Forecasting small population monthly fertility and mortality data with seasonal time series methods

Jorge Miguel Bravo, Edviges Isabel Felizardo Coelho

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

86 Downloads (Pure)

Abstract

Forecasts of small population monthly fertility and mortality data are a critical input in the computation of subnational forecasts of resident population since they determine, together with internal and international net migration, the dynamics of both the population size and its age structure. Demographic time series data typically exhibit strong seasonality patterns at both national and regional levels. In this paper, we evaluate the short-term forecasting accuracy of alternative linear and non-linear time series methods (seasonal ARIMA, Holt-Winters and State Space models) to birth and death monthly forecasting at the local and regional level. We adopt a backtesting time series cross-validation approach considering a multi-step forecasting approach with re-estimation. Additionally, we investigate the model’s performance in terms of forecasting uncertainty by computing the percentage of actual monthly births and death counts which fall out of prediction intervals. We use a time series of monthly birth and death data for the 25 Portuguese NUTS3 regions from 2000 to 2018, disaggregated by sex.
Original languageEnglish
Title of host publicationAs Ciências Sociais Aplicadas e a Interface com vários Saberes
EditorsWendell Luiz Linhares
PublisherAtena
Pages158-176
Number of pages18
Volume2
ISBN (Print)978-85-7247-979-0
DOIs
Publication statusPublished - Jan 2020

Keywords

  • Small population forecasts
  • SARIMA
  • Backtesting
  • seasonality
  • State Space models

Fingerprint

Dive into the research topics of 'Forecasting small population monthly fertility and mortality data with seasonal time series methods'. Together they form a unique fingerprint.

Cite this