EDA ja EDK ETTEKANDEKOOSOLEK

A NEW PROBABILISTIC METHOD FOR FORECASTING MORTALITY:
A CASE STUDY OF ESTONIA

19.08.2026 

Värska, EDK suveseminar

Etteknde teeb: David A. Swanson

Department of Sociology, University of California Riverside;Center for Studies in Demography and Ecology; University of Washington; Population Research Center; Portland State University; Social Science Research Center; Mississippi State University

Abstract

A new probabilistic mortality forecasting approach is introduced that, unlike the Lee-Carter Method (and its variants), is directly linked to the fundamental demographic equation, the cornerstone of demographic theory. This is an important consideration in developing accurate forecasts. Because it forecasts "years lived," this new approach directly yields life expectancy and a corresponding future life table (examples are provided in the Appendix, along with a description of the steps needed to construct them), unlike the Lee-Carter Method and its variants. In an ex post facto evaluation using Estonian data from the Human Mortality Database, the new approach was found to provide accurate forecasts of "years lived by age" (_nL_x) for both point and interval measures over 15 years. In a broader context, the accuracy of Estonia's point _nL_x forecasts is compared with that of Algeria, Australia, Japan, and the United States. Probabilistic _nL_x forecasts for Estonia from 2029 to 2039 are then provided, the results are discussed, and the next steps for evaluating this approach are suggested.

PRESENTATION