Access to groundwater is a key element for agriculture in many developing countries. This is the case in India, where more than half of the population works in agriculture and 60% of agricultural production relies on groundwater irrigation. The access to groundwater resources can increase farmers’ incomes and foster food security, through an increase in agricultural productivity.
Despite the relevant policy implications of quantifying the impact of access to groundwater on welfare, getting precise estimates of this causal relationship is far from being straightforward. An interesting attempt has recently been made by Sheetal Sekri (2014), who aims at estimating the precise impact of access to groundwater on poverty and the prevalence of conflict in rural India.
The identification of the causal effect of any policy or factor A (in this case low-cost access to groundwater) on outcome B (in this case poverty and conflict) faces a major well-known difficulty: we cannot observe the same individual at the same time being affected and not being affected by A. Yet, exactly this would be the ideal situation in order access the counterfactual impact of A on a given individual. To circumvent this difficulty, several evaluation techniques have been proposed. Broadly speaking, these techniques rely on finding individuals without factor A (i.e. without low-cost access to groundwater) but being similar in all other relevant aspects to the individuals who do have access. We can then estimate the impact of having access to low-cost groundwater by comparing the outcomes of individuals with access to low-cost groundwater against individuals without access.
Sekri exploits an exogenous change in the technology required to access groundwater. The required technology varies with the depth from which water is extracted: while water within 8 meters below ground can be extracted by a low-cost pump, deeper water requires more costly extracting methods. The pumps required to extract water beyond 8 meters below ground cost three times more than the low-cost pumps used to extract water within 8 meters.
Given that this technological change is exogenous, i.e. not a choice of the individual, it should not be correlated with individual characteristics. This implies that villages on each side of the technological cutoff of 8 meters depth have different costs of groundwater access but are likely to be similar in unobserved characteristics that may affect incomes and the prevalence of disputes. Therefore, in order to identify the impact of groundwater the author uses a regression discontinuity design, which compares the relevant outcomes between villages with groundwater depth just above 8 meters below ground and villages with groundwater depth just below 8 meters. Any jump in the headcount ratio or the prevalence of conflict around 8 meters can then be assigned to the change in access to groundwater (see Figure 1).
Results show that poverty measured by the number of individuals below the poverty line is 9 to 10% higher in villages with more costly access to groundwater (see Figure 2). The reduction on rural poverty in villages with cheaper access to groundwater is explained by an increase in agricultural productivity. Moreover, self-reported conflicts related to irrigation water increase 25% around the cutoff.
These results reveal that policies for water resources management may face two conflicting goals. On one side, sustainability considerations suggest a restriction on access to groundwater resources for agriculture. Given that irrigated agriculture is the main global consumer of groundwater resources, their unconstrained use may end up in depletion and other negative environmental effects such as ecosystem degradation. On the other side, the study shows that from a welfare point of view access to groundwater has relevant benefits. Therefore, policy makers searching for sustainable irrigation methods should bear in mind that a ban on the use of water resources can be quite harmful for rural communities in the short term.
 Wells, Water, and Welfare: The Impact of Access to Groundwater on Rural Poverty and Conflict, American Economic Journal: Applied Economics 2014, 6(3): 76–102.
 A comprehensive description of these techniques can be found in: Khandker, S., Gayatri K. and S. Hussain. Handbook on impact evaluation: quantitative methods and practices. World Bank Publications, 2010.
 For the interested reader, the paper provides a wide array of specifications. Notably, it includes an instrumental variable estimation in which the ownership of more costly pumps is instrumented by the change in technological feasibility of low-cost pumps.
 The use of the headcount ratio to measure poverty has been discussed on a previous post: https://poresp.wordpress.com/2013/11/05/questioning-the-use-of-the-head-count-ratio/