After discussions with the National Statistical Institute on the possibility of incorporating uncertainty and economic structural relationships into their projections, AIReF has undertaken the task of producing its own populations forecast (in opposition to the official projections) including a measure of uncertainty and using as much as possible information coming from structural relationships.
Most experts tend to base their analysis on exogenously given population projections, which are faced with two main limitations. First, these projections usually are offered without any measure of the uncertainty surrounding them, which is problematic since AIReF’s budgetary forecasts are probabilistic. Second, official population projections tend to be based on number of methodological hypotheses that tend to be poorly motivated. In practice, for Spain, the latest population projections from the statistical institute have come closer to be a worst case scenario rather than a central forecast.
AIReF’s forecast methodology is based on a two-pronged approach, intending it to be statistically sound but also consistent with basic economic theory. In particular, we have represented fertility and mortality curves through time as a multiple time series model. Instead of letting them drift freely we have anchored them to a long term level based on a panel of European countries, under the hypothesis that, in the long run, convergence between countries similar along a series of characteristics will be more important than more idiosyncratic characteristics. Migration, on the other hand, has been forecast using a gravity model developed specially for this project.
Further details are provided in the two working papers
WP1: Predicting Spanish Emigration and Immigration
WP2: Introducing Uncertainty on Fertility and Survival in the Spanish Population Projections: A Monte Carlo Approach