Building Monthly Equity Factors from High-Frequency Order Flow
Summary
The document describes turning intraday order-book and trade data into monthly equity-selection factors. It introduces volume order imbalance (VOI), order imbalance ratio (OIR), and market price bias (MPB). MPB classifies trades by whether buyers or sellers initiated them, then uses volume and turnover to estimate average prices between observations. The factors are evaluated in single-factor backtests.
The reported pattern changes with sampling frequency: VOI and OIR relate positively to returns at high frequency but reverse direction after aggregation, while MPB is negatively related to returns. The authors suggest short-term retail chasing and possible large-trader activity can make volume signals misleading in the short run, followed by price reversals. MPB is reported as the strongest factor, with a mean IC of -5.23%, annualized long-short return of 21.24%, and Sharpe ratio of 2.68. The document provides no sample period, universe, transaction-cost assumptions, or detailed validation, so these figures should not be treated as evidence of out-of-sample performance.
Key ideas
- VOI measures the difference between submitted buy and sell volume, while OIR captures order imbalance in ratio form.
- MPB classifies trades by initiating side and uses volume and turnover to estimate average transaction prices.
- VOI and OIR show positive return relationships at high frequency but reverse after conversion to monthly factors.
- The authors attribute the reversal to short-term chasing or large-trader activity followed by price reversals.
- MPB is reported to have the strongest single-factor backtest results, though validation details are absent.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.