摘要
By introducing proper pricing instruments, farm level water can be used more efficient. But it demands a proper estimation strategy to analyse the efficiency and the input use behavior of farms under the new pricing system. As most production relationships are stochastic in nature, excluding random errors and noise from model specifications often leads to criticism. By using a probabilistically constrained programming formulation, an empirical estimation of stochastic data envelopment analysis (DEA) is done to analyze the efficiency of irrigation water use in the agricultural production system in the Krishna river basin, India. The results illustrate that water demand is higher for farms at, or close to the frontier and lower for those with low efficiency levels. In a second step, a simulation model is developed by using the frontier and economic efficiency derived from stochastic DEA to analyse the impact of water pricing on water use efficiency and water use behaviour. It is shown that an increase in the water price would not cause sizeable profit loss, if the pricing system is administered on a volumetric basis, but water demand would decrease substantially.