Final Report Summary - FAMILIY POLARIZATION (STRATIFIED FAMILY DYNAMICS:POLARIZING TRENDS IN COUPLE BEHAVIOUR AND PARENTING)
Publishable summary of achievements
The point of departure for my FAMPOL project was the evermore evident lack of fit between the reigning theories of family change and observable -- albeit recent -- changes in family outcomes. Gary Becker's theoretical framework would predict a decline in marriages, heightened couple instability, and a fall in fertility in tandem with the decline of traditional gender specialization in the home. The Second Demographic Transition (SDT) thesis, as articulated by Ron Lesthaeghe and Dirk van de Kaa, predicts similarly a 'family erosion' scenario but in this case caused by 'postmodern' value change which promotes individualism and self-realization over long-term and binding commitments.
On many key indicators of family life one begins to observe a reversal of past trends over the past two decades or so. In a number of countries we see an upward trend in fertility and marriages, as well as a fall in divorce. To be sure, this trend is far from universal, and is most clearly evident in Northern Europe and North America. It was this reversal that triggered the key question which my ERC project has sought to answer: Is the family returning and if so which are the conditions that promote the reversal of past trends?
This said, my project was designed to pursue two parallel objectives: one, to demonstrate empirically the degree to which the reversal is genuine and here to stay; two, to develop an alternative theoretical model which is better able to understand and explain contemporary trends in family dynamics.
I have dedicated considerable effort to develop and formalize a core theoretical model. Building on multiple equilibrium theory from economics, the model takes as its point of departure the dynamics of women's role change. Contrary to the Becker theory, I predict a curvi-linear (U-shaped) function, according to which we should expect family erosion (i.e. a rise in divorce propensities and singlehood, declining marriages and fertility) in the initial stages of women's role transformation. But as society adapts, both at the micro-level of partnerships and at the macro-level of social institutions (i.e. the labor market and the welfare state), we should observe a return of the family. Put differently, when gender egalitarian norms take hold both within partnerships and in working life, this is when we should see an upturn in marriage propensities, fertility, and in couple stability. I have articulated and developed this theoretical model in several publications. A purely theoretical and formalized version (with Francesco Billari) has been published in Population and Development Review. A less formalized version was published also in my 2016 monograph, Families in the 21st Century.
Empirically, my ERC project has tested this model on all the key dimensions of family life. As the attached bibliography shows, I and my collaborators have dedicated considerable effort to identifying the precise conditions that promote a return to fertility levels which correspond to citizens' preferences. This latter point is crucial: in my framework fertility is not an objective per se. To the contrary, I take a 'welfare perspective', meaning that the relevant fertility measure is one that captures citizens' desired number of children.
In one study (co-authored with Bruno Arpino and Lea Pessin, and published in European Sociological Review), we approached the issue applying multi-level estimation for a large number of countries over the past 3 decades. Here we find very clear evidence that the return to higher fertility is closely connected with the degree to which gender egalitarian norms have penetrated society (and in particular with the degree to which both men and women are similarly 'gender egalitarian'). In another study (with Roberta Rutigliano) I broach the issue from a very different angle. Here we focus on the micro-level in two countries which represent especially stark differences in gender egalitarianism -- Norway and Spain. And here we examine the probabilities of first and second births in both married and cohabiting couples. We apply very complex simultaneous equation estimation so as to minimize bias that may arise from endogeneity and selection (e.g. choice of partnership may be driven by fertility desires). Our findings are quite surprising. Contrary to our expectations we find almost identical fertility probabilities (for first births) among Spanish and Norwegian cohabiting couples. We also uncovered a surprising degree of couple stability among Spanish cohabitors. This suggests that on one important dimension also Southern Europe (or at least Spain) is moving towards a 'return to family' equilibrium. And yet, the study demonstrates very clearly that this move is very partial since second birth probabilities in Spain remain very low, be it in married or cohabiting unions.
My ERC project has also focused heavily on changing divorce propensities. Here again, the key question is whether we observe greater conjugal stability in tandem with the adoption of gender egalitarian norms. In two empirical studies (with the participation of Daniela Bellani and Lea Pessin) we utilize German and US micro-level panel data so as to be able to follow partnerships over 15 years. And here we pay special attention to social class differences (measured by education level). This is because we would expect the higher educated to be the frontrunners in the adoption of gender symmetric attitudes and behavior. For both countries, but much more significantly so in the US, we find – as predicted – a substantial marital stability dividend among couples who display gender symmetric behaviour in the home (i,.e. in terms of domestic work), and this is clearly far more pronounced within higher educated couples.
In a third study (co-authored with Anders Holm), I examine the divorce effects from a very different angle. Here the point of departure is whether divorce risks increase significantly when the female partner becomes the dominant income earner. Utilizing Danish and Swedish registry data, I follow two marriage cohorts: the oldest partnered in 1980; the youngest in 2000. And we compare divorce odds for both low and high educated couples. For both countries we find a clear shift: the risk of divorce declines to less than half in the young cohort, compared to the older one. And the ‘social class’ effect is very evident. The divorce risk is roughly twice as high for the low educated. Indeed, it disappears altogether for the higher educated in the youngest cohort. I am currently extending these analyses also to German data so as to be able to test the female-dominance effect in a society that remains far more gender traditional.
Given that in many countries cohabitation has become functionally equivalent to marriage, I decided to test the hypothesis related to partnering by focusing on its opposite, namely on those who never partnered. In collaboration with Daniela Bellani and Lesia Nedoluzhko, we applied a multi-level modelling design comparing both across time (1980s-2012) and countries. The key question here was whether levels of lifelong singlehood vary by the degree to which gender egalitarianism is diffused. In other words, the research design here is very similar to that I used in the fertility study. In this study we find very clear evidence in favour of my theoretically argued U-shaped function: lifelong singlehood rates are almost 5 times higher in gender unequal societies than in those where gender egalitarianism has become normative. Additionally, we find particularly strong effects when controlling for education. In the past, highly educated women faced the highest risk of non-partnering; today it is the opposite – but only in those nations that represent advanced levels of gender egalitarianism.
From the very beginning, the project has also included a strong focus on parenting and child outcomes. These issues have been addressed on a number of distinct dimensions. Firstly, I have analyzed how parental effects – both in terms of differentials in parenting intensity and in terms of their socio-economic conditions – influence children’s educational attainment across childhood. I had initially expected to find heightened polarization in both parenting and child outcomes in light of the widening gap between low and high educated in terms of couple stability (and rising income inequalities as well). But most of my many analyses did not support this hypothesis. In the last section of my monograph (Families in the 21st Century), I estimate both differentials in parenting intensity (especially using time dedication information) and children’s transitions from the lower to the higher education levels across a number of countries – using data that allow me to control also for the children’s cognitive abilities. The latter is important because it allows me to identify inequalities that are not related to skill differences – i.e. I can identify net mobility probabilities.
The analyses however, do not point towards any rising degree of inequality when comparing cohorts born in the 1960s- 1980s. To the contrary, in the Scandinavian countries especially, I find a significant degree of equalization. In a related study (with Julia Cordero), we analyze how the intensity of mothers’ and fathers’ time dedication to their children at the pre-school stage influences their later probability of completing upper secondary education in Germany (using the GSOEP panel data). Here we find clear evidence that it is more the content and style of parenting, rather than the sheer volume, which is key to children’s educational attainment. The key lies in active and intellectually stimulating activities such as reading and playing with the children.
And finally – in collaboration with a team of US child research experts (Jane Waldfogel and others), I have undertaken one comparative Denmark-US study which examined the degree to which external childcare services and kindergarten attendance help equalize children’s educational abilities across the social classes. In this study we find a significant equalizing effect in Denmark, but not in the US. All the analyses suggest that this difference is primarily due to differences in early childcare service quality. In a parallel second study, I have (in collaboration with Johanna Dammrich) conducted a much broader nation comparison – here again estimating early childcare effects on children’s later abilities (measured at ages ca. 10-11 and 15). In this study we test the ‘quality’ thesis more directly and find very clear and strong evidence that early childcare is equalizing across socio-economic strata only when three conditions are met: a low child:staff ratio, that the staff has higher level pedagogical education, and that attendance is intensive (30+ hours per week).
As is evident from both the attached bibliography and the foregoing summary, I have very actively sought to include my PhD supervisees in my research as well as in publications. I believe that the ERC project has been a success in terms of helping produce high-quality PhD's. Diederik Boertien, Pablo Gracia, Marta Seiz, Francesca Luppi, and Sander Wagner and Lea Pessin have all completed the PhD. And all of them have succeeded in obtaining positions in excellent universities. Diederik Boertien and Pablo Gracia are at the European University Institute, Marta Seiz is at the CSIC in Madrid, Lea Pessin is now a post-doc at Penn State University, and Sander Wagner is at ENID in Paris.
Roberta Rutigliano is on schedule to defend in the Spring, 2017 (she already has completed two of the three empirical chapters). Alessandro Di Nallo and Luize Ratniece are scheduled to defend in the Fall of 2017 (both have completed the first empirical chapter). Jorge Cimentada has just completed his first year of doctoral studies. One doctoral student, Stefano Filauro, was forced to interupt studies due to family problems in his native Italy.
The point of departure for my FAMPOL project was the evermore evident lack of fit between the reigning theories of family change and observable -- albeit recent -- changes in family outcomes. Gary Becker's theoretical framework would predict a decline in marriages, heightened couple instability, and a fall in fertility in tandem with the decline of traditional gender specialization in the home. The Second Demographic Transition (SDT) thesis, as articulated by Ron Lesthaeghe and Dirk van de Kaa, predicts similarly a 'family erosion' scenario but in this case caused by 'postmodern' value change which promotes individualism and self-realization over long-term and binding commitments.
On many key indicators of family life one begins to observe a reversal of past trends over the past two decades or so. In a number of countries we see an upward trend in fertility and marriages, as well as a fall in divorce. To be sure, this trend is far from universal, and is most clearly evident in Northern Europe and North America. It was this reversal that triggered the key question which my ERC project has sought to answer: Is the family returning and if so which are the conditions that promote the reversal of past trends?
This said, my project was designed to pursue two parallel objectives: one, to demonstrate empirically the degree to which the reversal is genuine and here to stay; two, to develop an alternative theoretical model which is better able to understand and explain contemporary trends in family dynamics.
I have dedicated considerable effort to develop and formalize a core theoretical model. Building on multiple equilibrium theory from economics, the model takes as its point of departure the dynamics of women's role change. Contrary to the Becker theory, I predict a curvi-linear (U-shaped) function, according to which we should expect family erosion (i.e. a rise in divorce propensities and singlehood, declining marriages and fertility) in the initial stages of women's role transformation. But as society adapts, both at the micro-level of partnerships and at the macro-level of social institutions (i.e. the labor market and the welfare state), we should observe a return of the family. Put differently, when gender egalitarian norms take hold both within partnerships and in working life, this is when we should see an upturn in marriage propensities, fertility, and in couple stability. I have articulated and developed this theoretical model in several publications. A purely theoretical and formalized version (with Francesco Billari) has been published in Population and Development Review. A less formalized version was published also in my 2016 monograph, Families in the 21st Century.
Empirically, my ERC project has tested this model on all the key dimensions of family life. As the attached bibliography shows, I and my collaborators have dedicated considerable effort to identifying the precise conditions that promote a return to fertility levels which correspond to citizens' preferences. This latter point is crucial: in my framework fertility is not an objective per se. To the contrary, I take a 'welfare perspective', meaning that the relevant fertility measure is one that captures citizens' desired number of children.
In one study (co-authored with Bruno Arpino and Lea Pessin, and published in European Sociological Review), we approached the issue applying multi-level estimation for a large number of countries over the past 3 decades. Here we find very clear evidence that the return to higher fertility is closely connected with the degree to which gender egalitarian norms have penetrated society (and in particular with the degree to which both men and women are similarly 'gender egalitarian'). In another study (with Roberta Rutigliano) I broach the issue from a very different angle. Here we focus on the micro-level in two countries which represent especially stark differences in gender egalitarianism -- Norway and Spain. And here we examine the probabilities of first and second births in both married and cohabiting couples. We apply very complex simultaneous equation estimation so as to minimize bias that may arise from endogeneity and selection (e.g. choice of partnership may be driven by fertility desires). Our findings are quite surprising. Contrary to our expectations we find almost identical fertility probabilities (for first births) among Spanish and Norwegian cohabiting couples. We also uncovered a surprising degree of couple stability among Spanish cohabitors. This suggests that on one important dimension also Southern Europe (or at least Spain) is moving towards a 'return to family' equilibrium. And yet, the study demonstrates very clearly that this move is very partial since second birth probabilities in Spain remain very low, be it in married or cohabiting unions.
My ERC project has also focused heavily on changing divorce propensities. Here again, the key question is whether we observe greater conjugal stability in tandem with the adoption of gender egalitarian norms. In two empirical studies (with the participation of Daniela Bellani and Lea Pessin) we utilize German and US micro-level panel data so as to be able to follow partnerships over 15 years. And here we pay special attention to social class differences (measured by education level). This is because we would expect the higher educated to be the frontrunners in the adoption of gender symmetric attitudes and behavior. For both countries, but much more significantly so in the US, we find – as predicted – a substantial marital stability dividend among couples who display gender symmetric behaviour in the home (i,.e. in terms of domestic work), and this is clearly far more pronounced within higher educated couples.
In a third study (co-authored with Anders Holm), I examine the divorce effects from a very different angle. Here the point of departure is whether divorce risks increase significantly when the female partner becomes the dominant income earner. Utilizing Danish and Swedish registry data, I follow two marriage cohorts: the oldest partnered in 1980; the youngest in 2000. And we compare divorce odds for both low and high educated couples. For both countries we find a clear shift: the risk of divorce declines to less than half in the young cohort, compared to the older one. And the ‘social class’ effect is very evident. The divorce risk is roughly twice as high for the low educated. Indeed, it disappears altogether for the higher educated in the youngest cohort. I am currently extending these analyses also to German data so as to be able to test the female-dominance effect in a society that remains far more gender traditional.
Given that in many countries cohabitation has become functionally equivalent to marriage, I decided to test the hypothesis related to partnering by focusing on its opposite, namely on those who never partnered. In collaboration with Daniela Bellani and Lesia Nedoluzhko, we applied a multi-level modelling design comparing both across time (1980s-2012) and countries. The key question here was whether levels of lifelong singlehood vary by the degree to which gender egalitarianism is diffused. In other words, the research design here is very similar to that I used in the fertility study. In this study we find very clear evidence in favour of my theoretically argued U-shaped function: lifelong singlehood rates are almost 5 times higher in gender unequal societies than in those where gender egalitarianism has become normative. Additionally, we find particularly strong effects when controlling for education. In the past, highly educated women faced the highest risk of non-partnering; today it is the opposite – but only in those nations that represent advanced levels of gender egalitarianism.
From the very beginning, the project has also included a strong focus on parenting and child outcomes. These issues have been addressed on a number of distinct dimensions. Firstly, I have analyzed how parental effects – both in terms of differentials in parenting intensity and in terms of their socio-economic conditions – influence children’s educational attainment across childhood. I had initially expected to find heightened polarization in both parenting and child outcomes in light of the widening gap between low and high educated in terms of couple stability (and rising income inequalities as well). But most of my many analyses did not support this hypothesis. In the last section of my monograph (Families in the 21st Century), I estimate both differentials in parenting intensity (especially using time dedication information) and children’s transitions from the lower to the higher education levels across a number of countries – using data that allow me to control also for the children’s cognitive abilities. The latter is important because it allows me to identify inequalities that are not related to skill differences – i.e. I can identify net mobility probabilities.
The analyses however, do not point towards any rising degree of inequality when comparing cohorts born in the 1960s- 1980s. To the contrary, in the Scandinavian countries especially, I find a significant degree of equalization. In a related study (with Julia Cordero), we analyze how the intensity of mothers’ and fathers’ time dedication to their children at the pre-school stage influences their later probability of completing upper secondary education in Germany (using the GSOEP panel data). Here we find clear evidence that it is more the content and style of parenting, rather than the sheer volume, which is key to children’s educational attainment. The key lies in active and intellectually stimulating activities such as reading and playing with the children.
And finally – in collaboration with a team of US child research experts (Jane Waldfogel and others), I have undertaken one comparative Denmark-US study which examined the degree to which external childcare services and kindergarten attendance help equalize children’s educational abilities across the social classes. In this study we find a significant equalizing effect in Denmark, but not in the US. All the analyses suggest that this difference is primarily due to differences in early childcare service quality. In a parallel second study, I have (in collaboration with Johanna Dammrich) conducted a much broader nation comparison – here again estimating early childcare effects on children’s later abilities (measured at ages ca. 10-11 and 15). In this study we test the ‘quality’ thesis more directly and find very clear and strong evidence that early childcare is equalizing across socio-economic strata only when three conditions are met: a low child:staff ratio, that the staff has higher level pedagogical education, and that attendance is intensive (30+ hours per week).
As is evident from both the attached bibliography and the foregoing summary, I have very actively sought to include my PhD supervisees in my research as well as in publications. I believe that the ERC project has been a success in terms of helping produce high-quality PhD's. Diederik Boertien, Pablo Gracia, Marta Seiz, Francesca Luppi, and Sander Wagner and Lea Pessin have all completed the PhD. And all of them have succeeded in obtaining positions in excellent universities. Diederik Boertien and Pablo Gracia are at the European University Institute, Marta Seiz is at the CSIC in Madrid, Lea Pessin is now a post-doc at Penn State University, and Sander Wagner is at ENID in Paris.
Roberta Rutigliano is on schedule to defend in the Spring, 2017 (she already has completed two of the three empirical chapters). Alessandro Di Nallo and Luize Ratniece are scheduled to defend in the Fall of 2017 (both have completed the first empirical chapter). Jorge Cimentada has just completed his first year of doctoral studies. One doctoral student, Stefano Filauro, was forced to interupt studies due to family problems in his native Italy.