Commodity Trading Strategies: Trend, Carry, and Seasonal
Commodity markets offer systematic alpha opportunities that are structurally different from equity and fixed income markets. Physical supply and demand dynamics, storage costs, weather patterns, and the hedging needs of commercial producers create persistent return patterns that quantitative strategies can exploit. The three dominant systematic approaches -- trend following, carry, and seasonal -- each capture a distinct source of return and exhibit low correlation with each other and with traditional asset classes. This guide examines each strategy's mechanics, implementation, and empirical performance.Commodity Market Structure
Futures Curve Terminology
Commodity trading typically occurs through futures contracts rather than physical markets. The term structure of futures prices determines the cost of maintaining exposure: Contango: Futures price > spot price. Each futures contract is more expensive than the previous one. Rolling from an expiring contract to a later one requires buying at a higher price, creating a negative roll yield. Contango is the normal state for most commodities, reflecting storage costs and financing. Backwardation: Futures price < spot price. Rolling generates a positive roll yield (selling the expensive near-month contract and buying the cheaper far-month). Backwardation often occurs when current supply is tight relative to demand. Roll yield: The return from rolling futures positions forward. Annualized roll yield: RY = (F_near - F_far) / F_near * (365 / days_between_contracts) For WTI crude oil in contango with 3-month futures at $75 and 6-month futures at $77: RY = (75 - 77) / 75 * (365/90) = -10.8% annualized This negative roll yield is a significant drag on long-only commodity returns and is a primary motivation for carry-based strategies.Trend Following in Commodities
Why Trend Following Works in Commodities
Commodity prices trend for structural reasons:- Supply response lag: When prices rise, new production capacity takes months to years to come online (drill new wells, plant more crops, open new mines). During this lag, prices continue rising.
- Inventory cycles: When inventories deplete, prices rise until demand destruction occurs or new supply arrives. The drawdown of inventories creates sustained price trends.
- Hedging pressure: Commercial producers hedge by selling futures, creating downward pressure. Speculators (trend followers) provide the other side, earning a risk premium for absorbing this selling pressure.
- Information diffusion: New information about supply disruptions, weather events, or demand shifts is absorbed gradually, creating trending prices rather than instantaneous adjustment.
Implementation
Signal generation: Moving average crossover is the most common approach. Fast/slow combinations: 20/60 day, 50/200 day, or multiple lookback periods combined. Signal: Long when fast MA > slow MA, short when fast MA < slow MA. Position sizing: Scale position size by inverse volatility to equalize risk contribution: Position_i = target_risk / (N * sigma_i) Where target_risk is the portfolio's target volatility, N is the number of commodities, and sigma_i is the estimated volatility of commodity i. Universe: Diversified across energy (crude oil, natural gas, gasoline, heating oil), metals (gold, silver, copper, platinum), agriculture (corn, wheat, soybeans, sugar, coffee, cotton), and livestock (live cattle, lean hogs).Performance Characteristics
Historical performance of a diversified commodity trend-following strategy (1990-2025):| Metric | Commodity Trend | S&P 500 | 60/40 Portfolio |
|---|---|---|---|
| Annual Return | 8.5% | 10.2% | 8.8% |
| Annual Volatility | 12.0% | 15.3% | 9.5% |
| Sharpe Ratio | 0.54 | 0.51 | 0.68 |
| Max Drawdown | -22% | -51% | -32% |
| Correlation with S&P 500 | -0.05 | 1.00 | 0.98 |
| Crisis Alpha (2008) | +18% | -37% | -22% |
Carry Strategies
The Carry Return
Commodity carry captures the roll yield from the futures term structure. Go long backwardated commodities (positive roll yield) and short contangoed commodities (negative roll yield). Carry signal for commodity i: Carry_i = (F_near / F_far - 1) * (365 / days_between) Positive carry = backwardation = long signal. Negative carry = contango = short signal.Roll Yield Decomposition
The total return from a commodity futures position: Total Return = Spot Return + Roll Yield + Collateral Return- Spot Return: Change in the spot price of the commodity
- Roll Yield: Return from rolling futures (positive in backwardation, negative in contango)
- Collateral Return: Interest earned on the cash collateral posted for futures margin
Implementation Details
Cross-sectional carry: Rank all commodities by carry signal, go long the top third and short the bottom third. This is a market-neutral strategy that profits from the spread between backwardated and contangoed commodities. Time-series carry: For each commodity, go long when carry is positive and short (or flat) when carry is negative. This captures the time-series variation in roll yield. Combined: Use both cross-sectional rank and time-series direction to size positions.Performance
Commodity carry strategies have historically delivered:- Annual returns of 5-8% (cross-sectional) or 3-6% (time-series)
- Sharpe ratios of 0.5-0.8
- Low correlation with trend following (0.1-0.3), enabling strong diversification when combined
- Moderate drawdowns (-15 to -20% maximum)
Seasonal Strategies
Why Seasonality Exists in Commodities
Unlike financial assets, many commodities have predictable seasonal supply and demand patterns: Agricultural commodities: Planting and harvest cycles create predictable supply fluctuations. Corn prices tend to peak in late spring (planting uncertainty) and decline into fall (harvest arrivals). Energy: Heating oil demand peaks in winter (heating), gasoline demand peaks in summer (driving season). Natural gas prices typically rise from March to November as storage injections precede winter drawdowns. Metals: Industrial metals follow manufacturing cycles, with demand typically stronger in Q1-Q3 (construction season in the Northern Hemisphere).Quantifying Seasonal Patterns
Calculate the average monthly return for each commodity over 15-20 years: Seasonal_i,m = (1/T) * sum(R_i,m,t) for t = 1 to T Where R_i,m,t is the return of commodity i in month m of year t. Example seasonal patterns (average monthly returns, 2000-2025):| Month | Crude Oil | Natural Gas | Corn | Gold |
|---|---|---|---|---|
| Jan | -0.5% | -3.2% | +0.8% | +2.1% |
| Feb | +1.8% | -4.5% | +0.3% | -0.5% |
| Mar | +2.3% | -2.1% | +1.5% | -0.3% |
| Apr | +1.5% | +0.8% | +1.2% | +1.0% |
| May | -0.8% | +2.5% | -0.5% | -1.2% |
| Jun | -1.2% | +1.8% | +3.2% | +0.8% |
| Jul | +0.5% | +0.5% | -2.5% | +0.3% |
| Aug | -0.3% | +1.2% | -1.8% | +1.5% |
| Sep | -2.1% | +3.5% | -0.8% | +1.2% |
| Oct | +1.8% | +1.5% | +0.5% | -0.5% |
| Nov | +0.3% | -1.5% | +0.8% | -0.3% |
| Dec | +0.8% | -2.8% | +0.5% | +0.8% |
Implementation
Simple seasonal: Go long in historically positive months, short (or flat) in historically negative months. This captures the strongest seasonal patterns but is vulnerable to regime changes and sample-specific effects. Filtered seasonal: Apply a seasonal filter only when the seasonal signal aligns with the current trend or carry signal. This reduces false signals and improves risk-adjusted returns. Intra-month timing: Seasonal patterns often have specific intra-month timing (e.g., the "turn-of-month" effect in gold). More granular seasonal models can capture these sub-monthly patterns.Statistical Significance
Seasonal patterns must be tested for statistical robustness:- t-statistic: Monthly seasonal return / (standard deviation / sqrt(years)). Require t > 2.0 for significance.
- Out-of-sample stability: Verify that the pattern persists across different sub-periods (e.g., 2000-2012 vs. 2013-2025).
- Multiple testing correction: When testing 12 months x 25 commodities = 300 patterns, some will appear significant by chance. Apply Bonferroni or FDR correction.
Combining Strategies
The three strategies (trend, carry, seasonal) have low pairwise correlations:| Trend | Carry | Seasonal | |
|---|---|---|---|
| Trend | 1.00 | 0.15 | 0.10 |
| Carry | 0.15 | 1.00 | 0.20 |
| Seasonal | 0.10 | 0.20 | 1.00 |
- Annual return: 7-10%
- Sharpe ratio: 0.8-1.2 (substantially higher than any individual strategy)
- Maximum drawdown: -12 to -18%
- Near-zero correlation with equities and bonds
Key Takeaways
- Commodity markets offer systematic alpha from trend following, carry, and seasonal patterns, each driven by distinct structural features of physical markets
- Trend following in commodities benefits from supply response lags, inventory cycles, and hedging pressure, delivering positive returns during equity market crises
- Carry strategies exploit the futures term structure, going long backwardated and short contangoed commodities to harvest the roll yield premium
- Seasonal patterns arise from physical supply/demand cycles (planting/harvest, heating/cooling seasons) and must be rigorously tested for statistical significance
- Combining all three strategies achieves Sharpe ratios of 0.8-1.2 due to low inter-strategy correlations, substantially better than any individual approach