Unmasking the other face of flexible working practices: a systematic literature review.

Flexible working practice (FWP) has been acclaimed as the practice of the 21st century, and it is likely to continue as humans engage with new technologies in the world of work. Studies have signaled caution in the adoption of FWP, drawing attention to its many downsides. In this paper, we undertake an approximately 11-year systematic review (2011 – mid-2021) of scholarship that has examined FWPs. We focus on the downsides and unintended consequences regarding this concept. Following the rapid growth in remote working arrangements in the digital era, we map out the intellectual structure of these studies and uncover the hidden costs of FWP as well as its implication for employees, managers of technological platforms, organizations, and society. We identify the gaps in understanding these pitfalls and propose a holistic approach with health, legal, and spatial dimensions while indicating directions for future research.


Introduction
Flexibility and mobility have become attractive aspects of working models, having evolved over the past decades in response to the changing needs of the modern workforce (Rubery et al., 2016) and advancement in digital technologies (Valenduc & Vendramin, 2016). Flexible working practices (FWPs) denote working without rigid boundaries around working spaces, schedules, and contracts (Cooper & Baird, 2015;Groen et al., 2018). The digital era has amplified such flexible working arrangements (Hirsch-Kreinsen, 2016), and the media has made FWPs attractive to the modern workforce (Yu et al., 2019). Early approaches to FWP have included telecommuting, part-time work, flexitime, telework, and freelancing, but these have since become contemporary practice . This is reflected in the resurgence of research in FWPs over the last two decades, which coincides with global economic changes that have led to greater intersectoral dependencies (Uddin et al., 2021). These changes have produced uncertainty in the operating conditions of businesses, leading to new forms of organizational structuring, including new modes of working.
These trends in FWP have attracted wide attention since the COVID-19 pandemic has precipitated a need to work away from the office, with the boundary between work and personal times becoming blurred, thus raising questions as to exactly what flexibility in these work patterns means. The literature has examined the various elements of FWP and highlighted the benefits. However, the benefits of "flexibility," a contested concept, mask several pitfalls (Rubery et al., 2016). Unfortunately, there is a dearth of studies that provide an organized structure of the pitfalls in the literature to highlight salient issues and inform the direction of research in response to the potential dangers inherent in adopting FWP in contemporary organizations. Studies have acknowledged social isolation (Mulki & Jaramillo, 2011), tensions with work--life balance (Como et al., 2020), and other negative effects on team working (van der Lippe & Lippenyi, 2020). However, this limited focus portends future challenges for the sustainability of FWP in an increasingly digital world of work. The global digital divide lends further credence to the need for understanding the pitfalls associated with FWP as organizations grapple with the consequences of the COVID-19 pandemic and shifts to flexible working (Chen & Wellman, 2004).
The amplification of FWP benefits overlooks its pitfalls and longterm effects along several dimensions. This study critically examines the hidden dangers of FWP as individuals engage digital technologies for various remote working arrangements. FWP contrasts with conventional working models based on fixed contracts and working hours within rigid and regulated labor markets (Choi, 2018;Rubery et al., 2016). Terms associated with FWP include "co-working spaces," "on-call employment," "on-demand work," "self-employment," "telework," "remote work," "mobile work," "telecommuting," and "virtual work," (Groen et al., 2018;Marica, 2019;Tudy, 2021). These connotations are often used interchangeably in the literature to argue for benefits such as promoting collaboration and innovation among businesses (Spinuzzi, 2012), energy cost and office rental reductions (Richardson & McKenna, 2014), reducing turnover intentions while increasing job satisfaction (Bentley et al., 2016), balancing work-life requirements (Golden et al., 2012), helping women with caregiving needs (Carlson et al., 2010), and reducing traffic congestion (Vallicelli, 2018).
However, the adoption of flexible working (FW) models raises significant implications for organizational structures, digital technology platforms, workforce well-being, and physical workplace designs (Bentley et al., 2016;Johnson et al., 2020). The COVID-19 pandemic has also revealed new concerns, thus highlighting the hidden problems associated with FW models (Furmańczyk & Kaźmierczyk, 2020). These problems include worker monitoring and control measures, which have raised regulatory concerns (Choi, 2018). The call for regulation implicitly signals the existence of problems with FWP that affect the rights of employees and employers (Pedersen, 2017). In this paper, we investigate the pitfalls of FWP based on a systematic review of the literature over a period of approximately 11 years (from 2011 to June 2021). Given various approaches to literature reviews, ranging from systematic to integrative and quantitative to qualitative, as well as diverse purposes, which cover areas such as presenting the state of knowledge through historical overviews, eliciting themes, informing policy and practice, and theory building (Snyder, 2020), we adopt a qualitative approach to show the state of the field of research on the pitfalls of FWPs. We chose this approach because it allows for an exploratory understanding of the issues raised and elicit key themes while establishing the basis for future research agendas (Tranfield et al., 2003). Our review is driven by the following research question: What pitfalls are associated with FWP, and how have these pitfalls been presented in the extant literature over the previous decade? To begin with, we examine the different ways in which the downsides of FWP have been conceptualized in the literature. Subsequently, we present our methodological approach while offering relevant justifications. Following this, we present our findings and discuss implications for theory, practice, and policy to provide a contribution-focused review (Kraus et al., 2020). Finally, we conclude the paper by reflecting on the arguments raised and offer future research directions.

Flexible working practice and its pitfalls
FWP has been studied empirically (see Ravalet & Rérat, 2019;Zarei et al., 2021) as well as conceptually through arguments that examine the future of work (Vealey, 2016;Waples & Brock Baskin, 2021) and through personal narratives (Boncori, 2020;Obenauf, 2021). Although these studies have considered the positive and negative dimensions of FWP, a growing number of studies pay closer attention to the downsides as the adoption of FWP continues to take center stage in the world of work. The negative impact of FWP is shown across a range of areas such as work relationships, home-work conflicts, and the health, organizational, and economic domains. Although these issues are interconnected in the literature, the arguments can be found at the level of individuals, groups/organizations, and digital technologies.

Individual-level pitfalls
At the individual level, the arguments against FWP have mainly focused on the gendered nature of its impact (Drew & Humbert, 2012;Boncori, 2020), poor work-life balance (Como et al., 2020), health problems (Cech & O'Connor, 2017), and the fragmentation of work relationships (Hafermalz & Riemer, 2021;Soga et al., 2021). For instance, FWP has been shown to exacerbate work-life conflicts, resulting in damages to family structures (Bellmann & Hübler, 2021;O'Connor & Cech, 2018) through a blurring of boundaries and extended demands on worker time. The effects of FW transcend the domestic front and are visible in the negative impact on the career progression of women who struggle to fit into an "ideal worker" profile, consequently losing out on their career development (Cech & Blair-Loy, 2014;Jacobs & Padavic, 2015;O'Connor & Cech, 2018).
The other significant disadvantage for individuals relates to health problems (Lockwood & Nath, 2021;Müller et al., 2018), including stress, mental health impairment, and burnout (Peasley et al., 2020). These issues are the result of overwork, exhaustion, and other workload pressures as individuals remain switched on to their digital technology platforms for work (Turkle, 2008;Cech & O'Connor, 2017). Although these FWP pitfalls seem to be universally recognizable, we highlight some contextual variations as captured in the literature (see Table 1).

Group/organization-level pitfalls
The pitfalls of FWP have also been studied at group and organization levels, where scholars have suggested that home working hampers visibility and social interaction, which can affect trust in working teams (Allen et al., 2015;Hafermalz & Riemer, 2021). There are also unintended consequences of using digital technologies and platforms to support home working, such as exclusion and perceptions of surveillance (Soga et al., 2020), issues around communication and connectivity (Chadee et al., 2021), and the erosion of cohesion in organizations (Bentley et al., 2016). Further, there are negative effects on work commitment in teams with corresponding adverse effects on job satisfaction as businesses deploy FW arrangements (Jacobs & Padavic, 2015;Zarei et al., 2021). These are expressed in relation to an increased job insecurity (Kolasa et al., 2021), high attrition rates as a result of withdrawal behaviors (Stirpe & Zárraga-Oberty, 2017), and the nonengagement of employees (Golden et al., 2012) with resultant effects on business turnover and career progression (Gascoigne & Kelliher, 2018). Other issues relate to the quality of leadership and the effects on employee remuneration and career progression, such as a lack of promotions and unpaid overtime, all resulting in an increase in employee turnover as employees are forced to find other sources of income (Nohe & Sonntag, 2014;Junior et al., 2020).

Digital technology-related pitfalls
Furthermore, at the organizational level, setting up the requisite technological infrastructure to support teleworking can be capital intensive (Donnelly & Proctor-Thomson, 2015) despite offering competitive capabilities (Elia et al., 2021). For instance, in the absence of direct feedback in face-to-face situations, technological systems are necessary for reviewing and monitoring virtual teams, sharing information, and fostering work groups. These are additional burdens for managers (Alward & Phelps, 2019) and may result in depersonalized approaches to managing work and technology-induced fatigue (Suh & Lee, 2017). Other challenges include a low Internet speed, poor or inadequate equipment, technological reliability problems, a lack of training or awareness, and disinterest and doubts regarding the usefulness of technological platforms for FWP (Chadee et al., 2021). In addition, digital platforms potentially create concerns for employees and employers regarding data security, cybercrimes, and privacy (Balashunmugaraja & Ganeshbabu, 2020;.

FWP pitfalls and interconnections across levels of impact
The literature has shown that the issues concerning FWP are not restricted to discrete areas of impact but span different levels of impact. For instance, the adoption of technological systems creates not only opportunities but also challenges for managing workers as well as for group relationships in remote working. Recent studies have emphasized a debilitating effect of FWP on the home as it becomes a multipurpose  site for domestic activities, professional work, supporting children's online learning, and leisure, among others (Boncori, 2020;Crawford, 2020;Soroui, 2021). Overlapping functions in the same space have thus imposed demands on the home to respond to the requirements of different needs. This inflexibility of the home space has caused some families to seek different housing solutions (Ravalet & Rérat, 2019). The aforementioned research reflects the various problems associated with FWP; however, there is a lack of an organized framework that maps out the intellectual structure of these pitfalls. In the present study, we seek to address this gap through a systematic analysis of the literature over the past decade. Such an analysis is necessary owing to the ever-increasing adoption of remote working arrangements in the digital era (Yu et al., 2019).

Methodology
To achieve our objectives, we chose two databases for our investigation: Scopus and Web of Science. These databases are considered the most extensive sources of academic articles found in the different fields of the social sciences (Chadegani Arezoo et al., 2013) and are frequently used for reviewing academic scholarship (Gomezelj, 2016;Vieira & Ferreira Gomes, 2009).

Search procedure strategies
We began an initial exploratory search by setting keywords that allude to the downsides of FWP, which is the focus of the study, such as "flexible work* and danger*/challeng*/disadvantag*/problem*/issu*/ downsides" (Snyder, 2019). We reviewed the articles returned (n = 299) from both databases to elicit other synonymous keywords that directly refer to the objective of our study (Khlystova et al., 2022), such as "remote working," "teleworking," "telecommuting," "working from home," "mobile working," and "distant working." To reduce bias and human error, we made use of data extraction forms that provide a visual representation and historical record of the methodological choices made while also being a repository from which our analysis emerged (Tranfield et al., 2003). To ensure that no relevant papers were excluded, we deployed key-strings with Boolean operators and documented each step of the process (see Appendix B for a full reference of our search procedure) to help us select a comprehensive and relevant sample for the review, accounting for duplicates (Tranfield et al., 2003).

Inclusion and exclusion criteria
In our search, we adopted several criteria such as the relevance to our study objective, subject area, language, and time period (Tranfield et al., 2003). An initial cursory reading of abstracts was used to identify peerreviewed articles that were directly related to the objective of the study. We narrowed down the search criteria by refining results for subject areas to business and management, followed by the social sciences, and then the arts and humanities, which were our domains of research inquiry (Appendix B). Second, we limited our search to articles written in English alone to conduct the analysis within our own language of competence. Third, we defined our search to cover the period spanning from 2011 to June 2021. Several considerations underpinned the starting point of our time range selection, constituting our rationale for excluding papers before 2011. FWP began to gain the attention of organizations at the turn of the decade (the 2010 s) as a result of various legal provisions that created conditions for its adoption (Cooper & Baird, 2015). Some examples include the 2009 Fair Work Act in Australia, the 2010 Telework Enhancement Act and the 2013 Flexibility for Working Families Act in the US, and the 2014 Children and Family Act in the UK. These regulations reflect the changing perspectives regarding FWP and the need to introduce supportive structures for employees while laying the groundwork for the emergence of FW models in the subsequent years. Thus, beginning our search from 2011 gave us the opportunity to examine the negative effects of FWP in studies that were conducted at least two years after the initial legislation in 2009. Fourth, to capture only peer-reviewed articles, we excluded document types such as conference proceedings, book reviews, chapters, articles without author names, and other non-peer-reviewed materials. In addition, we excluded all in-press articles owing to access-related restrictions. Subsequently, we accounted for duplicate articles (n = 64) from the two databases used. We found a steady increase in the number of published articles that discussed the downsides of FWP, as shown in Fig. 1. These publications occur in journals that focus on technology (e.g., New Tech., Work & Feeling "tied with the computer to a greater extent than at workplace," "difficult seeing colleagues or other people," "some tasks cannot be done from home," "less interesting or not enough work," "missing access to needed data or documents," "bad physical conditions," "missing physical equipment," and "difficult to keep focus when alone" Overall, we ran 42 searches in SCOPUS (see Appendix B) and acquired a total of 227 refined articles that met the requirements. To obtain only articles that were relevant to our research objective, we grouped these articles into three categories based on our reading of all the abstracts: 1) focused articles (i.e., those directly related to the topic of the study); 2) peripheral articles (i.e., those indirectly related to the topic); and 3) unrelated articles. This grouping was used to extract the total number of focused articles, reducing it from 227 to 34. Subsequently, we deployed the same strategy for our search in the Web of Science database. Following the same procedure, we ran the 42 searches. From the 468 articles retrieved and after adjusting for 64 duplicates, we obtained a total of 79 focused articles. This brought the total number of articles from the two databases to 113. Thereafter, we examined each article closely (n = 113) to extract organizing categories, namely, the research topic, research questions, gap and aims, approach (conceptual/theoretical/empirical), data collection and analysis methods, findings, location (context) of study, articulation of downsides, issues/challenges mentioned by the author(s), and overall trend of arguments presented by all the articles. This allowed us to understand the status quo of the research in the field  and synthesize our results to make a contribution to the literature (Kraus et al., 2020). We provide an overview of the data collection process and strategies adopted for selecting and reviewing articles (see Table 1 in Appendix A).

Findings
Following Tranfield et al. (2003), we detail the findings using several categorizations that emerged from our study to provide a "'descriptive analysis' of the field" (Tranfield et al., 2003, p.218) under each section. These include the role of digital technologies in FWP downsides, the types of FWP and their corresponding downsides, the analytic structure for the pitfalls in FWP, and the significant trends in a timeline, identifying the focus of researchers over the epoch of the present study. Thus, we deploy a performance analysis that is descriptive in nature (Donthu, Reinartz, et al., 2021) and examine scholarship within a field while profiling the contribution of the various research constituents . To begin with, the distribution of the selected articles over the period under study (see Fig. 1) shows a marked rise in research on the downsides of FWP toward the end of the period (2020-2021). This increasing interest might be explained by the COVID-19 pandemic, which necessitated social distancing measures and the adoption of digital platforms for remote working. The prolonged period of social distancing provides sufficient time for the emergence of the medium-to long-term effects of FW. Therefore, the data during this period reflect consonance with the widespread adoption of FWP underpinned by digital technologies.

The role of digital technologies in FWP downsides
Our sample reveals the extent to which digital technologies are mentioned in the downsides of FWP. Multiple studies have mentioned the detrimental health impact on remote workers because of the protracted use of digital platforms (Johnson et al., 2020;Matli, 2020). These platforms include those used for communication with colleagues such as Microsoft's Yammer software, emails, and mobile phone apps, which keep employees always connected. With the advent of artificial intelligence (AI) technologies, where customers use self-service or are served by robots (Mariani & Borghi, 2021), there are often employees on standby to deal with customer frustrations when such systems stop working (Carillo et al., 2021;. In addition, employees have expressed fears about losing their jobs as a result of the adoption of smart technology, robotics, and AI, although AI accounts for only 5% of technologies mentioned in our dataset (see Fig. 2). The technologies that have been studied are primarily platforms that are used for collaborative work, as Fig. 2 shows, with emails (19%) constituting the majority. Similarly, 19% of the technological platforms mentioned by the authors are "mobile phone apps," as used in broad terms, although other technologies could form part of these mobile applications.
According to Ter Hoeven et al. (2016), email and other communication media such as videoconferencing and instant messaging are associated with stress, particularly for jobs that require periods of focused attention. McDowall and Kinman (2017) identified the workload associated with email traffic and digital housekeeping, that is, fixing or resolving technical issues (Suh & Lee, 2017), which create challenges for employees in terms of managing time commitments. In addition, the rapid adoption of technological platforms for FWP involves L.R. Soga et al. sharp learning curves for employees who must quickly adapt to these new technologies, also contributing to stress. These have implications for managers who lead remote teams as knowledge gaps may hamper collaborative efforts (Mihhailova et al., 2011). Safety issues also affect employee well-being and productivity. Bokovnya et al. (2020) highlighted cyber-attacks (including "Zoombombing") and hacking as problems associated with the use of digital communication platforms. These challenges are also mentioned for platforms such as Google classrooms, Moodle, Chamilo, and Blackboard, which are used in flexible learning environments (Rahmadi, 2021).

Number of relevant articles per year
Although a broad range of technological platforms is discussed in our sampled papers, we observe a dearth of reference to technological acceptance models and theories (e.g., TAM) and its variants. These studies have focused on the downsides of FWP and their associated technologies rather than on the adoption of the technologies in themselves. In Table 2, we summarize the key theories used by the authors in our dataset and the purpose for which said theories were deployed. These primarily include organizational behavior and management theories as well as some psychological theories. This leaves scope for contributing to our understanding of the pitfalls with respect to TAM and its variants.

Types of FWP
We found a lack of uniformity among scholars regarding the meaning of FWP, with various connotations used interchangeably (Capdevila, 2013;Jacobs & Padavic, 2015;Tudy, 2021). These include telecommuting (Allen et al., 2015), remote work (Junior et al., 2020), coworking spaces (van der Lippe & Lippenyi, 2020), and on-call work (Marica, 2019). Using a thematic analytical strategy, as also suggested by Tranfield et al. (2003), we organized these various connotations into four clusters-remote work, spatiotemporal work, on-demand work, selfdirected work-which also offer a structure to the domains of the pitfalls observed in the literature. In Fig. 3, we present a summary of these four clusters.

Remote work
The first group includes concepts such as teleworking, telecommuting, remote work, virtual work, part-time job, work from home, FW arrangements, flexible hours/schedules, flex time, flex leave, flex career, and flex place. These concepts relate to work performed away from traditional office spaces and time-related work, with an emphasis on flexibility (Jacobs & Padavic, 2015;Waples & Brock Baskin, 2021). We found that digital platforms such as Yammer, Microsoft Teams, Slack, Zoom, Moodle, and email that allow employees to communicate either synchronously or asynchronously outside the office space are primarily used for remote work. These technologies are also among those that have organization-level subscriptions to allow for unlimited usage.
One key characteristic associated with remote work is vulnerability to family life interruptions as undefined work time and schedules create personal conflicts for employees (Como et al., 2020). In addition, remote work contributes to isolation, loneliness, a lack of engagement, and low commitment because of prolonged periods of losing social contact (Fuller & Hirsh, 2019;Mulki & Jaramillo, 2011). Moreover, the research has suggested that flexible work schedules might leave employees vulnerable to irregularities in income, particularly for flex-time contracts (Marica, 2019), and, therefore, impact their ability to make future financial plans (Rafnsdóttir & Heijstra, 2013).

Spatiotemporal work
This type of FWP considers work in relation to space and time, with a focus on a "sharing culture" (Yu et al., 2019). Concepts such as coworking spaces, digital working hubs, job sharing, and office clubs give a different understanding of flexible work as requiring a sharing of spaces. Through this work model, clusters of workers equipped with technological tools emerge in response to the need for collaboration. Digital technology is a significant component of these working arrangements and is positioned to foster creativity and innovation (Eisenberg & Krishnan, 2018). We found that digital technologies that are mainly used for scheduling and autonomous processes are critical to spatiotemporal forms of FWP. Several examples include GoToMeeting, L.R. Soga et al. some AI technologies (although the studies do not mention the specific kinds), Basecamp, and Web-based platforms such as WebEx. These technologies offer functionalities that allow individuals to organize themselves within limited space resources or to synchronize schedules for collaborative purposes in shared spaces. In addition, they enable communication among individuals or groups working on projects. These technologies are often provided by the organization but used freely by the workers. However, such technologies also allow the organization to monitor worker activity, calendar availability, and resource usage that sometimes give users the feeling of being watched.

Digital technology platforms mentioned by authors
The digital technologies that are used for spatiotemporal work are often those touted to offer high degrees of transparency or openness, which also underpins the aforementioned concept of a sharing culture. However, this use of these digital technologies and co-working spaces does not preclude individuals from experiencing increasing levels of stress that such openness generates (Ingusci et al., 2021). Furthermore, sharing cultures (e.g., job shares) can create conditions for conflicts when it comes to accounting for individual work contributions (van der Lippe & Lippenyi, 2020). Similarly, employees do not always feel comfortable working in different temporal spaces as humans are creatures of habit (Bellesia et al., 2019). Clusters formed from these working models also risk developing groupthink as individuals work collaboratively over short periods (Furtmueller et al., 2011). In addition, employers are faced with financial implications in developing technological systems for enabling, managing, and monitoring these clusters (Donnelly & Proctor-Thomson, 2015).

On-demand work
On-demand work is captured as "on-call," "on-demand," and "zerohours" work, referring to the allocation of focused tasks for a specific time. This contrasts with spatiotemporal work where the concept of "ondemand spaces" refers to the flexibility inherent in space usage and no individual is uniquely associated with a specific workspace (Elfering et al., 2020;Yu et al., 2019). On-demand work typifies the unpredictability inherent in work terms where workers can be called at any time, without requiring any fixed terms of engagement with the organization (Marica, 2019). This is analogous to freelance workers who work as delivery agents or drivers in response to emergent orders of customers. Our data indicate that digital technologies that facilitate on-demand work are primarily used for determining worker availability, negotiating with potential workers, or sending real-time notifications. Examples include WeChat, some AI applications, and various technological applications that are referred to collectively by the authors as "mobile phone apps." These digital applications often allow workers to be visible to potential employers who may need them at short notices. In addition, such applications facilitate synchronous and asynchronous communication between work contracts.
However, the temporary nature of on-demand work raises significant concerns for organizations in terms of gauging the commitment of employees and their sense of belonging. In addition, employers must confront the challenge of monitoring the work of on-demand employees, particularly where they are geographically dispersed (Marica, 2019). This is often accompanied by additional administrative costs (O'Connor & Cech, 2018). From the standpoint of workers, studies have identified one of the downsides to on-demand work to be the instability of income owing to the impermanent nature of these jobs (Eaton, 2012;Jacobs & Padavic, 2015;Stirpe & Zárraga-Oberty, 2017). This leads to uncertainty for workers and dependence on employers who offer jobs based on emerging needs. In addition, we also identify a lack of clarity from the literature concerning regulatory or legislative support for on-demand workers. For instance, the legislation in the UK and Switzerland fails to define on-demand work, and employers are not legally obliged to guarantee a minimum number of working hours (Marica, 2019). These regulatory gaps leave on-demand workers unprotected and exposed to exploitation.

Self-directed work
This type of FWP refers to individuals who have the freedom to either create their own jobs or be associated with established organizations. The concepts used to reflect this type of FWP include self-employment, freelancing, self-directing, gig workers, and skunk workers. The main distinction between on-demand and self-directed work is the relative level of autonomy involved. For "on-demand" practices, workers are not attached to a specific organization, and their working hours cannot be defined in advance (Eaton, 2012). However, self-directed workers have a higher degree of freedom in deciding their terms of engagement, including specified working timeframes such as short-term projects or "gigs" (Furtmueller et al., 2011;Tudy, 2021). For instance, skunk workers refer to an independent group of individuals brought together within an organization to provide innovative solutions and ideas. Taking their available resources and autonomy into consideration, such workers can overcome constraints of routine procedures (Biron et al., 2021). Given their freedom from traditional organizational constraints, skunk workers operate in conditions that may facilitate a narrow preoccupation with how the self is seen and appraised, rather than a wider organizational focus (Furtmueller et al., 2011;Junior et al., 2020). This may, in turn, contribute to looser ties with coworkers and, therefore, adversely affect organizational commitment (Bellesia et al., 2019;Biron et al., 2021). For freelancers, the unpredictability of work exposes them to lulls in business activity and to periods when there are multiple projects to deliver. Consequently, there may be irregularities in income flow (Elfering et al., 2020) and stressful periods of conflict (Haley & Miller, 2015). This is shown to affect worker well-being, particularly when combined with work-family conflicts (Como et al., 2020). We found that digital technologies that support self-directed work are those that enable video, audio, and textual communication as well as organization of meetings, for example, through Google Meet, emails, Zoom, Skype, and Google Hangouts. Other digital platforms allow individuals to promote their services to find work and provide opportunities for payment transactions. In addition, these digital technologies also offer users freemium access, that is, usage without paid subscription unlike those platforms with organization-level subscription for employees. Although self-directed workers enjoy the freemium model that these technologies offer, the workers are subject to the terms and conditions that accompany such access, including limited functionality and potential datarelated privacy issues.

Analytic structure for pitfalls in flexible working
We further analyzed the data to elicit the pitfalls associated with FW, which were both directly stated and implied in the data. We coded these pitfalls and organized our findings into first-and second-order codes. The first-order codes are tags used by the studies to conceptualize the pitfalls in FWP, which we organized into 12 s-order categories to more succinctly capture the domains within which the pitfalls of FWP occur. Some of these represent negative unintended consequences or secondary effects. We observed relations between the categories that emerged and captured these along six dimensions: economic, health, sociocultural, spatiotemporal, technical, and political dimensions. In addition, we mapped these dimensions to different levels of influence as the data suggest (see Fig. 4 for a summary of analytic categories).
Moreover, by performing a content analysis of the papers that directly referenced the pitfalls of FWP, we found a higher occurrence of particular areas of concern such as gender issues and inequality (representing 10% of the papers), work-family conflicts (10%), social isolation (16%), health impairment (16%), stress (20%), and a negative impact of digital technology use (18%). These patterns indicate the direction of research over the period of study and reveal a growing concern with the negative impact of digital technologies on individuals' health and other social dynamics.
To strengthen the internal validity of our analysis, we generated a word cloud (see Fig. 1 in Appendix A) to account for any other relevant codes, which helped in adding another layer of confidence to our inductive analytic process. Thus far, our findings indicate that FWP poses multidimensional pitfalls at different levels. Further, greater focus has been given to sociocultural dimensions, technical dimensions, health issues, and economic issues; we found that political and spatial dimensions have not received as much attention in the literature. These less frequently discussed areas signal opportunities for future research. From our analysis, we also identified geographic variations in the empirical studies that have examined the problems of FWP. These studies highlight contexts in which the research on FWP has raised concerns and provide a snapshot of the various countries where the studies have been conducted (Fig. 5). The US (17%) and UK (16%) account for the highest number of empirical studies, as illustrated. These findings reveal gaps in the global conversations regarding FWP challenges across Europe, Africa, South America, Southern Asia, and the Far East. These data also potentially evince the variance in FWP adoption within and across various continents, where some countries seem to be ahead of others.
The countries indicated in dark colors have received the most attention. This spatial representation gives insight into the contextual L.R. Soga et al. nature of the findings around the pitfalls associated with the adoption of FWP. It is worth noting that some empirical studies (n = 15) represent research conducted in a continent (e.g., Europe, where n = 3) without a more precise reference to a specific country or countries; these were not represented in the map. Others did not highlight any specific geographical location for the research. The map also shows a dominance of studies focused on the Global North (especially the US and UK), with only India following closely behind (6%). In addition, we found a divergence in the perspectives of the author(s) toward FWP. Overall, 32% of the studies in our data set reflect a clearly negative position on FWP and its associated technologies, whereas a total of 68% hold either a positive view or are undetermined about FWP's impact on workers (see Fig. 2 in Appendix A). This lends credence to our research objective herein: to unveil the downsides that appear to have been taken for Analytic structure of pitfalls in FWP  L.R. Soga et al. granted.

Significant trends associated with research on pitfalls in FWP
Our analysis reveals notable triggers, which serve as key milestones that may have had a bearing on the direction of research on the pitfalls of FWP. We observe the advent of smart cities, the emergence of a sharing culture, and the COVID-19 pandemic as contextual elements that provided an impetus for the evolution of research over the period under investigation. We note these three key "moments" on the timeline expressed in Fig. 6.
The advent of a "sharing culture" and "smart cities" in 2010 attracted global interest regarding the role of digital technologies in working practices (Katrini, 2018;Vallicelli, 2018). This notion of sharing challenges the conventional understanding of work boundaries while presents avenues for cooperation among organizations (Kim et al., 2020). The sharing culture is further developed through the idea of smart cities, thereby opening up more opportunities for touch-points between FWP and the environment, leading to positive and negative outcomes (Nedovic-Budic & Williams, 2013). FW models offer benefits for the environment in terms of reductions in carbon footprint from work commutes (Hopkins & McKay, 2019). This decreased commuter activity lowers air pollution and traffic congestion.
However, we recognize that this might not be applicable to teleworkers whose primary work involves frequent commuting within and across cities. Rather, a high number of teleworkers might increase congestion (Ravalet & Rérat, 2019). Regarding "smart cities," these lead to a greater focus on digital technologies as facilitators of communication, with implications for engagement at work (Hopkins & McKay, 2019). The introduction of smart tools, such as smart transit and remote health services, also improves worker performance (Esmaeilian et al., 2018). Individuals use these technological tools to achieve greater flexibility and innovation (Toleikienė et al., 2020). The higher levels of integration associated with smart cities facilitate FWP as individuals leverage the technologies and transition to less conventional ways of working (Vallicelli, 2018). Here, we present a timeline of the dominant themes identified with respect to the pitfalls of FWP (see Fig. 6) to illustrate how the research has developed over the period under investigation.

COVID-19 trends
As indicated on the timeline (Fig. 6), there are similarities in the issues raised in the studies published before and after 2019. However, we observe three key areas of concern that reveal the impact of the COVID-19 pandemic, namely, home-work functions, a high dependence on digital technologies, and the emergence of debates on autonomy versus sharing cultures. These concerns catalyze research on FWP as various organizations continue to adopt FW models (Matli, 2020). For instance, the research suggests that 38% of jobs in Canada and 37% in the US that were performed on site before the COVID-19 pandemic could now be carried out remotely (Como et al., 2020). Our analysis of the data shows that a significant proportion (46.9%) of the research occurred after 2019. This high research interest in the downsides of FWP between 2019 and 2021 could be attributed to the pandemic, as evinced by the timeline (see also Table 2 in Appendix A).

COVID-19 impact: Changes to work-home functions.
The data reveal research concerns on work-family conflicts that were exacerbated by the COVID-19 pandemic alongside other pressures that created additional demands on working parents in particular (Bhumika, 2020). These demands were further intensified by the need of parents to support children in remote learning in addition to focusing on their regular work tasks (Kuc-Czarnecka, 2020). The pandemic has also produced challenges for women in relation to achieving work-life balance, with increased stress and anxiety levels being highlighted (Peasley et al., 2020). Fuller and Hirsh (2019) observed gender-related challenges for working mothers, noting a bias against mothers adopting FW patterns, which can result in wage penalties and negative assessments of an employee's job commitment. These challenges relate to extra work efforts, work-life conflict, and unpaid domestic work (Bhumika, 2020). (2011-mid-2021)  L.R.  impact: Dependence on digital technology as a dominant tool of communication. The COVID-19 pandemic has made the critical role of digital technologies in sustaining FWP prominent. The research shows that a growth strategy could give impetus to digitization, depending on the business model of a firm and portfolio considerations (Bouncken et al., 2021;Correani et al., 2020). However, digital transformation has far-reaching effects on the structure and governance of ecosystems, management of innovation, and competitive relations across firms (Appio et al. 2021); the adoption of relevant business and working models (D'Ippolito et al., 2019); and the long-term innovation capability of an organization through its effect on relational capital (Usai et al., 2021).

Evolution of research on FWP pitfalls
These effects of digitization also extend to individual-and grouplevel factors, such as health, exclusion, and group collaborations as the research has suggested that digitization heightens inequalities and, in particular, social exclusion (Warschauer, 2004) within firms. Our review shows that digital technology has been studied in relation to leading remote workers (Toleikienė et al., 2020), dealing with digital exclusion and inequalities (Kuc-Czarnecka, 2020), teleworker dispersion (Ruiller et al., 2019), and health issues (Johnson et al., 2020). However, concerns remain regarding its increasingly dominant role in human collaborations and communications (Chadee et al., 2021;Hafermalz & Riemer, 2021), with difficulties being present in controlling its impact across multiple geographical locations (Johnson et al., 2020;Rahmadi, 2021). Organizational dependence on digital technologies also results in exclusion for groups without access to said technologies (Soga et al., 2020) or with little technological know-how (Hafermalz & Riemer, 2021), thus leading to unemployment in some cases (Kolasa et al., 2021). 4.4.1.3. COVID-19 impact: Autonomy versus sharing. Yu, Burke, andRaad (2019) established that organizations are operating in an era of "smart cities" and a "virtual sharing culture." However, the COVID-19 pandemic has introduced a rupture in time and generated new concerns for researchers (Matli, 2020). Technology facilitates increased collaboration while, paradoxically, the pandemic's social distancing requirements create social isolation through imposed remote working schemes (Chadee et al., 2021). Social isolation and loneliness emerge as unintended consequences and create conditions for a sense of detachment or "autonomy" from organizations (Aguilar Gordón, 2020). This tension between a technology-enabled sharing culture and a technologyinduced sense of autonomy raises concerns in terms of straddling virtual working and social connectivity (Almonacid-Fierro et al., 2021;Baughman, 2019).

Discussion
Our data analysis reveals wide geographical variations in the adoption of FWP, as depicted in Fig. 5. Some countries, such as the UK and the US, reflect a greater adoption of FWP compared to other countries such as Ghana and South Africa. The data suggest that Finland is a global leader with 92% of companies adopting flexible schedules (Wood, 2019). The US and the UK also have high adoption rates with 79% of companies offering FW approaches (Lockwood & Nath, 2021;Mangla, 2021). Regarding opportunities for learning, the dominance of empirical research in the Global North and the dearth of studies in other contexts limit our understanding of the effects of FWP in different settings. This is particularly important as the digital divide creates conditions for uneven impacts in the adoption of FWP. These conditions are worsened by global events such as pandemics that can change the dynamics of working practices and precipitate shifts to flexible forms for which some regions may be less prepared. Therefore, it is important for each country to develop FW models that suit their contextual needs.
Second, our analysis reveals contradictions about FWP regarding its multidimensional aspects and societal implications. The data suggest that some studies have considered FW as a good practice for work-life balance (Wheatley, 2012a;Wheatley, 2012b), while others claim that it leads to work-life conflicts (Rafnsdóttir & Heijstra, 2013;Thornton, 2016). These different positions indicate a lack of clarity when it comes to broad attitudes toward FWP, particularly when viewed against the backdrop of COVID-19 and concerns related to digital technology use. In this new digital era intensified by the pandemic, the need to shift to virtual work contributes to a deepening of the digital divide-a segregation between two worlds where one has already stepped into digitization and the other remains marginalized (Hasan et al., 2021). Specifically, owing to a lack of technological capacity or resources, the latter is still outside the virtual game and struggling to access it.
A combination of benefits from embracing digital technologies and cost savings for organizations suggests that we can expect to see FWP rise (Como et al., 2020). By contrast, the growing adoption of FWP masks the more long-term negative effects on individuals and groups. More specifically, organizations can miss the critical warning signs that signal a need to reconsider their deployment of FWP if they fail to examine the unique needs of their employees. There are also environmental impacts because of the widespread use of FWP. Direct impacts are observed in the reduction of carbon footprint as workers travel less (Ellder, 2020) and the decline in the use of public spaces because of social distancing and virtual working (Yu, Burke, & Raad, 2019). However, teleworkers and freelancers can work from various locations, including public spaces such as parks and gardens (Tudy, 2021). Therefore, we may conceive of a new use of these spaces, such as for networking and working purposes. This leaves the purpose of the public space-although originally designed to facilitate public health and wellbeing-open to debate (Bentley et al., 2016).
In addition, the implications of FWP relate to its financial dimensions. Significant events outside human control (such as the COVID-19 pandemic) can shift some cost burdens (e.g., expenditure on utilities) from organizations to workers (Jacobs & Padavic, 2015). Although there is insufficient evidence to confirm that the cost savings have been transferred to employees, there have been some government policies that attempt to alleviate the same through tax refunds, signifying governmental recognition of FWP's cost implications (Rubery et al., 2016). Indeed, decisions to transfer cost savings come with uncertainty regarding what would be considered equitable allocation principles. This may explain the limited literature on cost reimbursements for FW employees, who potentially bear the burden of additional expenses. There is scope for research on the financial impact of FW on both organizations and workers.
On a wider scale, other pitfalls are associated with FWP in terms of its impact on infrastructure, including capacity underutilization for existing facilities that were originally designed for on-site work (Yu et al., 2019). The implication is that organizations would bear the acquisition costs of workspaces without the benefit of their use. In some cases, there may be conversion costs to cater for repurposing the workspace. This has significant implications for cities and city planners, who frame policies in response to changing business needs to accommodate FW trends within the ambit of legal provisions.
The ambiguity around the idea of "flexibility" that our findings evinced makes FWP a contested phenomenon. This ambiguity continues to challenge employees and employers as both groups contend with the pitfalls associated with FWP. In addition, the capacity to respond flexibly differs, which is particularly salient when it comes to physical infrastructure, creating hidden costs for workers and organizations. There are calls to regulate "flexibility," which raises definitional concerns when viewed against the emergence of co-working spaces and sharing cultures. "Flexibility" here is no more flexible as rules are introduced to control how employees share space and resources, especially as the digital technologies that enable flexibility are not necessarily flexible themselves (Lucas Jr & Olson, 1994). This lack of clarity reflects the inherent ambiguities in FWP as enabled by digital technologies (Kolasa et al., 2021).
Given the pitfalls raised and calls for the regulation of FWP, we can question the extent to which regulations can be successfully enforced without losing the prized flexibility, leaving room for future studies to address this topic. Although there is a strong argument for FWP, our critical analysis of the literature about its pitfalls shows that it creates systemic vulnerabilities along individual, organizational, and societal dimensions, which require protective regulatory measures. Therefore, we argue for a conceptualization of FWP that accommodates its limits to freedoms but recognizes an underlying equitable basis. Nevertheless, we still have much to learn about the real price of flexibility and worker freedom in a digital world.

Reflections and conclusion
Through our review of the downsides of FWP over the past decade, we observe health, sociocultural, economic, spatial, technical, and political dimensions. We highlight changing work-home functions, the role of digital technology as a dominant tool for communication, and the emergence of debates on flexibility, autonomy, and sharing cultures. We develop a framework of FWP, synthesizing it into the four broad categories of remote work, spatiotemporal work, on-demand work, and selfdirected work, thus offering a theoretical framing for future research on the disadvantages of FWP. Some authors argue for regulating FW models to address its inherent problems or reduce its impact on individuals and organizations. However, the widespread deployment of digital platforms for FWP creates conditions wherein such regulation is merely reactive to its adoption and attendant unintended effects. In addition, we call for research on the impact of the digital divide and the understanding of different contexts as well as on the conceptual ambiguities, regulatory, and spatial dimensions associated with FW.
Our findings also offer practical implications for businesses and managers in addressing the pitfalls of FWP. We highlight the need for research on a wider range of contexts to enhance the existing understanding of FWP to avoid the blind adoption of practices from other contexts. This study is aimed at moving practitioners toward adaptations that have a contextual fit. In addition, we raise practical implications for the managers of technological platforms. During an era of ubiquitous technologies (see Fig. 2), managers of these platforms should cautiously adopt digital technologies instead of deploying readily available, offthe-shelf options, particularly because competitive pressure or cost considerations may lead to blind adoption (Lee, 2004). There are additional implications for managers concerning specific user needs and the unintended consequences of the technologies deployed. Managers would benefit from seeing the learning curve in organizations as a serious tech-support issue without assuming that individuals already possess technological know-how. Moreover, there are also implications for policymakers regarding technology use as well as labor laws to protect employee health, safety, and privacy. In addition, regulations related to the design and (re)use of spaces can help mitigate the risks associated with shifts to more remote forms of working, which is particularly relevant for building design and town planning.
We note some limitations to this review: first, our sample size (n = 113) attests the paucity of attention paid to the pitfalls of FWP in academic research. Therefore, we call for continued investigations on the issues raised herein. Such research should include dimensions that are implicitly discussed in the literature as well as those highlighted in this study, such as the political and spatial issues associated with FWP. Second, we highlight the need for research in the Global South and in countries that have only recently deployed FWP. Based on our findings, we speculate that there might be an increased adoption of FWP and call for future research to consider how the technological, historical, and socioeconomic contexts of countries affect the emerging practices and meaningfulness of flexible working. This expands the scope for further learning on the pitfalls of FW. Third, we find that considerably few studies have presented concerns for particular technological platforms or software. As indicated by our findings, only 32% of the authors have a negative view of FWP and the technologies supporting it, which means that a critical assessment of digital platforms has not been presented as the focus is more on the impact and challenges of the use of digital technologies.
Additionally, while almost all the studies have explored both the positive and negative consequences of technology use, none have examined the downsides of changes in (or to) technological platforms on employee behavior and work. This is particularly important as organizations shift from one technology to another/newer versions or developers introduce updates. Therefore, the following two questions are raised here: 1) how can the changes in (or to) digital platforms affect organizations, employees, and employee-employer relationships? 2) How do specific digital platforms (as illustrated in Fig. 2) impact FWP contexts? Surprisingly, researchers have not explicitly focused on the changes in traditional hierarchies and the dynamic nature of manager--employee relationships as a result of technology-enabled FWP. Generally, authors have understood technology as a tool that facilitates FWP, thus perpetuating the instrumental view of technology. Given the active role that technological platforms continue to play in organizational life, other conceptualizations of technology are required, such as those offered by object-oriented ontology, actor-network theory, and other practice theories, to further our understanding of the effects of technology in FWP.
In summary, understanding the pitfalls associated with FWP also provides opportunities for innovative responses that create more equitable and sustainable work practices, policies, and societies. In this light, the COVID-19 pandemic represents a significant inflection point for research into FW pitfalls, holding the promise of a more in-depth understanding.