preview

Distribution Of The Risk Burden Essay

Decent Essays

1. Introduction:

Equitable distribution of the risk burden must be structured using advanced statistical methods to divide the portfolio into homogenous classes. thus, each insured belongs to a particular class can pay the pure premium Proportional to the risk degree of this class.

Over the last decade, generalized linear model (GLM) was a common statistical tool for a prior classification see; e.g. Dionne and Vanasse (1989), Haberman and Renshaw (1996),
Pinquet (1999), Bermúdez et al. (2001) and Boucher and Denuit
(2006). The merits of this model are twofold. Firstly , the random deviations from the mean may have a distribution not normal (kaas et al. 2008). Secondly, the transformed mean of dependent variable may be a linear function of the explanatory variables which determined by the link function.

However, this model have two drawbacks; firstly, It consider the random variables is independent and that may be not fulfilled in case of longitudinal, repeated measurements and spatial data. Secondly, the model is not convenient for unknown nonlinear effect of independent variables because the GLM models consider the independent variables have linear effect . The first shortage can be covered by GLMM models also this models used to superimpose credibility on a GLM setting see; e.g. Klinker (2011) and Antonio (2007). The second shortage can be covered by GAM models that permit us to overcome the linearity assumptions inherent to GLMs.

The spatial models give us

Get Access