Inferences on mortality using the Heligman-Pollard model: the Mexican case

Abstract 
The distribution of the parameters of the Heligman-Pollard model is calculated based on the Mexican mortality and Chebyshev-type confidence intervals are generated. In order to estimate life expectancy at birth, death probability predictions are made from 2016 to 2025, where a VAR(p) models are used. The results show that the mortality by sex and for the entire population are distributed as Gamma and Weibull respectively. We estimated coherent life expectancy and compared them with official data. Our forecasts are different and we consider that it is attributable to the statistical model employed and to the more updated datasets.