2012/12/03 by Bailey K. Fosdick, Adrian E. Raftery, Fosdick, Bailey K. +1 · 1 citation
Decision Sciences · Economics, Econometrics and Finance · Social Sciences · #Applications (stat.AP) #Economic Growth and Productivity #Economics of Agriculture and Food Markets #FOS: Biological sciences #FOS: Computer and information sciences #Insurance, Mortality, Demography, Risk Management #Populations and Evolution (q-bio.PE) #demographic modeling and climate adaptation
paper · pdf · doi:10.48550/arxiv.1212.0462
openalex publication_date 2012/12/03 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The United Nations (UN) Population Division is considering producing\nprobabilistic projections for the total fertility rate (TFR) using the Bayesian\nhierarchical model of Alkema et al. (2011), which produces predictive\ndistributions of TFR for individual countries. The UN is interested in\npublishing probabilistic projections for aggregates of countries, such as\nregions and trading blocs. This requires joint probabilistic projections of\nfuture country-specific TFRs, taking account of the correlations between them.\nWe propose an extension of the Bayesian hierarchical model that allows for\nprobabilistic projection of TFR for any set of countries. We model the\ncorrelation between country forecast errors as a linear function of time\ninvariant covariates, namely whether the countries are contiguous, whether they\nhad a common colonizer after 1945, and whether they are in the same UN region.\nThe resulting correlation model is incorporated into the Bayesian hierarchical\nmodel's error distribution. We produce predictive distributions of TFR for\n1990-2010 for each of the UN's primary regions. We find that the proportions of\nthe observed values that fall within the prediction intervals from our method\nare closer to their nominal levels than those produced by the current model.\nOur results suggest that a significant proportion of the correlation between\nforecast errors for TFR in different countries is due to countries' geographic\nproximity to one another, and that if this correlation is accounted for, the\nquality of probabilitistic projections of TFR for regions and other aggregates\nis improved.\n