Analysis

This technical economic analysis connects the October EIA annual oil forecast, the September IEA refinery report, the October EIA gas balance and the August IEA Indonesian truck brief. Evidence is checked as of 9 October 2026. It examines planning through the remainder of 2026 and 2027, without presenting a trained forecast, a measured investment return or a universal freight surcharge.

The distinction between evidence types is central. August refinery volumes and September diesel prices describe completed periods, with agency estimation where applicable. EIA's annual oil and gas values for 2026–2027 are dated forecasts with inputs closed on 1 October. The calculations below are our assumed sensitivities, not agency forecasts. Their value is to expose mechanisms and required inputs, rather than to attach unsupported probabilities to outcomes.

Crude costs and refinery product constraints

EIA's 6 October edition forecasts annual Brent averages of USD 96 per barrel in 2026 and USD 84 in 2027. Those numbers inform an annual planning baseline; they do not quote today's diesel. IEA's 11 September report estimates August refinery runs at 81.4 million barrels per day, up from July but below a year earlier. Together, the reports make a useful analytical distinction: a crude input benchmark and available finished-product supply are different variables.

A refinery produces a joint product slate. Processing another barrel requires available equipment and yields several fuels, so diesel output does not rise one-for-one with crude intake. Quality requirements, maintenance and transport can widen the delivered diesel premium even when the crude benchmark eases. EIA reports September US retail diesel at USD 6.29 per gallon. That monthly US price is useful context, but substituting it for another country's terminal or pump price would break the geographical comparison. In a working cost model, retain crude, refining spread, distribution, tax and currency as distinct inputs.

Gas supply, LNG delivery and industrial costs

The October gas outlook forecasts US dry production at 112.20 billion cubic feet per day in 2026 and 116.13 in 2027, with LNG exports at 17.6 and 18.6 billion cubic feet per day. It forecasts annual Henry Hub prices at USD 3.48 and USD 3.16 per million British thermal units. Production rising faster in absolute volume than exports is consistent with the agency's supply explanation, but it does not prove that every importing customer's cost falls.

Delivered LNG adds liquefaction, fuel use, shipping and regasification to the feedgas reference; a contract may index part of its bill differently. Storage also addresses timing, not just annual volume. A cold period or pipeline bottleneck can change local availability despite a comfortable national balance. For a factory that buys both gas and transport, treat the gas invoice and carrier surcharge as separate exposures. Adding their percentage changes directly would double-count or misweight them unless their shares of the same total cost are known.

Transparent freight sensitivities

Assume a baseline transport cost of 100 units, with fuel accounting for 30 and other costs for 70. These shares are illustrative, not measured fleet averages. Hold cargo, distance and non-fuel costs constant. If purchased fuel price rises 20%, total cost becomes 70 + 30 × 1.20 = 106, a 6% increase. The cost shock is diluted by the fuel share; it is not a 20% rise in the entire invoice.

If fuel consumption then falls 10% for the same work, the fuel factor becomes 1.20 × 0.90 = 1.08. Total cost is 102.4, or 2.4% above baseline. Alternatively, with unchanged consumption and fully foreign-currency fuel procurement, an assumed 10% increase in local currency required per foreign-currency unit compounds with the 20% fuel rise: 1.20 × 1.10 = 1.32, giving 109.6 total cost. Finally, an assumed 15% increase in fuel-consuming distance with unchanged payload and consumption per kilometre yields 1.20 × 1.15 = 1.38 and total cost 111.4. This last example holds other costs fixed solely to isolate fuel: a real detour could also increase labour and vehicle time.

Measurement, pass-through and limits

The August IEA truck brief warns that average Indonesian truck fuel consumption is not yet well understood. Its focus on data and testing reinforces a general requirement: measure the work denominator before claiming efficiency. Illustratively, 30 litres per 100 kilometres with a 20-tonne payload equals 0.015 litres per tonne-kilometre. With identical distance and consumption but only 15 tonnes, it becomes 0.020, one-third higher. These are assumed examples, not Indonesian fleet measurements; empty return trips need a separate allocation.

Finally, operating cost and price charged are different. A surcharge may update with delay, cover only an indexed portion or use a baseline set before these reports. An operator can absorb costs temporarily; later contract renewal can transmit them. A usable review should retain the actual fuel invoice, payload and distance logs, contract index, exchange-rate exposure and adjustment dates. The scenarios omit taxes, financing, demand response and new-equipment capital cost. No probability distribution is estimated. They support transparent budget discussion and sensitivity checks, while procurement decisions still require local quotations and contract terms.