Do Different Models Induce Changes in Mortality Indicators? That Is a Key Question for Extending the Lee-Carter Model

The parametric model introduced by Lee and Carter in 1992 for modeling mortality rates the USA was a seminal development forecasting life expectancies has been widely used since then. Different extensions of this model, using different hypotheses about data, constraints on parameters, appropriate methods have led to improvements model’s fit historical data future. This paper’s main objective is evaluate if differences between models are reflected indicators’ forecasts. To end, nine sets indicator predictions were generated crossing three block-bootstrap samples with each size fifty. Later predicted indicators compared functional ANOVA. Models block bootstrap procedures applied Spanish data. Results show block-bootstrap, interaction effects all indicators. Although it not our objective, essential point out that sample effect should be present they must realizations same population, therefore procedure lead do influence results. Regarding significant effect, follows that, although addition terms improves adjustment probabilities translates into an indicators, checked their interest.