Volatility in International Migration Flows of Nordic Countries: Estimating Past Trends and Lessons for Forecasting with Uncertainty

• Activity Number: 298
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): Guy Abel*+
• Companies: Wittgenstein Centre (IIASA, VID/OAW, WU), Vienna Institute of Demography
• Keywords: migration ; time series ; volatility ; forecasting ; Bayesian ; model averaging

Abstract

International migration flows are often volatile over time, making it the most difficult component of population change to forecast. This paper aims to model the volatility of migration flows in Nordic countries using a range of Bayesian time series models. In doing, past levels of variation in migration are estimated and forecasts with uncertainty are produced.

The migration data are explored in three parts. First, models with a variety of variance structures are fitted to crude net migration rates. Second, the same set of models are applied to series of immigration counts and emigration rates. Third, forecasts are conducted on truncated samples of all three migration measures. The estimated model averaged volatility levels in each migration series are presented. The performance of the model averaged forecasts from the truncated series are then discussed, comparing the predictive distributions of the net migration based on 1) the net migration series and 2) the difference of the immigration and emigration series. This paper demonstrates the use of volatility models and model averaging methods to provide realistic measures of uncertainty in international migration forecasts.