Dynamic limit order placement strategies: survival analysis with a multiple-spell duration model

Abstract

This study investigates the multiple events that occur in the life of each limit order by utilising a survival analysis methodology with a multiple-spell duration model. The estimates suggest that the hazard rates of limit order event transitions are determined by a number of factors and their impacts depend on whether the initial order event is a limit order submission, partial execution or revision. The differences in estimates across initial order events increase as exchange latency reduces in recent years. Using a multiple-spell duration model to examine the full spectrum of events that occur in the life of a limit order is thus shown to be informative and essential in modelling dynamic limit order placement strategies.