Documentation and Knowledge Center of Fundación Mapfre > A Log-normal chain ladder model closely aligning with Mack's assumptions

Notes: Sumario: The article introduces the log-normal Mack chain ladder (LMCL) model, a stochastic formulation of the chain ladder method that is closely aligned with Mack's classical assumptions and compatible with maximum likelihood estimation. Simple iterative procedures are developed for parameter estimation, and reserving methods with and without bias correction are proposed. In addition, analytical estimators of the mean squared error of prediction are derived. The model is empirically compared with the traditional chain ladder method, Hertig's model, and other approaches, showing good performance even for triangles with high volatilityRelated records: En: European Actuarial Journal. - Cham, Switzerland : Springer Nature Switzerland AG, 2021-2022. - 13/04/2026 Número 16 issue 1 - abril 2026 , 68 p.Materia / lugar / evento: Modelos actuariales Provisiones técnicas Seguros no vida Inferencia estadística Modelos estadísticos Otros autores: Springer Other categories: 6