Resumo

ANDREOZZI, Lucía. Probabilistic projections of the Argentine population. Pap. poblac [online]. 2021, vol.27, n.110, pp.41-78.  Epub 13-Jan-2023. ISSN 2448-7147.  https://doi.org/10.22185/24487147.2021.110.30.

In this study, models for functional data (MDF) are applied to the mortality and fertility data from Argentina to obtain their corresponding probabilistic forecasts and, from them, population projections. This exercise is interesting given the difficulties presented in the 2020 census round, as it could provide provisional figures pending its final completion. The application of these models to data from Argentina based on data from the period 1980-2010, the year of the last available census, shows that when there is information for not very long periods of time and an acceptable quality of the data sources, can be used successfully to make probabilistic forecasts. Second, the alternative consistent forecasting methodology is applied to mortality data in order to avoid a divergence between the two genders. Finally, from the fertility forecasts (based on two fertility hypotheses) and mortality (independent and coherent), stochastic forecasts of the population for men and women are elaborated. In all the cases evaluated, the estimated quantities are above the official ones. The results here are consistent with demographic theory and mainly correspond to probabilistic consistency.

Palavras-chave : Mortality; fertility; functional data; consistent forecasts.