A 70-place rank swing without changing the data

The ranking’s biggest movement comes from what we choose to value.

Saudi Arabia’s position in our experimental index ranges from 51st to 121st across the captured weight grid. Every underlying observation stays the same. The country cohort stays the same. The only change is how much weight we give health, essential services, material resources, civic voice and air quality.

That is a 70-place Rank Drift, the largest observed span among the 170 economies tested. This is not an isolated sensitivity: 76 economies move by at least 20 places between their best and worst tested positions. The median tested span is 18 places.

These are deliberately bounded choices

Each domain receives between 10% and 30% of total weight; weights sum to 100%. We evaluate the 381 combinations at five-percentage-point increments. No domain disappears, and no domain receives a majority of the score. Substantial movement can still emerge when a country’s domain profile is uneven.

Rank Drift makes that fragility visible beside the headline position. A small value means a rank is relatively stable over this grid. A large one means an ordinal claim depends more heavily on the tested priorities. Neither says whether the priorities are defensible.

Do not turn the grid into a probability distribution

The tested choices are not a survey of people’s preferences. A country appearing at a rank often in the grid does not make that rank likely. Nor is the observed span a confidence interval: it varies normative weights, not sampling error.

The grid also leaves out alternative indicators, cohort choices, percentile normalization and uncertainty in the original observations. Its extremes need not be the extremes over every continuously varying weight. If even this restricted exercise changes a country’s position substantially, a one-number ranking deserves a visible qualification.

Explore the six diagnostic models →