Effects of People-oriented Leadership and Subordinate Employability on Call Center Withdrawal Behaviors

A major organizational problem faced by call center management is personnel withdrawal behaviors, that is, lateness, absenteeism, and turnover (Hutchinson, Purcell, & Kinnie, 2000; Kleemann & Matuschek, 2002; Malhotra & Mukherjee, 2004; Rose, 2002; Schalk & van Rijckevorsel, 2007). An absenteeism rate of 5% has been reported for call centers versus the national USA average of 3.5% ( Management Today, 1999) and the turnover rate has surpassed 30% (Stuller, 1999). Such withdrawal behaviors are attributable to two broad factors. First, since call center operations are relatively low-value-added and subject to intense price competition, they converge toward a low-skill, low wage model of production, and employment relations which frequently include successive short-term contracts. Call centers offer jobs that require modest formal education insofar as the appraised candidate shows competencies in computer literacy, numeracy, and interpersonal communication (Batt, 2002). Thus, a high prevalence of female workers and youngsters probing temporal alternatives is observed (Holman, Batt, & Holtgrewe, 2009). Given the frequent conflict between work and home duties, absenteeism is greater among women than men (Scott & McClellan, 1990) and, because career exploration tends to happen at the beginning of one’s career, frequent turnover is to be expected among younger workers (Hechanova, 2013). Second, call center employees are subjected to high levels of stress (e.g., Tuten & Neidermeyer, 2004). This is so because the work tasks and the interactions with customers impose role overload and role conflict (Cordes & Dougherty, 1993; Kleemann & Matuschek, 2002; Singh, Goolsby, & Rhoads, 1994; Witt, Andrews, & Carlson, 2004). Wegge, van Dick, Fisher, West, and Dawson (2006) described the specific challenges posed by the organization of work (working in shifts, postures, computer malfunctioning, high noise level) and divided attention consuming demands (listening and speaking, inputting data and reading from the screen). Role conflict arises from the demands to be quick and simultaneously provide a service of high quality; conflict also arises between the negative emotions that surge when customers complain and the obligation to express positive feelings (Deery, Iverson, & Walsh, 2002, 2010; Grandey, Dickter, & Sin, 2004; Marcoux, 2012). In addition, the intensity of automated performance monitoring increases emotional labor and its perceived purpose affects job satisfaction (Welles, Moorman, & Werner, 2007). A well-known relationship exists between hindrance stressors and absenteeism and turnover (Gupta & Beehr, 1979; Podsakoff, LePine, & LePine, (2007) and, of course, between job satisfaction and withdrawal behaviors (Mobley, 1977; Tett & Meyer, 1993).

In this context, a number of studies have supported the benefits accruing from the implementation of high commitment management practices (see Clark, 2007; D’Cruz & Noronha, 2011; Harney & Jordan, 2008; Malhotra & Mukherjee, 2007; Schalk & Van Rijckevorsel, 2007; Schawfeli, Bakker, & van Rehnen, 2009) coupled with the creation of state-of-the-art workplaces seeking to communicate “people values” in call centers (Barnes, 2007). Using a nationally representative sample of call centers in the USA, Batt (2002) found that quit rates were lower in establishments that emphasized employee participation in decision making and in teams. Supervisor coaching (Liu & Batt, 2010) and support (Liaw, Chi, & Chuang, 2010) positively influence employee performance in call centers. However, whereas the influence of supportive leadership on absenteeism has been documented in other organizational contexts (e.g., Biron & Bamberger, 2012), its effects on absenteeism and turnover intention have not been demonstrated in call centers.

The Present Study

This study addresses supportive leadership using Lawrence, Lenk, and Quinn’s (2009) specific concept of people-oriented leadership. This is one of the four leadership orientations consistent with the Competing Values Framework of organizational culture (Cameron, Quinn, DeGraff, & Thakor, 2006; Hartnell, Yi Ou, & Kinicki, 2011; Quinn & Rohrbaugh, 1983), the other three being change-, results-, and process-oriented leadership orientations. People-oriented leadership entails encouraging participation, developing people, and acknowledging personal needs. We test the following hypothesis:

Hypothesis 1. Call-center supervisor’s people-oriented leadership diminishes subordinate’s turnover intention which, in turn, determines subordinate’s uncertified absenteeism.

Hypothesis 1 assumes the validity of one of the five possible modes of relationship between withdrawal behaviors identified by Seitz and Miner (2002): the spillover model of Beehr and Gupta (1978), which suggests that the two types of withdrawal behavior are positively correlated. This mode is tested in the present research taking the following form:

Increased turnover intention → increased uncertified absenteeism That is, we expect people-oriented leadership to influence uncertified absenteeism through turnover intention. The rationale of Hypothesis 1 has two components. First, supportive leadership has been shown to lead to decreased turnover: (Batt, 2002) and Mobley (1977) demonstrated that turnover intention is a precursor of actual turnover and mediates the job satisfaction-turnover relationship. Second, call-center employees with higher turnover intentions can be expected to experience greater degrees of freedom to be unjustifiable absent from work than employees who do not plan to leave the organization. This is postulated because the risk of losing the job as a consequence of being unjustifiable absent will entail fewer expected losses for workers who are already thinking in leaving the organization.

Of the other four modes of relationship between withdrawal behaviors identified by Seitz and Miner (2002), three were ignored in this study: the compensatory behaviors model formulated by Hill and Trist (1955), which indicates that as one of the behaviors occurs, the probability of occurrence of the other is reduced; the alternate forms model (Mobley, 1977; Rice & Trist, 1952; Rosse & Miller, 1984), which implies that the two behaviors are alternate forms of the same construct and hence essentially substitutable for one another; and the progression of withdrawal model, which postulates that the behaviors are enacted in an ordered sequence from least to most severe (Baruch, 1944; Melbin, 1961; Rosse, 1988). However, a fifth mode could not be ignored considering its robustness (Seitz & Miner, 2002). This is the independence forms model of March and Simon (1958), which proposes that the various withdrawal behaviors are independent of each other, from which it follows that turnover intention does not affect uncertified absence and vice versa. A meta-analysis of the literature seeking to provide estimates of the interrelationships between withdrawal behaviors found lack of support for a withdrawal construct encompassing lateness, absenteeism, and turnover; calculated a correlation of .25 between absenteeism and turnover; and reported some support for the progression of withdrawal model from lateness to absenteeism to turnover (Berry, Lelchook, & Clark, 2012). However, the meta-analysis did not consider turnover intention, which may have complex relationships with turnover (Li, Lee, Mitchell, Hom, & Griffeth, 2016) and Mobley (1977) did not specify the role of absenteeism in the job satisfaction-turnover sequence. Hence, we also tested the following hypothesis:

Hypothesis 2. Independent effects of supervisor’s people-oriented leadership occur on subordinate’s turnover intention and uncertified absenteeism.

But the influence of the people-oriented supervisor may not be the same across different types of subordinates. Perceived employability, or “an employee’s perception of how easy it is to find new employment” (Kirves, Kinnunen, De Cuyper & Mäkikangas, 2014, p. 46), represents the subjective dimension of a person’s career identity, personal adaptability, and social and human capital (Fugate, Kinicki, & Ashforth, 2004) and expresses his/her self-esteem as a worker. Not surprisingly, workers’ perceived degree of employability strengthens turnover intentions (Brody & Rubin, 2011; Stroh & Reilly, 1997; Van der Heide & Van der Heyden, 2006). Hence, highly employable call-center workers, likely to entertain ideas of quitting, can be expected to be less organizationally committed and less amenable to influence from their supervisors than employees who are more dependent on the organization and have a greater desire of remaining in their jobs. Therefore, the present study was also designed to generate evidence relevant to the following hypothesis:

Hypothesis 3. Call-center supervisor’s people-oriented leadership is associated with diminished subordinate’s turnover intention and uncertified absenteeism more strongly among low-employability than high-employability workers.

Organizational Context

The organization studied (Org) is part of an international service conglomerate that ranks second in the world and has 15% of the Latin American market of call centers. Org has operated in Peru during more than 10 years providing employment for more than 5,000 workers in Metropolitan Lima. These are divided into a Foreign Mobiles division dedicated to attend a foreign phone company and a division that has clients in various fields (banking, insurance, government, etc.). The Foreign Mobiles division has an average absenteeism of 7.9% which represents a loss of 5.5% of its income. How huge is its personnel turnover can be inferred from the fact that the median employee time with Org is 12 months.

Participants

The 728 client service representatives of Org’s Foreign Mobiles division were invited to participate in an online survey (June 2016). All of them had responsibilities entailing calling clients and responding to them. Their employee ID served to link their responses to their personal and work data in personnel files.

Measures

People-oriented leadership.Lawrence et al.’s (2009) Competing Values Framework Managerial Behavior Instrument includes a 9-item scale dealing with people-oriented leadership. The four orientations were derived through factor analysis from the Competing Values Framework Managerial Behavior Instrument. The Spanish version used here was obtained through forward-back translation of the 9 items. León, Burga-León, and Morales (2017) replicated Lawrence et al.’s (2009) hierarchical confirmatory factor analysis in the same setting of the present study, targetting three first-order factors and four second-order factors. The indicators of model adjustment were good (Tucker-Lewis Index = .965) or very good (RMSEA = .018, SRMR = .028). Only 379 employees filled in this part of the questionnaire.

Uncertified absenteeism. Org measures its employees’ absenteeism using an attendance marker, vacation control, and administration of medical dispenses, that is, objective indicators. The measure used in this study refers to uncertified absence and is automatically calculated. The period January-May 2016 was covered (N = 728).

Turnover intention. The following items were translated into Spanish by Alarco (2010): “Lately, I have many wishes to abandon this organization”, “Despite the obligations I have with this enterprise, I want to abandon my job”, “I would like to remain in this organization as long as I could” (inverse scoring), and “If I could, I would leave this job today”. Respondents used a five-point Likert scale. Only 268 employees filled in this part of the questionnaire, which came at the end of the questionnaire.

Employability.De Cuyper and De Witte (2009, p. 159) measured self-rated employability (SRE) as follows: “We … presented the items … alternating (those) referring to … quantitative and … qualitative SRE. The items for quantitative SRE were as follows: ‘I am optimistic that I would find another job with another employer, if I looked for one’, ‘I will easily find another job with another employer instead of my present job’, ‘I could easily switch to another job with another employer, if I wanted to’, and ‘I am confident that I could quickly get a similar job with another employer’”. The items for qualitative employability were comparable, except for the use of “a better job”. The author used a Spanish version tested by Alarco (2010) in Peru that included a five-point Likert response scale (N = 268).

Other variables. Other data used in the research included gender, age, education (1= secondary, 2 = university), number of children, number of hours worked per day by each employee, employee’s time at Org, and working shift (1 = morning, 2 = afternoon/night).

Analytic Strategy

Absence data in Org were highly skewed to the right and not normally distributed according to the Kolgomorov-Smirnov statistic (p < .001). Hence, bootstrapping with 1,000 samples was used in all the analyses. Residualized scores were obtained by setting constant the context variables (gender, age, etc.) and the total sample was divided at the median of employability into low-employability workers (n = 127) and high-employability workers (n = 128). The residualized scores were subjected to saturated path analyses to test hypotheses 1 and 2. The moderating effect of worker employability on the leadership-withdrawal relationship (Hypothesis 3) was evaluated comparing the path-analysis results of the two sub-samples and using the Baron and Kenny’s (1986) regression approach applied to the (not residualized) original scores.

Preliminary Analyses

Only 255 cases had scores on all the study variables. The internal-consistency reliabilities of people-oriented leadership, employability, and turnover intention were, respectively, α = .95, α = .88, and α = .78. Non-responders to the Competing Values Framework Managerial Behavior Instrument presented greater absenteeism (Mean = .0494) than responders (Mean = .0427), a significant difference (t = -2.118, p = .035). Similarly, non-responders to the employability and turnover questionnaires presented greater absenteeism (Mean = .052) than responders (Mean = .038), a significant difference (t = -2.515, p = .013).

Table 1 shows the means, standard deviations, and correlations between the study variables. People-oriented leadership correlated negatively with turnover intention, whereas education positively influenced employability. It can be seen that absenteeism negatively correlates with age, number of hours worked, and time in Org, whereas turnover intention increases with education, time in Org, and in the abnormal (afternoon and night) shifts. The positive relationship between employability and turnover intention is consistent with the literature (e.g., Van der Heide & Van der Heyden, 2006); on the other hand, the 0 correlation between the withdrawal behaviors is consistent with March and Simon’s (1958) model.

Table 1

Descriptive Statistics and Intercorrelations (n = 379 for variables 1-8, n = 728 for variable 9, n = 268 for variables 10-11)

Hypothesis Testing

Path analyses. The evidence contradicted Hypothesis 1 (effects of people-oriented leadership on absenteeism mediated by turnover intention) derived from the spillover model of Beehr and Gupta (1978) and Hypothesis 2 (independent effects of people-oriented leadership on turnover intention and absenteeism) derived from the March and Simon’s (1958) model: only direct leadership-intention relationships were observed (see Figure 1A). In contrast, positive results were obtained regarding Hypothesis 3: whereas people-oriented leadership determined both turnover intention and absenteeism among low-employability workers (Figure 1B), it did not among highly employable workers (Figure 1C).

Multiple regression. These analyses were conducted to cross-check the leadership x employability interaction. It can be seen in Table 2 that the results upheld Hypothesis 2 regarding turnover intention and Hypothesis 3 regarding absenteeism. Whereas people-oriented leadership was associated with reduced turnover intention across the board and neither people-oriented leadership nor employability affected absenteeism under regression models 1 and 2, their interaction did under model 3. It can also be noted in Table 2 that turnover intention was codetermined by educational level and time in the organization and absenteeism by age (with negative sign), number of children, and time in Org (with negative sign).

Table 2

Standardized Coefficients from the Regression of Turnover Intention and Absenteeism on Supervisor’s Leadership and Other Study Variables, per Regression Model (N = 255)

General linear model. To obtain a visual depiction of the leadership x employability interactive effects on absenteeism, the authors implemented a general linear model in which the remainder of study variables were held constant. Table 3 presents the results of the linear model and Figure 2 depicts the adjusted means for the four groups. All the absenteeism mean differences of the interaction were significant: the inferior and superior 95% confidence limits of the mean were .041 and .041 for low people-oriented leadership-low employability group, .036 and .036 for low people-oriented leadership-high employability group, .031 and .032 for the high people-oriented leadership-low employability group, and .043 and .043 for the high people-oriented leadership-high employability group.

Table 3

General Linear Model Results (N = 255)

The lack of spillover from turnover intention to absenteeism observed in the study contradicts Beehr and Gupta’s (1978) model regarding relationships between withdrawal behaviors and upholds March and Simon’s (1958) independent forms model. Past studies have tested these models using absenteeism and turnover, but not turnover intention (Berry et al., 2012). Studies in call centers of other countries are needed to establish the external validity of the findings, that is, whether they are particular to Peru or recur across international contexts.

The main findings of the study, however, are those showing summative effects of leadership and employability on turnover intention and interactive effects on absenteeism. Supervisor’s people-oriented leadership was associated with reduced turnover intention of subordinates regardless of their perceived level of employability. These results can be understood considering that the people-oriented supervisor, by being considerate with the needs of his/her subordinates, fomenting their participation in decisions, and promoting their development, probably makes the job and Org more attractive to them, thus leading to reductions in their turnover intentions regardless of whether they are more or less employable. The crossover leadership x employability interaction is more difficult to explain. Whereas people-oriented leadership was associated with reduced absenteeism among workers who perceived themselves as being less employable, this leadership orientation was simultaneously associated with an increased number of uncertified absences on the part of workers who saw themselves as more employable. The contrast between the main effects of leadership on turnover intention vis-à-vis the interactive leadership x employability effects on uncertified absenteeism can be understood considering that turnover intention is an inconsequential pencil-and-paper behavior whereas uncertified absenteeism is a work behavior and as such has important consequences for the worker. At Org, unjustified absences not only are immediately penalized but also count at the time of renewing the employment contract.

Since the degree of turnover intention of the worker does not bear on his/her decision to be absent at a particular moment, recent theoretical developments which have focused on proximal turnover states and have distinguished between enthusiastic leavers and stayers and reluctant leavers and stayers (Hom, Mitchell, Lee, & Griffeth, 2012; Li et al., 2016; Woo & Allen, 2014) do not appear to have potential for helping in the explanation of the leadership x employability crossover interaction entailing absenteeism observed in our study. On the other hand, Zimmerman (2008) reported the predictive validity of some of the Big Five personality factors regarding turnover behaviors; Woo and Allen (2014) demonstrated the relevance of positive versus negative affectivity; and Tuten and Neidermeyer (2004) showed that pessimists report lower turnover intent than optimists in call centers. The crossover interaction entailing absenteeism could be explained by greater positive affect and optimism among workers who are high in employability. Perhaps the highly employable worker under a people-oriented supervisor presents higher absenteeism because he/she expects fewer sanctions than one under a non-considerate supervisor and the worker who perceives him/herself to be scarcely employable presents lower absenteeism because if he/she needs more the approval of, and wishes to reciprocate, the consideration of the people-oriented supervisor. These interpretive hypotheses are highly speculative, yet testable. Practical implications of the results cannot be inferred given the ambiguous status of the underlying determinants.

The main limitation of this study is its questionnaire response rate. Only about half the client service representatives of the Foreign Mobiles division responded to the leadership questionnaire and even less to the items on employability and turnover intention. Moreover, non-responders presented greater absenteeism than non-responders. On the other hand, it is possible that more clear-cut, sharper results would have been obtained from a more representative sample which increased the power of the study design. Replication studies should emphasize subject recruitment mechanisms that improve the response rate. Such studies are needed because the study findings reveal greater complexity in the determination of withdrawal behaviors than previously construed. If people-oriented supervision is associated with increased absenteeism among a type of subordinates and with decreased absenteeism among another type, the effects will cancel each other. Thus, there is a need for understanding the underlying determinants as a pre-condition to deriving practical recommendations.

Transcending such limitations, the study contributes to the literature questions not previously asked and three novel findings with the potential to generate further theorization and research: the independence of uncertified absenteeism from turnover intention, negative effects of people-oriented leadership on subordinate turnover intention regardless of subordinate level of employability, and leadership x employability crossover interactive effects on subordinate uncertified absenteeism. Whether these relationships are specific to the call center setting and the individual country in which the research was conducted are questions with the potential to trigger further research. However, since the research was conducted in an international organization with offices throughout the world, it is highly likely that the findings are widely generalizable.