Herfindahl-Hirschman Index in Commodity Markets: How Supply Concentration Quantifies Mineral Volatility Risk
The Herfindahl-Hirschman Index (HHI) is one of the most direct ways to convert supply chain geography into a numeric risk factor. For commodity volatility modeling, HHI captures the structural fragility that determines how a mine closure, export ban, or political disruption in a single country translates into price dispersion across exchange-traded minerals.
HHI Construction and Interpretation for Commodities
The HHI sums the squared market shares of all producers in a given market. For mineral supply chains, the relevant unit is typically country-level production share:
HHI = Σ (sᵢ)²
where sᵢ is the percentage share of global production held by country i. The index ranges from near zero (perfectly fragmented supply) to 10,000 (single-country monopoly). The U.S. Department of Justice considers markets with HHI above 2,500 to be highly concentrated. Most critical minerals exceed this threshold by a wide margin.
Cobalt's production HHI exceeds 4,000 due to the Democratic Republic of Congo controlling over 70% of global mine output. Lithium's production HHI sits near 3,500, reflecting the dominance of Australia and Chile in spodumene and brine extraction respectively. By contrast, copper's production HHI falls below 1,500, distributed across Chile, Peru, China, the DRC, and the United States with no single country exceeding 28%.
The squaring operation is what makes HHI analytically useful rather than just descriptive. It penalizes concentration nonlinearly: a country holding 60% of supply contributes 3,600 to the index, while two countries each holding 30% contribute only 1,800 combined. This convexity mirrors how volatility responds to concentration. A disruption affecting a 60% producer generates far more than twice the price impact of a disruption affecting a 30% producer.
From Concentration to Volatility: The Transmission Mechanism
High HHI minerals exhibit structurally wider volatility distributions because the probability of a supply-relevant event is concentrated in fewer jurisdictions. The transmission channel operates through three layers.
First, substitution elasticity collapses when supply is concentrated. Cobalt consumers cannot rapidly switch to alternative sources when DRC output is disrupted because no alternative source of comparable scale exists. This inelasticity amplifies price moves for any given supply shock.
Second, concentrated supply chains tend to correlate with concentrated processing. China refines over 65% of global cobalt, over 60% of global lithium, and over 35% of global nickel. Processing HHI often exceeds mining HHI, creating a second chokepoint that compounds the volatility effect of upstream concentration.
Third, geopolitical event risk scales with concentration. A single regulatory change, export restriction, or infrastructure failure in a dominant producer country can move global supply by 10% or more. For minerals with HHI above 3,000, the Volterra model treats supply geography as a persistent feature rather than a one-off shock variable.
HHI as a Feature in the Volterra Model
The Volterra dataset incorporates geographic concentration indices as static and slow-moving features within its XGBoost architecture. HHI values at the mining and refining stages enter the model alongside 96 daily GDELT GKG news files, market microstructure data, and cross-commodity signals to produce volatility probability forecasts at five risk levels.
Concentration indices do not change daily, but their interaction with event-driven features is where predictive power emerges. A GDELT spike referencing DRC mining policy generates a different signal for cobalt (HHI above 4,000) than an equivalent spike referencing Chilean mining policy generates for copper (HHI below 1,500). The model learns these interaction effects through walk-forward cross-validation, achieving a mean AUC of 0.815 across the full mineral coverage set.
For minerals with HHI above 2,500, the Volterra model's baseline volatility probability tends to sit higher at every horizon. This means the 7-day, 14-day, and 30-day forecasts for cobalt and lithium rarely register LOW risk, reflecting the structural reality that concentrated supply chains produce wider realized volatility distributions even in the absence of active disruptions.
Practical Implications for Desks and Risk Managers
HHI provides a framework for normalizing volatility signals across minerals with different supply structures. An ELEVATED signal on copper carries different information than an ELEVATED signal on cobalt, precisely because the baseline concentration differs. Options desks calibrating vol surface positioning across multiple minerals should weight concentration-adjusted baselines into their skew and term structure assessments.
For procurement teams, HHI directly informs contract structure decisions. Minerals with HHI above 3,000 warrant longer-duration fixed-price contracts or more aggressive hedging overlays because the tail risk of a single-source disruption is structurally embedded.
Risk managers using VaR frameworks should note that standard historical simulation underestimates tail risk for high-HHI minerals. The realized return distribution for cobalt includes regime shifts that standard lookback windows may not capture. The Volterra signal framework addresses this by incorporating concentration as a structural prior rather than relying solely on historical price data.
Figures from the Volterra daily pipeline. Full historical backfill available on AWS Data Exchange.
HHI is not a forecasting tool in isolation. It is a structural feature that conditions how every other signal, from news flow to exchange microstructure, translates into forward-looking volatility probability. Understanding where your mineral sits on the concentration spectrum is a prerequisite for interpreting any volatility model output correctly.