How to compare crude cargoes: landed cost, yield value and working capital

Start with a common commercial basis

The most consequential error in a crude comparison often occurs before arithmetic. One offer may be priced at loading and another at discharge; one may include freight while the other requires the buyer to charter a vessel. A differential can refer to a different benchmark, quotation window or quality adjustment. PPAC maintains public crude and product price references, but those references are not substitutes for the contract’s executable pricing formula. Preserve the formula and its dates as data fields, not free-text comments. [S6]

A refinery is also a specific asset rather than an abstract national processing barrel. EIA’s February 2025 India review describes a multi-asset refining system; it does not establish the current marginal value of a grade at an individual plant. That value needs site constraints, assays, unit availability and product specifications. Treat a generic landed-cost screen as a shortlist, followed by engineering and commercial validation. [S7]

A larger discount does not necessarily reduce landed cost

The apparent $2/bbl advantage in B’s differential is more than consumed by its other assumed costs. B lands $0.65/bbl more expensive. At one million barrels, that difference is $650,000 before quality effects. Do not add the freight row to an offer that already includes the same freight; adjust the bridge to the actual contract basis. Similar care is required for loss allowances, which must not duplicate quantities already measured at delivery.

Connect the cost bridge to saleable product value

A low landed price is not the objective by itself. Assume A yields $88/bbl of saleable product value and requires $3/bbl of variable conversion plus $1/bbl product logistics. Its simplified contribution is $9.25/bbl. Give B $87.20/bbl product value, $3.20 conversion and $1 logistics, and its contribution becomes $7.60/bbl. These numbers are invented for a calculation; neither is an operating estimate for an Indian refinery.

The resulting $1.65/bbl advantage to A combines landed price, product value and variable cost. A real site model would need to resolve detailed yields and constraints rather than simply assign a gross product value. IndianOil’s FY2024-25 investor presentation reports a historical GRM, but its disclosed measure should not be compared directly with this simplified contribution without a definition bridge. Similar units do not guarantee similar economic boundaries. [S4]

Add the cash cycle and the timing basis

At an assumed 10% annual simple funding rate, an additional 15 days financing $74.75/bbl costs $0.307/bbl. For a one-million-barrel cargo, it adds about $307,192. The relevant clock depends on supplier payment, voyage, storage, conversion and customer collection, not just sailing time. Treasury should supply the actual marginal funding or opportunity cost, credit constraints and collateral requirements.

Price timing needs equal attention. A cargo priced on a loading-month average may be sold into product prices established later. A benchmark hedge, where appropriate to the organisation’s mandate, does not automatically eliminate grade, product, freight, currency or timing basis risk. The data model should make each exposure visible before any risk-management instrument is considered. The first protection is an accurate description of the exposure.

What a defensible procurement memo contains

Use the model as a repeatable decision record. Retain the original quote, source time and changed assumptions; document why a lower-ranked cargo was chosen if reliability or operational fit justifies it. After delivery, compare expected and realised costs and investigate deviations. That feedback is what turns a cost calculator into an intelligence process: the organisation learns whether its assumptions were conservative, stale or systematically wrong.

Treat the comparison as a versioned procurement record. Keep the quotation window, assumed arrival date, refinery run plan, product realisations and currency basis beside each cargo. A discount quoted against one assessment window cannot be combined casually with freight or product prices from another. Re-run the same cargo ranking when those inputs change rather than presenting a single spreadsheet result as an enduring preference. The purpose of the model is to reveal the variable that changes the decision: quality adjustment, yield value, logistics, financing or a constraint in the refinery. That variable becomes the next diligence question for the buyer, not a reason to claim that one crude origin is always cheaper.