Is the PDS Already a Cash Transfer? Rethinking India’s Food Subsidy Policies

Abstract

Critics argue that India’s mismanaged Public Distribution System (PDS), which sells subsidised cereals to poor families, should be replaced by cash transfers. Others fear cash may be misused. Using National Sample Survey data, this article demonstrates that families treat additional PDS subsidies wholly as a source of cash – exactly like a cash transfer. More worryingly, cereal consumption has not increased, despite higher real subsidies. Moreover, neither the PDS nor cash transfers are likely to raise total food expenditure in poor families. Finally, therefore, the paper explores how higher food consumption and other objectives of PDS subsidies may be achieved.

Acknowledgements

The author is grateful to Jaideep Sengupta for valuable comments on this work. Data used in the paper will be made available on request.

Notes

1. Throughout the paper, the terms ‘foodgrains’, ‘cereals’ and ‘grain’ are used inter-changeably.

2. However, this is likely to change after the passage of the new National Food Security Bill in September, 2013.

3. The NSS reports contain both the quantity in kilograms and the value in rupees of purchases of rice, wheat and all/total cereals by households. Dividing the value by the quantity of purchases provides a ‘unit value’ or average price paid by households per kilogram of rice, wheat or total cereals. The use of unit-value prices for cereals is unavoidable since the official CPIAL and CPIIW reports have only recently started providing a sub-index of cereal prices – for earlier periods, cereal price indices are not available. Also, the CPIAL and CPIIW reports do not provide prices for rice and wheat separately, but group them together with other cereals.

4. The difference between the market price and PDS price of rice multiplied by the quantity of rice purchased by a household from the PDS, represents the nominal income transfer due to rice purchases; the calculation for wheat purchases is identical.

5. Note that the growth rates of per-capita income based on these MPCE figures are generally acknowledged to be lower than the estimates of per-capita income growth arrived at by using the national income accounts (see Deaton & Dreze, 2009).

6. Therefore, although the actual composition of households comprising the bottom four MPCE deciles may have altered somewhat during the 11-year period, this change in composition should not have reduced cereal consumption but actually have improved it, since households in every decile were richer by the end of the period.

7. ‘Total cereal consumption’ here refers to consumption of rice, wheat and all other cereals, bought or received from any source, including the PDS.

8. Note that for conditions 3, 4 and 5 and the related variables, the tests of significance are not relevant since the changes in these variables have been in the opposite direction to that expected. For instance, the analysis concludes that total cereal consumption in poor households has not increased (condition 5). Since the cereal consumption figures show a decline, non-significance would simply suggest that this decline was insignificant and that total cereal consumption had stayed constant – which is exactly what the analysis claims. Similarly, condition 3 only requires that relative cereal prices do not rise (the figures show a fall), and condition 4 that real incomes do not fall (the figures show a rise).

9. The figures for total rice and wheat consumption (from all sources, including the PDS) are not shown in Tables 1 and 2. However, the analysis shows that the total quantity of rice and wheat consumed also declined between 1999–2000 and 2009–2010 for seven out of the eight deciles considered here. Even in the first rural MPCE decile, in which the total consumption of rice and wheat rose over this period, this rise was more than offset by a fall in the consumption of other cereals, resulting in a decline in total cereal consumption.

10. Note that the pattern of per-capita calorie consumption in poor households may differ somewhat from the pattern of real food expenditure due to changes in the sources of calories. A shift from more expensive to cheaper sources of calories, for instance from fats to cereals, might cause per-capita calorie consumption to rise even if real food expenditure per capita was falling.

11. Rural households too are increasingly relying on the market for cereal purchases. In 2009–2010, the ratio of home-grown produce to total consumption among rural households was just 25 per cent for rice and 37 per cent for wheat (National Sample Survey Office, 2013). Moreover, these figures represent the average across all MPCE deciles – for the poorest rural households, who are least likely to own land, the share of home-grown produce in total consumption may be considerably lower.

12. White and Masset (2007) suggest that the Tamil Nadu Integrated Nutrition Project is a successful example of such an integrated programme.