Title: Models for high-frequency trading volume data
Authors: Eduardo Rossi - University of Pavia (Italy) [presenting]
Paolo Santucci de Magistris - Aarhus University (Denmark)
Abstract: In finance theory prices are often supposed to follow an Ito semimartingale while no explicit assumptions are made on thedynamic evolution of trading volumes. The trading volume of a given security, i.e. the number of shares that change owners in a giveninterval of time, depend on the amount of new information available about the company. The intradaily trading volume series arecharacterized by large movements, so that the realizations are overdispersed. Models of high-frequency trading volume dynamics should becharacterized by large variance of the unconditional distributions. To this purpose, we propose a discrete-time model which has amultiplicative structure with time-varying parameters. We present the estimation and testing procedures and we validate the model throughan extensive empirical analysis.