Anticipating Continued Global Fertility Decline via Neural Forecasting
Abstract
As fertility falls across much of the world, major projection systems disagree about whether decline will give way to recovery, reflecting different assumptions about post-transitional fertility. We ask instead what the historical record suggests, using NeuralTFR, a recurrent neural network that learns fertility dynamics solely from pooled histories of 230 countries and territories. In out-of-sample forecasts from three origins, it predicts fifteen years ahead as accurately as the United Nations' Bayesian model, and its prediction intervals score slightly better. Its forecasts reveal a simple nonlinear relationship, which a spline regression reproduces with similar accuracy: fertility tends to rise slightly from the lowest levels and decline from higher ones, with recent trends affecting the size of these changes more than their direction. By 2038, these dynamics leave very low fertility, below 1.5 births per woman, more widespread than the United Nations expects, without the extreme lows of the most pessimistic projections.
Related
- The cost of ensembling: Is it always worth combining global retail forecasting models?
- Series-wise Cross-validation for Global Time Series Forecasting
- Dynamical-Model-Informed Stochastic Flow Matching for Global Probabilistic Subseasonal Forecasting of Means and Extremes
- Physics-Guided Flow Matching: A Generative AI Framework for Global Total Electron Content Forecasting
- Simultaneously small continued fraction entropy and base b entropy
BibTeX
@misc{ciganda2026anticipating,
title = {{Anticipating Continued Global Fertility Decline via Neural Forecasting}},
author = {Daniel Ciganda and Facundo Morini and Francisco Piriz and Henrik-Alexander Schubert and Ugofilippo Basellini and Mikko Myrskylä},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2605.23858},
eprint = {2605.23858},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.23858},
note = {CC BY-SA 4.0},
}
RIS
TY - UNPB
TI - Anticipating Continued Global Fertility Decline via Neural Forecasting
AU - Daniel Ciganda
AU - Facundo Morini
AU - Francisco Piriz
AU - Henrik-Alexander Schubert
AU - Ugofilippo Basellini
AU - Mikko Myrskylä
PY - 2026
DA - 2026/10/29
PB - arXiv
DO - 10.48550/arxiv.2605.23858
UR - https://arxiv.org/abs/2605.23858
AB - As fertility falls across much of the world, major projection systems disagree about whether decline will give way to recovery, reflecting different assumptions about post-transitional fertility. We ask instead what the historical record suggests, using NeuralTFR, a recurrent neural network that learns fertility dynamics solely from pooled histories of 230 countries and territories. In out-of-sample forecasts from three origins, it predicts fifteen years ahead as accurately as the United Nations' Bayesian model, and its prediction intervals score slightly better. Its forecasts reveal a simple nonlinear relationship, which a spline regression reproduces with similar accuracy: fertility tends to rise slightly from the lowest levels and decline from higher ones, with recent trends affecting the size of these changes more than their direction. By 2038, these dynamics leave very low fertility, below 1.5 births per woman, more widespread than the United Nations expects, without the extreme lows of the most pessimistic projections.
ER -