Using simple but flexible stochastic population forecasts to extend official population forecasts

Rebecca Graziani, Università Bocconi
Marco Marsili, Instituto Nazionale di Statistica (ISTAT)
Eugenio Melilli, Università Bocconi

In this paper we start from a formal framework for developing expert-based stochastic population forecasts, which starts from standard scenario-based population projections. Referring to a formalization developed in a previous paper, we use this approach to extend official Italian population projections developed by ISTAT over their actual forecast interval and turn projections into a stochastic version. The full probability distribution of forecasts is derived from 1) an initial scenario based on actual scenarios as available from standard deterministic projections); 2) expert opinions on developments conditional to the realization of high-low scenario.