Designing appropriate systems


back to issue

IT is possible to build a lean and mean efficient system such that hunger can be abolished in India today. A nation with a high growth rate, recognized software skills, billions in exchange reserves and grain stocks bursting at the seams can certainly do it. Unfortunately, the critics, drawing upon the grandiose public supply scenarios proposed by the proponents of the National Food Security Bill, have painted a cataclysmic picture. Imagine that 75% of the rural population and 25% of the urban population start getting three kilograms of food grains, and possibly seven kilograms as soon as the curtain rises. This translates into astronomical quantities of food grain and money. Though more than 60 million tonnes is normally the figure bandied about, the really bizarre numbers even go up to a mechanically calculated 100 million tonnes.

The number of thousands of crore the budget will thus require depends on which end of the quantity scale you favour; in any case the figure is very high. The economy, already in deceleration, goes into a tailspin. Also, already benefiting from employment guarantees, rural labour households will stop working and agriculture will suffer a lethal blow. Experienced economists and political leaders give the story great credibility. Unlike the normal discussion of ‘reforms’, underlining difficulties and roadblocks in implementing government policies, there are no shades of grey and no real solutions proposed in this discourse.


Reality will in all likelihood unfold differently. The legislation will take time. States with higher legal entitlements and parties which represent them, perhaps knowing that the actual benefits are in fact much lower, will play act at championing the underdog. Others will take their cue from the ‘reformers’ and ask for cash transfers instead. All this will take time. The more important reality is that even in states with universal coverage of the public distribution system (PDS) and subsidized grain, which is all that the present National Food Security Bill is about, the actual offtake is normally not more than 40%, if deficit and surplus states are taken into account – in deficit states this could be higher by half and in surplus states similarly lower.

Our finance ministry is conservative and it is extremely unlikely that the expenditures demanded, based as they are on theoretically derived numbers, will be approved. There is, however, a more compelling fallacy in the big numbers. Conventionally, per capita rations are multiplied by eligible families to estimate the requirements. But grain preferences are going down; even poor households in India consume more non-grains, and not grains, as their income goes up. The income elasticity for grain is less than a fifth of one per cent and those for non-grains between one to two per cent for poor families. If you take that into account, grain offtake will be much lower than talked about. I asked an economist friend who wrote a report on this and he said that he used National Sample Survey (NSS) estimates, but not behavioural trends. His analysis reveals that not more than two-fifths of the eligible families with the kind of numbers talked about will actually draw from the PDS.

Actually this leads to a major hiccup. The ‘special groups’ in the legislation, lactating poor mothers, the poor girl child, the women-headed households and so on, the ones who really deserve to be covered, are left to the discretion of the states. Now I know that many economists are keen to federalize the legislation in the name of ‘efficiency’. But the states have not quite covered themselves with glory on these counts, particularly the fast growing and high income ones. It is not for nothing that in most countries this kind of legislation has a strong federal or central government imprint. In our country, the most important priority is being undermined – the need to plug leakages. It will not cost much, but is important.


The argument that declining availability of labour for agriculture will become a stumbling block to growth too is overstated. There is overwhelming historical evidence that rising agricultural wages lead to technological progress in agriculture, resulting in much better use of land and other scarce resources. Union Minister of Agriculture, Sharad Pawar, well informed on the subject, was spot on in highlighting it a few months ago. He pointed out that decentralized mechanization, not just tractors and so on, is the way forward. One would be badly advised to raise the issue of labour shortage now. All in all, a better fed labour force will be an asset and it is high time to make a beginning.

A meaningful food security programme needs to be grounded in consumer behaviour. The poor are not just objects to be stuffed with stored (rotting?) grain. The income elasticity of demand even in the seventies was low for the poor and remained at less than 0.3 even in the nineties. But for commodities like milk and milk products, eggs and meat, edible oil and sugar, the estimates of expenditure elasticity are high for poor households, in some cases above two, while they are below one (although not very low) for the non-poor. There is voluminous literature on the declining consumption share of grains by poor households in India.


The literature on food security has been given an immediate focus in policy by the welcome inclusion of abolition of hunger as an objective by the UPA government in its short run policy agenda. This has led to two kinds of pressures on food demand exercises. The first is to raise the bar on poverty levels by the state governments and others. The other is to follow recent global work which tends to argue that almost the entire Indian population is poor.

The debate on poverty is central to food security. Some swear by global estimates. The fact that poverty estimates are significantly affected by price adjustments at the state level and not so much by consumption differences, has been documented by the data mining exercise by S. Gangopadhyaya and Amaresh Dube. In fact, when the price indices are purchasing power parity (PPP) prices for defining poverty lines, as in World Bank estimates, instead of price indices for each income class in rural and urban areas, you get the ‘questionable’ result that almost the entire Indian population is below the poverty line.

As a student I had worked on the original purchasing power parity estimates (Irving Kravis was my teacher), and know that PPP estimates for the country, even after being adjusted for rural/urban price differences, do not capture the consumption pattern of the poor. The work that R. Radhakrishna and Atul Sarma did at the Sardar Patel Institute in the late sixties, of estimating price indices separately for the rich and poor in rural and urban areas in India, gives robust results of a kind not seen elsewhere. It gave the Indian poverty estimates a robustness which earlier norms like Rs 20 per month did not. Incidentally, there is no available record of the calculations behind this Rs 20 norm, or for that now stated as Rs 32 per day. Professor Pogge, presenting the Rajiv Gandhi Memorial Lecture, 2011, gives the global correlates of this deficiency in an extremely forceful manner.


The Tendulkar Report has two merits and a shortcoming. It gives up the old official poverty line, but not wholly so. The new line is the urban line within the old poverty line, sometimes called the Alagh Poverty Line, since in the late seventies it was laid down by a Task Force of the Planning Commission that I chaired. This is now mandated as the new poverty line for both rural and urban areas. Though I have been demanding a fresh computation for two decades, since India has moved on from the seventies, all my efforts failed and the many committees set up for the task declined to follow up on the suggestion. Tendulkar claimed to have done so, but he went only part of the way. There is, however, an advantage in the method he suggested. Some indicators of malnutrition map on elegantly as a statistical distribution onto the new Tendulkar expenditure poverty line, again seen as a distribution.

The old line went from calories required, depending on sex, work status and so on, to the poverty line. The emphasis then was on grain, although it was known that some calories came from non-grains as well. Now if the objective is to ensure food, it is crucial to keep in mind that food as Tendulkar points out is not just grains any more. Tendulkar and Radhakrishna are good at this since they were a part of all the earlier exercises. Its great merit is that if we choose to move away from the political arguments of kilos of grain and rupees per kilo to severely undernourished mothers and the girl child, we now have the tools to do so. There are many issues that are debatable, but we have to begin somewhere.


Living as we did with the memory of ship to mouth food supplies, the concerns of emerging India in this century could not be visualized and it is futile to paste them on a tattered 1979 poverty line. The experience since 1991 is that reform by stealth failed in substantial measure. The Tendulkar Committee has made us cross that rubicon. Recognizing it now requires that we articulate the concerns of different sections of our people in a definable and contestable manner in the design of reform. This generation is doubly blessed for it has to once again reinvent a desirable future. It is futile to dodge this issue and the statistical machinery will have to come to grips with it.

The more important issues, however, relate to designing and managing the system. Coordination between different agencies and a zen for outcomes, i.e., abolishing hunger, will have to be the driving force, rather than proposing/advocating a gargantuan state within the state. We have failed to work on the problem of decision making tools for human security, since this is a new area, needing fresh work, and limiting ourselves to food security is more manageable. It will be the base of a management system. The national decision making level will be pivotal for food security, since critical decisions will have to be taken at this level. Some of the key issues on which we will have to decide on are indicated below.

1. Purchase of food items in international markets, or establishing access, for example through ‘future’ markets, or recourse to bilateral or multilateral agencies, for example ‘food aid’ or ‘cereal facilities’.

2. Decisions on adjustment of domestic stocks through national policies. These may include purchase or sale of public stocks and attempts to influence private inventories, as also the related questions of desired levels of domestic prices of food items.

3. Use of support prices, tariff mechanisms and domestic taxes, restrictions and subsidies.

4. Optimal internal stock movements and the related question of domestic availability and price spreads in regional markets.

5. Access and vulnerability, the question of classes of consumers, for example in mofussil areas, or categories like women, children, the unemployed and the destitute or disabled.

6. Short-run decisions relating to financing, credit and foreign exchange requirements of operational food policies.

7. Decisions with a medium term horizon like assessment of food demand, incentives and support policies for domestic producers, development of improved processing and marketing infrastructure, standardisation, nutritional and quality aspects, and employment and income supplements for marginal populations and areas.


It is obvious that strengthening the national decision making systems will be required at core economic policy making centres relating to food security, for example the cabinet, finance ministry, planning ministry and civil supplies, rural development and agriculture ministries. Second, the same will be required for sectoral levels, for example, health ministry and local government ministry. Finally, national level parastatals and institutes may require such information, such as the central bank, agricultural bank and nutrition and agricultural technology related institutions.

It needs to be noted that while the decisions required and agencies taking them have been listed separately, the information system itself will play a coordinating and integrating role. A decision to import food, for example, has to be based on simultaneous consideration of expected international and domestic prices and availability, both with government and likely behaviour of private trade. While the output for decision makers in charge of the decision support system will be a select list of indicators required by them, these will have to be integrated in processes of reasoning and analysis.


At the state level, the requirements of information for making decisions will be more direct and pressing. These will include data of a continuous kind on prices and quantities marketed in agricultural markets and retail prices by location. If base level models show, as is likely, a strong relationship between local food production and local prices, then estimates for area, production and yield and forecasts for the current agricultural season will be required. Data on government stocks and improvements will be needed.

If possible, an effort should be made to develop indicators of private stocks. Direct data from trade channels is a possibility. Alternately, stock movements or market arrival or turnover data may be used. If none of this is available, estimates may have to be inferred from price data and movements in it, for which a back-up relational model may be required. At the market level, it will be important to standardise figures in terms of quality, weight and related considerations. Quality noise can distort price data in a significant manner.

On the demand side, benchmark analysis of household budgetary surveys of the kind conducted by statistical survey agencies or the nutrition survey of the kind undertaken by the National Institute of Nutrition will be used to prepare indicators of ‘vulnerable’ or ‘at risk’ segments of the population. Their magnitude and location will have to be constantly updated. Three kinds of broad requirements can be indicated to outline the nature of work requirements for information system for decision support. Each would need considerable operational detailing. The first would relate to expansion of supply potentials, the second to marketing, trading, distribution and information infrastructure, and the third to improvements in access to food by different segments of the population.


Marketing information and distribution infrastructure relates to food security by ensuring access to available supplies in an efficient manner. Regional spreads in food prices emerge from market imperfections. Generally, regions of food deprivation or access difficulties, i.e., upper hills and rural areas outside the valleys and deltas, are also areas of poor communication and agriculture marketing infrastructure. Improved information on food prices is the first step, but wider and more efficient markets are important. Modern communication and data networking facilities should be used not only to collate, analyse and transmit data from rural markets, but for interactive communication as well.

As the rural economy diversifies, there is every reason for the peasant living in the upper hills or other backward areas, for example, to have available data on prices of crops to be considered for sowing, in other regions and in other countries, with which his output will compete at harvest time. Thus, extra superior long cotton prices in the Philippines or Pakistan may also be an important input for his decision.

Information systems will of course have to be built up on storage facilities, stock and movement of public stocks of food items. Efficient deployment of such stock through regions and time is important from the cost saving angle of public interventions. Inventory management under uncertainty and optimal transportation models become important here – as also improved storage processes. Seasonal aspects of the crop season and post-harvest assessments of stocks and prices are generally necessary for sound policy decisions.


Problems like iron deficiency in women, female rural illiteracy, low nutritional levels in some areas, and emerging trends of decline in protein/energy intake in some population segments hit by adjustment, may need to be isolated and reversed targeting of deliveries and social adjustment programmes of short run employment and income generation may need to be developed. Data bases which will provide labour market trends, information on successful community intervention strategies and filtering data mechanisms for isolating the truly needy populations at the household level, will be required. Information will also have to be planned for taste patterns, food storage, cooking and consumption habits and local technologies, either traditional or new, with application possibilities for resolving food gaps in an efficient manner. Delivery systems which work, particularly market channels integrated with community efforts or public objectives, will need to be actively searched for.

All of this may sound foreboding, but it really isn’t. The real issue is to get going with a clear vision and to learn from mistakes. A friend and well-wisher once told me, ‘Nothing ever works in India, Yoginder.’ I agreed, but had to say, ‘But nothing ever fails in India also.’