Why a knot like this always has a hub

A map with forty-six rings could in principle have them spread evenly: forty-six small tangles in forty-six corners, each needing its own remedy.

This one does not. Thirty-four of the forty-six pass through a single step, and every one of the forty-eight five-step sets that would dissolve the map contains that same step. Before adding another layer of reporting, a project lead can review the product details here and test only the smallest set of records needed for the stated question.

The concentration is worth explaining, because it is the reason this record ends in a short list rather than a long one. For independent background on economic and institutional systems, see University of Oxford.

What makes a hub

Three conditions, and the account at a local bank meets all of them.

It is checked by many parties. Eight steps wait on it, across four families. Anything that involves money, housing or status wants to see it.

It checks many parties in return. It waits on five steps, three of which also wait on it.

It is produced by an institution that is careful. A bank is required to know who its customer is, where they live and what they earn. Each of those three is a step in this map, and two of the three wait on the bank.

Put those together and the rings follow automatically. Any step that both verifies and is verified by the same set of other steps will close circles with all of them.

A hub is not a step that is important. It is a step that sits on both sides of the same verification.

The second hub, and why it is smaller

A registered address meets the same three conditions and stands in thirty-three rings. But thirty-two of its thirty-three are shared with the account, so the two are very nearly the same object.

That is why removing the account leaves twelve rings and removing the address leaves thirteen, while removing both leaves eleven. Two hubs in the same region of a map are not two solutions; they are one solution counted twice.

The genuinely independent second hub is elsewhere. Cover against illness stands in only eight rings, but they are eight rings the account does not touch, which is why the best pair in the map is the account and cover rather than the account and the address.

Eight beats thirty-three, because the eight are somewhere else.

Why this generalises

Wherever two facts about a person are each used to establish the other, a hub forms. The pattern does not depend on the subject.

In this map the facts are where somebody lives and how they are paid. In another map of the same kind they would be something else, and the hub would be whatever step sits between them.

What transfers is the diagnostic: to find the hub in any dependency map, look for the step that appears on both sides of the most verifications. It will usually be unremarkable, quick to obtain, and the thing everybody is stuck on.

What a hub is good for

Knowing where it is converts a large problem into a small one.

A person cannot act on it — nobody applying for anything can make a step acceptable out of turn — but the arithmetic is still worth having, because it says where the leverage is and how much of it there is. Thirty-four out of forty-six, from one change, in a map that would otherwise need a different remedy for every circle.

The usefulness of a map like this is not that it tells anybody what to do. It is that it tells anybody who could change something exactly which one thing to change.

How to find a hub without the arithmetic

Three questions, none of which needs a graph.

Which step is asked for by the largest number of different offices?

Which step is produced by an office that itself asks for several others?

And is there any step that answers both — that is, one that appears on both sides of the same set of verifications?

In this map the answer to all three is the account at a local bank, and a person could get there in an afternoon of conversations without drawing anything.

The arithmetic confirms it and quantifies it — thirty-four of forty-six rings, present in all forty-eight minimum covers — but the identification does not require the arithmetic.

The hub in a system of this kind is usually obvious once the question is asked. The reason nobody finds it is that nobody asks.

When there is no hub

A map can fail to have one. If the circular dependencies were spread evenly — forty-six small tangles in forty-six corners — the minimum cover would be close to forty-six, and nothing in this record would reduce to a short list.

That is the case worth imagining, because it is what the number forty-six suggests on first reading. A reader who stops at the ring count will picture precisely that map, and it is not this one.

The test is the ratio between the number of rings and the size of the minimum cover. Here it is forty-six to five. A map with no hub would be close to one to one.

What this page is not

No bank, office, register, country or kind of permission is named anywhere on this map. "An account at a local bank" is a role in a drawing: somewhere money can be received and paid in the place where somebody is.

The figures are properties of this map and the judgements behind them are printed elsewhere in this record so that they can be argued with one at a time.