The mothballed-refinery decision: the weave, worked

A real decision: with several South African refineries mothballed, should a player restart domestic refining or build for an import-led future? It is the cleanest illustration of the weave — Technical × Economic × Political onto one decision — and of what separates a band-1 analyst from a band-3 engagement manager.

The three threads

🔧 Technical — does it physically work? Refinery condition, conversion complexity, the diesel/petrol yield slate against where demand is actually going.

💰 Economic — do the numbers work? Refining margins versus the landed cost of imported product; the capex to recommission against the import-terminal alternative.

🏛 Political — do the rules and power work? Security-of-supply politics, fuel-price regulation, the strategic-autonomy argument for domestic refining, and who carries the cost.

The band that matters

An analyst can frame and size a single thread, sourced and defensible. The engagement-manager move is to thread all three onto the one decision and name the binding constraint. Here, the binding constraint is not total refining capacity — it is structural diesel-import dependency under EV uncertainty. Find that, and the verdict (act / wait / hedge / reprice / phase) follows. Miss it, and you produce three competent threads that never become a decision.

The rule the case teaches: a confident number with no provenance is band-1 work wearing a band-2 costume. Defensibility — refusing to over-claim on an unsourced assumption — is senior behaviour, not caution.

Internal field note — sanitised for cross-team learning. More field notes