Residual momentum
MomentumAdvanced
Residual momentum ranks stocks on the recent residual from a factor model, not on total return. The point, developed in work such as Blitz, Huij, and Martens, is to remove the part of the past return that came from market, size, value, or similar exposures. The portfolio then buys stocks with strong residual momentum and shorts stocks with weak residual momentum.
Not investment advice. Nothing here is a recommendation to buy or sell anything. Backtests and historical studies are not guarantees, and a rule that worked in one sample can fail in the next. Trading can lose money. Links are for education, not endorsements. Do your own research.
- Rules
- Each month, estimate a simple factor regression on a trailing window and take the residual return. Rank on the cumulative residual, often skipping the most recent month. Go long the top decile and short the bottom decile. Re-estimate the model on past data only. Neutralize the book to the same factors if the implementation says so.
- Where AI fits
- AI can fit the residual model on a rolling window that ends before the holding month and can refuse to use restated fundamental data. It can show whether the residual book is actually less exposed to the market than a total-return momentum book. Changing the factor list after seeing performance is fitting.
- Markets
- Broad individual-stock universes with a stable factor model and borrow for the short leg.
- Timeframes
- Monthly.
- Risks
- The residual is only as good as the model. A missing factor leaves that bet inside the 'residual.' Short residuals can be unborrowable. Estimation error in a short window creates noise rankings. This is not a cleaner RSI.