The targeted public distribution system, intended to provide subsidised food to poor households, is the largest welfare programme in India, with a budget corresponding to about 1% of the net national product. Several studies have found the system to be inefficient and costly in assisting the poor. This paper analyses the case for, and against, replacing a reformed version of this system with a targeted and differentiated cash transfer scheme. Such a scheme could cover about two-thirds of households, and make far larger transfers to the poorest compared to the actual subsidy embedded in the current system, eliminating the risk of large exclusion errors. Further, the overall budget can be held at the present outlay level. It is argued that most of the objections to such a transfer scheme can be circumvented at the design stage.
