Optimized Execution via Order Book Simulation

Abstract

In modern financial markets, the importance of trade execution cannot be overestimated. Due to intense competition for profit opportunities, trading costs can represent a significant portion of overall returns and thus must be taken into account both when a specific trade is being executed and when a general investment strategy is being designed. In this paper, we empirically demonstrate that by combining market and limit orders we are able to obtain a superior execution price than by using market orders only. More precisely, our primary contributions are: - The development of an order book reconstruction method for investigating execution strategies. - The introduction of a simple but realistic family of execution strategies, in which the goal is to buy or sell V shares within time T, and the only choice available is the limit order price. - An extensive experimental study of this family of strategies using Instinet order book data across several stocks, which clearly demonstrates significant opportunities for optimization within our family of strategies. - The introduction of name-specific optimal risk-return frontiers, which demonstrate the inherent trade-off between price improvement and the risk of non-execution. - A detailed examination of a number of natural factors that can influence the optimization, including the order volume V, execution window T, time of day, and order book depth. We also discuss how our findings in the simple framework studied here suggest richer execution strategies.