Previous Page Table of Contents Next Page


National Food Security
Household Food Security

At the national level, food security essentially refers to the capacity of a country to provide sufficient food for its population, and it will depend on such factors as food production, imports, food aid and within-country food distribution (in India’s case, this is state-level food security). The ability to absorb shocks to such a system from drought, floods, civil strife, etc. is also relevant. At the household level, food security is here defined as access to food, that is adequate in terms of quality, quantity, safety and cultural acceptability, for all household members (Gillespie and Mason 1991). While many definitions of household security exist, they virtually all hinge on a household’s ability to get sufficient food to the household door. The two main factors determining household food security - poverty and food prices - are investigated here. The cost of achieving and maintaining food security is important and can be revealed through an analysis of the proportions of either income or expenditure that is allocated to food.

National Food Security

Food production
Drought and food distribution
Per capita food availability


Agriculture remains by far the most important area of the India economy, accounting for about 63% of the labour force in 1985-1987, although its importance has decreased over the years. In 1989 it accounted for 28.2% of GDP, compared to 42.8% in 1973, indicating low agricultural labour productivity. About two-thirds of the crops cultivated for domestic markets are grains, mostly rice, wheat, sorghum and maize.

In India cropping activities are year-round provided water is available for crops. In Northern India there are two distinct crops seasons - the kharif (July to October), and the rabi (October to March). In some parts of the country there are no such distinct seasons. Based on the crops, the country can be divided into the following agricultural regions.

i) The rice region including a very large part of the north-east and south-east of India, with another strip along the western coast.

ii) The wheat region, occupying most of the northern, western and central India.

iii) The millet-sorghum region, comprising Rajasthan, Madhya Pradesh and the Deccan Plateau in the centre of the Indian Peninsula.

iv) The temperate Himalayan regions of Kashmir, Himachal Pradesh and Uttar Pradesh where potatoes are as important as cash crops and tree fruits form a large part of agricultural production.

v) The plantation crop regions of Assam (tea) and South Indian hill regions (coffee).

Rice is grown in high rainfall areas or where supplemental irrigation is available to ensure good yields. In southern states, namely Andhra Pradesh, Tamil Nadu and Kerala, rice is grown in more than one season and mostly under irrigation or under sufficient rainfall. Together these three states have over 6.0 million hectares representing over 17% of the all-India area under rice. Important alternative crops in Andhra Pradesh are pulses, groundnut, millets, maize, sugarcane and tobacco. In Kamataka, alternative crops to rice are ragi, millets, cotton etc. In Maharashtra rice is grown mostly in the Konkan area, over 1.3 million hectares, along with ragi, pulses, oil seeds, groundnut etc. In the states of Gujarat, Jammu and Kashmir, Rajasthan and Himachal Pradesh, rice is a minor crop. Rice is grown on more than 50% of cropped area in Orissa and on 4.3 million hectares (11.7% of the all India rice area) in Madhya Pradesh. In West Bengal and Uttar Pradesh rice is grown on 14.2 and 12.4% respectively of the all-India area.

Generally wheat and gram are concentrated in the sub-tropical region in northern India, where they can substitute for each other. The core of the wheat region responsible for 70% of the area and 76% of production comprises Punjab, Haryana, Delhi, Uttar Pradesh and Gujarat in the west and Bihar and West Bengal in the east.

With the increase in population, irrigated area is increasing, and as agricultural science advances, most of the extensive cropping patterns are giving way to intensive cropping. The development in minor irrigation works has provided farmers with opportunities to crop their land all year around with high-yielding varieties. This intensive cropping requires an easy and ready availability of balanced fertilizers and plant protection chemicals and an appropriate price policy for inputs.

Food production

In the early 1970s, a global food crisis, oil price hikes, unfavourable weather conditions, droughts (especially in Western India) and severe floods (during 1974 in the North East) all combined to threaten the two pillars of India’s food strategy - a political commitment to self-reliance in food and a determination to prevent famine. Ultimately, however, these beliefs were reinforced by these challenges.

The ‘green revolution’ package of high-yielding crop varieties, irrigation and complementary inputs, which enabled farmers to increase per acre yields, has helped agriculture keep pace with population growth, and led to national self-sufficiency (see below).

At the national level to the present though, India has made rapid strides in agricultural production, with food production increasing from 108.4 million tonnes in 1970-71 to a record production of 176.5 million tonnes in 1990-91 (GOI 1991), though this has been offset by an increase in population from 551 million to 844 million during this period. The per capita production of wheat in particular, and to a lesser extent rice, increased most significantly, as the use of inputs, and the area irrigated expanded. The food production per capita index, with 1971-80 as base (= 100), indicates an increase early in the 1980s followed by a levelling off in the mid-1980s, a drop in the drought years 1986-87, then a marked increase to 1989, before dropping again in 1990 (see Figure 10).

In the early 1990s, however, there are signs that such a strategy is running into diminishing returns, as a result of decreasing fertiliser productivities and long-term extraction of soil minerals. The rate of growth of foodgrain production is said to be decelerating (Sarma and Gandhi 1990), raising the question of sustainability of growth and possible exhaustion of the green revolution potential. Furthermore, although evidence has been presented to show that the technologies involved (at least since the mid-1980s) are largely scale-neutral (Hazell and Ramaswamy 1991), despite an early bias towards richer farmers, they are still only effective if a constant supply of water and fertilizer can be assured. In the (usually poorer) villages dependent on rainfed coarse grains - such as maize, sorghum and other millets - there has been no ‘revolution’ and here lies the challenge for the future. Initial biases towards wheat and rice and towards certain states, e.g. the Punjab and Haryana, need to be balanced, and ways sought to promote agricultural development in poorer rainfed areas.

The regional development of food production has been uneven. To some extent this simply reflects differences in the crop composition and seasonality of agriculture. There have also been regional variations in performance for individual crops which have accentuated existing regional disparities. Initial gains accrued particularly to the north-western states of Punjab, Haryana and the western half of Uttar Pradesh, which were environmentally more favourable and better equipped to give high returns to the new wheat technology package. Irrigated areas of Tamil Nadu were also able to exploit the new rice technology. Much poorer performance has been experienced in the states of eastern India with predominantly rain-fed agriculture and in the dryland agriculture of the states of Andhra Pradesh, Kamataka, Maharashtra and Madhya Pradesh - states characterized by a history, of instability in agricultural production and by a high incidence of poverty. Other aspects of agrarian structure such as gross inequities of land distribution with high degrees of landlessness, more or less disguised sharecropping and other forms of tenancy and exploitation through the marketing system as well as poorly developed irrigation, drainage and flood protection and a low uptake of fertilizer may explain the particularly dismal record of rice yields in the Eastern region.

Dryland agriculture is characterized by high female participation rates, very low and highly variable yields, a limited generation and heavy concentration of control over investible surplus, generalized lack of creditworthiness, small and sporadic creation of marketed surpluses and the distress commercialization (under conditions of debt and relatively low prices) of the agriculture of small and marginal farmers. Little progress has been made in developing strains of seed with high moisture stress resistance which would be suitable for use in dryland areas. Indeed, until recently there has been a bias in research funding away from these crops.

Drought and food distribution

The 1986-87 drought was the worst since before Independence, driving the agricultural strategy off-course, and necessitating the importing of cereals on a significant scale in 1988-89 (even though the drought broke in mid-1988). The precise definition of drought is difficult to provide. As Healthcote (1973) put it, "There are probably as many definitions of drought as there are uses of water". In 1987, India experienced what was labelled as the country’s worst drought both in terms of intensity and geographical spread - 252 districts out of 412 in the country received deficient rainfall. During 1986-87, food production dropped to 143 million tonnes and then in 1987-88 to 138.4 million tonnes - see trend in Figure 10 (GOI 1991).

Figure 10. Food Production Index

The adverse impact of the drought on the country’s economy in terms of the drop in agricultural production, industrial output, reduced purchasing power and the concomitant rise in unemployment, particularly of the rural labour force, was considerable. The results of a diet and nutrition survey during the 1987 drought however showed that widespread hunger and its consequences (experienced earlier in the 1960s and 1970s) were averted in 1987.

This was achieved through the wide geographical distribution of grain buffer stocks within the Public Distribution System (PDS).. These buffer stocks had become the symbol of self-reliance during the seventies and eighties, as a food security net for the population. The Public Distribution System (PDS) involves the distribution of subsidised foodgrains via a network of fair-price shops to the eligible poor with ration cards. Figure 11 shows how public distribution rose steadily from 1975, until buffer stocks were run down during the drought of 1987-88 (to the point where imports were necessary again in 1988). While buffers stocks may be costly to maintain, the 1987-88 drought demonstrated that such a food security strategy is cheaper than importing food on a large scale. States differ as regards the coverage, efficiency and effectiveness of the PDS, with states with a better infrastructure (e.g. Punjab and Haryana) - and urban areas per se.- often being better served, despite their lower relative need vis a vis poorer States such as Bihar, Orissa and Uttar Pradesh. A detailed analysis of the Public Distribution System is provided in Part II.

India has not experienced famine on any significant scale since Independence, a period including the worst drought this century (1986-88) for wide areas of western and central India. This is historically and comparatively an impressive achievement. Post-Independence policy was formed in the aftermath of the Great Bengal Famine. The dramatic measures to prevent famine in Bihar in 1966/67, to contain the food crisis in 1973/74, as well as to feed 10-12 million Bangladeshi refugees in 1971 indicate the overwhelming priority of famine prevention. After the so-called "Bihar Famine" of 1966/67, this has also been accomplished without significant direct external assistance. The prevention of famine in India now has a political importance that in the short run will also outweigh any other consideration. There is therefore a battery of famine warning and prevention measures, representing a continuation and strengthening of earlier Famine Code practice (Clay et al. 1988).

Figure 11. Foodgrain Use - Imports, Procurement, Distribution

Source: GOI 1991
An important qualification to this success, however, is establishing and maintaining a rudimentary food security net for both rural and urban poor throughout India, has also been accomplished without significant inroads into the massive continuing problem of structural poverty.

Per capita food availability

The per capita availability of food is a function of population and food production with adjustments made for exports, imports, wastage, livestock feed, seeds and stocks at the beginning of the year.

Cereals. The per capita net daily availability of cereals has not varied significantly in the last 20 years, being 417g in 1970-71 and 438 g in 1989-90 (GOI 1991) (see Figure 12). From the early 1950s to the early 1980s, per capita foodgrain availability fluctuated within a narrow band of 150-170 kgs per year, with a statistically insignificant trend. Thus, foodgrain production growth seems to have been absorbed by population growth, reduction in imports and increase in stocks (Sarma and Gandhi 1990).

Figure 12. Net Cereal Availability - Per capita

Source 1991 GOI
Pulses. Pulses account for less than 8% of total food grain production in the country. The pulse production has not kept pace with population growth. Per capita availability declined from 51.2g in 1970-71 to 33.1g in 1987-88 before rising to 36.5g in 1989-90 (GOI 1991). This was not only due to a reduction in land under pulse cultivation but also to a decline in yield per hectare - there has been no breakthrough in technology as with wheat and rice.

Based on the 1988 population, body weights and physical activity pattern, the daily per caput energy requirement has been estimated to be 2200 kcals. (Reddy 1990). After considering post-harvest and distributional losses, the per caput requirement of foodgrain would come to 605 g/day with the total food grain requirement at production level being 178 million tonnes. The food grain production in 1990-91 is 176.5 million tonnes, close to the estimated requirement indicating virtual national-level self sufficiency in this respect at present.

Household Food Security

Food prices and expenditure

Agriculture development and food production are central to national food security. In addition, there are several other factors that determine the food security at household level. These include food prices, food availability, employment and purchasing power, government subsidies and the food rationing system.


The poorest groups in India tend to be asset-less or nearly asset-less. Landlessness is increasing, both absolutely and relatively (from 9.6 per cent of rural population in 1971 to 11.3 per cent in 1982: GOI 1988). The poor derive income from wage work in which they may be simultaneously underemployed, underpaid and forced into unwaged work. The path of diversification of the rural economy varies from region to region and is often insufficient to absorb surplus labour from agriculture. In addition, there is still a particular institutionalized poverty amongst scheduled castes and tribes, and in regions where such people are concentrated the rates of poverty reduction are also very low. Within poor households, and more extensively throughout society in northern India, women and children are particularly disadvantaged and female-headed households are amongst the very poorest (see Maternal and Child Care).

Actual measurement of poverty in the Indian context has come under debate in recent years. There are two conventional approaches: one based on average annual household income (in relation to an income poverty line defining a predominantly caloric threshold) and the other on calorie intake. Although income is widely used to measure economic welfare, it has serious drawbacks, a major one being that it may have substantial fluctuations which are averaged out in the long run. Consumption expenditure may be a better indicator of the actual economic position of a household (Kakwani and Subbarao 1990). The National Sample Surveys provide reasonably comparable time series data on the levels and distribution of household consumption expenditure (see ‘Food prices and expenditure’ section).

The Indian Planning Commission estimates the population below a national poverty line, defined in 1981 in the Sixth Five Year Plan as being the mid-point of the monthly expenditure class having a daily per capita calorie intake of 2,400 kcals. in rural areas and 2100 kcals in urban areas. The incidence of poverty (so defined) declined steadily from 51.5% in 1972-73 to 29.2% in 1987-88 - a fall of over 22% (see Table 10). Estimates by economists, however, suggest a much more modest decline of a little above 10% from 56.3 to 45.9%. The Planning Commission reports a decline in total numbers of the poor from 291.5 million to 232.4 million while other estimates suggest a substantial increase from 308 million to 361.2 million2. While both estimates suggest a decline in the prevalence of poverty, the quantum of the decline implied by the Planning Commission is far in excess of those from alternate and possibly more realistic estimates.

2 The reasons for the disparity between the two estimates lies in the different definition of the poverty line and "adjustments/corrections" applied to the NSS data. The Planning Commission draws the line at Rs 49.08 and Rs 56.64 for rural and urban areas respectively at 1973-74 prices. Further, NSS food consumption data are scaled upwards by the Planning Commission to adjust for apparent underestimations, as estimated by comparison of NSS data with those from the Central Statistical Organization. This exercise results in "pulling up" a large number of persons who would otherwise be below the poverty line - a procedure disputed by many economists. Alternate estimations of poverty use state-specific cost of living indices for the middle range of the population "falling within the range of monthly per capita total expenditure which encloses the state-specific poverty norm in the base year" (Suryanarayana, 1991). The state-specific criteria are used to yield state-specific poverty lines and head count ratios from state-specific NSS distributions.
Table 10: Percentage of population below national poverty line

















Source: Planning Commission
These considerations are important in assessing the relative success of poverty alleviation programmes in the last two decades. Furthermore, as argued by Suryanarana (1991), if crossing the poverty line is used as the sole estimate of the success of poverty alleviation programmes, administrators may be inclined to focus only on the segment of population just below the poverty line. There is therefore a need to use distributionally-sensitive poverty indices in assessing the effectiveness of poverty alleviation programmes. Some attempts at using a similar approach are reflected in Subbarao’s (1989) assessment of the changes in nutrient intakes among the poor and the "ultra poor"3.
3 Lipton (1983) takes 20% as being the irreducible proportion of income spent by the poorest in a wide range of countries on non-food essentials. Given the tendency of food requirement figures to over-estimate real food needs, especially in relation to people living in warm climates, Lipton suggests that it might be more realistic to work with 80% of the FAO/WHO weight-adjusted dietary energy requirements (FAO/WHO 1973). This leads to a ‘double 80’ rule, according to which people are judged to be ‘ultra poor’ if 80% of their income is insufficient to purchase 80% of these FAO/WHO energy requirements.
Another way of investigating change is to focus instead on the extent to which the rich and the poor in rural India have been able to increase their consumption over time. The following pattern emerges from such an exercise, reported in the national newspaper, The Hindu (Ranade December 18, 1991):

Table 11: Percentage increases in real consumption expenditure by income status









(State-wise consumption data from the NSS were deflated by the respective consumer price indices for agricultural labour in conjunction with the price indices computed by Minhas et al. (1987) for the middle and total rural population. The requisite data for 1983-88 are only just becoming available).

Ranade notes that the quadrupling of the rate of growth in consumption of the rural poor was near-synchronous with the large increase in ‘targeted’ poverty alleviation programmes in the early 1980s. Any ‘trickle down’ effect of overall growth during this period was unlikely, given the deceleration in the rate of growth of the rural rich. This also concurs with evidence provided by Kakwani and Subbarao (1990) which suggests that the improvement between 1977 and 1983 was not associated with increased foodgrain production, but rather a decrease in within-state inequality along with the likely beneficial effects of several large-scale poverty-alleviation programmes. The improvement observed later, in the mid-1980s, on the other hand was likely to have been much more driven by economic growth and favourable relative food price changes (see Figure 14). Kakwani and Subbarao (1990) went on to show that this reduction in inequality was more pronounced for the ultra-poor than the poor. Ranade however questions whether the type of poverty alleviation programmes implemented in the early 1980s could be sustained given the poor state of the economy in the early 1990s. There are also inter-state variations in the effectiveness of these programmes (see Part II).

Kakwani and Subbarao (1990) also found that between 1972 and 1983 inequality at the all-India level, is driven largely by within-state, not between-state, inequality, although this appears slowly to be changing - between-state inequality is rising. In 1983, the states could be ranked thus, in order of declining poverty (using the head-count method): Bihar, West Bengal, Tamil Nadu, Orissa, Madhya Pradesh, Uttar Pradesh, Maharasthra, Assam, Kamataka, Kerala, Rajasthan, Andhra Pradesh, Gujarat, Haryana, Punjab. In fact, this ranking is almost identical to that of ‘ultra poverty’. Looking at the map of India, poverty is apparently concentrated in the north east of the country (Bihar, West Bengal) and, apart from Tamil Nadu in the south, stretches out to cover the neighbouring states of Orissa, Madhya Pradesh and Uttar Pradesh.

Food prices and expenditure

The trends in food price indices are shown in Figures 13 and 14, based on data from the Economic Survey of India (GOI 1991). While absolute food prices have risen, their levels relative to the consumer price index have fluctuated connsiderably during the 1980s, with a slight overall declining trend (the ratio of food to consumer price indices is an indicator of a consumer’s propensity to buy food as opposed to non-food articles). A downwards trend in relative food prices is likely to have a disproportionately beneficial effect on the poorest groups, with the highest price elasticities of foodgrain demand. Mean elasticities are estimated (Sarma and Gandhi 1990) to be 0.48 for rural, 0.23 for urban, and 0.42 for national, but elasticities differ sharply by quartile, being close to 1.0 for the bottom quartiles and about 0.1 for the top quartiles.

Figure 13. Food Price Indices

Figure 14. Relative Food Price - FPI/CPI*

Source: GOI 1991
Two thirds of the total expenditure of rural households is spent on food as compared to slightly more than half for urban households (Sarekshama 1989) (see Figure 15). Between 1972 and 1987, these proportions changed with relatively more money being spent on non-food items in both rural and urban areas; the main change occurring between 1972 and 1978.

National Sample Survey consumption data indicate a small improvement in the distribution of total expenditure between 1977/78 and 1983 in both the rural and urban areas, and between the rural and urban areas. For the poor population, the data indicate that foodgrain consumption of the rural bottom quartile rose somewhat between 1970/71 and 1983, but that of the urban bottom quartile was virtually stagnant. This is associated with a small increase in the real total; expenditure of the rural bottom quartile (combined with high elasticities) but near stagnation in that of the urban bottom quartile.

Figure 15. Food Expenditure - as % total expenditure

Source: Sarekshama (1991).

Previous Page Top of Page Next Page