Pobreza e desigualdade no Brasil: análises sobre sua evolução e diferenças regionais

Abstract

This doctoral thesis comprises three empirical studies that add to the discussion on the evolution of poverty and inequality in Brazil and their regional differences. The first study explores regional differences in poverty behavior through an unobserved component model. From this model, trend and cycle components were extracted that represent variations of a common and idiosyncratic nature in the state poverty indices. Thus, the determinants of poverty are assessed for different sources of unobserved shocks that influence their behavior. It is noteworthy among the results that the commonality in the evolution of poverty is predominant in most states, which would evidence the predominance of the effects of central policies and/or national shocks on the income of the poorest individuals in different regions of Brazil. However, there are some notable exceptions for some states in the North and South regions. In the case of the North, idiosyncratic contributions were estimated that act contrary to the national trend, worsening the situation of poverty in this region. In the southern states, an inverse situation was detected, where idiosyncratic contributions reinforce the national trend, intensifying poverty reduction in this region. Some econometric exercises were applied in an attempt to discover which factors could be associated with commonality in the behavior of state poverty indicators and regional differences. Among the results, there appears to be a substantial correlation between the common component and income inequality, captured by the Gini index. In regional differences, in addition to the aforementioned variable, there were significant correlations with the GDP per capita, employed population and urbanization variables. Furthermore, it was found that regional differences may be related to the pro-poor growth pattern of the states. For example, in the southern states pro-poor growth was considerably stronger than in the rest of the country. The second study performs a similar analysis to the first article, but with a focus on regional differences in terms of income inequality. For a panel of measures of income distribution over the period 1976-2019, a latent factor model is estimated that breaks down the variations of these measures into three components: a national factor, which captures the commonality of variations in all indicators of all indicators. States; state factors, which represent the unique commonalities in the variations of each state’s indicators (separately); and idiosyncratic factors, which correspond to the unique variations of each indicator in each state. The results indicate that changes in the income distribution of the richest states are more strongly associated with the national factor, while for the poorest states the state factor was the main responsible for the inequality dynamics. Furthermore, measures of inequality that give greater weight to lower strata of the income distribution seem to be completely disconnected across states. This implies that policies to combat inequality, depending on the interest group to be treated, must be individualized. Finally, the third study complements the previous research seeking to identify which characteristics of individuals and the market would be associated with changes in income distribution. The context chosen for analysis refers to the recent Brazilian economic crises of 2008 and 2014. Based on the quantile decomposition model proposed by Firpo et al. (2018), estimates indicate that the wages of the highest paid workers were the most sensitive, probably as a result of less rigidity at the top of the wage distribution. On the other hand, the composition effect works in the opposite direction, offsetting the structural effect. This means that although the highest paid workers were most affected, the remuneration of endowments helped to nullify the effect of the crisis. It is noteworthy, in general, that the middle class was the most affected by these crises. Finally, it is observed that these crises contributed to the reduction of inequality.