Feedback Trading

Jón Dańıelsson; Ryan Love;

RELATED

Related Papers     Patents    Grants    Venues    Experts    Cite    Share    Save

ABSTRACT

Order flow has been found to carry information to the market. When assessing how informative order flow is, the VAR methodology is typically employed, using impulse response functions. However, in such analyses, the direction of causality runs explicitly from order flow to asset return. If data are sampled at anything other than at the highest frequencies then any feedback trading may well appear contemporaneous; trading in period t depends on the asset return in that interval. The implications of contemporaneous feedback trading are examined in the spot USD/EUR currency market and we find that when data are sampled at the one and five minute frequency, such trading strategies cause the price impact of order flow to be significantly larger than when feedback trading is ruled out. ∗Both authors; Financial Markets Group, London School of Economics and Political Science. Thanks to the Bank of England for providing the transactions data used in this study. Financial assistance from the Economic and Social Research Council is gratefully acknowledged (grant number L138251036). The paper has also benefitted from comments by Greg Connor, Charles Goodhart and Andrew Patton. All errors are the joint responsibility of the authors. For correspondence: Jón Dańıelsson, Financial Markets Group, London School of Economics, Houghton Street, London, UK. WC2A 2AE; tel: ++44-(0)207 955 6056; email: J.Danielsson@lse.ac.uk. This paper is also available at www.riskresearch.org. Order flow, one way buying or selling pressure, has a contemporaneous impact on prices, while at the very highest frequencies, the converse is not true. However, when aggregated over time, order flow and prices can be expected to impact on each other simultaneously, a phenomenon we call contemporaneous feedback trading. The existence of contemporaneous feedback trading implies that such models cannot be estimated with traditional techniques. This is unfortunate, since empirical models with feedback trading can be expected to give firmer support for theoretical models than models that ignore such trading strategies. In this paper we argue that feedback trading is an inevitable consequence of time aggregation of order flow models, and propose an estimator of such models by using instrumental variable techniques. We find that the price impact of trades is much stronger when feedback trading is incorporated, further supporting market microstructure theories generally, and the validity of the order model specifically. It is well known in the theoretical microstructure literature that order flow conveys private information to the market as a whole. In this way, information is aggregated via the trading process implying that order flow has permanent effects on prices. See Kyle (1985), Glosten and Milgrom (1985), Easley and O’Hara (1987), and Evans and Lyons (2002b) for examples of such models. These models imply that if trades carry private information then the informativeness of trades can be accessed by their price impact. Furthermore, in empirical models when data are employed at the very highest frequencies, order flow, by definition, can only be affected by the lags of price changes. However, when data are aggregated, transactions and order entry are simultaneous, frustrating empirical investigations. The existence and profitability of feedback trading strategies has been considered in a number of papers. De Long, Shleifer, Summers, and Waldmann (1990) build a model of feedback trading with rational speculators who will buy (sell) when the price rises (falls). The profitability of a number of feedback trading strategies in stock markets is considered in Jegadeesh and Titman (1993) and the existence of high frequency positive feedback trading in the US treasury market is documented in Cohen and

 Paraphrase     Questionfy     Wikify     Simplify     Summarize