An interpretable piecewise model for human mortality: the Christopher–Samba model
摘要
Modelling mortality schedules has a long history in demographic research, with many models developed over time. This study proposes a new mortality model—the Christopher–Samba (C–S) model—that captures mortality patterns across the entire age spectrum. The proposed piecewise model integrates two components: The first part is an exponential function describing mortality up to age 5, while the second part focuses on mortality above age 5. The parameters of the second component, a reparametrized version of the classical logistic model, reflect the ages at which the annual probability of death reaches 1% and 10%, representing the benchmark ages for the onset of increased mortality and severe mortality, respectively. These thresholds naturally vary across populations. Mortality data from the Human Mortality Database were used to demonstrate the application of the proposed C–S model. The C–S model fits well across countries with diverse mortality schedules. Fitted parameters reveal substantial variation in the ages at which the annual probability of death reaches 1% and 10%, underscoring population differences in mortality dynamics. In summary, with its simplicity, strong fit, and meaningful parameter interpretations, the proposed C–S model is a valuable addition to mortality modelling literature.