Introduction

Abstract

Today it is customary that every single transaction of a financial asset traded on major financial markets around the world is recorded electronically with detailed information about the time of occurrence, price and volume and other relevant characteristics. Recently, many of these so called high frequency data sets have become available at relatively low cost to academic researchers. Hence, the last fifteen years saw an unprecedented upsurge in both theoretical and empirical work related to the analysis of market micro structure issues using transaction data sets that are steadily increasing in size.¹ This upsurge went hand in hand with an equally unprecedented progress in computer technology, that made empirical analysis of huge data sets with ordinary desktop computers possible. It seems only natural then, that this innovation in both, the quality of data available for research and the development of the relevant economic theory was accompanied by the introduction of econometric methods which were tailor-made for the analysis of many policy issues related to market microstructure, e.g. how trading mechanisms should be structured in order to enhance market viability, or how parallel trading of a security on several markets affects prices.