Abstract
AbstractIn this paper, we use TMB to study spatial variation in weather-generated claims in insurance. Our motivation is twofold. By comparing with INLA, we first find that TMB is a robust and efficient approach to deal with spatial variation of covariates and the dependent variable in a case with sparse data. Second, we demonstrate how examining the spatial pattern of random effects may offer auspicious suggestions for model extensions, represented by added covariates accounting for relevant spatial characteristics. Both the approach and the results represent useful input in reaching an efficient spatial diversification of premium rates in non-life insurance.
Publisher
Springer Science and Business Media LLC
Subject
Applied Mathematics,Control and Optimization,Computer Science Applications,Economics, Econometrics and Finance (miscellaneous)