Assessing time series models for forecasting international migration: Lessons from the United Kingdom

Abstract

Migration is one of the most unpredictable demographic processes. The aim of this article is to provide a blueprint for assessing various possible forecasting approaches in order to help safeguard producers and users of official migration statistics against misguided forecasts. To achieve that, we first evaluate the various existing approaches to modelling and forecasting of international migration flows. Subsequently, we present an empirical comparison of ex post performance of various forecasting methods, applied to international migration to and from the United Kingdom. The overarching goal is to assess the uncertainty of forecasts produced by using different forecasting methods, both in terms of their errors (biases) and calibration of uncertainty. The empirical assessment, comparing the results of various forecasting models against past migration estimates, confirms the intuition about weak predictability of migration, but also highlights varying levels of forecast errors for different migration streams. There is no single forecasting approach that would be well suited for different flows. We therefore recommend adopting a tailored approach to forecasts, and applying a risk management framework to their results, taking into account the levels of uncertainty of the individual flows, as well as the differences in their potential societal impact.

Global bilateral migration projections accounting for diasporas, transit and return flows, and poverty constraints

“…Potential future trends in global migration are therefore of interest. A number of different forecasting methods have been developed to predict future migration flows between countries or regions; relying, for instance, on time-series extrapolation (De Beer 1993), expert elicitation (Lutz, Sanderson, and Scherbov 1998), or combinations of both in Bayesian frameworks (Bijak and Wiśniowski 2010;Bijak et al 2019). Such methods can produce probabilistic forecasts that account for different sources of uncertainty and make extensive use of past migration statistics to inform likely future outcomes (Azose and Raftery 2015;Azose,Ševčíková, and Raftery 2016).…”

A brief guide to producing a national population projection

“…One option is to apply time series or econometric models. Examples of these types of models can be found in Bijak et al (2019), de Beer (2008, and Raymer and Wiśniowski (2018). Statistics Norway uses a model based on neoclassical economic theory to project various immigration streams (Cappelen et al 2015).…”

Short- to medium-run forecasting of mobility with dynamic linear models

“…Since we use national aggregates of mobility counts, the data are not prone to the modifiable areal unit problem. As stressed by Bijak et al (2019), the variability of migration data can be a major source of uncertainty in migration forecasts. In the context of mobility there could be biases for certain groups, such as students who stay registered at their parents' home.…”

“…A related problem with long time series is changes to administrative borders, where adjustments have to be made to achieve consistency (Husby et al 2014). Bijak et al (2019) carry out an empirical comparison of methods for forecasting (international) migration. They recommend a three-step process for migration modelling: (1) understand the features of the particular migration flow; (2) assess the available data; and (3) select a modelling approach appropriate for the type of migration and the available data.…”