Factor IC Test

This report conducts a 7-factor IC test on S&P 500 constituents with a 21-trading-day holding period and 39 cross-sectional observations.

FactorIC MeanICIRt-statp-valueIC Positive %Long-Short (Monthly)
EP (Earnings Yield)-0.0624-0.74 ⭐-4.57 ⭐0.00005 ⭐17.9%+3.87%
BP (Book-to-Price)+0.0056+0.08+0.480.635546.2%+0.82%
FCF Yield (Free Cash Flow Yield)-0.0589-1.43 ⭐-8.79 ⭐0.00000 ⭐7.7%+3.04%
ROE (Return on Equity)-0.0414-1.23 ⭐-7.58 ⭐0.00000 ⭐12.8%+1.11%
Mom (12-1M Momentum)+0.2241+2.70 ⭐+16.62 ⭐0.00000 ⭐100.0%-8.07%
Vol (60D Volatility)+0.1320+1.05 ⭐+6.48 ⭐0.00000 ⭐79.5%-7.61%
Size (Market Cap)+0.0479+1.25 ⭐+7.70 ⭐0.00000 ⭐87.2%-2.06%

Key Takeaways:

  • Momentum factor continues to dominate: ICIR 2.70, IC positive rate 100%, positive contribution across all cross-sectional periods. Q5 (high momentum) monthly return +7.95%, Q1 (low momentum) -0.12%, long-short spread -8.07% (negative value means high momentum outperforms low momentum). The momentum effect is extremely significant in the current market environment; trend-following strategies remain effective.
  • EP factor reversal persists: ICIR -0.74, negative IC indicates that high earnings yield stocks have recently outperformed low earnings yield stocks, consistent with a growth-to-value rotation. Q1 monthly return +6.41% far exceeds Q4’s +0.61%.
  • FCF Yield is the strongest value signal: ICIR -1.43, t-stat -8.79, extremely significant. High FCF yield stocks systematically outperform, Q1 monthly avg +5.64% vs Q3’s +0.19%. IC positive rate only 7.7%, meaning nearly every cross-sectional period points in the same direction.
  • BP factor completely ineffective: ICIR only 0.08, p-value 0.64, cannot reject the null hypothesis. Book-to-price has no predictive power for returns; the rising share of intangible assets has diminished the relevance of traditional book value metrics.
  • Volatility factor: ICIR +1.05, positive IC means high-volatility stocks have higher returns, likely a risk premium manifestation. Q5 monthly avg +8.15% vs Q1’s +0.54%.
  • Size factor: ICIR +1.25, IC positive rate 87.2%, large-cap stocks slightly outperform recently but the effect is modest.
  • Quality factor (ROE) reversal: ICIR -1.23, low ROE stocks recently outperform high ROE stocks, possibly reflecting speculative capital flowing into lower-quality names.

Factor IC

Sector Momentum Decomposition

The momentum factor is decomposed by GICS sectors to examine differential momentum effects within each sector. Communication Services, Energy, and Materials are excluded due to fewer than 30 samples.

SectorSampleIC MeanICIRIC Positive %Q5 Monthly ReturnQ1 Monthly ReturnLong-Short (Monthly)
Consumer Discretionary48+0.2522+2.11100%+2.15%-4.05%-6.21%
Consumer Staples36+0.2088+1.3487.2%+0.02%-4.65%-4.67%
Financials76+0.3094+2.63100%+7.80%+0.61%-7.19%
Health Care59-0.1280-1.2115.4%-2.66%+4.84%+7.51%
Industrials79+0.2659+2.88100%+5.51%-1.63%-7.14%
Information Technology73+0.5240+2.58100%+27.19%-1.35%-28.54%
Real Estate31+0.1242+0.8474.4%+4.95%+1.12%-3.83%
Utilities31+0.1993+1.0176.9%-1.66%-2.29%-0.63%

Sector Momentum Highlights:

  • Information Technology momentum extremely strong: ICIR 2.58, IC mean 0.524, the highest across all sectors. Q5 (high momentum) monthly return reaches +27.19%, Q1 (low momentum) only -1.35%, long-short spread -28.54%. The winner-takes-all pattern within tech is extremely pronounced; momentum strategies yield the most extreme returns in this sector.
  • Industrials momentum strongest: ICIR 2.88, the highest across all sectors. IC positive rate 100%, Q5 monthly avg +5.51% vs Q1’s -1.63%, momentum effect is uniform and persistent.
  • Financials momentum robust: ICIR 2.63, IC positive rate 100%. Q5 monthly avg +7.80%, momentum strategies perform strongly in the financial sector.
  • Health Care momentum reversal: ICIR -1.21, negative IC, the only sector with momentum reversal. Q1 (low momentum) monthly avg +4.84% outperforms Q5 (high momentum) at -2.66%. Previously sold-off stocks are showing significant mean reversion; sector rotation signals warrant attention.
  • Consumer Discretionary momentum effective: ICIR 2.11, IC positive rate 100%, Q5 outperforms Q1 by approximately 6.21%.
  • Consumer Staples momentum weakens: ICIR 1.34, but Q5 monthly avg only +0.02%, insufficient differentiation within the sector.
  • Utilities momentum weak: ICIR 1.01, long-short spread only -0.63%, high homogeneity within the sector, momentum effect is minimal.
  • Real Estate momentum moderate: ICIR 0.84, IC positive rate 74.4%, momentum has some effect but is not strong.

Sector Momentum

Summary

This factor test shows the market continues to exhibit a strong momentum-driven characteristic, consistent with the recent pattern:

  1. Momentum factor effective market-wide: ICIR 2.70, IC positive rate 100%, the strongest factor currently; trend strategies remain effective.
  2. Value factors diverge: EP (ICIR -0.74) and FCF Yield (ICIR -1.43) show significant reversal effects, with high-valuation stocks facing recent pullback pressure; BP is ineffective, as the intangible-asset-driven economic structure renders traditional book value metrics meaningless.
  3. Tech sector extreme dispersion: Information Technology ICIR 2.58, Q5 monthly return +27.19%, a winner-takes-all pattern within the sector.
  4. Health Care counter-trend reversal: The only sector with momentum reversal (ICIR -1.21); previously weak stocks are bouncing, requiring attention to sector rotation signals.
  5. Quality factor (ROE) reversal: ICIR -1.23, low ROE stocks recently outperform high ROE stocks; speculative capital may be flowing into lower-quality names.

Overall assessment: The market is in a momentum-dominated strong trend phase. Among value factors, FCF Yield and EP continue to provide effective contrarian signals, while BP has lost relevance. At the sector level, particular attention should be paid to the extreme momentum in tech and the reversal signals in Health Care. All factors are within normal volatility ranges with no anomaly alerts triggered. The TTM financial panel is in a mid-quarter quiet period with no changes in factor data.