Modelling Security Market Events in Continuous Time: Intensity based, Multivariate Point Process Models

A continuous time econometric modelling framework for multivariate market event (or 'transactions') data is developed in which the model is specified via the vector stochastic intensity. This has the advantage that the conditioning sigma-field is updated continuously in time as new information arrives. We introduce the class of generalised Hawkes models which allow the estimation of the dependence of the intensity on the events of previous trading days. Analytic likelihoods are available and we show how to construct diagnostic tests based on the transformation of non-Poisson processes into standard Poisson processes using random changes of time scale. A proof of the validity of the diagnostic testing procedures is given that imposes only a very weak condition on the point process model, thus establishing their widespread applicability. A continuous time bivariate point process model of the timing of trades and mid-quote changes is presented for a NYSE stock and the empirical findings are related to the theoretical and empirical market microstructure literature.