As international migration continues to evolve in scale, complexity and unpredictability, governments face increasing pressure to respond swiftly and effectively often under tight operational constraints and high public scrutiny. Anticipation is therefore becoming a core capability of migration governance. At the same time, forecasting migration remains inherently challenging: flows are shaped by rapidly changing geopolitical and economic conditions, policy interventions can alter the very outcomes that are being forecast, and data availability and quality vary widely across migration categories. In this context, forecasting is not about eliminating uncertainty; it is about making uncertainty actionable supporting preparedness, informing choices, and improving the timeliness and transparency of decisions.
This Handbook provides a practical guide to strengthening migration forecasting systems and preparedness strategies across OECD countries. It draws on country practices, state‑of-the‑art research and the lessons of the OECD’s Migration Anticipation and Preparedness (MAP) Task Force. Bringing together perspectives from national institutions and academia, it addresses the questions that policymakers, analysts and practitioners consistently face: what forecasting should be used for, why migration needs to be forecast by category rather than only in aggregate, how to select models and data that fit specific purposes, how to evaluate performance credibly, and how to communicate results to policymakers, especially uncertainty so that forecasts can be used responsibly.
The Handbook is structured around nine chapters that follow the forecasting lifecycle from problem definition to communication and system learning. After the introduction and general overview, Chapter 3 clarifies the policy question that forecasting is expected to answer and the main purposes forecasts can serve. Chapter 4 discusses how to choose models that fit policymakers’ needs, migration categories and operational constraints, including questions of model complexity and update cycles. Chapter 5 focusses on data foundations and system requirements, including ways to assess data sources, incorporate qualitative information, and operationalise policy indicators. Chapter 6 addresses model development, covering traditional time‑series approaches, Bayesian modelling, structured integration of expert judgement (including Delphi surveys), and the role of machine learning and causal methods. Chapter 7 sets out how models can be evaluated, validated and adapted through back-testing, calibration and robustness checks. Chapter 8 focusses on communication and how to present forecasts and uncertainty to policymakers, and how to support effective use in practice. Finally, Chapter 9 looks ahead to emerging developments in migration forecasting, including new approaches to uncertainty quantification and the growing importance of interactions across categories and countries.
This project was financially supported by the Directorate General for foreign nationals in France (DGEF), Ministry of the Interior, France, whose contribution made this Handbook possible. The Handbook was edited by Yves Breem (DGEF, French Ministry of the Interior) and Taehoon Lee (OECD). It was written by Jakub Bijak (University of Southampton), Yves Breem, Marcello Carammia (University of Catania), Stefano M. Iacus (Joint Research Centre; formerly Harvard University), and Taehoon Lee, in alphabetical order. The Handbook also benefited from presentations, participation and written comments provided by members of the OECD Migration Anticipation and Preparedness (MAP) Task Force. In addition, it benefited from comments by Jean-Christophe Dumont and Jonathan Chaloff (OECD) and Guillaume Mordant (DGEF, French Ministry of the Interior).
With practical tools, decision frameworks and real-world examples, this Handbook aims to support countries in building forecasting systems that are fit for purpose, resilient to changing conditions and connected to decision making. By strengthening the link between forecasting and preparedness, it seeks to help governments respond more effectively to uncertainty grounded in evidence, informed judgement and continuous learning.