Gate: flat when BAA-AAA spread > trailing-252d median (lag-1). IS 2020-2023, OOS 2023+ (earnings depth). Bootstrap n=5000.
| Variant | Segment | n | Ret (cum) | Sharpe(252) | MaxDD | p | CorrMkt | CorrEqMom | ResidSharpe |
|---|---|---|---|---|---|---|---|---|---|
| PEAD top-quintile ungated (baseline, hold60) | IS_2020-2023 | 756 | +16.27% | 0.32 | -51.54% | 0.296 | 0.86 | 0.70 | -0.40 |
| PEAD top-quintile ungated (baseline, hold60) | OOS_2023+ | 905 | +210.57% | 1.97 | -20.66% | 0.000 | 0.94 | 0.75 | 1.78 |
| PEAD x VIX gate (replication) | IS_2020-2023 | 756 | +14.56% | 0.30 | -51.54% | 0.302 | 0.86 | 0.70 | -0.43 |
| PEAD x VIX gate (replication) | OOS_2023+ | 905 | +168.57% | 1.81 | -19.16% | 0.000 | 0.96 | 0.74 | 1.48 |
| PEAD x credit gate (spread <= 252d median) | IS_2020-2023 | 756 | +37.63% | 0.91 | -17.80% | 0.057 | 0.35 | 0.37 | 0.75 |
| PEAD x credit gate (spread <= 252d median) | OOS_2023+ | 905 | +122.78% | 1.61 | -19.98% | 0.002 | 0.79 | 0.71 | 0.78 |
| EqMom baseline (12-1 top decile, hold21, net10bps) | IS | 2363 | +195.59% | 0.62 | -54.71% | 0.028 | 0.88 | 1.00 | -0.06 |
| EqMom baseline (12-1 top decile, hold21, net10bps) | OOS | 2667 | +1004.91% | 1.06 | -38.55% | 0.001 | 0.77 | 1.00 | 0.44 |
| EqMom x credit gate (sanity replication) | IS | 2363 | +131.75% | 0.73 | -31.33% | 0.012 | 0.43 | 0.59 | 0.46 |
| EqMom x credit gate (sanity replication) | OOS | 2667 | +423.60% | 0.96 | -26.39% | 0.001 | 0.40 | 0.74 | 0.61 |
| PEAD x double gate (VIX calm AND credit calm) | IS_2020-2023 | 756 | +22.90% | 0.83 | -7.47% | 0.073 | 0.26 | 0.27 | 0.70 |
| PEAD x double gate (VIX calm AND credit calm) | OOS_2023+ | 905 | +44.77% | 1.14 | -10.29% | 0.014 | 0.52 | 0.40 | 0.46 |
Verdicts (OOS only, p<0.05 and resid>0 and corrEqMom<0.6 required for PASS)
- PEAD top-quintile ungated (baseline, hold60) (OOS_2023+): -- (p=0.000, Sharpe 1.97) -- reference.
- PEAD x VIX gate (replication) (OOS_2023+): REJECT (p=0.000, Sharpe 1.81, corrEqMom 0.74, resid 1.48)
- PEAD x credit gate (spread <= 252d median) (OOS_2023+): REJECT (p=0.002, Sharpe 1.61, corrEqMom 0.71, resid 0.78)
- EqMom baseline (12-1 top decile, hold21, net10bps) (OOS): -- (p=0.001, Sharpe 1.06) -- reference.
- EqMom x credit gate (sanity replication) (OOS): REJECT (p=0.001, Sharpe 0.96, corrEqMom 0.74, resid 0.61)
- PEAD x double gate (VIX calm AND credit calm) (OOS_2023+): MARGINAL (p=0.014, Sharpe 1.14 vs baseline 1.97, corrEqMom 0.40, resid 0.46) -- halved maxDD (-10.29% vs -20.66%) at cost of Sharpe; strong diversification (corr 0.40 vs 0.75). Same profile as round-15 VIX gate.