All data and replication materials were kindly made public by Francesca R. Jensenius, Professor of Political Science at the University of Oslo and Senior Research Fellow at the Norwegian Institute of International Affairs.
Since 1950, the Indian Parliament and India's state assemblies have guaranteed a minimum number of seats to Scheduled Castes (SCs). Ensuring ascriptive representation for the 16% of Indian citizens who belong to SCs was intended, in part, as a mechanism to equitably allocate resources along caste lines. To implement SC quotas, the federal government non-randomly selected constituencies in which only SC members can run for office, though all members of the constituency are allowed to vote.
The paper uses a dataset of constituency-level data of 3,134 state assembly constituencies from the 15 largest Indian states to compare development levels across reserved and non-reserved constituencies in 1971 and 2001. As reserved constituencies were non-randomly determined, Jensenius forms pairs of reserved and non-reserved constituencies, matching based on pre-selection characteristics to mitigate the effect of selection bias. She finds a null constituency-level effect on overall development, redistribution to SCs, literacy rates, SC employment patterns, and village amenities.
Development from Representation appeared in the American Economic Journal: Applied Economics, vol 7, no. 3 pp. 196–220.