Multivariate time changes for Lévy asset models: characterization and calibration (Q2654202)

The authors introduce a class of multivariate time changes; each one is a sum of an idiosyncratic and a common component. Since through these changes of time Lévy processes must be obtained, it is assumed that the time changes are subordinators. Then the corresponding time changed Brownian motions are characterized. In order to justify modeling choices and their usefulness for financial applications, three desired features are in mind: the existence of characteristic functions in closed form, the ability to capture a wide range of linear dependence and the possibility of calibrating marginal and joint parameters separately. It is demonstrated that, as long as the multivariate subordinator is applied to independent Brownian motions, the model is extremely parsimonious in terms of parameters.