Childhood Neighborhoods and Life-Time Fertility in Twentieth-Century Sweden. A k-Nearest Neighbor Approach

Vinicius de Souza Maia, Lund University
Finn Hedefalk, Lund University
Martin Dribe, Lund University

In addition to the family context, persons in the neighborhood and community are important for individual life course, socioeconomic and demographic outcomes such as education, career and family transitions. This paper aims to study how neighborhood conditions in childhood influence fertility outcomes in adulthood. Theories focusing on motivation, aspirations and attitudes or peer influence generally predict that low-SES children living in high-SES neighborhoods will benefit from social contact, whereas resource competition and relative deprivation theories give opposite predictions.. Moreover, research has shown that childhood neighborhood conditions are associated with SES and health over the life course. We hypothesize that growing up in a high-SES neighborhood delays entry into parenthood, but does not necessarily lower the number of children ever born, as high-SES is increasingly becoming associated with higher fertility. Conversely, low-SES neighborhoods are expected to increase risks of early childbearing, regardless of individual class origin. We also expect women to be more strongly affected than men. We model neighborhoods using a k-nearest neighbors approach that addresses common issues with spatial aggregation and include a longitudinal component, accounting for the cumulative consequences of neighborhoods throughout the life course. We study individuals growing up in Landskrona in southern Sweden, 1905-1967, and follow them through adulthood regardless of where they reside in Sweden for the period 1968-2015 using data from the Scanian Economic Demographic Database and Statistics Sweden. We construct variables for: (1) SES of nearest neighbors and (2) socio-spatial characteristics of the neighborhood, for varying sizes of neighborhood. We focus on class estimates of the neighborhoods for children in Landskrona and look at their age of parenthood and children ever born at age 50 . The empirical analysis is based on spatial regressions. Preliminary results seem to conform to our hypothesis.

No extended abstract or paper available

 Presented in Session 145. The Impact of Mothers, Neighborhoods and Disability on later-Life Outcomes