Latin American middle class

Shannon O'Neil links to a Brookings Institute study regarding the middle class in Latin America. The study is extraordinarily optimistic about how large the middle class is. Yet a quick look at some data left me scratching my head.

Take El Salvador. Brookings says 46.8% of the country was middle class in 2005. But ECLAC (Table 16.1) says 48.9% of the population was poor in 2001, and 47.9% was poor in 2009. So half the country is middle class and half is poor?

Take Guatemala. Brookings says over 33.8% of the population is middle class. ECLAC says 54.8% is poor.

Or take Mexico. Brookings says 60% of the country is middle class. ECLAC says 34.8% was poor in 2008.

If you accept both reports, then somehow the wealthy have largely disappeared in Latin America, and presumably the Gini coefficients we always see are wrong. Otherwise one or the other is off. This is especially true because poverty rates fail to describe those who are just above the poverty line but at any moment could fall back--perhaps Brookings would label them middle class.

Poverty is gradually--albeit very unevenly--decreasing in Latin America, but we need to be careful about announcing the existence of large middle classes.