Measuring Groundwater Irrigation Efficiency in Pakistan: A DEA Approach Using the Sub-vector and Slack-based Models

We estimate the efficiency of groundwater use in cotton production in the Punjab province of Pakistan. We use a survey data of 189 cotton producers comprising 98 tube-well owners and 91 water buyers in order to get the differential impact of tube-well ownership on groundwater use efficiency. We use data envelopment analysis to compute the technical, scale, cost and allocative efficiencies for tube-well owners and water buyers relative to a meta-frontier and groupfrontiers. The DEA sub-vector and slack-based models are used to compute groundwater use efficiency. The results indicate low levels of technical inefficiencies with water buyers being more inefficient relative to tube-well owners. However, groundwater use inefficiency is more pronounced than the respective technical efficiency. The sub-vector and slack-based estimates are highly correlated suggesting the robustness of the results. The results on returns to scale indicate that the majority of cotton growers are operating at increasing returns to scale, suggesting that efficiency can be improved by expanding the scale of operation. We use a second-stage bootstrap truncated regression to investigate the factors that influence technical efficiency and groundwater use efficiency. We find that the level of education, seed quality and extension services have positive significant impacts on technical and groundwater use efficiency. We suggest that knowledge of crop water requirements and the use of improved crop varieties can play role in improving the efficiency of groundwater use.