Stochastic Mortality across generations and regions: the case of Italy

In this thesis, I use doubly stochastic processes (or Cox processes) in order to model the random evolution of mortality of several individuals in Italy. I mainly focus on stochastic models, analogously of what has been studied in the credit risk literature. The calibration of parsimonious time-homogeneous affine processes to three different generations of the North and the South of Italy is provided. Adequate space is also dedicated to the analysis of the Infant Mortality Rate across regions in Italy. The estimate of the curves of the rates in order to capture features in the long-run is provided and according to the panel data technique we will estimate the impact of two important variables on the Infant Mortality Rate: the GDP per capita and the Cesarean section rate.