Using G2 to measure income inequality in two Latin American upper middle income countries

Abstract

In his paper, Dr. Joseph Gastwirth argues that an income inequality indicator that combines the Gini coefficient with a measure that captures right skewness is a better measure to detect changes in the distribution that disproportionately favor upper income levels. Gastwirth proposes the indicator, G2, in which the mean/median is multiplied by the Gini coefficient, as a better measure of income inequality in such circumstances. In this paper, I consider whether G2 is a useful measure in Brazil and Mexico, two upper middle income countries that have historically high income inequality that appears to have decreased in the recent past. This paper suggests that G2 may be somewhat better than the Gini coefficient at capturing changes both in increasing income inequality and decreasing income inequality. In order to provide context for interpreting these results, I examine the quality of data on income inequality in both countries by analyzing the income-related questionnaires of the main household income and expenditure surveys for both countries. I find that although Mexico's household income survey does have more detailed questions on non-wage income than Brazil's household income survey, income estimates still fall below consumption estimates in both countries, which suggests consumption may be a better source of information on income inequality in both Brazil and Mexico.