Building a Liquidity Premium Factor from Order Book Data

Summary

The note describes a stock-selection factor derived from buy-side order book data. It simulates executing a specified trade amount against displayed orders, adds interpolated virtual orders, and compares the resulting market value with the value implied by trading at an average price. The relative difference is aggregated over a 21-day window to form the factor.

The reported analysis says the factor is negatively related to traditional liquidity measures and has a 72-day half-life, with information decay fastest in the first 30 days. It reports average IC near 7% and IC_IR near 0.5 across parameter settings. Stock-selection and long-short results are also summarized before and after controlling linearly for market capitalization and turnover. Two variants weight recent market values by recency or by absolute daily returns; the volatility-weighted version has the strongest reported results. The source provides summary statistics but not the underlying paper, detailed test design, costs, or implementation checks, so the figures alone do not establish live tradability.

Key ideas

• The factor compares simulated execution value against value at an average price using buy-side order book data.
• The daily relative difference is accumulated over a 21-day window.
• The note reports negative correlation with traditional liquidity factors and faster information decay early in the period.
• Recency weighting and absolute-return weighting are proposed variants, with the latter reporting stronger performance.

Tags

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