Before the COVID-19 pandemic, WFH was not a widespread practice in Mexico. Leyva and Mora (2021) estimated that just 10.6% of formally employed individuals in Mexico participated in WFH before the pandemic. Even during the lockdown, Mexico’s telework rate remained relatively low because the government recommended voluntary confinement rather than enforcing a mandatory lockdown, thus leaving businesses to decide on their restrictions (Jaimes Torres et al., 2021). However, the Cranfield Network on International Human Resource Management Survey (CRANET) indicated an increase in Mexican companies adopting telework. The CRANET survey revealed that 19.4% of Mexican companies reported that at least 20% of their employees teleworked before the pandemic. This rate surged to 63.6% during the pandemic and stayed at 45.8% afterward. While it is challenging to determine the exact rate of telework adoption in Mexico, the CRANET survey indicates a growing trend.

Oviedo-Gil and Cala Vitery (2023) argue that workers’ acceptance of WFH depends on their ability to perform the job outside of the office, employers’ willingness to allow this modality, and the personal convenience of this setup. In Mexico, Ordoñez-Parada (2018) investigated the antecedents of job satisfaction among 247 teleworkers and found that work environment conditions were crucial to job satisfaction, independence, and working hours. Studies conducted during the peak of the pandemic outbreak in Mexico, when mortality numbers and hospitalizations were among those at the top of the world, also pointed to the importance of work environment conditions (Barriga Medina et al., 2021; Jaimes Torres et al., 2021).

Wang et al. (2023) investigated responses from more than 2,000 teleworkers across the US. They found that employees seek flexibility concerning time and physical location, good living conditions, short or no commute, and little work-family interference while WFH. Stoian et al. (2022) investigated almost 300 employees in different countries of the European Union. The availability of stable infrastructure (at the workplace and home) that enables flawless WFH performance, a good relationship with the supervisor, opportunities for knowledge exchange, work-life balance, no professional isolation, and saving commuting time were pivotal factors for participants.

Eliminating commuting time through WFH is particularly important in Mexico and other Latin American countries with less-than-ideal infrastructure and transportation from working-class neighborhoods to city centers, as Lara-Pulido and Martinez-Cruz (2023) pointed out. Hess and Narteh-Yoe (2020) suggested that about 80% of commuters in Mexico City wish to change their daily routine to reduce long commuting time. Oviedo-Gil and Cala Vitery (2023) ranked Mexico City as the second worst place in job quality in LATAM, primarily because of substantial commuting time. Ton et al. (2022) reported that 71% of research participants wished to continue WFH even after lifting all pandemic-related restrictions, as they experienced an improvement in their work-life balance owing to a dramatic decrease in commuting time. These prior findings led us to hypothesize that certain location-related variables related to the conservation of personal resources (e.g., time, effort) associated with WFH will moderate the relationship between job satisfaction and productivity.

Lapshun and Madero (2023) noted that WFH increases employees’ perceived productivity compared to working from other locations outside the employee’s residence, as long as the latter is suitable for telework. Cuerdo-Vilches et al. (2021) also noted the importance of work dwelling quality (noise level, air quality, lighting) on teleworkers’ productivity. Thompson et al. (2022) pointed out that one of the primary advantages of WFH lies in being sheltered from distractions at work. Based on Karasek’s (1979) job demands-control model, we predicted that not having a dedicated workspace while WFH might interfere with employees’ ability to control their jobs. Indeed, working from remote locations other than one’s residence poses challenges and distractions that make WFH virtually impossible for many employees. These conditions were particularly challenging during the pandemic, when most co-sharing workspaces were closed, and other public places, such as coffee shops, dramatically cut their schedule. Although not every household meets the bare minimum conditions necessary for productive WFH (Berniell & Fernandez, 2021; Cuerdo-Vilches et al., 2021), working from one’s residence seems most likely to provide a distraction-free location suitable for telework. Therefore,

Hypothesis 1a: The relationship between employee satisfaction with teleworking and employee self-perceived productivity will be moderated by whether employees telework from home, such that the relationship will be stronger for those who telework from home than those who telework from other locations.

Thompson et al. (2022) researched two groups of teleworkers and one group of non-teleworkers, asking them why they would or would not like to telework. Three of the 2,054 reasons shared were voiced most frequently: avoid commuting, increase productivity, and meet family demands. Therefore, Thompson et al.’s data support the idea that cutting commuting time is one of the primary reasons behind WFH. Indeed, commuting long distances, particularly in the heavy traffic conditions that characterize many Latin American metropolitan areas, is likely to take a toll on employees’ resources like time and energy. We did posit that traveling more than 50 kilometers (equivalent to approximately 31 miles) from home to work is a critical threshold because this distance typically requires commuting time of at least one hour each way. Indeed, well before the COVID-19 pandemic, Graizbord (2015) highlighted the time wasted commuting in megacities like Mexico City. Wang et al. (2023), Stoian et al. (2022), and Lara-Pulido and Martinez-Cruz (2022) have also called attention to commuting distance as one of the primary reasons for WFH. Due to high gas prices, long commuting distances are also a financial burden (Lara-Pulido & Martinez-Cruz, 2022). Not surprisingly, Ton et al. (2022) reported that teleworkers believed their productivity increased when they cut commuting. Unlike in some European cities, where public transportation is reliable and widely available, commuting trips in Mexico City are subject to random traffic fluctuations and often occur in crowded, small vehicles (Graizbord, 2015). Hence,

Hypothesis 1b: The relationship between employee satisfaction with WFH and employee self-perceived productivity will be moderated by the commuting distance between home and the workplace, such that this relationship will be stronger for those whose workplace is at least 50 kilometers away from their home than for those living closer to their workplace.

Control over time is vital in the pursuit of basic needs satisfaction (Ordoñez-Parada, 2018). Gaining control over one’s own time might undoubtedly be a factor that drives employee satisfaction with WFH (Knight & Westbrook, 1999). Indeed, commuting time during peak hours is a big unknown in large urban areas in Latin America, even in short distances. Navigation system statistics suggest that commuting by car during rush hours involves up to 195 hours a month on the road in Mexico City (TomTom, 2020). Due to the high commuting cost relative to low-income levels, Arredondo (2017) estimated the financial burden of commuting to be as high as 40% of the annual middle-class income. Therefore, commuting time is probably the most straightforward operationalization of the personal resources consumed by commuting because distance is not always a reliable indicator of time owing to varying traffic conditions. Thus,

Hypothesis 1c: The relationship between employee satisfaction with WFH and employee self-perceived productivity will be moderated by commuting time between home and the workplace, such that this relationship will be most vital for those whose commuting time will take at least two hours a day (e.g., one hour each way) than for those whose commuting time is shorter.

Raisiene et al. (2022) investigated satisfaction with WFH among educators and managers. Educators felt stressed by WFH when WFH drastically cut their social interaction. They also admitted to experiencing low self-management capabilities while WFH. Hajal (2022) noted that the propensity to telework correlates with employees’ ability to do so effectively. In certain occupations like education and the commercial sector, face-to-face, live interactions have a critical impact on productivity, and therefore suppressing the ability to interact face-to-face should understandably reduce the workers’ ability to excel in their jobs. According to Karasek’s (1979) job demands-control theory, the inability to exercise complete control over one’s job by enacting face-to-face live interactions that might dramatically boost productivity should attenuate the satisfaction with the WFH-productivity relationship. Even though some employees enjoy the convenience of WFH, not all of them might feel productive because the WFH modality curtails their perceived ability to control their jobs. Consequently.

Hypothesis 2a: The relationship between employee satisfaction with WFH and employee self-perceived productivity will be moderated by occupation and sector of employment, such that this relationship will be weakest for those employed in occupations requiring a great deal of personal interaction (e.g., services, education, sales) than for those employed elsewhere.

Hunton and Harmon (2004) suggested that the factors discouraging telework include the psychosocial meaning of joining others at a physical workplace. In their reviews of the role of office design in employees’ motivation, Thompson et al. (2022) and Marques de Macêdo et al. (2020) argued that the office is often a place for collaboration, inspiration, and creation and not just a space where people do the job. Thus,

Hypothesis 2b: Employee satisfaction with WFH and employee self-perceived productivity will be moderated by the type of office space, such that this relationship will be weakest when office spaces are designed to motivate and inspire meaningful interactions.

Gentilin (2021) asserted that one of the characteristics of the knowledge economy is the increase in the number of occupations that provide considerable flexibility to their incumbents, primarily in occupations that require high skills and education. Indeed, evidence gathered during the COVID-19 pandemic suggests that telework is most likely to exist in developed economies (e.g., 37% of total jobs in the US), whereas it is less frequent in emerging economies (Berniell & Fernandez, 2021; Saltiel, 2020). Berniell and Fernandez (2021) investigated teleworkable jobs in Latin America. They concluded that most remote jobs are in large urban areas. The ratio of teleworkable jobs in these urban areas often exceeds 20%, thus approaching the ratio of teleworkable jobs found in developed countries. The higher industrialization of urban areas likely explains this figure, owing to the availability of suitable technology and telecommunications infrastructure in urban areas and the higher density of corporate headquarters. Consider, for instance, that approximately 44% of the Latin American population does not enjoy internet connectivity (Palomino Pichihua & Ruiz Sánchez, 2023) and that most of this population lives outside urban areas. The jobs that best enable WFH in Latin America are understandably concentrated in highly populated urban areas. Specifically, jobs in functional areas such as administration, marketing, and indirect services that do not necessarily require in-person interactions (López-Igual & Rodríguez-Modroño, 2020) seem most “teleworkable.” In contrast, the degree of job control in hands-on functional areas related to production, operations, supply chain, and logistics appears lower when working remotely. According to Karasek’s (1979) job demands-control theory, this reduction in perceived control will attenuate the relationship between job satisfaction and perceived productivity because even though employees might be satisfied with the comfort of working from home, they might feel that distance impedes a series of interactions and activities closely tied to being productive on their job. Therefore,

Hypothesis 2c: The relationship between employee satisfaction with WFH and employee self-perceived productivity will be moderated by the functional area of the job, such that this relationship will be stronger for jobs in functional areas such as administration, marketing and indirect services that do not necessarily require in-person activity than for jobs related to production, operations, supply chain, and logistics, which are more likely to demand hands-on and in-person activity (a summary of the hypotheses can be seen in Figure 1 below.

We collected data by surveying Mexican white-collar workers with WFH experience. We employed a snowball sampling technique where trained undergraduate students contacted employed participants over 18 years old who held a social security number. Participants received an explanation of the survey’s purposes and assurances of the strict confidentiality of their responses before giving consent. The survey was hosted on the Qualtrics platform during the summer of 2023.

We measured the employees’ perceptions regarding two crucial factors while working remotely: satisfaction with WFH and self-perceived productivity. Three research faculty members, chosen for their expertise in survey development, helped us assess survey relevance and readability. We initially consulted each expert separately, incorporating their feedback to eliminate or refine the wording of specific items. In the second phase, the experts convened to discuss the items collaboratively until they reached a consensus. The procedure generated 12 items: seven assessing job satisfaction and five assessing self-perceived productivity while WFH. We employed a 5-point Likert-type scale ranging from 1 = strongly disagree to 5 = strongly agree to evaluate the items. The Appendix lists the final 12 items.

The survey gathered 20 additional items: one integrity check, six demographic questions, one item assessing the preference for working from home, three related to worker location, four concerning commuting experiences, and five regarding the nature of the job.

To ensure that results were not contaminated by careless responding, participants who failed an integrity-check item, completed the survey too quickly (2 standard deviations below the mean), or showed overly consistent patterns in their responses were excluded. As a result of this screening process, we reduced the final sample size to 426 participants from 439 initial responses. Table 1 summarizes the survey respondents’ demographics.

We further conducted confirmatory factor analysis (CFA) to check our measurement model’s discriminant validity. Our data failed to pass the Doornik-Hansen multivariate normality test (p-value = .00), so we conducted CFA using the Satorra-Bentler-adjusted goodness-of-fit indices. The computed values for the root mean square error of approximation (RMSEA = .04), the standardized root mean square residual (SRMR = .027), and the comparative fit index (CFI = .979) were adjusted to account for the effects of non-normally distributed data, all of which exceeded the recommended thresholds (Hair et al., 2019). Following Henseler et al. (2015), we used the heterotrait-monotrait (HTMT) ratio to assess discriminant validity. The estimated HTMT ratio (.27) remained far below the conventional threshold of .90, thus supporting the discriminant validity of the underlying constructs.

After validating the two critical variables in our analysis—job satisfaction with remote work (SATISF) and self-perceived productivity while working remotely (PRODUCT)—we applied a linear regression model to predict their corresponding factor scores. Subsequently, we normalized these predicted values to a range between 0 (minimum) and 100 (maximum). The primary goals of normalization were twofold: to enhance the interpretability of the predicted factors and to prevent the necessity of centering the variables during moderation analysis.

After checking our measurement model, we conducted a moderation analysis to test our hypotheses. Moderation analysis aimed to determine whether the relationship between job satisfaction with remote work (SATISF) and self-perceived productivity while working remotely (PRODUCT) was influenced by moderator factors (Z). Equation 1 shows the regression model used to test our hypothesis.

In Equation (1), PRODUCT is the dependent variable, SATISF is the independent variable, and Z_i represents the moderator variables depicted in Figure 1. According to Dawson (2013), the coefficients β₁ and β₂ indicate the individual main effects of SATISF and the moderator variable (Z) on the dependent variable, whereas β₃ denotes the moderation effect. Therefore, β₃ is the critical coefficient representing the two-way interaction (SATISF*Zi) effect. Thus, determining the shape, magnitude, and statistical significance of β₃ was the primary focus of the analyses.

To avoid biased coefficient estimates owing to endogeneity, we included an indicator variable in the regression to control for the respondents’ self-revealed preference for work from home (WFH), as it is well known that omitted variables correlated with the error term (µ) are a source of endogeneity. This variable comprises three groups: the first captures those who preferred WFH to hybrid work (the second category) or in-situ work (the third category). Also, we included gender as a control variable. As running the regression analysis requires omitting one group to avoid perfect multicollinearity, we omitted the first (now the baseline comparison group). Thus, the coefficients should be interpreted relative to the group preferring WFH and to the group of male respondents. The following section describes how we tested each one of the moderator variables (Z_i) included in our hypotheses (e.g., type of remote work location, job characteristics).

Table 3 summarizes the moderation regression results for our first set of hypotheses. All three models had an adjusted R-squared value above .70, indicating that the model explained at least 70% of the variance in the outcome productivity variable. We observed positive and statistically significant relationships with SATISF coefficients across all models. Again, the self-revealed preference variable consistently showed negative and highly significant coefficients across the three models. In line with the expectations, the negative impact of remote work was more significant for those who prefer in-person work than those who prefer hybrid work.

Despite our cross-sectional, all-self-report design, the statistically significant moderators uncovered in our study are hardly explained by method variance, which is unlikely to inflate this non-linear relationship (Crampton & Wagner, 1994; Saunders, 1956). However, future research should capture longitudinal trends in the relationship between satisfaction with WFH and productivity over time. While longitudinal trend analysis is not always practically feasible when conducting organizational research, future studies should utilize cross-lagged, or panel designs, to monitor and further validate the evolution of the findings reported here. Theorists have posited that teleworkers craft their work and nonwork roles to satisfy their needs while enhancing job outcomes over time (Biron et al., 2023). For instance, prolonged social isolation might reduce satisfaction with WFH, hence augmenting the employee’s desire to return to the office and interact with others, even when such interactions do not necessarily make the employee more productive. Conversely, teleworkers might become more productive as they learn to master WFH technology to conduct online interactions that are as consequential in terms of productivity as face-to-face interactions can be. We advocate for further research on how to use technological innovations that augment the effectiveness of remote, electronic interactions, thereby attenuating teleworkers’ perceived productivity limitations while WFH.

Examining the effects of third variables on the employee satisfaction with WFH-productivity relationship is crucial to thoroughly understanding how WFH influences individuals, groups, and organizations. Our research highlighted the influence of contextual factors, including commuting time and distance, that consume plenty of scarce personal resources and thus make satisfied teleworkers feel productive. Differences in the availability, comfort, and cost of commuting, including public transportation, will likely alter the relationship under study and warrant further research.

From a practical standpoint, organizations concerned with work-life balance should remain open to WFH arrangements, but probably not at the expense of setting employees up for failure in a WFH arrangement where they see themselves unable to be productive. Future research should explore reliable and valid ways to inquire about teleworkers’ honest perceptions of their ability to remain productive while WFH. Questions might explicitly inquire about the need for facilitating conditions reported to make WFH more effective in prior research (e.g., instrumental and emotional support, temporal flexibility; Ahmad et al., 2022), which might also strengthen the relationship between satisfaction with WFH and self-perceived productivity. Research on how to best conduct these inquiries seems particularly important when those satisfied with their WFH arrangement are understandably afraid of losing their WFH privileges by being called back to work at the corporate office if they express their honest views on how WFH limits their control over tasks that increase their productivity.