The Information Coefficient (IC) measures a signal's predictive accuracy as the cross-sectional Pearson or Spearman rank correlation between the signal's forecast scores and the subsequent realized returns across a universe of assets.
IC ranges from −1 to +1. An IC of 0 indicates no predictive skill. A negative IC means the signal consistently predicts in the wrong direction. In cross-sectional equity research, IC values of 0.02–0.07 are considered meaningful — small correlations but, when applied over many independent bets at high breadth, capable of generating consistent active returns.
IC in the Fundamental Law
IC is the critical skill input to the Fundamental Law of Active Management (Grinold & Kahn): IR ≈ IC × TC × √N. A strategy's Information Ratio scales linearly with IC, so even a modest improvement in signal quality compounds significantly over breadth.
Key uses
- Comparing and ranking competing alpha signals on a common scale
- Detecting signal decay via rolling IC at increasing forward-return horizons
- IC-weighted aggregation — weighting signals in a composite forecast proportional to their recent IC
Full guide: Information Coefficient — how to measure quant signal skill