Welcome Back? Job Performance and Turnover of Boomerang Employees Compared to Internal and External Hires

Abstract

“Boomerang” employees are workers who leave an organization and are later rehired by that same organization. Although many organizations rehire former employees, only a handful of studies have examined this phenomenon. The present study uses a large, longitudinal data set to examine the performance and turnover of boomerang employees rehired into management positions (n = 1,318). Further, we provide some of the first comparisons between boomerang employees and two traditional sources of employees: external hires (n = 20,850) and internal promotions (n = 8,546). Evaluations of job performance before and after being rehired revealed that boomerang managers’ performance tended to remain the same—rather than increase or decrease—after being rehired. Furthermore, boomerang managers performed similarly to internally and externally hired managers in the first year on the job, but both internal and external hires improved more than rehires over time. Internal and external hires were also less likely to turn over from the organization than rehires. Finally, supplemental analyses indicated that boomerang managers who turned over a second time tended to do so for reasons similar to their initial turnover reasons. The overall results call into question some of the assumed benefits of rehiring and suggest that organizations consider factors such as the reason for initial departure, the time horizon of performance, and the availability of other types of hires before considering boomerang employees.

Keywords

staffing boomerang employees rehiring internal labor markets external hiring

A key staffing decision organizations often make is whether to promote internal job candidates or hire external candidates. On the one hand, internal hires may possess unique knowledge, skills, and relationships within the organization that may enable them to perform effectively in a higher level position. Internal hires are also familiar with the organization’s culture, norms, and systems and, thus, may “get up to speed” more quickly than hires who are external to the organization (DeOrtentiis, Van Iddekinge, Ployhart, & Heetderks, 2018). On the other hand, external hires may provide new perspectives or attributes the organization lacks. In addition, external hiring does not create another position to be filled as is the case when organizations promote from within. Research has found evidence of internal–external hire differences with respect to variables such as starting pay, job performance, promotions, and turnover. For instance, although internal hires tend to perform as well as or better than external hires, employees hired from the outside tend to receive higher starting salaries (Bidwell, 2011; Chan, 1996; DeOrtentiis et al., 2018).

A staffing practice that has received less attention is rehiring former employees. “Boomerang” employees are workers who leave an organization and are later rehired by the organization (Shipp, Furst, Harris, & Rosen, 2014). Historically, organizations rarely considered rehiring former employees. Choosing to leave an organization was viewed as disloyal and, once an employee left the company, they were gone for good (Kelly, 2017). However, in an age when the average employee will work for more than 12 different employers during their career (U.S. Bureau of Labor Statistics, 2019), organizations are becoming more open to rehiring former employees (Almeda & Schawbel, 2015). This has been referred to as the “new normal” (Browne, 2016). Tight labor markets and skill shortages also have contributed to a shift in attitudes about rehiring (Rowlings, 2017). For example, organizations have “alumni” programs to keep former employees connected to the organization as well as to provide them product-oriented “perks” (e.g., Microsoft Corporation, n.d.; Texas Instruments Corporation, n.d.). In addition, some organizations do not require boomerang employees to go through a formal rehiring process and may restore employees’ prior seniority or benefit levels (Hirschman, 2000; Swartz, n.d.).

The popular press often extols the potential benefits of rehiring former employees. First, because boomerang employees are a known entity, hiring them is thought to be less risky than hiring first-time employees (Florentine, 2015; Hirschman, 2000). Second, boomerang employees are familiar with the job, understand the organization’s culture and values, and may have relationships with existing employees. Thus, rehiring is thought to reduce the time and costs associated with onboarding, training, and socialization into the organization (Almeda & Schawbel, 2015). Third, employees who return after working for another organization, or upon completing a degree, may bring fresh perspectives and new knowledge and skills to the organization (Rowlings, 2017). Fourth, when boomerang employees choose to return to their former employer (perhaps after concluding “the grass is not greener on the other side”), this signals to existing employees that the organization is a relatively good place to work and, in turn, may increase employee commitment (Hirschman, 2000). Finally, rehiring past employees is consistent with the procedural justice principal of reconsideration (Gilliland, 1993) and with the idea of giving second chances.

Despite the increased prevalence of rehiring and its assumed benefits, surprisingly few studies have examined this staffing strategy. In fact, a literature review revealed only two published studies that focused on boomerang employees. First, Shipp et al. (2014) compared boomerang employees who returned to a consulting firm with employees who left but did not return (i.e., “alumni” employees). They found that boomerang employees tended to have shorter initial tenures than those who did not subsequently return (3.8 vs. 7.1 years). In addition, rehires were more likely to have left due to alternate job opportunities, whereas alumni employees were more likely to have left due to dissatisfaction with the job or organization. Second, Swider, Liu, Harris, and Gardner (2017) studied National Basketball Association (NBA) players who returned to play for a former team. Results revealed that prior performance was positively related to rehire performance. Also, post-rehire performance was higher when rehires returned to play for the same coach and when the player—rather than the organization—initiated the initial turnover.

Thus, boomerang employment appears to be an area for which practice has greatly outpaced research. The present study was conducted to help address this practice–research gap and in doing so contributes to the limited research in this area in several ways. First, we replicate, clarify, and extend findings from the very limited existing research on boomerang employees. For example, the only other study we are aware of that examined whether rehires’ job performance increases, decreases, or remains the same used professional athletes (Swider et al., 2017). As we discuss later, the professional sports context is unique from other work settings, including the key role of physical abilities and age-related declines in such abilities. Thus, we investigate whether these findings hold in a more typical employment and work setting with boomerang managers. In doing so, we suggest that behavioral consistency theory may be a helpful lens for understanding boomerang employees.

Second, we begin to insert boomerang employees into the staffing literature that has, to date, focused primarily on two sources of employees available to organizations: internal applicants (e.g., employees promoted from within) and external applicants (e.g., recent college graduates) (Bidwell, 2011; DeOrtentiis et al., 2018; Rosenbaum, 1979; White, 1970). In doing so, we address a key question that appears to have received no attention in organizational scholarship: How do rehires perform compared to internal hires and first-time external hires, both initially and over time? We suggest that human capital theory provides a relevant basis to understand these group differences. Further, our comparisons between rehires and internal and external hires seek to inform future staffing scholarship, including whether researchers should include hire type when examining the effectiveness of different recruitment and selection procedures. These comparisons also provide useful information to organizations whose applicant pools often include former employees, current employees, and potential new employees.

Hypothesis and Research Question Development

Figure 1 presents a framework for comparing the “lifecycle” of a boomerang employee to that of internal and external hires, and we use this framework to illustrate the relations we examine. Moving from left to right in Figure 1, the boomerang lifecycle includes (A) an initial period with an organization, (B) a separation event, (C) time away from the organization (e.g., returning to school, working for another organization), (D) a rehire event, (E) post-rehire performance, (F) a (potential) post-rehire promotion, and (G) a (potential) second separation. Prior research has examined some of the events within this lifecycle. Shipp et al. (2014) focused primarily on boomerang employees’ initial period with the organization (A) and reasons for their initial separation (B). Swider et al. (2017) examined rehires’ initial period with the organization (A), as well as their separation (B), time away from the organization (C), and post-rehire performance (E). In addition to these events, we examine boomerang employees’ post-rehire behavior more closely, including their promotion rates (F) and the likelihood of a second separation from the organization (G). We also use the framework to make comparisons between rehires and the lifecycles of internal and external hires.

Figure 1

Comparisons of the Boomerang Employee Lifecycle With Internal and External Hire Lifecycles

Changes Within Boomerang Employees

To begin, we compare boomerang employees’ initial job performance (A) to their post-rehire performance (E) to address whether the performance of boomerangs improves, declines, or remains the same after being rehired. We are aware of only one study that has examined the relationship between boomerang employees’ pre- and post-rehire job performance. Swider et al. (2017) found a correlation of .37 between NBA players’ performance before and after returning to their former team. Further, they reported (but did not test) that rehire performance was lower than initial performance (we calculated the standardized mean difference [d] to be –.71). They also found that rehire performance was lower when the team initiated the initial departure (i.e., involuntary turnover) and higher when it was player-initiated (i.e., voluntary turnover).

Swider et al.’s (2017) study provides some important initial insights about boomerang employees. However, key differences exist between rehiring professional athletes and rehiring managers in organizations. For instance, the selection of athletes tends to focus on maximizing short-term gains (e.g., winning a championship within a given season), whereas organizations tend to focus on longer term goals when selecting and developing managers. Furthermore, the performance domain of a professional athlete—and the Knowledge, Skills, Abilities, and Other Characteristics (KSAOs) that determine performance—is much narrower than the diverse behaviors and determinants that comprise the domain of managerial performance (Mumford, Campion, & Morgeson, 2007). In turn, performance changes among boomerang athletes are more straightforward to assess given the more objective measures by which athletes’ performance is evaluated (e.g., points scored). Finally, declines in physical ability associated with age can lead to lower performance over time (e.g., Maertens, Putter, Chen, Diehl, & Huang, 2012). As such, the finding of lower rehire performance among boomerang NBA players may have been due, in part, to increasing age and decreasing physical abilities (e.g., reaction time, recovery from injuries) rather than to a boomerang effect. In contrast, physical abilities do not tend to be a major determinant of managerial performance. Thus, the first goal of the present research was to examine whether the findings of Swider et al. generalize to boomerang managers in a retail setting. We present three competing hypotheses to evaluate whether boomerang employees demonstrate better, similar, or lower performance after being rehired.

Several theoretical perspectives suggest that boomerang employees will demonstrate better performance after being rehired. First, rehires may benefit from experiences acquired since leaving the organization. Specifically, experience is thought to provide workers an opportunity to develop knowledge and skills, which, in turn, are proximal antecedents of job performance (Schmidt, Hunter, & Outerbridge, 1986). For example, boomerang employees who left to work for a different organization or to attend school may develop additional knowledge and skills that allow them to perform better upon being rehired. Second, theory and research on information processing suggests that workers with prior job experience tend to possess more comprehensive and well-defined cognitive schemas than workers without this experience (Kirschenbaum, 1992). Thus, rehires possess schemas that enable them to better understand and integrate job-related information, as well as minimize information processing demands (Louis, 1980). Finally, prior experience in the organization may serve as a realistic job preview (RJP) and help calibrate boomerang employees’ expectations about the job and organization. In turn, calibrated expectations can help reduce role ambiguity (Wanous, 1978) and activate coping mechanisms that help rehires adjust to less desirable aspects of the job (Locke, 1976; Meglino, Denisi, & Ravlin, 1993). Based on these ideas, we propose the first competing hypothesis:

Hypothesis 1a: Boomerang managers will demonstrate better performance after being rehired.

In contrast, behavioral consistency theory (Owens, 1976; Wernimont & Campbell, 1968) would predict that boomerang managers tend to demonstrate similar performance before and after being rehired. This theory suggests that past behavior is a good predictor of future behavior and has been used as a rationale for selection procedures that collect and assess samples of applicants’ behavior to predict future performance on the job (e.g., Schmitt & Ostroff, 1986). For example, research has found that performance on behavioral-based procedures such as assessment center exercises (Arthur, Day, McNelly, & Edens, 2003), work sample tests (Roth, Bobko, & McFarland, 2005), and behavioral interviews (Taylor & Small, 2002) tends to be a good predictor of job performance. In the same way, we expect the pre-rehire performance of boomerang employees to be a good predictor of performance upon returning to the organization.

According to behavioral consistency theory, past behavior is thought to predict future behavior when the past behavior is characterized by high physical, content, situational, psychological, and behavioral fidelity with the future job (Thornton & Kedharnath, 2013). In comparison to behavioral-based selection procedures, rehiring provides an even higher fidelity setting in which behavior is evaluated. Particularly when boomerang employees return to the same or a similar job, as is the case in this study, the content and context of their work may be similar to their initial employment, thus heightening the possibility of behavioral consistency (Mischel & Shoda, 1995). For instance, the products or services an organization provides are likely to be similar upon returning, as are the policies, norms, and culture. The psychological similarity between initial and rehire employment also may be stronger than that between behavior assessed during the selection process and future behavior on the job. In particular, selection settings often elicit maximal performance rather than typical performance. In contrast, organizations observe and can evaluate boomerang employees over a longer period, which is likely to represent their typical performance (Sackett, Zedeck, & Fogli, 1988). Additionally, both the KSAOs required by the job and boomerang employees’ traits that influence performance (e.g., cognitive ability, personality) likely will be similar during their initial and rehire employment periods (Deary, Whalley, Lemmon, Crawford, & Starr, 2000; Roberts, Walton, & Viechtbauer, 2006). A final reason to expect behavioral consistency is that the high fidelity between initial and return employment may result in rehires resuming work habits they developed prior to leaving (Ouellette & Wood, 1998). Taken together, this leads to our second competing hypothesis:

Hypothesis 1b: Boomerang managers will demonstrate similar performance before and after being rehired.

The last possibility is that boomerang employees will demonstrate lower performance upon being rehired. For example, changes in organizational culture (e.g., due to new leadership) or changes to how work is performed (e.g., due to technology) could diminish the value of rehires’ prior knowledge of their job or organization (Swider et al., 2017). In addition, as noted previously, workers develop cognitive schemas that help them make sense of workplace information and events. Negative transfer of learning can occur when schemas workers developed in one context are not relevant to their new context (Woltz, Gardner, & Bell, 2000). For example, the schemas rehires developed in the interim organization(s) for which they worked may be incongruent with their former organization and could interfere with performance upon their return (Dokko, Wilk, & Rothbard, 2009). This leads to the competing hypothesis that:

Hypothesis 1c: Boomerang managers will demonstrate lower performance after being rehired.

Next, we build upon Swider et al.’s (2017) findings by examining whether the reason for boomerang employees’ initial departure (B) affects their post-rehire performance (E). A key distinction in the turnover literature is between voluntary versus involuntary turnover (Bluedorn, 1978). Voluntary turnover is when employees initiate their departure, such as by leaving to take a position in another organization or to attend school. Involuntary turnover is initiated by the organization, such as terminating employees whose performance is substandard or due to the closing of a location. Although rehiring employees who turned over voluntarily may be more common, research suggests that rehiring terminated employees can be beneficial in some circumstances (e.g., Rodgers, Helburn, & Hunter, 1986), and popular press articles indicate that some organizations do, indeed, rehire such employees (e.g., Zimmerman, 2006).

Behavioral consistency theory would predict that boomerang employees who initially left the organization voluntarily will demonstrate better post-rehire performance than employees who initially left involuntarily. Specifically, rehires whose initial turnover was voluntary left of their own accord, such as for a higher paying job or to pursue a degree, which is not a signal of a potential performance problem. In contrast, rehires whose initial turnover was involuntary more likely were terminated for performance-related issues, such as substandard performance or counterproductive work behaviors, which may reoccur when they rejoin the organization.

Hypothesis 2: Boomerang managers whose initial turnover was voluntary will demonstrate better post-rehire job performance than boomerang managers whose initial turnover was involuntary.

Differences Between Boomerang Employees and Other Types of Hires

So far, we have focused on possible changes within boomerang employees. We now consider possible differences between rehires and external and internal hires. External hires are selected from outside the organization. This includes individuals who previously were employed by another organization or who are entering the labor market for the first time (e.g., recent graduates). Rehires are similar to external hires in that neither group was employed by the organization at the time of hire. Internal hires are workers who already are members of the organization and (a) were selected into a higher level position from a lower level position (i.e., an internal promotion) or (b) were selected into a similar position from a different part of the organization (i.e., a lateral transfer). In the present study, we focus on current employees who were promoted into an entry-level management position. Rehires are similar to internal hires in that both groups possess experience with the current organization.

Job performance differences

We compare boomerang employees’ post-rehire job performance (E) to other types of hires to address whether the performance of boomerangs exceeds, matches, or lags behind other types of hires. On the basis of human capital theory (Becker, 1964), we expect that boomerang managers may initially outperform externally hired managers due to the organization-specific human capital rehires retained from their initial time in the organization. Whereas general human capital represents knowledge, skills, abilities, and other characteristics that impact performance regardless of the organization (e.g., intelligence), organization-specific human capital reflects knowledge and skills tied to a particular firm (DeOrtentiis et al., 2018). For example, boomerang employees may better understand the routines and processes required to perform their jobs, as well as possess established relationships with other organizational members that external hires would not yet possess. Furthermore, prior experience in the organization may facilitate the socialization of rehires and help them transition into their former role more quickly (Beus, Jarrett, Taylor, & Wiese, 2014). In contrast, external hires lack prior experience in the organization and thus may face greater uncertainty when they first enter the organization and take longer to “figure things out.”

However, theory and research on skill acquisition (e.g., Ackerman, 1987; Ackerman, Kanfer, & Goff, 1995) suggest the existence of a learning curve, such that performance often improves substantially when workers are first exposed to a new task or job but then levels off after they have acquired the requisite knowledge and skills. Applied to the present context, the advantage rehires possess (e.g., with regard to their organization-specific human capital) should diminish as external hires gain exposure to the job and organization. If so, rehires’ performance may level off more quickly than external hires, who may have more room to improve. Thus, we expect performance of external hires to improve more over time than rehires’ performance.

Hypothesis 3: (a) Boomerang managers will initially outperform externally hired managers, but (b) external hires will improve more over time.

Initial performance differences between boomerang managers and internally promoted managers may be less straightforward. For example, because internal hires currently work for the organization, their organization-specific human capital is more current, which may enable them to outperform rehires. On the other hand, rehires already possess knowledge of the management role in the organization, whereas internal hires from line positions lack prior experience in that role. Thus, due to their previous time in the management position in the organization, rehires may initially outperform internal hires, who are new to the position. Further, it is unclear whether rehires or internal hires might have a greater opportunity to improve over time. Given these different possibilities, we pose the following research question:

Research Question 1: (a) Do initial performance differences exist between boomerang managers and internally promoted managers, and if so, (b) does one group improve more over time than the other group?

Promotion rate differences

In addition to performance, we investigate whether rehires are promoted more or less quickly than the other hire types (F). There are at least two reasons other than performance why rehires may be more likely to be promoted than internal hires. First, rehires have shown a willingness to leave the organization, and thus, companies may reward rehires with promotions (in title or salary) as a way to retain those employees. In contrast, by taking a position in their current organization and not leaving for another organization, internal hires already have displayed an aversion to leaving (Bidwell, 2011). In addition, because internal hires were recently promoted, they may not expect to receive another promotion right away.

Second, rehires initially may have left the organization for another job or to pursue additional education. If so, such experiences may increase rehires’ general human capital and, in turn, better position them to leave the organization once again (e.g., for a higher paying position). This results in an incentive for organizations to reward rehires with promotions—beyond what their performance merits—to preserve their general human capital and maintain competitiveness with other organizations (Bidwell, 2011). Further, organizational decision-makers may value work experiences outside the organization to develop managerial skills (DeRue & Wellman, 2009; Dragoni, Oh, Vankatwyk, & Tesluk, 2011), making boomerang managers more promotable (De Pater, Van Vianen, Bechtoldt, & Klehe, 2009). In contrast, internal hires’ time in the organization should enable them to develop organization-specific human capital (Bidwell, 2011). Because this human capital holds less value outside a particular organization, internal hires are thought to possess minimal leverage for gaining rewards (Coff & Kryscynski, 2011). Taken together, organizations may reward rehires with promotions to retain those employees, whereas organizations may not have to use promotions to retain internal hires.

Hypothesis 4: Boomerang managers will be more likely to be promoted than internally promoted managers.

Whether organizations are more or less likely to promote rehires over external hires is less clear. Several of the reasons noted above for why rehires may be promoted more quickly than internal hires also appear relevant to external hires. For example, external hires’ willingness to leave their previous organization, as well as the general human capital they accumulated from their prior experience, may lead organizations to promote external hires at similar rates as they promote rehires. Alternatively, because rehires chose to return to the organization, they may present a lower risk for leaving again compared to external hires. Thus, organizations may promote external hires at a higher rate. For instance, after working for another organization(s), rehires may determine that their former organization was not so bad after all. Given these alternative possibilities, we explore potential promotion rate differences between rehires and external hires as a research question.

Research Question 2: Are boomerang managers the same, more, or less likely to be promoted than externally hired managers?

Turnover differences

Finally, we examine the probability of (subsequent) turnover among boomerang employees in comparison to internal and external hires (G). Rehires’ prior experience in the organization may serve as an RJP and help calibrate their expectations about the job and organization (Phillips, 1998; Wanous, 1978). These expectations, in turn, may help reduce role ambiguity and activate coping mechanisms that help rehires adjust to less desirable aspects of the job (Locke, 1976; Meglino et al., 1993). If so, rehires may be less likely to leave again compared to external hires, who are less likely to know what to expect. However, boomerang employees already have demonstrated a willingness to leave the organization, and they may exhibit this same propensity after being rehired. This would be consistent with the “hobo syndrome” (Ghiselli, 1974), which suggests that some workers tend to move from job to job regardless of conditions or available alternatives (Judge & Watanabe, 1995). Similarly, research suggests that some workers are predisposed to being satisfied or dissatisfied across jobs or organizations (Arvey, Bouchard, Segal, & Abraham, 1989; Fisher & Hanna, 1931). Taken together, it is unclear whether rehires and external hires may differ in terms of turnover, which led us to pose the following research question:

Research Question 3: Are boomerang managers more or less likely to turn over than externally hired managers?

In contrast to rehires, internal hires have chosen to remain with the organization and may be highly embedded both on-the-job (e.g., long-lasting work relationships) and off-the-job (e.g., community involvement; Mitchell, Holtom, Lee, Sablynski, & Erez, 2001). Previous research has shown that embedded employees could incur notable mobility costs (e.g., financial, social) by moving to a different organization (Lee, Burch, & Mitchell, 2014) and are less likely to leave than those who are less embedded (Jiang, Liu, McKay, Lee, & Mitchell, 2012). Furthermore, internal hires likely were promoted, at least in part, due to high levels of performance in their previous position. Receiving contingent rewards, such as a promotion, may make internal hires particularly unlikely to turnover (Harrison, Virick, & William, 1996). Thus, internal hires may be less likely to turn over than rehires, who, by previously leaving the organization, have shown they are willing to risk potential mobility costs. Moreover, because internal hires’ experience in the organization is more recent than rehires, internal hires may possess more realistic expectations about the job and organization, making them less likely to turn over than rehires.

Hypothesis 5: Boomerang managers will be more likely to turn over than internally promoted managers.

Method

Organizational Setting and Sample

The setting for this study was a large retail organization with locations throughout the United States. Each location had a lead manager, one or more assistant managers, one or more manager trainees, and numerous hourly employees. The organization hires a large number of managers each year and considers current employees, external applicants, and former employees when selecting managers. Further, there is enough turnover among managers (30%) to study rehires and compare them to other types of hires. For these reasons, this organization provided an ideal setting for the current study.

We analyzed data on 30,714 employees who initially were hired/rehired into the management trainee position. Of these cases, 1,318 (4%) were former employees who left and later were rehired as manager trainees, 20,850 (68%) were external hires selected from outside the organization, and 8,546 (28%) were internal hires promoted from a lower-level position within the organization.

Measures and Analytic Approach

Job performance

We used supervisor ratings to measure job performance. The lead manager of each store evaluated assistant managers and manager trainees on 5 core competencies (i.e., time management, communication, leadership, drive for results, and flexibility) and 10 job responsibilities (e.g., sales growth, inventory management, and customer service) based on a comprehensive job analysis of the tasks and KSAOs of store management jobs. Supervisors also considered accomplishment of specific goals in their ratings. Supervisors rated each competency, responsibility, and goal accomplishment on a 5-point scale, where 5 represents the highest level of performance. For the present study, we had access to the overall average of these ratings, rounded to the nearest whole number. Because the organization uses the performance evaluations for administrative purposes (e.g., to determine promotion, pay, and termination), they took the process very seriously. For example, they trained supervisors to conduct performance review meetings, provided a posttraining reference booklet, and required them to conduct rating calibration meetings (where managers meet to compare and defend their ratings to other managers) to maintain organizational standards across locations.

A company technical report assessing this performance measure revealed a single-rater reliability of .46 for the aggregate performance ratings, which is comparable to meta-analytic estimates of the interrater reliability of supervisor ratings (e.g., .52, Viswesvaran, Ones, & Schmidt, 1996). However, our analyses incorporated multiple performance ratings for each manager, resulting in higher reliability estimates. For example, our analyses used an average of 2.1 ratings per manager prior to rehire and 2.6 ratings for each manager after rehire, and the associated reliability estimates were .64 and .69, respectively. Another strength of these ratings is that they were not highly (negatively) skewed as can be the case with performance evaluations (e.g., the mean is 3.2 on a 5-point scale). On the other hand, variance was somewhat modest (SD = .53), which suggests estimates of relations involving manager job performance may be conservative.

We used repeated measures analysis of covariance (ANCOVA) to test our competing hypotheses related to initial and rehire performance differences as well as our hypothesis relating initial turnover reason to rehire performance. This procedure allowed us to evaluate changes in performance level within the context of the boomerang lifecycle by controlling for initial tenure, reason for initial turnover, and time away from the organization. We included a variable for time that indicated whether performance differed across the two measurement occasions (to test H1). We also included interactions between time and our three control variables to partial out performance change associated with those components of our model. Both initial tenure and time away were mean-centered (Delaney & Maxwell, 1981). Finally, we conducted a planned comparison of performance differences across reasons for turnover (to test H2) at the second measurement occasion (i.e., rehire performance of those who initially turned over voluntarily compared to involuntarily).

We used latent growth modeling (LGM) in AMOS Version 25.0 (Arbuckle, 2014) to test our hypotheses and questions comparing the job performance of rehires to that of internal and external hires. LGM enabled us to simultaneously test between-group differences in initial performance and in performance over time. We used 4 years of annual job performance evaluations. These four (first-order) measurements served as indicators of three (second-order) latent variables.¹ The latent intercept represents managers’ initial performance and provides information concerning the mean and variance of the intercepts for the growth curves of all the managers. The latent slope represents the rate of change in performance (e.g., at Year 1) and provides information concerning the mean and variance of the slopes across managers. The latent curve represents the acceleration in the rate of change in performance and provides information concerning the mean and the variance of the curvature for all the managers’ growth curves. Finally, the model included a variable that reflected rehires (n = 732) versus internal hires (n = 6,015) and a variable that reflected rehires versus external hires (n = 13,600) as predictors of the latent intercepts, slopes, and curves.

Prior to testing the full structural model, we examined the relationship between time and performance in an unconditional model (i.e., a model without the hire type predictors). We first specified a linear growth function by fixing the latent slope factor loadings of the four performance ratings to 0, 1, 2, and 3 and removing the latent curve variable. The fit indices for the linear model were as follows: χ² = 542.57, df = 5, p < .001; comparative fit index (CFI) = .86; and root-mean-square error of approximation (RMSEA) = .07. Overall, the linear model demonstrated less than ideal fit to the data based on conventions of good fit such as χ²/df less than 3 (Iacobucci, 2010), CFI greater than .95, and RMSEA of .06 (Hu & Bentler, 1999).

We also assessed a quadratic (i.e., nonlinear) model to test the possibility that the trajectory of performance reflects an inverted U shape, such that managers’ performance improves from Year 1 to Year 2 but then levels off in Years 3 and 4 (Chan, 1998). We specified a quadratic growth function by fixing the latent curve factor loadings of the four performance ratings to 0, 1, 4, and 9. The quadratic model fit the data significantly better than the linear model, Δχ²(4) = 536.26, p < .01. The fit indices for the quadratic model were χ² = 6.31, df = 1, p = .01; CFI = 1.00; and RMSEA = .02. Therefore, we used a quadratic model to test hypotheses and research questions.²

Promotions

We coded whether managers received a promotion from manager trainee to assistant manager.³ We also determined time to promotion by recording the days between date of entry into the manager trainee position and the date of promotion to assistant manager. For external hires, time to promotion reflected the number of days from initial hire date to promotion date. For internal hires, time to promotion reflected the number of days from initial promotion into the manager trainee position to the date of promotion to assistant manager. And for rehires, time to promotion reflected the number of days from rehire date to promotion date.

We used survival analysis (via Cox Regression analysis in SPSS 22.0) to test hypotheses and questions involving promotions. Survival analysis incorporates information regarding (a) whether an event occurs and (b) the amount of time each case is observed to estimate the probability of an event’s occurrence. The present analyses compared the promotion probability of rehires versus internal and external hires and involved a variable that indicated whether each participant was promoted (1) or not (0), time to promotion (i.e., from entry into the management trainee position until promotion, turnover, or the last day of the data set), and hire type (i.e., rehire vs. internal hire or rehire vs. external hire) as a categorical covariate. In addition, because performance likely played a role in promotion decisions, we controlled for (mean) job performance in the manager trainee position.

Turnover

We coded whether managers remained with or left the organization prior to the end of the data set. Also, when an employee left the organization, their manager entered a reason (from a standard set of approximately 30 reasons) into the human resources information system. Consistent with prior research (e.g., Lyness & Judiesch, 2001), we used this information to distinguish between voluntary versus involuntary turnover. Example voluntary turnover reasons included quitting for personal reasons, for another job, and to attend school. Example involuntary turnover reasons included termination for violation of a company policy, excessive tardiness or absenteeism, and inattention to duties. In a few instances for which it was unclear whether the reason reflected voluntary or involuntary turnover (e.g., “no hours/no earnings”) or when involuntary turnover was clearly not due to the behavior of the employee (e.g., “store or department closed”), we excluded these cases from analyses involving reason for turnover.

As with promotions, we used survival analysis to test hypotheses and research questions involving turnover. This analysis included a variable that indicated whether each participant turned over (1) or not (0), tenure (i.e., from entry into the management trainee position until turnover or the last day of the data set), and hire type (i.e., rehire vs. internal hire or rehire vs. external hire). In addition, to be consistent with the promotions analyses, we controlled for job performance.

Results

Changes Within Boomerang Managers

Table 1 presents descriptive statistics and correlations for the study variables. The first set of hypotheses addressed boomerang managers’ job performance and retention before and after being rehired. Competing Hypotheses 1a, 1b, and 1c predicted that boomerang managers would demonstrate higher, similar, or lower levels of performance, respectively, before and after being rehired. Rehires’ job performance prior to their initial departure (M = 3.15, SD = .52) was not significantly different from their post-rehire performance (M = 3.19, SD = .62) (t = .87, p = .39). Further, the standardized mean difference between initial and rehire performance was very small (d = .06).⁴ Additionally, the repeated measures ANCOVA indicated no significant change in performance, F(1, 184) = 3.73, p = .06, controlling for initial tenure, initial turnover reason, and time away from the organization (see Table 2). A planned pairwise comparison revealed marginal means of 3.01 for initial performance and 3.17 for rehire performance, and the mean difference of –.16 (SE = .08) was not significant (p = .06). To illustrate, 62% of managers received the same performance rating on their final performance evaluation of their initial tenure and their first evaluation after rehire. Overall, results support H1b rather than H1a or H1c and suggest that—consistent with behavioral consistency theory—the performance of boomerang managers tended to remain the same (rather than increase or decrease) upon being rehired.

Table 1

Descriptive Statistics and Correlations for Study Variables

	N	M	SD	1	2	3	4	5	6
Boomerang manager variables
1. Initial performance	511	3.15	.52
2. Reason for initial turnover	975	.95	.22	.11*
3. Initial tenure	1,318	757.04	954.48	.18**	−.01
4. Time away	1,317	1.77	1.86	.02	.06	−.06*
5. Post-rehire performance	733	3.17	.57	.44**	.11**	.07	−.00
6. Reason for post-rehire turnover	471	.83	.37	.01	.10	.05	−.01	.05
Boomerang versus other hires variables
1. Rehires versus internal hires	9,864	.13	.34
2. Rehires versus external hires	22,168	.06	.24	n/a
3. Mean performance	20,391	3.20	.53	.03**	.01
4. Turnover	30,714	.30	.46	−.13**	−.02*	−.19**
5. Tenure	30,519	811.80	721.79	.16**	.11**	.24**	−.27**
6. Promotion	30,714	.25	.43	.04**	.06**	.46**	−.18**	.55**
7. Time to promotion	30,574	587.95	541.35	.16**	.08**	−.09**	−.20**	.73**	.02**

Note: Reason for initial turnover was coded 1 = voluntary turnover and 0 = involuntary turnover. Rehires = 0 and internal or external hires = 1. Time away was in years. Initial tenure, tenure, and time to promotion were in days.

p < .05. **p < .01.

Table 2

Repeated Measures Analysis of Covariance for Initial and Rehire Performance

Variable	SS	df	F	p
Within-person effects
Time	1.45	1	3.73	.06
Initial Tenure × Time	.33	1	.85	.36
Time Away × Time	.29	1	.76	.39
Reason for Initial Turnover × Time	.20	1	.52	.47
Error (time)	71.28	184
Between-person effects
Initial tenure	3.09	1	16.82	.00
Time away	.16	1	.87	.35
Reason for initial turnover	1.63	1	8.88	.00
Error	33.82	184

Note: N = 188. SS = Type III sum of squares; df = degrees of freedom; F = test statistic. For within-person effects, a significant effect of time would indicate a change in within-person performance between the repeated measures. For between-person effects, the dependent variable was the average of the mean performance ratings before and after rehire. Initial tenure was measured as years from the hire date to the initial turnover date and centered. Time away was measured as years from initial turnover to the rehire date and centered. Reason for initial turnover was coded 0 = involuntary and 1 = voluntary.

Hypothesis 2 predicted that boomerang managers whose initial turnover was voluntary would demonstrate better post-rehire performance than boomerang managers whose initial turnover was involuntary. Consistent with prior research (Swider et al., 2017), rehires who initially left voluntarily did, indeed, have higher average post-rehire job performance ratings than rehires who initially left involuntarily (n = 528; M_difference = –.26, SD_difference = .10, t = 2.50, p = .02; d = .48). However, using ANCOVA to control for initial tenure and time away, a planned pairwise comparison of the marginal means of rehire performance for employees who initially turned over voluntarily versus those who turned over involuntarily indicated reason for initial turnover was not a significant predictor of rehire performance (M_difference = .28, SE_difference = .15, t = 1.87, p = .07).⁵ Thus, results do not support H2.

Boomerang Managers Versus Other Managers

Job performance

The next set of hypotheses addressed potential differences between boomerang managers and internally and externally hired managers. Hypothesis 3 predicted that boomerang managers would (a) initially outperform externally hired managers but that (b) external hires would improve more over time. Table 3 presents descriptive statistics and correlations for the job performance ratings used as input for the LGM analyses, and Table 4 displays the model fit statistics and parameter estimates for the model centered at Years 1 to 4. Focusing on the model at Year 1, the external hire dummy variable was not a significant predictor of the intercept (y = –.04, SE = .02, z = −1.64, p = .10), which suggests that rehires tended to perform similar to external hires in the first year on the job. Thus, Hypothesis 3a was not supported. In support of Hypothesis 3b, hire type was significantly related to changes in performance (i.e., the slopes), such that external hires improved more over time than rehires (y = .14, SE = .04, z = 3.21, p < .01). Because the model includes a latent quadratic term, both intercepts and slopes differ each time performance is measured (Biesanz, Deeb-Sossa, Papadakis, Bollen, & Curran, 2004). Centering the model at Years 2, 3, and 4 revealed that external hire performance was significantly better than rehire performance at each of these times. Interestingly, external hires continued to improve more than rehires at Year 2, but this effect leveled off by Years 3 and 4, as depicted in Figure 2. Finally, the relationship between external hire type and the latent curve was not significant (y = –.03, SE = .015, z = −1.955, p = .051), indicating that the rehires’ curve was similar to that of external hires. In practical terms, rehires initially outperformed external hires by 1% on average, but external hires improved enough by Year 3 to outperform rehires by an average of 4%.

Table 3

Descriptive Statistics and Correlations for the Job Performance Ratings (Years 1 Through 4) by Manager Type

Manager Type/Performance Rating	N	M	SD	1	2	3
Boomerang managers
1. Year 1	689	3.17	.64
2. Year 2	377	3.24	.72	.40
3. Year 3	152	3.34	.70	.20	.32
4. Year 4	82	3.24	.71	.32	.38	.41
Internally hired managers
1. Year 1	5,689	3.15	.59
2. Year 2	4,017	3.34	.67	.32
3. Year 3	2,422	3.38	.68	.20	.33
4. Year 4	1,459	3.37	.69	.15	.21	.36
Externally hired managers
1. Year 1	13,022	3.13	.63
2. Year 2	9,871	3.32	.70	.36
3. Year 3	6,285	3.39	.71	.23	.36
4. Year 4	4,356	3.36	.71	.14	.19	.34

Note: Total Ns for boomerang, internally hired, and externally hired managers were 732, 6,015, and 13,600, respectively. All correlations were statistically significant (p < .05).

Table 4

Latent Growth Model Comparing Hire Types’ Performance Rating Means, Slopes, and Curves Over 4 Years

Variable/Time	$χ^{2}$	df	CFI	RMSEA	Intercept (SE)	Slope (SE)	Curve (SE)
	9.69*	3	1.00	.01
Boomerang versus internal hires
Centered at Year 1					−.02 (.03)	.14 (.04)**	−.03 (.02)
Centered at Year 2					.09 (.03)**	.08 (.02)**	−.03 (.02)
Centered at Year 3					.14 (.03)**	.02 (.03)	−.03 (.02)
Centered at Year 4					.13 (.05)*	−.04 (.06)	−.03 (.02)
Boomerang versus external hires
Centered at Year 1					−.039 (.024)	.14 (.04)**	−.03 (.02)
Centered at Year 2					.07 (.03)*	.08 (.02)**	−.03 (.02)
Centered at Year 3					.11 (.03)**	.02 (.03)	−.03 (.02)
Centered at Year 4					.10 (.05)*	−.04 (.06)	−.03 (.02)

Note: Ns for boomerang, internally hired, and externally hired managers were 732, 6,015, and 13,600, respectively. The Intercept, Slope, and Curve columns represent parameters (and standard errors) between a hire type dummy variable and the latent intercept, slope, and curve, respectively. Hire type predictors were dummy coded such that boomerang managers were the reference group. With the intercept centered at Year 1, slopes were coded 0, 1, 2, and 3 representing Years 1, 2, 3, and 4, respectively. With the intercept centered at Years 2, 3, and 4, the slopes were −1, 0, 1, and 2; then −2, −1, 0, and 1; and then −3, −2, −1, and 0, respectively.

p < .05. **p < .01.

Figure 2

LGMs With Type of Hire as a Predictor of Job Performance Over Time

Research Question 1a asked whether initial job performance differences exist between boomerang managers and internally promoted managers, and Research Question 1b asked whether rehires or internal hires improve more over time. As Table 4 shows, although rehires performed similar to internal hires at Year 1 (y = –.02, SE = .03, z = –.87, p = .39), internal hires improved more over time (y = .13, SE = .04, z = 3.11, p < .01). As with the comparisons between rehires and external hires, centering the model at Years 2, 3, and 4 revealed that internal hire performance was significantly better than rehire performance at each of these times. Also, internal hires continued to improve more than rehires at Year 2, but differences in performance change leveled off by Years 3 and 4. Finally, the relationship between hire type and the latent curve was not significant (y = –.03, SE = .02, z = −1.82, p = .07). In practical terms, rehires initially outperformed internal hires by less than 1% on average, but internal hires improved enough by Year 3 to outperform rehires by 4% on average. Taken as a whole, the above results suggest that despite rehires’ similar initial job performance to internal and external hires, these two hire types improved more than rehires.⁶

Promotions

Hypothesis 4 predicted that boomerang managers would be more likely to be promoted than internally promoted managers, and Research Question 2 asked whether boomerang managers would be more or less likely to be promoted than externally hired managers. Table 5 displays the results of the survival analysis, and Figure 3 displays the probability of promotion for each set of hires. The overall model, which controlled for differences in job performance, was statistically significant in predicting promotion decisions, $χ^{2}$ (3) = 5,285.75, p < .001. Further, in support of Hypothesis 4, rehires were 1.24 times (i.e., 1/.80) more likely to be promoted than internal hires (Wald = 7.59, p = .01). In response to Research Question 2, external hires were 1.16 times more likely to be promoted than rehires (Wald = 3.77, p = .052).

Table 5

Results of Survival Analyses Predicting Promotions and Turnover

Analysis/Variable	$χ^{2}$	df	N	B	SE	Wald	Exp(B)
Hire type as a predictor of promotion	5,285.75**	3	19,636
Job performance				1.58	.02	5,241.36**	4.85
Rehires versus internal hires				−.22	.08	7.59**	0.80
Rehires versus external hires				.15	.08	3.77	1.16
Hire type as a predictor of turnover	1,535.34**	3	20,245
Job performance				−1.01	.03	1,422.49**	0.36
Rehires versus internal hires				−.81	.07	139.02**	0.44
Rehires versus external hires				−.71	.07	119.00**	0.49

Note: $χ^{2}$ = chi-square statistic. B = unstandardized coefficient. SE = standard error of the unstandardized coefficient. Exp(B) = odds ratio. Rehires were the reference category in each of the categorical variables.

p < .01.

Figure 3

Cumulative Likelihood of Promotion by Hire Type

Turnover

Research Question 3 asked whether boomerang managers would be more or less likely to turn over than externally hired managers, and Hypothesis 5 predicted that boomerang employees would be more likely to turn over than internally promoted managers. Table 5 displays the survival analysis results, and Figure 4 displays the probability of retention for each set of hires. The overall model was statistically significant in predicting turnover, $χ^{2} (3)$ = 1,535.34, p < .001. Related to Research Question 3, rehires were 2.04 times (i.e., 1/.49) more likely to turn over than external hires (Wald = 119.00, p < .001). Further, in support of Hypothesis 5, rehires were 2.27 times (i.e., 1/.44) more likely to turn over than internal hires (Wald = 139.02, p < .001).

Figure 4

Cumulative Survival Rate by Hire Type

Supplemental analyses of reasons for initial turnover

We considered additional ways to categorize the turnover reasons beyond the voluntary–involuntary distinction. Previous literature has distinguished reasons for departure in terms of avoidable versus unavoidable turnover (Campion, 1991), functional versus dysfunctional turnover (Dalton, Todor, & Krackhardt, 1982), and what each reason signals regarding embeddedness (Mitchell et al., 2001). However, these turnover distinctions were not broadly applicable across the turnover reasons in our data. Additionally, these categories arose when viewing turnover as an outcome. Instead, our primary interest was turnover reasons as predictors of future (i.e., rehire) behavior.

Consequently, we evaluated the reasons for pre-rehire turnover to develop a categorization focused on predicting post-rehire outcomes. This approach revealed a categorization of turnover reasons that represented their valence to the organization for predicting performance and turnover after rehire. Each of the authors independently evaluated the valence of each turnover reason, coding them as positive, neutral, or negative for each outcome. For predicting rehire performance, we coded the valence of the reasons as positive if they might develop human capital (e.g., left to go back to school), as neutral if the reasons lacked information relevant to human capital or job performance (e.g., left for personal reasons), and as negative if the reasons indicated poor performance (e.g., inattention to duties). For predicting rehire turnover, we coded the valence as positive if the reasons represented the potential to develop positive attitudes towards the organization (although we did not identify any positively valenced reasons), as neutral if the reasons lacked information relevant to attitudes toward the organization (e.g., left for personal reasons), and as negative if the reasons indicated negative attitudes or counterproductive work behavior (e.g., insubordination). Initial coding resulted in 93% agreement regarding the valence of each reason for both outcomes, and the coders resolved the few discrepancies through discussion. When considering performance as the criterion, there were 15 negative reasons (n = 3,464), 11 neutral reasons (n = 1,801), and 4 positive reasons (n = 4,159). When considering turnover as the criterion, the reasons sorted into 17 negative reasons (n = 7,051) and 13 neutral reasons (n = 2,373). Finally, we also summed together the performance- and turnover-related valences to evaluate the overall valence to the organization. This resulted in 15 negative (n = 3,464), 1 somewhat negative (n = 99), 11 neutral (n = 5,190), and 3 somewhat positive (n = 671) turnover reasons. Additional information regarding the turnover reasons and codes is provided in the online supplement to this article.

We first explored the relationship between the performance-related valence of boomerang managers’ reason for initial turnover and post-rehire performance. The correlation with post-rehire performance was .10 (p = .02), which suggests a slight tendency for boomerangs who initially left for more positive reasons to perform better when rehired than those who initially left for negative reasons. We also evaluated this relationship within the framework of the boomerang employee lifecycle by entering the variables used to test H1 and H2 (excluding voluntary turnover) into a hierarchical regression. After accounting for initial performance, initial tenure, and time away from the organization, we added dummy variables to represent the neutrally and positively valenced turnover reasons (i.e., negatively valenced reasons were the referent group).

Table 6 presents the results, which revealed that although turnover reason valence did not significantly increase the variance explained by the model (ΔR² = .02, ΔF = 2.66, p = .07), rehires who initially turned over for neutral reasons outperformed those who initially left for negative reasons (β = .20, t = 2.03, p = .04). Similarly, rehires who initially turned over for positive reasons outperformed those who initially turned over for negative reasons (β = .22, t = 2.22, p = .03). In contrast, there were no differences in post-rehire performance between rehires who initially turned over for neutral as compared to positive reasons (β = .01, t = .07, p = .95).

Table 6

Hierarchical Regression Analysis of Mean Rehire Performance on Turnover Reason Valence

Variable	Step 1 β	Step 2 β
Initial tenure	.09	.10
Time away	−.07	−.09
Mean initial performance	.34**	.32**
Neutrally valenced turnover reasons		.20*
Positively valenced turnover reasons		.22*
R ²	.14	.17
Adjusted R²	.13	.14
ΔR ²	.14**	.02

Note: N = 188. Initial tenure was measured as years from the hire date to the initial turnover date. Neutrally valenced turnover reasons were coded 0 = negative or positive reasons (n = 129) and 1 = neutral reasons (n = 59). Positively valenced turnover reasons were coded 0 = negative or neutral reasons (n = 88) and 1 = positive reasons (n = 100). Time away was measured as years from initial turnover to the rehire date and was mean-centered. Rerunning this analysis with neutrally valenced reasons as the reference category showed that rehire performance did not differ between neutrally and positively valenced reasons (β = .01, p = .95).

p < .05. **p < .01.

Next, we conducted a survival analysis to explore the relationship of initial turnover reason valence with post-rehire turnover. We evaluated the bivariate relationship between post-rehire turnover and a categorical variable that represented the negatively and neutrally valenced turnover reasons identified when considering turnover as the criterion. This relationship was not significant (Wald = .97, p = .32). We also assessed this within the framework of the boomerang lifecycle by controlling for initial employment tenure, time away from the organization, and post-rehire job performance. As Table 7 shows, the overall model was statistically significant, The overall model was statistically significant, $χ^{2} (4)$ = 35.37, p < .001, but the valence of turnover reasons did not significantly predict post-rehire turnover (Wald = .74, p = .39).

Table 7

Results of Survival Analysis Using Turnover Reason Valence to Predict Turnover Among Rehires

Variable	$χ^{2}$	df	N	B	SE	Wald	Exp(B)
	35.37**	4	527
Initial tenure				.03	.04	.50	1.03
Time away				−.13	.05	6.54*	.88
Mean rehire job performance				−.78	.15	26.67**	.46
Neutrally valenced turnover reasons				.13	.15	.74	1.14

Note: $χ^{2}$ = chi-square statistic. B = unstandardized coefficient. SE = standard error of the unstandardized coefficient. Exp(B) = odds ratio. Neutrally valenced turnover reasons were coded 0 = negatively valenced reasons and 1 = neutrally valenced reasons utilizing likelihood of turnover as the criteria for coding.

p < .05. **p < .01.

Finally, we used the turnover reason valence framework to further examine the usefulness of behavioral consistency as a theory for predicting boomerang employee behavior. In particular, behavioral consistency theory would predict that boomerang employees who turn over a second time will tend to do so for similar reasons as those that led to their initial turnover. For example, rehires who initially turned over for negatively valenced reasons will be more likely to turn over a second time for negatively valenced reasons than for other reasons.

Table 8 presents a cross-tabulation of the overall (i.e., summed) valence of turnover reasons for rehires who turned over a second time. Results showed that rehires’ reasons for turning over a second time tended to be similar to their initial turnover reasons, $χ^{2} (9)$ = 52.58, p < .001. For example, among rehires who turned over a second time, 64% who initially turned over for negatively valenced reasons did so for negative reasons again. Similarly, 59% of employees who initially turned over for neutrally valenced reasons did so for neutral reasons again. Exhibiting a slightly different pattern, employees who initially turned over for positively valenced reasons tended to turn over the second time for neutrally valenced reasons (53%). This finding appears to result because it would be unlikely for people who initially left to pursue a degree or enter the military to leave for that same reason again.

Table 8

Rehires’ Initial Turnover Reason Valence Compared With Subsequent Turnover Reason Valence

	Subsequent Turnover Reason Valence
Initial turnover reason valence	1	2	3	4
1. Negatively valenced reasons	40	0	21	2
2. Somewhat negatively valenced reasons	1	0	2	0
3. Neutrally valenced reasons	61	2	101	7
4. Somewhat positively valenced reasons	20	0	46	21

Note: N = 324. Values represented cases within each cell. $χ^{2} (9) = 52.58$ , p < .001. $χ^{2}$ = chi-square statistic. Spearman correlation = .32, p < .001. Valence represented the overall valence to the organization based on summing valence associated with performance and with turnover as the criterion.

Discussion

Employers are increasingly considering former employees to expand applicant pools in response to tight labor markets and skill shortages. Despite this, we know very little about the outcomes of people who return to a previous organization. Although rehiring is thought to be advantageous to organizations (e.g., compared to external hires, rehires are better known and can be trained and onboarded more quickly), empirical evidence to support such beliefs is scant.

The present study contributes to the limited research on boomerang employment by examining whether some of the initial findings in this area generalize to a commercial work setting with boomerang managers. In contrast to Swider et al.’s (2017) finding that players’ performance tended to decrease upon returning to their former team, the present results suggest that boomerang managers’ performance tends to remain the same after being rehired. Additionally, initial turnover reason was not a good predictor of rehire performance when controlling for initial performance. Finally, boomerang employees who left the organization a second time tended to do so for similar reasons to those of their initial departure.

This study also extends previous research by investigating how the performance and promotion and turnover rates of boomerang employees compare to that of internally and externally hired employees. We found that although boomerang managers perform similarly to both internal and external hires in the first year on the job, both internal and external hires improve more over time than rehires. Boomerang managers also are more likely to turn over than other types of hires. Interestingly, despite demonstrating less performance improvement and a higher propensity to turn over, boomerangs are more likely to be promoted than internal hires, and they are as likely to be promoted as external hires. This suggests that organizations may use promotions to try to retain boomerangs, even though such employees may not be as effective as other employees over the long term. Taken as a whole, these findings call into question some of the proposed benefits of rehiring former employees.

Implications for Theory, Future Research, and Practice

The present results have several implications for theory. For one, they suggest that behavioral consistency theory (Wernimont & Campbell, 1968) provides a useful lens to understand and predict the behavior of boomerang employees. Despite their time away from the organization, boomerang employees tend to demonstrate similar behaviors when they return. Their performance is similar, they are more likely to turn over again, and if they leave for a second time, they tend to leave for similar reasons. This is in contrast to the possibility that boomerang employees’ performance may improve due to new competencies or decline due to degradation of unused competencies. Also, although we noted that rehiring is consistent with the procedural justice principle of reconsideration (Gilliland, 1993) and the idea of giving people second chances, our findings suggest that rehires whose initial performance is lower do not “turn over a new leaf” upon returning. Instead, in line with behavioral consistency theory, lower initial performers tend to have lower performance upon returning and those who initially leave for negative reasons tend to turn over for negative reasons once again.

Our findings also suggest that human capital theory (Becker, 1964) provides a relevant basis to understand promotion differences between boomerang employees and other types of hires. For example, despite similar levels of initial job performance, rehires and external hires are more likely to be promoted than internal hires. This finding is consistent with the idea that organizations may reward certain types of employees (e.g., rehires) with promotions for reasons other than performance, such as their higher levels of general human capital or in response to the external market for their skills. Further aligned with human capital theory, research on skill acquisition (e.g., Ackerman, 1987; Ackerman et al., 1995) found that performance often improves substantially when workers are first exposed to a new task or job but then levels off after they have acquired the requisite knowledge and skills. The present results suggest that this phenomenon also helps explain differences between rehires and internal and external hires. Specifically, although boomerang managers perform similar to internal and external hires initially, rehires’ performance levels off more quickly. Thus, it appears that different types of hires tend to exhibit divergent performance trajectories.

Finally, our results highlight the need for high-quality indicators of human capital. Although prior research has found that human capital relates to performance (e.g., Ployhart, Van Iddekinge, & MacKenzie, 2011), our evaluation of turnover reasons as proxies of knowledge and skill growth (i.e., positive turnover reasons) did not explain differences in performance as compared to neutral turnover reasons. This result aligns with recent findings that prior work experience in other organizations does not tend to relate to performance in a new organization (Van Iddekinge, Arnold, Frieder, & Roth, 2019). One potential remedy is to look for variables that offer greater insight into actual performance or knowledge as opposed to those that indicate opportunities to perform or learn, as work experience and human capital proxy variables often do.

The present findings also have implications for future research. For example, we proposed an organizing framework for understanding the different events within the lifecycle of a boomerang worker and comparing those to other hiring sources. We hope this framework will help stimulate and guide future research that examines the various stages of, and impacts on, boomerang employment. Furthermore, our findings indicate that boomerang employees often behave differently from other types of hires. This suggests that future studies should distinguish rehires from other types of hires when researching human resources issues. For instance, hire type could influence conclusions about the validity of selection procedures or the effectiveness of different types of rewards (e.g., promotions). More specifically, rehire applicants may complete assessments they previously completed, which may affect how they respond or how they perform relative to applicants who are taking the assessments for the first time (e.g., external applicants). Additionally, our findings respond to and reinforce calls for research to identify more effective ways to categorize turnover reasons than traditional categories such as voluntary and involuntary turnover (Campion, 1991). Future researchers may further test how the valence of turnover reasons offers insights into rehiring and performance management.

Finally, this study has some key implications for practice. As noted, although the popular press often extols the benefits of rehiring, our study suggests such benefits may be short-lived. For instance, other types of hires tend to improve more over time and are more likely to stay with the organization. Our results especially highlight advantages of promoting existing employees. For example, internal hires are less likely to turn over (20.9%) than both rehires (36.6%) and external hires (33.5%). Further, internal hires tend to require lower starting salaries than external hires, and organizations are thought to feel less pressure to promote internal hires due to their commitment to the firm (e.g., Bidwell, 2011; DeOrtentiis et al., 2018). Thus, there appear to be several reasons why organizations should consider promoting from within (although this practice also has limitations, such as having to fill the position from which internal hires were promoted).

Although the present findings raise some concerns about rehiring employees who have departed, they also identify situations in which rehiring may be more effective. For one, boomerang employees whose initial turnover was voluntary outperformed internal and external hires during the first year (see Note 4). These findings highlight the importance of detailed record-keeping concerning reasons for initial departure (Campion, 1991). Additionally, rehiring high performers may be especially beneficial for situations that require fast onboarding to the job or for which the average tenure is relatively short. As an example, hiring former employees may be a good approach for staffing short-term projects or for temporary work.

Potential Limitations and Future Directions

Our study possesses some limitations that future research needs to address. First, although we were able to examine various events within the boomerang employee lifecycle, we did not have data on some events. For example, we did not have detailed information about what rehires did during their time away from the organization. It would be interesting if future research could collect more specific information about what boomerang employees do while they are away and whether and how their activities affect changes in post-rehire behavior. Given that our results appear to support behavior consistency theory, future research could explore whether specific types of experiences produce deviations from previous behavior, further identifying experiences organizations should consider when deciding whether to rehire former employees.

Second, our research focused on what would appear to be the main stages through which boomerang employees would progress. However, we did not examine some of the more specific factors and decisions that may exist within this lifecycle. For example, we did not examine factors that may cause boomerang employees to consider returning to a former employer in the first place or factors that influence organizations’ decisions about whether to rehire former employees. We hope future research will investigate such questions and perhaps expand our framework to include prehire events or decisions.

Third, our study evaluated performance trajectories of different hire types over time using mean supervisor performance ratings. The modest reliability likely attenuated relations involving this measure. Future research should seek to replicate and extend these findings with more robust performance measures. Further, future research may examine whether specific facets of performance differ over time or according to hire types. For example, differences may exist between measures of task performance, citizenship behavior, and counterproductive work behavior. Additionally, future studies may also consider whether there are differences between subjective and objective measures of performance for boomerang employees. For example, perhaps supervisor ratings of boomerang employees are biased by past evaluations or beliefs about boomerang employment. If so, objective performance measures might differ from subjective ones.

Finally, the present study identified some potential boundary conditions of boomerang hiring, such as whether rehires’ initial departure was for voluntary or involuntary reasons and whether performance was measured soon after being rehired or later on. Future studies could explore additional potential boundary conditions. For example, rehires initially turned over from a particular location with a particular staff, cliental, and so forth. Upon being rehired, some employees may return to their original location, whereas others may be hired into a different location. Thus, future research could examine whether and how being rehired to the same versus a different location affects post-rehire behavior.

Conclusion

Organizations and hiring managers, faced with a “war for talent,” typically had to look to either internal or external labor markets to find new employees. However, both internal and external hiring possess potential limitations. Today, organizations are increasingly considering a third source of candidates: former employees. Boomerang employees are thought to capture positive features of both internal hires (e.g., they know the organization’s culture and routines) and external hires (e.g., they may bring new knowledge and perspectives to the organization). Yet very little is known about the effectiveness of rehiring former employees. The results of the present study call into question some of the assumed benefits of rehiring and suggest that organizations often may be better served by promoting existing employees or hiring external applicants into management positions. However, research on boomerang employment is in its infancy, and there is much more we need to know. We look forward to seeing additional work that helps researchers and organizations better understand this emerging staffing issue.

Supplemental Material

Online_supplement_final – Supplemental material for Welcome Back? Job Performance and Turnover of Boomerang Employees Compared to Internal and External Hires

Supplemental material, Online_supplement_final for Welcome Back? Job Performance and Turnover of Boomerang Employees Compared to Internal and External Hires by John D. Arnold, Chad H. Van Iddekinge, Michael C. Campion, Talya N. Bauer and Michael A. Campion in Journal of Management

Footnotes

Acknowledgements

The authors thank Filip Lievens, Abbie Shipp, and Brian Swider for their helpful feedback on our ideas. An earlier version of this article was accepted for presentation at the 80th Annual Meeting of the Academy of Management, Vancouver, BC, Canada.

Supplemental material for this article is available with the manuscript on the JOM website.

ORCID iD

Chad H. Van Iddekinge

Notes

References

Ackerman

P. L.

1987. Individual differences in skill learning: An integration of psychometric and information processing perspectives. Psychological Bulletin, 102: 3-27.

Ackerman

P. L.

Kanfer

Goff

1995. Cognitive and noncognitive determinants and consequences of complex skill acquisition. Journal of Experimental Psychology: Applied, 1: 270-304.

Almeda

Schawbel

2015. The corporate culture and boomerang employee study. WorkplaceTrends. Retrieved from https://workplacetrends.com/the-corporate-culture-and-boomerang-employee-study

Arbuckle

J. L.

2014. Amos (Version 25.0) [computer program]. Chicago: IBM SPSS.

Arthur

Jr. Day

E. A.

McNelly

T. L.

Edens

P. S.

2003. A meta-analysis of the criterion-related validity of assessment center dimensions. Personnel Psychology, 56: 125-153.

Arvey

R. D.

Bouchard

T. J.

Segal

N. L.

Abraham

L. M.

1989. Job satisfaction: Environmental and genetic components. Journal of Applied Psychology, 74: 187-192.

Becker

G. S.

1964. Human capital: A theoretical and empirical analysis, with special reference to education. New York: National Bureau of Economic Research.

Beus

J. M.

Jarrett

S. M.

Taylor

A. B.

Wiese

C. W.

2014. Adjusting to new work teams: Testing work experience as a multidimensional resource for newcomers. Journal of Organizational Behavior, 35: 489-506.

Bidwell

2011. Paying more to get less: The effects of external hiring versus internal mobility. Administrative Science Quarterly, 56: 369-407.

10.

Biesanz

J. C.

Deeb-Sossa

Papadakis

A. A.

Bollen

K. A.

Curran

P. J.

2004. The role of coding time in estimating and interpreting growth curve models. Psychological Methods, 9: 30-52.

11.

Bluedorn

A. C.

1978. A taxonomy of turnover. Academy of Management Review, 3: 647-651.

12.

Browne

2016. LinkedIn VP explains why ‘boomerang employees’ are becoming the new normal. Business Insider. Retrieved from https://www.businessinsider.com/linkedin-vp-explains-why-boomerang-employees-are-becoming-the-new-normal-2016-10

13.

Campion

M. A.

1991. Meaning and measurement of turnover: Comparison of alternative measures and recommendations for research. Journal of Applied Psychology, 76: 199-212.

14.

Chan

1998. The conceptualization and analysis of change over time: An integrative approach incorporating longitudinal mean and covariance structures analysis (LMACS) and multiple indicator latent growth modeling (MLGM). Organizational Research Methods, 1: 421-483.

15.

Chan

1996. External recruitment versus internal promotion. Journal of Labor Economics, 14: 555-570.

16.

Coff

Kryscynski

2011. Invited editorial: Drilling for micro-foundations of human capital-based competitive advantages. Journal of Management, 37: 1429-1443.

17.

Dalton

D. R.

Todor

W. D.

Krackhardt

D. M.

1982. Turnover overstated: The functional taxonomy. Academy of Management Review, 7: 117-123.

18.

Deary

I. J.

Whalley

L. J.

Lemmon

Crawford

J. R.

Starr

J. M.

2000. The stability of individual differences in mental ability from childhood to old age: Follow-up of the 1932 Scottish Mental Survey. Intelligence, 28: 49-55.

19.

Delaney

H. D.

Maxwell

S. E.

1981. On using analysis of covariance in repeated measures designs. Multivariate Behavioral Research, 16: 105-123.

20.

DeOrtentiis

P. S.

Van Iddekinge

C. H.

Ployhart

R. E.

Heetderks

T. D.

2018. Build or buy? The individual and unit-level performance of internally versus externally selected managers over time. Journal of Applied Psychology, 103: 916-928.

21.

De Pater

I. E.

Van Vianen

A. E. M.

Bechtoldt

M. N.

Klehe

. 2009. Employees’ challenging job experiences and supervisors’ evaluations of promotability. Personnel Psychology, 62: 297-325.

22.

DeRue

D. S.

Wellman

2009. Developing leaders via experience: The role of developmental challenge, learning orientation, and feedback ability. Journal of Applied Psychology, 94: 859-875.

23.

Dokko

Wilk

S. L.

Rothbard

N. P.

2009. Unpacking prior experience: How career history affects job performance. Organization Science, 20: 51-68.

24.

Dragoni

Vankatwyk

Tesluk

P. E.

2011. Developing executive leaders: The relative contribution of cognitive ability, personality, and the accumulation of work experience in predicting strategic thinking competency. Personnel Psychology, 64: 829-864.

25.

Fisher

V. E.

Hanna

J. V.

1931. The dissatisfied worker. New York: Macmillan.

26.

Florentine

2015. Seven reasons to hire back the right former employees. CIO. Retrieved from https://www.cio.com/article/2985257/hiring/7-reasons-to-hire-back-the-right-former-employees.html

27.

Ghiselli

E. E.

1974. Some perspectives for industrial psychology. American Psychologist, 29: 80-87.

28.

Gilliland

S. W.

1993. The perceived fairness of selection systems: An organizational justice perspective. Academy of Management Review, 18: 694-734.

29.

Harrison

D. A.

Virick

William

1996. Working without a net: Time, performance, and turnover under maximally contingent rewards. Journal of Applied Psychology, 81: 331-345.

30.

Hirschman

2000. Reserve space for rehires. Society for Human Resource Management. Retrieved from https://www.shrm.org/hr-today/news/hr-magazine/pages/0100hirschman.aspx

31.

L. T.

Bentler

P. M.

1999. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6: 1-55.

32.

Iacobucci

2010. Structural equations modeling: Fit indices, sample size, and advanced topics. Journal of Consumer Psychology, 20: 90-98.

33.

Jiang

Liu

McKay

P. F.

Lee

T. W.

Mitchell

T. R.

2012. When and how is job embeddedness predictive of turnover? A meta-analytic investigation. Journal of Applied Psychology, 97: 1077-1096.

34.

Judge

T. A.

Watanabe

1995. Is the past prologue? A test of Ghiselli’s hobo syndrome. Journal of Management, 21: 211-229.

35.

Kelly

2017. The pros and cons of hiring a boomerang employee. Recruiter. Retrieved from https://www.recruiter.com/i/the-pros-and-cons-of-hiring-a-boomerang-employee/

36.

Kirschenbaum

S. S.

1992. Influence of experience on information-gathering strategies. Journal of Applied Psychology, 77: 343-352.

37.

Lee

W. T.

Burch

T. C.

Mitchell

T. R.

2014. The story of why we stay: A review of job embeddedness. Annual Review of Organizational Psychology and Organizational Behavior, 1: 199-216.

38.

Locke

E. A.

1976. The nature and course of job satisfaction. In Dunnette

M. D.

(Ed.), Handbook of industrial and organizational psychology: 1297-1349. Chicago: Rand McNally.

39.

Louis

M. R.

1980. Surprise and sense making: What newcomers experience in entering unfamiliar organizational settings. Administrative Science Quarterly, 25: 226-251.

40.

Lyness

K. S.

Judiesch

M. K.

2001. Are female managers quitters? The relationships of gender, promotions, and family leaves of absence to voluntary turnover. Journal of Applied Psychology, 86: 1167-1178.

41.

Maertens

J. A.

Putter

S. E.

Chen

P. Y.

Diehl

Huang

2012. Physical capabilities and occupational health of older workers. In Hedge

J. W.

Borman

W. C.

(Eds.), The Oxford handbook of work and aging: 215-235. New York: Oxford University Press.

42.

Meglino

B. M.

Denisi

A. S.

Ravlin

E. C.

1993. Effects of previous job exposure and subsequent job status on the functioning of a realistic job preview. Personnel Psychology, 46: 803-822.

43.

Microsoft Corporation. n.d. Microsoft alumni network. Retrieved from http://microsoftalumni.com

44.

Mischel

Shoda

1995. A cognitive-affective system theory of personality: Reconceptualizing situations, dispositions, dynamics, and invariance in personality structure. Psychological Review, 102: 246-268.

45.

Mitchell

T. R.

Holtom

B. C.

Lee

T. W.

Sablynski

C. J.

Erez

2001. Why people stay: Using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44: 1102-1121.

46.

Mumford

T. V.

Campion

M. A.

Morgeson

F. P.

2007. The leadership skills strataplex: Leadership skill requirements across organizational levels. The Leadership Quarterly, 18: 154-166.

47.

Newman

D. A.

2009. Missing data techniques and low response rates: The role of systematic nonresponse parameters. In Lance

C. E.

Vandenberg

R. J.

(Eds.), Statistical myths and urban legends: 7-36. New York: Routledge.

48.

Ouellette

J. A.

Wood

1998. Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior. Psychological Bulletin, 124: 54-74.

49.

Owens

W. A.

1976. Background data. In Dunnette

M. D.

(Ed.), Handbook of industrial and organizational psychology: 609-644. Chicago, IL: Rand McNally.

50.

Phillips

J. M.

1998. Effects of realistic job previews on multiple organizational outcomes: A meta-analysis. Academy of Management Journal, 41: 673-690.

51.

Ployhart

R. E.

Van Iddekinge

C. H.

MacKenzie

W. I.

Jr.

2011. Acquiring and developing human capital in service contexts: The interconnectedness of human capital resources. Academy of Management Journal, 54: 353-368.

52.

Roberts

B. W.

Walton

K. E.

Viechtbauer

2006. Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132: 1-25.

53.

Rodgers

R. C.

Helburn

I. B.

Hunter

J. E.

1986. The relationship of seniority to job performance following reinstatement. Academy of Management Journal, 29: 101-114.

54.

Rosenbaum

J. E.

1979. Tournament mobility: Career patterns in a corporation. Administrative Science Quarterly, 24: 220-241.

55.

Roth

P. L.

Bobko

McFarland

L. A.

2005. A meta-analysis of work sample test validity: Updating and integrating some classic literature. Personnel Psychology, 58: 1009-1037.

56.

Rowlings

2017. Boomerang employees on the rise. Yoh. Retrieved from http://www.yoh.com/blog/boomerang-employees-who-are-they-and-why-they-are-on-the-rise

57.

Sackett

P. R.

Zedeck

Fogli

1988. Relations between measures of typical and maximum job performance. Journal of Applied Psychology, 73: 482-486.

58.

Schmidt

F. L.

Hunter

J. E.

Outerbridge

A. N.

1986. Impact of job experience and ability on job knowledge, work sample performance, and supervisory ratings of job performance. Journal of Applied Psychology, 71: 432-439.

59.

Schmitt

Ostroff

1986. Operationalizing the “behavioral consistency” approach: Selection test development based on a content-oriented strategy. Personnel Psychology, 39: 91-108.

60.

Shipp

A. J.

Furst

H. S.

Harris

T. B.

Rosen

2014. Gone today but here tomorrow: Extending the unfolding model of turnover to consider boomerang employees. Personnel Psychology, 67: 421-462.

61.

Swartz

. n.d. Hiring an employee back after they have worked for you. Monster. Retrieved from https://hiring.monster.ca/hr/hr-best-practices/workforce-management/employee-performance-management/hiring-an-employee-back-after-they-have-worked-for-you.aspx

62.

Swider

B. W.

Liu

J. T.

Harris

T. B.

Gardner

R. G.

2017. Employees on the rebound: Extending the careers literature to include boomerang employment. Journal of Applied Psychology, 102: 890-909.

63.

Taylor

P. J.

Small

2002. Asking applicants what they would do versus what they did do: A meta-analytic comparison of situational and past behaviour employment interview questions. Journal of Occupational and Organizational Psychology, 75: 277-294.

64.

Texas Instruments Corporation. n.d. “Perks at Work” is coming to TIAA. Retrieved from https://tialumni.org/perks-work-coming-tiaa

65.

Thornton

G. C.

III Kedharnath

2013. Work sample tests. In Geisinger

K. F.

Bracken

B. A.

Carlson

J. F.

Hansen

J.- I. C.

Kuncel

N. R.

Reise

S. P.

Rodriguez

M. C.

(Eds.), APA handbook of testing and assessment in psychology, Vol. 1. Test theory and testing and assessment in industrial and organizational psychology: 533-550. Washington, DC: American Psychological Association.

66.

U.S. Bureau of Labor Statistics. 2019. Number of jobs, labor market experience, and earnings growth: Results from a national longitudinal survey summary (USDL-19-1520). Washington, DC: U.S. Department of Labor.

67.

Van Iddekinge

C. H.

Arnold

J. D.

Frieder

R. E.

Roth

P. L

. 2019. A meta-analysis of the criterion-related validity of prehire work experience. Personnel Psychology, 72: 571-598.

68.

Viswesvaran

Ones

D. S.

Schmidt

F. L.

1996. Comparative analysis of the reliability of job performance ratings. Journal of Applied Psychology, 81: 557-574.

69.

Wanous

J. P.

1978. Realistic job previews: Can a procedure to reduce turnover also influence the relationship between abilities and performance? Personnel Psychology, 31: 249-258.

70.

Wernimont

P. F.

Campbell

J. P.

1968. Signs, samples, and criteria. Journal of Applied Psychology, 52: 372-76.

71.

White

H. C.

1970. Chains of opportunity: System models of mobility in organizations. Cambridge, MA: Harvard University Press.

72.

Woltz

D. J.

Gardner

M. K.

Bell

B. G.

2000. Negative transfer errors in sequential cognitive skills: Strong-but-wrong sequence application. Journal of Experimental Psychology and Learning, 26: 601-625.

73.

Zimmerman

2006. The boom in boomerangs. Workforce. Retrieved from https://www.workforce.com/2006/01/25/the-boom-in-boomerangs

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB