Have we been too optimistic about the falling inequality in Brazil?

May. 2023

The current consensus says there was an unprecedented decline in 2003-15, against the global trend. New estimates from Mauricio De Rosa, Ignacio Flores and Marc Morgan suggest the story is more complicated.

Household surveys show a falling income inequality in Brazil (and in most of Latam) during the commodity boom, but they tend to miss the very rich. Also, the total income from the surveys does not match the figures in the national accounts (i.e. the things in the GDP).

This paper aims to address those missing pieces. They use admin data to reweight the surveys (to make them more representative at the top) and then rescale incomes (to make the total consistent with the corresponding aggregates in the national accounts).

After adjusting the raw survey data, the Gini index for Brazil tells a different story: (a) the inequality increases by nearly 10 points, and (b) the index no longer falls smoothly over 2003-15. Why? Mostly because the raw survey was underestimating earnings at the top.

Key result: a better measurement of the income that was missing at the top 1% (esp. capital income) probably offsets the fall in inequality observed for the bottom 99% of the distribution (which is closer to the raw results that justify the current consensus).

For all the measurement details and a discussion on the different trends in Mexico, Chile, and other Latam countries, check their working paper.

Originally published here.