Persistent Lead-Lag Networks Predict FX Order Flow and Average Prices

Summary

The document introduces a method for inferring lead-lag networks from the actions of individual agents, then applies it to trader-resolved foreign exchange data. It presents network persistence as an explanation for why one trader’s activity can help predict later order flow from other traders. The analysis also considers whether past prices shape traders’ actions and thereby make changes in the average price they pay predictable.

Random forest models are used to assess hourly predictions for retail investors. The reported findings indicate strong predictability for both order-flow direction and the direction of average transaction prices. The authors suggest these patterns could matter to brokers and order-matching systems, and interpret trader interactions as one reason activity may arise endogenously. The summary does not provide sample details, model settings, numerical performance measures, or evidence beyond the stated results, so the strength and generality of the findings cannot be independently assessed here.

Key ideas

• Lead-lag networks can represent relationships among individual traders’ actions.
• The inferred networks are reported to persist over time, supporting order-flow prediction.
• Past-price-dependent trading may make the direction of traders’ average paid prices predictable.
• Random forests show reported hourly predictability for retail order-flow and average transaction price directions.
• The authors connect trader interactions with the endogenous emergence of market activity.

Tags

# Statistically validated lead-lag networks and inventory prediction in the foreign exchange market


# Statistically validated lead-lag networks and inventory prediction in the foreign exchange market









We introduce a method to infer lead-lag networks of agents' actions in complex systems. These networks open the way to both microscopic and macroscopic states prediction in such systems. We apply this method to trader-resolved data in the foreign exchange market. We show that these networks are remarkably persistent, which explains why and how order flow prediction is possible from trader-resolved data. In addition, if traders' actions depend on past prices, the evolution of the average price paid by traders may also be predictable. Using random forests, we verify that the predictability of both the sign of order flow and the direction of average transaction price is strong for retail investors at an hourly time scale, which is of great relevance to brokers and order matching engines. Finally, we argue that the existence of trader lead-lag networks explains in a self-referential way why a given trader becomes active, which is in line with the fact that most trading activity has an endogenous origin.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.