International migration flows are considered the most difficult demographic component to forecast and, for that reason, models for forecasting migration are few and relatively undeveloped. This is worrying because, in developed societies, international migration is often the most influential component of population growth and in debates about societal and economic change. In this paper, we address the need for better forecasting models of international migration by testing a Bayesian functional model recently developed to forecast age and sex patterns of immigration and emigration in the United Kingdom on other types of migration flow data. For the applications, we obtained age- and sex-specific time series data from Sweden, South Korea and Australia. The performance of the forecasts are compared and assessed with observed time series data. The results demonstrate the generality and flexibility of the Bayesian functional model for forecasting migration, as well as highlighting areas for further research.
James Raymer is a Professor in the School of Demography at the Australian National University. From 2013 to 2016, he served as the Director of the Australian Demographic and Social Research Institute before leading the creation of the new School of Demography in 2015 and becoming the Head of School. Prior to that, he worked in the Division of Social Statistics and the ESRC Research Centre for Population Change at the University of Southampton, England, after obtaining his PhD in Geography from the University of Colorado, Boulder, United States. He has published articles in demography, applied statistics, regional science and geography, and has co-authored three books entitled The Indirect Estimation of Migration, Demographic Aspects of Migration and International Migration in Europe. He is currently leading two research projects funded by the Australian Research Council on ‘the demographic consequences of migration to, from and within Australia’ and ‘overcoming the problems of inconsistent migration data in the Asia Pacific.’
Arkadiusz Wisniowski is a Lecturer in Social Statistics at the University of Manchester. He is also a member of the Cathie Marsh Institute. Prior to this, he was a Research Fellow at the ESRC Centre for Population Change and the Southampton Statistical Sciences Research Institute, University of Southampton. His research concentrates on developing statistical methods for combining various sources of data. He developed a set of harmonised estimates of migration amongst 31 EU and EFTA countries for 2002-2008 in the Integrated Modelling of European Migration project and a model for probabilistic forecasting of population in the UK by age and sex. Currently he is working on combining traditional data sources such as census, administrative and survey with social media data to estimate and forecast internal movements of populations. He also has a general interest in time series analysis and forecasting, hierarchical models, Bayesian computational methods and demographic analyses.