, if employees intend to leave they would be less willing to make the investment as they may not obtain the returns (Hui et al., 2007). This theorization appears to be supported by the empirical investigation of this relationship according to Hui et al. (2007), who found evidence of a negative relationship between turnover intention and job performance. In addition, Werbel and Bedeian (1989) found that older employees with high intention to quit had poorer performance, whereas those who had low intention to quit put greater effort in their current job performance.

Research on the recently developed pre-quitting behaviors construct (Gardner et al., 2018) also supports this line of reasoning. Several items in the final pre-quitting behaviors scale (e.g. productivity has decreased more than usual, doing the minimum amount of work more frequently than usual) suggest that turnover intentions are negatively associated with job performance. According to Gardner et al. (2018), individuals with a strong intention to quit may exhibit less focus and concentration on the job. Further, they posit that because people have a finite supply of self-control (Muraven et al., 1998), they may find it difficult to consistently hide their negative feelings or behaviors, and this mental fatigue of living a “double life” (superficially positive but negative at the core) may adversely impact job performance. Gardner et al. (2018) also give the example of a salesperson who, intending to quit, may no longer generate the enthusiasm necessary for the product or service they sell. Thus, in line with extant research, a negative relationship between turnover intention and job performance is predicted.

H1.

Turnover intention is negatively related to job performance.

However, as it has been pointed out, the relationship between turnover intentions and performance appears to be complex and nuanced. The relationship depends, inter alia, on the mix between enthusiastic stayers and leavers and reluctant stayers and leavers. Li et al. (2016) found that reluctant stayers and leavers differed significantly from enthusiastic stayers and leavers regarding the relationship between intention to quit and turnover behavior. Further, job satisfaction and job embeddedness strongly influenced the intent to leave and job search behavior of enthusiastic stayers and leavers more than of reluctant stayers and leavers (Li et al., 2016). Thus, if there were more reluctant stayers and leavers, their turnover intentions may not be job-related and therefore, may not be associated with lower job performance. In fact, reluctant leavers might improve their job performance because their opportunity to continue to perform the job would be ending against their own wishes, and they might want to make the most of it and try to keep options open for a future return to that organization. This is only one of the various situations in which employees with turnover intentions may direct their efforts to enhance job performance rather than decrease it. Therefore, the role of two such potential situational moderators that would provide the boundary conditions in which the hypothesized relationship might change is investigated (Cortina, 2003).

One moderator for this study is organization-related, and another job-related since these are arguably the two most important situational dimensions that influence the termination of the employment relationship (e.g. Liu et al., 2012; Smith et al., 2011). For instance, the job and the organizational context are two of the three fundamental issues in the study of voluntary turnover (Liu et al., 2012). In addition, all four conceptual antecedents (Smith et al., 2011) of voluntary turnover could be categorized under two dimensions: job-related (which would include job satisfaction, job alternatives and job embeddedness) and organization-related (i.e. organizational commitment). Similarly, on the empirical side, De Moura et al. (2009), found support for two predictors of turnover intentions, one organization-related (i.e. organizational identification) and the other job-related (i.e. job satisfaction).

The organization-related moderator in this study is attitude to change, which “concerns an individual’s beliefs and intentions regarding the extent to which change is needed and whether there is organizational capacity to make that change work successfully” (Lee et al., 2009, p. 627), and the job-related moderator is job engagement, which is the positive contrast of burnout (Saks, 2006). The attitude to change construct has been selected because: firstly, it is predominantly influenced by organizational commitment (Chih et al., 2012), one of the four conceptual perspectives of voluntary turnover (Smith et al., 2011); and secondly, it relates to organizational change which could be one of the “shocks” that could precipitate turnover more than job dissatisfaction (Mitchell and Lee, 2001).

Moderating role of attitude toward change

By definition, turnover implies a change of organization, and clarifies that “every individual experiences change in a unique way” (Bouckenooghe, 2010, p. 501). Hence, an individual’s attitude toward change is required to fully understand the relationship between turnover intention and job performance. Attitude toward organizational change is defined as “an employee’s overall positive or negative evaluative judgment of a change initiative implemented by his or her organization” (Elias, 2009, p. 39). Consistent with research on attitudes, in general, research on attitudes toward change has considered the cognitive, affective and behavioral dimensions (Yousef, 2000).

Individuals with a less favorable attitude toward organizational change may react to planned changes with an intention to quit, but their job performance may not be adversely affected. Such employees would be included in the reluctant leaver’s category of the proximal withdrawal states theory, i.e. those who have a high preference to stay but perceive low control over their staying (Li et al., 2016) and feel forced to leave for reasons independent to how they feel about their job (Li et al., 2016), and the expectation is that their job performance would not be affected. Indeed, Li et al. (2016) found that reluctant leavers were like enthusiastic stayers in terms of aspects such as affective commitment and job satisfaction and therefore, their job performance may not decline but may, in fact, improve just like enthusiastic stayers.

Further, Devos et al. (2007) found that organizational changes resulting in severe job losses lead to negative attitudes toward organizational change, compared to organizational changes that do not result in job losses. Building on this finding, in situations of downsizing or right-sizing, employees may have a less favorable attitude toward change and may increase their job performance to stave off involuntary turnover (i.e. being fired) while simultaneously preparing for the worst by looking for alternative employment. We hypothesize that:

H2.

The relationship between turnover intentions and job performance is moderated by attitude toward organizational change such that when the attitude is less favorable, the relationship between turnover intention and job performance is more strongly positive compared to the situation when the attitude toward change is favorable.

Moderating role of job engagement

Job or work engagement is defined as “a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption” (Schaufeli et al., 2002, p. 74). Vigor is defined as a “high level of energy and mental resilience while working, the willingness to invest effort in one’s work, and persistence even in the face of difficulties” (Schaufeli et al., 2002, p. 74) Dedication is described as “a sense of significance, inspiration, pride, and challenge” (Schaufeli et al., 2002, p. 74). Finally, as a component of job engagement, absorption refers to “the state of being fully concentrated in one’s work, whereby time flies when working, and employees have difficulties with detaching from working” (Caesens et al., 2014, p. 1). Engagement implies that employees are experiencing positivity at work and would not be likely to quit. Hence, not surprisingly, research has generally found a negative correlation between engagement and turnover intentions (Halbesleben, 2010).

However, recent research (Caesens et al., 2014) found evidence of a curvilinear relationship between engagement and turnover intentions and argued that, based on the reciprocity principle, highly engaged employees would have high expectations of their organization and if these are not met, the employees might consider joining other organizations. Another argument advanced by Caesens et al. (2014) is based on Maertz and Griffeth’s (2004) framework of eight motivational forces in general, and calculative force in particular. Calculative force is described as the “rational calculation of the probability of attaining important values and goals in the future through continued membership” (Maertz and Griffeth, 2004, p. 669). Caesens et al. (2014) suggest that highly engaged employees have higher work goals, and therefore, would be more susceptible to the calculative force and might think their current organization may not be able to fulfill their needs and thus start looking for work elsewhere.

The above line of argument suggests that highly engaged employees may have turnover intentions, but this may not necessarily lead to lower job performance Instead, it may lead to higher job performance for the following reasons. Firstly, highly engaged employees, compared to less engaged employees, take greater pride in their work and would not lower standards and let their performance deteriorate. Secondly, even if their intention to quit has been triggered by issues at the workplace, highly engaged employees, by definition, have the mental resilience to persist in the face of difficulties. Thirdly, drawing on Caesens et al. (2014), highly engaged employees may have decided to quit due to calculative logic or force, which is cognitive and has no associated negative emotions. In fact, Maertz and Griffeth (2004) delineate an affective force distinct from the calculative force. Consequently, there would be no dissatisfaction or withdrawal behaviors and performance would not be lower. Rather, by applying calculative logic, highly engaged employees may increase their performance to burnish their resumes and track record of achievement, which would improve their chances of moving to a more suitable organization. Thus, we propose that when highly engaged employees have turnover intentions their performance increases, and this relationship is stronger, the more highly engaged the employees are.

H3.

The relationship between turnover intentions and job performance is moderated by job engagement such that when there is high job engagement, the relationship between turnover intention and job performance is more strongly positive compared to the situation when job engagement is low.

Figure 1 presents the framework of this study.

Method

Sample

To test the hypotheses, we collected survey data from employees and their supervisors at eight large research and education organizations in Vietnam that each had a minimum of 500 employees at all levels. Vietnam is an emerging economy in Southeast Asia with a prime population, i.e. more than 65% of the population under the working age range. White-collar workers are the spine of a knowledge economy and understanding their behaviors related to performance is vital for that context.

We followed a stratified sampling procedure to get a representative sample (Parsons, 2017) and approached the top management teams of the chosen eight organizations to gain access to their staff. One manager worked with us to decide which departments would participate in the survey and department heads circulated two surveys to all employees (for the employee survey) and first-line managers (for the manager survey). While employees answered questions on the predictor of performance and mediators in this conceptual model, supervisors rated the employees on their job performance.

1,026 completed questionnaires from employees and 1,011 matching responses from supervisors for job performance were received. In terms of demographic profile, women represented 57.3% of the total, and the age range was from 23 to 65 years, with an average age of 39 years. Nearly 75% of the respondents had either a master’s or a doctoral degree.

Measures

Turnover intentions

We measured this variable with two items. These items are similar in content to the versions of two-item scales used by Lapointe et al. (2013). An item from the scale was: “I often think about quitting this organization.” The responses for this scale were measured on a Likert-type scale. The reliability coefficient was α = 0.74.

Attitude toward organizational change

Three items from the scale developed by Dunham et al. (1989) were used in the survey. A sample item was: “I am usually eager with new ideas.” While we used four items in the survey, one item was dropped as its exclusion improved the scale reliability coefficient. Since it is desirable to include more indicators per factor for reasons of convergence and interpretability (Marsh et al., 1998), no more items were dropped. The reliability coefficient of the 3-item scale was α = 0.86. (The hypotheses testing results were similar in nature for both versions of the scale.)

Job engagement

Three items to measure job engagement were based on a scale developed by Saks (2006). A sample item from the scale was: “Sometimes I am so into my job that I lose track of time.” While we used five items in the survey, two items were dropped as their exclusion improved the scale reliability coefficient. The reliability coefficient of the 3-item scale was α = 0.82. (The hypotheses testing results were similar in nature for both versions of the scale.)

Job performance

There is no single performance measure appropriate to all purposes (Behn, 2003). In organizational psychology studies, employees rating their own performance (or self-report) is seen to be subjective, and managers rating employee performance is considered more objective (Podsakoff et al., 2003). That is why manager ratings are widely used to reduce performance bias (e.g. Scott and Zweig, 2021), and similarly, information on employee job performance was obtained from their supervisor. Since the sample consisted of employees across different job roles (professional, managers and administrative staff), each job’s items were worded differently. The performance of managers and researchers was measured with three items and the performance of professionals and administrative staff was measured by four items on a seven-point Likert-type scale. Since this study’s goal was to address performance across different job roles, new scales were developed for this study based on the contextual understanding of these organizations, and a review of items used to measure job performance in previous research. For example, the scale measuring the performance of managers referred to their effectiveness in planning, implementation and control functions. The scale reliability (Cronbach’s alpha) for the job performance scales for different job categories were as follows: 0.85 (professional), 0.90 (managers) and 0.90 (administrative staff).

The initial measures were in English. The committee approach, back-translation and a pre-test procedure were employed (Bui et al., 2021) to prevent any methodological problems associated with the translation from one language to the other (Sperber et al., 1994). The process was done by three different researchers to ensure all items made sense to the respondents.

Control variables

The possible confounding effects of age, gender, educational qualifications (four levels) and job category of the employee were statistically accounted for. Employees were classified into four job categories: teachers, researchers, managers and administrative staff. We created three dummy variables for teachers, researchers and administrative staff, with managers omitted to control the effect of the job category.

Discriminant validity

Three variables in the model - turnover intention, employee engagement and attitude toward change – indicated individual disposition and were self-reported by the employee, so we conducted a confirmatory factor analysis (CFA) to ensure the three measures were distinct from one another. The three-factor model was compared with the one-factor model and different combinations of the two-factor models. One-factor solution did not have a good fit (chi-square = 1,548.61, df = 20, p < 0.05; RMSEA = 0.286). Three-factor solution had an acceptable fit (chi-square = 102.17, df = 15, p < 0.05; RMSEA = 0.07). There was a significant improvement in model fit when we compared the three-factor model with the one-factor model (change in chi-square = 1,446.44, change in degrees of freedom = 5, p < 0.05). Similarly, compared to any of the three two-factor models, there was a significant improvement in model fit with the three-factor model.

Results

The descriptive statistics and correlations are shown in Table 1. Hierarchical multiple regression was used to test the main effect of turnover intention as well as its interaction effects. Supplemental analysis was performed to test the moderating effects. The results are shown in Table 2. As can be seen in the first block of variables in Table 2, age and gender did not impact much on individuals’ performance but qualifications did. It showed that the higher qualifications people had, the higher level of performance they were rated by their line managers.

The second block of variables in Table 2 shows that turnover intention was not related to job performance (β = 0.06, ns), and thus, H1 was not supported. In other words, employees who have the intention to leave might still perform, and intention does not necessarily adversely affect performance, possibly because they want to show they are responsible workers. In addition, the pool of employment in professional jobs is not large. People can leave and return to an organization. Many want to show they are dedicated employees for future job references. This is because professional networks have become stronger due to the development of social media (Davis et al., 2020) and people getting to know one another more easily. With such a null hypothesis, testing moderating relationships is even more important for understanding the boundary conditions under which the relationship in H1 might change.

H2 and H3 were about the moderating effects of attitude toward organizational change and job engagement. We tested these hypotheses by including two multiplication terms, turnover intentions and attitude toward organizational change and turnover intentions and job engagement, after mean-centering these variables. After accounting for the main effects of these variables, the interaction terms were significant in both cases in the third block of variables in Table 2. Thus, there was a preliminary likelihood of H2 and H3 being supported.

To identify the nature of moderating effects, the procedure outlined by Aiken et al. (1991) was followed as the values of job performance for low and high values of turnover intention under conditions of low and high values of the moderators were plotted. The low and high values were based on the criteria of one standard deviation below and one standard deviation above the mean value, respectively. Figure 2 shows the moderating effect of attitude toward change. For unfavorable attitudes toward organizational change, the relationship between turnover intentions and performance was positive. On the other hand, for favorable attitude toward organizational change, the relationship between turnover intentions and performance was negative. Thus, H2 was supported.

Similarly, Figure 3 shows the moderating effect of job engagement. When job engagement was low, the relationship between turnover intentions and performance was negative. However, when job engagement was high, the relationship between turnover intentions and performance was positive. Therefore, H3 was supported.

Discussion

This study aimed to examine the relationship between turnover intentions and job performance, which has been sparsely researched compared to the reverse relationship, between job performance and turnover intentions. Given the complexity of the relationship between these two variables, a more important purpose of this research was to identify the boundary conditions of the relationship between turnover intentions and job performance. This study focused on two potential moderators – attitude toward change and job engagement - and collected survey data from employees and obtained job performance data from their supervisors, thereby avoiding the problems associated with self-reported performance data.

Theoretical implications

This study shows there is no significant relationship between turnover intentions and job performance. There are two main reasons why a lack of significant relationships might not be an anomaly. Firstly, although there has been a vast amount of research on the relationship between performance and turnover intentions, as indicated by a recent meta-analysis on this topic (Zimmerman and Darnold, 2007), a converse relationship might not work the same way. For example, whereas employees with a high job performance might be more marketable and could plan to leave their jobs, employees may develop turnover intentions for several reasons (e.g. dissatisfaction with their jobs, abusive supervision, dual-career issues) that may not necessarily be related to their job performance. In other words, to express this in the context of proximal withdrawal states theory, all leavers may not be enthusiastic leavers, and some may, in fact, be reluctant leavers. Once an employee develops an intention to quit, their motivation to perform could change in ways that could boost or harm performance. Some employees may work toward improving their performance to negotiate better terms of employment elsewhere, whereas others may demonstrate withdrawal behaviors, thereby hurting their performance. Secondly, fewer studies treat turnover intention as the predictor of job performance, and it is possible the relationship may just not be strong in the overall workforce. For example, Bauer et al. (2006) reported a near-zero relationship between turnover intentions and job performance, with job performance measured with a time lag following the measurement of turnover intentions. While Bauer et al. (2006) did not primarily focus on the relationship between these two concepts (i.e. turnover intention and job performance) in their large-scope study, the reported null relationship is consistent with this study’s findings.

More importantly, a non-significant relationship between turnover intentions and job performance could indicate important boundary conditions under which the relationship might vary in different directions, with the possibility of an overall null effect. The findings contribute to the literature on turnover intentions and performance by emphasizing the roles of two such moderators: attitude toward change and job engagement. Miller et al. (1994) identified several reasons why employees could develop a negative attitude toward change and might not support such initiatives in the organization. It is unlikely that the reasons identified by the authors – individual inertia, political coalitions, investment in the status quo and lack of motivation to change one’s behavior – characterize an employee who would work towards improving their performance. This study’s findings suggest employees with a less favorable attitude toward organizational change might trigger positive behaviors. A possible reason might be that, under those conditions, the employee views these positive behaviors as instrumental in achieving other objectives, such as finding a job outside the organization.

Higher employee job engagement might activate the desire to take control of the situation even when employees are planning to quit the organization. One possible reason is that employees with an internal locus of control are likely to be more engaged in their jobs (May et al., 1997). This tendency to take control of the situation perhaps enables employees to focus on performance-oriented behaviors, even when they intend to leave the organization. While job engagement and turnover intentions have been treated as correlates in past research (Halbesleben, 2010), this study adds to the literature by demonstrating their joint effect on job performance.

Practical implications

These findings have important implications for practitioners. Firstly, it is useful to know that employees' turnover intentions might not adversely affect their performance, unlike the implications from previous research. Under certain conditions, employees with turnover intentions might perform better. This research identified two such conditions: a high level of job engagement and a less favorable attitude toward change. As Rich et al. (2010) suggested, job engagement might be an important factor that managers could influence to enable higher job performance. Several factors related to employee selection and the work context might affect employee engagement, for example, those with an internal locus of control and who perceive their own values to be congruent with the organizational values might be more engaged in their jobs (Rich et al., 2010). Perceived supervisor support might also enhance job engagement (Swanberg et al., 2011). Thus, managerial influence on factors that positively affect job engagement could ensure good performance even if employees intend to quit.

In the organizational change literature, positive employee attitude towards change is generally considered desirable (Chen and Wang, 2014). While we do not dispute the practical importance of having employees with a positive attitude towards change, this study’s findings suggest that a less favorable attitude toward change might not hurt employee performance. In fact, for employees who intend on leaving the organization, a less favorable attitude might motivate them to work harder and improve their performance. Of course, this study did not investigate results of such an employee orientation. It is possible a good employee performance might eventually lead to their exit from the organization as more attractive options become available.

Limitations and future research

One of this study’s key implicit considerations is that there is a temporal ordering such that intention to quit precedes job performance. However, since cross-sectional rather than longitudinal data are collected for this study, it is not possible to resolve the issue of time order of cognitions/behaviors. Since past research has separately examined each direction of the relationship, it might be helpful to conduct a longitudinal study that investigates whether one direction is more evident than the other. It is also possible that the relationship unfolds in both directions over time, for example, high/low job performance might increase turnover intentions, which in turn might lower job performance. In both these relationships, there might be important moderators, such as attitude toward change and job engagement might attenuate or enhance the relationship between turnover intentions and performance, and the vice versa.

Another significant limitation of this study is that, even though information on job performance was obtained from the supervisor rather than the employee, it is also helpful to use objective measures of individual performance. However, most jobs involve multiple tasks and require subjective assessment by a supervisor in performance appraisal processes; therefore, there is value in considering a supervisor’s response to items of a job performance scale. In addition, this study was conducted in the Global South which is culturally different from Western countries where much research on this topic has been performed in the past. Future research must attempt to replicate these findings in other socio-economic contexts.

In addition, drawing on regulatory focus theory (Higgins, 1998), future research could investigate the role of promotion-focus and prevention-focus, particularly across cultures, regarding attitude to change, turnover intentions and job performance. For instance, Leon et al. (2015) found that prevention-focused individuals were more likely to develop turnover intentions in response to news of an organizational change, such as a merger, and these individuals would also have lower openness to change and may therefore perform relatively better. Since Vietnam is one of the fastest-growing economies, its society could be in the process of becoming more promotion-focused (Inglehart and Oyserman, 2004), and it would be interesting to see if these results would hold if the study was rerun in a few years. Further, future research could be conducted on national samples that are culturally distant from Vietnam to establish the generalizability of the findings.

Figures

Figure 1

Framework of proposed hypotheses

Figure 2

Moderating effect of attitude toward change in the relationship between turnover intentions and job performance

Figure 3

Moderating effect of job engagement in the relationship between turnover intentions and job performance

Table 1

Descriptive statistics and correlations

Variable	Mean	SD	1	2	3	4	5	6	7	8	9	10
1. Turnover intentions	2.56	1.57	0.74
2. Attitude towards change	6.14	0.89	−0.24**	0.86
3. Job engagement	5.03	1.24	0.01	0.25**	0.82
4. Performance (supervisor-rated)	5.86	0.87	0.04	0.04	0.06	0.85+
5. Age	39.16	10.61	−0.05	0.03	0.18**	0.05	--
6. Gender (male = 0)	–	–	−0.06	−0.03	−0.22**	−0.01	−0.32**	--
7. Qualification	2.07	0.79	−0.01	0.04	0.20**	0.05	0.42**	−0.31**	--
8. Job category (professional)	–	–	0.00	0.00	0.05	−0.02	−0.08*	−0.01	0.21**	--
9. Job category (managerial)	–	–	0.00	0.00	−0.04	−0.18**	−0.03	0.00	−0.01	−0.22**	--
10. Job category (office)	–	–	−0.04	−0.03	−0.12*	0.09**	−0.18**	0.21**	−0.52**	−0.52**	−0.11*	–

Note(s): The diagonal elements are reliability coefficients, where applicable

*p < 0.05 (two-tailed)

**p < 0.01 (two-tailed)

⁺ Job performance was measured with specific items for different job categories

Source(s): Table by authors

Table 2

Summary of hierarchical regression analysis for the relationship between turnover intentions and performance

Variable	Performance
	b	std. error	β
Block 1
Age	0.00	0.00	0.04
Gender (male = 0)	0.04	0.06	0.02
Qualification	0.14	0.05	0.12**
Job category (professional)	−0.02	0.07	−0.01
Job category (managerial)	−0.69	0.14	−0.16**
Job category (office)	0.27	0.10	0.12**
Block 2
Turnover intentions	0.03	0.02	0.06
Attitude toward change	0.04	0.03	0.04
Job engagement	0.02	0.02	0.03
Block 3
Turnover intentions x Attitude toward change	−0.09	0.02	−0.18**
Turnover intentions x Job engagement	0.06	0.02	0.13**

Note(s): R² for step 1 = 0.05, p < 0.01; ΔR² for step 2 = 0.005, ns; ΔR² for step 3 = 0.029, p < 0.01

**p < 0.01 (two-tailed). *p < 0.05 (two-tailed). While a particular block of variables included variables from the preceding block(s), the information for only the new variables in that block has been shown in this table for brevity

Source(s): Table by authors