Despite India’s large-scale food and nutrition programs, it is estimated that 15 percent of India’s population is currently undernourished. A recent FOODSECURE working paper analyses consumption patterns of food grains delivered through the Public Distribution System (PDS) in India and investigates targeting errors as well as reasons for leakage and self-selection of staples using cross-sectional household data throughout India.
Through the PDS, the government provides food grains at subsidized prices to those with different income levels through three types of ration cards: AAY (the poorest people in India), BPL (those just below the official poverty line) and APL (those above the poverty line). The amount of grains available under each ration card varies per state, generally and officially, the per capita ration is: 35 kilograms per household in all states for those with the AAY ration card, 35 kilograms for those with a BPL ration card in most states, and below 35 kilograms for those with an APL ration card.
Regarding methodology, the study is based on the 68th round of the Indian National Sample Survey of Household Consumer Expenditure. The survey includes 101,651 households in 7,469 villages across India. Data was collected between July 2011 and June 2012 and it covers demographic data and household characteristics, as well as consumption quantity and value, total consumption expenditure, and PDS consumption, quantity and expenditure. Through a series of probit and tobit models, the study estimates the impact of the PDS on staple consumption for different ration card holders.
Based on the data collected, the study estimates that 80 percent of Indian households possess a ration card: only 5.5 percent of these have the AAY card, 38.5 percent the BPL card and 56 percent the APL card. Around 90 percent of the AAY and BPL ration card holders are found to use their cards, whereas APL cardholders predominantly do not use their cards (74 per cent). Those with an AAY ration card buy close to the full ration (averaging 30.4 kilograms per household). The study finds that those with BPL and APL ration cards consume significantly less (averaging 21.9 kilograms and 15.9 kilograms per household respectively). Moreover, around a third of AAY card holders do not buy any grains on the market, which means they consume exclusively from the fair price shops. This suggests that the poorest rely on the PDS much more than the non-poor. However, the study finds that 60 percent of AAY, 44 percent of BPL and 23 percent of APL ration card holders’ households monthly per capita expenditure falls below the official poverty line. The study highlights that these findings indicate significant targeting errors in the allocation of ration cards (for instance, that households with an APL ration card should actually have a BPL ration card).
The study uses a probit model to explain the reasons for the relatively low take up of grains through the APL ration card program, the study does this to investigate whether the low take up occurs due to leakage (supply driven) or due to self-selection (demand driven). The study finds no evidence of leakage having a deciding role on the low take-up observed in the APL group. Higher incomes and education levels can be associated with lower probability of APL take-up, which is probably due to negative self-selection (that they chose not to purchase the PDS grains) of the more affluent households. The study suggests that this is probably due to the inferior quality of PDS grains.
The study also uses a tobit model to estimate the quantity consumed from the PDS, separately for wheat and rice, across all ration card holders and different social groups (caste, migrants and female headed households). In this case the study finds that higher household expenditure in general has a significant and negative impact on consumption of both PDS wheat and rice. More specifically, the study estimates that an additional thousand rupees household monthly expenditure decreases, on average, a household’s consumption of rice by 1.2-1.4 kg and 0.5-0.6 kg of wheat. Again, the study suggests that this indicates negative self-selection by richer households. Additionally, the study finds that, for all ration card holders, migrant workers and female led households are not well covered by the PDS, and that they generally under-purchase from the PDS as compared to male dominated households. This suggests that there are some serious targeting errors in the PDS and the paper suggests that this may be related to the marginalization of female led households.
Based on these findings the study discusses a number of implications. Firstly, the paper suggests that more diversified rations, including for example pulses, eggs and vegetables in the PDS could be highly beneficial for the improved nutrition of the poor as they appear highly dependent on the PDS. Secondly, the results suggest that implementing cash transfers to the poor (as is being discussed), instead of providing subsidized grains, could improve the functioning and targeting of the PDS. This is because it would allow better targeting to the poor and poorest households and prevent negative self-selection by households.
In conclusion the study highlights two main limitations. Namely, that the data used in the study only provides limited information on household characteristics and institutional differences between districts (such as the distance to the PDS shop) which could have a significant impact the consumption patterns of households. Secondly, as the analysis used data from across India it does not account for differences in local cultures and production systems across regions and states. The study recommends that further research is undertaken on state and district levels.
The full working paper can be accessed here.
By: Bas Paris