Order-Data Equity Alpha Factors

Article BigQuant

Summary

This research summary describes equity signals built from timestamped order submissions and cancellations, which can reveal more of the trading process than completed trades alone. It focuses mainly on Shenzhen exchange data because that venue had a longer history of order-level records; differences in exchange data formats affect how orders and cancellations are represented. The proposed factors include a pre-open cancellation ratio, the relative size of buy and sell orders during the first half-hour, and measures of how submitted prices differ from recent trades.

The summary reports that a higher pre-open cancellation ratio was associated with lower average future returns. It also says a buy-order-based revision of the APB factor performed better than the original on RankIC, long-short returns, and portfolio grouping, including during a later period when the original factor declined. Relative price disagreement reportedly had weaker stock selection results than absolute disagreement, but lower correlation with other factors. These are summarized findings; the document supplies no underlying sample details, construction formulas, transaction-cost analysis, or full validation, so the reported results cannot establish robustness on their own.

Key ideas

Order-level submissions and cancellations can provide information about trading behavior beyond completed transactions.
The study mainly uses Shenzhen data and notes that exchange-specific data formats differ.
A higher pre-open cancellation ratio is reported to correspond with lower average future returns.
The relative sizes of early-session buy and sell orders are proposed as a signal of large-investor activity.
A revised APB factor based on buy orders is reported to outperform the original on several portfolio measures.
The available summary omits detailed tests, costs, and full factor definitions.

Equity Alpha Factors from Order-Level Data

Summary

Key ideas

Tags