Do Smaller States Lead to More Development? Evidence from Splitting of Large States in India
*Names in bold indicate Presenter
As states are the proximate determinants of local institutions driving development outcomes, a change in their boundaries provides an opportunity to evaluate the impact of these shifts on the provision of public goods and distribution of development outcomes. The splitting of states can have many effects on both the parent and the new state. For example, the division can unleash suppressed growth potential in the new state, stemming from efficiency in decision making, better management and allocation of public resources due to smaller size, conducive policies that meet local needs in a geographically and culturally homogenous areas, and control over natural resources found in the newly created states. Potential negative impact on the new state could be the increased vulnerability to the pressures of corporations and the greed of newly emergent regional elite. However, as majority of the impacts seem to be positive, the paper hypothesizes that the overall impact of state splitting on both states will be positive and will be reflected in the developmental indicators. The paper uses large datasets from National sample survey (120,000 households over multiple periods) and Indian Human Development Survey to test this hypothesis.
The paper uses three methods to show the impact of the split on development through policies and efficient administration. First it uses a difference-in-difference in which all the adjacent states are considered as comparison and the states that go through division as treatment. Second, this paper goes one step ahead and uses the methodology of synthetic control to get precise estimates of impact of division. The synthetic control which is a relatively new method in economics is relevant when there is no single untreated unit that provides a good comparison for the unit affected by the treatment or event of interest. Third, to show how the new states have done better (or not) compared to the old states, the paper uses a Regression Discontinuity Design based on geographical boundaries (Imbens and Lemieux, 2008). RD method was used to see the changes in developmental indicators on people living in districts on both sides of the newly formed state borders.