Reconstructing Order Flow from Synchronized Trades and Quotes

Summary

This study examines tick-by-tick trade and quote data for stocks on the Paris, London, and Frankfurt exchanges over a five-year span. It develops an algorithm to synchronize the two data sources and evaluates how reconstruction choices affect the order flow inferred from aggregated data. The authors also analyze trade signatures and refine the Lee-Ready procedure, including guidance on selecting its time lag.

The algorithm’s performance varies with the exchange and time period, and the analysis uses those variations to identify technical changes in the studied markets and to assess database quality. The authors show that reconstructed order flows influence the quantitative models later calibrated on them. Their findings are consistent with financial reasoning and an illustrative Poisson order-flow model. The results concern the listed exchanges and database; the document does not provide enough detail to assess generalization to other venues or data sources.

Key ideas

• Trade and quote timestamps must be synchronized to reconstruct order flow from tick data.
• The proposed synchronization algorithm’s performance can reveal changes in exchange systems and aggregated data quality.
• Choices made during order-flow reconstruction affect models calibrated on the resulting data.
• The study refines Lee-Ready trade classification by examining how to choose the time lag.
• Findings are compared with financial reasoning and an illustrative Poisson model.

Tags

# Reconstruction of Order Flows using Aggregated Data


# Reconstruction of Order Flows using Aggregated Data









In this work we investigate tick-by-tick data provided by the TRTH database for several stocks on three different exchanges (Paris - Euronext, London and Frankfurt - Deutsche Börse) and on a 5-year span. We use a simple algorithm that helps the synchronization of the trades and quotes data sources, providing enhancements to the basic procedure that, depending on the time period and the exchange, are shown to be significant. We show that the analysis of the performance of this algorithm turns out to be a a forensic tool assessing the quality of the aggregated database: we are able to track through the data some significant technical changes that occurred on the studied exchanges. We also illustrate the fact that the choices made when reconstructing order flows have consequences on the quantitative models that are calibrated afterwards on such data. Our study also provides elements on the trade signature, and we are able to give a more refined look at the standard Lee-Ready procedure, giving new elements on the way optimal lags should be chosen when using this method. The findings are in line with both financial reasoning and the analysis of an illustrative Poisson model of the order flow.

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.