BoC 2026-06-04 — Hold at 2.25%

Thesis

1. 1.Latest BoC statement ingested: 2026-03-18. Hawk-dove net -8.4 (net dovish). Constraint-norm 3.60 (>= 3.5 uncertainty-pause threshold). Constraint dimension is the #1 hold discriminator in the backtest — scenario / wide-range / too-early / more-data markers appear ~160x more in holds than cuts.
2. 2.Model score breakdown: CUT=3.0 / HOLD=9.8 / RAISE=0.0. HOLD dominates by 3.3x over CUT; RAISE shut out.
3. 3.Backtest continuation accuracy: HOLD->HOLD 80%, CUT->CUT 86%, RAISE->RAISE 100%. Transition detection from text alone: 0%. The predictor is a continuation CONFIRMER, NOT a standalone transition predictor. Treat as flag-and-confirm input.
4. 4.EHIQ probability 0.633 after continuation-prior haircut: BoC HOLD->HOLD historical base rate is ~71% in this period. Naked confidence 0.766 would overstate the edge over the continuation prior, so we discount.
5. 5.BoJ receipt-discipline motivation: the April 30 2026 BoJ MPM hawkish-dissent signal (Himino/Takata via central-bank-language analyzer; June FXY calls position) was a paid trade for EMJ but was NOT publicly timestamped pre-resolution, so could not be claimed on /predictions. BoC June 4 is the first central-bank-language call where the publicly-timestamped pre-resolution receipt discipline is being honored.
6. 6.Methodology lock — diagnostic input, not standalone trigger: per Phase C OOS FAIL on conviction_sensor (different domain, same discipline), TIQ output is treated as ONE analytical input to Eric's conviction call. Not a method-validated predictor. Pairs with separate macro inputs (CAD employment, CPI, oil, USD-CAD path).
7. 7.5-bank central-bank-language stack status (tonality_iq/central_banks/): BoJ + BoC have committed scored corpora as of 2026-05-22; Fed + BoE + ECB are structurally scaffolded but need scrape+score runs before June meetings. Only BoC June 4 is publishable at this discipline level today.
8. 8.Asymmetric-payoff sizing frame: 1-2% NAV for any associated CAD / USD-CAD / CGB position. Same sizing discipline as the BoJ trade (multi-statement-corpus signal, not single-print).
9. 9.Market probability blocked at free-source level: CME BAX implied probs require Reuters Eikon (~CAD 800/mo) or Bloomberg (~USD 2400/mo). Kalshi has no KXBOC series; Polymarket coverage sporadic. marketProb stays null; EHIQ-vs-market gap analysis is one-sided.
10. 10.Corpus gap (operator-owned): boc_corpus.json ends 2026-03-18; 2026-04-29 BoC HOLD-at-2.25% statement + Governing Council deliberations summary NOT yet ingested. Refresh path: run boc_scraper.py OR hand-paste from bankofcanada.ca/2026/04/fad-press-release-2026-04-29. After ingest, re-run boc_rate_predictor.py; signal likely lifts to ~0.65-0.70 (a 4-hold streak strengthens the continuation prior).
11. 11.Resolution rule: BoC announcement 2026-06-04 at 13:45 UTC (09:45 ET). Hold-at-2.25% = YES at 0.633. 25bp cut to 2.00% = NO. 25bp raise to 2.50% = NO. Out-of-corridor surprise (50bp+) = NO miss + flag for model recalibration.
12. 12.EMJ Capital position discipline: any associated FX / rate position must be disclosed on follow-up commentary. As of 2026-05-22, no current BoC-meeting-derived position is held.