Interstate Development and Disparity: An Unsupervised Learning Approach
Economic and Political Weekly Vol.59; No.38 - 21 Sep. 2024 pp. 74-80
The relationship between inequality and social welfare is an important yet understudied one. Using state-level data, a critical replication of a study conducted in the Economic Survey in 2021 is undertaken and it is found that the claimed positive inequality-welfare correlation does not hold. Then, unsupervised learning methods—principal component analysis and hierarchical clustering are used to classify states into groups based on a gamut of welfare indicators. It is found that some clusters form neatly around geographical divisions (north versus south) and confirm well-established developmental facts; significant heterogeneity among other regions (North-East) that do not adhere to established narratives is documented; and the development gap between states is found to be persistent and path-dependent. Finally, the need to strengthen rights-based legislations and establish a set of universal basic rights to course-correct for persistent inequalities is underscored.