Horizontal and Vertical Inequalities in India

Using the first nationally representative detailed income data for India from the 2005 India Human Development Survey (IHDS), Desai and colleagues (2010) reported Indian income inequality to have the high levels typical for lowand middle-income countries. The IHDS Gini of 0.52 puts India at a similar level as Brazil (0.49) and well above the levels observed even in the most unequal high-income countries included in the Luxembourg Income Study (LIS) Database, such as the United States at 0.37 (LIS 2011). This high level of inequality may have come as a surprise to researchers accustomed to the moderate levels often reported for India. Dreze and Sen (2002) report Ginis for rural areas hovering around 0.30 and for urban areas around the mid-0.30s in the last decades of the twentieth century. Concern has been raised as those coefficients rose recently. Datt and Ravallion (2009), for instance, reported urban Ginis rising to the high 0.30s by the end of the first decade of the century. Compared to those results, a coefficient above 0.50 is startling. The main international databases have also included Ginis for India in the low to mid-0.30s. Deininger and Squire (1996) report a mean Gini of 0.32 for 31 annual observations between 1951 and 1992. The World Inequality Data Base updates this for India in 2004 to 0.37. But all of these earlier calculations depended on expenditure-based data that usually provide results well below those for income. Deininger and Squire suggest adding 6.6 percentage points to expenditure-based Ginis to provide better comparability with income-based measures. But Atkinson and Brandolini (2001) are skeptical that a uniform additive fix will suffice—a stance that the IHDS results support. Galbraith and Kum (2005) adjust for the