Residual momentum is a momentum signal built from the part of a stock's past return that is not explained by the market and common factors such as size and value. It aims to capture stock-specific trends, instead of the factor trends that conventional momentum also picks up.
How it is built
- Regress each stock's monthly returns on market and factor returns over the past 36 months. See Fama-French factor models.
- Keep the residuals, the part of each month's return the factors did not explain.
- Add up the residuals for the past 12 months, skipping the latest month, and often scale by their volatility.
- Rank stocks on that score. See momentum investing.
An illustrative example
| Stock A | Stock B | |
|---|---|---|
| 12-month return | 30% | 20% |
| Return explained by market and factors | 25% | 5% |
| Residual return | 5% | 15% |
Ordinary momentum ranks A first, but much of its gain came from the market and factors it is exposed to. Residual momentum ranks B first, because its gain was specific to the company. The numbers are made up.
Why use it
- Blitz, Huij and Martens (2011) found it had a similar average return to conventional momentum with lower volatility and smaller drawdowns.
- Conventional momentum swings between factor exposures, which is a source of momentum crashes. Residual momentum is less exposed to them.
- It has lower turnover than short-horizon signals.
Cautions
- Results depend on the factor model and window, so the signal can be sensitive to choices. See overfitting.
- Regression estimates are noisy for short histories.
Frequently asked questions
What is residual momentum?
It is momentum measured on the portion of a stock's returns not explained by market and factor returns, instead of on its total return.
How is residual momentum different from normal momentum?
Normal momentum ranks on total past return, which includes market and factor effects. Residual momentum removes those and ranks on stock-specific returns.
Does residual momentum reduce crash risk?
Research suggests it has lower volatility and smaller drawdowns, but it does not remove risk.
Why skip the most recent month?
To avoid the short-term reversal effect. See short-term reversal.