Energy Regime Index

Pillar 1 — Energy. A 3-component composite grounded in Kilian (2009, AER): supply/geopolitical stress, demand pulse, and precautionary uncertainty. Updated .

Dominant driver:

( contribution). The composite reads because .

Composite stress

Line = 3-month trailing MA of the composite stress z. Rail beneath shows the (un-smoothed) regime classification. Higher = more energy-market stress. Shaded bands = NBER-dated US recessions.

Each line is one component's expanding-window z-score. Copper/Gold YoY is plotted with sign flipped so all three lines are oriented such that higher = more stress — the composite is their simple average. When lines diverge, the regime is channel-specific; when they co-move, the regime is broad-based.

Gulf fiscal breakevens — the OPEC+ asymmetry

Brent monthly average against IMF Article IV fiscal breakevens for Saudi Arabia (: $) and the UAE (: $). The structural read: through 2025, Brent ran in a band that was sub-Saudi-breakeven and well-above UAE-breakeven for most months — which is the structural reason UAE is the OPEC+ defector candidate, not Saudi. UAE clears its budget below $50; Saudi needs $90+. The April 2026 spike to $138 followed by the crash back to $100 is the Iran flare and its de-escalation, not the Gulf-budget signal. Sources: .

Natural gas — same molecule, four prices

The argument is the spread, not the bars. Same molecule (CH₄), four prices, one week. Spread: per MMBtu between Waha and JKM. Henry Hub at as of (live FRED DHHNGSP). Waha / TTF / JKM as of (paid feeds — NGI / ICE Endex / S&P Platts — refreshed manually).

Uranium — production geography (the locality lens)

Uranium is the cleanest specific-access commodity in the Energy pillar — only 16 countries produce it commercially, and the geography map onto the multi-polar fracture line directly. Western utilities cannot build their fuel cycle from one-side-only sources for long.

Friendly jurisdictions

Canada, Australia, US, Namibia

Mixed

Kazakhstan, Uzbekistan, S Africa

Unfriendly

Russia, China, post-coup Niger

tU world production in . Kazakhstan alone is % — the single largest source by far, and politically mixed: not Western-allied, not actively unfriendly. The Niger 2023 coup pulled ~5% of supply into uncertain status almost overnight, and Russia's Rosatom is % of pure mining (much higher when you include enrichment). The US is not in the top-15 at the mining stage. Western utilities run on imports, full stop. Source: World Nuclear Association.

Methodology

The Energy Regime Index operationalizes the Kilian (2009, AER) decomposition of oil-market shocks into three distinct channels — supply, aggregate demand, and precautionary (forward-looking) demand. Each component is a free-data proxy for one channel:

Component	Proxy	Channel
Brent – WTI spread	`BZ=F` − `CL=F` monthly close	Oil-specific supply / geopolitical stress — wider spreads reflect Mideast and logistics risk premia over US shale
Copper / Gold YoY	`HG=F` / `GC=F` monthly ratio, 12-month % change	Aggregate demand — industrial metal vs. safe haven isolates the demand component from monetary effects
OVX	`^OVX` monthly close	Precautionary uncertainty — the options market's ex-ante distribution of crude prices

Z-score convention. Each component is standardized with an expanding-window mean/std (minimum 24 months). The z at date $T$ uses only data through $T$ — same discipline as the BCI scorer. No look-ahead contamination.

Sign alignment. All three components are oriented so higher z = more stress: Brent–WTI wider (+), Copper/Gold YoY weaker (–, sign flipped), OVX higher (+).

Composite: $z^{\text{stress}}_t = \tfrac{1}{3}(z^{BW}_t - z^{CG}_t + z^{OVX}_t)$ — equal-weight. (Equal weights are a deliberate choice against overfitting factor loadings on a 3-component sample; when a larger data feed lands, a DFM-style loading estimation becomes viable.)

Phase classification.

$z < -0.5$ → CALM
$-0.5 \le z < 0.5$ → NORMAL
$0.5 \le z < 1.5$ → ELEVATED
$z \ge 1.5$ → STRESS

What's in v1, what's in v2

The originally scoped 7-component composite (adding TTF natural gas, Baltic Dry Index, Brent term structure, oil-producer GPR country detail, and Urals/WCS regional discounts) is deferred. Reason: the missing components are subscription-gated (Platts, ICE) or involve more complex data plumbing (full futures curve, GPR country microdata) — shipping them half-sourced would contaminate the index. The v1 composite lives on data that is free, robust, and directly maps to Kilian's decomposition.

Revisit once revenue supports a real commodity-data subscription; the v2 upgrade should land as one coordinated release with methodology-change notes, not drip-fed additions.

Canon

Kilian (2009, AER) — oil supply/demand/precautionary decomposition; Hamilton (2003, J. Econometrics) — oil-price nonlinearity and recession identification; Stock & Watson (2002, JBES) for expanding-window z-score discipline inherited from the BCI.

Inflation pipeline

Energy is the first stage of the producer-price → consumer-price inflation pipeline. WTI YoY enters PPI within ~1 month and feeds CPI YoY at ~3-6 months. The downstream nowcast lives on the Inflation Nowcast page — a 2-input composite (PPI + core PCE 6m annualized) calibrated by 60-month rolling-origin OLS at +0.88 OOS correlation. When Energy stress is elevated, the inflation nowcast tends to move higher 3-6 months later.

Inputs: Yahoo Finance (BZ=F, CL=F, HG=F, GC=F, ^OVX). Nightly rebuild. See the methodology index for the full indicator manifest.