Unemployment fell in every state. The gap between them depends on the yardstick

In the second quarter of 2021, 14.2% of the people in Brazil's labor force were looking for a job and not finding one. Five years later, in the same quarter of 2026, the figure was 5.4%. That is a 62% drop, and the 2026 number is the second lowest in the PNAD Contínua series (IBGE's continuous household survey), which begins in 2012. Only the 5.1% at the end of 2025 is lower.

The line below shows the whole path. The sawtooth pattern is seasonal: the rate usually rises every first quarter, when temporary year-end contracts expire. That is why every comparison in this piece is made between second quarters.

No one was left behind

The decline was not the privilege of one region. All 26 states and the Federal District ended the period with lower unemployment than they started. Espírito Santo had the largest proportional drop: from 11.6% to 2.3%, a reduction of 80%. Amapá had the smallest: from 16.1% to 9.8%, a drop of 39%. As a result, it now has the highest rate in the country.

In the chart, each line is a federative unit. The orange dot is 2021; the blue one, 2026. The longer the stroke between them, the bigger the drop in points.

Two yardsticks, two answers

In 2021, the gap between the state with the most unemployment, Pernambuco (21.8%), and the one with the least, Santa Catarina (5.9%), was 15.9 percentage points. In 2026, between Amapá (9.8%) and Santa Catarina (2.1%), it is 7.7 points. By that yardstick, the country became more like itself.

Now by the other one. In 2021, Pernambuco's rate was 3.7 times Santa Catarina's. In 2026, Amapá's is 4.7 times. By that yardstick, the distance grew.

Both calculations are right. What changes is the question. Percentage points answer how far one state would have to go to catch up with the other. The ratio answers how different a person's chance of being unemployed is in one place versus the other. When all rates fall, the first yardstick tends to shrink even if relative inequality stays the same or worsens. A measure that looks at all the states, not just the extremes, tells the same story: the coefficient of variation across the 27 rates rose from 0.27 to 0.40.

What these charts do not say is why some states fell more than others. Informality, the weight of public-sector jobs, agriculture and migration all enter that equation, and each calls for its own data. That is left for future pieces.