Building Stock Selection Factors from Intraday Price and Volume Data
Summary
This research report develops equity-selection factors from four types of intraday information: price behavior, price-volume relationships, pre-open auction data, and selected trading periods such as the close or high-volume intervals. It evaluates 46 candidate factors for weekly stock selection. The report argues that high-frequency inputs may offer less crowded signals, greater variety, and more independent observations than lower-frequency factors.
Factor information coefficient and long-short returns are used to identify 12 candidates, including realized skewness, intraday returns, closing-period volume share, price-volume correlations, auction-derived returns, and Amihud illiquidity. The selected factors show long-short excess returns in the study, but for all except Amihud illiquidity, positive-alpha returns were notably weaker than negative-alpha returns. After transaction costs, benchmark-relative long-only results fell substantially; most reported annualized excess returns were between negative one and nine percent. The findings depend on the historical market and implementation assumptions, and the report warns that changing market structure or increased participation could weaken the signals.
Key ideas
- The report constructs factors from intraday prices, price-volume relations, opening-auction data, and selected trading windows.
- It evaluates 46 factors for weekly stock selection using information coefficients and long-short returns.
- Twelve candidates are retained, spanning realized skewness, intraday returns, volume patterns, auction moves, and illiquidity.
- Most factors had weaker positive-alpha returns than negative-alpha returns, and trading costs reduced long-only excess returns.
- Historical factor performance may not persist if market structure or participation changes.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.