Centenarian population growth does not require extension of the fitted lifespan tail across five countries
Abstract Centenarian populations are growing rapidly, but an increasing number of people aged 100 years or older need not indicate extension of the extreme lifespan tail. We analyzed harmonized population counts and sex-specific period life tables from the Human Mortality Database for Japan, Sweden, France, Italy, and South Korea. We linked old-age population stocks to a modified stretched-exponential survival representation and evaluated 317 one-year-ahead centenarian-population predictions. The pooled mean absolute percentage error was 3.02%, compared with 3.02% for a direct mortality-data benchmark and 8.34% for naive persistence. Thus, forecasting accuracy primarily reflects explicit accounting for entry at age 100 and survival of existing centenarians rather than superiority of the fitted survival function. The representation also supplies a characteristic age (α) and a model-derived upper-tail descriptor (ω). Across all five countries, α increased, whereas ω showed no comparable long-term rise; recent estimates were approximately 117–122 years. Exploratory scenarios suggest further centenarian growth to 2050 but are sensitive to calibration assumptions. These findings distinguish growth in centenarian abundance from changes in the fitted lifespan tail without implying a fixed biological maximum lifespan.
Authors
- Byung Mook Weon (ORCID: https://orcid.org/0000-0002-5224-5590)
Institutions
- Sungkyunkwan University (KR)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-13
- DOI
- https://doi.org/10.21203/rs.3.rs-11326424/v1
- Primary Topic
- Insurance, Mortality, Demography, Risk Management
- Type
- preprint
Funders
- Ministry of Health, Labour and Welfare
- National Research Foundation of Korea
- Ministry of Science and ICT, South Korea