1
Implementing a Policy of Unlimited Leave: How
Does It Affect Employee Well-being?
Master Thesis for MSc BA Health Faculty of Economics and Business
University of Groningen
Name: Romée Reef Student number: S2517221
Date: June 21st 2020
2
Abstract
As many organizations around the globe aim to employ motivated and talented employees, more and more employers start incorporating measures of flexibility. Such concepts are said to be very attractive to employees, as they could result in employees having more time with their families and work when they feel their best. This study empirically tests the effects of the flexible working arrangement of unlimited leave, which allows employees to take paid time off as they desire, while still fulfilling their job responsibilities. The present research quantitively and qualitatively studied the effects of unlimited leave on employee well-being and performance at a large Dutch bank over the course of four months, researching whether there are differences between a group having unlimited leave and a control group with a regular leave policy. The study performed repeated measures ANOVAs to assess differences between the two groups with a final sample of, on average, 215 employees. I expected the unlimited leave policy to have a positive effect on both well-being (work-life balance, work engagement, and energy) and performance. Furthermore, I expected the employee’s perception of autonomy to have a strengthening effect on the relationship between unlimited leave and well-being and unlimited leave and performance. Unfortunately, none of this study’s hypotheses were supported, which can possibly be explained by the unique measures taken by Dutch government to contain COVID-19. This study did find a decrease in work-life balance, work engagement, and in leave days taken over the four month period for the experimental and the control group, which can likely be attributed to the COVID-19 pandemic. Also, I found a significant difference in work-life balance and energy between employees experiencing high versus low autonomy, though this difference cannot be attributed to unlimited leave but seems to be inherent to job characteristics.
3
Introduction
In our high-paced lives the present day, many individuals view their work to have a negative effect on the rest of life, creating pressure and restricting time they have for themselves, families, friends, and communities (Pocock, Skinner, & Pisaniello, 2010). Indeed, a quarter of women working full-time and one fifth of similar men were dissatisfied with their balance between work and private life according to a survey conducted in 2010, and many workers are working more than they would prefer (Pokock et al., 2010).
As a response, many organizations are trying to manage the well-being of their employees by offering more flexibility. In the last decade, numerous policies have been trialed all over the world, from a 38-hour working week in Sweden to working in the comfort of your own home in the United States (Lister & Harnish, 2019). The number and the variety of flexible working arrangements still continues to increase (Spreitzer, Cameron, & Garrett, 2017). A rather new principle that more and more large corporations have taken an interest in is the concept of “unlimited leave”, meaning that employees are allowed to take paid time off whenever they want, while still fulfilling the
responsibilities of their jobs (People Management, 2014). It has been claimed that giving employees total freedom to choose their leave days could be beneficial for both company and employees (De Jong & Arevalo Östberg, 2015). Among others, Netflix, Kronos, and Virgin have already introduced an unlimited leave approach in the last decade (People Management, 2014).
It would seem from the experiences of these prominent organizations that such a policy is potentially very beneficial (Ain, 2017). Businesses offering unlimited leave are said to be employing motivated and responsible employees (Griswold, 2013). Furthermore, implementing such a concept could result in employees having more time to spend with family and friends, and by doing so they are able to spend more time on their health and well-being (De Jong & Arevalo Östberg, 2015). Having such well-rested and happy employees is of importance to many companies around the globe. Studies have shown a positive relationship between recovery from work and well-being of employees (De Bloom, Kompier, Geurts, De Weerth, Taris, & Sonnentag, 2009). Also, being well-recovered could increase employees’ ability to perform their work (De Menezes & Kelliher, 2011).
self-4 regulation, and because of partly removing the worker/boss culture of authority and subordination (People Management, 2014).
Some hold the opinion that these concerns about unlimited leave can be overcome by a culture where people are judged by the value of the work they deliver, and not by the hours they work (People Management, 2014). Also, having trust in the people that work in the organization is an important determinant that is mentioned (Ain, 2017). There are people in practice who find flexibility a good motivator, as long as there is fairness across the organization (People Management, 2014). Finally, perceptions of autonomy could determine whether such policies can be used optimally (Gerdenitsch, Kubicek, & Korunka, 2015). In short, certain requirements are likely of influence in determining the success of an unlimited leave policy.
In summary, it seems that the effects of unlimited leave policies are still unclear, both in theory and in practice. There are many mixed opinions about the effectiveness of such policies and very few empirical studies on the matter. To the author’s knowledge, the only scientific paper covering the concept of unlimited leave is written by De Jong and Aravalo Östberg (2015), who conducted five interviews concerning the effects of unlimited leave on job satisfaction, productivity and work-life balance in high-tech US companies. But mainly, articles and websites covering the concept of unlimited leave today are based on opinions from leaders and other experts. Published research using data from employees having such a policy in place is lacking. The current study will address this gap by both quantitatively and qualitatively testing the effects of the implementation of unlimited leave on employee well-being (encompassing work-life balance, work engagement, and energy) and
performance at a large Dutch bank. This study also aims to discover whether results differ between employees experiencing high versus low autonomy. Finally, this study will assess whether employees take on a different amounts of leave with the policy in place.
This research will address the following questions:
1. How does the presence of an unlimited leave policy affect both employee well-being (work-life
balance, work engagement, and energy) and performance?
2. How do these effects differ between employees experiencing high versus low autonomy? The present study uses both survey and interview data. Survey data will be used to determine the changes in well-being and performance, and whether this differs between people perceiving to be more or less autonomous. The interview data will provide a more in-depth view as to why results have come to pass and to discover deeper-rooted feelings that associate with an unlimited leave policy. Interview data will thus be used to assess the conditions that may have influenced the results.
5 others, the closing of public places and schools, and the call for working from home as much as possible. The present study has responded to this pandemic and included questions regarding unlimited leave and COVID-19 measures in both surveys and interviews. Therefore, the current research will also assess whether having an unlimited leave policy has a different effect on well-being compared to not having an unlimited leave policy during the measures taken to contain COVID-19.
This study aims to contribute to theory by being among the first to study the effects of unlimited leave. As there already does exist a large body of literature focusing on flexible working arrangements, the present research aims to extend this literature by adding the concept of unlimited leave to theories and other policies regarding flexible working arrangements. Furthermore, the present study will be among the first to scientifically approach the concept of unlimited leave and to
investigate this concept in a statistical manner. As most work currently available serves more as practical guides or experiences from companies, this study extends this body of work by empirically testing the effects of such policies using longitudinal data and a large sample size.
The study is also of practical use. By showing the effects of unlimited leave in a large Dutch organization, many other companies could possibly use the results of the current study to make decisions about well-being of and suitable leave policies for their employees. Furthermore, as I will also show results concerning perceived autonomy, practitioners can gather information whether an unlimited leave policy is workable for their businesses.
The remainder of this paper is organized as follows. Firstly, in the literature review section, work that has been done regarding recovery from work, flexible working arrangements and autonomy will be discussed. This is followed by reviewing some requirements that have been highlighted in literature to be of importance for the success of unlimited leave. Also, within this section, I will highlight the measures regarding COVID-19 and the possible influence of these measures in this study. This is followed by the method section, in which I will describe the data that was used, as well as the analysis strategy. Then, in the results section, the outcomes of the data analyses and hypotheses testing will be presented. I will also address interview data that might explain certain results. Finally, in the discussion chapter, the study’s summary, implications, and limitations will be laid out, as well as directions for future research.
Literature review
Recovery from work and employee well-being and performance6 activities – both at work and at home – could possibly lower well-being, while leisure time in evenings reduces stress and improves being. Most research on recovery from work has focused on well-being indicators, but some studies have looked at employee performance (Sonnentag, Venz, & Casper, 2017). It is made clear in those studies that being recovered is an important concept for employee performance (Binnewies, Sonnentag, & Mojza, 2009).
Next to recovery on evenings and weekends, research has suggested that vacations can result in an increased state of being recovered (Sonnentag et al., 2017). It seems that not taking a holiday is associated with worse work-life interaction (Pocock et al., 2010). Because people often recover insufficiently during evening hours and weekends, vacations could be seen as a longer period to recover from work (De Bloom et al., 2012). This effect is also present when employees take only short vacations of four to five days (De Bloom et al., 2012). A meta-analysis also concludes a positive effect between vacations and well-being (De Bloom et al., 2009). In short, researchers seem to agree that vacations can help individuals recover in a way that cannot be achieved in evenings and weekends.
Flexible working arrangements
In recent years, the concept of flexible working arrangements has received attention to be used as a tool that can help individuals manage work and private responsibilities (Allen, Johnson, Kiburz, & Shockley, 2013). Due to modern technology, working is no longer restricted to standard working hours and places (MacEachen, Polzer, & Clarke, 2008). These possibilities lead to the development of flexible arrangements that could give employees the freedom to work where and when they desire. No clear cut definition of flexible working arrangements seems to be in place in literature, but most research makes a distinction in terms of flexibility in work times and work place. Examples of such arrangements are flextime (setting own working hours), telecommuting (working from other locations), job-sharing, and compressed work weeks (Kalev & Kelly, 2006).
The concept of unlimited leave can be regarded as a form of flexible working arrangements (De Jong & Aravalo Östberg, 2015). It seems that it can best be compared with an arrangement concerning work times, since unlimited leave also gives employees control over their time. The difference between the two is that employees taking leave mostly do not work several days in a row.
Flexible working arrangements and well-being
By providing employees control over where and when they work, flexible working
7 better fit between work and individual needs of employees, enhancing their well-being (Russell, O'Connell, & McGinnity, 2009).
But mixed results stem from research regarding flexible working and well-being (De Menezes & Kelliher, 2017). For instance, Allen and colleagues (2013) demonstrated the inconsistency of research results regarding flexible working arrangements and well-being. This could be explained by notions that not all flexible working arrangements can be compared (Allen et al., 2013; Berkery, Morley, Tiernan, Purtill, & Parry, 2017), but also because flexible working can have negative effects. The use of mobile devices allows employees to work anywhere and anytime, leading to the tendency to be available all the time – so called “always on” attitudes (Spreitzer et al., 2017). Employees having flexible working arrangements in place could experience pressures, such as being required to be “always on”, intensifying their work efforts, possibly lowering well-being (Kelliher & Anderson, 2010). It is mentioned by Allen and colleagues (2013) that indeed mixed results stem from flexible working, and that the relationship between flexible working arrangements and well-being still needs further evaluation.
Flexible working arrangements and performance
Looking at performance, one might say that flexible working arrangements allow employees to work when and where they are most productive and focused (Kalev & Kelly, 2006; Allen et al., 2013). Multiple studies show that adjusting work according to own need can lead to an increase in performance (e.g. Kelliher & Anderson, 2010). Also, social exchange theory has been used to explain changes in employee performance when they are offered flexible working arrangements. According to this theory, it is said that when employers offer flexible working arrangements to their employees, employees feel morally obliged to do something in return, increasing their productivity (Berkery et al., 2017). So, flexibility provides employees with the freedom to choose working hours when they are most productive; and they may want to return the favor of getting this freedom by being productive. Both are likely to lead to an increase in employee performance.
8 arrangements are shown to possibly lead to increased well-being and performance, unlimited leave policies also have the potential to have such effects.
Autonomy
A meta-analysis indicates that the design of work has a profound effect on employees’ behavior, attitudes, and well-being (Humphrey, Nahrgang, & Morgeson, 2007). Over 34% of the variance in performance and over 55% of the variance in job satisfaction was due to design of work (Humphrey et al., 2007). So, the way jobs are designed are said to contribute to different outcomes regarding employee well-being and performance.
Within the job characteristics literature, autonomy is the most studied job characteristic, and therefore it is regarded an important construct to include in the present research (Coelho & Augusto, 2010). Within this study, I argue that individuals experiencing higher autonomy are better able to organize their leave days according to specific needs, leading to increased well-being and
performance. Previous work indicates that autonomy can lead to enriched work experiences (Jiang, Di Milia, Jiang, & Jiang, 2020). It provides individuals with the flexibility and freedom to gather, learn, and process job information (Jiang et al., 2020). The current study defines autonomy as the freedom in three interrelated aspects: work scheduling, decision making, and work methods (Morgeson &
Humphrey, 2006). It would seem that those experiencing high autonomy have more freedom to determine their way of working, possibly leading to higher performance and a higher sense of well-being (Ali, Said, Yunus, Kader, Latif, & Munap, 2014).
Though positive effects of autonomy seem convincing, research has also found negative effects. Sonnentag and Fritz (2007) argue that, when experiencing more autonomy, one will be more inclined to continue thinking about one’s job after work and psychological detachment from work will become more difficult. This is confirmed by other authors. Bennet, Gabriel, Calderwood, Dahling, and Trougakos (2016) found that individuals perceiving more autonomy at work are less likely to detach and more likely to think proactively about work during nonwork time. Another argument could be that autonomous employees could perform less because of less supervision and a lack of measurable targets. For instance, looking at a call center, there are often clear targets regarding conversation duration and handling (Miciak & Desmarais, 2001). On the contrary, jobs with more autonomy often have less clear performance measures, providing an incentive for individuals in these jobs to sub optimize and to cut corners. Finally, it would seem that research covering work time autonomy finds mixed results regarding performance, because being highly autonomous in determining working times could interfere with communication and cooperation among colleagues resulting from employees not being at work during the same time period (Kattenbach, Demerouti, & Nachreiner, 2010).
9
Though this study mainly relies on quantitative hypotheses testing to determine the effects of unlimited leave policies, it is important to note that certain aspects play a role for the successful implementation of such arrangements. This study will use interview data to assess whether those aspects that are said to play a role within scientific literature were also of importance within this research project. Lastly, I will make a number of notions regarding the link between the COVID-19 pandemic and unlimited leave. Interview data will be used to get insight into the effects of the measures that were taken and the role of unlimited leave in this aspect.
Responsibility to managers
With an unlimited leave policy in place, managers get an extra responsibility in ensuring that all employees have the same opportunity to take time off, and in making sure that the office is fully staffed throughout time (Griswold, 2013). According to Griswold (2013), managers need to act as role models to make sure their employees actually take on the leave they need. In the case of unlimited leave, this would mean to take up leave days themselves, so that other employees may follow (De Jong & Aravalo Östberg, 2015). Managers should examine how employees should make use of such policies and also prevent them from abusing it (De Jong & Aravalo Östberg, 2015). In short, the role of a manager could have an impact in how an unlimited leave policy plays out.
Furthermore, it is said that managers are responsible for gathering continued feedback and evaluations about employee experiences with the policy (Younis, 2016). Younis (2016) stresses the importance of evaluation, feedback, and information sharing by managers to ensure the effectiveness of flexible working arrangements for both employer and employee. Lack of such evaluation and information sharing activities can likely impede positive effects of flexible working arrangements (Younis, 2016). Managers thus play an important role in making sure that employees can give input at all times, and that they are up to date about any developments regarding the policy.
Trust
Researchers mention the importance of trust regarding flexible working arrangements (De Jong & Aravalo Östberg, 2015). According to Younis (2016), it is paramount to have a climate of trust in place to be able to successfully utilize flexible working. Within this climate, regular communication and check-ups are required to prevent problems from occurring (Younis, 2016). In the study of De Jong and Aravalo Östberg (2015), trust is mentioned by participants as a feature that was already present within the organization before implementation of the unlimited leave policy. One could say that unlimited leave policies offer trusts to employees to determine their own leave days, but it seems important that a culture of trust is an inherent concept within the organization to aid the feasibility of unlimited leave (De Jong & Aravalo Östberg, 2015).
10 It is argued in literature that the extent to which individuals make use of flexible working arrangements depends largely on what their colleagues do (Lambert, Marler, & Gueutal, 2008). As such, it would seem that processes on team level affect the use of flexible working arrangements (Lambert et al., 2008). Since people are, at least partly, shaped by their environments, norms that are present within teams could possibly determine the effectiveness of flexible working arrangements. As mixed results stem from research regarding flexible working arrangements, it is possible that team norms are an explanation for negative effects of such arrangements (Pedersen & Lewis, 2012).
Regarding the concept of unlimited leave policies, it is likely for these team-level antecedents and norms to have an impact during implementation. One could argue that employees might actually be taking less leave days, because they do not want to be judged by peers about their amounts of leave (Griswold, 2013). On the other hand, employees could start imitating co-workers when some are taking up excessive amounts of leave. In short, it seems that processes on team-level are of influence concerning how an unlimited leave policy plays out.
COVID-19
Due to the COVID-19 pandemic, many measures were taken by governments and
organizations. In the Netherlands, among other measures, the call is made to stay at home as much as possible, to close all schools, and to keep a distance of one and a half meters to others (Rijksoverheid, n.d.). As a result, many organizations urged their workforce to work from home.
Working from home allows employees to better manage their work versus home activities, reduce time commuting to the workplace in busy traffic, and create a work schedule that is adapted to personal productivity (Allen et al., 2013). But there are also downsides mentioned to working from home, such as loss of contact with colleagues, and loss of structure (Halford, 2005; Nijp, Beckers, Van de Voorde, Geurts, & Kompier, 2016)
.
It is said by Nijp and colleagues (2016) that working from home could depend on whether this is voluntary or involuntary.11 This effect of actually working more when working from home is also shown in scientific literature (Baruch & Nicholson, 1997), leading to people working late and possibly decreasing well-being.
In short, it seems likely that the COVID-19 measures and the consequence of involuntarily having to work from home negatively influence well-being. I expect that having unlimited leave may buffer this effect. Compared to those who do not have an unlimited leave policy, I expect individuals with such a policy in place to more easily take leave, to deal with stressful situations from telework while at the same time taking care of children and homeschooling activities, and thus focus on their well-being.
Hypotheses development
Employee well-beingWithin this research, employee well-being will encompass three constructs. Firstly, as
work-life balance comprises elements such as a balance between stressful and relaxation activities and work
and leisure time (Syrek, Bauer-Emmel, Antoni, & Klusemann, 2011), this construct can be regarded a good indication of employee well-being. Furthermore, work engagement is often defined as the opposite of burnout (Schaufeli, Bakker, & Salanova, 2006). Employees with high work engagement have an energetic and effective connection with their work activities, and view themselves able to deal with the demands of their jobs (Schaufeli et al., 2006). Work engagement is thus considered to
represent the positive side of employee well-being (Mäkikangas, Hyvönen, & Feldt, 2017). Lastly,
energy is regarded in literature as a valuable tool in research about well-being, indicating that
opposites on the energy continuum (exhaustion and vigor) represent opposite well-being states (Mäkikangas et al., 2017). This implies that a high amount of energy at work would represent a high sense of well-being.
As said, from what is known about literature regarding flexible working arrangements, it seems that such arrangements could have both positive and negative effects to the well-being of employees. On the one hand, flexibility of working can lead to better balance between family and work (e.g. Jang, 2009; Uglanova & Dettmers, 2018) and can reduce stress (Almer & Kaplan, 2002). But studies also show that flexible working arrangements can lead to actually working more and at a more intense level (Uglanova & Dettmers, 2018). Nonetheless, these studies also highlight that, generally speaking, well-being seems to increase due to flexible working arrangements (Kelliher & Anderson, 2010).
12 is said that such a policy facilitates the fit between work and other life domains (Uglanova &
Dettmers, 2018).
I expect, that, consistent with the general argument in literature concerning flexible working arrangements and first results of unlimited leave studies, the positive effects will outweigh the
negative effects. As previously mentioned, there are minimum requirements that are likely of influence concerning implementation of such policies. Though they are expected to play a role, based on
research about unlimited leave policies and flexible working arrangements so far, I argue it likely for unlimited leave to have a positive effect on employee well-being. So, I hypothesize:
H1: The implementation of an unlimited leave policy has a positive effect on employee well-being (work-life balance, work engagement, and energy).
Employee performance
As said, studies regarding recovery and flexible working arrangements have also looked at performance outcomes (Sonnentag et al., 2017). Research often indicates a positive association between recovery and employee performance (Binnewies et al., 2009), and flexible working hours and employee performance (Eaton, 2003; Peretz, Fried, & Levi, 2018).
Though there are meta-analyses indicating inconclusive evidence between flexible working arrangements and employee performance (e.g. De Menezes & Kelliher, 2011), I use social exchange theory in building my argument regarding employee performance. Following this theory, I argue that an having an unlimited leave policy is seen as generous feature by employees. In turn, employees feel responsible for returning the favor. Furthermore, following the literature about recovery I argue that when employees can take days off according to their own need, they will feel more recovered and thereby performance is enhanced (Binnewies et al., 2009).
This is confirmed by the one empirical article about unlimited leave. De Jong and Aravalo Östberg (2015) find that employees perceive to be more productive due to an unlimited leave policy, which they call the “battery recharging effect”. These findings are thus in line with the general argument stemming from literature about flexible working arrangements and social exchange theory. Following these findings, I hypothesize:
H2: The implementation of an unlimited leave policy has a positive effect on employee performance.
Autonomy
13 otherwise. Also, the evidence concerning the negative aspects of flexible working arrangements and autonomy is pointed more towards work intensification and work-nonwork conflicts (Gerdenitsch et al., 2015). I expect these effects to be marginalized with the arrangement of unlimited leave, since employees are actually not working when making use of this policy.
Empirical evidence seems to support the notion that perceived autonomy is positively associated with being. Royer and Moreau (2016) find, in a study concerning psychological well-being of childhood educators, that 23% of variance in well-well-being is due to perceived autonomy. Crawford, Lepine, and Rich (2010) find that having more job resources – such as autonomy – is negatively associated with burnout. Within literature concerning flexible working arrangements, Gerdenitsch and colleagues (2015) researched flexible working arrangements in relation to autonomy. They demonstrate that whether flexible working arrangements are regarded positively or negatively is associated with the perception of autonomy of employees.
Aside from high-autonomy individuals experiencing higher well-being, I expect that a perception of high autonomy provides individuals with the freedom to take time off whenever they regard it most needed, giving them a better ability to increase their well-being. As is mentioned, using unlimited leave is more difficult with certain types of jobs (Griswold, 2013). For instance, within departments where occupancy rate is an important feature for unit performance, it could be more difficult for employees to take days off according to their specific needs (Miciak & Desmarais, 2001). It is likely for such individuals to perceive lower autonomy and to have a lower ability to take leave. On the contrary, those with more autonomy have more freedom to determine how, when, and where they work, leading to a better ability to take on leave according to their own needs. So, I hypothesize:
H3a: The implementation of an unlimited leave policy has a stronger effect on well-being for individuals perceiving high, compared to low autonomy.
Besides well-being, I expect the effect of an unlimited leave policy on job performance to be stronger for individuals perceiving higher, compared to lower autonomy. Meta-analytic evidence indicates, across 259 studies and 219,625 employees, that autonomy alone is positively related to job performance (Humprey et al., 2007). Whenever individuals experience more autonomy regarding how they can structure their work, they seem to be extra productive when they are working (Eaton, 2003; Peretz, Fried, & Levi, 2018). These positive effects are confirmed by a more recent meta-analysis, indicating that a higher sense of autonomy leads to better job performance (Muecke & Iseke, 2019).
14 specific needs, consequently increasing productivity when they are working (Eaton, 2003; Peretz, Fried, & Levi, 2018). So, I hypothesize:
H3b: The implementation of an unlimited leave policy has a stronger effect on employee performance for individuals perceiving high, compared to low autonomy.
Methodology
Data collectionThe present research was conducted in cooperation with a large Dutch bank, at which an unlimited leave policy was implemented in 2020. The project is led by a team of professors and associate professors from several universities, specialized in different fields of recovery from work stress. The time span of the study is one year, but because of time constraints, the present research will use data from the first four months of the study. A total of 580 participants were equally divided in an experimental group and a control group at the beginning of the study (January 2020). The control group has a policy in which they can take on a fixed amount of paid leave (194.4 hours based on a 36-hour working week), with the possibility of “buying” extra days off. The experimental group has an unlimited amount of paid leave days.
The company in question is organized or transitioning to be organized in an agile manner. The agile way of working, according to the British Computer Society, is a way of working in which an organization empowers its people to work where, when and how they choose – with maximum
flexibility and minimum constraints – to optimize their performance. Within the agile way of working, autonomous teams are ought to set their own goals. Since the company uses such a concept, the organization has a rather flat structure, with departments being very autonomous in determining their own objectives. No specific boundary conditions were set by the research team concerning the unlimited leave policy. It is left to the participating departments to determine how unlimited leave is dealt with.
15 Additionally, the research team has conducted face-to-face semi-structured interviews with a subgroup of around 20 participants belonging to the unlimited leave group several times throughout the project to gain in-depth knowledge about the implementation and the effectiveness of unlimited leave. The interviews that are used in this study took place in the first month of implementation (January 2020) and three months after (April 2020), and were conducted by different members of the research team. The interviews took about an hour each.
Baseline January February March April
Sample characteristics • Age • Gender • Family status • Educational level • Contractual working hours • Contract type • Job tenure Well-being • Work-life balance • Work engagement • Energy • Work-life balance • Work engagement • Energy • Work-life balance • Work engagement • Energy • Work-life balance • Work engagement • Energy
Performance • Subjective job
performance • Subjective job performance • Subjective job performance • Subjective job performance Autonomy • Decision-making autonomy • Work methods autonomy Leave days • Leave days
taken in 2019
• Leave days last month
• Leave days last month
• Leave days last month
Figure 1: Assessment of variables in the questionnaires
Participants
The sample of the present study is not randomized, due to organizational constraints and the goal to create a realistic sample. Managers of the organization were asked whether they would be interested to take part in the study (experimental or control condition), of which 54% agreed to allow their employees to take part. Furthermore, in order to keep the study realistic, it was necessary to select colleagues who work closely together to have the same leave policy. The average response rates were 75% in January, 60% in February, 55% in March, and 53% in April. Characteristics of
16 It can be seen from table 1 that the response rate of the experimental group is higher than that of the control group, which likely has to do with more motivation to participate by the unlimited leave group. Differences between the experimental and control group were related to age and work-life balance. The control group is generally older an experienced a higher work-life balance at the beginning of the study. Also, the control group experienced higher autonomy at the beginning of the study period.
Experimental (n) Control (n) Difference X²/t
Participants (January 1st) 290 290 Average age 41.41 (217) 44.92 (175) t(390) = -3.25*, p = .001 Gender (%) Male Female Non-binary Rather not say
68.70 (149) 30.40 (66) 0.50 (1) 0.50 (1) 76.10 (134) 22.20 (39) 0.00 (0) 1.70 (3) X²(3, 393) = 5.52, p = .137 Family status (%) Single
Living with partner
Living with partner and child(ren) Living with children
Other 16.60 (36) 24.00 (52) 50.70 (110) 4.60 (10) 4.10 (9) 14.20 (25) 29.00 (51) 49.40 (87) 2.80 (5) 4.50 (8) X²(4, 393) = 2.15, p = .708 Educational level (%) No education Primary education
Lower vocational education Middle vocational education High school
Higher vocational education Academic education 0.50 (1) 0.00 (0) 0.90 (2) 11.20 (24) 17.70 (38) 43.70 (94) 26.00 (56) 0.00 (0) 0.00 (0) 1.70 (3) 4.70 (13) 14.20 (25) 53.40 (94) 23.30 (41) X²(5, 391) = 5.64, p = .343
Average contractual working hours 37.02 (215) 36.78 (176) t(324.55) = 0.73, p = .465 Contract type (%) Permanent Temporary 96.30 (207) 3.70 (8) 96.60 (170) 3.40 (6) X²(1, 391) = 0.03, p = .869
Job tenure in years 4.63 (188) 4.47 (161) t(347) = 0.33, p = .742
Work-life balance first measurement (1-5)
3.30 (245) 3.45 (194) t(437) = -2.95*, p = .003
Work engagement first measurement (1-6)
3.86 (240) 3.88 (187) t(425) = -0.21, p = .834
Energy first measurement (1-7) 4.64 (240) 4.81 (187) t(425) = -1.24, p = .214
Performance first measurement (1-10)
7.02 (241) 6.87 (191) t(430) = 0.60, p = .551
Perceived autonomy (1-7) 5.06 (213) 5.32 (163) t(374) = -2.68*, p = .008
Leave days 2019 24.47 (238) 23.91 (188) t(424)= 0.87, p = .387
17 Regarding a survey question about leave days in 2019, responses that were not realistic were deleted. For instance, one respondent reported taking 402 leave days in 2019. Such values were deleted to avoid skewed means, by deleting outliers based on Z-scores > 3.29. In essence, this means that responses from 94 days onward were regarded missing.
Out of the sample of the experimental condition, 20 to 25 participants representing the sample were asked to participate in the interview sessions, with the goal of interviewing them both moments in time. For the most part, the same people were interviewed at the first and second time point, but some participants were different because of dropouts or interviewees leaving the company
Measures
Employee well-being was measured through three constructs: work-life balance, work
engagement, and energy, all of which are measured in each month of the study period. First of all, work-life balance is measured through five statements on a 5-point scale from 1) strongly disagree to 5) strongly agree (Syrek et al., 2011). Examples of statements are: during the last four weeks… …I
was satisfied with the balance between my work and private life; … I managed to achieve a good balance between stressful and relaxing activities in my life. In the fourth month of the study,
work-life balance comprised four questions. Since there was a dropping response rate over the course of the study, it was decided to shorten the questionnaire. The questions are summarized to create one construct concerning work-life balance.
Secondly, work engagement is measured through nine statements which participants have to respond to on a 6-point scale ranging from 1) never to 6) every day (Schaufeli et al., 2006). Examples of statements are: during the last four weeks… … at my job, I felt bursting with energy; … I was proud
of the work that I did. These nine statements were summarized to generate a final score regarding
work engagement. Also concerning work engagement, the fourth survey contained only six statements to shorten the survey.
Third, a question about the energy level experienced as by Weigelt, Wyss, Hoffmann,
Fellmann, and Lambusch (2019) was used, in which participants were asked to rate how much energy they felt during the last four weeks (figure 2). This is measured on a 7-point scale from 1) no energy to 7) full energy.
18
Employee performance was measured according to a subjective measure adapted from
Kessler and colleagues (2003), namely a survey question asking employees about their own perceived performance in the last month. Performance is measured every month on an anchored scale of 1 to 10, where 1 is the worst job performance anyone could have at the job and 10 is the performance of a top worker.
Autonomy was measured using questions of two constructs concerning autonomy at work.
These constructs – decision-making autonomy and work methods autonomy – are both measured with three statements on a 7-point scale, as to Morgeson and Humphrey (2006). Examples of statements are: in general, over the last year, my job … … allowed me to make a lot of decisions on my own (decision-making autonomy); … gave me considerable opportunity for independence and freedom in
how I do the work (work methods autonomy). These two autonomy concepts are measured solely in
the baseline questionnaire. The assumption here is that these autonomy principles do not change rapidly. Factor analysis was carried out to assess whether these two constructs can be summarized into one autonomy measure. As can be seen in Appendix A, factor analysis showed that items of both decision-making autonomy and work methods autonomy load into one component (as the eigenvalue of component 2 is < 1), in which all items have high factor loadings (table 2). This indicates that all autonomy questions can be summarized into one measure of autonomy. Cronbach’s alpha is .93.
Question Factor loading
Decision-making autonomy 1 .902
Decision-making autonomy 2 .873
Decision-making autonomy 3 .869
Work methods autonomy 1 .860
Work methods autonomy 2 .851
Work methods autonomy 3 .850
Table 2: Factor analysis autonomy Additional data
As mentioned, to test basic differences in whether the unlimited leave policy leads to employees taking more or less leave days, I will take along a survey question asking the amount of leave days a participant took the last month. This question is taken along in the months of February, March, and April. Furthermore, a survey question regarding COVID-19 is included in the
questionnaires of March and April, asking participants about how much the COVID-19 pandemic has impacted their working life and their decision to take on leave, on a scale of 1) not at all to 4) to a great extent.
19 Repeated measures ANOVAs were used to test this study’s hypotheses. Since listwise deletion occurs in SPSS when performing repeated measures ANOVAs, the analysis only takes along those participants that filled in the survey regarding a certain construct at each time point, reducing the n compared to the respondents of the construct for each month. The study therefore deals with a sample in the repeated measures of only those cases that are complete (average response rate = 37%). To assess whether this sample is a good representation of the full sample encompassing responses per time point individually, means per study variable at each time point were plotted to be compared with the repeated measures ANOVAs’ plots. As can be seen in Appendix B and C, the graphs and values of the full sample (Appendix B) appear similar to those of the smaller n of the repeated measures
ANOVAs (Appendix C). Regarding energy and performance, the graphs look rather different at first glance, but since energy is measured on a scale of 1 to 7 and performance on a scale of 1 to 10, the complete cases sample and the full sample are, also regarding these constructs, not much different. Since I regard the complete cases an accurate representation of the full sample, this sample containing only complete cases was used for data analysis.
This study also assessed whether the individuals within the sample of complete cases display similar characteristics to the full sample, and regarded them similar. Small differences concern to less people living alone in the experimental group within the complete cases sample (13% vs. 16.6%). Such differences in living situation are also present when comparing the control groups, as less people in the sample of complete cases are living alone (12.1% vs. 14.2%) and more people are living with a partner and children (52.5% vs. 49.4%). Also, within the control group, the complete cases sample contains slightly less women compared to men (20.6% vs. 78.7%) than the full sample (22.2% vs. 76.1%). Lastly, the control group of the complete cases sample has a slightly higher percentage of academically educated individuals (25.5% vs. 23.3%). As the differences mentioned above are very minor, I regard also the characteristics of the full sample and the complete cases sample similar.
Concerning the reporting of the analyses, first of all, this study presents outcomes of several preliminary analyses, such as mean differences over time and intercorrelations, to get a first idea about the development of study variables and how they associate. Furthermore, t-test were performed to assess whether study variables significantly differ between the experimental and control group at specific moments in time. The final number of participants is, on average, 131 in the experimental group and 84 in the control group.
20 1 and 2. This analysis also allows insight into the possible effect of unlimited leave during COVID-19 measures.
Secondly, to test hypothesis 3, repeated measures ANOVAs were performed encompassing a high-autonomy and low-autonomy group within the unlimited leave sample. A median split based on the median of the full sample was used to determine the groups. Since the median in the complete cases sample was slightly higher than the median in the full sample, a division based on the full sample is regarded most representative to determine what is high and low autonomy. The median of the autonomy variable in the full sample was 5. In order to create more equal groups, participants exactly on the median were incorporated in the low-autonomy group, as the low-autonomy group had a smaller n. The final number of complete cases in the experimental sample is, on average, 81 in the high-autonomy group and 50 in the low-autonomy group. Following presentations in tables, effects over time were plotted to ease interpretation of the results.
The results from comparing the multiple moments in time on well-being and performance are combined with the results of the semi-structured interviews to attempt to find out the “why” of the quantitative findings. Since the concept of unlimited leave is a rather nascent field, it is important to assess why people in the experimental group act and feel as they do, and what are key issues in implementation. In that way, the interview data is used to explain the findings of the survey data. Parts from interview transcriptions regarded relevant for explaining the study’s results, including possible COVID-19 consequences, are discussed in the results section, following the quantitative results.
Quantitative results
Preliminary analysesTo get a first look at the changes that took place regarding each of the study’s variables for both groups, means and standard deviations of the experimental and control group over time are presented in table 3. Furthermore, intercorrelations between study variables are reported in table 4. Lastly, data over time is presented for the high-autonomy and low-autonomy group within the experimental group in table 5.
21
January February March April Experiment M (SD) Control M (SD) Experiment M (SD) Control M (SD) Experiment M (SD) Control M (SD) Experiment M (SD) Control M (SD) Work-life balance n exp = 132 n control = 83 3.41 (0.81) 3.55 (0.87) 3.56 (0.74) 3.47 (0.93) 3.33 (0.91) 3.24 (0.85) 3.26 (0.87) 3.43 (0.87) Work engagement n exp = 128 n control = 83 3.91 (0.87) 3.76 (0.96) 3.91 (0.93) 3.64 (0.99) 3.98 (0.88) 3.77 (0.96) 3.64 (1.02) 3.52 (1.00) Energy n exp = 128 n control = 83 4.62 (1.35) 4.63 (1.36) 4.78 (1.29) 4.59 (1.40) 4.92 (1.11) 4.54 (1.26) 4.67 (1.16) 4.58 (1.23) Performance n exp = 129 n control = 83 7.23 (2.31) 6.93 (2.58) 7.11 (2.26) 6.64 (2.50) 7.05 (2.13) 6.87 (2.13) 7.12 (2.06) 7.11 (1.80) Leave days n exp = 137 n control = 89 2.39 (2.78) 2.08 (3.44) 0.99 (2.03) 1.20 (4.79) 1.49 (2.45) 1.01 (2.04)
Table 3: Means and standard deviations of study variables over time; ̄n = 215
Table 4 presenting correlations also shows some interesting results. Firstly, none of the studied constructs are associated with the amount of leave days taken. Furthermore, the amount of leave days at each month do not correlate, indicating no association between well-being or performance
constructs and leave days taken, and also that leave days of another month do not indicate anything for leave days of this month.
Regarding performance, correlations are slightly more peculiar, as a consistent significant association between being constructs and performance cannot be found. It does seem as if well-being and performance constructs correlate more often within the months of March and April, yet it is still an unexpected finding that well-being and performance are not consistently correlated.
22
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.
1. Work-life balance January .942
2. Work-life balance February .590* .930
3. Work-life balance March .238* .423* .914
4. Work-life balance April .317* .351* .617* .892
5. Work engagement January .263* .273* .196* .242* .918
6. Work engagement February .273* .421* .190* .182* .747* .926
7. Work engagement March .235* .328* .360* .260* .734* .787* .924
8. Work engagement April .182* .273* .220* .273* .651* .646* .729* .910
9. Energy January .431* .346* .197* .245* .534* .468* .456* .403* 1 10. Energy February .267* .431* .175* .206* .448* .574* .468* .415* .422* 1 11. Energy March .278* .401* .403* .349* .433* .466* .618* .419* .430* .492* 1 12. Energy April .287* .327* .468* .468* .458* .454* .562* .683* .438* .498* .613* 1 13. Performance January -.064 -.041 -.144* -.119 .131 .131 .104 .099 .208* .119 .035 .042 1 14. Performance February .077 .128 .037 .031 .098 .164* .173* .050 .038 .263* .148* .074 .190* 1 15. Performance March .037 .149* .089 .058 .154* .187* .248* .171* .102 .194* .294* .140* .269* .158* 1 16. Performance April .093 .077 .034 .216* .284* .205* .219* .359* .217* .160* .221* .325* .240* .118 .293* 1 17. Leave February -.052 .115 .065 .045 .113 .102 .116 .081 -.066 -.041 .016 .005 .014 -.073 -.057 -.045 1 18. Leave March .032 .110 .109 .128 .106 .052 .080 .046 .089 .020 .089 .053 .082 .006 .068 .041 .118 1 19. Leave April -.065 -.071 -.058 .079 .077 .067 .025 .018 -.071 .009 .043 .084 .022 -.056 .020 .063 .010 -.020 1
23
January February March April
High autonomy M (SD) Low autonomy M (SD) High autonomy M (SD) Low autonomy M (SD) High autonomy M (SD) Low autonomy M (SD) High autonomy M (SD) Low autonomy M (SD) Work-life balance n high = 81 n low = 51 3.57 (0.72) 3.15 (0.87) 3.66 (0.72) 3.41 (0.75) 3.38 (0.93) 3.86 (0.88) 3.30 (0.87) 3.20 (0.89) Work engagement n high = 80 n low = 48 3.99 (0.86) 3.76 (0.88) 4.00 (0.93) 3.76 (0.93) 4.05 (0.94) 3.87 (0.77) 3.80 (1.03) 3.39 (0.73) Energy n high = 80 n low = 48 4.84 (1.35) 4.25 (1.30) 4.98 (1.22) 4.46 (1.34) 5.03 (1.14) 4.75 (1.06) 4.88 (1.14) 4.33 (1.14) Performance n high = 80 n low = 49 7.44 (2.22) 6.90 (2.44) 7.25 (2.29) 6.88 (2.22) 7.21 (2.03) 6.80 (2.28) 7.10 (2.21) 7.16 (1.82) Leave days n high = 83 n low = 54 2.54 (2.89) 2.15 (2.62) 1.12 (2.33) 0.78 (1.44) 1.37 (1.80) 1.67 (3.21)
Table 5: Means and standard deviations of high-autonomy and low-autonomy groups within the experimental group over time; ̄n = 131
Hypotheses testing
Effects on well-being and performance
Results of repeated measures ANOVAs are shown in table 6 and 7. Furthermore, graphical representations of the development of study variables over time are presented in Appendix C (experimental and control group) and Appendix D (high-autonomy and low-autonomy group within the experimental sample) of this paper.
Over the study period, graphical representations in Appendix C generally show a decline in leave days taken up for both the experimental and the control group. The graph shows a steep decline for both groups between February and March. Consequently, leave days for the control group continue to decrease in April, while the experimental group’s leave days increase. Though the leave amounts of the experimental group rise, while amounts of the control group decrease, t-tests show that the
24 correction show a significant decrease between the months of February and March (p < .01) and February and April (p < .01).
Looking at the graph of work-life balance, results seem inconsistent, with a higher work-life balance score for the control group in the months of January and April, and a higher score for the experimental group in February and March. The pathway of the experimental group starts with an increase in work-life balance between January and February, after which scores decrease in both March and April. The control group’s work-life balance decreases from January until March, with a very steep decline in March, and quite a steep increase again in April. Though pathways of the two groups look rather different, repeated measures indicate no significant reason to assume work-life balance developing differently for the two groups. However, concerning work-life balance, there is also a significant time effect (F(2.43, 517.54) = 5.89, p = .001, η2 = .027), indicating that scores of work-life balance for both groups together significantly changed over time. A post-hoc pairwise comparison showed that this time effect is likely due to decreasing work-life balance scores between the months of February and March (p < .01).
Regarding work engagement, graphs show a different starting point for both groups, which consequently take more or less the same pathway, with a steep decline in the month of April. The control group’s pathway shows a decline in February, an incline in March and another, steeper, decline in April. The experimental group’s score very slightly increases in February and rises further in March, with a steep decline in April. T-tests show scores significantly differing in the month of February, likely because of the decline in the control group’s work engagement scores (t(209) = -1.98,
p = .049). Yet, as the repeated measures show, there is no reason to assume that work engagement
develops differently for the two groups over time. Once more, results do show a significant time effect over the four month period (F(2.71, 566.43) = 12.77, p = .000, η2 = .058). Post-hoc pairwise
comparisons show that the months of January, February, and March do not significantly differ from each other, but all do significantly differ from the month of April (p < .01).
Within the graph of energy, the control group’s pathway remains rather stable over time, while the experimental group’s energy rises until March, with a steep decrease in April. Results show the expected development from January through March, as one would expect an increase in energy in the experimental group due to unlimited leave, while scores of the control group lack behind. T-tests show that energy scores significantly differ within the month of March (t(159.35) = -2.24, p = .027).
However, as the experimental group experiences a steep decrease in April, it brings both groups back at rather the same score. It is likely for this reason that repeated measures do not produce any
significant results for the construct of energy.
25 group’s perceived performance decreases from January to February. However, within the months of March and April, the control group’s subjective performance starts to rise towards a rather equal score compared to the experimental group. Though differences in for example February look large, t-tests find no significant difference between the two groups at either moment in time, likely because performance is measured on a scale of 1 to 10. Repeated measures similarly find no significant group effect, time effect, or group x time effect over the study period, indicating that it is difficult to draw conclusions regarding performance scores.
Because none of the group x time interactions presented in table 6 are significant, there is not enough evidence to support hypothesis 1 and 2. Though well-being constructs do significantly differ between the groups at certain moments in time, there is not enough reason to assume that the unlimited leave group experiences higher levels of well-being over time. Also regarding performance, there is not enough evidence to support the notion that the unlimited leave group performs better over time.
F-value
Group effect Time effect Group x time effect
Work-life balance n experimental = 132 n control = 83 0.11 5.89* 2.59 Work engagement n experimental = 128 n control = 83 2.52 12.77* 0.84 Energy n experimental = 128 n control = 83 1.39 0.67 1.65 Performance n experimental = 129 n control = 83 1.51 0.64 0.47 Leave days n experimental = 137 n control = 89 0.62 9.84* 0.85
Table 6: Results of repeated measures ANOVAs for effects between groups (experimental and control group), across time (work-life balance, work engagement, energy, and performance: January, February, March, April; leave days: February, March, and April), and for group x time interactions; ̄n = 215; * p < .05
The role of autonomy
As mentioned, perceptions of autonomy could determine whether flexible working
26 Solely within the experimental group, plots in Appendix D show a pathway of leave days that is very comparable to the pathway of the experimental group in Appendix C. Both the high-autonomy and low-autonomy group follow more or less the same pathway, with a steep decline in the month of March, and a slight incline again in April. Repeated measures show a significant time effect of leave days (F(2, 270) = 11.13, p = .000, η2 = .076) for the experimental group (table 7). Within this group, the drop of leave days is significant between the months of February and March (p < .01) and February and April (p < .05).
Within the graph of work-life balance, both groups experience rising scores for work-life balance in February, and a decrease in March and April. The high-autonomy group seems to already experience more work-life balance at the beginning of the study, which is also statistically confirmed by t-tests (t(130) = -3.04, p = .003). However, since the plotted line of work-life balance of the high-autonomy group moves towards the low-high-autonomy group in the following months as its pathway declines more steeply, repeated measures do not find a significant group x time effect. Yet, a significant group effect is found (F(1, 130) = 4.29, p = .040, η2 = .032), indicating that work-life balance differs between individuals experiencing high versus low autonomy. Furthermore, a
significant time effect indicates that work-life balance for the experimental sample changes over the course of the study (F(2.54, 329.79) = 4.45, p = .007, η2 = .033). Post-hoc pairwise comparisons show that work-life balance significantly drops between the months of February and April (p < .01).
Graphs of work engagement in Appendix D show very similar patterns when divided according to autonomy compared to that of the experimental group in Appendix C. Both groups remain rather stable between January and February, with a slight increase in March and a steep decline in April. Consequently, repeated measures find a significant time effect for the experimental sample (F(2.57, 323.46) = 11.30, p = .000, η2 = .082). T-tests do show a significant difference between the high-autonomy and low-autonomy group in the month of April (t(126) = -2.21, p = .029), likely because in this month the work engagement score of the low-autonomy group dropped more steeply than the high-autonomy group. Post-hoc pairwise comparisons indicate that work engagement significantly drops within the month of April compared to the first three months (p < .01).
Regarding the energy plot, it seems that the pathways of the high-autonomy and low-autonomy group are not much different. Both groups’ scores increase until March, and decrease in April. Yet the high-autonomy group consequently experiences higher energy levels than the low-autonomy group. This group effect is confirmed by the repeated measures (F(1, 126) = 9.04, p = .003,
η2 = .067). Also, as previously shown, t-tests show a significant difference in energy scores for three
27 concerning energy (F(2.66, 335.61) = 2.90, p = .041, η2 = .023), where post-hoc pairwise comparisons show a significant drop between the months of March and April (p < .05).
Scores concerning performance do not show a clear pattern. Performance scores of the high-autonomy group slightly decrease over the four month period, while the low-high-autonomy group experiences slight decreases until March, but a steep increase in April. Within this sample also, no significant effect is found by repeated measures regarding performance.
F-value
Group effect Time effect Group x time effect
Work-life balance n high autonomy = 81 n low autonomy = 51 4.29* 4.45* 1.66 Work engagement n high autonomy = 80 n low autonomy = 48 3.23 11.30* 1.07 Energy n high autonomy = 80 n low autonomy = 48 9.04* 2.90* 0.62 Performance n high autonomy = 80 n low autonomy = 49 1.64 0.16 0.53 Leave days n high autonomy = 83 n low autonomy = 54 0.56 11.13* 0.83
Table 7: Results of repeated measures ANOVAs for effects between groups (high-autonomy and low-autonomy group within the experimental group), across time (work-life balance, work engagement, energy, and performance: January, February, March, April; leave days: February, March, and April), and for group x time interactions; ̄n = 131; * p < .05
To assess whether the group effects found are also present in the control group, I also performed repeated measures ANOVAs with solely the control group. Consistent with the results of the
experimental group, group effects concerning work-life balance and energy are also found within the control group.
28 hypothesis 3b is not supported, since no significant effects were found regarding performance
measures.
The role of COVID-19
As mentioned, COVID-19 consequences are likely to have an impact on the study’s results. Participants report COVID-19 to affect working life in a high manner, with an average score or 3.71 in March and 3.58 in April (scale 1-4). The way that COVID-19 has an impact on the decision to take on leave is rated 2.32 in March and 2.68 in April (scale 1-4). These numbers indicate that the presence of COVID-19 has not really had an impact on people’s decision to take on leave.
Qualitative results
This section will present the most remarkable parts of the interview sessions. It was divided in a pre-COVID-19 and post-COVID-19 part, since 1) the second-round interviews themselves were organized in such a way and 2) it became clear that the role of unlimited leave changed after COVID-19 measures.
Before the unlimited leave policy started, almost all interviewed participants’ first reactions were positive, and they were excited for the organization to try it. Following these reactions, most employees started thinking about practicalities, and about what such a policy would do with the amount of leave they would take. Some participants had hesitant feelings about the policy, as participant 1 indicates: “There was, was a group saying, oh, this is bad, I'll take less, there was a
group saying, well, this is good, I'll take more. And there was a group that just got obsessed with technicalities, such as how does it work, operational, who approves, ehm.” Mostly all participants do
indicate that unlimited leave does not mean that they can just take leave whenever they would like to. They have more freedom concerning the amount of leave they take, but still they still have to deliver and make deadlines.
Unlimited leave and well-being
Though I find no support for unlimited leave having a positive effect on well-being,
participants do expect the unlimited leave policy to give them more space to spend time on activities enhancing their well-being. For instance, participant 11 mentions well-being as a goal of unlimited leave: “That we are happier, a better balance between work and private. Ehm, yeah… Becoming
happier eventually, yeah..” Also, it is said that solely the fact of having an unlimited leave policy
leads to an increased sense of well-being, because knowing to have it gives people a sense of freedom, as is indicated by participant 2:“But going away for the weekend sometimes, knowing that I cóúld do
29
already gives me a better feeling. I think it will be really favorable for work happiness.” Also within
the second round of interviews, participants highlight the aspect of not having to calculate about leave days, as mentioned by participant 5: “You don’t have to calculate anymore: what do I still have, and
what do I still expect?”
Unlimited leave and performance
Even though quantitative results do not provide support for individuals having unlimited leave performing better, many participants do expect their organization to experiment with unlimited leave to increase organizational performance. This is said by participant 15 for instance: “A happy and
healthy employee will become more valuable in the end and less downtime.” Participant 11 brings the
role of social exchange theory forward: “Yeah I think you also get a lot back from that, that when your
employer gives you that space, that that will generate a lot. That I will then say sometimes: I will actually come back on Sunday again sometimes, because I feel good enough.”
Unlimited leave and autonomy
This study does not find support for the hypothesis that high-autonomy individuals benefit more from an unlimited leave policy. However, many participants mention that the concept of autonomy can hinder peoples’ ability to take leave. For instance, as participant 2 indicates: “It
becomes unfair when you are limited within your function to take on days because there is no other way, so I can imagine that within the office net. There, you have a certain need of, how do you call it, occupation.”
However, it seems that even in highly autonomous jobs, capacity problems could lead to hesitations of taking on leave, as for example participants 3 and 10 indicate: “We are pretty
understaffed for the amount of work we have and that means that it is always difficult to take on leave. Ehm, and that actually always also means and I can imagine that that is very much below the surface, that you always ask your colleagues for a favor to be able to go on a holiday. And that is of course not a very pleasant thing.” “I feel burdened to take on leave because I know that if I do it, X [colleague] has to work twice as hard.”
Unlimited leave and team-level processes
The role of team-level processes is frequently discussed in literature and is highlighted earlier in this paper as a possible determinant for the success of unlimited leave. Because of responsibility to managers and teams, some struggles regarding how to use and consult about unlimited leave come to the forth during implementation. For example, a manager indicates: “You ask quite a lot from your
30
difficult.” It is interesting to see that very different experiences stem from different participants, since
participant 15 indicates: “How I look at it, it flew out very natural;” and: “The team deals with that in
a very mature way.” Participant 18 actually mentions having noticed nothing since unlimited leave
was implemented: “I haven’t seen changes in people’s behavior, I haven’t seen changes in taking
leave, I actually haven’t witnessed much concerning this, that really something changed.” It would
thus seem that different teams dealt with consulting about leave periods in different ways, from formally sitting together to just individuals planning their own leave. Participant 3 mentions: “I know
from one team that they did that. That they sat with each other in advance like: how are we going to do this with each other? What exactly is the impact of this? So I know that one team did that. I also know that there is a team that deals with this very harmoniously. There is actually never conflict, so they solve all this very well. And I also know that there are teams that are really fighting, that you need a lot of consulting and structure to lead all this in the right direction.”
Unlimited leave and COVID-19
The measures taken concerning the spread of COVID-19 have had an impact on participants of the study and the role of unlimited leave. Nearly the whole organization works from home, and participants indicate that unlimited leave plays either no role, or a very minor role during these times. This seems to be the case for two reasons. First of all, many departments within the bank have an enormous amount of work coming in, because customers are concerned about for instance their mortgages or stocks, leading to quite some pressure to finish all the work. It even led to a manager announcing a leave stop in one department. Secondly, since measures also encompass closing of public places and the call to practice social distancing, many participants actually really want to work, because there is not much else they can do. This is also seen in the quantitative results, as the amount of leave taken up is a lot lower in the months of March and April.
COVID-19 and well-being
As quantitative results already show, COVID-19 measures have had a large impact on participants’ working life. Regarding well-being, it can be seen that work-life balance and work engagement significantly decrease either in March or April. These results are supported by interview participants, who indicate that most people are not too happy or even struggling with the situation of working at home. Such downsides from the situation are mentioned by participant 3: “Some find it
