Ensemble Properties of High Frequency Data and Intraday Trading Rules

Abstract

We demonstrate that a stochastic model consistent with the scaling properties of financial assets is able to replicate the empirical statistical properties of the S&P 500 high frequency data within a window of three hours in each trading day. This result extends previous findings obtained for EUR/USD exchange rates. We apply the forecast capabilities of the model to implement an explicit trading strategy. Trading signals are model-based and not derived from chartist criteria. In-sample and out-of-sample tests indicate that the model performs better than a benchmark asymmetric GARCH process, and expose the existence of small arbitrage opportunities. We discuss how to improve performances and why the trading strategy is potentially interesting to hedge volatility risk for S&P index-based products. nomalous scaling; Memory; Intraday returns; Intraday strategy.

Keywords: time series ensemble, trading rule, long memory

JEL Classification: C23, C53, C51