There has been some discussion of late in the popular press and in certain public forums on using direct cash transfers for at least some of the items distributed to the poor through the public distribution system (PDS). In this context, a recent study in Delhi is often being referred to. This study was conducted by the India Development Foundation (IDF) on behalf of Self-Employed Women's Association (SEWA)1. The IDF undertook this study for two reasons. First, with Aadhaar and talk of cash transfers, the findings of the study would have obvious policy implications. Second, it allowed us to address an important research question.
There are many studies on cash transfers. Most seem to suggest that conditional cash transfers are more effective in reaching a poverty goal than unconditional cash transfers. Yet this research is not undisputed, and mainly focuses on improving school enrolment rates in Latin American countries2. One of the most controversial issues, which has attracted explicit attention both from policy makers and in academic circles, relates to conditionality. In other words, should transfers be made only if the poor undertake certain specific decisions? Alternatively, should they be given the cash transfer without any conditions attached, for they will take those actions in any case? If the second were right, it would vastly reduce the cost of monitoring the decisions by the poor. Observe that the PDS to those below the poverty line (BPL) is a subsidy to the poor only if they buy the goods distributed through ration shops. On the other hand, if the subsidy amount is directly given to the poor households, it is an unconditional cash transfer without any explicit condition attached to it on how they can spend it.
There are many who say that direct cash transfers would not have the desired effect since poor households would spend the extra cash on things other than food. In particular, men in the poor households will spend more on alcohol instead of the food they were getting from the PDS. It is important to understand the assumption behind such fears. For one, the poor, it would seem, have behaviours that are inherently different from the non-poor. They, unlike us, generally prefer alcohol at the expense of food. Alternatively, there is something triggered by poverty that makes them more vulnerable to choose alcohol over food. Whether or not this is true is an obvious research question.
In our study, we first asked how many would want to get cash in lieu of their BPL cards. We randomly divided all those who came forward into three groups. Group 1 got the cash transfer through a bank account that was opened for the woman in the household; their BPL cards were stamped by the Delhi government stating that they were not eligible to get anything from the PDS shops for one year from the date of the stamp. A bank account was opened for the woman in Group 2, but no cash was given to her and she could use the BPL card in the PDS shop. Nothing was done for Group 3. The reason for creating Group 2 was to isolate the impact of the cash transfer from that of having a bank account.
We first checked that the three groups were more or less similar in their observed characteristics and in their pre-experiment behaviours. We studied their behaviours twice more - once after six months and then again after the completion of one year. We analysed the changes in their behaviour over this year to see if, and how, the behaviour of Group 3 people was different from the other groups.
Recall that a common criticism against unconditional cash transfer in lieu of an in-kind food subsidy (or a conditional cash transfer) is that households would spend less on food and more on other things that could lead to a loss in nutrition with harmful long-term effects, especially on children. Our study does not find any evidence for this. Moreover, we find that unconditional cash transfers provide opportunities for households to shift to other nutritious options in the non-cereal segment. The BPL card primarily subsidises cereal group items such as rice and wheat. Therefore, it is in line with expectations that unconditional cash transfers instead of a BPL card would increase the intake of other nutritious food such as pulses, milk, egg, fish, meat, fruit, vegetables, etc. What is remarkable, though, is that this increase in non-cereal food items is not compensated by a decrease in calorie intake from cereal food items. Our experiment provides empirical support to the hypothesis that unconditional cash transfers to poor people may be an important instrument to improve food security.
We also observed an interesting effect of cash transfer on those who did not receive it. This is the spillover effect. In particular, we found that the service quality of PDS shops improved over the period of the experiment. Our hunch is that this happened because the PDS shops now faced "competition" from private shops as some of their clients had now shifted to private shops. However, to fully understand this interesting result, more research is needed.
At the very least, the study suggests that there could be poor households that would prefer the cash transfer compared to the PDS and the objectives of the PDS would not be compromised. This is true in at least those cases in which the poor have access to alternative sources of supply, as is the case in most urban centres.