1 Introduction

Partners tend to coordinate some of their temporal and spatial decisions to optimize resource allocation and maximize their material and non-material well-being (Skinner, 2005; Van Klaveren et al., 2013). For instance, married couples achieve desirable saving and retirement objectives by coordinating their financial actions (Warren, 2013). They also synchronize their work schedules because they value their time together (Carriero et al., 2009). Such a temporal adjustment allows them to enjoy joint consumption and shared leisure activities as essential sources of union gains (Blau & Winkler, 2018). Several factors influence couples’ tendencies towards such coordination. For instance, the similarity of preferences between partners, their educational background, and their marital status play significant roles in shaping the level of coordination observed within a relationship (Van Klaveren & Van Den Brink, 2007; Voorpostel et al., 2010; Ashton, 2021). In this article we argue that apart from when and how much to work (i.e. synchronization of working schedules), partners may also engage in the synchronization (harmonization) of the physical location of their work. This represents an additional avenue for them to spend time together and foster a sense of togetherness. Recent COVID-19 pandemic-induced expansions in access to home-based work could potentially enable working couples to harmonies their home-based working arrangements and maximize their joint consumption to a much larger extent than prior to the pandemic. At the same time, however, the pandemic has led to an increase in childcare responsibilities due to the unprecedented disruptions in institutional childcare and education it brought about (Lyttelton et al., 2022). These changes, along with potential limitations in home workspace, may discourage parents from simultaneously working from home (Craig, 2020). Therefore, the question of whether parents’ working arrangements were synchronized – or desynchronized –is not a trivial one. Throughout this study, “synchronized working arrangements” refers to occasions when both parents either simultaneously work from home or attend workplace.

While a few studies (e.g., Genadek et al., 2020; Hamermesh, 2020) have examined how couples have managed their time during the pandemic, there is a significant gap in the literature regarding whether parents’ working arrangements were coordinated or uncoordinated during the Covid-19 pandemic. However, researchers from multiple disciplines have contributed to a growing body of literature aiming to understand the impact of the Covid-19 pandemic on households’ livelihoods and work-life dynamics (Cheng et al., 2021; Huebener et al., 2021; Nayal et al., 2022; Shafer et al., 2020; Vaziri et al., 2020). Nevertheless, the consequences of widespread telecommuting access and couples’ adaptation strategies to overcome pandemic-related challenges remain largely unexplored. While some studies have focused on families’ responses to policies, others have assessed the outcomes of coping with the pandemic crisis (Huebener et al., 2021; Mariani et al., 2020; Vaziri et al., 2020). A number of researchers have also examined parental responses and partner dynamics (Patrick et al., 2020; Lyttelton etal., 2020). Nonetheless, the degree to which partners coordinate actions to conform to the “new normal” and overcome pandemic-induced predicaments have been investigated rarely.Footnote 1

In this article, we seek to fill this research gap and examine whether parents’ working arrangements were synchronized or desynchronized during the Covid-19 pandemic. We also explore the role of some relevant sociodemographic factors that could have contributed to (de)synchronization. This is the first study of its kind to investigate parental coordination of work arrangements among dual earners.

Our study focuses on the work arrangements of heterosexual working parents with dependent children (the youngest child under 12) from a couple’s perspective. We approach the issue of (de)synchronization in three different ways using Familydemic Harmonized Dataset (FHD) data with retrospective information on the working arrangements of parents and their partners from March 2020 to May 2021 collected in Canada, Germany, Italy, Poland, Sweden, and the United States (US). First, we investigate whether partners’ transitions between different working arrangements—i.e., from working from home (WFH) to working at the workplace (WAW)—were synchronized or desynchronized, i.e. whether partners rather entered/exited WFH/WAW simultaneously or rather asynchronously. Second, we explore the role of several potential explanatory factors, such as the number and age of children, couples’ relationship status, and partners’ educational background similarities in couples’ transitions from desynchronized to synchronized episodes. Third, we assess the degree to which these factors influence the length of both desynchronized and synchronized spells. For all these analyses, we also consider the role of COVID-19 confinement policies, as e.g., simultaneous entries of partners to and exits from WFH could—to some extent—be attributed to changes in the stringency of these policies.

2 Theoretical Background and Hypotheses

2.1 Work Arrangement Coordination

The economic theories of marriage have attempted to explain temporal coordination among dual earners. These theories originally postulated specialization as the primary source of marital gains (Mansour & McKinnish, 2014; Van Klaveren & Van Den Brink, 2007). As Becker (1981) described, in a society with traditionally defined gender roles and values, men and women were attracted to each other and could generate conjugal gains based on their traditional, gendered specialization and exchanging them. However, rising female wages increased women’s labor market participation, and technological changes have gradually diminished such gains, at least for some partners (Blau & Winkler, 2018; Lundberg & Pollak, 2007; Stevenson & Wolfers, 2007). Instead, joint consumption of household goods and non-work/leisure time have emerged as the primary source of gains from marriage and cohabitation (Blau & Winkler, 2018; Lam, 1988). These gains are not marriage-specific, and most of them (i.e., complementarity in production and consumption) are also available to cohabiting partners (Blau & Winkler, 2018; Lundberg et al., 2016).

The economics literature has already empirically shown that working partners tend to synchronize their working hours (i.e., when and how much they work) as much as possible to maximize joint consumption (Hallberg, 2003; Carriero et al., 2009; Mansour & McKinnish, 2014; Jenkins and Osborn, 2004). This synchronicity enables them to maximize their satisfaction (utility) from spending their leisure time (i.e., non-work time) together.

It has also been shown that schedule synchronicity generates positive outcomes for families as it diminishes time poverty (Brannen et al., 2013). According to this body of literature, the ability to synchronize leisure time constitutes one of the pillars of time wealth, a state that enables individuals to “satisfactorily coordinate different temporal requirements” (Geiger et al., 2021; Mullens & Glorieux, 2023). As Reisch claimed (2001), synchronous leisure time significantly enhances well-being by increasing attention and involvement, both of which require time. She also argued that in addition to the quantity of non-work time, the extent to which we can synchronize leisure determines our quality of life. Temporal harmonization can also produce positive spillovers at the social level. In particular, Smith and McBride (2021) showed that, in the face of time scarcity, synchronized leisure could diminish time inequalities, potentially generating desirable outcomes for family life. Research also suggests that married individuals experience an increase in life satisfaction when they spend more time with their spouse. Hamermesh’s (2020) findings indicated that for married couples, any decrease in satisfaction caused by pandemic-induced limitations on personal freedom and income was partially offset by spending more time with one’s spouse.

We argue that partners may strive to maximize their joint consumption/leisure time spent together, not only by synchronizing their working hours but also by synchronizing (harmonizing) home-based working arrangements. When both partners harmonize the time they spend at home while working, they may enjoy improved schedule control and spend together the time they have saved by not commuting—for example, preparing and sharing a meal. This leads to higher marital/companionship gains from joint consumption. Americans allocated almost 60 percent of the time they saved not commuting to activities other than their primary jobs during the COVID-19 pandemic (Barrero et al., 2020). Furthermore, it is worth noting that individuals’ overall satisfaction and utility, in addition to joint consumption, are influenced by the specific individuals with whom they spend their time. In the context of pandemic lockdowns, married individuals tended to experience higher levels of well-being due to increased time spent with their most significant and "utility-enhancing" companions (Hamermesh, 2020).

This leads us to the formulation of our first and primary hypothesis:

H1

Parents in dual worker couples tend to synchronize their home-based working arrangements by jointly transitioning from WAW to WFH arrangements.

2.2 Factors Influencing Work Arrangement Coordination

Van Klaveren & Van Den Brink (2007) found that dual worker couples reap greater gains out of shared consumption than other couples because their preferences on consumption tend to better match those of their spouses/partners. Such harmonized preferences epitomize the positive assortative mating, where individuals with similar traits (i.e., higher education or similar occupations) are inclined to marry each other or at least enter a long-term relationship with joint commitments (e.g., having children together) (Blau & Winkler, 2018; Bryant & Zick, 2005). These couples obtain direct satisfaction from engaging in shared activities stemming from their common interests, but such levels of utility are hardly achievable alone or with individuals other than partners. This reasoning can be extended to a comparison of couples with similar educational attainments and couples with different educational levels. Thus, as Voorpostel et al. (2010) concluded one can reasonably expect that couples with similar educational attainments gain more from joint consumption and leisure than other couples. This leads us to our second hypothesis:

H2

Parents in dual worker couples with similar educational levels are more likely to synchronize their work-from-home arrangements than couples with different educational attainments.

It has also been argued that married couples (couples in registered unions) prefer togetherness and a joint lifestyle more than their cohabiting counterparts in non-registered unions (Ashton, 2021; Kalmijn & Bernasco, 2001). Approaching the issue from a psychological perspective, Luo (2017) attributed the preference for togetherness to couples’ original shared traits (i.e., positive assortative mating) and similarities developed thanks to converging preferences realized throughout their relationship. Registered unions are characterized as more stable relationships than non-registered ones (Musick & Michelmore, 2018; Sassler & Lichter, 2020). Thus, long-term couples such as married and registered cohabiting couples appear to maintain and expand their initial homogamy. Such a mechanism enhances couples’ emotional experiences, brings them closer together, and provides them with marital satisfaction (Gonzaga et al., 2007; Murray et al., 2002). Therefore, the more persistent the relationship, the stronger the desire will be for shared activities, togetherness, and temporal-spatial harmonization, including synchronization of working arrangements. Based on these arguments we formulate our third hypothesis:

H3

Working parents in registered unions are more likely to synchronize their home-based work arrangements than working couples in non-registered unions.

With all this said, parental coordination can sometimes lead to temporal desynchronization for reasons other than the maximization of utility through joint consumption. For instance, as Barnet-Verzat et al. found (2011), limited access to full-time childcare, high childcare costs, and living in societies with market-based childcare systems may push parents toward time and spatial desynchronization. Southerton (2020) argued that these desynchronized choices are often crucial for mothers’ participation in the labor force. Cheng et al. (2021) meanwhile, noted that COVID-19 pandemic-related policies—school closures, the introduction of online learning, social distancing orders, and other restrictive policies that disorganized childcare to an unprecedented scale—dramatically increased the need for childcare. They may also have limited parents’ ability to harmonize spatially (i.e., jointly working from the office or jointly working from home). In particular, couples with dependent children needed to secure parental supervision at home by at least one of the parents, and may therefore have found it more difficult than parents with older children to coordinate simultaneous work from the office or other workplace (Del Boca et al., 2020). For larger families, joint work from home might also have proved untenable due to a lack of space for all family members to work or teach/learn online simultaneously (Craig, 2020). All this leads us to our final two hypotheses:

H4

The number of children in the family is negatively related to the parents’ synchronized work from home.

3 Data and Methods

3.1 Data and Sample

To conduct our empirical analyses, we used the Familydemic Harmonized Dataset, the brainchild of an international collaborative study that focused on family-, gender-, and work-related aspects of the COVID-19 pandemic (Kurowska et al., 2023). The survey was conducted on large random samples (minimum of 2,000 respondents per country) of parents with dependent children between June and September of 2021 in six countries: Canada, Germany, Italy, Poland, Sweden, and the United States. Details of the survey design, method and sample characteristics can be found in the data descriptor published in Scientific Data (see Kurowska et al., 2023). The multi-national nature of this study enabled us to conduct our analyses while controlling for potential country-specific contexts.

In addition to families’ essential socio-demographic characteristics, this study collected information on various topics including the division of unpaid work and responsibilities (childcare and housework), various aspects of satisfaction, and some employment-related data that were suitable for the present investigation. The FHD also recorded monthly information on parents’ working arrangements. We transformed this segment of the data into a monthly panel dataset and generated event history data on partners’ exits from/entries to each working arrangement: working from home (WFH) and working at the workplace (WAW). Monthly information on working arrangements was collected for the respondent and their partner, enabling us to create person-period data structures for both partners. The duration of each working arrangement spell was measured in months and the sequences of work arrangement data stretched from March 2020 to May 2021. We eliminated all non-heterosexual partners from our person-period data due to small subsample size. Thus, our final restructured data restricted our sample to heterosexual dual-earner partners with at least one child younger than 12 years. This sample, of working parents with young children, was the most appropriate one for our investigative objectives because it represented parents with potentially more family and work-related challenges than other parents both before and during the pandemic.

We used this sample to conduct a three-stage regression analysis. First, we employed logistic regressions in a discrete-time setting to evaluate whether working partners coordinated their transitions when they switched from WAW to WFH. Second, we utilized a multinomial recurrent event regression analysis to identify which factors encouraged or discouraged working couples’ transitions from non-synchronized arrangements to synchronized ones. Third, using random effects regressions, we investigated how various factors could be related to the duration of couples’ working arrangements. Detailed descriptions of these stages are as follows.

3.2 Transitions from WAW to WFH

Using our person-period data we constructed our dependent and primary explanatory variables. The former was a binary indicator that captures the events of interest for respondents: transition to WFH from WAW. These indicators equaled one once a given work arrangement spell ended. A similar dichotomous variable marking the same events for respondents’ partners represented our key independent variable. These events were recurrent because respondents and their partner’s history could be divided into multiple work arrangement episodes. Under such circumstances, recurrent discrete-time models serve as the most appropriate quantitative methods.

We began our regression analyses by estimating the possibility of synchronization using the following logistic event history model:

$$\log \left( {\frac{{p_{ti} }}{{1 - p_{ti} }}} \right) = \alpha PExit_{ti} + \beta x_{ti} + u_{ti}$$
(1)

where \({p}_{ti}\) was the probability of an event for the respondent (i.e., entering a home-based working arrangement), \({PExit}_{ti}\) represented the coordination indicator capturing a similar event for respondents’ partners, and \({x}_{ti}\) combined time variant and time-invariant explanatory and control variables and their interactions with our coordination indicator. In this equation, the primary focus was the coefficient on partners’ event/exit (i.e., \({PExit}_{ti}\)). The synchronization occurred whenever parents’ event indicators—both the dependent variable and the key independent variable (PExitti)—simultaneously equaled one (that is, their work arrangement transitions coincided). From an empirical perspective, Eq. 1 seemingly fitted an individual-level regression model. However, because we combined respondents’ and partners’ person-period data, the likelihood of matching incidences (both partners’ events) was estimated. This allowed us to investigate the likelihood of working couples simultaneously transitioning between work arrangements (testing H1).

Our secondary explanatory variables included the measure of partners’ educational attainment similarity, relationship status, number of children, and age of the youngest child. To capture the degree of partners’ educational level similarities, we created a binary variable of one for couples with similar educational attainment (e.g., both partners finished college or both did not) and zero otherwise. We coded the relationship status as a dichotomous variable equal to one for married/registered cohabiting couples and zero for non-registered cohabiting couples. In the case of our US sample, the dichotomy effectively segregated married couples from cohabiting couples because registered cohabitation is not formally recognized in the United States.

In our regression models, we accounted for various factors that could influence individuals’ transitions between different working arrangements, aside from the desire to spend more time together with a partner. These factors include the policy environment, such as stay-at-home orders implemented during the pandemic, as well as job-related factors. In addition, we controlled for respondents’ age, the presence of children born during the period under study, and the type of occupation based on the International Standard Classification of Occupations (ISCO-08) 1-digit code. We also incorporated the policy stringency index to capture the severity of pandemic-related restrictions. Furthermore, our regressions included control variables for the couples’ country of residence and month-specific effects to capture any time-varying factors.Footnote 2

Including occupation types as a categorical variable was crucial to controlling for highly volatile and heterogeneous work arrangements across various occupations under restrictive pandemic-related policies. The adaptability of occupations with home-based work determines employees’ likelihood of WFH (Hatayama et al., 2020). Furthermore, in order to account for the policy environment, we used the policy Stringency Index (SI) as a macro time-variant covariate. The SI was created by a group of researchers at Oxford University’s Oxford Covid-19 Government Response Tracker (OxCGRT), a composite time-varying measure representing the differentiation of COVID-19-related policies from country to country. It combined nine COVID-related aspects of the government’s response, such as school closures, workplace closures, stay-at-home requirements, and public information campaigns (Hale et al., 2021). It used a scale of 0 and 100, with 100 representing the strictest response. We included the average monthly index in our regression models and matched it with our data structure.

To investigate the degree of harmonization between both parents’ events, we employed a random effects (RE) estimator, which accounts for individual-specific heterogeneity. This statistical approach enabled us to estimate Eq. 1 and compare the RE estimates of matching events with those obtained through regular logistic regression. The use of a discrete-time method was appropriate for our data, which involved recurrent events within each observation. This methodology is commonly utilized in studies that analyses data with multiple recurrent events per observation, as is the case in our present study. (Rabe-Hesketh & Skrondal, 2022; Steele, 2011).

3.3 Couples’ Joint Transitions from Non-synchronized to Synchronized States

Our first recurrent event analysis was useful for investigating whether working parents synchronized their transitions from WAW and WFH arrangements in a particular month during the pandemic. However, it was also crucial to study the underlying determinants of couples’ transitions from non-synchronized arrangement to a synchronized one (H2–H4, in other words). To understand these relationships, we created a couple-level categorical variable that would distinguish couples’ spells of harmonized working arrangements from desynchronized ones. This categorical variable represented three types of couples’ work arrangements (CWA): non-synchronized working arrangements (NS), synchronized WFH (sWFH), and synchronized WAW (sWAW). The multinomial variable used in our analysis aimed to explore the factors that contribute to the probability of transitioning from desynchronized working arrangements to harmonized ones. This variable also automatically transformed our parson-episode data structure into a couple-episode one, allowing us to perform an additional discrete-time recurrent event investigation at the couple level. Taking our cue from Steele (2011) and Rabe-Hesketh and Skrondal (2022), we utilized a multinomial logit regression (i.e., competing risk) model with the NS state as the base category. In the context of our three-level categorical outcome, this model would analyze the probabilities of experiencing each of the three events (Schmid & Berger, 2021). In this model, we included observable family characteristics (e.g., relationship status, number of children, and the age of the youngest child), the measure of COVID-19 policy stringency, and country and month identifiers. This method’s output corresponds to the following equations:

$$\log \left( {\frac{{P\left( {state = sycnhed\; WFH} \right)}}{{P\left( {state = desynchronized} \right)}}} \right) = \alpha_{1} + \beta_{1} x_{ti} + u_{ti}$$
(2)
$$\log \left( {\frac{{P\left( {state = sycnhed \;WAW} \right)}}{{P\left( {state = desynchronized} \right)}}} \right) = \alpha_{2} + \beta_{2} x_{ti} + u_{ti}$$
(3)

where \({\beta }_{j}\) represents the vector of regression coefficients.

Unlike the first model, this multistate one included more states and observations as it incorporated alternative states/events, i.e., non-synchronized work arrangements. It also provided a parsimonious tool for analyzing how individual observable factors and household characteristics could influence couples’ probability of transitioning to a particular type of synchronized CJA without requiring interaction terms. The estimated parameters of this model identified important factors determining the relative risks of transitions to each type of synchronized work arrangement (sWFH and sWAW) relative to non-synchronized arrangements (NS) throughout the period of analysis. As with the previous model, we estimated the likelihood of transitions between competing states using the RE approach to account for unobserved individual-specific heterogeneity.

3.4 Duration of (De) Synchronized Spells

Finally, in addition to couples’ mobility between different working arrangements and the incidence of synchronized episodes, the duration of (de)synchronized spells and their determinants are also factors in understanding the (de) harmonization of working arrangements among dual-earner couples. To investigate these aspects of synchronization, we created three couple-level continuous variables measuring the durations of couples’ synchronized and desynchronized working arrangements in months. This enabled us to calculate the average total duration (in months) that couples spent in sWAW, sWFH and NS during the entire period of analysis. We also used these measures as dependent variables in three RE regression models to explore the extent to which observable family and individual characteristics (e.g., relationship status, number of children, and the age of the youngest child) could determine the length of (de)synchronized episodes. We separately modeled the relationships between couples’ characteristics and the duration of these spells while controlling for policy stringency, occupations, country-specific differences and time trends by including appropriate indicators.

4 Results

4.1 Sample Characteristics

Table 1 (below) shows the weighted averages of a few of our sample’s characteristics by country. Our American sample was the youngest, with an average age of nearly 36 years. Among all working couples, American and Swedish parents reported the highest average numbers of children, while the youngest population belonged to the Canadian sample. Regarding marital status, about 11 percent of Americans and more than 30 percent of Swedish couples cohabited, respectively. The prevalence of similar educational levels was higher among American and Italian dual earners. Moreover, we presented our descriptive findings on the prevalence of working from home (WFH) among respondents, which was found to be the highest in Canada and lowest in Poland. Additionally, we analyzed the prevalence of synchronized WFH episodes and discovered that North American countries had a higher prevalence compared to their European counterparts. The percentage of working couples who experienced at least one month of synchronized working arrangement slightly varied across countries. For example, the corresponding shares for Canadian and Italian couples were nearly 63 and 75 percent. Finally, the country-level average duration of any synchronized work arrangement episodes ranged roughly from six to eight months.

Table 1 Weighted means of working couples’ selected characteristics
Full size table

4.2 Transitions from WAW to WFH—Regression Results

Table 2 summarizes pooled logistic and RE analyses of parents’ exits from WAW to WFH. As stated earlier, we were interested in the coefficient on partner’s events, particularly its sign and statistical significance. We empirically tested whether there was a statistical difference between our pooled logit and RE estimates, and listed corresponding P-value at the bottom of Column 2. This test result implied that controlling for individual heterogeneity was necessary and we should rely on our RE results (Column 2). The results of our regression models suggested that a positive relationship existed between partners’ exits from onsite work (i.e., WAW) to home-based work. The pooled logit and RE estimates of the coefficient on the partner’s event(s) from our base model were positive and statistically different from zero.

Table 2 Pooled logistic and RE estimates of parents’ transitions from WAW to WFH
Full size table

Regarding characteristics affecting individuals’ transitions to WFH, compatible educational levels, marital status, and having a newborn child were positively associated with the likelihood of such transitions. Having a partner with a similar educational attainment, being in a marital or registered partnership, and having a newly born child increased the likelihood of switching from WAW to WFH. However, these transitions seemingly were not correlated with the number of children, the age of the youngest child, and individual’s age. Moreover, among all working parents, European working parents were less inclined to transition from workplace to home-based working in comparison to Canadian dual-earner parents. These regression models also included month dummies and other control variables such as the policy stringency index, and occupation type, none of which have been shown in the tables for the sake of brevity. In separate models, which are not presented here, we examined potential country-level differences by interacting the partner’s exit identifier with binary indicators for different countries. Our findings indicated that the incidence of synchronized transitions to WFH was only more prominent among US couples compared to Canadian ones.

To assess the robustness of our main finding regarding working parents’ synchronized transitions to home-based work arrangements from WAW, we conducted an extended regression analysis. This analysis incorporated additional control variables to account for potential confounding factors. The inclusion of these control variables was based on existing empirical literature on the COVID-19 pandemic, which has established associations between working from home (WFH) and factors such as restrictive policies, childcare responsibilities, and relationship dynamics (Del Boca et al., 2020; Alon et al., 2020; Schmid et al., 2021). In particular, we included measures related to the division of childcare responsibilities within the household and changes in the quality of partners’ relationship since the onset of the pandemic. The former represented the extent to which the division of childcare was relatively equal or relatively unequal. A value of 1 indicated a relatively equal division of childcare responsibilities, while a value of 0 indicated a relatively unequal division. The latter was characterized by a categorical variable that captured three categories: deterioration, no change, and improvement.

The objective of including these additional variables was to ensure that the observed matched events were not driven by other factors such as pandemic restrictions, childcare division, or changes in the quality of partners’ relationship. To properly account for these relationships, we introduced interaction terms between these covariates, the policy stringency index, and the key binary independent variable representing the transition to home-based work (PExit). The crucial results of these extended regression models, including coefficients representing the main effects and interaction terms, are presented in Table 3. Notably, the positive coefficient associated with partners’ events (main effect) remained statistically significant and different from zero in these models,Footnote 3.Footnote 4

Table 3 Extended pooled logistic and RE models with interactions for robustness purposes
Full size table

4.3 Couples’ Joint Transitions from NS to sWFH and sWAW—Regression Results

The RE multinomial regression results for couples’ transitions to sWFH and sWAW from NS arrangements are listed in Table 4. The results observed in Column 1 offer valuable insights into our analysis, specifically addressing the empirical tests conducted to evaluate our secondary hypotheses, H2 to H4. The estimated coefficient associated with the number of children was positive, indicating a positive relationship. However, it is important to note that this coefficient was not statistically different from zero. Furthermore, the analysis revealed a negative association between couples’ transitions to sWFH states and the age of the youngest child. This implies that as the youngest child’s age increases, couples are less likely to transition to synchronized WFH arrangements. The estimation results also revealed that dual earners who shared similar levels of educational attainment exhibited a higher likelihood of exiting NS states and entering concurrent WFH spells. As shown in Column 1, being married or in a registered cohabiting relationship relatively increased the chance of transitioning to sWFH from NS arrangements. According to our results listed in Column 2, only couples’ likelihood of joint transition from NS to sWAW states decreased with age. The coefficient on the indicator of a newborn child was statistically different from zero in Eq. (3), with a negative sign denoting those couples who had a child during the period under analysis but were less likely to switch to sWAW from NS. To ensure the robustness of our key findings, we conducted further analysis by re-estimating our multinomial model with the inclusion of additional control variables, such as measures of childcare responsibilities and relationship quality. The results of this extended specification exhibited minimal differences compared to our original findings.

Table 4 RE multinomial logistic estimates of the determinants of dual-earners’ transitions from NS to sWFH and sWAW
Full size table

4.4 Duration of (De) Synchronized Spells—Regression Results

Finally, Table 5 contains the random-effects results on the duration of synchronized and desynchronized working arrangements. The presence of one additional child in the household appears to be associated with nearly a six-day decrease in couples’ contemporaneous WFH arrangement. At the same time, the synchronized period of WAW extended as the age of the youngest child increased. The duration of synchronized spells appeared to be longer among dual earners with similar educational levels. However, after engaging in a contemporaneous WAW arrangement, married and registered cohabiting couples exited this arrangement earlier than their non-registered cohabiting counterparts. According to these results, dual earners increased their synchronized home-based work arrangement by almost two days for an incremental change in age. As expected, the policy stringency measure was positively associated with the continuity of sWFH episodes and negatively correlated with the length of sWAW episodes. Nonetheless, the coefficient on the stringency index was not statistically different from zero in the sWAW model (Column 2).

Table 5 RE estimates of determinants of (de)synchronized work arrangements continuity
Full size table

Regarding non-synchronized spells, older couples tended to spend longer in such arrangements, which decreased along with the age of youngest child. In contrast to the length of harmonized working arrangements, the duration of NS episodes was shorter among working parents with similar educational level attainments than those with dissimilar ones. Finally, in the NS equation, the coefficient on the stringency index was positive, but it was not statistically different from zero.

5 Discussion

The recent shift to telecommuting and home-based work accelerated by the COVID-19 pandemic will likely endure. Numerous studies have examined a variety of the impacts flexible working arrangements have had on family life (Kim, 2020). So far studies have shown that partners tend to synchronize their working schedules (i.e. when and how much they work). As an unprecedented exploration in this domain, our research sheds light on the phenomenon of synchronized work arrangements among working parents. We add to this literature by investigating whether they synchronize from where they work. Our study also investigates how couple- and child-related characteristics moderate the coordination process of working arrangements. Furthermore, the methodological approach adopted in this study is unique and robust. Using Familydemic Harmonized Data from a survey conducted on large samples of parents in six countries (US, Canada, Sweden, Germany, Italy, and Poland) in mid-2021, we examined the working arrangement dynamics of dual-earner parents between March 2020 and May 2021.

Relying on theories suggesting that dual earners achieve more marital/union gains through joint consumption than specialization (e.g., Mansour & McKinnish’s work (2014)), we expected that couples would tend to synchronize their home-based working arrangements in order to be able to spend more time together. Indeed, our analysis, taking into account variations in the stringency of Covid-19 policies, revealed a notable pattern: when one parent made the transition from working at the workplace to home-based work, their partner also was more likely to make a synchronous transition to the same working arrangement. This tendency appeared to be even stronger among couples with similar educational levels than those couples whose educational attainments were different. Dual-earner parents with matching educational levels were also inclined to remain in their synchronized WFH arrangements for relatively longer spans of time than dual earners with non-matching educational attainments. These effects can be attributed to the correspondence of partners’ preferences.

Existing scholarly literature has posited that the benefits derived from shared consumption and leisure activities are maximized when partners exhibit a high level of concordance in their preferences (Voorpostel et al., 2010). Building upon this theoretical framework, we postulated that parents in married or registered unions would display a greater propensity to synchronize their home-based working arrangements compared to couples in non-registered unions. In our study, we found support for this hypothesis. Our results indicated that when exiting non-synchronized arrangements, married and cohabiting parents in other types of registered unions were more likely to transition to harmonized work from home (WFH) than non-registered cohabiting couples. This finding aligns with existing literature that emphasizes a stronger preference for togetherness and a shared lifestyle among married couples compared to cohabiting couples (Ashton, 2021; Kalmijn & Bernasco, 2001).

Our analyses also partially supported our expectations regarding the role of the number of children in couples’ working arrangement dynamics. Our most prominent finding on child(ren)’s characteristics relates to the time working parents spend in a given synchronized work arrangement. The higher the number of children, the shorter the harmonized home-based work arrangement. We suppose that in the case of larger families, congestion and a lack of space might have encouraged working parents to shorten their synchronized WFH arrangements. During the analysis period (the first year of the Covid-19 pandemic) many children were kept at home due to childcare closures and schools being moved online, thus limiting their parent’s opportunities to work from home at the same time (Lyttelton et al., 2022).

Our empirical analysis revealed a significant relationship between the age of the youngest child and the likelihood of parents simultaneously transitioning to work-from-home (WFH) arrangements. Specifically, as the age of the youngest child increased, the probability of parents making synchronized transitions to WFH arrangements decreased. This finding aligns with existing literature that emphasizes the impact of children’s age on parental leisure and work coordination (Barnet-Verzat et al., 2011). Moreover, we observed a noteworthy pattern concerning the duration of time when both parents worked outside of the home versus when only one parent worked outside. As the youngest child grew older, parents tended to spend more time in simultaneous work outside of the home, while the duration of time when only one parent worked outside diminished. This trend can be attributed to the evolving demands of childcare and the associated challenges parents face in planning their schedules in the presence of young children.

We also observed variations across countries in the coordination of work arrangements among parents. For instance, we found that once Canadian working parents made a simultaneous transition to WFH arrangements, they tended to remain in this state for a longer duration compared to dual earners from other countries, including the United States. These country-level differences in work arrangement coordination among parents warrant further investigation. Future studies should delve deeper into the intercountry variations and strive to identify the underlying factors that contribute to these differences. Exploring the role of cultural, institutional, and policy factors in shaping working couples’ harmonization processes across different nations would provide valuable insights and help us better understand the reasons behind these variations.

These findings offer valuable insights into the work-life preferences of parents in the post-pandemic context. As they perceive synchronized work arrangements as valuable, it becomes crucial to ascertain whether the positive inclinations of dual-earner parents toward WFH and location flexibility endure in the aftermath of the pandemic. Unsurprisingly, emerging evidence suggests that hybrid working arrangements, allowing employees to telecommute, remain popular among many (Mohammadi et al., 2023; Delbosc & Kent, 2023; Hopkins & Bardoel, 2023). The ability to synchronize work arrangements may contribute to this post-pandemic trend, providing parents with the opportunity for harmonized work arrangements. However, there is a noticeable gap in research that specifically examines how parents coordinate their work arrangements in the post-COVID-19 era. Hence, further research is needed to establish a robust connection between parents’ post-pandemic work-life preferences and work arrangements synchronization.

Like many other studies, this one is not without limitations. Some we tried to address by estimating alternative specifications with lags and leads, extending models with additional controls, and comparing pooled estimations against our RE estimations. Still, some limitations remain. For instance, information on parents’ work arrangements throughout the analysis could contain inaccuracies due to the retrospective nature of this information. Moreover, having access to days or weeks of transitions between different work arrangements could have enhanced our analytical procedures. Our empirical investigation could also have been augmented if our data contained an extended longitudinal segment. Lastly, it is important to acknowledge that the findings of our study are limited to a specific group of parents with at least one child up to the age of 12. Due to the nature of our data, we were unable to include parents having only older children. Therefore, the generalizability of our findings to parents with older children only may be limited.

It is also important to note that while we observed partners undergoing simultaneous transitions to WFH after considering the impacts of varying pandemic-related restrictive policies, our data did not enable us to account for variations in job suitability for home-based work—a crucial factor influencing joint WFH arrangements. Ideally, more comprehensive information with detailed information on job types would have allowed for a nuanced consideration of these differences. Regrettably, our dataset offered only a generalized classification of respondents’ occupations, with specific job categories (minor groups) suffering from considerable missing values, making them unsuitable for thorough analysis. This inherent constraint prevented us from completely ruling out potential interdependence between work arrangement transitions and occupations. Consequently, despite our endeavors to address this limitation, it is prudent to exercise caution when interpreting our observation regarding synchronized transitions to WFH. Regardless of the limitations, however, our study makes an important contribution to the field by shedding light on the harmonization process of working arrangements among working couples with young children as well as on some potential mechanisms underlying the harmonization. We also hope our study will stimulate more research on the harmonization of working arrangements in couples.

We believe that future research should focus on further exploring the underlying reasons for (de)synchronization of working arrangements among working couples, as well as the implications of such coordination for gender equality and families’ well-being. Parents could potentially employ the synchronization of work arrangements to bridge the gender gap in childcare and housework, particularly when they simultaneously worked from home. Working from home usually penalizes mothers by forcing them to increase their parenting time and domestic responsibilities (Kurowska, 2020). There is evidence implying that this gendered outcome intensified following the onset of the COVID-19 pandemic (Lyttelton Zang, & Musick, 2022). Nonetheless, couples working from home can share parenting tasks more effectively and equitably than those with non-synchronized work arrangements. Thus, such a setting could help the benefits of WFH (e.g., enhanced work-life balance) further materialize for working mothers.