DETECTION: TIGHTENING
Statistical Methods

Universal Z-Score Standardisation

Formula

Z(t) = [X(t) − μ(rolling N)] / σ(rolling N)
  • X(t) – Raw metric value at time t
  • μ(rolling N) – Rolling arithmetic mean computed over N periods (default: 25 years / 300 months)
  • σ(rolling N) – Rolling standard deviation over same window
  • N = 25 years – Chosen to span full debt supercycle and monetary regime cycles post-1971

Why It Matters

The 25-year window is a deliberate methodological choice. Shorter windows (1–5 years) are too sensitive to the most recent cycle and will not detect regime shifts accurately. Longer windows (40+ years) include structurally different monetary regimes (Bretton Woods, Volcker shock) that distort the baseline. 25 years — roughly three full business cycles — captures enough history to render the Z-Score regime-invariant while remaining relevant to the current monetary framework.

Institutional Use

The Federal Reserve H.4.1 analytical teams, BIS Quarterly Review researchers, and Goldman Sachs Financial Conditions Monitor all use equivalent rolling Z-Score normalisation for their liquidity and credit data series. The methodology is also used by the IMF in their "Early Warning Exercise" for financial stability surveillance.

How to Read It

Extreme Stress (≤ −2σ)Z < −2.0
Moderate Stress (−2σ to −1σ)−2.0 ≤ Z < −1.0
Normal Range (±1σ)−1.0 ≤ Z ≤ 1.0
Elevated (1σ to 2σ)1.0 < Z ≤ 2.0
Extreme Expansion (> 2σ)Z > 2.0
Internal methodology — data inputs vary by metricAll metric methodologies →
Macro Strategy Division

Get the Weekly Regime Digest

De-dollarization, liquidity & energy signals — synthesized from 15+ official sources. No noise.

Email cadence
Stored in Supabase · insert-only RLS · double opt-in
Terminal Active: Capture Mode