From Tables 1 and 2, we can observe that bipolarization has been stable in rural areas, but increased in urban areas since the 1990s.
While the above measures are based upon consumption expenditure, the NSSO also conducts an All-India Debt and Investment Survey, which can be used to estimate inequality in wealth. This survey is conducted at less frequent intervals, and for the period that we are interested in, three years are relevant: 1991, 2002, and 2012. Despite the availability of these data, wealth inequality in India has been underexplored (particularly compared to consumption expenditure inequality). Jayadev et al. (2007), Subramanian and Jayaraj (2013), and Anand and Thampi (2016) are some exceptions. Anand and Thampi (2016) present the most recent analysis and show that the RG of wealth as measured by total assets has increased from 0.65 to 0.74. If wealth is measured in terms of net worth, the RG increased from 0.66 to 0.75.
Having examined interpersonal inequality, we will now move to group-based inequality. Caste is one of the important forms of social stratification in the Indian context. The ‘caste system’ is incredibly complex, defying easy understanding and generalization, and continues to be the subject of considerable debate and controversy today. In the interest of space, it is not possible to go into these debates, but readers can refer to the following studies (and the references therein): Chatterjee (1998), Gupta (2000), Dirks (2001), and Rawat and Satyanarayana (2016). I will examine the broad caste groupings that secondary statistical data like the NSS allow us to analyze: Scheduled Castes (SC), Scheduled Tribes (ST), Other Backward Classes (OBC), and Others. The first three groups (particularly SCs and STs) have been historically disadvantaged against and lag behind on many dimensions: income, education, etc. In fact, the scheduled groups are referred to in this manner because they are given a special place (schedule) in the Indian constitution. Motiram and Naraparaju (2015) presented the cumulative distribution functions (CDFs) of consumption expenditure for these four groups in 2011–2012 and showed that the CDFs of the disadvantaged caste groups lie above that of Others in both rural and urban areas. The average consumptions are lower for the scheduled groups compared to OBCs, who themselves have lower average consumption compared to Others.
A serious controversy has erupted in recent times on poverty measurement in India. The official poverty lines recommended by two official committees (Tendulkar and Rangarajan to be artificially low and based upon methodologies that are indefensible.10 However, as the CDFs discussed above reveal (and as noted by Motiram and Naraparaju, 2015), irrespective of the poverty line one uses, there was an unambiguous ranking of poverty rates in both rural and urban areas in 2011–12.11 The Head Count Ratio (HCR) of poverty for the scheduled groups is the highest, followed by the same for OBCs, and then for the Others. Studies that have examined poverty in previous years using official poverty lines (e.g. Motiram and Vakulabharanam, 2011) have come to a similar conclusion.
What about poverty of the entire population? In Table 3, I present (from Motiram and Naraparaju, 2015) the various quantiles of real consumption and their growth in the period 2004–2005 to 2011–2012. As we can observe, all the quantiles, including the poorest, have experienced growth. This implies that poverty has fallen, irrespective of the poverty line that is used. However, the poorer groups have grown at slower rates compared to the middle and richer groups; this is particularly pronounced in the urban areas. Motiram and Naraparaju (2015) also show that the poor among disadvantaged caste groups have grown at slower rates compared to the overall average person (median). Motiram and Vakulabharanam (2012) carry out a decomposition exercise to examine whether inequality between scheduled and non-scheduled groups in consumption has increased or not.12 In this exercise, an inequality measure that belongs to the single-parameter entropy family of inequality indices (e.g. Log-Mean Deviation or Theil, see Shorrocks and Wan, 2005) is decomposed into two components: inequality between groups and inequality within groups. The former, and its contribution to overall inequality, can be used to understand whether group-based inequality has increased. After rising during the period 1993–1994 to 2004–2005, inequality between scheduled and non-scheduled groups has fallen since.
Table 3:Growth of rural and urban quantiles between 2004–2005 and 2011–2012, India.
Notes: Data are expressed in 2011–2012 prices. Real values are computed using price indices implicit in official poverty lines.
Source: Motiram and Naraparaju (2015).
Class is another important cleavage in the Indian context. Motiram and Naraparaju (2015) use the NSS data on consumption expenditure to divide rural India into seven classes based upon their ‘household type’13 and land possessed: Large farmers (greater than 10 hectares), Medium farmers (between 2 and 10 hectares), Small farmers (between 1 and 2 hectares), Marginal farmers (between 0 and 1 hectare), Self-employed in non-agriculture, Agricultural and other laborers, and Others. They show that the rankings of poverty rates and average consumption are as expected, with the lower classes characterized by higher poverty rates and lower average consumption. In urban areas, Motiram and Naraparaju (2015) use the household type (Self-employed, Regular Wage, Casual Labor, and Others) to divide the population. They show that the poverty rates show a clear ranking from highest to lowest: Casual Labor, Self-employed, and Regular Wage. Average consumption shows the reverse ranking. They also show that consumption of the poor among Casual Laborers has grown at a slower pace compared to the overall average (median). Vakulabharanam (2010, 2014) developed a rigorous class-schema based upon landownership and occupations to classify the Indian population into various classes. He used NSS consumption data and performed a decomposition exercise based on the Gini index to show that class-based inequality has been rising since the 1990s.14 Overall, during the high-growth period since early 1990s, India has witnessed an increase in interpersonal inequality in consumption and wealth. This is robust to the way we conceptualize and measure inequality (traditional: relative, absolute, and intermediate; polarization) although the results look much stronger with absolute and intermediate measures. We should also appreciate the (highly likely) possibility that the real increase in inequality is starker than what the data reveals, given the limitations that we discussed above. Class-based inequality has also increased. Caste-based inequality has come down in consumption and other domains (e.g. education) although there are still stark differences among caste groups.15 Having examined the Indian context, we now move to the Brazilian context.
Brazil in the Age of a Second Democratization
In the following discussion, I will draw upon secondary literature to examine inequality and poverty in Brazil in recent decades. Before doing so, it would be worthwhile to provide some background.16 The period since the mid-1980s saw a transition from a military regime and has been described as a phase of “modernization” within a democratic framework (Fausto and Fausto, 2014) or ‘redemocratization’ (Skidmore, 1996).17 Despite the end of military rule and the emergence of a constitution (in 1988), the 1980s and early 1990s saw enormous political and economic turmoil. Fernando Collor de Mello, the elected president resigned in 1992; the rate of inflation kept increasing throughout the 1980s (Baer, 1995 as cited in Skidmore (1996: 194)) and was as high as 2500% in 1993 (Ferreira de Souza, 2012); and five different plans were implemented during the period 1986–1991. The Real Plan (Plano Real) implemented in 1994 tried to control inflation through a new currency, tighter monetary policy, trade liberalization, and privatization. Although this plan was successful in controlling inflation, the Brazilian economy came under stress due to the two crises that hit the world in the late nineties, viz, East Asian and Russian, which led the Central Bank to adopt inflation targeting. In the 2000s, Brazil benefited from Chinese growth, which fueled a rise in the prices of Brazilian exports (commodities) and from increased foreign direct investment. The Brazilian government took advantage of better economic conditions to increase spending on social programs (more on this below). During the period 1998–2007, Brazilian