Falling Inequality in Brazil: The Unusual Suspects

Poverty - Inequality - Aid Effectiveness

Rising income inequality in the developed world has attracted a great deal of attention in both academic and policy circles. But inequality can fall, as well as rise. A recent World Bank study has found that income inequality fell (by more than one Gini point) in 39 of 91 countries for which data were available in the 1993-2008 period (World Bank, 2016). Most of these declines were found among emerging and developing countries. In Latin America and the Caribbean, inequality fell in sixteen of the seventeen countries where it can be consistently measured between 2003 and 2011 (Messina and Silva, 2017).

Brazil stands out. Long one of the world’s most unequal countries, Brazil surprised pundits by recording a massive reduction in household income inequality in the last couple of decades. Between 1995 and 2012, the country’s Gini coefficient for household incomes fell by seven points, from 0.59 to 0.52. For comparison, Jacobson and Occhino (2012) report that all of the inequality increase in the United States between 1967 and 2011 amounted to eight Gini points. Most of that inequality decline in Brazil was driven by falling dispersion in labor earnings, rather than by larger and better-targeted transfers, such as Bolsa Família and the like. Those transfer programs did play a role, but the bulk of the equalization came from a decline of almost 20% in the Gini coefficient for labor incomes, from 0.50 to 0.41.

What’s more: most of this wage compression took place during a period – the 2000s – when the Brazilian economy was growing and its labor market thriving. Minimum, median and mean wages rose, unemployment fell, and the formal sector grew at the expense of informality. What accounts for this remarkable performance? In an age of growing concern with rising inequality in most rich countries, are there any lessons to be learned from this experience in one of the world’s ten largest economies?

In a recent paper (Ferreira, Firpo and Messina, 2017), we examine the recent evolution of earnings inequality in Brazil and study possible determinants. The main suspect – on which a number of earlier papers, including Barros et al. (2010), had focused – was education. There had been a marked increase in the supply of skilled workers (in this context, those with completed high-school or more) and demand for those skills had failed to keep pace. The result was a decline in skill premiums – the story went – driving down wage inequality. Another key suspect was the sharp rise in minimum wages, which doubled in real terms between 1995 and 2012, leading to a growing spike nearer and nearer the median of the wage distribution (see figure 1).  Other suspects included changes in the sectoral composition of the labor market (possibly as a result of the commodity boom and resulting changes in the demand for Brazilian exports); changes in the demographic composition of the labor force (including greater female labor force participation), and geographical factors (such as continuing urbanization). 

The analysis uses recent decomposition methods based on re-centered influence function (RIF) regressions (Firpo, Fortin and Lemieux, 2009; and Fortin, Lemieux and Firpo, 2011) to estimate the quantitative impact of five groups of candidate explanatory factors on changes in the Brazilian earnings distribution. These factors are: i) human capital; ii) labor market institutions; iii) demographic characteristics of workers; iv) spatial segmentation; and v) sectoral distribution of the labor force. In the spirit of Oaxaca (1973) and Blinder (1973), but for different measures of inequality, we separate out what can be attributed to changes in the distribution of observable workers’ characteristics - the composition or endowment effect - and what is due to changes in the premiums associated with those characteristics, the pay structure effect. The analysis distinguishes between two sub-periods: 1995-2003, a period of mild inequality reduction but declining real income and wages; and 2003-2012, a period marked by strong growth in income and wages and a sharp reduction in inequality.

The main decomposition results for the Gini coefficient of labor earnings are shown in Table 1. The first result was rather surprising: the overall effect of educational dynamics was actually inequality-increasing!  A substantial increase in the schooling levels of Brazilian workers did indeed take place: the proportion of the working age population with at least 10 years of schooling doubled from 25 to 50% between 1995 and 2012. And this was followed by a decline in the returns to education, as hypothesized (see Figure 2).

This decline of the schooling premium during 2003-2012 was inequality-reducing (the pay structure effect). But it was too weak to offset what we sometimes call the ‘paradox of progress’: increases in educational attainment—even when its dispersion declines—may be inequality-enhancing because of the marked convexity of the earnings-education profile. As the distribution of education shifts to the right, the density mass at the range of years of schooling with the steepest returns increases (the composition effect). That contributes to an increase in average earnings, but also to rising earnings inequality.

A similar story of mutually offsetting effects also applies to the role of minimum wages. During the boom years of 2003-2012, rising minimum wages were unambiguously equalizing.  But in the earlier sub-period of 1995-2003, when labor demand was not growing as fast, minimum wage growth was accompanied by falling compliance, with an increase in the number of workers earning less than the minimum wage of a full six percentage points of the labor force. This ‘composition effect’ proved to be inequality-enhancing, so that the overall effect of Brazil’s minimum wage policy over the entire period was muted.  One possibly more general lesson from this experience is that the effectiveness of minimum wage policies is highly context-specific: even within the same country, their impact on employment varies over the business cycle, and in turn helps determine its effect on the wage distribution as a whole.

So what factors accounted for the largest share of Brazil’s inequality decline over this period? It turns out that the main equalizing change was a marked decline in the returns not to education, but to experience. Figure 2 shows the two profiles. The decline in remuneration to (potential) labor market experience accounted for 3-5 Gini points of the inequality reduction, depending on the specification. The importance of falling returns to experience for inequality dynamics is confirmed by the experiences in Argentina and Chile during the same period (Fernández and Messina, 2017) and is reminiscent of the discussion of age- (Behaghel and  Greenan. 2010) and experience- (Caselli, 2015)-biased technical change elsewhere.

The second biggest culprit – or rather, hero – of the story was a combined reduction in gender, racial, urban-rural, and formal-informal wage gaps. These gaps are estimated conditionally on all observed characteristics, of course, including the human capital and labor market institution effects. While we cannot attribute them causally to changes in labor market discrimination (across genders and races, say) or segmentation (between formal and informal sectors, or urban and rural areas), the observed patterns do suggest that the “playing field” in Brazil’s labor market for workers of comparable levels of observed education and experience, working in the same sectors, was becoming more level. When these falling conditional wage gaps are taken together, they account for approximately half of the nine-point decline in the Gini coefficient for labor earnings. Results are very robust to a battery of checks including decompositions of other distributional statistics.

Our results raise at least as many questions as they answer. What are the economic forces driving the decline in returns to experience in Brazil? Does the lower formal-informal wage gap reflect a corresponding narrowing of the productivity gap? Will the gender and racial wage gaps continue to close even as Brazil’s labor market struggles in the midst of the current two-year recession? Much more work is needed to shed light on questions such as these, and thus to enhance our understanding of how some countries do manage to reduce inequality.


Barros Ricardo, Mirela Carvalho, Samuel Franco and Rosane Mendonça (2010) “Markets, the state and the dynamics of inequality in Brazil.” in: L.F. Lopez-Calva and N. Lustig (eds): Declining Inequality in Latin America: A decade of progress? Washington (DC): Brookings Institution and UNDP.

Behaghel, Luc and Nathalie Greenan. (2010) “Training and Age-Biased Technical Change”, Annals of Economics and Statistics. pp. 317-342.

Blinder, Alan S (1973) “Wage Discrimination: Reduced Form and Structural Estimates”, Journal of Human Resources,  8: 436-455.

Caselli, Francesco (2015) “Experience-biased Technical Change”. Unpublished manuscript. London School of Economics. Available here: http://personal.lse.ac.uk/casellif/papers/EBTC

Fernández, Manuel and Julián Messina (2017) “Skill Premium, Labor Supply and Changes in the Structure of Wages in Latin America”, Inter-American Development Bank Working Paper Series 786

Ferreira, Francisco, Phillippe G. Leite and Julie A. Litchfield (2008) “The rise and fall of Brazilian inequality: 1981–2004”, Macroeconomic Dynamics12 (S2): 199-230.

Ferreira, Francisco, Sergio Firpo and Julián Messina (2017), “Ageing poorly? Accounting for the decline in earnings inequality in Brazil, 1995-2012”. Policy Research working paper; no. WPS 8018. Washington, D.C. : World Bank Group.

Firpo, Sergio, Nicole M. Fortin and Thomas Lemieux (2009), “Unconditional Quantile Regressions”, Econometrica, 77 (3): 953-973.

Fortin, Nicole M., Thomas Lemieux and Sergio Firpo (2011), “Decomposition methods in economics.” Handbook of Labor Economics, Vol 4, 1-10. (Amsterdam: Elsevier).

Messina, Julian and Joana Silva (2017). Wage Inequality in Latin America. Understanding the Past to Prepare for the Future. Forthcoming in World Bank Publications. Washington, DC:

Oaxaca, Ronald (1973) “Male-Female Wage Differentials in Urban Labor Markets”, International Economic Review 14 (3): 693-709.

World Bank (2016): Taking on Inequality: Poverty and Shared Prosperity Report 2016. Washington, DC: World Bank Publications.

Share this