Commodity Analysis

Overview

Analyze commodities from four dimensions — supply-demand balance, pricing model, inventory cycle, and futures structure — and output directional signals suitable for backtesting. Focuses on crude oil (global pricing anchor), gold (safe haven + inflation hedge), and copper (economic barometer).

Core Concepts

1. Crude Oil Supply-Demand Balance

Key supply-side variables:

Variable	Data Source	Frequency	Direction of Impact
OPEC production	OPEC monthly report	Monthly	Production cuts → oil price ↑
US shale output	EIA weekly report	Weekly	Higher output → oil price ↓
Rig count (Baker Hughes)	Baker Hughes	Weekly	Leads production by 3-6 months
Strategic Petroleum Reserve (SPR)	EIA	Weekly	SPR release → short-term oil price ↓

Key demand-side variables:

• IEA global oil demand forecast (quarterly)
• China crude imports (customs monthly data)
• US gasoline demand (EIA weekly report, implied demand)
• Global PMI (leads demand by 1-2 months)

Supply-demand balance signals:

# Simplified supply-demand judgment
if opec_compliance > 90% and us_rig_count_declining:
    supply_signal = "tight"  # bullish for oil
elif opec_compliance < 80% and us_production_rising:
    supply_signal = "loose"  # bearish for oil

if global_pmi > 50 and china_import_yoy > 5%:
    demand_signal = "strong"  # bullish for oil
elif global_pmi < 48 and china_import_yoy < 0:
    demand_signal = "weak"    # bearish for oil

2. Gold Pricing Framework

Four-factor model:

Factor	Weight	Logic	Indicator
Real rates	40%	Real rates ↓ → lower opportunity cost of holding gold → gold ↑	10Y TIPS yield
US dollar index	25%	USD ↓ → gold becomes cheaper in pricing terms → gold ↑	DXY
Safe-haven demand	20%	Risk ↑ → safe-haven buying → gold ↑	VIX + geopolitical risk index
Central-bank buying	15%	Central-bank purchases → structural demand support	WGC quarterly report

Practical rules:

• 10Y TIPS < 0%: strong support for gold (negative real rates mean negative holding cost)
• 10Y TIPS > 2%: pressure on gold (positive real rates reduce attractiveness)
• Correlation between DXY and gold is around -0.6, but not absolute (they both rose in 2022 due to safe-haven demand)
• Central-bank purchases >1000 tons / year (2022-2023 level): long-term structural bullish support

3. Dr. Copper as an Economic Predictor

Copper as a leading indicator:

• YoY copper-price change leads industrial production by about 2-3 months
• Copper / gold ratio is highly positively correlated with the US 10Y Treasury yield (r > 0.7)
• Copper breakout above the prior high confirms economic recovery

Copper fundamental tracking:

Indicator	Data Source	Threshold
LME copper inventory	LME daily report	<150k tons = tight
SHFE copper inventory	SHFE weekly report	WoW decline >10% = tight
Copper concentrate TC/RC	SMM	TC < $30/ton = tight mining supply
China copper imports	Customs monthly report	YoY growth >10% = strong demand

4. Inventory Cycle Analysis

Visible inventory vs hidden inventory:

• Visible inventory: published by exchanges (LME / SHFE / COMEX), transparent and trackable
• Hidden inventory: bonded areas / trader warehouses, opaque but potentially larger
• The true turning point in prices is the turning point in total inventory

Four inventory-cycle stages (using copper as example):

Active restocking (price↑ volume↑) -> Passive restocking (price↓ volume↑) -> Active destocking (price↓ volume↓) -> Passive destocking (price↑ volume↓)
      mid bull market                 late bull market                 mid bear market                 late bear / early bull market

Signal mapping:

Stage	Inventory Direction	Price Direction	Trading Signal
Passive destocking	↓	↑	Long (best buying point)
Active restocking	↑	↑	Keep long positions
Passive restocking	↑	↓	Close longs (warning)
Active destocking	↓	↓	Short or stay neutral

5. Futures Premium / Discount Structure

Contango (futures > spot, normal market):

• Supply is abundant, and the market prices in carrying costs (storage + funding)
• Roll yield is negative (roll yield < 0), unfavorable for long holders
• Deep contango (far month - near month > 5%) = severe oversupply

Backwardation (futures < spot, inverted market):

• Supply is tight, and spot premium reflects strong immediate demand
• Roll yield is positive (roll yield > 0), favorable for long holders
• Deep backwardation (near month - far month > 3%) = squeeze or extreme shortage

Term-structure signal:

# Spread ratio = (front month - second month) / front month
spread_ratio = (front_month - second_month) / front_month

if spread_ratio > 0.02:    # backwardation > 2%
    signal = "strongly bullish"  # spot shortage
elif spread_ratio < -0.03: # contango > 3%
    signal = "bearish"           # oversupply
else:
    signal = "neutral"

6. Seasonality

Oil seasonality:

• March-May: refinery maintenance ends + summer inventory build → seasonal rise (ahead of the "driving season")
• September-October: hurricane season (Gulf of Mexico) → supply disruption → higher volatility
• November-December: heating-oil demand → stronger diesel crack spread

Gold seasonality:

• January-February: Lunar New Year + Indian wedding-season physical demand → relatively strong
• July-August: traditional soft season → relatively weak
• October-November: Diwali + Christmas restocking → relatively strong

Copper seasonality:

• March-April: China construction season starts → demand recovery
• June-July: off-season inventory buildup → pressure
• September-October: "Golden September, Silver October" → demand recovery

Analysis Framework

Five-Step Commodity Analysis

1. Supply-demand sets direction: is the balance in surplus or shortage? Which way are marginal variables moving?
2. Inventory sets rhythm: which inventory-cycle stage are we in? Is a turning point close?
3. Term structure confirms: contango or backwardation? Does it confirm the supply-demand judgment?
4. Seasonality overlay: is seasonality currently a tailwind or a headwind?
5. Macro validation: do the dollar / rates / risk appetite support the directional judgment?

Composite Scoring Template

commodity_score = {
    "supply_demand": +1,    # supply-demand is tight
    "inventory_cycle": +2,  # passive destocking (best stage)
    "term_structure": +1,   # mild backwardation
    "seasonality": 0,       # neutral seasonality
    "macro_env": -1,        # stronger dollar is a headwind
}
# Total score = +3/5 = +0.6 -> bullish bias, but not a strong signal

Output Format

## Commodity Analysis Report — [Commodity Name]

### Supply-Demand Structure
- Supply side: [surplus / balanced / shortage] — [specific data]
- Demand side: [strong / stable / weak] — [specific data]
- Balance table: [inventory build X tons / drawdown X tons]

### Inventory Cycle
- Current stage: [active restocking / passive restocking / active destocking / passive destocking]
- Visible inventory: [LME X tons, SHFE X tons, WoW change]

### Term Structure
- Front-back spread: [contango X% / backwardation X%]
- Roll yield: [positive / negative]

### Composite Score
| Dimension | Score(-2~+2) | Basis |
|------|------------|------|
| Supply-demand | +1 | OPEC compliance rate 92% |
| Inventory | +2 | LME inventory hit 18-month low |

### Trading Direction
- Direction: [bullish / bearish / neutral]
- Confidence: [high / medium / low]
- Risk points: [specific risks]

Notes

• Commodity data sources are fragmented (EIA / OPEC / LME / SHFE, etc.). This skill provides the analytical framework; data should be retrieved through web-reader or entered manually
• Futures prices include roll costs, so direct comparison across different contracts must account for expiry-roll effects
• Seasonal patterns are statistical averages and may be completely overwhelmed by fundamentals in a given year
• Gold has both commodity and financial attributes, and the financial side (rates / dollar) usually dominates short-term pricing
• Copper’s financial characteristics have strengthened since 2020 (copper futures are used as a macro hedge), so pure fundamental analysis may be insufficient
• Inventory data is lagged (hidden inventories cannot be tracked in real time), so cross-check with price and basis behavior
• This framework is for research backtesting only and does not constitute investment advice