Weakly inefficient markets: Stability of high-frequency trading strategies

Abstract

Market participants who capitalize on high-frequency price dynamics and rely on automated trading are responsible, along with market makers, for the observed level of market efficiency. The remaining inefficiency is usually measured as the ratio of expected P&L, derived from the price signals, to its standard deviation. Such signals are also termed alpha in market slang. Signals and their volatility depend on time in a different manner, leading to temporal diversification and rise of multi-step strategies. It is shown that the coexistence of small market inefficiencies, multi-step strategies, and market impact lead to price randomization. In other words, high-frequency strategies redefine prices in their attempt to amplify weak price signals, and make markets more effective. In this paper we identify and explore discrete and continuous strategies. We further demonstrate that strategies within the domain of weak inefficiency are stable when incorporated into regular risk-return framework. In the presence of market impact we show how an efficiency edge propagates towards smaller time scales.