A measure of justice

The case for first measuring, and then increasing, social diversity in public spaces is built on the notion of fair and proportional representation of the major groups in society. These groups could be religious, linguistic, castes, tribes, racial, (or based on other markers of ethnicity) and gender based. In most societies, there are easily identifiable groups that are disproportionately poor, disadvantaged, and excluded from several social spaces. It can be demonstrated that these outcomes are not only due to the fact that these groups are less educated but also because of systemic discrimination, past and present.

Unfortunately, this is a problem with no easy solutions. International experience over time suggests that the magnitude of inter-group disparities and the extent of discrimination against stigmatised groups are independent of the country’s overall prosperity level, its rate of growth and the extent of market orientation of the country’s economy. Thus, growth enhancing policies and universal welfare programs, while important, cannot address under-represented groups, such as Dalits in India or blacks in the US. Hence, countries like the US, South Africa, Brazil and others have been experimenting with different policies.

To begin with, the DI measures the gap between the proportion of the group (say, Muslims or women, or Dalits) in an institution and its proportion among the “eligible population.” Thus, for a bachelor’s degree, all those high school graduates constitute the eligible population. This gap is then calculated as a proportion of the share of the group in the population. The gap can be positive (for over-represented groups) or negative. The DI quantifies under-representations, so it only takes into account negative gaps. There are three broad dimensions across which the DI is calculated separately: caste, religion and gender. The different numbers are integrated horizontally (across the three social categories, using semi-flexible weights, reflecting the needs of the institution) and vertically (across the different tiers) to yield one number — a composite measure of diversity. The next step is to classify institutions (whether public or private) according to whether diversity is low (DI value between 0 and 1/3), medium (between 1/3 or 2/3) or high (between 2/3 and 1), and tied to plan allocations by the government.

... contd.