Andrew Bailey — early-evidence cross-asset signal on sterling.

Per-speaker drill-down. 22 BoE Monetary Policy Report press conferences parsed since 2020-11 – 2026-02. ≈ 192K tokens. The same Powell-built lexicon was applied to Bailey's transcripts; the strongest signal is on GBP/USD direction, with 17 meetings scored walk-forward. CIs are wide given the small sample — we treat this as early-evidence, not validated. Live pre-registered call for May 8, 2026 is on the main page.

Pressers parsed
22
Meetings scored
17
GBP/USD edge
+39.3pp
GBP/USD Sharpe
0.92

Cross-asset edges (walk-forward, n=17)

Same Ridge model architecture as Powell + Lagarde, applied to BoE-relevant assets. GBP/USD is the strongest target — sterling moves more decisively on Bailey verbal cues than the equity indices do. FTSE 100 (large-cap, multinational) is essentially null; FTSE 250 (more domestic) shows a positive direction edge but the Pearson ρ is negative, which means the magnitude is poorly fit even though direction beats momentum. Disclosed honestly.

AssetClassEdge over momentumSharpe (ann)95% CIρStatus
GBP/USDFX+39.3pp0.92[-0.03, 2.05]+0.28early
FTSE 250Equities+27.2pp0.31[-0.81, 1.11]-0.20early
FTSE 100Equities+2.9pp0.13[-1.07, 1.04]-0.12early

Last 10 walk-forward predictions — GBP/USD

Real walk-forward predictions on Bailey's strongest target. 8 of 10 directional hits. The two recent misses (May 2025, Feb 2026) were both cases where the model leaned modestly long sterling and Bailey delivered a more dovish tone than the lexicon features picked up. The May 8, 2026 live call (-0.04% predicted, basically flat) is on the main page and resolves Friday.

Meeting datePredictedActualPredicted dirActual dirHit?
2023-11-02+0.632%+0.247%upup✓ hit
2024-02-01-1.254%-0.134%downdown✓ hit
2024-05-09-0.266%-0.065%downdown✓ hit
2024-08-01+0.122%+0.126%upup✓ hit
2024-11-07-0.115%-1.229%downdown✓ hit
2025-02-06+0.355%+0.208%upup✓ hit
2025-05-08+0.212%-0.463%updown✗ miss
2025-08-07+0.515%+0.440%upup✓ hit
2025-11-06+0.546%+0.237%upup✓ hit
2026-02-05+0.278%-0.386%updown✗ miss

Bailey only meets quarterly with this MPR cadence, so the n grows slowly. CIs will tighten gradually; we are explicitly not claiming this is validated yet, even though the point estimate is strong (Sharpe 0.92, lower CI just below zero).