Fertility & reproductive health strand abstracts

Strand organisers:

Professor Melinda Mills & Dr. Nicola Barban, University of Oxford 

Fertility & reproductive health:  Policy, economics & fertility. Monday 12 September 1:30pm 

Do family policies have an impact on the fertility rate? A comparative study of the Swiss cantons 
Marion Burkimsher, University of Lausanne (Affil.)   

Switzerland is a federal republic composed of 26 cantons. Each canton has considerable leeway in developing its own family policy with respect to financial support and childcare provision. The total fertility rate for Switzerland reached a minimum of 1.38 in 2001 and has been rising since then. However, the increase in TFR from 2001 to 2014 varies considerably between cantons, from 0.05 to 0.34. Different family policies are not the only potential explanation, however. Alternative explanations include differential migration, changes in GDP and employment rates, and the religious make-up and associated attitudinal differences of each canton. Variations in birth timing is another potential explanatory factor; postponement of childbearing, particularly for first births, has been depressing period fertility measures for the past 40 years, but this is now slowing in some parts. The structure of the population by parity at the turn of the millennium could also account for the differentials in TFR increases: if there were a high proportion of low parity women, then these could have increased the likelihood of an area to experience a rebound in fertility as a ‘catch-up’ process. Using comprehensive birth registration data for each canton, age- and birth order-specific fertility rates are calculated for the Swiss and foreign populations for the period 2000-2014. The contributory effect of each factor mentioned above to the upturn in TFR can then be assessed. The results of this study could point to where family policies can best be focussed to have maximum effect in raising the fertility rate 


The effect of the economic crisis on fertility behaviour in southern Europe: The example of Andalusia
Diego Ramiro-Fariñas 1, J. Viciana 2, Victor Montañes 2, 1 Spanish National Research Council, 2 Francisco Institute of Statistics and Cartography of Andalusia.   

This paper will try to describe and analyze the effect of the economic crisis on fertility behaviour by educational attainment and employment status in Southern Europe, using as an example one the most deprived areas in the South of Spain: Andalusia. Using the Longitudinal Database of the Population of Andalusia we estimated the fertility of women with permanent residence in Andalusia during the period 2002-2013 using educational attainment and employment status reported in the 2011 census. 


Low Educational Attainment and UK Fertility Change in Times of Austerity
Elspeth Graham, Albert Sabater, Department of Geography and Sustainable Development & ESRC Centre for Population Change, University of St Andrews    

Previous studies of the relationship between education and fertility in advanced societies have focused more on fertility postponement among highly educated women and less on fertility among women with low educational attainment. Recent fertility changes in the UK, including increases in the average age of mothers at first birth and a downward trend in live births since 2013, suggest that the economic recession may be having a negative impact on reproduction. However, it is not yet clear whether women with low education are differentially affected both by the economic downturn itself and by the austerity policies introduced by the UK government in response. This paper investigates recent fertility trends among women with low education and the possible effects of new welfare policies introduced in 2010. Using data from the Family Resources Survey, we examine the probability of having a(nother) birth before and after the implementation of austerity measures. A difference-in-differences approach is used to compare fertility change prior to (2007-2010), and after (2011-2014), the welfare reforms were introduced. Women aged 20-45 with low education are compared with all other women in the same age group to test for differential trends over the study period. Macro-controls that may affect fertility, such as regional house prices and unemployment rates, are included in the analysis. While we find differential fertility change over the period of economic recession, there is no convincing evidence suggesting that austerity policies have influenced the fertility of women with low education post-2010. 


Income, welfare and the transition to third birth in the UK
Juliet Stone, Ann Berrington, ESRC Centre for Population Change, University of Southampton   

Researchers have tended to assume that social differentials in patterns of childbearing in the UK result, at least in part, to the means-tested welfare system in the UK that provides a safety-net for low income (often young) parents. Popular and media discourse often assumes a U-shaped relationship between income and fertility, whereby large families are more common either among the poor, whose childbearing is assumed to be to be subsidised by welfare benefits, or among the rich whose income can be used to offset the costs of children. However, there is little empirical evidence as to how income and benefit receipt affect childbearing in the UK in the 21st century. This paper provides new understanding by investigating the extent to which household income and the proportion of income derived from benefits are associated with the propensity for women to have a third birth. Using data from the United Kingdom Household Panel waves 1 to 5 (2009-2014) we focus on two sub-sets of women: all those with two children at baseline (including lone parents), and secondly all those with two children living in a couple at baseline. We show that while an association between welfare receipt, overall household income and parity progression exists, this is largely explained by the timing of childbearing in different socioeconomic groups. We consider our findings with reference to social theories of fertility including those of Bourdieu, and how habitus and related preferences for social and cultural capital shape women’s childbearing behaviour. 


Fertility & reproductive health:  Contraception & abortion. Monday 12 September 4:45 pm 

Reconsidering (In)Equality in the Use of IUDs: The Changing Context of U.S. Contraceptive Choice across the Reproductive Lifespan
Megan M. Sweeney, Mieke C. W. Eeckhaut, Jessica D. Gipson, University of California, Los Angeles   

Scholars and policymakers frequently point to poverty’s association with early and unintended childbearing in the United States. Sawhill (2014) particularly highlights the potential for long-acting reversible contraceptives, such as the intrauterine device (IUD), to reduce poverty and inequality. Yet our understanding of social-class disparities in IUD use remains limited. Although policy discussions focus particularly on the impact of first birth timing on women’s subsequent socioeconomic attainment, little work directly investigates how contraceptive choice varies across three key stages of a woman’s reproductive lifespan: the time before a first birth occurs (childbearing “postponers”), the time between births (childbearing “spacers”), and the time after the end of intended childbearing (childbearing “limiters”). In our sample of 6,189 contracepting women of reproductive age in the 2008-13 rounds of the U.S. National Survey of Family Growth (NSFG), we find IUD use to be considerably more common among childbearing spacers than among childbearing postponers or limiters. We also show that our understanding of socioeconomic disparities in IUD use changes considerably when viewed through the lens of the reproductive lifespan. Use of this highly-effective, reversible method is associated with educational advantage among childbearing postponers and limiters, but is associated with educational disadvantage among childbearing spacers. We consider our findings in light of ongoing policy discussions about the potential for IUDs to reduce inequality and poverty among disadvantaged women and children. We argue that such discussions must take into account the multiple contraceptive goals women hold, which may include protection from disease as well as pregnancy, while remaining sensitive to the balance between ensuring contraceptive access and protecting against coercion. 


Sex-selective abortion: measuring the who and how often
Aditi Dimri 1, 2, Véronique Gille 3, Philipp Ketz1 2, 1 CORE, Université catholique de Louvain, 2, Paris School of Economics, 3 Université Paris-Dauphine   

Any imbalance in the aggregate sex ratio at birth gives us an estimate of the total number of sex-selective abortions. With increasing imbalance seen in Indian data, unknowns about the extent and nature of sex-selective abortions, which is an illegal procedure, is a glaring issue. In this paper, we provide a methodology to estimate using household level data, the fractions of populations undergoing sex-selective abortions, and also the number of repeated abortions. We use the increase in birth spacing time between two births as a consequence of abortion to identify these parameters of repeated abortion rates. Using maximum likelihood, we estimate a simple structural model of the joint probability of the birth being of a particular gender, and the birth happening at a fixed number of months given the gender of birth. We use three rounds of DHS India datasets to estimate the extent of sex-selective abortions over time and across subgroups of population. 


Contraceptive use behaviour before and after unintended birth in Colombia and Peru: does it change?
Ewa Batyra, London School of Social and Political Science 

In spite of high levels of contraceptive use, unintended childbearing is widespread in Latin America. In Colombia the proportion of births reported as unintended increased in the last 20 years from 35 to 55%, in Peru it has been stable at 55%. Contraceptive use has been widely studied as a determinant of unintended birth. However, little is known about contraceptive use behaviour after unintended birth, with scant evidence from low and middle income countries. This research, using the reproductive calendar of DHS, aims to examine whether contraceptive use behaviour changes as a response to unintended birth. This knowledge is important not only to understand high levels of unintended childbearing in Latin America, but also substantial levels of repeat unintended pregnancies. This study will apply transition matrices and discrete-time-competing-risk hazard models to examine how the transition to first contraceptive method adopted during 12 months’ postpartum period depends on the contraceptive use before conception. The analysis will be conducted both for women who reported their last birth as unintended and intended. Preliminary analyses reveal that large proportion of women who experienced an unintended birth changed their contraceptive use to more effective, but many continued not to use any method. Moreover, women who did not use contraception before pregnancy are most likely of all women to be non-users also after birth. This potentially puts them at the risk of subsequent unintended pregnancies. Lastly, the patterns of use differ according to birth intention status, signalling that unintended birth experience might affect future contraceptive use.  


Timing of abortions, births and relationship transitions in Finland
Heini Väisänen, University of Southampton   

Objective: Studies have shown union dissolution and fertility decisions are made jointly, but there is a lack of research taking abortion decisions into account, although relationship problems are common reasons for abortion. I study how decisions to terminate a pregnancy are intertwined with the decisions to end a romantic relationship and with childbearing. Methods: I used multi-level multi-process event-history modelling to analyse a set of Finnish register data of women born in 1965-69. Results: There were unobserved characteristics affecting both union stability and the likelihood of an abortion. Women with a tendency towards less stable unions also had a higher likelihood of abortion. A strong positive association between experiencing an abortion and union dissolution within the same year was found. This effect was less strong for cohabiting than married couples, as was the negative association between a conception leading to birth and a union dissolution within the same year. Relationship situation was an important factor in determining the likelihood of experiencing a birth or an abortion. Conclusions: Women in unstable relationships may not want children and thus are more likely to terminate an unintended pregnancy. Tendency towards unstable unions not being associated with higher fertility implies that commitment effect increasing the fertility of those experiencing many unions is not strong in Finland. The effects were different for married and cohabiting women, particularly in early years of the relationship, implying a different level of commitment. Contribution: This study is the first to show how these processes are intertwined using reliable data. 


Fertility: Social norms &fertility. Tuesday 13 September 09:00am 

Family and fertility in context: a cross-cultural analysis of whether cooperation within the family increases fertility
Rebecca Sear 1, Susan B Schaffnit 1, Kristin Snopkowski 2, 1 London School of Hygiene and Tropical Medicine, 2 Boise University    

The family is very diverse, taking a number of different forms cross-culturally. Commonly, however, it involves cooperation among family members to raise children successfully. This cooperation may lower the costs of reproduction for women such that they are able to maintain higher fertility in the presence of cooperative family members. We use several large, rich secondary datasets to test whether the presence of family increases fertility in both high- and middle-income contexts. We find evidence from middle-income contexts (using the Indonesia Family Life Survey and the Thailand Demographic and Health Survey) that family increases fertility. The family members most strongly associated with fertility are women’s parents-in-law. We further find evidence that these relationships are plausibly driven by cooperation between family members. But the picture is more complicated in high-income settings. Using both Gender and Generations Survey data from Europe, and the UK’s Millennium Cohort Survey, we demonstrate that associations between family and fertility are found in high income contexts, but they are not always positive. Further, and in contrast to middle-income contexts, we find that how support is measured in high income populations may be more important than who provides support: certain types of practical help appear to be negatively correlated with fertility, perhaps because they are an indication of need on the part of recipients. Greater emotional support from both family and non-family sources, however, is associated with higher fertility. We end by considering the role of family in the fertility decline which occurs as populations develop economically.


Fertility Is Low When There Is No Societal Agreement on a Specific Gender Role Model
Ansgar Hudde, Bamberg Graduate School of Social Sciences (BAGSS), University of Bamberg, Germany   

Many authors argue that recent rises in fertility in Western Societies can be explained by increases in gender equality and gender equal attitudes, but the mechanisms behind these relations remain unclear. This paper argues that fertility levels are not (only) related to the content of societal gender role attitudes, but to the variation in these attitudes. A great variation in attitudes among potential partners causes uncertainty and conflicts, which decreases people’s propensity to choose parenthood (again). How this idea is tested: macro-level regressions on 23 countries are run in which a measure for the average gender role attitude as well as the dispersion in that attitude are regressed on the level of fertility. This dispersion is measured as the standard deviation of an attitudinal variable, which is computed through factor analysis, in the given country. Attitudinal information is from the ISSP 2012. The analysis gives support to the hypothesis: the greater the variation in gender role attitudes, the lower the fertility. The association is considerably strong, significant, and holds against various robustness checks. The contribution of this paper is threefold. First, it sheds more light on the mechanism behind the widely discussed relation of gender relations and fertility. Second, it shows how the variation in a variable - independently of its average value or content - matters for a social outcome. Third, it applies a more reliable measure for fertility than most comparative studies, namely completed cohort fertility. 


Kin support and childbearing intentions: Polish mothers and fathers in Poland and in the UK
Joanna Marczak, London School of Economics and Political Science 

Reproductive decisions imply long-term direct and indirect costs and individuals can utilise family, market and the state’s assistance to achieve resources perceived as necessary for childbearing. Reliance on kin help, social assistance and employment markets may have various impact on childbearing choices. The evidence presented in this paper is based on 42 in-depth, semi-structured interviews with Polish mothers and fathers in the UK and Poland, conducted in 2011. The findings illustrate that specific socio-institutional settings are linked to different strategies parents employ to achieve desired resources related to their childbearing intentions. Respondents in Poland often resorted to the market and the family itself (nuclear and extended). The pervasiveness and importance of long-term kin support in the Polish familialistic context increases economic resources seen by respondents as necessary for childrearing and it has negative consequences for reproductive intentions because many parents in Poland expect to invest rather heavily in their children long into adulthoods. Polish respondents in London derive resources based on the market and the welfare state and their expectations about the necessity of long-term financial investment in children were overall lower than in Krakow. The findings also question assumptions about the meaning of categories used in surveys measuring childbearing intentions, e.g., ‘adequate housing’ may signify enough space for a child but it can also denote being able to provide a property for each adult child. The link between kin support and childbearing intentions is complex, context specific and could well change for an individual when they migrate. 


Childlessness and fertility by couples’ educational gender (in)equality in Austria, Bulgaria and France
Beata Osiewalska, Cracow University of Economics   

Examining procreative behaviour by education and sex has led to different results – a negative influence of educational level mainly concerns women, while among men the correlation is often positive or negligible. While the gender differences have been examined in depth, couples’ perspectives on fertility remains understudied and the question about how (in)equality between partners’ educational levels drives their reproductive behaviour is still not answered. Thus, with the use of GGS data for Austria, Bulgaria and France we aim to observe couples with female partners aged 24-45 and examine how their number of children ever born varies according to partners’ educational constellations. Various measures of educational endogamy are considered (robustness check). The hurdle zero-truncated Poisson model within the Bayesian framework is applied. The three chosen countries provide different conditions for work-family life balance. Austria has relatively high gender equality, low unemployment, but an inefficient childcare system. Bulgaria has a wide prevalence of the dual earner family model, but a relatively low level of gender equality and unadjusted family institutions. In France the level of gender equality is relatively high and the childcare system is one of the best in Europe. The results indicate the negative impact of hypogamy and homogamy in high education on couples’ fertility, which might be assigned to the postponement of having their first and subsequent children. Homogamous low educated partners have, on average, the highest fertility in all countries, but France. The effect of hypergamy, in turn, is insignificant and thus similar to homogamy in medium education. 


Fertility: Reproductive health. Tuesday 13 September 1:30pm 

Advanced maternal age not an independent risk factor for low birth weight or preterm birth
Alice Goisis1,2, Hanna Remes 3, Kieron Barclay 1,2,4, Pekka Martikainen 2,3, Mikko Myrskylä 1,2,3 1 Department of Social Policy, London School of Economics and Political Science, 2 Max Planck Institute for Demographic Research, Rostock, Germany, 3 Population Research Unit, Department of Social Research, University of Helsinki, 4 Department of Sociology, Stockholm University     

Advanced maternal age at birth is considered a major risk factor for birth outcomes. It is unclear to what extent this association is confounded by maternal characteristics. To test whether advanced maternal age at birth independently increases the risk of low birth weight (<2500g) and preterm birth (<37 weeks’ gestation), we compare between (comparing children born to different mothers at different ages) and within-family (comparing children that are born to the same mother at different ages) models. The latter approach reduces confounding by unobserved parental characteristics that are

shared by siblings. We use Finnish population registers of 124,098 children born between 1987-2000. In between-family models, when compared to maternal age 25-29, maternal ages 35-39 and 40 and above are associated with a 1.1 percentage point (95% CI: 0.8,1.4) and 2.2 percentage point (95% CI: 1.4,2.9) increase in probability of LBW, corresponding to 50% and 100% increase in the relative risk. The associations are similar for risk of preterm. In within-family models, the relationship between advanced maternal age and LBW or preterm is statistically and substantively negligible. Advanced maternal age does not independently increase the risk of LBW and preterm amongst mothers who have at least two live births in Finland. 


Who Seeks Help? Responses to Perceived Fertility Impairments in Malawi
Jasmine Fledderjohann 1, Jenny Trinitapoli 2, Francesco Billari 1, 1 University of Oxford, 2 University of Chicago 

Infertility, the inability to conceive or maintain a pregnancy after 12+ months of regular intercourse, has been linked to a range of negative consequences. Some of the highest rates of infertility in the world are found in sub-Saharan Africa (SSA), where the social and economic value of childbearing is particularly high. Yet how individuals perceive and respond to fertility impairments in SSA has not been well-documented. In this paper, we use wave 4 the Tsogolo la Thanzi survey of young women in Malawi to explore the prevalence and predictors of self-identified fertility impairments and help-seeking strategies. Nearly 13% of the sample reported experiencing fertility impairment. Of these, 85.3% sought some kind of help for their perceived fertility impairment. Visiting a hospital or clinic was the most common response, followed closely by going to a traditional healer. Around one-fifth employed multiple help-seeking strategies, highlighting the need for various help-seeking behaviours to be viewed in tandem rather than in isolation. 


Maternal Socioeconomic Status and Risk of Labour Induction in the UK
Sarah Carter, University of Southampton   

The World Health Organization recommends that rates of caesarean section do not exceed 10%. As the rate of caesarean section in the United Kingdom has risen well above this recommendation, concerns over the maternal and infant health implications of operative deliveries have increased and caesarean prevention has a become goal in UK hospitals. Additionally, links between maternal socioeconomic status (SES) and maternal and infant health make research into the impact of the socioeconomic gradient on childbirth imperative. This paper explores the relationship between SES and risk of caesarean section in the UK by examining labour induction, the beginning of the “cascade of interventions” leading to increased caesarean section rates. Logistic regression analyses were conducted using the first sweep of the Millennium Cohort Study, with models utilizing indicators of SES such as maternal educational attainment, electoral ward deprivation, housing tenure, and marital status. Final models controlled for maternal health mediators, including BMI and illness in pregnancy. After controlling for maternal health, the relationships between labour induction and two markers of SES – fewer educational qualifications and living in disadvantaged places – remained important. The results for maternal education and location imply that younger, less advantaged women are at higher risk of induction, suggesting a relationship between SES and induction. This association could identify women at higher risk of entering the cascade of interventions leading to operative deliveries, and may help enhance caesarean prevention programs. 


Female genital mutilation in Mali and Mauritania
Valeria Cetorelli 1, 2, Ben Wilson 1, Ewa Batyra 1, Ernestina Coast 1, 1 The London School of Economics and Political Science, 2 Johns Hopkins University 

There is almost no research examining long-run trends in FGM practices in Africa. Reconstructing these trends is essential if changes in FGM behaviours are to be understood, and if the effectiveness of national FGM policies is to be evaluated. In this research, we make use of under-exploited secondary survey data to generate new knowledge about FGM in Africa. We carry out a case study of two comparable countries, Mali and Mauritania. In 2005, Mauritania introduced a policy to make FGM illegal. The results show a decline in long-run circumcision rates in Mauritania, but not in Mali, for women born before 1996. For women born between 1997 and 2004, circumcision rates in the two countries are stable and parallel, and the rate is an average of 18 percentage points higher in Mali. However, for women born after 2005, there is a narrowing in the difference in circumcision rates between the two countries, which suggests that new legislation in 2005 may have had no impact on FGM practices in Mauritania. 


Fertility: Fertility dynamics. Tuesday 13 September 16:45pm 

The Link between Structural Adjustment Programs, Educational Discontinuities and Stalled Fertility in Sub-Saharan Africa
Endale Kebede, Anne Goujon, Wolfgang Lutz, Wittgenstein Centre for Demography and Global Human Capital (IIASA, VID/ÖAW, WU) 

One of the puzzling mystery of sub-African demography remains the interruptions of fertility declines in the earliest years of 2000s – so called stalls in the fertility transition. Many scholars have focused on the lack of family planning in the early 2000s. However, the possible linkage between the early-life educational conditions of women and the stalled fertility transitions has been mostly ignored. Goujon, Lutz and K.C. (2015) show descriptively that those countries that experienced stalls in their fertility declines also experienced stalls in their educational expansion some 20 years earlier. Those were times when Structural Adjustment Programs (SAP) were implemented in the region, causing discontinuities in the educational improvements for the cohorts of women who were of primary school age, due to budget cuts in social spending. The fact that those cohorts of women, who were forced to drop out of schools or denied to enroll due to SAP, have joined the reproductive age in the early 2000s would provide a plausible explanation for the fertility stalls occurring in the same period. Here, applying a multi-level poisson regression analysis to several Demographic and Health Surveys, we provide micro-level evidences that the educational discontinuities experienced during the SAP implementation period partially explain the stalled fertility transitions experienced in the region. The findings of this paper have potentially important input for policy makers. The fact that the most recent and most educated cohorts are replacing the cohorts who were affected by the SAP means that the fertility stall is a temporary phenomenon. 


Fertility, family and employment: A life course perspective on fertility of immigrants and their descendants in the UK
Tina Hannemann, Hill Kulu, University of Liverpool   

This study investigates fertility among immigrants and their descendants in the UK. Previous research has shown large differences in fertility behaviour between UK natives and certain immigrant groups. Some researchers attribute elevated fertility levels among certain UK-born ethnic minorities to cultural factors, whereas others argue that high fertility is the consequence of their poor education and labour market prospectives. This study aims at exploring these fertility differences using sequence analysis. The ethnic minority boost sample in the Understanding Society dataset and the detailed individual-level histories of union formation and dissolution processes and economic activity provide a great opportunity to model the trajectory of the three processes – fertility, family formation and employment - simultaneously. This life-course perspective will shed further light on the interdependencies between all three trajectories and provide valuable insights on the life decisions which lead to the fertility differences observed among the immigrant groups in the UK, while taking important individual socio-economic factors into account. This study further extends the analysis to descendants of immigrants to show if fertility differences persist or erode throughout generations. Additionally, the study is able to distinguish immigrants and their descendants from different origins which, given the wide range of ethnic groups in the UK, will provide more refined results on the role of cultural background and integration opportunities. 


Shifts to low fertility since the baby-boom: The role of parity progression ratios Kryštof Zeman, Éva Beaujouan, Zuzanna Brzozowska, Tomáš Sobotka, Vienna Institute of Demography (Austrian Academy of Sciences) & Wittgenstein Centre for Demography and Global Human Capital, Vienna    

The long-term decline in cohort fertility among women born between the mid-1930s and 1970 has been documented in most developed countries. Changes in the parity distribution have also been studied: childlessness has been on the rise, the share of women with one child has tended to increase, the proportion of women with two children has stopped rising and the importance of large families has diminished continuously. However, whether these declines were mainly driven by the diminishing importance of larger families or whether they occurred primarily due to declines in first and second birth parity progression ratios (and the concomitant rise in childlessness and one-child families) has not been sufficiently analysed. In this study we quantify the contribution of the changing parity progression ratios to cohort fertility decline in currently low-fertility countries. Using data from the Human Fertility Database (HFD) and the Cohort Fertility and Education (CFE) database we focus on women born after 1940 who (almost) completed their childbearing (i.e. aged 43 or more). We employ a decomposition method as originally developed by Pullum et al. (1989). It takes into account the sequential nature of childbearing as a chain of transitions across parities. The preliminary results show that for the 1940-55 birth cohorts the drop in completed fertility was driven by falling high parity transitions in most countries. In the 1955-1970 cohorts the fertility decline resulted mainly from the decreasing transition to first birth in Western Europe and from the falling progression to second birth in Central and Eastern Europe. 


The emerging positive educational gradient in fertility in Belgium: A multi-level regional analysis of a vanguard country
Jonas Wood 1, Karel Neels 1, Sebastian Klüsener 2, Mikko Myrskylä 2, 1 Centre for Longitudinal and Life-course Studies, University of Antwerp, Belgium, 2 Max Planck Institute for Demographic Research, Rostock, Germany   

There is increasing evidence for substantial shifts in the macro- and micro-level relationships between human development and fertility in highly developed parts of the world, including shifts in the long-standing negative relation between women’s socio-economic position and childbearing. Belgium seems to constitute a vanguard country in these trends with positive associations between tertiary education and second births. It has been argued, that these patterns might be related to Belgium’s role as a forerunner in implementing policies supporting the reconciliation of family and career goals. In order to improve our understanding of the emerging positive educational gradient in fertility in Belgium, this study relies on a unique combination of microdata covering all Belgian residents between 2002 and 2005 and information on economic development and social services at the municipality level. We assess how variation in these municipality-level conditions is related to fertility outcomes by education. Results show that the effect of being highly educated on second births varies considerably across the 579 municipalities, with municipality-specific effects ranging from no difference to a 250 per cent difference in second birth odds. Controlling for various individual-level characteristics, we show that a considerable part of the between-municipality variation in the educational gradient in second births can be related to economic conditions and social services at the municipality level. Whereas extensive social services positively correlate with second births for highly educated mothers, low educated mothers are found to be less likely to progress to a second child in more prosperous municipalities.