What Is Residualisation?

Residualisation removes the part of a variable explained by one or more control variables. It is commonly done by fitting a regression to the controls and using the residual: the component left unexplained by those controls. In quantitative research, it can be applied to a feature, a model prediction, a target, or portfolio returns.

For example, a cross-sectional signal can be residualised against market beta, volatility, size, momentum, or sector exposures before its information coefficient (IC) is remeasured. This tests whether the apparent signal is largely a known exposure. Residualisation depends on the selected controls and their point-in-time estimates; it does not prove causal independence or guarantee that a residualised strategy is tradeable.

See also

• Factor neutralisation

• Feature ablation

• Cross-sectional analysis

Research and literature

• Eugene F. Fama and Kenneth R. French, Common Risk Factors in the Returns on Stocks and Bonds, Journal of Financial Economics 33(1), 1993.