A framework for commodity prices

Source: Goldman Sachs Global Investment Research, Commodity Insights: 10 Lessons Learned with Jeff Currie, September 26, 2023.

AI-assisted summary, condensed from the original PDF notes. Proprietary datasets and model estimates have not been independently reproduced. Page references are to that PDF.

Read the whole curveCopy Anchor Link

The report separates commodity prices into two components:

Price = structural component + cyclical component

The structural component concerns the economics of future supply: project costs, financing, investment hurdles, and uncertainty. The cyclical component concerns physical availability now: inventories, delivery constraints, and the value of having supply on hand.

This is a diagnostic framework, not a formula that produces a reliable price forecast.

Two different questionsCopy Anchor Link

A timespread is the price difference between two delivery dates. The report relates these spreads to inventories, financing, and storage economics (pp. 5–7). Long-dated prices are its starting point for thinking about marginal supply investment (pp. 3–4, 10).

The distinction helps explain an apparent contradiction: prices can remain high while inventories build if investment concerns support the back of the curve. Conversely, a near-term shortage need not imply a lasting increase in project costs.

The feedback loopCopy Anchor Link

Higher prices can fund new supply, discourage consumption, and encourage fuel switching. Those responses take different amounts of time. The report argues that prolonged underinvestment and uncertainty can delay the supply response (pp. 11–16).

It also distinguishes OPEC's influence over near-term balances from the economics of the marginal producer. That is a useful lens; it should not be read as a claim that cartel decisions have no effect on longer-dated prices (pp. 17–19).

How much weight to put on itCopy Anchor Link

The evidence is mostly historical comparisons, correlations, and proprietary models. The inventory/timespread exhibits support a relationship, not a universal forecasting rule. Project-cost data and supply-elasticity estimates depend on samples and assumptions that cannot be reconstructed from the PDF alone.

The report's 2023 forecasts belong to their publication date. This brief retains the framework rather than presenting those forecasts as a current market view.

Use it to organize an explanation: identify which part of the curve moved, inspect the corresponding physical and investment evidence, and state what would disprove the account.

Official economic data sources provide context for further research.