Price reaction to order flow ‘news’ in Australian equities

Abstract

The process of ‘price discovery’ is examined here by looking at the unexpected component of the order flow. This ‘news’ in order flow is simply taken to be the innovations in orders, or the deviations that orders take from their one-step ahead expected values. The filtration caused by these innovations is the evolution of the information set, and hence estimation is via recursive least squares. We link these innovations to the ‘unpredictable’ component of price change; that is, the component of price change that is unrelated to predictable effects like bid-ask bounce or transient and liquidity effects. We find that ‘news’ in order flow can take several orders to become fully incorporated into prices. We also find that the effect is strong in both calendar and transaction time, that stocks that trade more frequently incorporate this news more quickly, and that the permanent (i.e., cumulative) effect of an innovation of 10% of average order size is usually a price change of between 0.03% and 0.08%. We also note that the rate at which price incorporates information appears unrelated to the size of the impact of the information. Finally, the effect of news is in general linear in order innovation size and absolute order size.