1. Introduction
The organization of work was profoundly disrupted during the Covid-19 pandemic, especially when the authorities decided to confine the population in order to contain the coronavirus. Due to this emergency situation, healthcare workers were summoned to the front line (
Taylor 2020), and many companies and public structures used telework to ensure the continuation of business at a distance despite the lockdown in order to avert a social and economic crisis. During early stages of the Covid-19 pandemic, based on a survey on 229 Human Resources (HR) departments,
Gartner (
2020) stated that one-half of the companies had more than 80% of their working staff working remotely. Several surveys carried out during the confinement period in France, between late March and early April, estimated that 25% of employees had to work remotely (e.g., Direction de l’Animation de la Recherche, des Études et des Statistiques—DARES Flash Survey (
DARES 2020);
Odoxa-Adviso Partners 2020). As a point of reference, according to the
DARES (
2019), 7% of employees were teleworking in France in 2017 (1.8 million remote workers), but only 3% of employees teleworked regularly and at least one day per week. This outstanding increase in telework has surpassed all the projections made before the pandemic, which certainly anticipated an acceleration in the use of telework, but in a more gradual way. Scholars and professionals outline that this “mandatory work-from-home global experiment” has quashed some of the most frequently heard concerns (i.e., the belief that there are few jobs that can be done from home, the fear of lack of control over people, and so on), and that this has been achieved by enhancing digitalization and technological innovation (
Kniffin et al. 2020;
Rudolph et al. 2020).
Directors, HR officers, managers, and employees had to deploy new ways of working 100% remotely in unprecedented conditions and within noticeably short timeframes. The type of telework experienced during the crisis and containment period was very different from the definition of telework
1 before the Covid-19 health crisis; some authors refer to it as “mandatory work-from-home” (
Kniffin et al. 2020). Indeed, some specific aspects are not negligible: It was a forced, unprepared, and brutally imposed work organization, which was not negotiated or formalized. Moreover, “mandatory remote workers” were required to work within their place of confinement and in the presence of their personal entourage. While some employees had already experienced telework before, no one had experienced telecommuting at home under such conditions.
The surveys carried out in France (e.g., Malakoff Humanis, Empreinte humaine, or the National Agency for the Improvement of Working Conditions surveys) reveal that telework may have been an important resource in overcoming this period (e.g., feeling of social usefulness, maintaining relations with professional circles); however, they also emphasize that it has been a source of stress and has led to difficulties in reconciling professional and private life, overwork, and fatigue. Several scholars also addressed the main work and organizational psychology issues enhanced by the Covid-19 pandemic (
Rudolph et al. 2020;
Vaziri et al. 2020), especially taking into account the extreme increase of remote work. A substantial concern pertaining to the consequences of mandatory working from home for well-being has been shared by researchers: There is evidence of poorer working conditions, increasing demands, and some difficulty in ensuring adequate resources (
Bakker and Demerouti 2007;
Kniffin et al. 2020).
During the work-from-home experience, managers, supervisors, and coordinators played an important role, emphasizing the importance of deepening the contribution of leadership and virtual leadership and stressing the need to identify indicators for selection and training. The focus was mainly on the “right way” through which people in leadership positions could play their role in coordinating followers and promoting the organization’s well-being and sustainability. Despite the fact that there is evidence of an important presence of destructive leadership behavior in organizations (
Aasland et al. 2010;
Schyns and Schilling 2013;
Tepper 2007), little is known about the relationship between the dark side of leadership and some consequences on employees’ well-being in remote working conditions. Thus, the main purpose of this study was to contribute to the literature about supervisors’ role during remote working, considering the potential negative consequences of destructive leadership behaviors on workers’ well-being.
The study of remote work during the Covid-19 pandemic allows us to identify what the warning signs are in case of the implementation of telework beyond the emergency situation. Indeed, despite the fact that telework experienced during the lockdown was different from the “traditional” one, some aspects like destructive leadership behaviors can be examined as generalizable risk factors for remote workers’ well-being. In other words, the mandatory working from home during the emergency can be considered as a window through which to investigate dynamics that are less observable under standard conditions (
Kniffin et al. 2020).
Therefore, the purpose of the research was to investigate the role of leaders during the Covid-19 mandatory work-from-home situation in order to find potential solutions and formulate recommendations for practice. In particular, the findings and related recommendations could be useful in case of both a future societal crisis and the implementation of remote working practices under more traditional conditions. Indeed, we can hypothesize that remote working will also be maintained by companies after the extreme conditions caused by the Covid-19 pandemic. Some recent works (e.g.,
Barbuto et al. 2020) highlighted the benefits of remote working acknowledged by both organizations and workers; for example, cutting economical costs related to “home–work–home” travelling, saving time, and planning work autonomously. Some other recent papers have focused more on negative consequences of remote working on workers’ well-being (e.g.,
Molino et al. 2020).
Using the job demands–resources (JD-R) and recovery framework (
Kinnunen et al. 2011), our paper aimed to analyze in detail the relationship between destructive leadership behaviors perceived by workers concerning their supervisors and workers’ emotional exhaustion during the Covid-19 mandatory work-from-home situation in a sample of French remote workers. By doing so, the paper takes into account the mediating role of one resource—namely autonomy—two demands—namely cognitive demands and off-work-hours technology-assisted job demand (off-TAJD)—and recovery strategies (
Sonnentag and Fritz 2007). We suggested that destructive leadership impacts recovery and exhaustion through alterations in job demands and job resources. In addition to these variables, we also consider social support to better explain remote workers’ emotional exhaustion.
2. Destructive Leadership and Its Consequences
This paragraph introduces the theme of destructive leadership, placing the construct in the context of current work dynamics. Several informal reports claim that during the experience of mandatory working from home, the distance was bridged (through information technologies), allowing people to “be close” in work relationships, both positively—in terms of support and sharing content and information—and negatively—in terms of control, excessive requests, and invasion of work into personal living spaces. In this context, the leader–follower relationship represents a special link that needs to be deepened (
Kniffin et al. 2020;
Rudolph et al. 2020).
The leadership role is crucial in determining organizational outcomes and employees’ well-being, especially during a crisis (
Grabo et al. 2017): Assuring the effectiveness of leadership in remote working is, therefore, very important. Most studies (pre-Covid-19, during pandemics, but also “in perspective”) focus on analyzing the effects of positive leadership to define organizational actions and personal behaviors suitable for maximizing the achievement of objectives and the well-being of people (
Grabo et al. 2017).
Despite the large number of studies about the positive effects of transformational, charismatic, or authentic leadership (
Grabo et al. 2017;
Wong et al. 2010), little research has considered the dark side of leadership and its effects; however, some scholars, lately, have pointed out this aspect: Misconduct on behalf of people in coordination or supervision roles is not rare (destructive and constructive behaviors often alternate) and has important costs for people and for organizations (
Aasland et al. 2010;
Dirican and Erdil 2020;
Fosse et al. 2019;
Tepper 2007). Furthermore, a recent review by
Kaluza et al. (
2020) highlighted the negative relationship between destructive leadership and both followers’ and leaders’ well-being.
However, what about the dark side of remote leadership in mandatory working from home during the Covid-19 pandemic? Our paper aims at partially answering this question. The expression “the dark side of leadership” refers to a wide spectrum of words used to connote leadership derailment (
McCall and Lombardo 1983): toxic leadership (
Lipman-Blumen 2005), abusive or destructive leadership (
Higgs 2009), or the dark triad of leader personality (
O’Boyle et al. 2012); all of them describe a broad variety of leaders who engage in negative behaviors (
Pelletier 2010), such as “intimidating, bullying, manipulating (Machiavellianism), micromanaging, arrogance (narcissism), and engaging in abusive or unethical behavior” (
Webster et al. 2016, p. 346). Destructive leadership has been defined as “a process in which over a longer period of time, the activities, experiences, and/or relationships of an individual or the members of a group are repeatedly influenced by their supervisor in a way that is perceived as hostile and/or obstructive” (
Schyns and Schilling 2013, p. 141).
Krasikova et al. (
2013) defined it as “volitional behavior by a leader that can harm or intends to harm a leader’s organization and/or followers” (p. 1310).
Destructive leadership can be defined as a specific form of negative leadership characterized by the implementation of systematic and repeated behavior that violates the organization’s interest and the followers’ well-being or job satisfaction (
Einarsen et al. 2007). Some scholars argued that this destructive behavior may not be meant to damage subordinates or organizations; instead, it is the consequence of a leader’s insensitivity, incompetence, and negligence (
Einarsen et al. 2007). Conversely, other scholars (e.g.,
Krasikova et al. 2013) defined destructive leadership by also considering volition, and differentiated it from ineffective leadership by pointing out that, usually, it is the nature of a destructive leader’s choice to follow an objective or to behave in a certain way that is damaging.
There is more and more scientific evidence that managers and supervisors can exercise their role in a dysfunctional way, with significant negative consequences for individuals and organizations (
Benson and Hogan 2008;
Chatterjee and Hambrick 2007;
Fosse et al. 2019;
Ghislieri and Gatti 2012;
Ghislieri et al. 2019;
Schyns and Schilling 2013). Destructive leadership negatively impacts people’s well-being, and may cause physical health issues, emotional harm, and psychological distress (
Webster et al. 2016). Among the consequences, we can include the feeling of being constantly diminished and reminded of previous errors or the devaluation of one’s opinions (
Tepper 2000), as well as job tension and emotional exhaustion (
Harvey et al. 2007), lower job satisfaction and engagement (
O’Donoghue et al. 2016;
Tepper 2000), lower job-embeddedness (
Dirican and Erdil 2020), reduced family well-being (
Hoobler and Brass 2006), workaholism (
Molino et al. 2019), absenteeism, turnover intention and low performance (
Tepper 2007), and physical health problems (
Einarsen et al. 2010;
Kelloway and Barling 2010;
Macki 2008;
Tepper 2007). Destructive leaders generate tension and impose a controlling work environment that induces work intensification in order to avoid negative feedback (
O’Donoghue et al. 2016).
3. The Job Demands–Resources and Recovery Model
Exhaustion, which is the negative consequence of intensive affective, physical, and cognitive strain, is considered the principal component of burnout (
Demerouti et al. 2010). Originally, the JD-R Theory (
Bakker and Demerouti 2007;
Bakker and Demerouti 2017) was introduced to explain how specific working conditions may impact exhaustion; according to the theory, the risks of exhaustion and loss of psychosomatic well-being are highest when job demands are high and job resources are limited. Job demands have been defined as “those physical, psychological, social, or organizational aspects of the job that require sustained physical and/or psychological (cognitive and emotional) effort or skill and are therefore associated with certain physiological and/or psychological costs” (
Bakker and Demerouti 2007, p. 312). They are not always detrimental; however, when they require a great expenditure of energy and resources that is not followed by an adequate recovery, they can lead to stress (
Meijman and Mulder 1998).
In addition to destructive leadership, among job demands, we considered cognitive demands and the use of technology for work ends. Cognitive demands refer to the degree to which working tasks require constant mental and cognitive effort (
Bakker et al. 2004). While physical demands are crucial in certain occupations, cognitive demands are critical for remote workers who are continuously involved in cognitive tasks, such as processing information, making decisions, and solving problems. Another crucial aspect for remote workers is the use of technology, which, in the last few years, has revolutionized the ways in which people work and communicate. While information and communication technologies have facilitated and accelerated several working processes and expanded the amount of available information, they have also exposed employees to higher risks of work-related stress and excessive use (
Ghislieri et al. 2018;
Vayre and Vonthron 2019); this detrimental process has also been found in a sample of Covid-19 remote workers (
Molino et al. 2020). Indeed, the increased use of information and communication technologies has produced expectations about individuals always being connected and responsive and working better and faster than in the past. Being constantly connected to work obstructs psychological recovery and, in turn, leads to exhaustion and health problems (
Derks et al. 2014). When an employee perceives that his/her organization asks him/her to do supplemental work duties during non-standard working hours thanks to the aid of technology, we are in the presence of a specific job demand that is called off-TAJD in the literature (
Ghislieri et al. 2017). Previous studies highlighted its positive association with work–family conflict (
Ghislieri et al. 2017) and exhaustion through the mediation of workaholism (
Molino et al. 2019). To the best of our knowledge, the mediation of recovery between off-TAJD and exhaustion has not been investigated yet. Firstly, we hypothesize a positive relationship between job demands and exhaustion, as illustrated by
Figure 1.
Hypothesis 1. (a) Destructive leadership, (b) cognitive demands, and (c) off-TAJD are directly and positively related to exhaustion in male and female remote workers.
Together with job demands, the JD-R Theory classifies job resources and defines them as “physical, psychological, social, or organizational aspects of the job that are functional in achieving work goals, reducing job demands and the associated physiological and psychological costs, or stimulating personal growth, learning, and development” (
Bakker and Demerouti 2017, p. 274). In this study, we examined the roles of two of the most investigated job resources, namely job autonomy and social support, which seem to be particularly relevant during Covid-19 mandatory working from home. Job autonomy refers to the degree to which the job gives the employee substantial liberty and freedom to perform his/her own tasks and schedule work activities (
Marchese and Ryan 2001). Social support refers to the level of support that an individual has perceived from his/her colleagues and relatives during the lockdown. Hence, we hypothesize a positive relationship between job resources and exhaustion.
Hypothesis 2. (a) Autonomy and (b) social support are directly and negatively related to exhaustion in male and female workers.
Based on the previous introduction to destructive leadership, we may also assume that abusive, hostile, and non-supportive supervisor behavior may impact working conditions (
Schyns and Schilling 2013). Specifically, leaders’ tendency to control their subordinates may deprive workers of their autonomy and discretion in how to perform working tasks. Moreover, destructive leaders often show expectations that are difficult (or impossible) to satisfy and make so many requests that work becomes particularly demanding and mentally strenuous. They also expect that workers are constantly available and also push them to work during off-work hours with the aid of technological tools.
Hypothesis 3. Destructive leadership is directly and positively associated with (a) cognitive demands and (b) off-TAJD and (c) directly and negatively associated with job autonomy in male and female remote workers.
As previously mentioned, in the absence of adequate recovery, individuals are exposed to stress. Indeed, recovery has been defined as “a process during which individual functional systems that have been called upon during a stressful experience return to their pre-stressor levels” (
Sonnentag and Fritz 2007, p. 205). If the recovery process is not adequate or sufficient, an individual needs to allocate extra effort to maintain a satisfactory job performance level; in the long term, this extra effort may turn into health problems (
Meijman and Mulder 1998). The recovery process can be explained by the Effort–Recovery Model (
Meijman and Mulder 1998) and the Conservation of Resources Theory (
Hobfoll and Wells 1998). The Effort–Recovery Model (
Meijman and Mulder 1998) starts from the assumption that physiological activation and fatigue are normal consequences of effort expenditure at work. In order to have an effective recovery process, the functional systems activated during work activities should not be strained anymore; otherwise, they cannot return to pre-stressor levels. Furthermore, the Conservation of Resources Theory (
Hobfoll and Wells 1998) states that the individuals attempt to defend and maintain their resources to protect themselves against stress. The recovery process is essential to restore internal resources, such as the energy, self-efficacy, and positive mood that have been expended during the working day, as well as to gain new resources (
Sonnentag and Fritz 2007).
Sonnentag and Fritz (
2007) proposed four recovery experiences, which are off-job activities that allow the recovery process and the replacement of resources. (1) Psychological detachment, which is the psychological and cognitive detachment from work, and (2) relaxation, the state of calm and low activation, include those activities that do not mobilize the functional systems and resources involved at work. In addition, (3) mastery and (4) control embrace all activities that foster the acquisition of new resources and the replenishment of the endangered ones. Mastery refers to activities performed during off-work time that challenge an individual in learning new skills, different from those used at work, while control is the degree to which a person can freely decide how to spend his/her leisure time.
Hypothesis 4. Recovery experiences are directly and negatively related to exhaustion in male and female remote workers.
Several studies have highlighted that inadequate recovery leads to stress (
Kinnunen et al. 2011;
Sonnentag and Bayer 2005), while sufficient recovery is positively related to well-being and positive job performance (
Fritz and Sonnentag 2005;
Sonnentag and Bayer 2005;
Westman and Etzion 2001). Moreover, recovery has the ability to buffer the relationship between work–family conflict and both psychological strain and life satisfaction (
Moreno-Jiménez et al. 2009), as well as between workload and work–family conflict (
Molino et al. 2015). In addition,
Derks and Bakker (
2014) showed that both detachment and relaxation moderate the relationship between smartphone use and work–home interference. Following the extension of the JD-R Theory proposed by
Kinnunen et al. (
2010), who investigated the mediation role of recovery in the JD-R model, we intended to understand whether recovery experiences mediate the relationship between specific job demands, resources, and exhaustion experienced during Covid-19 teleworking.
Particularly, we expect that a high level of working requests may prevent the recovery process. If the worker is involved in tasks characterized by mental fatigue during the working day, it will be more difficult for him/her to stop thinking about problems and decisions and detach from work at the end of the working day, as well as to relax and be involved in different activities. Moreover, the use of technologies for work purposes during off-work hours has a direct negative impact on the recovery, since until the individual is involved in working thoughts and/or activities, it is impossible for him/her to effectively activate the processes needed to restore depleted resources and gain new ones (
Derks et al. 2014).
Hypothesis 5. (a) Cognitive demands and (b) off-TAJD are directly and negatively related to recovery experiences in male and female remote workers.
On the contrary, we assume that job resources may support and foster the recovery process. Having control and autonomy with respect to how to organize and perform one’s own working tasks leaves the workers the chance to decide when to interrupt working activities and when a break is needed, and there are more opportunities to preserve recovery experiences. In addition, receiving support from others, both colleagues and relatives, relieves the individual from requirements and worries, leaving more space for the recovery process.
Hypothesis 6. (a) Autonomy and (b) social support are directly and positively related to recovery in male and female remote workers.
As described above, an inadequate recovery process may turn into stress and health problems (
Meijman and Mulder 1998). Therefore, the effects of job demands and job resources on recovery also imply an indirect impact on exhaustion; the effect will be negative for demands (cognitive demands and off-TAJD) and positive for resources (autonomy and social support).
Hypothesis 7. (a) Cognitive demands and (b) off-TAJD are indirectly and positively associated with exhaustion through the mediation of recovery experiences in male and female remote workers.
Hypothesis 8. (a) Autonomy and (b) social support are indirectly and negatively associated with exhaustion through the mediation of recovery in male and female remote workers.
Finally, we investigated the role of destructive leadership in the JD-R and recovery model. As hypothesized above, destructive leadership is expected to be positively related to cognitive demands and off-TAJD and negatively related to job autonomy. As a consequence, through an increase of job demands and a decrease of job autonomy, destructive leadership may also show an indirect negative effect on recovery and an indirect positive effect on exhaustion.
Hypothesis 9. Destructive leadership is indirectly and negatively associated with recovery experiences through the mediation of (a) cognitive demands, (b) off-TAJD, and (c) autonomy in male and female remote workers.
Hypothesis 10. Destructive leadership is indirectly and positively associated with exhaustion through the mediation of (a) cognitive demands, (b) off-TAJD, (c) autonomy, and (d) recovery in male and female remote workers.
In this general framework, this study intended to investigate male–female differences. Through their meta-analysis,
Purvanova and Muros (
2010) highlighted that women are slightly more emotionally exhausted than men (on the contrary, men are more depersonalized than women). With regard to recovery experiences, early studies did not report gender differences (e.g.,
Sonnentag and Bayer 2005;
Sonnentag and Fritz 2007) and, in general, this issue has received little attention in the literature so far.
Nolen-Hoeksema and Harrell (
2002) reported that rumination and cognitive activation are higher for women; thus, we should expect gender differences in psychological detachment. Nevertheless, the results are contradictory. For example, higher levels of psychological detachment have been underlined for both women (
Gluschkoff et al. 2017;
Shimazu et al. 2012) and men (
Kinnunen et al. 2010). Furthermore, many other studies did not find any significant difference (e.g.,
Bakker et al. 2015).
4. Method
4.1. Participants and Procedure
A convenience sample of 716 participants took part in this study. Participants filled in a self-report online questionnaire on the Google Moodle Platform at the beginning of April 2020 during the Covid-19 lockdown. All participants gave informed consent; the researchers ensured the confidentiality and analyzed data in an aggregated manner. The objective of the research, the voluntary and unpaid participation, the data treatment information, and instructions to complete the questionnaire were emphasized in the cover letter. Participants were mainly recruited by members of the research team who work in different universities in France. The study observed the Helsinki Declaration (
World Medical Association 2013); the used procedures could not threaten participants’ psychological or social well-being. The Bioethical Committee of the University of Turin approved the study (document n. 150561, 3 April 2020). Participants were informed about the voluntary and unpaid participation in the research, the confidentiality of the data, and the anonymity; we obtained informed consent from all of them.
The participants were all workers from different sectors and in different role positions who evaluated their working conditions and their supervisors’ behaviors during the mandatory working from home. The average age of the whole sample was around 36 years old (SD = 11.77, min. = 20; max. = 63). The participants were from different occupational sectors: 59.6% services, 16.3% education and research, and13.0% industry; the remaining participants (11.1%) were from other sectors, such as the arts sector. Among the participants, 49.6% worked in large organizations (250 or more employees), 21.1% in medium organizations (50–249 employees), 21.5% in middle-small organizations (10–49 employees), and 7.7% in small organizations (1–10 employees). Weekly remote working days were, on average, 4.5 (SD = 1.10), and weekly working hours were, on average, 29 (SD = 13.10). In the sample, 437 were women (61.0%) and 279 were men (39.0%).
In the female sample, 52.4% had children and 47.6% declared that they were primarily responsible for the care of the children. The average age was 36.88 years (SD = 11.82; min. = 20; max. = 63). Most of the female sample had a full-time job (84.4%) and a permanent contract (63.4%), worked in the private sector (74.2%), and had a bachelor’s or master’s degree (87.5%). As for the job profile, 50.1% were office workers, 44.4% middle managers, and 3.7% managers; the remaining participants had other job profiles. Mean seniority in the profession was 8.50 years (SD = 9.76). Only 38.2% of female participants had already remotely worked in the past, for an average of 2.17 days per week (SD = 3.77) and a total of 16.32 months, on average (SD = 23.32).
In the male sample, 43.4% had children and 31% declared that they were primarily responsible for the care of the children, elderly parents, or other family members. The average age was 35.65 years (SD = 11.67; min. = 20; max. = 63). Most of them had a full-time job (92.5%) and a permanent contract (71.9%), worked in the private sector (74.8%), and had a bachelor’s or master’s degree (90.4%). The job profile was office workers (40.5% of participants), middle managers (48.0%), and managers (6.8%); the remaining participants had other job profiles. The mean seniority in the profession was 7.44 years (SD = 8.55). Among male participants, 33.3% had already remotely worked in the past, for an average of 2.50 days per week (SD = 3.64) and a total of 12.98 months, on average (SD = 14.49).
4.2. Measures
Emotional exhaustion was measured using eight items of the Oldenburg Burnout Inventory (OLBI) (
Demerouti et al. 2010), applying a five-point Likert scale, from 1 = strongly disagree to 5 = strongly agree. Participants were asked to think about the last ten days and assess statements such as “during my work, I often feel emotionally drained.” Cronbach’s alpha was 0.77 for the whole sample, 0.78 for the female subsample, and 0.72 for the male subsample.
Recovery was detached using 12 items (
Sonnentag and Fritz 2007), applying a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree. Four dimensions define the factor structure of this scale: psychological detachment, relaxation, mastery, and control; each one was measured through three items. Participants were asked to think about what they did during the last ten days in the evening after they stopped working. For example, they were asked if they “...forgot about work” (psychological detachment), “...did relaxing things” (relaxation), “...sought out intellectual challenges” (mastery), and “...decided their own schedule” (control). For psychological detachment, Cronbach’s alpha was 0.91 for the whole sample as well as for the female and male subsamples. As for the relaxation dimension, Cronbach’s alpha was 0.92 for the whole sample, 0.93 for the female subsample, and 0.92 for the male subsample. Regarding the mastery dimension, Cronbach’s alpha was 0.84 for the whole sample, 0.85 for the female subsample, and 0.83 for the male subsample. Finally, for the control dimension, Cronbach’s alpha was 0.88 for the whole sample, 0.89 for the female subsample, and 0.86 for the male subsample.
Autonomy was measured using seven items adapted from
Karasek et al. (
1998), applying a six-point Likert scale from 1 = strongly disagree to 6 = strongly agree. Participants were asked to evaluate the degree of autonomy they had in their work during the last ten days; an example item is “Your work allows you to autonomously manage your activities.” Cronbach’s alpha was 0.93 for the whole sample and also for the male and female subsamples.
Social support was detached using three items ad hoc, applying a five-point Likert scale from 1 = very little to 5 = very much. We asked participants how much they felt supported by their colleagues, manager, and relatives over the last ten days. Cronbach’s alpha was 0.68 for the whole sample, 0.65 for the female subsample, and 0.73 for the male subsample.
Cognitive demands were measured using four items (
Bakker et al. 2003), applying a five-point Likert scale from 1 = never to 5 = always. Thinking about the last ten days, workers assessed how much their work had been cognitively demanding; an example item is: “Your work required a lot of concentration.” Cronbach’s alpha was 0.78 for the whole sample, 0.80 for the female subsample, and 0.74 for the male subsample.
Off-TAJD was assessed using four items adapted from the scale of
Ghislieri et al. (
2017), which asked how often during the last ten days the workers perceived that the organization required them to work beyond the traditional working hours with the aid of technologies. All items were evaluated on a five-point scale ranging from 1 = never to 5 = always. An example item is “How often does your organization require you to answer phone calls and emails during off-hours?”. Cronbach’s alpha was 0.89 for the whole sample, 0.89 for the female subsample, and 0.91 for the male subsample.
Destructive leadership was detached using four items (
Schmidt 2008), applying a five-point Likert scale ranging from 1 = never to 5 = always. Participants were asked to think about their direct supervisor (the formal one or, if they did not have a formal supervisor, the person who they informally recognize as supervisor) and to evaluate his/her behaviors during the last ten days. An example item is: “My supervisor invades the privacy of subordinates.” Cronbach’s alpha was 0.77 for the whole sample, 0.79 for the female subsample, and 0.73 for the male subsample.
4.3. Data Analysis
The IBM SPSS 25 statistics software was used to perform descriptive data analysis in the whole sample and in the male and female subsamples. Pearson correlations were performed to detect relationships between variables, and Cronbach’s alpha coefficient was calculated to verify scales’ reliability, while analysis of variance (t-test for independent samples) was performed to compare the means of the variables of the male and female samples.
In order to estimate the hypothesized model, a full multi-group structural equation model (MG-SEM) was calculated by Mplus7. Age, concern about Covid-19, and the presence of children, parents, or other family members in charge were used as control variables. The method of estimation was maximum likelihood. The item parceling technique (
Little et al. 2002) was applied to the two variables with the highest number of items (exhaustion and recovery) for reasons of parsimony, creating two parcels for exhaustion and four for recovery. Several goodness-of-fit criteria were used to assess the model (
Bollen and Long 1993): the χ
2 goodness-of-fit statistic; the root mean square error of approximation (RMSEA); the comparative fit index (CFI); the Tucker–Lewis index (TLI); and the standardized root mean square residual (SRMR). Finally, in order to test the mediation hypotheses, a bootstrapping procedure was used (2000 new samples extracted from the original sample). The measurement model was also tested through a multi-group confirmatory factor analysis (MG-CFA) that reported an acceptable fit to the data [χ2 (1066, NF = 437, NM = 279) = 2191.87,
p < 0.001, RMSEA = 0.05 (0.05, 0.06), CFI = 0.91, TLI = 0.90, and SRMR = 0.08].
5. Results
The findings evidenced significant differences between the female and male groups concerning the levels of recovery, exhaustion, and the concern about Covid-19. Analysis of variance showed lower levels of recovery among women (
M = 3.35,
SD = 0.84) than among men (
M = 3.49,
SD = 0.79) [
t(714) = 2.32,
p < 0.05]; in particular, two dimensions of recovery significantly differed: mastery and detachment. The level of mastery was higher for men (
M = 3.45,
SD = 0.98) than for women (
M = 3.24,
SD = 1.08) [
t(714) = 2.60,
p < 0.05], as well as the degree of detachment, which was higher for men (
M = 2.93,
SD = 1.20) than for women (
M = 2.72,
SD = 1.19) [
t(714) = 2.34,
p < 0.05]. The results also showed a significant difference for exhaustion: Women reported more exhaustion (
M = 2.89,
SD = 0.94) than men (
M = 2.63,
SD = 0.86) [
t(629, 260) = 3.75,
p < 0.001]. Finally, women (
M = 3.50,
SD = 0.99) were more worried about Covid-19 than their male counterparts (
M = 3.24,
SD = 1.15) [
t(525,933) = 3.05,
p < 0.01]. The other variables, namely autonomy, cognitive demands, destructive leadership, off-TAJD, and social support, did not show any significant differences.
Table 1 includes descriptive statistics, correlations, and alpha values (for the whole sample, α values range 0.68–0.93). As shown by
Table 1, exhaustion was significantly correlated with all variables considered in this model in both male and female workers. In addition, recovery was significantly correlated with all the other variables.
The MG-SEM of the hypothesized model fit adequately to the data [χ
2 (881, NF = 479, NM = 279) = 1771.49,
p < 0.001; CFI = 0.91; TLI = 0.91; RMSEA = 0.05 (0.05; 0.06); SRMR = 0.07].
Figure 2 shows the standardized parameters, and
Table 2 shows the item loadings for all latent factors. Hypothesis 1 stated that job demands are positively related to exhaustion in both subsamples. The findings showed that only cognitive demands (Hypothesis 1b) had a strong and positive relationship with exhaustion in both the female and male samples [F: β = 0.35,
p < 0.001; M: β = 0.37,
p < 0.001]. With regard to job resources, Hypothesis 2 was fully confirmed; indeed, the findings showed a significant and negative association of autonomy (Hypothesis 2a) [F: β = −0.19,
p < 0.001; M: β = −0.23,
p < 0.001] and support [F: β = −0.21,
p < 0.01; M: β = −0.26,
p < 0.01] (Hypothesis 2b) with exhaustion in female and male workers. Hypothesis 3 assumed that destructive leadership is positively related to cognitive demands (Hypothesis 3a) and off-TAJD (Hypothesis 3b), and negatively related to autonomy (Hypothesis 3c). This third hypothesis was fully confirmed. In both the female and male sample, the results revealed a strong and positive association of destructive leadership with cognitive demands [F: β = 0.34,
p < 0.001; M: β = 0.37,
p < 0.001] and off-TAJD [F: β = 0.22,
p < 0.001; M: β = 0.19,
p < 0.001] and a negative association with autonomy [F: β = −0.14,
p < 0.01; M: β = −0.14,
p < 0.01]. Hypothesis 4, which assumed a negative relationship between recovery and exhaustion, was confirmed in both the female and male samples [F: β = −0.45,
p < 0.001; M: β = −0.52,
p < 0.001]. As regards the negative association between job demands and recovery, Hypothesis 5 was fully confirmed in both samples: Cognitive demands [F: β = −0.13,
p < 0.05; M: β = −0.12,
p < 0.05] (Hypothesis 5a) and off-TAJD [F: β = −0.19,
p < 0.001; M: β = −0.21,
p < 0.001] (Hypothesis 5b) showed a significant negative association with recovery. Hypothesis 6 stated that job resources were positively related to recovery. This hypothesis was also fully confirmed in female and male workers: Autonomy [F: β = 0.28,
p < 0.001; M: β = 0.29,
p < 0.001] (Hypothesis 6a) and support [F: β = 0.29,
p < 0.001; M: β = 0.31,
p < 0.001] (Hypothesis 6b) were positively related to recovery. Outside of the hypotheses, the control variables had a significant relationship with recovery and exhaustion. In particular, age showed a strong and negative relationship with exhaustion across the two samples [F: β = −0.22,
p < 0.001; M: β = −0.28,
p < 0.001], as well as the concern about Covid-19, which was statically positively related to exhaustion [F: β = 0.12,
p < 0.05; M: β = 0.19,
p < 0.05]. Moreover, the care of children, parents, or other family members was significantly and negatively related to recovery in the female and male samples [F: β = 0.19,
p < 0.05; M: β = 0.13,
p < 0.05]. Finally, the variance explained by the model was in female and male samples, respectively, 68% and 90% for exhaustion, 31% and 34% for recovery, 12% and 14% for cognitive demands, only 0.05% and 0.06% for off-TAJD, and 0.02% in both samples for autonomy.
Subsequently, the bootstrapping procedure tested the indirect effects by extracting 2000 new samples (
Preacher and Hayes 2008). A significant mediation is confirmed when the confidence interval does not include zero.
Table 3 shows the statistically significant mediated effects for the female and male samples. Specifically, confirming the Hypothesis 7, the results pointed out that the association between cognitive demands and exhaustion was partially mediated by recovery in female and male workers [0.06 and 0.06] (Hypothesis 7a), as well as the relation between off-TAJD and exhaustion, which was fully mediated by recovery [0.09 and 0.12] (Hypothesis 7b). Moreover, as assumed in Hypothesis 8, recovery was a partial mediator in the relation between social support and exhaustion [−0.13 and −0.17] (Hypothesis 8b), as well as between autonomy and exhaustion [−0.13 and −0.16] (Hypothesis 8a). Furthermore, the findings confirmed Hypothesis 9, since destructive leadership was indirectly and negatively associated with recovery experiences through the mediation of job demands and job resources in women and men, showing that the total indirect effects are equal to −0.13 and −0.12, respectively. Destructive leadership was also indirectly and positively associated with exhaustion (Hypothesis 10) through the mediation of job demands, job resources, and recovery with a total indirect effect equal to 0.20 for female and 0.28 for male.
6. Discussion
In line with its objectives, this study investigated, in an original way, the indirect and positive association between the frequencies of destructive leadership behaviors perceived by remote workers and their emotional exhaustion. The study of remote work imposed during the lockdown, despite its peculiarities, may support the empirical understanding of the negative and generalizable consequences of destructive leadership on remote workers’ well-being beyond the Covid-19 crisis. Following the extension of the JD-R Theory proposed by
Kinnunen et al. (
2010), in order to explain the relationship between destructive leadership behaviors and emotional exhaustion, we considered the mediating role of one job resource—namely autonomy—two demands—namely cognitive demands and the request to use technology for work reasons during non-work hours—and recovery strategies (
Sonnentag and Fritz 2007). Moreover, this research included a further job resource, namely social support, as an antecedent of both recovery and exhaustion. In detail, we tested a cascade model in which recovery strategies mediate the relationship between all job demands and resources on the one hand, and emotional exhaustion on the other. As can be deduced, all considered job demands and resources are not only relevant for remote work during a pandemic, but they are also typical in traditional telework, allowing generalizations outside the Covid-19 pandemic situation to be made.
In detail, destructive leadership behavior perceived by remote workers appeared to be related to a higher level of cognitive demands and more requests to use technology for work reasons during non-work hours, which, in turn, were negatively related to recovery strategies, which were able to decrease the risk of emotional exhaustion. Furthermore, destructive leadership was negatively associated with autonomy, which, in turn, was positively related to recovery strategies. These results confirmed that destructive leaders’ behaviors have a crucial role in influencing workers’ well-being, and that the distance created by remote working does not protect followers from potential negative consequences of their supervisors’ conduct. In particular, leaders’ negative behaviors, such as engaging in abusive or unethical monitoring and having excessive or unattainable requests, impact working conditions in such a way that the recovery process is compromised. Consequently, in the absence of adequate recovery, individuals experience more exhaustion (
Meijman and Mulder 1998), which can be attributed to supervisors’ behaviors.
More generally, the results also confirmed that the relationship of the set of job demands and resources with exhaustion seems to be better explained if we also consider the recovery experiences, in line with
Kinnunen et al. (
2011) and the Effort–Recovery Model (
Meijman and Mulder 1998). In particular, the relationship between the two job resources, namely autonomy and social support, was partially mediated by recovery experiences. The findings related to autonomy are in line with Bakker and colleagues’ evidence (2004), which prudently warned about the limited capability of autonomy to buffer the effects of job demands, but suggested the negative association with exhaustion. Moreover, the negative relationship between social support and exhaustion reinforces the idea of its protective role in the well-being of male and female workers, particularly in the presence of destructive leaders.
In addition to this, the relationship between cognitive demands and the off-TAJD was also partially and totally mediated by recovery experiences, respectively. This result represents an original contribution, especially for the off-TAJD variable. Indeed, previous studies focused their attention on the direct relation to work–family conflict (
Ghislieri et al. 2017) or to workaholism and, in turn, to exhaustion (
Molino et al. 2019), but no contribution considered the mediation of recovery so far.
In addition to job demands, job resources, and recovery strategies (
Kinnunen et al. 2011;
Meijman and Mulder 1998), we controlled for three variables: age, the concern about the Covid-19 pandemic, and the responsibilities for the care of children, parents, or other family members. The findings showed that two out of the three control variables contributed to explaining exhaustion. Specifically, in line with a previous meta-analysis (
Brewer and Shapard 2004) that underlined a negative correlation between burnout and age, the findings confirmed the negative relationship with exhaustion in female and male remote workers. Furthermore, the concern about the Covid-19 pandemic was positively related to exhaustion, confirming the normal stressor–strain psychological association; indeed, Covid-19 is a source of fear, worry, and stress related to a real threat (
World Health Organization 2020). Finally, taking care of children, parents, or other family members was negatively related to recovery, a preliminary result of which further investigation is recommended.
The findings suggested that the set of variables chosen predicted exhaustion better for male than for female remote workers. One possible explanation is that the model included mainly working-related variables, while familial aspects were poorly investigated. Considering the persistence of the traditional family model in which women remain the main member responsible for the care of children and elderly parents (
Saraceno 2013), family workload, which was not so extensively explored in this model, might better explain female exhaustion. During the lockdown, families were forced to simultaneously manage work and family without the traditional institutional support (
Fisher et al. 2020); this situation led to further increases in the care burden for women (
Power 2020).
Furthermore, the analysis of variance showed a higher level of recovery among men than among women and, on the contrary, a lower level of exhaustion and concern about the Covid-19 pandemic among men than among women. The tested model contributed to almost partially explaining these gender differences, since recovery is a predictor of exhaustion, as well as the concern about the Covid-19 pandemic. This explanation is certainly not exhaustive. A recent work by
Gualano et al. (
2020) conducted in Italy during the lockdown highlighted a positive association between being female and both anxiety and sleeping disturbances. Another study (
Liu et al. 2020) conducted in China reported that one month after the outbreak of the pandemic, women showed the most worrying posttraumatic stress symptoms in the domains of re-experiencing, negative alterations in cognition or mood, and hyper-arousal than men. Other scholars outside the Covid-19 pandemic situation found that the levels of exhaustion differed slightly between women and men (
Purvanova and Muros 2010), probably due to the double burden (working and family) that mainly concerns women and that limits the chances of recovery especially for them (
Nolen-Hoeksema and Harrell 2002). Gender inequalities deserve further investigation in the future in light of their potential impact on well-being and also on perceived productivity and job satisfaction, as shown by
Feng and Savani (
2020), who conducted a study on dual-career parents working from home during the Covid-19 pandemic.
Women continue to be considered the main people responsible for the family; globally, they spend more hours than men in housework and care activities such as cleaning, cooking, helping children in schoolwork, caring for elderly parents, and so on (
Moreira da Silva 2019;
Pozzan and Cattaneo 2020). Notwithstanding, it is important to point out that substantial differences persist among countries (
Ranci Ortigiosa and Pavolini 2015). Gender roles are strongly related to the context and to cultural norms; therefore, gender expectations in terms, for example, of housework change a lot across nations. France is a country in transition from a traditional family paradigm to a more egalitarian model. Nevertheless, French women still spend more time than men on childcare and other housework activities (
Pailhé et al. 2019). In this light, gender differences in these dynamics will require future cross-cultural studies.
7. Limitations and Future Studies
This research presents a solid starting point for examining how destructive leadership, job demands, and job resources are related to exhaustion through recovery among remote workers. While original and informative, the study has several limitations that must be noted. First, remote working from home as experienced during the crisis and the containment period is not reflective of telework as it was practiced before the crisis. Knowing that both the government and employees support the continuation of this work arrangement in France, our research model could be tested with post-pandemic remote workers. Furthermore, the study used a relatively small sample restricted to French remote workers. Complementary and cross-cultural studies could allow us to develop and refine knowledge in this area of research.
Second, our study focuses on the role of management style, job demands, and job resources among remote workers. Research comparing the effects of these processes in telework vs. on-site work situations within the same sample (comparing different work periods) or between comparable samples (comparing remote workers to non-remote workers) would allow us to expand our results. Third, the focus was solely on exhaustion, omitting other dimensions, such as remote workers’ health and quality of life, which are equally important. Future research could explore the role of supervisors, job demands, and resources in achieving the benefits of telecommuting (e.g., engagement, well-being, work–life balance). Moreover, particular attention should be paid to the role of organizational culture when it allows, or even feeds, destructive leadership behavior. Further studies could lead to a better understanding of how and to what extent tele-pressure (intended as a strong impulse to quickly answer to work-related messages;
Van Laethem et al. 2018) is considered as a tool of control in remote work and how it affects remote workers’ health and quality of life.
Fourth, the questionnaire was self-reported, entailing potential social desirability biases, and the study was cross-sectional. Longitudinal designs are preferred to make better conclusions about causal relationships among variables. In that sense, methods such as diary studies could enrich this study’s findings. Finally, additional studies should be designed to examine how the work domain encroaches on personal life and the role of remote working. The participants’ supervisors, colleagues, spouses, or partners should be interviewed to better understand how remote workers jointly manage the demands of work and home when they are teleworking at home. Particular attention could be paid to the potential differences between men and women, as well as gender inequalities, when teleworking at home.
8. Conclusions and Practical Implications
This paragraph introduces the implications for practice; some recommendations for selection, training, and counseling are discussed. The findings of the present paper could be useful for better management of potential future confinement periods. Furthermore, outside the lockdown caused by the Covid-19 pandemic, the results of this research could be generalizable and offer a starting point for understanding the dynamics of remote working. In particular, the findings allow us to identify interventions to prevent the negative consequences of destructive leadership on remote workers’ well-being.
The management of remote work at home is a major issue when considering its impact on employees’ health. This paper confirms that, although physically distant, destructive leaders may be “too close” to their followers through an increase in cognitive requests and a pressing use of technology, thus reducing autonomy to the point of limiting recovery and contributing to the increase of exhaustion. Therefore, it is important to weigh the opportunities and risks of telecommuting in terms of work organization, engagement, and efficiency (e.g., team supervision, performance appraisal, identification with company standards and norms), but also health and quality of life (e.g., workload, exposure to psychosocial risks, work–life balance, gender equality).
The risk of a negative interpretation of leadership must be understood already in the selection and, even more, socialization phase, paying attention to abusive and demanding behavior that violates followers’ personal life, also through the use of technology. This role interpretation can be fueled by organizational cultures that are demanding and inclined to test the loyalty of collaborators through excessive requests and tele-pressure (
Van Laethem et al. 2018). To reduce the impact of destructive leadership behavior, training is a fundamental element: In this regard, not only performative or prescriptive training is required, but targeted interventions are also important—with real working groups, wherever possible—which are capable of bringing out problematic behavior and the causes of that behavior by focusing on the roles of individuals and that of the organization (
Ghislieri and Gatti 2012).
Managers play a strategic role in the implementation and the success of remote work. The benefits of telecommuting are more likely to be realized when organizations adopt formal policies and when supervisors learn how to manage new ways of working (
Lautsch and Kossek 2011). Middle managers need to adjust to “managing by results” rather than by observing work hours, and as a result, to develop new skills and competences. The deployment of remote work implies that they are comfortable with the idea of giving their subordinates a certain amount of freedom in organizing their daily work tasks and in the progress of the projects and missions entrusted to them. Telework also requires managers to be able to delegate, to adopt decentralized work management, and to develop new performance evaluations and assessment systems (
Mello 2007). Training programs for managers may allow them to develop and implement new managerial practices, with attention to toxic behavior and to the organizational habits toward control and power over followers. These would enable them to establish a sustained and trusting relationship with their remote workers, and to make them aware of the difficulties encountered by remote workers (
Lautsch and Kossek 2011).
The training should also target followers by providing support in dealing with destructive leaders. Despite repeated calls for research aimed at defining effective strategies for coping with destructive leadership, evidence is scant (
Martinko et al. 2013;
Pelletier 2010;
Yagil et al. 2011). In this regard, in terms of remote work, it is essential to consider that coping strategies are influenced by “distance”. In the study of
Webster et al. (
2016), which was based on the categorization of coping responses by
Yagil et al. (
2011) and
Skinner et al. (
2003), a fundamental strategy for facing “the dark side of leadership” is the search for social support. However, this strategy can become difficult with distancing, leaving people more at the mercy of leadership toxicity, with a prevalence of dysfunctional reframing responses that are harmful to well-being (e.g., helplessness, self-blame, etc.). For this reason, training interventions on virtual teams (
Kniffin et al. 2020) must be strengthened, having in mind their function of containment, identity support, and social support, as well as their effectiveness in achieving the objectives.
Training that focuses more on self-management of work and the establishment of physical, temporal, and psychological boundaries between work and non-work is also essential. A skill that workers should develop concerns the ability to disconnect from work and to separate it from the personal and familial life by modifying the organization and through management of the spaces and times invested in the different activities (
Vayre and Pignault 2014). Studies in the field show that remote work has positive consequences on work–life balance provided that remote workers develop skills in terms of work scheduling and self-management and rigorously structure telework periods (e.g.,
Greer and Payne 2014;
Vayre and Pignault 2014). However, the rules and social norms related to the use of digital technologies and the regulation of these uses also need to be discussed, negotiated, and defined within work organizations, taking into account the workload and health of employees, their right to rest, and their right to disconnect (e.g.,
Vayre and Vonthron 2019).
Finally, psychological counseling tools (even at a distance) can be useful in all cases in which leaders and followers encounter difficulties in managing the consequences of the toxic components of their relationship. Preventing the potentially damaging effects of remote work requires a major transformation of the organizational culture (e.g.,
Greer and Payne 2014;
Vayre 2019). It implies the development of a real company policy concerning this modality of work (i.e., clear, transparent, voluntarist). It also requires the development of managerial policies and practices (e.g., promotion of autonomy at work, establishment of a relationship of trust), the redefinition of processes for evaluating efficiency and performance at work, and the evolution of the representations conveyed about remote workers. This research could contribute with useful information to the promotion of remote workers’ quality of life and well-being, which may be developed in interventions and preventative measures. Telework experimentations, experience feedback from remote workers as well as their managers and colleagues, the creation of spaces for dialogue, and collective decision-making are an effective way to achieve these objectives and improve the existing system.