The growth in frequency of cohabiting unions, and of births outside of formal marriage, is one of the major demographic trends of recent decades, but consistent long-run data on the subject is sparse. Cross-sectional surveys provide information on trends in union status at survey and vital registration statistics give approximate information on the share of fertility in informal unions. The paper will present data on trends in partnership and fertility from a more complete data source—the fertility and union histories of a combined file of the General Household Survey from 1979-2007. Dated fertility and marital histories are available throughout the series, partial histories of cohabitation from 1979-1998, and near-complete cohabitation histories from 2000-2007. We use these data to examine detailed trends in male and female partnership, and also the interrelationships between fertility and partnership trends, for women. For example, we look at such issues as the distribution of births by order and partnership type, and the timing of births in the sequence of partnerships.

Email: E.Beaujouan@soton.ac.uk M.Ni-Bhrolchain@soton.ac.uk

The rise of non-marital fertility across European regions since 1960: state policies, political borders, and natural experiments
Sebastian Kluesener, Max Planck Institute for Demographic Research; Brienna Perelli-Harris, University of Southampton; Sanchez Gassen, Max Planck Institute for Demographic Research

This paper combines Political Demography and Political Geography approaches to investigate the role of political structures andprocesses in shaping geographic patterns of non-marital fertility in Europe since 1960. Earlier research by Susan Watkins (1991) has shown that as a result of nation building processes, variation in non-martial fertility rates (as well as nuptiality and marital fertility rates) decreased between 1870 and 1960 across (western) Europe. But this decrease occurred mainly within states. For the future she expected national boundaries to become less deeply etched on the demographic map of Europe as a result of European integration processes. Our paper investigates, to what extent this prediction has materialized over the last decades, focusing on the recent rise of non-marital fertility across Europe. In a first step we display maps of Europe in 1960, 1975, 1990, and 2007, which provide at high geographic detail the increase in non-marital fertility over space and time. Second, we examine the role of state policies and political borders in producing spatial patterns of non-marital fertility. For this we employ inequality measures and field-orientated spatial modeling techniques. First results deliver support for Watkins`s hypothesis, that the role of political borders as demographic divides has weakened at least over the last two decades. We are also able to identify several potential natural experiment settings for further study on the role of policies in shaping spatial patterns of non-marital fertility.

Email: kluesener@demogr.mpg.de

Exploring the geography of fertility in Scotland
Francesca Fiori, Elspeth Graham, Zhiqiang Feng, Centre for Population Change, University of St. Andrews

This study originates from the observation of significant regional variations in fertility within Scotland. Building on a previous studyby Boyle, Graham and Feng (2007), it seeks to shed further light on the clustering of relatively higher and lower fertility at the local scale by examining differences between groups of women defined by age. The paper addresses two research questions: (1) Is the fertility of three age groups of women differentially associated with the demographic and socio-economic composition of the resident population in their local area?; and (2) Does geographical variability by age persist after controlling for local population characteristics? For the analysis, Output Area (OA) data from the 2001 Scottish census were combined with birth counts for the years 2000-2002 from GROS vital events registration. The relationships between births in a small area and the composition of the resident population were assessed by means of negative binomial regression models. GIS software was used to map geographical variations in the number of births. The findings suggest that fertility in an OA is significantly associated with the characteristics of the resident population but in different ways for women of different ages. When the residuals from the age-specific regression models were mapped using the G* Statistic to identify clustering, the persistence of significant local pockets of higher than expected, and lower than expected, fertility remained. The paper concludes by reflecting on both the limitations of the study and possible interpretations of the findings.

Email: ff20@st-andrews.ac.uk

Time bomb or damp squib? fertility in contemporary Northern Ireland
Patrick McGregor, Patricia McKee, University of Ulster

Northern Ireland has been and continues to be deeply divided on the basis of religion. There is considerable interestgenerated by the publication of the decennial census: broadly speaking it is assumed that voting patterns will directly reflect religious affiliation. Formerly higher Catholic fertility was offset by higher migration but this was against a background of a trend decline in fertility. This paper examines and compares contemporary fertility in the two communities. The model employed uses the notion of identity as developed by Akerlof and Kranton (2000). In this case fertility is a weighted average of that prescribed by the social group and that determined by standard neoclassical considerations. The data are drawn from the Northern Ireland Longitudinal Study. A logit analysis of births in the period 1997 – 2007 is performed and the marginal effects of religion and the proportion of Catholics in an area are computed at the individual level. Interpretation is facilitated by graphing the results against age, time and the socio-economic characteristics of the locality. The key identity variable for Catholics is their proportion in the locality concerned. This acts as a reinforcement mechanism of identity but is largely absent on the Protestant side. Although the latter's sense of identity is as strong as Catholics, it does not have a demographic dimension. The residential segregation in Northern Ireland has meant that demographic change within the two communities is distinct though subject to a similar economic environment. This makes simple extrapolation of current trends unreliable.

Email: ppl.mcgregor@ulster.ac.uk

Accounting for near-replacement fertility in Accra, Ghana
Allan G. Hill, Harvard Center for Population & Development Studies & University of Southampton; Kelly Blanchard, Ibis Reproductive Health, Cambrdge, Mass.; Naomi Lince, Ibis Reproductive Health, Johannesburg

A conundrum noted by several authors concerns the apparent inconsistency between the level of fertility and the prevalence ofcontraceptive use in several African populations. Several explanations have been advanced, none very convincing. Most make use of some variant of Bongaarts's proximate determinants framework and investigate the effect of under-reported levels of induced abortion, misunderstandings about the meaning of contraception and "traditional methods" of avoiding conceptions, possibly including withdrawal and herbal or homeopathic methods. Bongaarts suggested that part of the explanation is the strength of the commitment to smaller families that determines the use-effectiveness of a variety of contraceptive methods. In Accra, the commitment of women to a small number of children is indisputable. Using both quantitative and qualitative data, we describe the fertility transition in Accra and current patterns of reproduction from detailed fieldwork begun in 2003 and completed in 2009-10. With careful measurement of the principal determinants including special studies of the contribution of induced abortion, we show that women in Accra have found a variety of ways to manage their low fertility, principally by avoiding regular coitus. Reservations about the side effects and costs of modern contraceptives figure large in the women's calculations. Although the methods chosen to limit the number of births seem effective, we note the considerable negative consequences on health and the substantial costs of health maintenance with the chosen strategies for managing fertility. These costs are particularly important for women with strong attachments to the labour force throughout their reproductive careers.

Email: ahill@hsph.harvard.edu

Effects of maternal socio-demographic characteristics on birth weight distribution in Greece: a quantile regression analysis
Georgia Verropoulou, University of Piraeus & Institute of Education; Cleon Tsimbos, University of Piraeus

Aims: The study aims at exploring effects of the socio-demographic characteristics of the mother on the distribution of birthweight in Greece. Data and methods: For the purposes of the analysis restricted access nationwide vital registration data at individual level are used. The data include 103,266 single live births registered in 2006 to women aged 20 or higher. The response variable is birth weight, recorded in grams. The method involves application of Quantile Regression Models. The effects of the independent variables are estimated at 6 different quantiles of the response variable (0.05, 0.10, 0.25, 0.50, 0.75 and 0.90). Results: The findings indicate that important risk factors tending to reduce birth weight include female sex of the newborn, primiparity, age of mother over 35, illegitimacy, prior history of stillbirths, infant and child deaths, and living in metropolitan areas. Regarding socio-economic factors, being a housewife is also associated with lower birth weight while higher educational attainment has a protective effect. Finally, being an immigrant is associated with significantly higher birth weight. The results indicate that the impact of most factors differentiates substantially across the birth weight distribution and effects are greater at the left tail. For instance, illegitimate newborns have lower birth weight by 180 grams at the 0.05 quantile compared to legitimate babies while the difference is only 66 grams at the 0.90 quantile. Quantile regression gives more accurate results compared to an OLS model which assumes a constant effect across birth weight outcomes.

Email: gverrop@unipi.gr

Determinants of quality of antenatal care utilisation in Lesotho: a multilevel analysis
Rebecca Vassallo, Amos Channon, University of Southampton

Antenatal care plays a vital role in identifying and managing pregnancy complications and referring women to skill-attendeddelivery. The 2004 Lesotho Demographic and Health Survey indicated that 90% of women in Lesotho made use of some kind of antenatal care. Despite this high uptake percentage, controversies over the effectiveness of routine antenatal care, the complexity of this service, and differential access to care of varying standards highlighted the need to consider the characteristics and quality of the antenatal care utilised. Whilst indices of quality antenatal care were identified in the literature, none of these measures were populated using DHS data, and therefore the composition of an index of personal-level antenatal care utilisation became a principal objective of the study. Scores on the constructed index were allocated to each case and the descriptive analysis identified the need to eliminate cases with a value of 0 to better adhere to the normality assumption. This restricted the analysis to the quality of antenatal care utilised by women with access to some kind of antenatal care. Multilevel modelling was the analytical tool used to identify possible determinants of the index score. Associations between the index score and spatial variables, socio-economic variables, reproductive and demographic timing variables, previous health care use and transportation access barrier indicators, and media exposure were found. Moreover, a significant result for a=0.05 for a random intercept at the enumeration area-level indicated the existence of important similarities in the quality of antenatal care utilised by women within the same cluster.

Email: rv1g09@soton.ac.uk

Rising overmedicalisation of births in Southern India: a demand vs supply phenomenon
Tiziana Leone, London School of Economics; Ramila Bisht, J. Nehru University; Alice Goisis, London School of Economics; Ernestina Coast, London School of Economics

Overmedicalisation of childbearing is a common phenomenon in developed countries. Recently, numerous studies havehighlighted the increasing reliance on c-sections in developing countries. This is even more worrying as resources are scarce and when not necessary they create a further risk for mother and child's health. India is not immune to this phenomenon and in particular Southern States have seen sky rocketing c-section rates reaching twice as high levels than the recommended WHO 15%. What is not still clear is whether this increase is determined by demand or supply. The aim of this study is to understand the interaction between health systems and individual socio-economic factors in determining the probability of a c-section after controlling for major demographic and biological risk factors. Using multilevel modelling this study analyses data from the 2007-08 Reproductive and Child Health Survey from three Southern Indian states with high c-section rates: Andra Pradesh, Kerala and Tamil Nadu. Results show that after controlling for key risk factors, private institutions have a higher risk to perform c-sections. On the demand side higher education levels rather than wealth seems to increase the likelihood of a c-section. District level effects are significant in all the states demonstrating the need to control for health systems factors. This study confirms the need to invest in further research to understand whether the quest for increased institutional deliveries in a country with a high maternal mortality might not be compromised by a push for overmedicalisation.

Email: T.Leone@lse.ac.uk

The value of children and theories of fertility transition
Melanie Frost, University of Oxford

Debate over why people choose to have a certain number of children has caused a great deal of controversy over the years, and the validity of competing theories of fertility transition turn largely on this question. Demographers generally favour either cultural or economic explanations for fertility transition, but rarely ask individuals why they have children, especially in the developing world. Using Young Lives data from four less developed countries (Ethiopia, Vietnam, Peru and India) this paper looks at the value parents place on their children and how these relate to the number of children they have. These childbearing motivations range from social and psychological to practical and economic. Latent class analysis is used to construct indices of the strength of parents' different motivations (economic, kinship, comfort and esteem) and multilevel modelling is used to look at the effect of these on both children ever borne and sex composition.The results are interpreted in relation to what they mean for competing theories of fertility transition.

Email: melanie.frost@qeh.ox.ac.uk

Is there a global two-child norm?
Stuart Basten, David Coleman, University of Oxford

Since the earliest projections, the UN has supposed that global fertility will converge on something around 2.0. Similarly, surveysof ideal family sizes and fertility intentions in the industrial world almost always show a very heavy concentration around 2, with the tendency being to higher rather than lower fertility. Family tickets, cars and hotel rooms are primarily designed for 2 adults and 2 children. This presentation examines some parts of the world where the two-child norm is being challenged. The first example of China can be viewed as an exception - at least in policy terms - but other evidence from DHS data and, particularly from India, suggest that significantly lower fertility ideals are far from exclusive to China. Indeed, the phenomena of below replacement level fertility can be found in many urban centres across the developing world. As the world becomes increasingly urban, and as the global demographic influence of 'The West' diminishes, could it be time to challenge this assumption of a global 'two child' future?

Email: stuart.basten@spi.ox.ac.uk

The relationship between kin and availability and and fertility: a systematic review
Rebecca Sear, University of Duham

The question of why people have the number of children that they do has still not been fully answered, despite decades ofresearch on this topic. Here, we test the hypothesis that kin may significantly influence fertility rates, thereby helping to answer this question. The theoretical framework we use is that of evolutionary anthropology. In this field, a consensus is emerging that humans are cooperative breeders, i.e. women need significant help from others to raise their children as human children are too costly for mothers to rear alone. To test the hypothesis that kin availability influences fertility, we present the results of a systematic review of all published studies which have investigated correlations between the presence of kin and female fertility. This review demonstrates substantial evidence that the presence of family members is correlated with fertility, but these results do not always show consistent relationships. In some cases, particularly in high fertility societies and particularly for a woman's in-laws, kin increase fertility rates. In other cases, most commonly in low fertility societies and for a woman's own parents, kin slow down fertility. This literature is, however, very varied, with different studies using different methodologies and measures of fertility and kin availability. More comparative work is needed to elucidate exactly how kin might influence fertility rates. We conclude by presenting a brief outline of a research programme which aims to investigate kin influences on fertility using existing large-scale demographic datasets, using comparable methodologies in all datasets to facilitate comparisons between populations.

Email: rebecca.sear@durham.ac.uk

Navigating new socio-demographic landscapes using anthropological demography to understand the 'persistence' of high and early fertility among British Pakistanis
Kate Hampshire, Mwenza Blell, Bob Simpson, University of Durham

British Pakistanis (and Bangladeshis) continue to have earlier and higher fertility than other ethnic groups in the UK, despite arecent rise in educational and employment opportunities for British Pakistani women. In this paper we present findings from an interview-based study with 91 British Pakistani women and men in the Northeast of England, in order to throw some light on the reproductive choices and constraints that underlie this apparent demographic anomaly. Drawing on detailed reproductive narratives, we show that, contrary to common assumptions within the literature, high and early fertility does not necessarily indicate a passive acceptance of 'cultural norms' or a reluctance to engage with new social and economic opportunities. For some young women at least, early motherhood represents a deliberate and strategic choice: a way of managing different sets of family-building and other aspirations, embedded within a complex interplay of relationships between individuals, couples and wider families. This study contributes to the emerging 'inter-discipline' of anthropological demography, which offers important insights into population processes that neither discipline can do alone.

Email: k.r.hampshire@durham.ac.uk

How the determinants of childbearing intentions change over the life course
Maria Iacovou, University of Essex; Alexandra Skew, University of Essex; Lara Patricio Tavares, University of Bocconi

There is a substantial literature on the processes via which people decide whether to have children, how many they want, andwhen to have them. Many factors influencing childbearing intentions have been identified, including those related to individuals' own family of origin; social networks; partnership dynamics; and enabling and constraining factors such as people's economic situation. However, no research has yet examined how the processes determining childbearing intentions change over the life course. We identify four phases: (1) an early stage during which initial childbearing intentions are formed, but before an individual lives with a partner or actually expects to have children; (2) the period around the beginning of a co-residential partnership, when the prospect of children starts to become more real; (3) the period when a couple starts actively trying to have children, or actually has one or more children and decides whether or not to have more; and (4) a stage towards the end of a person's biologically fertile years. There is a substantial literature on the processes via which people decide whether to have children, how many they want, and when to have them. However, no research has examined how the processes determining childbearing intentions change over the life course. We identify four phases: (1) an early stage during which initial childbearing intentions are formed, but before an individual lives actually expects to have children; (2) the period around the beginning of a co-residential partnership, when the prospect of children becomes more real; (3) the period when a couple starts actively trying to have children, or actually has one or more children and decides whether to have more; and (4) a stage towards the end of a person's biologically fertile years. Using BHPS data, we find that the determinants of childbearing intentions vary markedly between these stages. Personality traits, and the characteristics of their family of origin, are key determinants of the formation of intentions in the early stage, but not later. The period around the beginning of a co-residential partnership is characterised by a reduction in the expected number of children, as childbearing becomes a possibility rather than a hypothetical event. The third phase sees people accommodating their expectations to those of their partner, and adapting them in response to learning about parenthood. In the fourth phase, intentions might be expected to decline among those still childless and/or unpartnerned, but we find only weak effects, suggesting that adaptation to the biological clock may begin relatively early in life.

Email: maria@essex.ac.uk

Interrelationships between childbearing and housing transitions in the family life course
Hill Kulu, University of Liverpool

Research has examined the effect of family changes on housing transitions and childbearing patterns within various housingtypes. While most research has investigated how an event in one life domain of a family depends on a state in another domain, the interplay between the two life domains has been little studied. This study examines the interrelationships between childbearing and housing transitions. We use rich longitudinal register data from Finland and apply event-history analysis. We first investigate the effect of children on housing changes and childbearing patterns by housing type. We then model childbearing and housing transitions jointly to control for unobserved characteristic of women, which may simultaneously influence their fertility behaviour and housing choices. Finally, we investigate the timing of childbearing and housing changes with respect to each other in order to deepen our understanding of the relationship between these two domains of the family life course.

Email: hill.kulu@liverpool.ac.uk

A cross-national comparative study of intergenerational continuities in childbearing
Mike Murphy, London School of Economics

A number of studies have shown that fertility patterns of parents and children are positively correlated, although the relationship isfrequently designated as 'weak'. Evidence from both historical data sets and recent large-scale surveys in a number of developed countries up to the later part of the Twentieth Century was brought together and presented in Murphy (1999). This showed that the association was tending to become stronger over time. This presentation updates and broadens the geographic coverage to enable additional data to be added to the database on intergenerational continuities in childbearing published in Murphy (1999), and in particular to establish whether these trends have continued into this century. The principal indicator used will be product moment correlations of fertility in successive generations, which is the only comparable variable widely available. Data from a number of recent large-scale data sources and programmes such as the Fertility and Family Surveys (FFS), International Social Survey Programme (ISSP) 2001 round, the Generations and Gender Programme (GGP) and the UK Understanding Society survey considerably extend the sample sizes, covariate availability and range of countries for which data are available. The presentation will also discuss how far patterns vary between socio-economic groups, such as by educational level, to assess how far these attenuate the strength of the relationship and vary between different countries as discussed for earlier periods in Murphy and Wang (2001) using hurdle and negative binomial multilevel models.

Email: M.Murphy@lse.ac.uk

New approaches for children demand estimating: A fertility control motivation analysis
Olfa Frini, University of Manouba, Tunisia

Attitude toward fertility regulation embraces, in general, both approval and disapproval of family planning. The paper assessesmechanisms of family planning demand and determines those which explain largely fertility regulation using 2001 Tunisian ONFP survey's data. Therefore, to estimate the demand for fertility control four indicators are considered: age marriage decision, motivation, attitudes and access. Motivation is identified by contraceptive use motivation. Attitude is reflected by the marriage period at the first birth control use. Access to fertility control is reflected through contraceptive method choice in Tunisian social setting where contraceptives are costless and easy available. It tests the theoretical view embodied in the Easterlin synthesis framework of fertility determination. Various regression techniques are used. In particular, a survival model is applied to estimate age marriage and marriage duration at the first use of contraceptive; a probit model with sample selection is applied for contraceptive usage by the Heckman technique; and a non-ordered logit multinomial model for contraceptive method modality choice. Beside variables commonly used new kinds are considered as the network family and social and cultural environment variables. Results showed that Tunisian women are aware of fertility regulation especially whose make a trade-off between children quantity and quality. Demand factors explain age marriage but supply factors explain marriage duration at the first birth control use. In Tunisia, familial and social network and cultural environment intervene in children formation preferences and fertility regulation motivation and contraceptive methods choice more than socioeconomic and economic factors.

Email: frini.olfa@planet.tn

Immigration, integration and fertility
Sylvie Dubuc, Lorraine Waller, University of Oxford

The paper offers a comprehensive presentation of fertility estimates of post- World War II immigrants and their descents in the UK. This includes a comparison between second generation and contemporary immigrants, an intergenerational (1^st/2d generations) analysis and comparison between immigrants' fertility and that in their country of origin. The applied methodology provides a solution to, at least partly, overcome the documented problem of overestimation when measuring fertility of immigrants using Period measures. Results reveal intergenerational fertility transitions that strongly contribute to the fertility convergence between ethnic groups and indicate degrees of fertility "assimilation" or "intergenerational adaptation" to the UK mainstream childbearing behaviour. Despite inter-ethnic converging trends, the children of immigrants still show ethnic specific fertility profiles that may reflect distinct social characteristics. Findings of fertility by educational attainment of the mothers contribute to clarify the role of socio-economic background/characteristics in shaping fertility of the ethnic minority groups in the UK.

Email: sylvie.dubuc@spi.ox.ac.uk

Fertility patterns of recent migrants to the UK: Exploring the role of partner's migration status.
Lorraine Waller, University of Oxford

Hypotheses of migrant fertility patterns each place differing emphases on the importance of origin, destination or the migration process itself. Less consideration has been given to the migration characteristics of partners and their role in the migration-fertility nexus. This research uses a pooled sample of UK Labour Force Survey data, from 2001-2009, to analyse the migration characteristics of the partners of recent migrants to the UK. Firstly, analyses explore with whom migrants are partnering, in terms of whether partners are non-migrants or migrants; and where applicable, whether partners are from the same country of birth and arrived to the UK in the same time period. Secondly, the information on partner's migration status is analysed with respect to the childbearing behaviours of recent migrants to the UK, for key migrant groups. A couple-centred approach provides further insight into the nature of the migration undertaken and how this relates to different fertility behaviours amongst migrant groups. The research adds to existing theory by addressing which migration characteristics are important when those of individuals within a couple differ from one another, for the UK context.

Email: lorraine.waller@spi.ox.ac.uk

Migrant fertility in England and Wales: origins, generations and convergence
Ben Wilson, Wendy Sigle-Rushton, London School of Economics

Compared with the native-born, fertility rates in England and Wales are higher for foreign-born women (Tromans et al. 2009; Coleman et al. 2002). However, there are several problems with this comparison. It is difficult to untangle quantum and tempo effects from period fertility measures (Ni Bhrolchain 2008), and foreign-born women are a heterogeneous group, with different characteristics from the native-born (Blau 1992; Kahn 1994; Fernández and Fogli 2009). Furthermore, the fertility of foreign-born women has been shown to vary considerably according to age at migration (Andersson 2004; Toulemon 2006). The fertility of first-generation adult immigrants is expected to differ from that of their children, who may be born in England or Wales (secondgeneration migrants), or born abroad (the 1.5 generation). Compared with the first-generation, the fertility of these children is more likely to converge with that of ancestral natives (i.e. native-born children of native-born parents), (e.g. Parrado and Phillip Morgan 2008). Several theories reinforce this expectation, in particular the effects of socialisation within the new destination culture (Sobotka 2008). The convergence hypothesis therefore predicts that birth risks for different generations can be ordered as follows: first generation, 1.5 generation, second generation, and ancestral natives. Using data from the ONS Longitudinal Study, this research will create models of birth risk, taking account of exposure to risk, country of origin, age at migration, and other socio-demographic characteristics. This presentation will describe the results, and discuss their implications, both for further research, and for existing statistics such as population projections.

Email: b.m.wilson@lse.ac.uk