Modelagem aditiva generalizada para tarifação do seguro agrícola através da produtividade por município da cana-de-açúcar em Sergipe

Abstract

This study aims to propose the use of generalized additive model to calculate the pricing of the insurance contract for the cultivation of sugarcane in the State of Sergipe. The methodology used in this research integrates the advances of the actuarial calculation for the premium rate, due to the unavailability of information. The data used were obtained from the IBGE Automatic Recovery System (CIDER), relating to municipalities producers of sugarcane of Sergipe State in the period from 2001 to 2014. The Shapiro-Wilk normality test was applied to verify which municipalities had abnormality, and this way, to model the productivity of semi-parametric form. Then calculated the coefficient of variation for relative risk analysis. It was used the generalized additive model (GAM), adjusting it with spline smoothing to calculate estimates for pricing. It can be concluded that the model showed to be fair to estimate expected values of the agricultural insurance payments for the productivity of sugarcane in the municipalities of Sergipe. Still there was evidence that the high compensation amount paid by insurance companies can contribute to the little development of this industry by the private initiative in Sergipe