Building a Limit-Order-Book Liquidity Premium Factor for Stock Selection

Summary

The document describes a liquidity premium factor built from buy-side limit order book data. For a chosen trade amount, the method simulates execution against displayed orders, adding interpolated virtual orders as needed, and compares the resulting market value with the value of trading at an average price. The relative difference is aggregated over a rolling 21-day period. The factor is presented as distinct from traditional liquidity measures and as a way to capture changing market conditions.

The reported analysis gives an average information coefficient near 7% and an information ratio around 0.5 across parameter settings. It also compares stock-selection and long-short results before and after removing linear size and turnover effects. Two variants weight recent observations or days with larger absolute price moves more heavily; the volatility-weighted version reports stronger results in the cited comparisons. These are source-reported historical findings, with no full paper, sample details, or implementation discussion included here, so they do not establish out-of-sample performance.

Key ideas

• The factor compares simulated execution value from buy orders with value at an average price.
• The relative difference is aggregated over a rolling 21-day window.
• The reported factor has a roughly 72-day half-life, with information decay fastest early in that period.
• Time weighting emphasizes recent observations, while volatility weighting emphasizes days with larger absolute returns.
• The reported tests show stronger results for volatility weighting, but the excerpt omits detailed sample and testing specifications.

Tags

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