Financing Rural Infrastructure in Developing Countries: the case of India

Motivated by the robust empirical evidence on the positive growth and poverty eradication outcomes of public investment in rural infrastructure, this paper investigates variations in utilization by subnational state governments in India of a recent non-concessional lending facility for financing rural infrastructure projects. Contrary to prior expectations that only states in a robust fiscal situation would voluntarily approach a non-concessional window, a fixed effects panel regression establishes that the scheme was accessed in years of fiscal stress. A second exercise on irrigation funding in a low rainfall state shows allocations to high rainfall rather than low rainfall districts within the state. This is not an efficient allocation, in the light of empirical findings for India and China of higher returns to rural infrastructure in low-potential rainfed areas. Together, the results point to the need for fiscal conditionalities on the borrowing government, going beyond default guarantees. Lending has to be made conditional on upfront evidence of sectoral or general fiscal recovery mechanisms, linked to the sectoral pattern of use, with the time-pattern of disbursement dictated purely by project considerations. Poorer states, and less endowed regions within states, will need technical assistance to identify financially viable projects, since projects readily available off the shelf are typically available for better endowed regions, where the demands in terms of technical complexity and community involvement are lower.