Diversification and Correlation Across Commodities

Diversification means spreading risk across markets that do not move together, so that one bad trade or one bad sector does not decide your year. Within commodities, markets in the same complex (WTI and Brent, gold and silver, corn and wheat) are closely related, while different sectors (metals, energy, grains, softs, livestock) often move more independently. Correlation is not fixed, and in a crisis many correlations rise at once, so you need sector limits as well as a long list of markets.

What correlation measures

Correlation is a number from -1 to +1 that describes how two markets move together:

• +1: they always move in the same direction in proportion.
• 0: no consistent relationship.
• -1: they always move in opposite directions.

Calculate correlation on returns (daily or weekly percentage changes), never on price levels. Two unrelated markets that both trended upward for a year will show a high correlation of prices that means nothing. Also choose a window deliberately: a 60-day correlation shows the recent relationship; a one-year or longer window shows the more stable picture. Compare both.

How commodity markets relate

The relationships below are general tendencies. Measure them yourself with current data, because they change.

Relationship	Typical strength	Why
WTI and Brent	Very high	Same global oil market
Crude, heating oil and gasoline	High	Products are refined from crude
Gold and silver	High	Shared monetary drivers
Corn, wheat and soybeans	Moderate to high	Shared weather, land and feed substitution
Copper and crude oil	Moderate, variable	Both tied to global growth
Gold and crude oil	Low and unstable	Different drivers
Livestock and metals	Low	Unrelated supply and demand
Coffee, cocoa and sugar vs metals	Low	Driven by crop-specific weather

The maths of diversification

For two positions with weights w1 and w2, volatilities s1 and s2 and correlation r, portfolio volatility is:

Portfolio volatility = square root of (w1² x s1² + w2² x s2² + 2 x w1 x w2 x r x s1 x s2)

With many equal-risk positions and an average correlation r between them, a useful shortcut is:

Portfolio volatility = single-market volatility x square root of (1/N + (1 - 1/N) x r)

So lowering correlation helps more than simply adding markets. Ten markets from five sectors usually diversify better than ten energy contracts.

Count bets, not markets

Holding WTI, Brent, heating oil, gasoline and natural gas is five positions but perhaps only two independent bets: oil and gas. A useful habit is to estimate your effective number of bets by grouping highly correlated markets (for example, those with a return correlation above about 0.7) and counting each group as one.

Correlations rise in crises

Diversification is weakest exactly when you need it most. In the second half of 2008, most commodities fell together as the financial crisis spread. In March 2020, energy and industrial metals fell sharply together, and even gold dropped for a short time as investors sold whatever they could to raise cash. The usual causes are forced selling, a rising US dollar and margin calls across markets.

The practical answer is to assume that in a crisis your positions will be more correlated than in normal times, and to cap total risk accordingly. This is why the Turtle rules limited units in closely correlated markets and in one direction overall, as described in the trend-following lesson.

Diversify across methods and asset classes too

You can also diversify by:

• Time frame: a 20-day and a 55-day breakout system enter and exit at different times.
• Strategy type: trend following and short-term mean reversion often perform at different times.
• Asset class: commodities can behave differently from stocks, currencies and crypto. DanioX publishes signals and backtests for stocks, forex and commodities, which lets you compare how similar methods behave in different markets.

Each extra layer of diversification needs its own testing; mixing untested strategies does not reduce risk.

Practical diversification rules

1. Measure return correlations over 60 days and one year at least once a month.
2. Group markets with correlation above about 0.7 and treat each group as one bet.
3. Set a sector risk cap (for example 1.5% open risk per sector) and a total risk cap (for example 5%).
4. Prefer adding a new sector over adding another market to a sector you already hold.
5. Review correlations again after any market shock, because they can change quickly.