A project where every workstream is waiting on evidence another has not produced yet does not look like a project in trouble. Every status line is honest, every owner is working, and the date moves anyway.
On a project of any size the sequence is rarely a line. The financing decision waits on a consent condition. The consent condition waits on a term in a contract that has not been agreed. The contract term waits on the financing case, because nobody will commit to a number before they know what the money costs.
Each of those three positions is correct. Each owner is doing the responsible thing, which is to decline to commit ahead of the evidence. The result is a circle, and a circle has no first step.
This is not a planning failure. A well-run project produces exactly this shape, because sound governance means declining to approve ahead of evidence — and three parties declining in a ring is what discipline looks like from the inside.
A status report asks each workstream whether it is blocked, and each one answers accurately: awaiting input. Nothing in that format can represent a dependency that comes back round to the workstream that raised it.
The tell is not a red line. It is a date that moves by the same two weeks, in a different workstream each month.
Every line reads amber, and no line reads red.
The loop has no owner, because it belongs to no workstream.
The slip is recorded against whichever workstream is asked last.
Somebody has to commit to an assumption in writing before the evidence for it exists, and carry the consequence if it changes. That is a governance act, not an analytical one — which is why producing another paper does not move it, and why it usually waits for the person who is authorised to be wrong.
Handle maps the sequence as a dependency graph rather than a plan: every decision, and the specific evidence it will not proceed without. Circles become visible the moment the evidence is named rather than assumed.
In practice one edge of the circle is almost always cheaper to cut than the others, and it is rarely the one under the most pressure. A financing case can usually carry a stated assumption about a contract term; a signed contract term cannot easily carry an assumption about financing.
Name the evidence each decision actually requires — not the document it usually arrives in.
Find the cycle, and price each edge: what it would cost to commit early, and who carries it.
Cut the cheapest edge, in writing, with the assumption and its owner recorded.