Thesis title: A simplified framework for the analysis of profitability of Italian with-profit policies
The present work aims at devising a flexible quantitative tool to investigate the profitability of the Italian with-profit policies possibly on supervisory data.
We consider a stylized life-insurance company whose with-profit business model is inspired by a real company that is managed with a slightly negative duration gap and, accordingly to the aggregate features of its pool of contracts, we model the principal risk factors which affect the asset side.
In order to investigate the effects of book value Italian accounting rules we follows the approach in Mauer et al. (2016), that model the returns of the available-for-sale and held-to-maturity asset compartments. Given the aggregate features of a real pool of contracts and the inventory asset portfolio, we then adapt Gerstner et al. (2008) general ALM framework to the Italian with-profit business model.
In addition, given the notable share of whole-life contracts, we introduce a novel stochastic logistic model based on a non-homogeneous Poisson process for the modelling of lapse (surrender) risk; thanks to the modular structure of Gerstner et al. (2008) ALM model, this is expected to improve the sensitivity analyses of lapse risk.
Correlation analysis between risk drivers is easily achieavable (by means of the concept of affine dependence) and is left for future research.
To gain an hindsight on the risk-return features of the segregated funds investigated, an exploratory ARMA(1,1) time-series model is calculated based on effective supervisory data.
Montecarlo numerical implementation is partially available and will be the object of future work.