Demographic projections are central to understanding how populations evolve and to planning for the future. Governments, international organisations, and researchers rely on forecasts of fertility, mortality, and migration to anticipate demand for schools, healthcare systems, housing, and pensions. Yet producing reliable projections remains a complex task. Population processes are shaped by changing social norms, economic conditions, and individual decision-making, all of which introduce uncertainty and make it difficult to capture future trends accurately.
Fertility, in particular, presents a persistent challenge. While overall birth rates are routinely projected, the underlying processes that drive them are more difficult to model. Many existing approaches rely on aggregate data—summarising births by age and time—without fully accounting for the individual pathways that lead to family formation. As a result, important aspects of demographic behaviour may be overlooked, limiting both the accuracy and the interpretability of projections.
A new study co-authored by Jakub Bijak with colleagues at the University of Southampton, Dr Joanne Ellison and Professor Erengul Dodd, addresses this challenge by introducing a more detailed way to model fertility dynamics as a staged process of decision making. Published in Demography, the article — “Can Incorporating Parity Information Improve the Reliability of Completed Cohort Fertility Projections?” — develops a novel statistical framework that incorporates the concept of parity, or the number of children a woman has already had, into fertility forecasts.
The key insight behind the research is that childbearing is inherently sequential. Decisions about whether to have a second or third child depend on prior births, and are made one at a time. This means that fertility cannot be fully understood as a single aggregate process. By explicitly modelling these birth-order dynamics, the authors show that it is possible to generate projections that are both more realistic and more informative.
By using a Bayesian modelling approach and data from 16 countries, the study demonstrates that incorporating parity information can improve the plausibility and, in many cases, the predictive performance of fertility projections. Beyond improving forecasts of overall fertility levels, and providing realistic assessment of their uncertainty, the method also provides insights into how different components of fertility—such as transitions to first births or progression to larger families—contribute to demographic change.
Co-author Jakub Bijak highlights the broader implications of this approach:
“Fertility is not just about how many children are born, but about the sequence of decisions that lead to those births. By incorporating parity into our models, we can better reflect the realities of family formation and produce projections that are both more credible and more useful for understanding future population changes.”
By linking methodological innovation with a deeper understanding of demographic behaviour, the study contributes to ongoing efforts to refine how population forecasts are produced. As demographic patterns continue to shift across countries, with fertility decline observed in many parts of the world, such advances are essential for ensuring that projections—even if uncertain—remain robust, transparent, and relevant for both research and policy.
You can read the article here.
