Negotiating with a government counterparty, the most expensive assumption you can make is that the people in the meeting are the people who decide.
They usually are not. In most ministries, the official leading the discussion is carrying a mandate rather than an authority. They can advance a position, they can report back, and in many cases they can kill something quietly — but they cannot say yes. The yes lives somewhere else, often in an agency that has never appeared in the process and whose objection, when it arrives, will be procedural rather than substantive.
What this changes
If you assume the room decides, you negotiate as though concessions buy agreement. They do not. They buy a slightly better report to whoever actually decides, and you have spent them without knowing what the decision-maker cares about.
The practical consequence is that early concessions in government negotiations are often wasted. They are made to establish goodwill with a party that has no capacity to reciprocate, and by the time the real authority engages, the negotiator has already moved off their opening position for nothing.
Mapping the structure
The work of finding out who decides is unglamorous and mostly consists of asking better questions than "who signs?"
- Who has to be consulted, formally or by convention? Convention matters more. An agency with no statutory role in the approval can still stop a project by declining to give an informal clearance that everyone expects it to give.
- Who bears the consequence if this goes wrong? That person's objection carries more weight than the org chart suggests, because they will be asked about it later.
- What is the political cost of approval, and who pays it? If the answer is a minister facing an election, the timeline is not a scheduling question.
- Who has killed a comparable project before? Precedent within an institution is extremely durable.
The corollary
Once the structure is mapped, the sequence changes. You engage the people who can say no before you engage the people who can advance the file, because the objection you cannot see is the one that ends the process eight months in.
This is also why compressing the assessment phase is a false economy. Every hour not spent understanding who decides is spent later negotiating against the wrong party.