Publication: Análisis del riesgo de caída de cartera en seguros: metodologías de “inteligencia artificial” vs “modelos lineales generalizados”
Loading...
Files
Official URL
Full text at PDC
Publication Date
2017-01-13
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Complutense de Madrid
Citations
Exportar
Abstract
Han transcurrido varios años desde que se comenzó a hablar de Solvencia II y hoy es una realidad; cuyo objetivo es el desarrollo y establecimiento de un nuevo sistema que permita determinar los recursos propios mínimos a requerir a cada aseguradora, en función de los riesgos asumidos y la gestión que se realice de ellos. Así mismo, engloba un conjunto de iniciativas para la revisión de la normativa existente, la valoración y supervisión de la situación financiera global de las entidades aseguradoras y modos de actuación interna de las mismas. Uno de los temas más controvertidos bajo esta regulación es cómo conseguir una adecuada evaluación de los riesgos asumidos por las entidades. Esto se traduce en lograr identificar las causas que puedan suponer una pérdida en sus recursos; así como en innovar en el campo técnico para lograr una correcta cuantificación de los riesgos posibles en los que podrían estar expuestas las entidades. El objetivo de este trabajo es mostrar la posibilidad de utilizar dos enfoques metodológicos distintos para la evaluación de riesgos: uno no paramétrico para lo cual se recurrirá a las técnicas de Inteligencia Artificial y, en contraste, la aplicación de los Modelos Lineales Generalizados provenientes de la estadística paramétrica. De esta forma, lograr establecer una serie de reglas de decisión básicas, a manera de herramienta de clasificación, que puedan ser capaces de determinar los perfiles de clientes susceptibles a la cancelación de su póliza. La aplicación práctica de ambas metodologías, se llevará a cabo con la finalidad de analizar el Riesgo de Caída de Cartera; el cual hace referencia a uno de los tantos riesgos medibles que el sector habrá de tener en cuenta de acuerdo a Solvencia II...
Now, Solvency II is a reality after several years talking about it. Its objective is the development and establishment of a new system to ensure minimum capital requirements for each insurance company, depending on the risks assumed and the best management of them. It also includes a set of initiatives for existing legislation review; for evaluating and monitoring overall financial situation of insurers, as well as internal action procedures to control it. With this background, one of the most controversial issues presented by Solvency II is how to get a proper assessment of risks assumed by the entities. First, entities have to identify the causes that may lead to a loss on entities resources, and then they have to innovate on technical fields for the best estimation of the potential risks they might be exposed. The aim of this paper is to show the possibility of using two types of methods: a non-parametric using Artificial Intelligence techniques; in contrast to the results obtained with the parametric statistics by using Generalized Linear Models. Thus, to achieve a set of basic decision rules, as a classification tool that may be able to determine the profiles of policy customers susceptible to cancellation. A practical application of both methodologies will be done in order to analyze Lapse Risk which refers to one of risk that insurance entities must take account under Solvency II regulatory. The relevance of both empirical applications will be able to have an approximation of probability of customer cancellation by those patterns. It would become an improvement in management of lapse risk, contributing to balance and stability of entity solvency levels...
Now, Solvency II is a reality after several years talking about it. Its objective is the development and establishment of a new system to ensure minimum capital requirements for each insurance company, depending on the risks assumed and the best management of them. It also includes a set of initiatives for existing legislation review; for evaluating and monitoring overall financial situation of insurers, as well as internal action procedures to control it. With this background, one of the most controversial issues presented by Solvency II is how to get a proper assessment of risks assumed by the entities. First, entities have to identify the causes that may lead to a loss on entities resources, and then they have to innovate on technical fields for the best estimation of the potential risks they might be exposed. The aim of this paper is to show the possibility of using two types of methods: a non-parametric using Artificial Intelligence techniques; in contrast to the results obtained with the parametric statistics by using Generalized Linear Models. Thus, to achieve a set of basic decision rules, as a classification tool that may be able to determine the profiles of policy customers susceptible to cancellation. A practical application of both methodologies will be done in order to analyze Lapse Risk which refers to one of risk that insurance entities must take account under Solvency II regulatory. The relevance of both empirical applications will be able to have an approximation of probability of customer cancellation by those patterns. It would become an improvement in management of lapse risk, contributing to balance and stability of entity solvency levels...
Description
Tesis inédita de la Universidad Complutense de Madrid, Facultad de Ciencias Económicas y Empresariales, Departamento de Economía Financiera y Contabilidad I, leída el 16/12/2015