Gender Differences in Commuting: New Evidence from Spain

In recent years, the phenomenon of commuting has been increasingly studied in the literature. The reason for this is its great impact in social and environmental terms as well as on the quality of life of citizens, since its duration and characteristics result in monetary and time use burdens, and it has a decisive influence on the balance between work and family life (Kahneman & Krueger, 2006; Kahneman et al., 2004), among other aspects.

A significant number of studies on this topic have focused on the analysis of gender differences in commuting and have generally found that women tend to have shorter commuting times than men, with evidence of a statistically significant gender gap in commuting time for a wide range of countries including the US, Ireland, Canada, Sweden, the Netherlands, and the UK (e.g. Crane & Takahashi, 2009; Hanson & Pratt, 1995; McQuaid & Chen, 2012; Mok, 2007; O’Kelly et al., 2012; Sandow & Westin, 2010a; van Ommeren & van der Straaten, 2008). Such a finding is relevant because, it may be associated with narrower job search areas, which could lead to poorer wages and other non-monetary working conditions and a possible increase in the risk of overeducation (Blau & Kahn, 2017; Le Barbanchon et al., 2021; Petrongolo & Ronchi, 2020), as a result of worse matches in the labour market (Crane & Takahashi, 2009; Ruppert et al., 2009; Van Ommeren & Rietveld, 2005; Waldfogel, 2007), or even lower labor force participation, especially in the case of married women (Black et al., 2014; Farré et al., 2020; Moreno-Maldonado, 2021).

The aim of this article is to further analyze the origin of gender differences in commuting by examining this phenomenon in the case of Spain. Previous analyses on this topic for Spain are very scarce and refer exclusively to specific areas such as the regions of Madrid, Valencia or Andalusia (Rodríguez & García, 2012; Salom & Delios, 1998; Torrado et al., 2018). This article overcomes this limitation by examining the commuting patterns of men and women, and identifying the factors that explain the potential gender differences in commuting times for the country as a whole.

In any case, the relevance of the study in the context of the related literature lies mainly in the fact that the empirical analysis is approached in a novel way using Oaxaca-Blinder econometric decomposition techniques, which are new to the field of research on gender gaps in commuting. These techniques have the advantage of decomposing the gender gap in commuting into two components: one that measures how much of it is due to the fact that men and women differ in the characteristics that determine commuting behavior and quantifies the specific contribution of each of the variables considered, and a second that captures the part of the difference that arises from the fact that both sexes have consistently different mobility patterns, even when comparing observationally similar individuals with the same characteristics. To this end, the analysis uses microdata from a Spanish nationwide representative survey, whose richness of information allows an adequate control of the numerous personal, family, geographic and work-related elements identified in the literature as potential determinants of gender differences in commuting, and provides new evidence on the origin of the gender gap in commuting. The survey provides information on commuting in terms of time, which has advantages over the alternative of measuring it in terms of distance for several reasons. The first is that the measurement of commuting based on time tends to be more refined, as it often includes time spent on ancillary activities such as finding a parking space (Bovy & Stern, 1990), whereas self-reported commuting distance often does not correspond to the actual distance traveled (Rietveld et al., 1999). The second is that travel time gives a better picture of the links between commuting, the mode of transport used and the efficiency of transport networks (Crane, 2007). Finally, it is believed that while commuting distance is more closely related to the direct economic costs of travel, commuting time is more closely linked to the opportunity cost of time spent (Alonso, 1968; Muth, 1969), which may be particularly relevant given the tighter time constraints of women (Doyle & Taylor, 2000). In this sense, it should also be noted that the specific way of measuring commuting could be particularly important in gender analyses, since the differences observed between men and women are greater in terms of commuting distance than in terms of duration (Crane, 2007; MacDonald, 1999) so that commuting time and distance offer different views of the phenomenon (Giménez-Nadal & Molina, 2016).

The decomposition analysis of the gender gap in commuting allows testing the first of the three basic hypotheses of the analysis: differences in commuting time could be explained by the different average characteristics of men and women in various dimensions related to socio-economic and family background, geographical distribution, use of different modes of transport and work-related elements. Additional disaggregated analyses allow us to test the second main hypothesis, that the gender gap in commuting may be different for certain groups characterized by more egalitarian behavior and where roles are comparatively less gendered. In this sense, three specific groups are examined separately given previous evidence from other countries on the significant influence of family responsibilities on commuting (Giménez-Nadal & Molina, 2016; Sandow & Westin, 2010a; Wheatley, 2013), the smaller gender gap in commuting observed for women with higher levels of education (Lee & McDonald, 2003; Sandow & Westin, 2010a; Vance & Hedel, 2008), and the fact that for women with partners it is often the man’s employment that determines the location of the couple’s residence, which conditions the woman’s job search (Madden, 1981; Salom & Delios, 1998). These groups are: women with higher levels of education, women without family responsibilities, and women without a partner. Finally, to test the third major hypothesis that women’s lower commuting may be due to different preferences than men’s, we also examine whether gender differences are at least partly due to differences in men’s and women’s tolerance for commuting, which should be reflected in different degrees of association between commuting and satisfaction in different domains.

In summary, the evidence obtained confirms that in Spain, as in other advanced countries, women on average have shorter commuting times than men. These unadjusted differences, although relatively small at first sight, are interestingly due to two relevant countervailing effects. The first stems from the average differences between men and women in various dimensions related to commuting time, which together tend to increase the commuting time of women. These include both the characteristics of their jobs (which are of lower quality and less stable) and the modes of transport they use to get to work (with a higher prevalence of public transport), which tend to be associated with longer commutes, with the exception of women’s greater family responsibilities, which are associated with shorter commutes. Second, when comparing men and women with the same characteristics, women systematically have shorter commutes. Taken together, this evidence suggests that there may be social or cultural conditioning factors that limit women’s commuting and thus worsen their employment prospects. This possibility is reinforced by the evidence from the disaggregated analysis that there are no distinct patterns of mobility by gender for certain groups of women (namely those with higher education, without family responsibilities, or without a partner) who are assumed to be more similar to men in their work patterns and gender role perceptions. Finally, the evidence also shows that there is no different association between commuting and satisfaction levels for men and women, suggesting that women’s shorter commuting times are not plausibly the result of choices associated with a hypothetical lower tolerance for commuting.

The article is organized as follows. After this introduction, the second section reviews the literature related to gender differences in commuting. The third section describes the database used and the methodology of the empirical analysis. The fourth section reports the findings. Finally, the last section presents the main conclusions.

The analysis of commuting patterns has received considerable attention in the literature, reflecting the fact that this type of mobility generates considerable impacts on the environment, the territory and the quality of life of individuals (Gottholmseder et al., 2009; Kahneman & Krueger, 2006; Kahneman et al., 2004; Wener et al., 2003). For all these reasons, commuting today occupies a central place in territorial planning or the provision of infrastructure and means of transport (Wener et al., 2005).

The available evidence shows that commuting patterns vary according to many individual and family factors, such as age, marital status, number of children, educational attainment, or disposable income, with gender being a particularly important factor (de la Hoz, 2008; McQuaid & Chen, 2012; Paull, 2008; Sandow & Westin, 2010a). Thus, it is observed that women’s commutes are generally comparatively shorter in both distance and time, with evidence on the existence of a statistically significant gender gap in commuting time for a large number of advanced countries, such as the US (Crane, 2007; Crane & Takahashi, 2009; Giménez-Nadal & Molina, 2016; Hanson & Johnston, 1985; Hanson & Pratt, 1995; Turner & Niemeier, 1997; White, 1986), Ireland (O’Kelly et al., 2012), Canada (Mok, 2007), Sweden (Sandow, 2008; Sandow & Westin, 2010a), the Netherlands (Gimenez-Nadal & Molina, 2014; van Ommeren & van der Straaten, 2008), and the United Kingdom (Dickerson et al., 2014; McQuaid, 2009; McQuaid & Chen, 2012; Nafilyan, 2019; Roberts et al., 2011), and on the fact that such gender differences in commuting are heterogeneous across countries (Giménez-Nadal et al., 2020a).

In this sense, whether the lower commuting observed among women is a desirable outcome or not is not an uncontroversial issue. On the one hand, shorter commutes reduce financial, time and psychological costs (Blau & Kahn, 2017; Le Barbanchon et al., 2021; Petrongolo & Ronchi, 2020) and allow for a better work-family balance (particularly for women with children or dependents). However, lower commuting may be due to some degree of discrimination or other cultural barriers that limit women’s work opportunities (Giménez-Nadal & Molina, 2016; McQuaid & Chen, 2012). Therefore, shorter commutes may be associated with poorer relative working conditions, if they are involuntary and limit their job search area (Ruppert et al., 2009; Van Ommeren & Rietveld, 2005; Waldfogel, 2007). In general, an advantage of longer commutes is that they allow for a better match with the individual’s educational level. In this sense, different studies show that longer commutes are associated with higher wages and better job characteristics and conditions (Brown et al., 2015; Manning, 2003; Mulalic et al., 2014; Plaut, 2006; Roberts & Taylor, 2017), so that commuters have, on average, higher wages than non-commuters (So et al., 2001). In addition, longer commutes provide a greater amount of job opportunities, so higher wages are an incentive to commute significantly longer (Sandow & Westin, 2010b).¹ Thus, Le Barbanchon et al. (2021) show that women have a lower willingness to commute and that gender differences in the valuation of commuting are a relevant source of the gender wage gap, explaining between 9% and 16% of it. In turn, according to Farré et al. (2020), longer commutes also have a comparatively significant negative effect on female labor force participation, because although such an effect also is observed for men, its magnitude is comparatively much smaller for them. Furthermore, the negative relationship between commuting times and labor force participation is particularly strong for married women and women with children. This evidence emphasizes the influence of gender norms on the roles that women play in society, on shaping women’s commuting patterns and, by extension, on their work outcomes. Overall, to consider whether less commuting is a desirable outcome or not, it is essential to keep in mind that longer commutes will be perceived positively if they are freely chosen, but negatively if they are experienced as unwanted (Sandow & Westin, 2010a).

Three main areas of analysis can be identified in the literature on the origins of gender differences in commuting. Starting with the first of these, gender economics, the different degree of household responsibilities between men and women has been considered the main potential explanation of the phenomenon (Household Responsibility Hypothesis). In this regard, the available evidence suggests that part of the gender differences in the time and distance commuted is indeed due to the presence of children (Sandow & Westin, 2010a; Wheatley, 2013) and a generally greater involvement of women in household activities compared with men (Giménez-Nadal & Molina, 2016; Hanson & Johnston, 1985; Singell & Lillydahl, 1986; White, 1986). This suggests that women adjust their commuting patterns to meet family needs, thereby limiting themselves to jobs closer to home and narrower local labor markets² (Hanson & Pratt, 1995; Turner & Niemeier, 1997).

In light of this evidence, the origin of gender differences in commuting may be largely cultural and social in nature (Crane, 2007; Marcén & Morales, 2021; Sandow, 2008) and may be rooted in the fact that women often have to combine work and family responsibilities and, in some cases, tend to prioritize their domestic and family responsibilities and to perceive their work role as secondary (Lee & McDonald, 2003; Rosenbloom & Burns, 1993), which would lead to women’s entrapment in terms of their mobility (entrapment-of-women theory). This is consistent with the fact that the gender stereotypes that exist to a greater or lesser extent in most societies (Leonard, 2001) lead, for example, to women being more likely to take care of family emergencies (Adkins, 1995; Rosenbloom & Burns, 1993), thus prioritizing their family responsibilities over the pursuit of paid work under the same conditions as men, which entails not being able to carry out commuting arrangements comparable to those of male workers (Cristaldi, 2005). This is in line with the economic theories of the household, which argue that the distribution of time devoted to the different daily activities is carried out seeking to maximize the income and joint satisfaction of all household members based on elements such as comparative advantage, productivity and the preferences of each member, which may differ between men and women (Becker, 1991; Lakshmanasamy, 2003; Mattila-Wiro, 1999); Becker, 1991; Lakshmanasamy, 2003), as well as with the bargaining models, which assume that in the search for joint maximization of household utility a decision-making negotiation takes place in which each member uses his or her bargaining power, with the possibility of imbalances between men and women due to elements such as the different resources they control inside and outside the household (Lundberg & Pollak, 1993; McElroy, 1990).

One element with a potentially significant impact on commuting is educational attainment, which is positively correlated with individual commuting in terms of both distance and travel time (Lee & McDonald, 2003; Sandow & Westin, 2010a; Vance & Hedel, 2008). Such a relationship can be explained by job search in more specialized labor markets and a lower aversion to longer trips as a result of prioritizing residential location (Groot et al., 2012). The gender gap in commuting tends to be comparatively smaller for individuals with higher levels of education (Lee & McDonald, 2003; McQuaid, 2009; Sandow & Westin, 2010a; Vance & Hedel, 2008), which could be related to the fact that higher levels of education among women are associated with a lower acceptance of traditional gender roles (see, for example, Kane, 1995 and Garrido, 2018), which in turn could lead them to adopt more egalitarian work patterns. This is consistent with the significant influence of social/cultural beliefs about gender roles on women’s work behavior and outcomes, as evidenced by the fact that countries with more anti-egalitarian views tend to have lower female employment rates and larger gender pay gaps (Fortin, 2005).

Another potentially influential element is the family structure, since in the case of women with a partner it is often the man’s employment that determines the location of the couple and conditions the woman’s job search (Madden, 1981; Salom & Delios, 1998). This may be the case even in households where both spouses have higher education and full-time jobs, i.e., regardless of the presence of comparable employment patterns (Green, 1995; Van Ommeren et al., 1998; Wheatley, 2013), although this is a controversial issue on which there is no conclusive evidence (Hoogstra et al., 2011, 2017). The influence of family structure on women’s relative commuting is consistent with the fact that being married can significantly influence residential and work location decisions, as well as commuting time and distance (Lee & McDonald, 2003). In this sense, single workers have been found to have longer commutes than married workers, and married workers whose spouses also work have shorter commutes than those whose spouses do not work (Johnston-Anumonwo, 1992; Lee & McDonald, 2003; Turner & Niemeier, 1997). Thus, a larger gender commuting gap is observed when controlling for marriage, as women tend to coordinate dual roles (Giménez-Nadal & Molina, 2016; Hanson, 2010; McGuckin et al., 2005), resulting in greater spatiotemporal constraints (Kwan, 1999; Rapino et al., 2011). In contrast, commuting differences between men and unmarried women are generally smaller or insignificant (Hersch & Stratton, 1994; Kwon & Akar, 2021; Turner & Niemeier, 1997). Along the same lines, Cristaldi (2005) shows that differences in the commuting gap are also observed across female cohorts. In general, younger women, who are less likely to be married, have significantly longer commutes than older women, who are more likely to be married, less educated, and have greater family responsibilities. However, the evidence is inconclusive, with some studies finding that married women tend to commute further in response to wage increases (England, 1993) and that the presence of children in some cases leads to an increase in commuting time (Iwata & Tamada, 2014) to maintain the children’s ties (Hofmeister, 2005) or to be closer to the grandparents’ homes, which significantly reduces the burden of childcare (Iwata & Tamada, 2014).

In labor economics, on the other hand, the analysis of the determinants of the commuting gap has focused on the role of job characteristics, concluding that the nature of women’s employment, characterized by occupational and sectoral distributions that are relatively different from those of men (Hanson & Pratt, 1995), is often associated with different labor market areas (Sandow & Westin, 2010a; Wheatley, 2013). This is because female-dominated jobs are often better spatially distributed, allowing for shorter commutes that facilitate the coordination of work and family responsibilities (Hanson & Pratt, 1995; Salom & Delios, 1998). Similarly, women often work part-time and are paid comparatively low wages, which reduces their ability to afford the monetary costs associated with increased travel (Van Ommeren & Rietveld, 2005).

There is a close relationship between commuting and migration, two forms of mobility that are central to the adequate functioning of the labor market (Deding & Filges, 2010; Lux & Sunega, 2012) by allowing the spatial adjustment of labor supply and demand. In fact, depending on distances/times and costs, migration and commuting can be substitutes (e.g., Royuela & Vargas, 2009; Sandow, 2008), in a mechanism that improves labor conditions. It should be noted, however, that in the case of Spain residential mobility is lower than in other European countries (Romani et al., 2003; Sánchez & Andrews, 2011), given the presence of rigidities both in the labor market (high unemployment rates that discourage workers from changing jobs) and in the residential market (a strong preference for homeownership and a scarce and expensive supply of rental housing).

Extensive literature has addressed the related decisions of commuting and migration, an extremely complex issue in terms of endogeneity, reverse causality, the dynamic nature of these processes, and the context of uncertainty that characterizes them (Haas & Osland, 2014). For example, although commuting is not the most important factor in the decision, many studies have concluded that the length of the commute is positively related to migration (Deding & Filges, 2010; Deding et al., 2009) because, from an individual point of view, very long commutes are suboptimal (Van Ham & Hooimeijer, 2009). In the case of two-earner households, where mobility decisions are made jointly (Plaut, 2006), a change in residence affects the commuting patterns of both workers and may have different consequences (Nivalainen, 2004; Roberts & Taylor, 2017). Authors such as Plaut (2006) or Roberts and Taylor (2017) conclude that commuting decisions within a household are complementary, i.e. both spouses decide to adjust commuting in such a way that both increase or reduce their commuting. However, the evidence is inconclusive because these decisions may also be substitutive, i.e., one spouse tends to commute significantly longer than the other (Davis, 1993). Thus, some authors (Brown et al., 2015; Sandow & Westin, 2010b) show that the relocation of the family residence is associated with commuting times/distances that are longer than before the migration, although to a much lesser extent in the case of women (Clark et al., 2003), implying a widening of the gender gap in commuting. However, Axisa et al., (2012a, 2012b) find that over time after the change of residence, as the length of residence increases, the commuting distance could decrease for both spouses through an adjustment process, implying a narrowing in the gender gap in commuting.

Finally, urban economics offers a third set of explanatory factors, as commuting patterns also reflect individuals’ location choices and the characteristics of the territories in which they live, which may differ for men and women if they have different geographical distributions (Østbye et al., 2018; Sandow & Westin, 2010a). In particular, one territorial factor that is likely to be relevant is population density, as higher density is usually associated with longer commuting times due to factors such as traffic congestion and greater reliance on public transport (Groot et al., 2012). In this area of analysis, there is also evidence that the mode of transport may play a particularly important role, as the relatively greater use of public transport by women tends to increase the length and distance of their commutes, and thus their job search area and employment opportunities (Crane & Takahashi, 2009; De la Hoz, 2008).

On a separate vein, studies that have examined how commuting affects different aspects of workers’ subjective well-being have found that longer commuting times have an overall negative impact on individuals’ level of life satisfaction (De Vos et al., 2013; Ettema et al., 2010; MacKerron, 2012; Simón et al., 2020; Stutzer & Frey, 2008). Thus, longer commutes are not fully compensated by improvements in the labor or housing markets, leading to higher disutility (Ingenfeld et al., 2019; Jacob et al., 2019; Wheatley, 2014). Several studies have also looked at the possible differential effect of commuting by gender in areas such as perceived stress, mental health, flexible working hours, impact on leisure time, and subjective well-being (Chatterjee et al., 2020; García et al., 2007). Evidence suggests that commuting may have a comparatively detrimental effect on women’s satisfaction and stress levels, suggesting a lower tolerance for travel time (Wener et al., 2005). Tao et al. (2023) examine this relationship from a household perspective and show that although both men and women experience a decrease in their subjective well-being, increases in women’s commuting time not only reduce their own life satisfaction but also negatively affect their husbands’ subjective well-being. This may be due to reduced support for family tasks when wives commute longer hours (Brömmelhaus et al., 2020). Nevertheless, such evidence is not entirely conclusive, as some studies suggest that the effect of gender is negligible (Clark et al., 2020; Lorenz, 2018; Morris & Zhou, 2018), not significant (Gottholmseder et al., 2009; Lucas & Heady, 2002) or significant only for those individuals with very long commutes (Jacob et al., 2019).

To conclude, analyses of the gender gap in commuting for the Spanish case are very scarce. With a few exceptions they are mainly descriptive and use aggregate data. Moreover, all of them have focused on specific territories, such as the region of Madrid (Rodríguez & García, 2012), that of Valencia (Salom & Delios, 1998), or the main urban areas of Andalusia (Torrado et al., 2018). These analyses confirm, in line with studies for other countries, that women’s commuting time to work is lower than men’s but differ in their conclusions about which are the main explanatory factors: occupational and sectoral segregation of women (Torrado et al., 2018) or greater domestic responsibilities (Rodríguez & García, 2012 and Salom & Delios, 1998).

Following the literature review in Sect. 2, the empirical analysis tests three major hypotheses. The first hypothesis is that the observed gender differences in commuting time could be explained by the different characteristics that men and women might have on average in different dimensions related to socioeconomic and family, geographic distribution, modes of transport and job elements. The second main hypothesis is that in certain groups where gender roles may be comparatively minor women’s commuting behavior may be more similar to that of men. Finally, a third hypothesis is that women’s shorter commutes may be the result of their different preferences, as women may have lower a tolerance for commuting, and that this may be reflected in different degrees of association with satisfaction in different domains compared to men.

The aim of this article is to examine the differences in commuting times between male and female employees in Spain. A better knowledge of the origins of the gender gap in commuting times, which, albeit with international differences in its magnitude, tends to be observed in many advanced countries, could contribute, among other circumstances, to a better understanding of one of the elements that could influence the worse employment outcomes suffered by women.

A particularly novel aspect of the research is the use, for the first time in this field of study, of econometric decomposition techniques. These techniques make it possible to quantify how much of the gap can be explained (in total and separately) by each of the individual, family, territorial and work-related factors identified as relevant in the literature, and to test whether different commuting patterns exist when men and women with the same characteristics are compared. To this end, the empirical analysis uses microdata from a survey with a wealth of information that allows for a comprehensive control of the various elements that the literature has identified as possible determinants of gender differences in commuting, including those related to several dimensions of family responsibilities. The analysis is based on time-based measures of commuting, which have important advantages over those based on distance.

The empirical analysis tests three main hypotheses. The first one is that the observed raw gender differences in commuting time could be explained by the different characteristics that men and women have on average with respect to different individual, family, territorial, and job-related elements associated with commuting time. The second hypothesis is that there may be no gender differences in commuting patterns for certain groups of women with weaker gender roles, such as those with higher education, those without family responsibilities, and those without a partner. Finally, the third major hypothesis is that the lower commuting of women could be due to different preferences as compared to those of men and a lower tolerance of women towards commuting, and that this could be reflected in different degrees of association between commuting time and satisfaction on different domains.

As in other advanced countries, women in Spain have comparatively shorter commuting times than men. The results of the decomposition analysis show that Spanish women as a whole have relative characteristics that, ceteris paribus, are associated with longer commutes, such as the modes of transport they use (with a predominance of public transport) and the characteristics of their jobs (generally of lower quality and less stable), with the exception of their greater family responsibilities, a factor typically associated with shorter commutes. Although this evidence confirms that gender differences in the relative endowments of characteristics help to explain the raw gender differences in commuting time, it does not support the hypothesis that the lower commuting time of women is due to the fact that men and women differ in the characteristics that are important determinants of the phenomenon. Consequently, according to the decomposition analysis, women systematically show significantly shorter commuting times than observationally comparable men with the same personal, family, territorial and occupational characteristics. This evidence is consistent with the hypothesis expressed in the literature that there are social or cultural determinants that may limit women’s ability to commute in practice. This evidence is further supported by the finding that women with higher education, those without family responsibilities and those without a partner have commuting patterns similar to those of comparable men, confirming the hypothesis that there are no gendered commuting patterns in groups with lower acceptance of traditional gender roles. Finally, with regard to the third hypothesis, it does not appear that the relationship between commuting and satisfaction with various life domains varies by gender. Thus, this result does not support the possibility that women’s shorter commuting times are allegedly voluntary, related to different preferences and, in particular, to a lower tolerance for commuting.

The shorter average commuting time of women suggests that the average size of the local labor markets in which they participate is comparatively small, which may condition their access to a wider variety of higher quality jobs. The evidence obtained suggests that women’s shorter commutes are related to their spending more time on household tasks, together with unobservable elements that are likely to be related to traditional gender roles and women’s functions in the family structure. Any policy that addresses these factors could therefore contribute to reducing the commuting gap between men and women and possibly lead to less differentiated employment opportunities. Among the various measures that would be in line with both this objective and the European strategies for reducing gender gaps in the labor market (European Commission, 2017; European Union, 2019) are, for example, those related to the provision of sufficient and affordable services offering quality care for children (including free education for 0–3 year olds, non-teaching staff in schools for children who require sporadic care, after-school activities and homework support, and holiday care services); an adequate supply of quality and affordable long-term care services related to population aging and dependency (such as residential care, day centres, and permanent or emergency home care); and, finally, a more flexible organization of working time, which would reduce the pressure on women to work closer to home to care for their families and cope with contingencies, and could lead to greater spatial flexibility. As Petrongolo and Ronchi (2020) point out, two relevant aspects in this context are the impact of new technologies, both in terms of technological and organizational changes that allow more flexible forms of work organization in established sectors, and the rise of the collaborative economy. While both of these trends could prove positive for women by reducing the costs of flexibility, there is also a risk that they will deepen their specialization in occupations that are more permeable to non-standard work arrangements, thereby reinforcing a potential comparative advantage in non-market work, which in turn could have a negative impact on gender norms and aspirations. Similarly, policies aimed at providing better care services in the context of ageing, dependency and childcare need to be carefully designed and implemented to avoid simply commodifying currently unpaid female care work in ways that create low-paid, low-status jobs, mostly occupied by women, thereby reinforcing the stereotype of care work as female work.

To conclude, some limitations of the analysis, which derive from the characteristics of the dataset employed, should be mentioned. First, since the survey does not include information on the unemployed, it is not possible to account for the selection into labor market participation and possible related biases. Second, the econometric techniques used in the analysis do not allow for the identification of causality in the observed relationships and, in the specific case of the examination of the determinants of satisfaction, the estimations could be affected by endogeneity and reverse causality problems that cannot be sorted out with our cross-sectional data. Finally, the results observed for Spain in the specific period examined cannot be automatically extrapolated to other countries and periods. All these points encourage further research in this area, expanding the range of countries and periods considered, and based on different types of data and econometric techniques.