Title: Further developments of the ratio model of Cox-type intensities for high frequency financial data
Authors: Ioane Muni Toke - CentraleSupelec (France) [presenting]
Abstract: A model of ratio of Cox intensities sharing a common baseline intensity has been recently proposed. We have shown that such a model is particularly well designed for high frequency financial data, where the common baseline intensity may represent a rapidly varying market activity. We are, for example, able to estimate relative effects of parameters of interest on the intensity of submission of orders in a limit order book. We now build on this model in several directions. In a first direction, we investigate the variations across time of the influences identified by the ratio model, and show the benefits of the model as an econometric tool. In a second direction, we further investigate limit order book modeling by developing a combined model of Cox- and Hawkes-type intensities (with multiple exponential kernels for the latter type) for the submission of orders. The estimation of such a model with (ultra) high frequency data is a challenging task that is carried out on several stocks and exchanges.