战略与国际研究中心-Long_128页_737kb

Long-Term Immigration Projection Methods: Current Practice and How to Improve It

Core Content

This report, authored by Neil Howe and Richard Jackson with contributions from Jennifer Blakeslee and Keisuke Nakashima, examines the current state of long-term immigration projection methods and proposes an improved approach. It highlights the lack of theoretical and empirical rigor in existing practices and advocates for a more systematic, data-driven model to enhance the accuracy and reliability of immigration forecasts.

Main Points

Current Projection Practice

• Ad Hoc and Judgmental: Most national and international projection agencies rely on informal methods, often using "expert opinion" or "historical experience" to make assumptions about future immigration.
• Limited Theoretical Justification: Few projections are grounded in explicit theories of immigration behavior, despite the wealth of theoretical frameworks available in social science and policy disciplines.
• Variability in Assumptions: The range of assumptions about immigration can significantly affect long-term population projections. For instance, the U.S. Census Bureau's 2100 projections vary by 417 million people depending on immigration assumptions.
• Focus on Total Numbers: Most agencies project total immigration numbers (gross or net) rather than detailed demographic or socioeconomic distributions.

Theoretical Insights

• Six Theoretical Frameworks: The report outlines six major frameworks for understanding immigration:
  • Neoclassical
  • World Systems
  • New Economics
  • Social Network
  • Dual Labor Market
  • Policy Frameworks
• Empirical Research: A growing body of empirical studies is testing the explanatory power of these theories against historical migration trends.
• Key Variables: Theories suggest that immigration is influenced by demographic, economic, and social factors, including population age structure, wage differentials, education levels, urbanization, and public opinion.

Proposed Driver-Based Model

• Driver-Based Approach: The report proposes a model that links immigration trends to a wide range of "drivers," including:
  • Built-in Demographic Drivers: Age structure of origin countries.
  • Modeled Demographic Drivers: Growth rates of youth populations in origin countries and aged dependency ratios in destination countries.
  • Economic and Development Drivers: Wage and living standard differentials, education levels, and development indicators.
  • Nonpolicy Drivers: Income inequality, trade, technology, and environmental changes.
  • Policy Drivers: Public opinion and the skill level of immigrant populations relative to native-born workers.
• Order of Drivers: Drivers are introduced in order of presumed certainty, allowing projection agencies to set thresholds between plausible and less plausible assumptions.
• Scenario Building: The model enables the creation of various scenarios, which can help policymakers understand the long-term implications of different immigration trends and policies.
• Potential Benefits: A more accurate projection model can improve understanding of future population changes, which are critical for addressing policy challenges related to retirement systems, healthcare, and economic growth.

Key Information

• Importance of Immigration: Immigration is now a dominant factor in population growth in developed countries, with net immigration accounting for about 40% of U.S. population growth and 90% of EU-15 population growth.
• Need for Improvement: Despite the significance of immigration, current projection methods lack a robust theoretical and empirical foundation.
• Challenges: Immigration projections are more uncertain and complex than fertility or mortality projections due to the variety of influencing factors and the lack of a unified theory.
• Future Directions: The report suggests that developing a formal, driver-based model could lead to more reliable and useful projections, potentially involving collaboration between immigration theorists, empirical researchers, and projection experts.

Conclusion

• The report emphasizes that while progress has been made in fertility and longevity projections, immigration remains a critical and underdeveloped area.
• A systematic, theory-informed, and data-driven approach is necessary to improve the accuracy of long-term immigration projections.
• The proposed driver-based model could serve as a foundation for more comprehensive and reliable demographic forecasting, aiding policymakers in addressing future challenges.

Recommendations

• Formalize Projection Models: Agencies should move away from ad-hoc assumptions and adopt formal, theory-based models.
• Incorporate Probabilistic Methods: Probabilistic approaches can help assess the likelihood of different immigration scenarios, improving the clarity and precision of projections.
• Integrate Theoretical Insights: Combining diverse theoretical frameworks can lead to a more holistic understanding of immigration trends.
• Collaborative Efforts: Developing such models may require multi-year collaborative projects involving various experts and agencies.

Appendices and References

• The report includes an appendix listing members of the CSIS Working Group on Long-Term Immigration Projections.
• It references a wide range of academic and policy sources, including works by Stephen Castles, Mark J. Miller, Douglas S. Massey, and Eurostat, to support its analysis and recommendations.