Residential context, migration and fertility

Kulu, H. and Washbrook, E. Residential context, migration and fertility.

Abstract

This study examines fertility variation by residential context. While there is a large literature on fertility trends and determinants in industrialised countries, little research has investigated spatial fertility variation. We study fertility variation across regions with different size and within urban regions by distinguishing between central cities and suburbs of the cities. We use vital statistics and longitudinal data from Britain and apply event history analysis. We investigate to what extent do the socio-economic characteristics of couples and selective migrations explain fertility variation between residential contexts and to what extent do contextual factors play a role. Our analysis shows that fertility levels decline as the size of an urban area increases; within urban regions suburbs have significantly higher fertility levels than the city centres.

Further details

It is a stylised fact that observed in many countries that fertility levels are higher in rural areas and small towns and lower in large cities. One explanation for this regularity is that households with high fertility preferences selectively migrate towards less densely populated areas because they view such areas as more appropriate, or more cost effective, for the raising of children. Under this explanation the direction of the relationship runs from fertility to residential location choice. An alternative view is that factors related to the immediate living environment in settlements of different sizes influence desires for children (or the ability to realise those desires). Here the hypothesised relationship runs in the opposite direction, from location to fertility. Surprisingly little evidence of the relative importance of the two explanations has been provided in the literature, perhaps because individual-level longitudinal data is required to untangle different influences on behaviour.

To explore this issue, we first constructed total fertility rates (TFRs) for local authority districts (LADs) using data from the Office of National Statistics (ONS) birth statistics and from the 2011 UK Census. We then assigned each LAD to an area classification ranging from central London to rural area, distinguishing between central cities and suburban areas for cities and towns with more than 200,000 people. We document the expected spatial variation in fertility – the larger the settlement size the smaller was total fertility, suburbs have significantly higher fertility levels than the city centres.

We then link this data to information on women aged 16–49 from the 18 waves of the BHPS. We use event-history analysis to calculate parity-specific birth rates (for the first three transitions). We model the time to conception (subsequently leading to a birth) and focus on whether the risk varies across four types of area: London central city; London suburbs; other towns and cities; and small towns and rural areas. We explore whether the observed variation can be explained by differences in the types of women living in different areas, considering characteristics such as education, employment, ethnicity and partnership status. We include indicators for a recent move within or between labour-market areas to assess the effects of selective migration. We also include in the model a woman-level residual (random effect) to control for unmeasured time-invariant characteristics that influence her fertility behaviour.

In terms of the grant’s hypotheses, we find that fertility rates vary systematically across areas, increasing as settlement size decreases. This variation is driven by differences in first- and third-birth rates as we found no differences in second birth probabilities. We found evidence that some couples do move to suitable areas in anticipation of childbearing, particularly prior to the birth of second- and higher-order children. However, selective migrations did not explain any of the variation in spatial fertility as the share of internal migrants was small. In addition, the ability of differences in the characteristics of women resident in different types of area to explain fertility variation was limited, and this conclusion was robust to allowance for time-invariant unobserved fertility preferences. The results, therefore, are supportive of the view that contextual factors in smaller settlements lead to higher fertility, for example because of lower costs of childrearing and social and cultural factors.