Active-Buying Factors for Industry Rotation

Article BigQuant

Summary

This report studies whether active-buying measures derived from high-frequency trading data can help rank industries for rotation. It constructs buy-share, buy-intensity, net-buy-share, and net-buy-intensity factors using volume and turnover data from the open, close, or full trading day. The analysis compares equal-weighted and market-cap-weighted industry aggregates, weekly versus monthly calculation windows, holding periods, and time-series standardization.

The reported results favor weekly factor windows with a two-week holding period. Full-day factors show significant information coefficients against two-week industry returns, while monthly calculated equal-weighted factors perform poorly across holding periods. Standardization can improve ICs but may reduce long-short returns; for market-cap-weighted factors, the report favors the original full-day buy factors. Example portfolios report a 7.66% annualized long excess return for a standardized equal-weighted net-buy-intensity factor, and a 5.65% long excess return plus a -10.13% short excess return for an unstandardized market-cap-weighted buy-intensity factor. Results vary by year: full-day factors generally fare poorly in 2015, and closing-period factors later lose effectiveness. The report flags model specification and systemic market risk; its summary provides no detailed methodology or independent validation.

Key ideas

The study builds active-buying factors from volume and turnover data for different parts of the trading day.
Weekly factor windows paired with a two-week holding period are reported as the strongest setup.
Full-day factors show significant information coefficients against two-week industry returns.
Time-series standardization may improve IC while weakening portfolio returns.
Market-cap-weighted factors show stronger short-side effects, but results vary across years.

Intraday Active-Buying Factors for Industry Rotation

Summary

Key ideas

Tags