Determinants of unfair inequality in Brazil, 1995 and 2009

Abstract

This article analyses the trend of unfair inequality in Brazil (1995-2009) using a nonparametric approach to estimate the income function. The entropy metrics introduced by Li, Maasoumi and Racine (2009) are used to quantify income differences separately for each effort variable. A Gini coefficient of unfair inequality is calculated, based on the fitted values of the non-parametric estimation; and the robustness of the estimations, including circumstantial variables, is analysed. The trend of the entropies demonstrated a reduction in the income differential caused by education. The variables "hours worked" and "labour-market status" contribute significantly to explaining wage differences imputed to individual effort; but the migratory variable had little explanatory power. Lastly, the robustness analysis demonstrated the plausibility of the results obtained at each stage of the empirical work.

Inequality of opportunity in Brazil: decomposition of the sources of injustice between 2001 and 2014

“…In the analysis for Brazil, Annegues et al (2015) also find divergences between the trends shown by traditional measures of inequality and measures of opportunity inequality. The authors calculated indices of entropy between 1995 and 2009 and found that measures of IOp in the country remained stable, with effort accounting for approximately 0.19, using the parametric method, and 0.21, using the non-parametric method.…”

“…The non-parametric method requires a greater number of observations in the sample, with a positive bias when we subdivide the sample into groups of circumstances with few observations. As shown in the literature, the discrepancies between the parametric and non-parametric is that the non-parametric methods tend to yield higher values of IOp than the parametric ones (Ferreira and Gignoux, 2011; ANNEGUES et al , 2015). To estimate the lower-bound IOp, the parametric method is seen as an adequate approach (Bourguignon et al , 2007; Ferreira and Gignoux, 2011).…”