Title: Flexible panel regression model for bivariate count/continuous data with insurance applications
Authors: Yang Lu - Aix-Marseille University (France) [presenting]
Abstract: A semi-parametric bivariate panel regression model is proposed that is suitable for mixed count/continuous variables. We develop a polynomial expansion approach for the distribution of the correlated bivariate random effect. Besides the distributional flexibility it offers, the model allows for closed form forecast updating formulas on an individual basis. This property is especially important for insurance applications, in which the future individual insurance premium should be regularly updated according to ones own past claim history. An application on vehicle insurance claims is provided.