Women's work–life preferences: reconceptualization and cross-country description over time

Research on the determinants of women's employment patterns and childbearing has traditionally emphasized the role of constraints. For example, cross-country variation in family policy has often said to be the reason for cross-country differences in women's employment and fertility (Esping-Andersen 1999; McDonald 2000; Rindfuss et al. 2003). The constrained-oriented theories on women's work–life choices are challenged by Hakim (1998, 2000, 2002, 2003a,b), who argues that women's preferences, rather than constraints such as family policy, are decisive for women's decisions. In particular, Hakim claims that women have heterogeneous preferences, which vary in a continuum from home-centered (focus on family only) to adaptive (focus on combination of work and family) and work-centered (focus on work) preferences, and that women's responsiveness to social policies and employment conditions varies depending on their preferences.

The discussion around Hakim's preference theory has so far concentrated on the relative importance of constraints and preferences for women's choices. Hakim's conceptualization of preferences is left with lesser attention, even though several commentators have pointed out the many weaknesses. For example, McRae (2003: 328) draws attention to the difficulty to distinguish between work-centered women and ‘adaptive women at the full-time end of the employment continuum’ as well as between home-centered women and ‘adaptive women at the part-time end’. Further, Johnstone and Lee's (2009a,b, 2016) work shows that adaptive women are more heterogeneous than what is suggested by preference theory. Finally, there is a large variation in how fast the mothers return to employment after the childbirth (e.g. Gutiérrez-Domènech 2005; Pronzato 2009) and how many hours they work (Lewis et al. 2008). It is possible that different preferences lie behind these different strategies to combine work and family.

Taken together, the literature suggests that more work needs to be done to better understand the nuances of the adaptive preferences. In the current paper, it is argued that we need to study when the women who want to combine work and family want to return to employment after the childbirth, and how many hours they would like to work. Based on this information, it is possible to distinguish between three types of adaptive women: (1) the home-oriented adaptive women who wish to stay at home a longer time after the childbirth and then wish to return to either full-time or part-time employment. (2) The truly adaptive women who wish to return to work soon after childbirth but who want to work part-time and (3) the work-oriented adaptive women who likewise want to return to part-time employment relatively soon, but when the child grows older wish to switch to full-time employment.

To study the fruitfulness of the reconceptualization, a cross-sectional descriptive study of women's preferences over time and between the countries is conducted by employing data from the International Social Survey Programme (ISSP) for 14 countries. The results indicate that there are considerable differences between the countries and over time in how women want to combine work and family, and that many of these differences would remain hidden if we would employ Hakim's conceptualization.

Hakim's (1998, 2000, 2002, 2003a,b) preference theory challenges the constraint-oriented theories about women's employment and fertility and shifts the focus to preferences. According to Hakim, women have heterogeneous preferences when it comes to work and family: the home-centered women prefer homemaking, work outside home only if they have to and often have large families. For the work-centered women, career/work comes first, family is at the second place and these women often remain childless. Finally, the adaptive women wish to combine work and family, without giving priority to either of them (or are drifters/have unplanned careers). As for the size of the groups, Hakim assumes that 10–30% of the women are home-centered or work-centered, respectively, and 60–80% are adaptive. Importantly, Hakim also claims that women are responsive to such policies, which increase their possibilities to choose according to their preferences.

The evidence in support for preference heterogeneity is convincing, and studies also show that there are cross-country differences in what women want (e.g. Evans and Kelley 2001; Hakovirta and Salin 2006; Kangas and Rostgaard 2007; Vitali et al. 2009; Cotter et al. 2011; Janus 2013a,b). Hakim's claim about the increasing influence of preferences, however, has been fiercely criticized (e.g. Crompton and Harris 1998; McRae 2003; Crompton and Lyonette 2005), especially for her ignorance of constraints. Indeed, studies which investigate the importance of preferences for women's employment decisions find that not only preferences, but also family policies, labor market structures, such as the availability of part-time work and the attitudes of the husband/spouse influence women's labor market participation (Kangas and Rostgaard 2007; Janus 2013a,b). Moreover, it is well known that the presence of children (Corrigall and Konrad 2007; Kan 2007; Stähli et al. 2009), or individual level factors such as education (Debacker 2008) are influential in the decision-making process.

In addition, as pointed out by Fagan (2001: 244), ‘preferences are expressed from the vantage point of current circumstances, needs and the range of perceived feasible alternatives, and would be expected to change in different circumstances’. For instance, Johnstone et al. (2011) showed that between two surveys in 2000 and 2003, employment aspirations had changed for around 40% of the surveyed women. Further, several studies evidence that the causality between preferences and behavior is reciprocal (Crompton and Lyonette 2005; Kan 2007), and that women adjust their preferences after changes in, for example, external circumstances, childbearing and relationship status (Crompton and Harris 1998; Johnstone et al. 2011).

Many studies on preferences do not employ Hakim's classification, but ask women about their preferred workings hours (Fagan et al. 2001), reasons for working part-time (Gash 2008) or preferred working hours in comparison to the current ones (Yerkes 2013). These studies focus on short-term preferences, and are useful to understand the determinants of preferences and the association between preferences, constraints and behavior. Yet, women with different long-term work–life preferences may give the same answer if they are asked about their current working hour preferences. For example, two women who in a survey indicate that their current working hour preference is part-time work, may vary considering their working time preferences in five years. One of the women may never wish to work full-time and has a permanent preference for part-time work. The other woman, however, even though she currently wishes to work part-time, may want to start working full-time within the next five years. To detect these differences, it is important to study women's long-term preferences.

Hakim's theory outlines one possibility to distinguish between women based on their long-term preferences, and a majority of those studies, which divide women into theoretically pre-determined groups, employ Hakim's classification (e.g. Vitali et al. 2009; Janus 2013a,b). However, Hakim's conceptualization of the adaptive women as women who ‘prefer to combine employment and family work without giving a fixed priority to either’ (Hakim 2003a: 357) is lacking in analytical sharpness. Indeed, the adaptive group is ‘a “safe” but effectively meaningless category, including any woman who doesn't fit the other two categories’ (Johnstone and Lee 2009a: 13). The residual-character of the adaptive group has consequences for our understanding of what women want. For example, Johnstone and Lee's (2016) attempt to classify women into Hakim's three categories shows that a strict definition of the adaptive women leaves a large number of women unclassified, whereas a more inclusive definition increases the internal heterogeneity of the adaptive group.

The research on women's work–life choices further highlights the need to separate between the different adaptive women. It is well known that women's return to the labor market after the childbirth differs (Gutiérrez-Domènech 2005; Pronzato 2009), and that there are differences in the number of hours mothers work (Lewis et al. 2008). Thus, there are different ways to combine work and family. If we assume that behavior is partly determined by corresponding preferences, we would expect that the adaptive preferences are more complex than what is suggested by Hakim.

Here it is suggested that, based on the research on women's work–life choices, the adaptive women vary depending on how soon they want to return to employment after the childbirth, and how many hours they would like to work. Three different types of adaptive women are distinguished: for the home-oriented adaptive women family has a higher priority than work when the children are young. These women prefer to stay at home until the youngest child is relatively old, but would then like to return to the labor market. Consequently, home-oriented adaptive women are responsive to policies which support homemaking in the first few years following the childbirth (e.g. long parental/care leaves and tax reductions for breadwinners). Only when the children grow older do these women prefer policies such as available and affordable child care, which facilitate the work–family conflict.

For the work-oriented adaptive women work has a high priority and thus, these women prefer to stay at home a relatively short time after the childbirth. Unlike the work-centered women, however, the work-oriented adaptive women would like to work part-time when the children are young, and switch to full-time employment only when the children grow older. In other words, long and generous parental leaves have little relevance for these women, whereas the availability of part-time work and affordable childcare is of importance.

The truly adaptive women are torn between work and family. On the one hand they do not want to stay at home for extensive periods after the child birth, but on the other hand they would also not want to work long hours and be absent from home. Thus, their preferred solution is part-time work independently from the age of the child, and they are responsive to the availability of part-time work and affordable childcare.

In reality, women's adaptive preferences are extremely complex, as the exact time the women want to stay at home, as well as their preferred working hours differ enormously. Even though the suggested three groups cannot capture all these nuances, differentiation of the different types of adaptive women nevertheless allows us to get a more detailed understanding of what women want in the long-term than what so far has been the case. The specification is valuable for researchers, who want to study the link between preferences, constraints and behavior. For example, it is possible that home-oriented and work-oriented adaptive women are influenced differently by limited childcare supply: the odds of home-oriented adaptive women to opt for homemaking may be high, whereas the work-oriented women may be more inclined to decide not to have children at all. Moreover, policy-makers are likely to benefit from a better understanding of the adaptive preferences in order design policies, which in an optimal way support the adaptive women's heterogeneous needs.

Hakim claims that to understand women's work–life preferences (i.e. what women want for their own lives), we should not employ questions about general attitudes (i.e. what kind of behavior women in general approve or disapprove). Hakim further argues that, based on her own research, the correlation between attitudes and preferences is weak, and that the impact of attitudes on behavior is weaker than the impact of preferences (see, for example, Hakim 2003c). Even though Hakim's argument about the need to distinguish between personal preferences and general attitudes is sound, the availability of data on personal preferences is limited. Moreover, the good availability of cross-country survey data on attitudes has lead several researchers to employ such data in their studies on preferences.

Table 1 shows the questions Hakim employs to operationalize women's preferences, and the questions from ISSP and European Social Survey (ESS), which in previous studies have been used to investigate what women want. The ISSP's Family and Changing Gender Roles-module is the most frequently employed dataset. For example, Evans and Kelley (2001) employ question 1b, Hakovirta and Salin (2006) study questions 1b and 1c separately, Kangas and Rostgaard (2007) build an index of questions 1a–c, 2 and 3, and Janus (2013a,b) combines the answers to questions 1a–c.

Vitali et al. (2009) employ data from the ESS and operationalize women's preferences with help of three questions. The first two questions clearly refer to attitudes, but the third question corresponds closely to the third question in Hakim's list. When Vitali et al. (2009: 420) compare their results for the UK for 2004 with Hakim's results for 1999, they conclude that the results are quite similar. Thus, against Hakim's claims the approximation of preferences with attitude-questions might not be such a bad idea.

The current study employs data from the four rounds (1988, 1994, 2002 and 2012) of the ISSP Family and Changing Gender Roles-module. The long time span and repeat questions make it possible to study changes in over more than two decades (ISSP Research Group 1990, 1997, 2013, 2014). The study includes 14 Western European and Anglo-Saxon countries for which there are data at least for the 2002 and 2012 rounds. For some countries, there are data also for 1988 and 1994.

The concentration is on work-orientations of women aged 15/16–50 years. For older women the thoughts about combination of work and family might to an increasing extent reflect the work–life choices they have made. To operationalize women's preferences, questions 1b and 1c as shown in Table 1 are employed.

According to Hakim, the home-centered women ‘are most inclined to have larger families, and these women avoid paid work after marriage unless the family is experiencing financial problems’ (Hakim 2003a: 359). Following Hakim's conceptualization, women who think that women should not work at all outside the home independently from the age of the youngest child are classified as home-centered. As the work-centered women are ‘Committed to work or equivalent activities’ (Hakim 2003a: 358), all women who think that women should work full-time independently from the age of the youngest child, are classified as work-centered.

The ISSP-questions do not ask about the time when the women want to return to the labor market after the childbirth or about the number of hours, the women would like to work when they return to employment. It is, however, possible to roughly estimate the number of home-oriented adaptive, truly adaptive and work-oriented adaptive women based on the ISSP-data.

Respondents, who think that women should not work at all outside the home when there is a child under school age, but should work part-time or full-time when the youngest child starts school, are classified as home-oriented adaptive women. Respondents who think that women should work part-time independently from the age of the youngest child are classified as truly adaptive. Finally, respondents who think that women should work part-time when there is a child under school age, but full-time when the youngest child start at school, are classified as work-oriented adaptive.

Respondents for whom the answer to one of the two questions was missing are excluded. Information on the number of women in the sample before and after the exclusions can be found in the appendix. According to a small number of respondents (on average, 2.2% in 1988, 1.5% in 1994, 1.6% in 2002 and 1.3% in 2012), women should work less when the youngest child starts at school than when there is a child under school age. As these preferences are marginal (except for Austria in 1988, Eastern Germany in 1994, and Finland in 2012, when 7.2%, 6.2% and 4.4% of the women, respectively, gave such an answer), these women are excluded from the graphical presentation of the results. Weights were used for all countries for which they were available.

To demonstrate that the chosen measure of preferences is not simply a measure of women's actual work–life patterns, preferences are compared with actual work–life decisions in 2012. The work–life decisions are operationalized with the following question: ‘Did you work outside the home full-time, part-time, or not at all when a child was under school age?’ The women could choose between the following answers: ‘work full-time’, ‘work part-time’, ‘stay at home’ and ‘does not apply’. All women who ever had had children were asked to answer the question.

In all countries, the number of home-centered women has been below 5% during the whole observation period. The home-oriented adaptive preferences, however, were popular among a substantially large group in 1994 and 2002, but the size of this group has declined below 10% by 2012.

Truly adaptive preferences were the most common ones in all countries in the beginning of the observation period. For example, in 2002 as many as 48% of the Eastern German, 43% of the Danish, 29% of the Finnish, 36% of the Norwegian and 50% of the Swedish women preferred this option. Ten years later, the corresponding numbers were 44% for Eastern German, 21% for Danish, 22% for Finnish, 23% for Norwegian and 29% for Swedish women. Thus, in 2012 the truly adaptive women are the largest group in Eastern Germany, and either the second or third largest group in the Nordic countries (after the work-centered adaptive or/and work-centered women).

Simultaneously with the decreases in the number of home-oriented and truly adaptive women, there has been an increase in the number of work-oriented adaptive and – in particular in the Nordic countries – work-centered women. In 2012 as many as 25–31% of the women hold work-oriented adaptive preferences. As for the work-centered preferences, in Norway and Sweden the number of work-centered women has increased between 1994 and 2012 with over 30 percentage points. In Denmark and Finland where there are data only starting from 2002, an increase with 24 and 16 percentage points, respectively, can be observed. Thus, in 2012 as many as 35–43% of the Nordic women were work-oriented. In this respect, Eastern Germany is somewhat of an outlier: in 2012, the number of women with work-centered preferences was only 25%.

In English-speaking countries, the government has traditionally provided families with monetary benefits (e.g. tax allowances), rather than with services (childcare) and time-rights (parental leave). Even though the government has not supported role compatibility, the levels of fertility and female labor force participation have been high (Esping-Andersen 1999; McDonald and Moyle 2010). Women's preferences, however, are very different in the Anglo-Saxon social policy context in comparison to the Nordic countries and Eastern Germany.

As in the Nordic countries and Eastern Germany, also in Australia and the UK the number of women with home-centered preferences has been low throughout the observation period. In Ireland, 12% of women were home-centered in 1988 and 1994, but by 2002, the number even in Ireland had declined to 6%.

Even though the home-oriented adaptive preferences have become less common throughout the years, in 2012 as many as 28% of the women in Ireland, 36% in the UK and 37% in Australia held these preferences. Thus, the contrast to the Nordic countries and Eastern Germany is considerable. Concerning the truly adaptive preferences, in Australia an increase from 22% to 34% between 1994 and 2012 can be observed. In Ireland, there is an inverted-U pattern: more than 30% of the women in 1994 and 2002 were truly adaptive, but in 1988 and 2012 the numbers were slightly lower. Finally, in the UK there is a slight decline from 1988 to 2012, but as in Ireland, a higher number for 2002. Again, the contrast to the Nordic countries and Eastern Germany is notable. Finally, as in the Nordic countries and Eastern Germany, there has been a substantial increase in the popularity of the work-oriented adaptive preferences, and in 2012 as many as 25–32% of the women in Australia, Ireland and the UK are work-oriented.

Finally, in 2012 only 11% of the Irish, 6% of the British and 3% of the Australian women had work-centered preferences. Thus, in Ireland the work-oriented adaptive women are the largest group, whereas in Australia and the UK the home-oriented adaptive preferences are the most common ones.

In German-speaking countries, social policies have for long supported the male breadwinner model of family, which is often said to be the reason for the low fertility and female labor force participation (Esping-Andersen 1999). During the past two decades, social policies have given more emphasis to work–family compatibility, but fertility rates nevertheless remain at low levels, even though increases in female labor force participation can be observed (Adema et al. 2014). Interestingly, women's preferences in these countries are also relatively home-oriented, but we also notice considerable changes through the decades.

In 1988 as many as 34% of Western German, and 28% of Austrian women were home-centered. By 2012, popularity of the home-centered preferences in both countries had declined to a level below 10%. In Switzerland, where there are data for 2002 and 2012, the number of home-centered women has been below 10%.

The number of home-oriented adaptive women has declined in Switzerland (from 28% to 13% between 2002 and 2012) and Western Germany (from 34% to 16% between 1988 and 2012). The number of home-oriented adaptive women in these two countries is thus higher than in the Nordic countries and Eastern Germany, but lower than in Australia, Ireland and the UK. Interestingly, in Austria the number of home-oriented adaptive women has increased with 10 percentage points between 2002 and 2012, which is in contrast to the declining trend in the other observed countries.

The number of truly adaptive women has increased in Switzerland (from 55% to 67% between 2002 and 2012) and Western Germany (from 26% and 56% between 1988 and 2012). Thus, the number of truly adaptive women is the highest in these two countries among the observed countries. In Austria, in turn, there was an increase in the number of truly adaptive women between 1988 and 2002 (from 24% to 47%) but a decline by 2012 (to 30%).

As for the number of work-oriented adaptive women, there is an increase over time. Yet, the three countries differ from the rest of the observed countries with the particularly low number of work-oriented adaptive women (11% in Switzerland, 18% in Austria and Western Germany in 2012). Finally, the number of work-oriented women has been low, below 10%.

France, Spain, and the USA belong to completely different types of family policy regimes. France is often grouped together with the Nordic countries, the USA belongs to the Anglo-Saxon family and Spain is often classified together with the German-speaking countries or included in a separate category with the Southern European countries (Esping-Andersen 1999; Leitner 2003). The fact that women's preferences in these three countries are similar suggests that even other factors than the family policy regime are consequential for the preference formation.

Even in France, Spain and the USA the number of home-centered women has been low. The number of home-oriented adaptive women has declined to 16% in Spain, 19% in the USA and 23% in France in 2012 (levels comparable to those observed in Western Germany).

The number of truly adaptive women has remained stable in the USA, around 19%. In France, there has been a decline from 32% to 23%, and in Spain, we see an inverted U-shaped pattern (increase from 28% in 1994 to 34% in 2002, followed by a decline to 29% in 2012). In other words, in these countries the number of truly adaptive women is similar as in all other observed countries except for Switzerland and Eastern and Western Germany.

As for the work-oriented adaptive women, there has been an increase: in Spain from 19% to 40%, in France from 17% to 38% and in the USA from 27% to 36% between 2002 and 2012. Consequently, it is in these three countries where the share of the work-oriented adaptive women was the highest in 2012.

Finally, in France and the USA there has been an increase in the work-centered preferences. In France, the increase has been a minor one. In the USA, already in 1988 as many as 15% of the women were work-centered, and by 2012 the number had increased to 22%. Spain is a special case in that as with the work-oriented adaptive preferences, there has been an inverted U-shape pattern. However, in 2012 as many as 14% of the women were work-oriented. Thus, the numbers are higher than in Australia, Ireland, the UK, and in the three German-speaking countries, but still much lower than in the Nordic countries.

As for the home-oriented adaptive women, the countries divide into two groups. In the first group (Australia, Austria, Switzerland, Western Germany and the UK), over 50% of the women stay at home when the children are below school age, and those who are in the labor market work mostly part-time. In the second group (France, Ireland, Norway, Spain and USA), less than 50% stayed at home and it is common that those women who work are in full-time employment. Notice that these two groups overlap partly with the family policy regimes: in the first group we find the German-speaking countries which traditionally have supported homemaking, whereas in the second group we find Norway and France which support mothers’ work. The Anglo-Saxon countries can be found in both clusters.

As for the truly adaptive women, we can distinguish between three groups. In the first group (Australia, Austria, Switzerland and Western Germany), over 50% of the women work part-time when the children are young. Moreover, over 20% have opted for homemaking, which is particularly interesting as the German-speaking countries have traditionally supported homemaking. In the second group, (Denmark, Norway, Sweden, Eastern Germany, the UK), over 50% of the truly adaptive women are in part-time employment, but a majority of those who are not in part-time employment opted for full-time employment. This is partly in accordance with the support for working mothers in the Nordic countries and Eastern Germany. Finally, in the third group of countries (Finland, France, Ireland, Spain and the USA), less than 50% of the truly adaptive women are in part-time employment. In Finland and France, a large majority of the truly adaptive women who are not in part-time work, work full-time. In Ireland, Spain and the USA, however, it is common that the truly adaptive women work full-time or stay at home.

As for the work-oriented adaptive women, we would expect to find several similarities to the truly adaptive women, given that both groups of women wish to work part-time when the children are young. Indeed, for several countries (Finland, Ireland, Norway, Sweden, UK and Spain) the work–life choices of the truly adaptive and work-oriented adaptive women are similar. In other countries, there are some differences between the two groups. In general, a much lower share of the work-oriented adaptive women work part-time in comparison to the truly adaptive women and those who do not work part-time, opt for full-time employment. This is true even for Australia, where several of the truly adaptive women stayed at home if they did not work part-time. There are two exceptional cases, however. In Austria as many as 78% of the work-oriented adaptive women work part-time, which is 25 percentage points more in comparison to the truly adaptive women. In Western Germany only 38% of the work-oriented adaptive women work part-time, which is 28 percentage points lower share in comparison to the truly adaptive women.

The gap between preferences and behavior is smallest for the work-centered women: in almost all countries, over 70% of the work-centered women worked full-time when the youngest child was below school age. In Western Germany and Switzerland the share of work-centered women in full-time work was somewhat lower, but still above 50%. The findings indicate that there are fewer constraints for the work-centered women than for other women, even though the constraints might be larger in countries where family policy has traditionally supported homemaking.

The findings show that adaptive women want to combine work and family in different ways, and that Hakim's threefold category thus does not capture all the important aspects of women's preference heterogeneity. The results also indicate that the number of women who prefer a lifetime of homemaking is currently below 10%, and the findings call into question Hakim's claim that 10–30% of the women are home-centered. Yet, the concept of home-centered women remains useful to understand women's preferences in Austria, Western Germany and Ireland in 1988 and 1994. Further, the size of the work-centered group is not always within the by Hakim suggested interval of 10–30%. For example, in Australia, Austria, Switzerland, the UK and Western Germany the size of this group in 2012 is very low, below 10%. At the other extreme, we find the Nordic countries, where the size of the work-centered group is around 40%.

The findings also show that there is a gap between preferences and behavior. In addition, there are substantial cross-country differences in the size of the gap, the gap is smallest for the work-centered women, and there is some overlap between family policy regimes and the cross-country differences in the gap. The findings thus add to the evidence, which shows that constraints are an important determinant of women's behavior.

Future research on women's preferences needs to give more attention to the heterogeneity of the adaptive preferences. It is also important to discuss the appropriate conceptualization and operationalization of preferences and in particular, more research is needed to determine if attitudes are a good enough proxy for preferences. In addition, it is essential to understand the determinants of preferences. Longitudinal data is needed to understand preference formation, the changes in preferences over time, and the role of constraints, which meditate the relationship between preferences and behavior.

I am grateful for the anonymous reviewers for their constructive comments, which considerably helped me to clarify my thoughts and improve on the paper.

No potential conflict of interest was reported by the authors.

Elina Schleutker received her doctorate in political science from the Heidelberg University. Her previous work has been published, for example, in Comparative Population Studies and in Zeitschrift für Soziologie.

Number of women included in the study by country and year.