Sustainability Implications of Working-From-Home (WFH): A Systematic Review of the Travel Behavior Literature

SLR Search Strategy

The article identification process was carried out between 15 August 2022 and 14 January 2023. Three databases were searched: Web of Science, Scopus, and Ebsohost. These were selected because of their well-established breadth and depth. The search concepts, terms, and strings that we employed are summarized in Table 1. Recognizing that small changes in the search terms can produce large differences in results, we refined the search terms based on an iterative process seeking to ensure a comprehensive selection of studies. The effectiveness of the search protocol was evaluated by adopting a gold-set testing process (Whiting et al. 2008). This involved iteratively testing the ability of our search to retrieve known articles that relate to our research question. In this manner, the search protocol was able to pick up a set of “golden papers.”

OPEN IN VIEWER

Table 1.

Search Concepts, Terms, and Strings.

Search concepts	Search terms and strings
Terms that describe WFH	(“Working from home” OR “work from home” OR “WFH” OR “flexible work*” OR “remote work*” OR “telewor*” OR “telecommut*” OR “hybrid work*” OR “flexible employment*”) AND
Terms that describe travel	(“Travel behaviour” OR “travel pattern*” OR “travel habit*” OR “commuting behaviour” OR “sustainable travel” OR “activity pattern*” OR “transportat*” OR mobil*OR trave*)

Abbreviation: WFH = working-from-home.

Timeframe

We included articles from 2000 to 2022 (inclusive). This timeframe is technologically comparable as it witnessed the widespread adoption of mobile phones and the emergence of videoconferencing. Also, this timeframe allowed us to include a sample of studies conducted before, during, and after COVID-19. ¹ The articles in our database were categorized into before, during, and after COVID-19 groups based on the data they employed. For instance, if a paper was published in 2020 but relied on data from 2018, it was categorized as “before COVID-19.”

Screening Criteria

Table 2 lists inclusion and exclusion criteria. Following these criteria, two researchers separately screened articles by the title, and abstract using the Covidence software (Kellermeyer, Harnke, and Knight 2018). Then they compared notes and reached consensus over a small number of inconsistencies in judgment. The full texts of the short-listed articles were checked against the inclusion and exclusion criteria. Only studies that were relevant to the research questions were selected for inclusion in the dataset. Regarding the “Outcome measures” inclusion criterion, we included articles if they measured at least one of the travel-related behavioral outcomes listed in the table.

OPEN IN VIEWER

Table 2.

Inclusion and Exclusion Criteria.

Inclusion criteria	Exclusion criteria
Outcome measures: travel distance, travel time, travel mode, trip frequency, residential location choice, cycling, walking, public transport ridership, car use, and car ownership.	Other outcomes/non-travel behavior outcomes (e.g., health, household consumption, energy/water/internet use).
Language: articles published in English.	Language other than English.
Study design: studies relying on primary data (qualitative, quantitative, descriptive, longitudinal, cross-sectional), or modeling.	Unclear study design.
Publication type: Peer-reviewed articles, book chapters, and conference papers.	Opinion/commentary/discussion/review article, conference abstract, protocol, thesis, preprint.
Sample: employed adults.	Studies that rely on other population subgroups.
Interventions: all intervention types that include WFH variants. Flexible working arrangements are included if linked to WFH.	Interventions that do not include any WFH aspects.

Data Extraction

Article metadata was extracted into a Microsoft Excel file using the following checklist:

Author(s) name(s) and year of publication

Data Analysis

A thematic analysis was conducted—separately for each outcome (Popay et al. 2006). Four steps were applied to synthesize the data: (1) read the results of each study and extract relevant findings; (2) develop a preliminary synthesis by describing the studies and identifying patterns; (3) categorize studies by travel behavior outcome and organize factors linked to the outcomes; and (4) thematize items and explore the relationships of each factor within and among themes. The characteristics of the included studies were described and summarized in the tables and the figures (see “Description of Included Studies” section). Note that in this case, it was not possible to conduct a meta-analysis due to the diversity of measures used to estimate the outcome variable (travel behavior).

Final Database

Figure 1 depicts the PRISMA flow diagram. The initial search returned 428 publications (after removal of duplicates). After screening by title and abstract, eighty-one studies were included for full-text screening. Out of these, forty-eight studies met the final inclusion criteria, while thirty-three were excluded. The most common reasons for exclusion were (a) unclear research design and (b) unclear empirical methodology. Discussion papers and policy review documents were excluded as well.

OPEN IN VIEWER

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

Trendline of Studies

The plot in Figure 2 shows the number of publications by year (2000–2022). It is clear that:

Until 2017, the rate of WFH publications was consistently low, with just a few articles per year.

OPEN IN VIEWER

Figure 2.

Number of published studies by year.

Many studies that appeared during the pandemic utilized empirical data collected before the COVID-19 outbreak. This suggests that the increase in WFH-related publications since 2017 reflects the pandemic only in part. Interest in WFH appears to have preceded the lockdowns by several years.

Study Settings

Most studies (84 percent) were set in wealthier regions, including North America (n = 18), Europe (n = 18), Australia (n = 7), and East Asia (n = 5). This means that the evidence is biased toward cities that have long been characterized by housing unaffordability and gentrification (Burke and Hulse 2010; Thackway et al. 2023). In addition, WFH might be popular among workers in more developed countries who have better access to Internet technologies and infrastructure. Furthermore, a prevalence of white-collar jobs in developed countries tends to promote WFH opportunities.

Dependent Variables

The included studies examined the following outcome measures (i.e., dependent variables):

Independent Variables

The included studies controlled for some or all of the following variables:

Study Participants

Study participants included:

Individuals who worked from home, at least some of the time (de Abreu e Silva and Melo 2018b; Melo and de Abreu Silva 2017) and/or individuals who commuted to a workplace (Zhu and Mason 2014). And,

Study Methods

The methods employed in the included studies were primarily quantitative, regardless of the travel behavior being measured (Dissanayake and Morikawa 2008; Ellder 2020). Two studies assessed WFH in multiple cities across the globe(Balbontin et al. 2021; Matthews and Williams 2005). While the majority of the studies focused on a single city, some studies addressed the travel behavior of people with flexible work arrangements in multiple cities within the same country (Ellder 2020; Hensher, Beck, and Balbontin 2021a). Data sources included existing national household travel surveys. Importantly, many post-pandemic studies used purpose-built surveys to capture information about the travel implications of the pandemic (Caldarola and Sorrell 2022; Choo, Mokhtarian, and Salomon 2005; van der Loop, Haaijer, and Willigers 2019a; Zhu and Mason 2014). Two studies used scenario modeling based on simulations or other types of models to project the future impacts of WFH on travel behavior (Lila and Anjaneyulu 2017; Moeckel 2017). Most studies set in the post-COVID-19 period were based on modeling and prediction, with many drawing on hypothetical data. Qualitative studies based on interviews were in the minority (n = 2, both based on interviews) (Lim 2005; Lyons and Haddad 2008). Only one study used a mixed methods approach, combining quantitative and qualitative methods (Vaddadi et al. 2022).

Peak Hour Travel

Studies typically defined “flexible working” in terms of schedule rather than location. Some studies considered the total avoidance of travel during business hours. The synthesized evidence for all three periods showed that workers with flexible schedules (by day or week) tended to postpone the morning peak hour commute to later in the day (Bussière and Lewis 2002; Ellder 2020; He 2013; van der Loop, Willigers, and Haaijer 2019b; Shaz and Corpuz 2009). This resulted in a spreading out of peak-hour traffic. A reduction in peak-hour travel was more pronounced during the COVID-19 pandemic period (Campisi et al. 2022b), possibly because of the strict lockdowns implemented in many countries. Compared with workers with fixed work schedules, workers with flexible work arrangements were less likely to travel during rush hour (He 2013).

A Dutch study using 2000–2016 data found that when individuals worked from home and avoided peak-hour travel, travel delays were reduced. Specifically, the study reported a reduction of 20 min during the morning peak, 12 min during the afternoon peak, and 10 min over the entire day for each person with flexible work arrangements (van der Loop, Willigers, and Haaijer 2019b). Full-time WFH has a higher impact on the daily congestion average compared to part-time WFH (Ellder 2020; van der Loop, Willigers, and Haaijer 2019b).

Overall, studies found that WFH for at least two to three days resulted in a substantial reduction in peak-hour travel. This was highlighted in studies that used data from the pandemic era but even after COVID-19 restrictions eased, peak hour travel and congestion did not reach pre-pandemic levels, as many people continued WFH/hybrid working. These trends may have been reversed in the post-pandemic period, but studies are too limited to substantiate this trend, and the norms around WFH behavior are likely still evolving.

Studies found that, among people with flexible work arrangements, the propensity to travel during the peak hour was influenced by the worker's age, gender, ethnicity, race, and income (Lachapelle, Tanguay, and Neumark-Gaudet 2018; Shaz and Corpuz 2009; Yang, Steiner, and Srinivasan 2016). For instance, teleworkers who were female, older, less educated, poorer, and carless were less likely to travel during the peak (Ellder 2020). Meanwhile, full-time teleworkers with children were more likely to travel during the morning peak (Lachapelle, Tanguay, and Neumark-Gaudet 2018). ² This is because parents need to drop off and/or pick up children at school, which indirectly impacts their work start/end time. Even though professionals and managers were shown to have the most flexibility, they were still the ones most likely to travel during rush hours (Yang, Steiner, and Srinivasan 2016). Also, a more flexible work schedule may translate into scheduling personal appointments such as banking, doctor's visits, and suchlike in the early morning or evening.

Travel Distance

Several measures were used to assess travel distance in the reviewed studies. The choice of a specific measure depends on the study's research objective, data availability, and level of detail needed. Very few of the reviewed studies used travel time to measure how long workers were traveling for different activities. Most studies used the distance traveled for work and non-work activities on a daily, weekly, monthly, or yearly basis, or the one-way commuting length.

Travel distance (i.e., length) was typically measured in miles or kilometers. This was derived from respondent-reported distances, network measurements, or calculations by professionals. Typically, these calculations determined the distance from the home centroid to the workplace or other activity locations. Travel duration or travel time was measured in minutes and included the total time spent on a trip, covering both the time spent in travel and any waiting times (Hensher, Beck, and Balbontin 2021a).

Travel time as a measure of distance was used in two studies, which found that teleworking was associated with a reduction in the overall travel time but not the total amount of travel distance (Elldér 2017; Melo and de Abreu Silva 2017). From a behavior change perspective, distance and time are related, but are not synonymous. Individuals may have a better understanding of travel time, rather than distance, and may also be more responsive to variations in time, rather than distance. For instance, peak and off-peak travel differ in terms of time, but not necessarily in terms of distance.

Commute Frequency

This metric relates to the commute distances, but it is distinct, as trips vary in duration and length. The studies included in this review have generally found that WFH is associated with fewer but longer commuting trips. The relationship holds in all three eras—but it was particularly strong during the COVID-19 pandemic as many people were forced to stay home (Beck and Hensher 2022; Hensher, Beck, and Balbontin 2021a). In settings where the proportion of people with WFH increased from 25 percent to 75 percent, the average number of weekly commuting trips decreased from 2.1 to 0.6 trips per person (Hensher, Beck, and Wei 2021b).

Walking and Cycling

Before the pandemic, WFH was associated with greater use of non-motorized modes (Lachapelle, Tanguay, and Neumark-Gaudet 2018), and full-day teleworkers were more likely to use active modes of travel than part-day teleworkers (Ellder 2020). The time freed up by the commute was perhaps being used to walk and cycle more. COVID-19 intensified this trend (Campisi et al. 2022b; Ellder 2020; Hensher, Beck, and Balbontin 2021a). In Argentina, the active travel mode share increased from 11 percent to 32 percent (Balbontin et al. 2021) whereas in Australia this share increased from 8 percent to 12 percent (Balbontin et al. 2021). Governments that imposed WFH, social distancing, and other forms of isolation still encouraged people to engage in physical activity outdoors to maintain physical health and relieve the psychological stress of the pandemic. Many people at the time began taking shorter trips around their residential areas (Wang and Ozbilen 2020). These trends may be sustained after the COVID-19 pandemic. Compared to part-time teleworkers, those who work full-time are more likely to use active modes of travel (Ellder 2020).

Public Transport Use

The evidence linking flexible work arrangements and public transport use was limited. However, it appeared that teleworkers workers were more likely to ride public transport for their commuting trips on the days they traveled to their workplace compared to workers who commuted five days a week (Campisi et al. 2022b; de Abreu e Silva and Melo 2018b). This might be because workers with flexible work arrangements are not constrained by time and the opportunity to start late and finish early supports public transport use (public transport being slower than the private car). The opportunity to work while traveling also increases public transport use among workers with flexible work arrangements (WFH for part of the day) compared to workers who are required to travel to a workplace and follow a traditional schedule. Finally, as a fixed route/fixed schedule mode, public transport is perhaps seen as more suitable for (a) long-distance commuting, which can be stressful by car, ³ and (b) non-work-related travel by highly flexible people, who can work anywhere and are not constrained by schedules.

The COVID-19 outbreak introduced major fears of contamination in enclosed environments, prompting a substantial decline in public transport use (Beck and Hensher 2022; Hensher et al. 2022). For example, a study set in Australia reported that monthly trips on public transport fell by 80 percent and suggested that this was in part due to fear of contagion but also owed to an increased prevalence of WFH (Beck and Hensher 2022). But for those who worked outside the home and could not drive, transit remained an important means of mobility even during the pandemic period (Parker et al. 2021).

While the COVID-19 pandemic caused all types of travel to initially fall across many different travel work arrangements, the declines in travel by transit were greater among higher-income workers (Brough, Freedman, and Phillips 2021; Jiao and Azimian 2021). Wealthier and more educated neighborhoods lost more transit riders (Brough, Freedman, and Phillips 2021). Other factors that contributed to low public transport use during the pandemic included shrinkages in public transport services, restrictions on the maximum number of people on board, a general fear of COVID-19 infection, and a failure of public transport systems to properly adapt to a modified business environment (Campisi et al. 2022a).

As the pandemic is coming to an end, public transport demand is recovering; however, some researchers estimate that the share of transit commuters will likely remain lower than before the COVID-19 outbreak (Conway et al. 2020; Javadinasr et al. 2021). There will likely be inertia here as many communities are still living with COVID-19, as of this writing (Huang, Loo, and Axhausen 2023).

Car Ownership and Use

While teleworkers may have used public transport more than the general population, the car was still their primary mode in the period before the pandemic (de Abreu e Silva and Melo 2018a; Glogger, Zängler, and Karg 2008; Javadinasr et al. 2021). Tasks such as transporting children to school and shopping were mostly completed by car (de Abreu e Silva and Melo 2017; Zhu and Mason 2014). Also, it is the case that when one household member takes up WFH, this frees up his or her car which can then be used by other household members (Guerin 2021b). The COVID-19 pandemic strengthened the role of private cars in cities (Meena 2020; Zannat et al. 2021). In some places, car use kept growing even after the lockdowns were lifted (Javadinasr et al. 2022), due to the higher risk of catching the virus on public transport and the outward population move mentioned earlier.

However, the relationship between car ownership/use and residential location is complex, and there is self-selection at play, meaning that more car-dependent households are more likely to choose living in remote areas (Guerin 2021b). Most studies reviewed here controlled for trip characteristics in predicting mode choice, particularly car ownership and use. Controls included the number of trips per day, travel time, travel distance, trip timing (weekend vs. weekday), and number of vehicles per household. All of these were shown to have a positive association with car use (de Abreu e Silva and Melo 2017; Zhu and Mason 2014). One study found that even during the pandemic, the correlation between the number of cars per adult in a household and the probability of commuting by car as a driver remained (Hensher, Beck, and Balbontin 2021a).

Environmental Sustainability

Travel behaviors of people WFH have clear implications for aggregate and local traffic flows, with impacts on transport pollutants and greenhouse gas emissions. To the extent that reduced peak hour travel is associated with better-flowing traffic, greenhouse gas emissions are reduced through greater efficiency in fuel combustion. More efficient combustion and, in some cases, reduced overall travel also have air quality impacts. From an environmental perspective, WFH can thus constitute a policy option for cities and regions struggling with air pollution, as well as congestion (Shabanpour et al. 2018). However, if travel (and traffic) is more spread throughout the day, rather than concentrated in a few peaks, the frequency and timing of public transport will need to change. Increasing off-peak public transport service might incentivize the uptake of more sustainable modes of travel among those working from home. However, this comes with higher financial costs, which may only be palatable to the public if combined with other strategies to reduce car-based travel. An unintended consequence of WFH may be the phenomenon of induced demand whereby reductions in congestion induce new car trips, thus negating initial gains in environmental sustainability.

While WFH may suppress mobility (i.e., the number of trips taken), any remaining commutes may become longer; people who work from home live up to 75 percent farther from their jobs on average. This is not ideal in terms of air quality, especially if those who work from home rely on a private car. Typically, longer commutes are found where the urban environment is characterized by sprawl. Strategies to densify cities while providing affordable housing next to major employment centers can enable shorter commutes for those with occupations that are amenable to WFH. In addition to longer commutes, WFH is sometimes found to result in longer, more frequent, and more car-based trips for non-work purposes (Kitou and Horvath 2006; Zhou and Hagen 2006). The impact depends on the proximity of the teleworker's home to schools, retail outlets, and other essential destinations (Thakuriah et al. 2012). For WFH to benefit the environment, easy access to non-work activities must be present. This is not a short-term solution and is further complicated by income inequalities which affect the quality and density of amenities that can be sustained in different locations.

Social Sustainability

Travel distance varies across the working population, with younger workers, female workers, and working mothers tending to travel more each day. WFH, particularly when combined with relocation to more spacious homes in remote and underserviced places, may further exacerbate travel inequalities at the societal level. Inequalities may grow at the intra-household level too, if relocation is driven by the needs or desires of only one household member (Einstein et al. 2023). For example, in a traditional family household, relocation may benefit the male partner by providing space for a home-office but lead to more travel related to house chores and childcare for the female partner. Another aspect to consider is that, while the studies in our sample focused on travel behavior, WFH could also lead to more stress among parents and other caretakers (Xiao et al. 2021).

For some groups, including rural workers, women with small children, and people with disabilities, WFH might be the only option (Ameri et al. 2022). On the positive side, WFH may mean, in locations that are already rich in amenities, greater utilization (and market potential for) local services. Rather than accessing services and amenities around the place of work or en route to/from work, people may engage more with their locality. This may enhance community well-being and public health.

Economic Sustainability

Off-peak travel can have tangible benefits for individual, household, and government budgets. Congestion is an externality with societal economic costs. By avoiding excessive work commutes, individuals can save on car maintenance, workplace parking, toll, and insurance costs, and free up time for leisure or work (Elldér 2017; Hensher, Beck, and Balbontin 2021a). A study from Europe estimated that people could save 178 Euros and 2 to 14 h per week on average if they travelled off-peak (Bieser et al. 2021; Glogger, Zängler, and Karg 2008).

The time and money formerly wasted sitting in traffic can be used for productive or recreational activities. Finally, off-peak travel helps cities “sweat their assets”—that is, make more efficient use of the existing road infrastructure networks. On the other hand, if commute-related travel is spread throughout the day, rather than concentrated into a few peaks, and if demand for public transport declines, this may reduce government revenue from public transport, potentially undermining its financial viability. Further, if workers move to the exurbs en masse if given the opportunity to WFH, businesses such as restaurants and shops may follow them. This might spur more private car use, ultimately resulting in higher infrastructure costs for all modes. Policymakers and planners should be proactive in responding to these shifts.

Another implication of WFH is a potential growth in active travel for non-work activities among teleworkers when compared to office-based workers. This may lower public health expenditures, as active travel is known to promote good health (Besser and Dannenberg 2005) and reduce the risk of premature death (Davis et al. 2005). While we did not examine the comparable costs of WFH, some studies in this review concluded that it increases productivity and job satisfaction (Criscuolo et al. 2020; Moglia, Hopkins, and Bardoel 2021).

Implications for Policy and Practice

Much land-use and transport planning practice remains centered around traditional ways of working and commuting (Mualam, Israel, and Max 2022). To harness WFH's environmental, social, and economic sustainability potential, urban planning will need to adapt and change. Several policy options, listed below, could assist cities in their efforts to embrace WFH:

Adapting public transport service. During the preparation of transport plans, planners should consider (a) reductions in commute-related travel, (b) shifts in travel hours, and (c) increases in walking and cycling due to WFH. To accommodate these changes, public transport services may need to be more distributed throughout the day rather than concentrated in the commute peaks. Public transport networks may need to extend to outer suburbs, perhaps in the form of feeder buses. Further, public transport funding may need to be reallocated to providing better infrastructure for walking and cycling in order to accommodate a growing demand for these travel modes.

Directions for Future Research

Further research is needed to identify the long-term impact of residential location choices on the frequency of travel, both for work- and non-work-related activities. Importantly, the study found a considerable lack of in-depth qualitative research around teleworkers’ travel behaviors. It is not clear how much WFH shapes how and when people travel to work, or whether WFH might be considered a travel mode choice in its own right, and a sustainable one at that. We recommend that WFH be considered as a commuting mode-choice and explicitly treated as such in future research studies. ⁴ Other re-conceptualizations of WFH are possible too. For instance, WFH may be considered as a sustainable lifestyle choice (as opposed to a practical or employer-driven solution). Research that provides an in-depth understanding around the mechanisms and dynamics that are observed at the macro-scale is still needed. Studies should help to uncover the WFH circumstances and/or social practices that catalyze or undermine sustainable mobility outcomes.