In this paper we motivate, specify and estimate a model in which the intra-day volatilty process affects the inter-transaction duration process and vice versa. In order to solve the estimation problems implied by this interdependent formulation, we first propose a GMM estimation procedure for the Autoregressive Conditional Duration model. The method is then extended to the simultaneous estimation of the interdependent duration-volatility model. In an empirical application we utilize the model for an indirect test of the hypothesis that volatility is caused by private information that affects prices when informed investors trade. The result that volatility shocks significantly increase expected inter-transaction durations supports this hypothesis.
Inter-transaction duration and volatility financial market microstructure ultrahigh frequency data autoregressive conditional duration