Download Gender Revolution & Fertility: Analyzing Cohort Fertility Rates & Labor Force and more Study Guides, Projects, Research Political Geography in PDF only on Docsity!
The Two-Part Gender Revolution, Women’s Second Shift and
Changing Cohort Fertility
Tomas Frejka, Frances Goldscheider, Trude Lappegård
Abstract: The two parts of the gender revolution have been evolving side by side at least since the 1960s. The first part, women’s entry into the public sphere, proceed- ed faster than the second part, men’s entry into the private sphere. Consequently, many employed mothers have carried a greater burden of paid and unpaid family support than fathers throughout the second half of the 20 th^ century. This consti- tuted women’s “second shift,” depressing fertility. A central focus of this paper is to establish second shift trends during the second half of the 20 th^ century and their effects on fertility. Our analyses are based on data on cohort fertility, male and fe- male labor force participation, and male and female domestic hours worked from 11 countries in Northern Europe, Western/central Europe, Southern Europe, and North America between 1960/70 and 2000/2014. We find that the gender revolution had not generated a turnaround, i.e. an increase in cohort fertility, by the end of the 20 th century. Nevertheless, wherever the gender revolution has made progress in reduc- ing women’s second shift, cohort fertility declined the least; where the second shift is large and/or has not been reduced, cohort fertility has declined the most.
Keywords: Two-part gender revolution · Transformation of male breadwinner family model · Women in public sphere · Men in private sphere · Women’s second shift
1 Introduction
The on-going transformation of the male breadwinner family model has been among the most remarkable societal changes of the past 50 to 60 years. This ap- pears to be occurring in two parts: The first part, the growth of female labor force participation, is accompanied by a second part, the growth in men’s involvement in the tasks at home. These overlapping parts are, in essence, the gender revolution ( Goldscheider et al. 2015).
Comparative Population Studies Vol. 43 (2018): 99-130 (Date of release: 31.10.2018)
(^100) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
This revolution has not been without consequences. The first part has evidently been unbalancing the family, reducing fertility and disrupting unions as men and women have struggled with women’s new roles that combine responsibilities in both the public sphere of employment and the private sphere of the family ( Cherlin 2016; Rindfuss/Choe 2015, 2016; Arpino et al. 2015; Lesthaeghe 2010; Sobotka et al. 2011a; Frejka et al. 2008; Olah/Bernhardt 2008; Engelhart et al. 2004; Brewster/ Rindfuss 2000; Hochschild/Machung 1989; Lesthaeghe/van de Kaa 1986). The sec- ond part might be rebalancing the family, as men’s increased domestic involve- ment reduces the pressure on women, perhaps stabilizing unions and creating the potential for fertility to increase ( Anderson/Kohler 2015; Esping-Andersen/Billari 2015; Goldscheider et al. 2015). The size and duration of the gap between the two stages, which has been called women’s “second shift” ( Hochschild/Machung 1989), in which women attempt to meet both work and major home obligations with rela- tively unimportant involvement of their partners, may thus be very consequential for fertility and family stability. Two early analyses of this proposition vis-à-vis fertility thus far did not find fertil- ity turnarounds or reversals ( Frejka et al. 2016a/b). However, that research focused on the second half of the gender revolution – the increase in men’s involvement in household tasks – and did not examine either the effects of the first half of the gender revolution or those of the gap between the two halves, both of which have varied across countries and over time. This means that it is important to consider, as well, the on-going fertility depressing effects of the first half, as female labor force participation increased, particularly in countries that provide little support to reduce work-family conflict, such as high quality childcare and job-protected, paid family leave. It is also important to consider the relationship between these two trends and the extent to which their trajectories open up larger or smaller “second shifts,” i.e., to what extent an increase in women’s employment is or is not balanced by increas- es in men’s domestic contributions. In this paper, we examine the fertility of women born in different birth cohorts, who belong to different generations where fertil- ity decisions have been made within different gender revolution contexts. Hence, this paper will examine the connections between these two trends – the growth of female labor force participation relative to men’s and the growth in men’s domes- tic participation relative to women’s – and their relationships with cohort fertility across a wide range of industrialized countries. We ask: Do countries with patterns that minimize the “second shift” experience less drastic declines or even increases in fertility than those seen in countries whose patterns maximize the second shift? The minimizing of the second shift is generated by an increase in men’s domestic share offsetting increases in female labor force participation. Our focus is on the impact on fertility of these two parts of the gender revolution, although we will also consider whether state policies that provide support to parents, such as subsidized high quality childcare and paid parental leave, might also help maintain fertility.
(^102) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
or on differences among countries or other geographic/legal units. Gender role change has been studied in terms of both behavior (e.g., Cooke 2009) and attitudes (e.g., Brinton/Lee 2016), as well as in terms of country-level gender characteristics (e.g., Mills 2010). Vis-à-vis fertility, for the individual level, scholars have studied the transition to parenthood (e.g., Bernhardt/Goldscheider 2006) as well as parity progression (e.g., Dommermuth et al. 2017; Goldscheider et al. 2013; Fiori 2011; Brodmann et al. 2007; Cooke 2009). At the macro level, similarly, although most studies examine the period total fertility rate, which is a cross-sectional measure (cf. Rindfuss/Choe 2015, 2016), a few others have looked at cohort fertility (e.g., Frejka et al. 2016a/b). In this review, we will focus primarily on macro studies, and on actual behavior rather than attitudes. We distinguish between the gendered behaviors of women and men by considering not only trends in women’s role in the public sphere (female labor force participation) and in men’s roles in domestic activities relating to home and children, but also the gap between trends in these two phenomena (women’s second shift). Importantly, although many have studied the effects of the growth of female labor force participation, few examine the fertility consequences of the second half of the gender revolution, much less the weight of women’s second shift.
2.1 Female labor force participation and fertility
Prior to the middle of the 20 th^ century, few married women worked outside the home, at least in Sweden and among the white majority in the United States, two countries recognized as leaders in the growth of female labor force participation ( Stanfors/Goldscheider 2017). The late nineteenth century and the first half of the twentieth century were the period of the “separate spheres,” when married women remained in the home if at all possible, even as the growth in industrial and com- mercial jobs drew men away from the subsistence household economy. During this period women pursued many productive roles, including replacing their husbands’ chores (e.g., drawing water, maintaining kitchen gardens), in addition to their own (e.g., spinning, sewing, washing), together with increased responsibilities for the daily activities of children. These roles evolved over time, and generally became less onerous, with the growth of cities, together with the spread of running water and electrification and children’s education ( Cowan 1983). Nevertheless, precondi- tions for eventual employment roles were developing, as women’s education in- creased, as well as their employment experience prior to marriage ( Stanfors/Gold- scheider 2017; Goldin 1990). The growth of labor force participation by women during the second half of the 20 th^ century, especially by married women and even married mothers, as a result, was rapid. This trend has been frequently documented ( Pott-Buter 1993; Rosenfeld 1996; Spain/Bianchi 1996). In the early years of the increase in female labor force participation, a research consensus emerged that it was the result of the growth in demand for “female” workers ( Oppenheimer 1970; Goldin 1990). More recent stud- ies have highlighted the positive effects of subsidized childcare (e.g., Pettit/Hook 2005).
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 103
The negative impact of increasing female labor force participation on fertility was widely recognized from the beginning of the surge in the employment of mar- ried women (e.g., Stycos/Weller 1967). There is clearly a two-way relationship: women in couples that for whatever reason have few children are more likely to be employed, and women who plan substantial labor force engagement expect fewer children ( Waite/Stolzenberg 1976). Empirical evidence confirms that the dominant causal arrow is from (plans for) employment to (plans for) childbearing, as a second income has become increasingly important for couples since the 1980s ( Oppenhe- imer 1974). It is also possible that the consequences of the growth in female labor force par- ticipation have changed ( Matysiak/Vignoli 2008). For instance, in response to this negative impact on fertility, many countries enacted policies that have attempted to alleviate the work-fertility dilemma, such as paid parental leave and expanded, subsidized high quality childcare ( Thévenon/Gauthier 2011; Gauthier 2010). But more universally, as employment changed from an occasional or emergency role for women into one that most women (and men) expect to be life-long, there was apparently a massive change in the relationship between female labor force par- ticipation and fertility, at least using cross-sectional measures of fertility and coun- try-level data. This appears to have occurred between the 1970s and 1990s. Some studies documented a shift from the expected negative relationship to a positive relationship, at least in the cross-section, using period total fertility rates ( Ahn/Mira 2002; Brewster/Rindfuss 2000); another study using different data and methods found the negative relationship had attenuated to no relationship ( Engelhardt et al. 2004). Others have found a U-shaped relationship, with a negative impact on fertility of attitudes favorable to female labor force participation early in the period of rapid growth followed by a positive impact ( Arpino et al. 2015). It is clear that this relation- ship needs further exploration. McDonald (2000) argues that the strong family-friendly policies in Northern Europe reduce the conflict between women’s work and family roles. By providing some combination of excellent childcare and job-protected family leave, women can maintain a strong connection to the labor force and have several children. (This also relieves pressure on their spouses to increase their domestic responsibilities). There is a substantial consensus in cross-national studies that the provision of high-quality, subsidized childcare supports fertility. An examination of 16 European countries finds that the most consistent positive effect on fertility is the level of childcare coverage ( Baizan et al. 2016). A more limited study comparing Denmark and Spain also cites Danish childcare support as an important source of the fertility difference between these two countries ( Brodmann et al. 2007). A study of Canada ( Beaujot et al. 2013), which now has very different family policies in Quebec and the rest of Canada, finds that childcare has helped Quebec experience a stronger recovery from low fertility than was the case in the rest of Canada. Detailed stud- ies of Nordic countries also agree that their childcare programs have helped them maintain fertility (e.g., Olah/Bernhardt 2008; Rindfuss et al. 2010), although some are concerned that public policies alone might not be able to sustain fertility in the longer run ( Ronsen/Skrede 2010).
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 105
a negative effect on fertility while Puur et al. (2008) found a positive effect, a contra- diction that was resolved by focusing on the very different measures of gender role attitudes that were employed in the two studies ( Goldscheider et al. 2010). But is this trend in male behavior vis-à-vis the second half of the gender revolu- tion enough to have a positive impact on fertility? In particular, will it be enough to offset the negative effects of the first half of the gender revolution? Are there nascent characteristics of fertility trends that validate the hypothesis of increasing childbearing in countries in which the gender revolution is most advanced? In the international comparative analysis that follows, the fertility of countries more ad- vanced on one or both halves of the gender revolution is compared with the fertility of a range of countries that have made less progress. In this paper, we will focus only on the implications of the gender revolution for cohort fertility; much more work needs to be done on its impacts on union for- mation, dissolution and timing and on out-of-union childbearing and child rearing (see the review in Goldscheider et al. 2015). First, after reviewing trends in cohort fertility, we will document the progress of both halves of the gender revolution for 11 industrialized countries in Europe and North America. Second, we will explore whether there are characteristics of fertility trends that support the hypothesis of decreasing fertility linked with the increase in the growth of female labor force par- ticipation, and increasing childbearing in periods and places in which the second half of the gender revolution is most advanced.
3 Data, Measures, and Methods
3.1 Data
In order to analyze the relationship between cohort fertility trends and the two halves of the gender revolution, we use data from three major sources: national information on cohort fertility and labor force participation, which is often avail- able annually; and surveys of time use from various sources, which are collected in many countries (but much more sporadically). Based on the availability of at least some data on each of these measures, we were able to analyze trends between ap- proximately 1960 and 2010 for 11 industrialized countries: the Northern European countries of Finland, Norway, and Sweden; the Western/central European coun- tries of France, Germany, the Netherlands, and the United Kingdom; the two Eng- lish-speaking countries of North America, Canada and the United States; and two Southern European countries, Italy and Spain.
3.2 Measures
3.2.1 Cohort fertility
In this analysis we will employ the cohort total fertility rate as our ultimate depend- ent variable, covering the fertility of cohorts born starting in 1920. The cohort total
(^106) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
fertility rate at age 40 (CTFR 40) is used instead of the actual CTFR at the end of the reproductive period at age 50. The considerable benefit of using the CTFR 40 is that it allows us to follow cohort fertility for ten additional cohorts. This is justified by the fact that only small amounts of childbearing occur when women are in their 40s – 0.6-2.6 percent in the 1940s-1950s birth cohorts in the countries analyzed in this ar- ticle – although these shares will be increasing in future cohorts. More importantly, the concern is focused mainly on the CTFR trends and these are almost identical at ages 40 and 50 ( HFD 2018). Information on cohort total fertility rates (CTFR) was obtained from the Human Fertility Database ( HFD 2018), a joint project of the Max Planck Institute for Demo- graphic Research and the Vienna Institute of Demography. The more commonly used measure, the period total fertility rate (PTFR), reflects not only changes in the quantum of fertility but also changes in the timing of childbearing ( Frejka 2011; So- botka et al. 2011a; Bongaarts/Feeney 1998). A PTFR decline might be the result of a decline in the quantum of fertility modified by later childbearing as, for instance, during the world-wide economic crisis late in the first decade of the 21st^ century across all the countries of Europe ( Comoli 2017; Sobotka et al. 2011b), while a PTFR increase might be caused by earlier childbearing or by a slowing down of childbear- ing postponement due to an economic recovery. In contrast to the PTFR, cohort total fertility rate trends (CTFRs) are a true expression of trends in fertility quantum ( Frejka 2011; Sobotka et al. 2011a; Bongaarts/Feeney 1998). Cohort measures in demography can be identified by the date of birth, a rule we follow in our fertility tables and graphs. However, when ascertaining the fertility ef- fect of the gender revolution and of its components, namely of trends of women’s labor force participation and trends of men’s involvement in domestic tasks, as well as of the trends of the relationship between these two components (see below), we center the cohort fertility values at age 27. This means that when associating trends of the gender revolution components (which are period measures) with fertil- ity levels and trends, we graph the completed cohort fertility rates (to age 40) of the respective cohorts as of when they reach age 27. Specifically, the gender revolution measures around 1960 are associated approximately with the completed cohort fertility rate of the 1933 birth cohort because this cohort reached age 27 in 1960; analogously for measures around 1970 the 1943 total cohort fertility rate is used as the 1943 birth cohort reached age 27 in 1970, and so on. (This will become clearer as we discuss the tables and graphs relating cohort fertility to period measures of gendered behaviors.) Why age 27? It is the age at which most women are in the early career building stages, yet already have enough information about their (likely) male partners to be able to estimate how many children are consistent with their likely second shift of work at home and in the labor force. We realize this is somewhat arbitrary, and women’s decisions about work and family might have been formed earlier (particu- larly among the older cohorts) and continued to evolve as they moved through their childbearing years. Our results, however, were not very sensitive to centering our cohort completed fertility rates at alternate ages, from age 20, when many women may already have plans for their next decades, to age 35, when at least some time
(^108) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
domestic hours (including childcare) for all men and women aged 20 to 59 years 2 on our 11 countries for the period 1961 to 2004. 3 We had to be less precise with dates, given the wide variability in when time-use surveys were conducted in these coun- tries, and made liberal use of interpolation to fill holes between surveys. Again, we created a sex ratio of these hours (but this time: male to female), the TDH-SR (total domestic hours sex ratio). Because in no country have men’s average domestic hours exceeded women’s, the TDH-SR, like the LFP-SR, ranges between 0 and 1. Finally, we constructed a third gender revolution measure, which we call the “second shift ratio” (SSR). It is the ratio of our labor force participation ratio (the LFP-SR) divided by the domestic hours ratio (the TDH-SR), and normally ranges from 1 to 2, but exceptionally even exceeds 2. If these two ratios were equal, the combined ratio would equal one, meaning that women’s participation in the public sphere relative to men’s is roughly balanced by men’s participation in the private sphere relative to women’s. Values above one indicate that the gender gap in the public sphere has closed more than the gender gap in the private sphere, leaving women on average with a second shift of domestic work beyond their involvement in paid employment.
3.3 Methods
This is a descriptive analysis, which rests on our examination of trends between the 1960s and the early years of the 21st^ century in these four indicators for these 11 countries. As more data become available (especially for men’s involvement in the private sphere), it might become possible to use more advanced statistical tech- niques to test what are essentially our hunches about the patterns we see in these trends, as we will spell out in the analysis that follows. At this point, however, we simply describe patterns and make linkages among them for clusters of countries.
4 Results
We first examine trends in total cohort fertility and then link these trends to overall changes in the three ratio measures for our full set of countries. We group countries
(^2) Hence, the data include those with and without partners, parents and nonparents. With the exception of childcare, it is not clear how country variations in levels of parenthood and union membership would affect our results. It could introduce a small amount of reverse causality, given that gender equality appears to contribute to continued childbearing at the couple level, and contribute, as well, to union formation and retard union dissolution ( Goldscheider et al.
(^3) Unfortunately, we were not able to use more recent analyses of private sphere trends ( Altintas/ Sullivan 2016, 2017), as the first focused only on a subset of domestic chores, to the exclusion both of tasks normally done by men and of childcare, where men’s participation has recently been rising rapidly, at least in the United States ( Aguiar/Hurst 2007). The second study, while it did focus on childcare, was restricted to men with at least one child under the age of five.
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 109
into geographic regions. Each regional group generally shows similar patterns in these ratios, although with important exceptions.
4.1 Completed cohort fertility as of age 40
The levels and trends in cohort fertility evident in Table 1 and Figure 1 are shown both as of the year these women reached age 27 and the year of birth. For exam- ple, the values for 1960 are those for the birth cohort of 1933, who reached age 27 in 1960 (and reached their near complete fertility at age 40 in 1973). Similarly, the values for 2000 are those for the birth cohort of 1973, who reached age 27 in 2000 (and age 40 in 2013). In Northern Europe, the youngest 20 to 30 cohorts in Finland and Norway, born during the 1950s, 1960s and early 1970s, reached a fairly stable plateau close to replacement of around 2.0 births per woman (Fig. 1A). Sweden has actually expe- rienced an approximately replacement level of cohort fertility for over 50 cohorts. 4 The two countries in North America experienced relatively high fertility in the 1930s birth cohorts, with CTFRs above 3.0 followed by sharp declines between the cohorts born in the early 1930s up to those born around 1950 (Table 1 and Fig. 1C). They did not approach the replacement level of the north European countries until the birth cohorts of the early 1950s (like Norway), after which Canada continued a slow decline and the United States, unlike Northern Europe, experienced a slight increase. Somewhat different patterns characterize both Western/central Europe and Southern Europe (Table 1 and Fig. 1B, 1D). These two regions experienced much lower fertility among more recent cohorts than the two previously described re- gions (except France). France most closely resembles the countries of Northern Europe, reaching a similarly slightly declining CTFR plateau for the younger 1950s, 1960s and early 1970s birth cohorts at a level around 2.0, after prior declines. Co- hort fertility in Germany, the United Kingdom, and the Netherlands declined almost continuously among the youngest birth cohorts born between the late 1940s and early 1970s. This was also the case for the two countries in Southern Europe, except that their cohort fertility declined further and faster, reaching lowest-low CTFRs of 1.4 (Italy) and 1.3 (Spain) in the early 1970s birth cohorts. An important result of these differing patterns of change is the reversal in coun- try fertility levels that has been so frequently remarked upon. Italy and Spain had the highest fertility in the 1940s birth cohorts and early in the 21st^ century have the lowest fertility in the mid-1970s cohorts, as a result of their sharp declines. Finland, Norway and Sweden, which had relatively low fertility among the 1940s cohorts, now have among the highest fertility in the mid-1970s cohorts in this set of countries as a result of their relative stability.
(^4) For each of these countries, childbearing for women in their late 30s in the most recent cohort might have been depressed by the world-wide Great Depression, which hit when they were ap- proximately aged 34.
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 111
In sum, among the recent youngest approximately 40 cohorts, born between the 1930s and the early 1970s, fertility declined in all the analyzed countries (Table 1, last column, lower panel; Fig. 1). Further, the overall propensity for contemporary fertility to fall off among western countries is confirmed by the moderate fertility declines among the youngest 15 to 20 birth cohorts, those born in the 1960s and early 1970s, in all countries with the exception of the United States (Table 1, column 1990-2000, lower panel; Fig. 1). Given these various country patterns of cohort to- tal fertility rates, the question we now ask is how these fertility patterns of the late twentieth century relate to changes in our three measures of gender change – labor force participation, domestic hours, and women’s second shift.
4.2 Labor force participation: The LFP-SR
The massive change in the public sphere gender balance between 1960 and 2010 is evident in Table 2 and Figure 2, which show, by region and country, the sex ratio of
Fig. 1: Cohort total fertility rates at age 40, Northern, Western/Central Europe, North America, Southern Europe, birth cohorts 1920-
1920 1930 1940 1950 1960 1970 1980 Birth cohort
Finland Norway Sweden
A - Northern Europe Cohort total fertility rate at age 40
1920 1930 1940 1950 1960 1970 1980 Birth cohort
France Germany Netherlands United Kingdom
B - Western/Central Europe Cohort total fertility rate at age 40
C - North America Cohort total fertility rate at age 40
D - Southern Europe Cohort total fertility rate at age 40
1920 1930 1940 1950 1960 1970 1980 Birth cohort
United States Canada
1920 1930 1940 1950 1960 1970 1980 Birth Cohort
Italy Spain
Source: HFD 2018
(^112) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
female to male labor force participation, the LFP-SR. Clearly, at least at the macro- level, the bivariate relationship between changing cohort fertility and the changing ratio of female to male labor force participation is inverse. Further, in most cases the pattern of change at least loosely follows that for cohort fertility (in reverse), with
Tab. 2: Ratio of female to male labor force participation rates, LFP-SR, selected countries, absolute values and rates of change, 1960-
Country and Ratio of female to male labor force participation rate LFP-SR in year Region 1960 1970 1980 1990 2000 2010 2014
Northern Europe Finland 0.620 0.722 0.862 0.908 0.929 0.946 0. Norway 0.305 0.549 0.714 0.845 0.899 0.936 0. Sweden 0.435 0.640 0.832 0.952 0.937 0.931 0. Western/Central Europe France 0.471 0.530 0.655 0.760 0.830 0.880 0. Germany 0.568 0.452 0.559 0.703 0.802 0.860 0. Netherlands 0.206 0.229 0.367 0.660 0.783 0.867 0. United Kingdom 0.411 0.579 0.660 0.770 0.820 0.850 0. North America Canada 0.374 0.531 0.602 0.806 0.859 0.910 0. United States 0.475 0.521 0.659 0.795 0.845 0.863 0. Southern Europe Italy 0.369 0.297 0.421 0.566 0.628 0.697 0. Spain 0.210 0.260 0.310 0.521 0.658 0.816 0.
Country and Average annual rate of change in period (in percent) Region 1960- 1970- 1980- 1990- 2000- 2010- 1970 1980 1990 2000 2010 2014
Northern Europe Finland 1.5 1.8 0.5 0.2 0.2 0. Norway 5.9 2.6 1.7 0.6 0.4 0. Sweden 3.9 2.6 1.3 -0.2 -0.1 0. Western/Central Europe France 1.2 2.1 1.5 0.9 0.6 0. Germany -2.3 2.1 2.3 1.3 0.7 0. Netherlands 1.1 4.7 5.9 1.7 1.0 0. United Kingdom 3.4 1.3 1.5 0.6 0.4 0. North America Canada 3.5 1.3 2.9 0.6 0.6 0. United States 0.9 2.3 1.9 0.6 0.2 -0. Southern Europe Italy -2.2 3.5 3.0 1.0 1.0 1. Spain 2.1 1.8 5.2 2.3 2.2 1.
Source: Authors’ calculations based on World Bank Databank 2016; OECD 2016; Mitchell 1998; ILO 1960
(^114) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
end of the period in 2014. The LFP-SR increased rapidly to settle at a level around 0.90 in 1990 in Northern Europe; in Western/central Europe and in North America the process was more gradual, reaching close to 0.90 only in 2010; and in Southern Europe, levels all along were lower than elsewhere, reaching 0.70 to 0.80 early in the 21 st^ century. Hence, in more recent years there has been substantial convergence. This pattern of convergence is even evident within regions. The three Northern Eu- ropean countries show quite different LFP-SRs in 1960, ranging from Finland’s high ratio of 0.62 down to Norway’s 0.31, with Sweden in between. However, their mas- sive increases after 1960 led to LFP-SRs above 0.80 in the 1980s, and essentially to total convergence at about 0.90 around 1990, even reaching 0.95 in 2014. (Norway lagged slightly for most of this period.) As these trends suggest, however, there was considerable variation in tempo- ral patterns. In the early part of the period (1960-1970), some countries were rap- idly accelerating (Norway, Sweden), with annual rates of change of 5.9 percent and 3.9 percent (Table 2). Some were barely increasing (the Netherlands, France, and the U.S.), with annual rates of change of around 1 percent; and some were even decreasing (Germany,^7 Italy), with annual rates of change of -2 percent. Yet all these countries experienced declines in cohort fertility during that period, and the rates of fertility decline only very loosely matched the rates of increase in the LFP-SR. For example, between 1960 and 1970, Finland experienced the slow- est increase in the LFP-SR (admittedly from a high base) but the fastest decline in cohort fertility. Nevertheless, the overall patterns of fertility change and change in the sex ratios for labor force participation suggest that the first half of the gender revolution is far from complete in many of these countries, and its negative effects on fertility are likely far from complete, as well.
4.3 Domestic work hours: The TDH-SR
The two-stage approach to the gender revolution posits that change came first in the public sphere, with the increase in female labor force participation, and that this was followed by change in the private sphere, as men, after some delay, increased their engagement in the tasks of the home ( Anderson/Kohler 2015; Esping-Anders- en/Billari 2015; Goldscheider et al. 2015). Our results suggest, however, that this is an oversimplification, at least to some extent. From the earliest days of the gender revolution, as early as the 1960s, the general pattern is that men have been increas-
(^7) The data for Germany combine the former East Germany with the former West Germany. This is somewhat problematic in this case, as the gender systems in these two (then) countries evolved in very different ways during the 40+ years that they were separated. The former East Germany emphasized gender equality (particularly in the public sphere); the proportions of women who were reported to be housewives were very different ( Klein/Nauck 2005). In the former East Germany, only 9 percent of women and 10 percent of men prefer that when a cou- ple become parents “one should stay at home” during the preschool years, while in the former West Germany, 56 percent of women and 63 percent of men gave that response ( Dannenbeck 1992). Unlike in the US, in West Germany most couples are not concerned about domestic equality ( Bellani et al. 2017).
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 115
ing their share of hours in the private sphere relative to women (Table 3 and Fig. 3). Although much of the increase in the sex ratio of total domestic time (TDH-SR) in the early years reflects women’s massive decline in housework hours, men did ex-
Tab. 3 : Ratio of male to female total domestic hours worked, TDH-SR, selected countries, absolute values and rates of change, 1960-1969 to 2000-
Country and Ratio of male to female total domestic work TDH-SR in period Region 1960-69 1970-75 1976-84 1985-89 1990-94 1995-99 2000- “Central years” 1965 1972.5 1980 1987.5 1992 1997 2002
Northern Europe Finland 0.486 0.578 0.583 0. Norway 0.341 0.452 0.504 0.556 0.591 0. Sweden 0.592 0.642 0. Western/Central Europe France 0.268 0.310 0.362 0.415 0.467 0. Germany 0.330 0.368 0.405 0.442 0.480 0.505 0. Netherlands 0.331 0.350 0.390 0.406 0.472 0. United Kingdom 0.244 0.300 0.358 0.416 0.458 0.500 0. North America Canada 0.372 0.502 0.464 0.522 0. United States 0.291 0.413 0.472 0.531 0.588 0.671 0. Southern Europe Italy 0.184 0.217 0.251 0. Spain 0.
Country and Average annual rate of change in period (in percent) Region 1965- 1972.5- 1980- 1987.5- 1992- 1997- 1972.5 1980 1987.5 1992 1997 2002
Northern Europe Finland 2.3 0.2 0. Norway 3.8 1.4 2.2 1.2 1. Sweden 1.6 1. Western/Central Europe France 1.9 2.1 1.8 2.6 2. Germany 1.4 1.3 1.2 1.8 1.0 1. Netherlands 0.7 1.4 0.9 3.0 0. United Kingdom 2.8 2.3 2.0 2.1 1.8 1. North America Canada 4.0 -1.0 2.6 1. United States 4.7 1.8 1.6 2.3 2.6 -1. Southern Europe Italy 3.7 2.9 2. Spain
Note: Values in italics are interpolated
Source: Authors’ calculations based on Kan et al. 2011
The Two-Part Gender Revolution, Women’s Second Shift and Changing Cohort Fertility • 117
early years. The timing is similar, however, both beginning to increase as early as the 1960s. Although the general level of the TDH-SR remained lower than the LFP-SR throughout the period, the regional variation in TDH-SR trends fairly closely re- sembles the regional variation we observed for the LFP-SR, with high levels for the North European countries and low levels for the South European countries, with the Western/central countries in between. The major differences in regional patterns between these two sets of ratios (LFP and TDH) are for the two North American countries, and most dramatically, for the United States. While these two countries ranked among those in the middle for the LFP-SR, they rank among the North Euro- pean leaders for the TDH-SR, again, particularly the United States, despite the dip or “stall” ( England 2010) in the recent period. Hence, the overall patterns of change in gender behavior in the private sphere of the family appear to do little to account for change in cohort fertility. Taken together, one could even argue that increasing men’s share (the TDH-SR) reduces fertility, instead of increasing it, given that as with the LFP-SR, the relationships are inverse – increases in men’s share, decreases in cohort fertility. But what might be operating is not the absolute trends in the LFP-SR and the TDH-SR, but their relationships with each other – women’s second shift – to which we now turn.
4.4 Relating the two ratios: the second shift
For our final measure of gender change, we calculated a “ratio of ratios” for these countries (Table 4 and Fig. 4), dividing the public sphere ratio by the private sphere ratio for each country and time. This is our indicator of the size of women’s “second shift,” albeit at the macro level, as it relates how close women have come to equality in the public sphere (the numerator) to how close they have come to equality in the private sphere (the denominator). Hence, in this case, the “second shift ratio” (SSR) throughout is almost always above one, 8 indicating more equal gender sharing in the public sphere than in the private sphere. The higher the value, the greater the second shift. The overall impression of the SSR data shown in Table 4 is that women’s burden is considerably greater than men’s. There was little pattern and little change prior to 1990, as shown by the largely random pattern of percent change in the lower part of Table 4. Thereafter, however, the major trend is (slightly) down, indicating a reduc- tion in women’s second shift. It is worth noting that the SSR in the year 2000 has converged around the value of 1.5 (except for the Southern European countries). In an evaluation of the overall picture of combined activities in the public and the pri- vate spheres, this might be interpreted as women expending about 50 percent more effort than men towards their families’ well-being.
(^8) The Netherlands’ strange early value (0.691) reflects women’s unusually low ratio of labor force participation rather than anything unusual in Dutch men’s domestic participation.
(^118) • Tomas Frejka, Frances Goldscheider, Trude Lappegård
Tab. 4: Second shift ratio, SSR, selected countries, absolute values and rates of change, 1960 to 2000
Country and Region Second shift ratio (LFP-SR/TDH-SR) 1960 1970 1980 1990 2000
Northern Europe Finland 1.772 1.565 1. Norway 1.613 1.578 1.595 1. Sweden 1. Western/Central Europe France 1.757 1.712 1.807 1. Germany 1.720 1.230 1.379 1.524 1. Netherlands 0.691 1.048 1.659 1. United Kingdom 1.687 1.927 1.843 1.763 1. North America Canada 1.426 1.200 1. United States 1.632 1.264 1.396 1.421 1. Southern Europe Italy 2.820 2. Spain 1.
Country and Region Average annual rate of change in period (in percent) 1960-1970 1970-1980 1980-1990 1990-
Northern Europe Finland -1.2 0. Norway -0.2 0.1 -0. Sweden Western/Central Europe France -0.3 0.5 -0. Germany -3.4 1.1 1.0 0. Netherlands 4.2 4.6 -0. United Kingdom 1.3 -0.4 -0.4 -1. North America Canada -1.7 3. United States -2.6 1.0 0.2 -0. Southern Europe Italy -1. Spain
Source: Authors’ calculations based on Tables 2 and 3
