An Examination of the Role of Time and its Impact on Price Revision

Abstract

We consider a new class of time series models (introduced by Engle and Russell (1998)) used in statistical applications in finance. These models treat the time between events (durations) as a stochastic process and the corresponding durations are modelled using a theory similar to that of autoregressive processes. On a sample of six stocks listed on the ASX, We find evidence in support of the important role that both the deterministic and stochastic components of time play in both our quote revision and signed trade equations, and it is the stochastic indicator of time that has a greater influence than the time-of-day periodicities.

Keywords: Autoregressive, Conditional expectation, Intensity, Hazard func-tion, Stochastic process, Prediction, Estimation, Irregular data, Transaction data, Finance, Mid-quote price, Trade indicators, Autocorrelations, Volatility

JEL Classification: C22, G12