Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Distributional aspects of caloric consumption: implications for food policy

R. Radhakrishna. Department of Economics, University of Hyderabad, Hyderabad, India

Policy implications


A major thrust of this paper is to develop a quantitative model for linking nutritional intake (calories) with agricultural production and prices. The population has been stratified by regions (rural and urban), and expenditure classes and demand models for calories have been estimated for each group. The redistributional effect on calorie intake has been quantified under alternative assumptions concerning the supply. The effects of fluctuations in agricultural production are quantified by linking land holding distribution with income distribution. Implications of utilization of agricultural surplus are traced.


Although enough food items are now produced to provide every person with an acceptable level of caloric consumption, nearly half the population suffers from caloric deprivation, and this is alarmingly large for the bottom strata (Dandekar and Rath 1971; Rao and Vivekananda 1979). In fact, in periods of upswings in grain production, stocks accumulate even before the food deficiencies of the poor are met. The real problem is that hunger is overwhelmingly the result of income inequality and the lack of purchasing power among the lower strata of the population. Viewed thus, the food problem would appear to be mostly a result of distribution, and the solution to it would not merely lie in producing more food.

The paper analyses the calorie intakes of various strata. It examines (a) the extent of caloric deprivation among the bottom classes; (b) the effects of income and prices on calorie intake; and (c) the implications of the results for policy priorities. Since distribution is one of the determining factors of caloric deprivation among the poor, the analysis is focused on expenditure classes. We present in the next section the basic model employed and in the last section the results.


Demand Model

It is postulated that the demand patterns of any group can be approximated by an extended version of the Linear Expenditure System given in Nasse (1970). The Nasse model has the form:


(i = 1, ......................., n)

(t = 1, ......................., T)


E() = 0

E(') =

where pit and qit represent the price and quality of the ith item in period t; mt is the total expenditure in period t; bI and Cij are parameters and is the stochastic term.

The Nasse model fulfils the adding-up property if and Slutsky symmetry conditions if and only if Cij = Cji. Fulfilment of the second order conditions imply that

Expenditure elasticities are given by


(I = I, ...................., n)

Direct and cross price elasticities are given by:



Caloric Consumption Function

The quantities consumed at a given level of expenditure and prices can be obtained from equation (1) and the calorie intake can then be estimated by converting the quantities into calories. Let kj be the quantities of calories contained in one unit of the ith item. The caloric consumption function is given by


Ignoring the time subscript and rearranging the terms, equation (4) can be written as


The marginal propensity to consume calories is given by in equation (5). Caloric elasticity with respect to total expenditure is given by


The caloric elasticity with respect to the jth price is given by



Murty and Radhakrishna (1981) have provided, for India, parameter estimates for the Nasse model computed from the data on consumer expenditure provided by the National Sample Survey Organization (NSSO) for the rounds starting from number 2 (1952/53) to number 25 (1970/71). Six of the nine commodity groups considered by the model belong to food, and the model has been estimated separately for five expenditure classes in rural and for five in urban areas. Some of the details of the expenditure classes are given in table 1.

TABLE 1. Percentage Distribution, Per Capita Expenditure. and Caloric Consumption by Size of Class

Class intervals
(rupees per capita
per month)
of population
during 1970/71
Monthly per capita
expenditure during
1970/71 (rupees)
Per capita
caloric intake
(kcal per day)
1 0-18 16 14 13 1,461
2 18 28 31 22.98 1,978
3 28-43 32 34.51 2,454
4 43-75 17 53.43 3,268
5 75 and above 4 121.93 4,454
All 34.37 2,423    
1 0-18 4 15.00 1,261
2 18-28 17 23.59 1,590
3 28-43 30 35.19 2,021
4 43-75 31 55.49 2,503
5 75 and above 18 120.89 3,099
All 54.64 2,239    

TABLE 2. Calories Contained in One Rupee Worth of Expenditure on Commodity Groups during 1970/71 (kcal)

Commodity groups Rural classes Urban classes
1 2 3 4 5 1 2 3 4 5
Cereals 4,410 3,840 3,480 3,480 3,330 3,780 3,330 3,180 2,940 2,730
Milk and milk products 1,290 810 810 780 750 690 690 720 660 630
Edible oils 1,440 1,620 1,620 1,590 1,470 1,650 1,860 2,010 2,040 1,740
Meat, fish, and eggs 690 570 510 480 360 810 630 420 390 300
Sugar and gur 3,390 3,210 3,090 2,940 2,550 2,700 2,610 2,550 2,460 2,310
Other food 1,680 1,740 1,680 1,620 1,440 2,160 1,590 1,710 1,560 1.170

The following procedure has been adopted in working out the calories contained in one rupee worth of expenditure on each of the six food commodity groups. For 1961/62, NSSO provides quantities consumed and their values at a fairly desegregate commodity level for a number of expenditure classes. These quantities have been converted into calories, and then the calories contained in one ropes worth of expenditure on each commodity group is estimated. The expenditures are at 1961/62 prices, and using commodity group-specific price deflators, the calories contained in one rupee worth of expenditure at 1970/71 prices have been estimated.

These estimates exhibit a few broad features of interest. First, the cereal group is the cheapest source of obtaining calories, and the meat, eggs, and fish group the costliest. Roughly, a rapee expended on cereals yields the same quantity of calories as that of Rs 2-2.5 expended on items belonging to the "other food" group, and Rs 5-10 on meat, fish, and eggs. Second, a rural consumer derives more calories from one rupee spent on any commodity group than his counterpart in urban areas. Third, calories contained in one rupee worth of expenditure on any commodity group decline as one moves from bottom to top classes. For instance, from a ropee spent on cereals, the bottom classes derive 4,410 kcal and 3,780 kcal respectively in rural and urban areas, while the corresponding figures for the top classes are respectively 3,330 and 2,730 kcal. Thus within each commodity group, the upper strata replaces calorie-intensive items by items with a lower calorie content.

Table 3 shows the shares of various commodity groups at mean level, which is obtained by relating the quantities estimated from the demand models to the results given in table 2. The cereal group is the major source of calories for all classes both in rural and urban areas. It accounts for 87 per cent of the calorie intake by the bottom rural class and 67 per cent for the rural top class, while the percentages for corresponding urban classes are 77 and 26. the "other food" group is also an important source of calories for the higher classes, particularly in urban areas. Only in the top urban class does one notice some diversification in the sources of calories; otherwise, cereal and other food groups account for most of the calorie intake.

TABLE 3. Percentage Distribution of Calorie Intake by Sources in Bottom and Top Classes


Percentage distribution of calorie intake


Rural bottom class (class 1)

Rural top class (class 5)

Urban bottom class (class 1)

Urban top class (class 5)






Milk and milk products





Edible oils





Meat, fish, and eggs





Sugar and gur





Other food





Making use of parameter estimates of the demand models and the results given in table 2, calorie consumption function, i.e. equation (4), has been estimated for each expenditure class. Per capita calorie intake during 1970/71 has computed at mean level from the calorie consumption function separately for each expenditure class (see table 1). The per capita per diem calorie intake works out to be 2,423 kcal, for rural areas, 2,239 kcal for urban areas, and 2,394 kcal for India as a whole. Clearly, though at the aggregate level there is no caloric deficiency, deprivation is observed for the bottom two rural and bottom three urban classes if we take the norm as 2,250 kcal. The extent of deprivation is around 35 per cent for the bottommost rural class and 44 per cent for the bottommost urban class. Since the expenditure classes considered by the demand models are broad, and the population distribution is also not uniform for rural and urban areas, neither the percentage of population affected by calorie deficiency, nor rural-urban comparisons of the intensity of deprivation, could be deduced. However, an earlier study has shown that the percentage of population affected by calorie deficiency is higher in urban than rural areas (Rao and Vivekananda 1979).

Table 4 provides the average propensity to consume calories by size of classes obtained by dividing the mean calorie intake by the corresponding mean expenditure. It can be seen that the average propensity to consume calories vary markedly across the classes. For instance, it works out to be 3,044 kcal for bottommost rural class and 1,228 kcal for the topmost class, while for the bottom and top urban classes it works out to be 2,519 and 769 kcal respectively. The low propensities observed for higher classes is obviously due to their higher propensity to consume non-calorie food items.

TABLE 4. Average and Marginal Propensities to Consume Calories (expressed in kcal per rupee)




Expenditure classes

Average propensity to consume calories

Marginal propensity to consume calories

Average propensity to consume calories

Marginal propensity to consume calories


























TABLE 5. Caloric Intake Elasticity with Respect to Prices and Total Expenditure

Caloric elasticity with respect to the prices of: Caloric
with respect
to total
Cereals Milk and
fish, and
and gur
Clothing Fuel
1 - 0.827 - 0.005 - 0.028 - 0.016 - 0.020 - 0.079 - 0.004 - 0.032 - 0.002 1.007
2 - 0.716 - 0.020 - 0.033 - 0.012 - 0.028 - 0.119 - 0.002 - 0.024 - 0.003 0.957
3 - 0.534 - 0.078 - 0.041 - 0.024 - 0.044 - 0.128 - 0.003 - 0.024 - 0.002 0.879
4 - 0.246 - 0.069 - 0.031 - 0.009 - 0.055 - 0.084 - 0.027 - 0.024 - 0.048 0.595
5 - 0.240 - 0.036 0.030 - 0.014 - 0.023 - 0.129 - 0.020 - 0.011 0.002 0.442
1 - 0.715 - 0.005 - 0.037 - 0.040 - 0.025 - 0.172 0.002 - 0.007 - 0.006 1.005
2 - 0.569 - 0.015 - 0.049 - 0.024 - 0.018 - 0.126 - 0.001 - 0.024 0.003 0.862
3 - 0.310 - 0.088 - 0.077 0.005 - 0.082 - 0.286 - 0.006 - 0.030 - 0.015 0.880
4 - 0.169 - 0.029 - 0.053 0.005 - 0.041 - 0.178 - 0.013 - 0.023 - 0.049 0.561
5 - 0.076 - 0.083 - 0.056 - 0.016 - 0.038 - 0.281 - 0.020 - 0.011 - 0.057 0.505

The marginal propensity to consume calories computed at mean levels varies substantially across the classes; the low values obtained for rural and urban top classes are striking (table 4). The caloric elasticity with respect to total expenditure (TIC) also exhibits a more or less similar pattern. The caloric elasticity is unity for bottom classes implying that calorie intake expands proportionately with total expenditure; and it declines to about half for top classes implying that a 1 per cent increase in per capita expenditure results in a half per cent increase in calorie intake. Among price effects, cereal price has a strong influence on calorie intake and its magnitude declines as we move from lower to higher groups (table 4). For instance, a 1 per cent rise in cereal price will result in a cut of 0.8 and 0.7 per cent in the calorie intake by the bottom rural and bottom urban classes respectively and the corresponding percentages for the top classes are only 0.2 and 0.1. Thus, per capita expenditure and cereal price mostly determine the calorie intake level. Only in the case of upper classes does price level of the "other food" group have some influence on calorie intake.

Policy implications

It does seem that the total quantity of food utilized for domestic consumption can provide the minimum calorie requirement. Despite this, a large section of the population is deprived of the minimum. The real problem appears to be the distributional imbalance in the cereal consumption, since cereal is the major source of calorie intake. It is likely that it will continue to be a major source for some time to come at least for the lower strata.

What then are the policies to alleviate the caloric deprivation of the poor? The redistribution strategy of limiting the consumption of the upper income groups and transferring the purchasing power to the poor has received wide attention in Indian plan formulations. Simple calculations using the caloric elasticity would show that in order to eradicate caloric deficiency, it would be necessary to increase the per capita expenditure by 54 and 14 per cent for the bottom first and second rural classes respectively, and by 78 per cent, 48, and 13 per cent for the bottom first to third urban classes respectively, through slicing the expenditures of the higher income groups. However, such consumption transfers would result in additional demand for cereals, since the cereal group takes a comparatively greater share in the marginal rupee spent by the lower classes. If the excess demand for cereals is not eased through supply expansion, cereal price tends to rise, and the impact of the price rise would be heaviest on the poorest sections of the population not covered by purchasing power transfers (see Murty and Radhakrishna 1981).

It is thus clear that meeting the caloric deficiency by consumption transfers presupposes stepping up of the aggregate cereal consumption. On the other hand, a number of schemes have been proposed for nationwide public distribution that are workable with the existing supply. The implementation of the schemes is still a hope. It must be pointed out that public distribution as tried earlier was essentially designed to serve through rationing the requirements of the cities and industrial centres that cover also the rich class. The rural areas and their vulnerable sections had little access to the public distribution system. Such inadequate coverage would be worse than no rationing at all, since the vulnerable population outside the rationed areas may then be without assured supplies.


Dandekar, V.M., and N. Rath. 1971. "Poverty in India: Dimensions and Trends," Economic and Political Weekly, pp. 25-48, 106-146.

Murty, K.N., and R. Radhakrishna. 1981. "Agricultural Prices, Income Distribution and Demand Patterns in a Low-income Country." In Robert E. Kalman and J. Martinez, eds., Computer Applications in Food Production and Agricultural Engineering. North-Holland Publ. Co.

Nasse, P.S. 1970. "Analyse des Effets de Substitution dans un Systme complet de fonctions de demande." Annales de l'Insec; 5: 81-110.

Rao, V.M., and H. Vivekananda. 1979. "Food Problem and Policy Priorities." In C.H. Shah, ed., Agricultural Development of India, pp. 178 198. Orient Longman, India.

Contents - Previous - Next