Our analysis relies on the Multinational Time Use Study (MTUS) data set, coordinated by the Centre for Time Use Research (CTUR) at University College, London, and incorporated in the Integrated Public Use Microdata Series (IPUMS) of the Institute for Social Research and Data Innovation of the University of Minnesota (Fisher et al., 2019). The MTUS is a dataset aimed at harmonizing time-use surveys worldwide and includes harmonized information on 69 activities performed by individuals during the day, for randomly selected samples from 25 countries over 5 decades, including travel activities. In addition, the MTUS collects information on individual and family-level socio-demographics. Data is gathered via completion of personal diaries as well as individual and household questionnaires.
There is a growing literature using time-use surveys to analyze transportation behavior (Gimenez-Nadal and Molina, 2014; 2016; Jara-Díaz and Rosales-Salas, 2015; Gimenez-Nadal et al., 2018a, 2018b, 2022; Echeverria et al., 2022). Such surveys have become the “gold standard” to study other uses of time, including paid work, housework, and childcare (Aguiar and Hurst, 2007; Gimenez-Nadal and Sevilla, 2012), as the analysis derived from time use data yields more reliable and accurate estimates in comparison to time use information obtained from stylized questions (Robinson and Godbey 1985; Juster and Stafford 1985).
We select those countries from the MTUS with available information on travel to/from work for both members of the dual-earner couple. The MTUS currently includes four developed countries whose time use surveys include that information, for the 2000s and 2010s: Spain (2002-2003 and 2009-2010), Italy (2002-2003 and 2008-2009), Korea (2009), and the United Kingdom (2000-2001 and 2014-2015).[4] We restrict the analysis to heterosexual working couples aged between 21 and 65 years old (Aguiar and Hurst, 2007, Gimenez-Nadal and Sevilla, 2012) and with both members reporting positive time devoted to commuting during working days (Gimenez-Nadal and Molina, 2019; Gimenez-Nadal et al., 2018a, 2018b; Molina et al., 2020). The final sample is 6,751 couples (see Table A.1 for sample composition by country).
Our analysis focuses on two sets of dependent variables. First, we are interested in the commuting times of each member of the couple, defined as the minutes per day that they devote to travel to/from work (i.e., Commuting time).[5] Second, we are interested in the proportion of commuting time by private transport (car, truck, or motorbike), public transport, and active transport (walking and cycling). To construct these variables, defined at the individual level, we sum for each individual the commuting time (in minutes) by each mode of transport, and divide it by the total time spent in commuting during the day by the individual.
Table 1 shows the time devoted to daily commuting and the proportion of that time done by private, public, and active modes of transport, by country and gender. We also report the results from statistical mean tests of gender differences (t-type test of equality of means, where H0 refers to the equality of means for men and women). We observe that, on average, men commute for longer times than do women, in all countries of the sample. The largest differences in commuting times between men and women are found for the UK (17.9 minutes), followed by Italy (8.6 minutes), Spain (7.1 minutes) and Korea (4.6 minutes). These differences are statistically significant at the 1% level, indicating that men in double-earner couples commute about 37% more than their female partners in the UK, 16% more in Italy, 14% more in Spain, and 8% more in Korea, respectively.
Regarding the mode of transport used for commuting, we observe important and cross-country-consistent differences by gender. For all countries of our sample, men spent on average a larger proportion of their commuting time using private transport (ranging from 73.5% in Spain to 79.8% in the UK), while women ravel a larger fraction by public transport (ranging from 6.4% in Italy to 14.1% in Spain) and active transport (ranging from 20% to41.8% in the UK). Further, gender differences are larger in the proportion of commuting by private car, followed by differences in active transport, while gender differences in the use of public transport are much smaller. In turn, these differences are consistently larger for Korea, followed by Spain, Italy, and the UK, irrespective of the mode of transport. Mean differences are statistically significant at the 1% level, except for the proportion of commuting time by public transport in the UK.
To account for the observed heterogeneity across individuals and couples, we include individual and household characteristics, including the number of children in the family. For personal characteristics, we consider age and the highest level of formal education achieved (primary education/uncompleted secondary, completed secondary, and higher education). Roberts and Taylor (2016) find that younger, more educated men and women in the UK commute longer, but the gradient is steeper for men (women) in the case of education (age). We also incorporate the hours of paid work per week and occupational category (Schwanen and Dijst, 2002; Gutierrez-i-Puigarnau and van Ommeren, 2010; Gimenez-Nadal and Molina, 2014), which has been found to be significantly related to the commuting distance gap in German couples (Chidambaram and Scheiner, 2020).
We include the number of children under age 18 in the household to proxy for household responsibilities, following Hanson and Johnston (1985), Johnston-Anumonwo (1992), Lee and McDonald (2003), McQuaid and Chen (2012) and Roberts and Taylor (2016). Children and childcare activities may impose differential constraints on the commuting of partners, and are affected by the opportunity cost of time. Evidence on the relationship between the presence of children and commuting is mixed. Some studies find that having a child is not a significant factor in commuting distance for men or women (Kim, 1994, 1995). In addition, having children is not related to the gender gap in commuting distance in the US (Chidambaram and Scheiner, 2020). In contrast, one study for the US suggests that increases in the number of children are likely to lead to longer commuting (total) distances for two-earner couples, while likely to decrease the share of women’s commuting distance in total household commuting (Kwon and Akar, 2021). In the case of the UK, having children bears no relationship to men’s commuting times but is associated with shorter commuting time for women (Roberts and Taylor, 2016).
We include a set of household characteristics, such as the residential location (urban/suburban or rural/semi-rural), ownership of a home (own outright, mortgage, or rent) and ownership of at least one motorized vehicle (either car or motorcycle)[6]. Evidence for the US (Plaut, 2006) and the UK (Roberts and Taylor, 2016) shows that owners commute further than renters, reflecting rigidities in the housing market. Car ownership is associated with shorter commuting times for women in the case of renters, and to longer trips for men homeowners in the US (Plaut, 2006), while the number of cars in the household is associated with shorter commuting times for women in the UK (Roberts and Taylor, 2016). However, in terms of the gender gap in commuting between partners in Germany, car availability is not related to commuting distance. In addition, there is no indication of greater gender equality in commuting in urban areas (Chidambaram and Scheiner, 2020).
Panel (A) of Table 2 reports summary statistics of the socio-demographic characteristics by gender.[7] Men are, on average, 42.7 to 45.8 years old, depending on the country, and are slightly older than women (40.6 to 42.8 years old). There is a larger proportion of men (women) with primary (or uncompleted secondary) education in Spain, Italy and the UK (Korea). In Spain (Italy and Korea), the proportion of men (women) with secondary education is larger, while in the UK it is of the same order. In Spain, Italy, and the UK (Korea), there is a larger proportion of women (men) with higher education. In all countries the proportion of men and women who only achieved a primary educational level is the lowest. In Spain (Italy) there is a larger fraction of individuals with higher (secondary) education, while the distribution of individuals across secondary and higher levels of education is quite similar in the UK. Men work more hours per week than do women. Gender differences in paid work are the greatest in the UK (11.6 more hours per week), followed by Korea (7.1 hours), and Spain (5.9 hours).
Panel (B) of Table 2 reports the proportion of couples without children and with one, two, or more than two children. In Spain, Italy, and Korea approximately 40% of couples do not have children, while that figure is 50% in the case of the UK. ISpain and Italy have a larger proportion of couples with one child (31% and 34%, respectively) than with two children (26% and 23%, respectively). Korea and the UK have a larger proportion of couples with two children (35% and 24%, respectively) than with one child (21% for both countries). In all countries, the percentage of couples with more than two children is comparatively low (between 4% and 6%).
Panel (C) of Table 2 shows that Italy has a lower proportion of couples living in urban/suburban areas in Italy (62%), followed by Spain (72%) and Korea (94%). At the same time, 68% of couples in the Korean sample are home-owners, while this is 78% in Italy and 84% in the UK. In addition, almost all Italian couples in the sample own at least one motorized vehicle (99%), while only 60% of couples in the UK are owners of a car or a motorcycle.
4We do not consider time use surveys in previous decades (e.g., 1980s, 1990s) because we want to give an up-to-date view of commuting behavior within couples.
5Because information on travel distance is not available in the MTUS, we rely on information regarding individual commuting times.
6 Information on ownership of a home and motorized vehicles is not available for Spain (2009-2010); ownership of motorized vehicles is not available for Korea, while urban location is not available for the UK (2014-2015).
7 Information on working hours is not available for Italy (2002-2003).