Related Experiment Videos
[Uncertainty variants in population forecasts for the Netherlands]
Summary
Netherlands population forecasts have inherent uncertainty. A statistical confidence interval can be derived, showing the Low and High variants approximate a two-thirds confidence interval for future population sizes.
Area of Science:
- Demography
- Statistical Modeling
Context:
- The Netherlands Central Bureau of Statistics (CBS) uses deterministic models for population forecasts, publishing Low, Medium, and High variants.
- The probabilistic interpretation and reliability of these forecast intervals are currently unknown.
Purpose:
- To derive a statistical confidence interval for total population size based on forecast errors.
- To assess the reliability of the CBS's existing Low and High forecast variants.
Summary:
- Population forecast errors are assumed to be serially correlated.
- A first-order autoregressive model was estimated using CBS population forecasts published since 1950.
- The interval between the Low and High variants approximates a two-thirds confidence interval for the next two decades.
Impact:
- Provides a method to quantify the uncertainty in deterministic population forecasts.
- Suggests that existing forecast intervals offer a limited, but quantifiable, level of confidence.
- Informs the interpretation and use of national population projections.