This study employs the positive mathematical programming (PMP) approach to estimate groundwater derived demand for irrigation using a cross-sectional dataset of 200 predominantly groundwater irrigated farms from the Punjab province of Pakistan. First, we find that the PMP optimal solution uses less water than what is available (being extracted) in order to make farmers allocate all the available land to different crops. Second, when water supplies are constrained farmers allocate land to different crops based on their total returns, not on the irrigation water requirements. The study results suggest that the limiting/constraining groundwater extractions would induce farmers to reconsider their irrigation water demand. The study findings suggest an introduction of Rs. 0.04/m3 of groundwater would not decrease farm income rather it would make farmers aware of the economic value of water. We suggest that although water pricing can induce an efficient use of groundwater extractions, additional policies are also required that improve irrigation water use efficiency.