🧭 The Order Grid

Force is not goodness and integrity is not virtue. A state can be capable, bound by law, and still do great harm. This describes conditions, never worth.

Two things follow. A state near the top of the force axis is a state that can act, which is as true of the ones you admire as of the ones you do not. And a rule of law reading describes what the law is able to bind, not what the government of the day chooses to do inside it.

The reading is also a single recent cross-section, so this board cannot yet show a country going backwards. A state whose institutions are eroding right now still shows the position its institutions had when they were last measured. That is the largest thing wrong with the board, and it is fixed by a time series rather than by an adjustment.

Force. Recognised power in the Power Atlas (0.75) -- spending, productive economy, reach and standing -- blended with an administrative leg (0.25) carried half and half by tax revenue and government consumption, both as a share of GDP. Each a percentile against contemporaries. Two fiscal readings rather than one because tax revenue is the central government series and reads federal states low for a definitional reason; a country with neither reading is scored on recognised power alone rather than on a guess.

Integrity. V-Dem rule of law (0.65), constitutional durability (0.25) and freedom from suspension (0.10). Percentile against contemporaries.

Distance. Distance from the corner on absolute terms, 0 to 100, where higher is further away. Force is the state's share of world recognised power; integrity is the raw 0 to 1 composite. Not derived from the percentile axes, which would put the leader of any field a whisker from the ideal.

Constitutional durability is the age of the constitution in force, on a log scale against the oldest still standing. 171 states are dated from the constitutional chronology, fifteen of them from curated adoption years the chronology cannot supply because their founding document predates the state system it is keyed to. Norway is the clearest case: the constitution dates from 1814, the entry in that system from 1905. Each curated year carries its own source and note.

• Correlates of War National Material Capabilities v6.0 and Maddison Project Database 2023, via the site Power Atlas
• V-Dem rule of law via Our World in Data
• Comparative Constitutions Project, Chronology of Constitutional Events v6.0
• World Bank Open Data, tax revenue and general government final consumption (% of GDP), via build-country-indicators.py

Each axis is cut into thirds, so the nine positions are labels on a continuous score. A country sitting a fraction of a point from a band edge holds its position by less than the measurement error in its own inputs, while its score is not in doubt at all. Every row carries how far it is from a different position, and which of the two axes decides it, so a knife edge does not read as a settled fact.

72 of the 173 states placed here sit within five points of a band edge, and 12 within one. Those 12 are named by a margin thinner than the difference between two adjacent countries in the same list. The scores behind them are not close calls; the labels are.

Read the Held by column with the position, never without it. A state 15 points clear of any edge is where the board says it is. A state 0.4 from one is a state the next data refresh may move, without anything about it having changed.

The administrative leg is what a state raises and spends as a share of GDP, and that understates any state funded off the tax line. Singapore is the clear case: tax revenue 13.6% of GDP and government consumption 10.3%, both near the bottom of the field, for a state nobody would call administratively weak. It funds itself substantially from land sales, returns on its reserves, and mandatory savings that are not counted as tax. The reading is accurate and the inference from it is wrong, which is why the leg is a quarter of the axis rather than half.

This cuts the other way too. The tax series counts what the CENTRAL government takes, so a federal state whose provinces collect the rest reads lower than a unitary one at the same real tax take: Germany and the United States both sit near the bottom fifth on that leg, below Brazil. Pairing it with general government consumption softens the effect without curing it, and the pairing is why the leg is two readings rather than one.

Two states here have no single founding document: the United Kingdom and New Zealand. That is a fact about them, not a gap in the data, and the board says so in the constitution column rather than leaving a blank.

The United Kingdom's constitution is 17 separate documents and roughly 280,000 words, assembled over centuries, and the chronology records no replacement and no suspension of it since 1789. Durability here measures how long the arrangement in force has survived, so an order that has never been replaced scores level with the oldest written constitution, which is the United States at 236 years.

That is worth stating plainly because it is the argument. A designed constitution and an accumulated one can reach the same place, and a board that could only see the designed kind would have missed half the evidence.