Abstract
This paper introduces a new approach for adjusting the diurnal variation in the trade durations. The model considers that durations are multiplicatively decomposed into a deterministic time-of-day and a stochastic component. The parametric structure of the diurnal component allows the duration process to change smoothly over the time-of-day. In addition, a testing framework consisting of Lagrange multiplier tests is proposed for specifying the diurnal component. Our methodology is applied to the IBM transaction durations traded at the New York Stock Exchange.