High-frequency Statistical Arbitrage Strategy Based on Stationarized Order Flow Imbalance

Abstract

From a high-frequency perspective, the order book data records each market participant’s price expectations for the current underlying assets, and therefore it records the fundamental mechanism of pricing. In this paper, we propose a stationarized indicator based on the classical indicator – Order Flow Imbalance (OFI), and empirically find that there is a stronger linear relationship between the new indicator and the mid-price. Furthermore, we develop a statistical arbitrage strategy based on this new indicator. Compared with the classic indicator, the strategy built with the new indicator has better performance.