Modelling Adverse Selection on Electronic Order-Driven Markets

The vast majority of models that decompose the bid/ask spread assume the quote-driven, specialist structure of the NYSE. This paper critically evaluates these models to construct a model specific for an electronic order-driven exchange. The model not only captures adverse selection and the impact of order flows on price discovery but it includes the imbalance of supply and demand inherent in the public limit order book. With this new model we investigate the change to anonymity on the Australian Securities Exchange (ASX). Following the change to anonymity, both adverse selection and the demand/supply imbalance have an increased impact on prices while order flow has a decreased influence, suggesting the change to anonymity has improved market efficiency. The model also uncovers a change in traders’ behavior once their fear of front-running is reduced. We show that the model is stable and robust across high liquidity stocks as well as stocks with as few as 5 trades per day.