Long-Term Immigration Projection Methods: Current Practice and How to Improve it
Abstract
Assumptions about international migration are an increasingly important component of the demographic projection module. Yet most official immigration projections both in the United States and abroad rely on ad-hoc assumptions based on little theory and virtually no definable methodology. The purpose of the present report is threefold: to assess the state of immigration projection practice; to explore theoretical insights and empirical research about immigration; and to discuss how these insights and research could be used to create a superior projection model. The first chapter describes the current projection methods of leading national and international projection-making agencies, from the U.S. Census Bureau and U.S. Social Security Administration to the United Nations and the World Bank. The second chapter scans the wide array of “theoretical frameworks” that social scientists from a variety of disciplines have proposed to explain international migration flows. It also surveys the growing body of empirical literature on efforts to statistically test the power of these theories in explaining historical trends in international migration. The third chapter describes our proposed framework for a long-term immigration projection model. The model is “driver-based,” meaning that immigration is projected based on observed associations between migration behavior and other conditions, such as multinational trends in population, wages and living standards, trade and capital flows, age distribution, education, urbanization, and market orientation or “globalization.” We conclude that this approach has the potential to greatly improve long-term projections of net immigration to a major developed country like the United States.