Smoothing, Decomposing and Forecasting Mortality Rates

Area of Science:

• Demography
• Actuarial Science
• Biostatistics

Background:

• The Lee-Carter (LC) model is a foundational tool in mortality forecasting.
• Existing LC model variants address limitations individually, not comprehensively.
• A unified approach is needed to overcome multiple drawbacks simultaneously.

Purpose of the Study:

• Introduce the Three-Component smooth Lee-Carter (3C-sLC) model.
• Address multiple limitations of the original LC model concurrently.
• Enhance the accuracy and interpretability of mortality forecasts.

Main Methods:

• Decompose mortality into childhood, early-adult, and senescent components.
• Apply a smooth variant of the LC model to each component.
• Incorporate component-specific schedules for projections.

Main Results:

• The 3C-sLC model demonstrates superior performance compared to a smooth LC model.
• The model captures mortality developments more effectively.
• Wider prediction intervals are generated, reflecting increased uncertainty.

Conclusions:

• The 3C-sLC model offers a unified solution for smoothing, decomposing, and forecasting mortality.
• It provides valuable insights for actuaries, demographers, and epidemiologists.
• This approach enhances the robustness of mortality projections.