Stochastic Population Forecast: A Supra-Bayesian Approach to Combine Experts Opinions and Observed Past Forecast Errors

Rebecca Graziani, Università Bocconi
Eugenio Melilli, Università Bocconi

We suggest a method for developing stochastic population forecasts based on a combination of experts opinions and observed past forecast errors. The supra-Bayesian approach (Lindley, 1983) is used. The experts opinions on a given vital are treated by the researcher as observed data. The researcher specifies a likelihood, as the distribution of the experts evaluations given the rate value, parametrized on the basis of the observed past forecast errors of the experts and expresses his prior opinion on the vital rate by assigning a prior distribution to it. Therefore a posterior distribution for the rate can be obtained, in which the researchers prior opinion is updated on the ground of the evaluations expressed by the experts. Such posterior distribution is used to describe the future probabilistic behavior of the vital rates so to derive probabilistic population forecasts in the framework of the traditional cohort component model.