Stopping, Spacing and Postponing in the British Fertility Transition: Insights from Census Data
Hannaliis Jaadla, University of Cambridge
Alice Reid, University of Cambridge
Eilidh Garrett, University of Essex
Ian Timæus, London School of Hygiene and Tropical Medicine
First, we calculate age-specific fertility rates using the own children method, for each census year and for regional and social groups. Second, we use the age of the youngest child as an indicator of stopping behaviour, exploring proportions of women with the youngest child over a certain age (eg 5 or 10 years), by age of mother. This also allows us to compare different census years as well as regional patterns and different social groups. Our final method is only available for the 1911 census which asked about the children ever born and dead children of married women. We impute the ages of dead and absent children and thereby reconstruct full birth histories which can be used to estimate duration-specific fertility rates, parity progression ratios, and indices of the length of birth intervals. Initial explorations suggest that widespread stopping at low parities may only have emerged later in the twentieth century. Results for regions and social groups will be produced and the implications for theories of fertility control will be discussed.
Stopping, spacing andpostponing in the British fertility transition: insights from census data
The orthodoxstory about fertility decline during the demographic transition is one of a movefrom natural, uncontrolled fertility to a regime where couples embark onchildbearing soon after marriage and stop once they reach a desired or targetnumber of children. The concept of natural fertility has been challenged repeatedlyand mounting evidence exists of non-parity specific fertility control beforethe demographic transition, but the idea that the European fertility declinewas achieved through stopping behaviour is rarely questioned. Factors that mayaccount for this include: the strong evidence that many later transitions werecharacterised by stopping behaviour; the lack (in many countries) of availableevidence with which to construct age-specific and particularly parity-specific fertilityschedules; a tendency to give attention to the fact that populations containheterogenous groups of women behaving in different ways; and the predominanceof small-scale studies in which the sample size precludes much disaggregation. Thegrowing evidence, however, of different forms of fertility transitions, inAfrica in particular, characterised by spacing (delaying the next birth toensure a preferred birth interval) or postponing (avoiding conception untilcircumstances are more favourable) prompts us to ask whether such behaviourmight also have characterised fertility declines in the nineteenth century, whenthe absence of reliable contraception rendered a target family size aproblematic concept.
Thispaper uses the complete count census data from 1851 to 1911 (excluding 1871)for England and Wales (over 17.5 million records in 1851 and more than 36million records in 1911) using a version of the Integrated Census Microdata whichhas been further enhanced to increase its suitability for demographicanalysis. We will start by illustrating age-specific fertility ratesreconstructed using the own children method, demonstrating that these fail toshow the classic stopping pattern of hollowing out of fertility at older ages,even among the wives of professional men and textile workers who led the moveto smaller families (Figure 1). Without parity-specific information it isimpossible to tell whether this inability to detect stopping is due to failedstopping behaviour, to the dilution of those practicing stopping with those whowere not controlling their fertility, or to the widespread presence of adifferent form of fertility control, such as spacing or postponing.
Thispaper will therefore proceed to alternative ways of exploring the existence ofspacing, postponing and stopping behaviour in the British fertility transitionusing complete count census data. Our first approach will use the age of theyoungest child in the household as an indicator of stopping behaviour, exploringproportions of women with the youngest child over a certain age (eg 5 or 10years), by age of mother. This will allow us to compare different census yearsas well as regional patterns and different social groups.
Oursecond approach is to reconstruct full birth histories for married womenenumerated in the 1911 census of England and Wales. The method is based an extensionto the own children method proposed by Luther and Cho (1988) which combinesthe information about women''s number of children ever born (CEB) and number ofdeceased children (CD) with the ''own children'' birth history of each womanprovided by the list of household members. The 1911 census was the firstBritish census to ask CEB/CD questions, and while it did not collect thisinformation on all women, it did ask married women questions about childrenborn to their current marriage, and also included a question about the durationin years of these marriages. Using the answers to these questions, one cancomplete women''s birth histories by probabilistically imputing ages to theirdead children and to any children not present in the household. From thesehistories, we will calculate measures such as marital or intervalduration-specific fertility rates, parity progression ratios, and the medianlength of closed birth intervals. Initial explorations suggest that widespreadstopping at low parities may only have emerged later in the twentieth century.Results for regions and social groups will be produced and the implications fortheories of fertility control will be discussed.
Luther NY and Cho L-J (1988) Reconstruction of birth histories from census andhousehold survey data. Population Studies, 42, 451-72.
Figure1: Age-specific marital fertility rates for the leaders in the Britishfertility transition: the wives of professional men (social class 1) andtextile workers, 1851-1911.
Presented in Session 1059: History