Dependence modeling of multivariate longitudinal hybrid insurance data with dropout

Financial services industries, such as insurance, increasingly use data from their broad cross-section of customers and follow these over time. In other areas medicine, engineering, communication systems, it is well known that following subjects time may result in biased data, for example, the so-called ”dropout effect”. This paper introduces techniques to address dropout commonly encountered insurance domain. Specifically, context, multivariate claims outcomes be related a customer’s or decision lapse policy. Insurance are also naturally hybrid with both discrete continuous components, which adds complexity model calibration. Decision makers industry will find our work provides helpful guidance integrating customer loyalty, especially bundled coverages. generalized method moments technique estimate dependence parameters where associations represented using copulas. useful large sets. The describes how joint new information insurers can better manage portfolios risks. An application Spanish insurer set presented.