annual rate of 3.0%. In the subsequent years, it displayed high growth rates in 2008 (5.1%) and 2009 (7.5%).18
Given this background, we now turn to an examination of inequality in Brazil. Latin America has historically been characterized by very high levels of inequality.19 In a comparison to various regions of the world in 2004, Lopez-Calva and Lustig (2010, Figure 1-1) show Latin America to be characterized by the highest level of inequality. Latin America has an RG of about 55%, whereas the corresponding figure for South Asia and East Asia are about 40%. Brazil is no exception to this Latin American feature, and has typically been characterized by high levels of inequality, e.g. second highest in the world in 1984 (Ferreira et al., 2008). To understand Brazilian inequality, scholars have used various household surveys, but the Pesquisa Nacional por Amostra de Domicílios (PNAD) is quite popular (Ferreira de Souza, 2012). Like the surveys conducted by NSSO in India discussed earlier, these are large cross-sectional surveys. However, unlike the Indian surveys, these are conducted at more frequent intervals, almost annually.
Table 4:Inequality in household per-capita income, Brazil, 1993–2009.
| Year | Relative Gini |
| 1993 | 0.60 |
| 1995 | 0.60 |
| 1997 | 0.60 |
| 1999 | 0.59 |
| 2001 | 0.59 |
| 2003 | 0.58 |
| 2005 | 0.57 |
| 2007 | 0.55 |
| 2009 | 0.54 |
Source: 1993–2007 estimates are from Barros et al. (2010). 2009 estimate is from Ferreira de Souza (2012).
To the best of my knowledge, the literature on Brazilian inequality has focused upon relative measures (e.g. RG) and ratios of income shares of rich and poor quantiles (e.g. richest 20% to poorest 20%) to estimate inequality. I will therefore (and unlike in the discussion on India in the second section) only discuss relative measures for Brazil. Barros et al. (2010) present the RG of income inequality using the PNAD surveys for the 30-year period 1977–2007. Inequality reached its peak level (about 63%) in 1989 and its lowest level (about 55%) in 2007. In Table 4, I present inequality estimates for the period since 1990s from Barros et al. (2010) and Ferreira de Souza (2012). From this table, it is clear that inequality has been falling steadily since the 1990s, although it continues to be high compared to international standards. Among Latin American countries, Brazil is not unique in this regard. Lopez-Calvo and Lustig (2010, Figure 1-2) consider 17 countries in Latin America and show that during 2000–2006, inequality fell in 12 of them, with Ecuador registering the sharpest fall.
Several scholars (e.g. Ferreira de Souza, 2012; Soares et al., 2016) have pointed out that poverty has been falling in Brazil in recent times. The Brazilian government has suggested two different thresholds of household per-capita income, one for poverty (100 Real) and the other for extreme poverty (50 Real) (Soares et al., 2016). Both poverty and extreme poverty have decreased sharply. During the period 2004–2013, the former fell by about 14 percentage points (22.4% to 8.9%) and the latter fell by about 4 percentage points (7.6% to 4%) (Soares et al. 2016). Average income has been rising in Brazil since the mid-1990s, particularly since 2003. The mean real income (household income per-capita, in US dollar Purchasing Power Parity (PPP) terms) grew from $221 in 1995 to $245 in 2003 and then to $372 in 2009 (Ferreira de Souza, 2012, Figure 1). What is more interesting is that the income of all the centiles, including the poor, grew. In fact, the incomes of the poor grew at a much faster rate than those of the middle and richer groups. The growth incidence curve from Ferreira de Souza (2012, Figure 2) has a distinctly downward slope.
Moving from interpersonal to group-based considerations, Soares et al. (2016) classify Brazilian households into four groups based upon occupations: agricultural, pluriactive, non-agricultural rural, and non-agricultural urban, based upon source of income and official designation (of rural or urban).20 In the early 2000s, the ranking of deprivation of these groups (highest to lowest) was as follows: agricultural, pluriactive, non-agricultural rural, and non-agricultural urban. Soares et al. (2016, Figure 1A) show that poverty has fallen steeply during 2004–2013 for all the groups; except for Pluriactive households, extreme poverty has also fallen for all groups (Soares et al., 2016, Figure 1B). The performance for the most deprived group, agricultural households, is particularly impressive — poverty has fallen from almost 50% to about 25%, and extreme poverty has fallen from about 20% to about 5%. The share of agricultural households has been falling, whereas the share of non-agricultural urban households has been rising (as is expected from the perspective of structural transformation). On the contrary, the shares of Pluriactive and Non-agricultural rural households have been stable (at about 8% and 4%, respectively). Soares et al. (2016) argue that the stability of the share of pluriactive households, combined with the lack of progress in eradicating extreme poverty among them, is a cause for concern and that these households should be the target of specific policies.
Inequality based upon race and ethnicity is also of considerable interest in the case of Brazil and has in fact attracted attention in the media recently, e.g. The Economist (2008) and Barbara (2015). Racial/ethnic inequality has deep historical roots going back to the colonial period and slavery. In the interest of space, I will not go into these details and instead refer the readers to Skidmore (1996) and Bucciferro (2017). Bucciferro (2017) presents a comprehensive analysis of the evolution of racial inequality on many dimensions: education, life expectancy, occupation, and income. In general, racial differences have narrowed over time, but continue to be high. In the period of concern to us, progress was modest, e.g. the difference between years of education of whites and non-whites has fallen from 2.1 years in 1998 to 1.8 years in 2008. Bucciferro (2017, Figure 5) presents the evolution of incomes of various groups (Black/Pardo, Black, Indigenous, Asian, Pardo (Brown)) relative to that of whites since the 19th century. The ratio of average non-white to white earnings rose from 0.417 to 0.505 during 1998–2008.
We now move to a comparison of India and Brazil and some explanations for their different performance.
Comparison and Explanations
Over the past two-and-half decades or so, India has grown rapidly, but disparities have also grown. Interpersonal inequality has increased, in whatever way we conceptualize and measure it, particularly in urban areas. Class-based disparities have also increased. One idea that has captured the imagination of both scholars and policymakers in India is ‘inclusion’ or ‘inclusive growth’. In fact, inclusive growth was made an objective by the erstwhile Planning Commission, although its conceptualization was too broad and not operationalized. Several scholars have tried to operationalize the idea of inclusion and examine whether Indian growth has been inclusive. Motiram and Naraparaju (2015) use a framework developed in the literature on pro-poor growth to argue that Indian growth in the 2000s has not been inclusive. They do so by examining whether the growth of quantiles of the poor is adequate or not (in comparison to the growth of the average person). They find that poverty has declined, but the poor have grown at slower rates compared to the middle and richer groups. The shortfall in the growth of the poor is particularly severe in urban areas. They also extend this framework to examine subgroups of the population and find that growth of the poor belonging to disadvantaged groups (castes and classes) has been inadequate. Other scholars (e.g. Jayaraj and Subramanian, 2012a, 2012b; Suryanarayana and Das, 2014) have used other methodologies to assess inclusion, and come to the same conclusion. I have discussed the details of these