Ind. Jn. of Agri. Econ.

Vol.68, No.4, Oct.-Dec. 2013

ARTICLES
Unravelling Food Basket of Indian Households:
Revisiting Underlying Changes and Future Food Demand
S.K. Srivastava*, V.C. Mathur**, N. Sivaramane, Ranjit Kumar***,
Rooba Hasan†, P. C. Meena***
ABSTRACT

The study empirically revealed striking difference in the consumption pattern of Indian households
across rural and urban sectors, geographical regions and income categories. The findings were in
conformity with Engel’s law and Bennett’s law of consumption. Expenditure elasticities of food
commodities were estimated using LA-AIDS demand system and food demand was projected for the year
2020. The expenditure elasticities of high value agricultural commodities (HVACs) like milk, non-
vegetarian products, fruits, etc. were higher than staple food, i.e., cereals. Further, wide inter-regional
variations in the household demand for food commodities necessitates to match the demand and supply at
disaggregate level and to remove the bottlenecks in production of food commodities in the respective
region to fulfil the demand.
Keywords: Consumption pattern, Food commodities, Expenditure elasticities, Future household
demand.
JEL: Q11, Q18, C42.

INTRODUCTION

In the last few decades, Indian economy has undergone a significant structural
change leading to increased employment opportunities, steady growth in income,
improved access to quality food, increasing urbanisation, etc. The immediate impact
of these changes falls on the demand and the composition of food basket, particularly
for the households who are at the bottom of the pyramid. Several scholars have
studied the increasing diversification of food basket and changing dietary pattern of
the consumers (Kumar and Mathur, 1996; Paroda and Kumar, 1999; Dastagiri, 2004;
Keyzer et al., 2005; Mittal, 2006). In the developing country like India, increasing
income and its impact on food consumption needs a special mention, where a major
portion (more than 50 per cent) of income is spent on food products.

                                                            
*Scientist (Agricultural Economics), National Centre for Agricultural Economics and Policy Research, New
Delhi-110 012, **Professor, Division of Agricultural Economics, Indian Agricultural Research Institute, New Delhi-
110 012, ***Senior Scientists, National Academy of Agricultural Research Management (NAARM), Hyderabad-500
030 and †Assistant Professor, Rajiv Gandhi South Campus (Banaras Hindu University), Barkachha, Mirzapur (Uttar
Pradesh), respectively.
The authors are grateful to the anonymous referee for his constructive comments and suggestions on the earlier
draft which helped them to improve the paper.
 
536 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

The Indian economy has grown by about 7 per cent per annum during 1990-91
and 2009-10. However, the growth has remained uneven across different sectors of
the economy. The agriculture and allied sectors, which provides employment to 52
per cent of the workforce and contributes 14.2 per cent of gross domestic product
(GDP at 2004-05 prices), could grow only by 2.7 per cent per annum during the same
period. Sluggish growth in agriculture and allied sectors has aggravated the disparity
in the income and thus consumption pattern of rural and urban households in the
country. The Report on the Status of Food Insecurity in Rural India (MSSRF, 2008)
indicates that in 2004-05, about 13 per cent of the rural population in India consumed
less than 1,890 Kcal per consumer unit per day. In the hindsight of skewed income
distribution, the consumption pattern and therefore nutritional status vary widely
across income categories. Besides, the food habits of the Indian households also
differs across different regions because of large cultural diversity and consequent
differences in food histories across different regions of the country (Minhas, 1991).
It is thus pertinent to unravel the trend and pattern in food consumption of Indian
households across different income classes, geographical regions and rural and urban
sectors to provide in lucid manner a baseline for formulating suitable food and
nutritional security policies. The present study also unwinds the expenditure
elasticities of food products in different geographical regions with a purpose to
estimate future food demand. This would also help in giving directions for the food
industry, a sunrise sector to reorient their strategies to meet the future demands.

II

DATA AND METHODOLOGY

The consumption pattern of food commodities was studied using unit level
National Sample Survey Organisation (NSSO) data on consumption expenditure
pertaining to the years 1987-88 and 2004-05. The quantity consumed and expenditure
on individual commodity were aggregated into food groups, i.e., cereals, pulses,
edible oils, milk and milk products (MMP), non-vegetarian products, vegetables and
fruits, and consumption pattern (in terms of per capita consumption and budget share)
of each food group was examined across different income categories, geographical
regions and rural and urban sectors. The commodities, not expressed in ‘kg’ unit,
were converted using suitable conversion factors1 and aggregated thereafter.
Aggregation of individual food commodities in the respective food group is useful
even with the usual limitations of adding foods in quantity terms.

Regional Demarcation

The regional variations in food consumption was examined by dividing India into
five geographical regions, viz., North (Uttar Pradesh, Punjab, Haryana, Delhi,
Chandigarh, Himachal Pradesh, Uttarakhand and Jammu and Kashmir), West
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 537
 

(Rajasthan, Gujarat, Maharashtra, Madhya Pradesh, Goa, Dadra and Nagar Havelli
and Daman and Diu), South (Karnataka, Kerala, Tamil Nadu, Andhra Pradesh,
Puducherry, Lakshadweep and Andaman and Nicobar Islands), East (West Bengal,
Bihar, Odisha, Jharkhand and Chhattisgarh) and North-East (Assam, Arunachal
Pradesh, Manipur, Nagaland, Sikkim, Meghalaya, Mizoram, and Tripura). It is to be
noted that in the year 1987-88, Chhattisgarh was not separated from Madhya Pradesh.
As Chhattisgarh is geographically close to the Eastern region with contiguous borders
with Odisha and Jharkhand, district level data was extracted instead of using state
level data for the year 1987-88 and aggregated to form the respective regions.

Income Categories

Income category-wise food consumption pattern was examined by dividing

households into four expenditure classes2 based on poverty line. The households were
classified as, very poor with consumption expenditure of less than 75 per cent of
poverty line, poor with consumption expenditure below poverty line up to 75 per cent
of poverty line, middle class with consumption expenditure ranging from poverty line
up to 150 per cent of poverty line and rich with consumption expenditure more than
150 per cent of poverty line (Radhakrishna and Ravi, 1990, Kumar et al., 2011).
Monthly per capita consumption expenditure at poverty line was taken as Rs. 356 for
rural areas and Rs.538 for urban areas (Government of India, 2007). To compare the
results pertaining to the years 1987-88 and 2004-05, all the expenditure values were
expressed at 2008 prices by using consumer price index (CPI) for agricultural labour
for the rural sector and CPI for Urban Non- Manual Employees for the urban sector.

Divisia Indices for Temporal Consumption Changes

Temporal change in consumption of food commodities between 1987-88 and

2004-05 was studied by estimating Divisia Quantity Index (DQ). DQ is a combined
measure of change in consumption of all the food groups together (Selvanathan and
Selvanathan, 2006). DQ was estimated as follows:

n
DQ = ∑ wi (1n Qit – 1n Qit-1) ....(1)
i=1

where,
Qit = per capita consumption (kg) of i-th commodity in period t (2004-05),
Qit-1 = per capita consumption (kg) of i-th commodity in period t-1 (1987-88),
wi = arithmetic mean of budget share of i-th commodity in period t and t-1, and
n = food groups, i.e., cereals, pulses, edible oils, MMP, non-vegetarian products.
538 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

It is to be noted that DQ represents change in quantity of all defined food groups

together during the periods under consideration. As within the food basket, the
consumption pattern of all the food commodities do not follow similar trend, relative
quantity log-changes were calculated to know the change in consumption of
individual food groups (cereals, pulses, edible oils, MMP, non-vegetarian products,
vegetables and fruits) relative to overall food consumption change (DQ). This will
also provide an idea about the changing importance of food commodities within food
basket during the period under consideration.

RQLCi = DQ1 – DQ ....(2)

where,
RQLCi = relative quantity log change of i-th food group, and
DQi = ln Qit- ln Qit-1

Similarly, Divisia Price Index (DP) was also estimated to know the changing
structure of prices of food commodities during the period under consideration.

n
DP = ∑ wi (1n Pit – 1n Pit-1) ....(3)
i=1

where
Pit = unit price of i-th food commodity in period t (2004-05), and
Pit-1 = unit price of i-th food commodity in period t-1 (1987-88)

Unit price of each food group is the ratio of expenditure to the quantity consumed
of the respective food commodity group. DP was estimated in nominal as well as real
(at 2008 prices) terms. Further, attempts were made to know the co-movement of real
price and quantity of food commodities by estimating divisia price-quantity
correlation coefficient (pt) as follows:

∑ w ( DP − DP)( DQ − DQ)
i i i
ρt = i =1 ....(4)
⎡ n 2⎤⎡
n
2⎤
⎢∑ i w ( DQi − DQ ) ⎥ ⎢ ∑ wi ( DPi − DP ) ⎥
⎣ i =1 ⎦ ⎣ i =1 ⎦

Expenditure Elasticity using LA-AIDS Model

Expenditure elasticity explains the likely responsiveness of income on the

consumption of food commodities and is used to project their future demand.
Expenditure elasticity of food commodities were estimated separately for each region
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 539
 

and rural and urban sectors separately using cross sectional data pertaining to the year
2004-05. The study used multi-stage (two stage) budgeting framework with Almost
Ideal Demand System using Linear Approximation (LA-AIDS) developed by Deaton
and Muellbauer (1980) to model the consumption behaviour of the households. The
multistage budgeting technique addresses a common problem in empirical estimation
of system demand models requiring a sizeable number of equations, given the wide-
variety of consumption goods jointly purchased by the households (Mustapha et al.,
1994; Tiffin and Tiffin, 1999).
In the first stage, the expenditure elasticity of food was estimated with ordinary
least squares (OLS) procedure using the following linear function:

lnF = α + β lnpf + γ lny +δ1z1+ δ2 z2+δ3 z3+ e …. (5)

where,
F = monthly per capita food expenditure,
γ = monthly per capita total expenditure,
pf = unit price of food,
z1 = age of household head
z2 = household size, and
z3 = child proportion

In the second stage, a household allocates a portion of food expenditure for the
consumption of food groups such as cereals, pulses, edible oils, MMP, vegetables,
non-vegetarian products, etc. The natural approach would be to include purchase of
food in the right hand side as regressor. This raises the second major problem, which
is simultaneity, given that such purchasing decisions are endogenous. To address this,
the predicted value rather than actual value is used as regressor (Tiffin and Tiffin,
1999; Dey, 2000). Another problem arises that several commodities have a
consumption value of zero for several households due to variation in the preference,
infrequent purchasing and/or misreporting (Keen, 1986). To overcome the problem of
zero observations, two-step Heckman estimation procedure was used, wherein, first a
probit regression model is computed in order to estimate the probability that a given
household consumes the given commodity (Heien and Wessells, 1990). This
regression is used to estimate the inverse mills ratio (IMR) for each household, which
is used as an instrument in the second regression. The structural form of LA-AIDS,
which was employed in the second stage of demand system is as follows:

where,
si = budget share of i-th commodity in food expenditure, i = 1, 2, 3, …..n
540 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

pij = price of j-th food group in i-th food equation,

I = stone price index of the respective commodity (eqn. 10),
F̂ = predicted value of food expenditure from eqn (8),
IMRi = inverse mills ratio with respect to i-th food commodity,
z1i = age of household head, and
z2i = household size

ln I= ∑ϖ
i
i ln pi ....(7)

where, ϖi is the mean of the expenditure share of the i-th commodity. Further, it
is to be noted that although simultaneity bias has been reported in the estimation
using LA-AIDS, this model is widely used because it is relatively easy to estimate
and interpret, satisfies axioms of choices exactly, compatible with aggregation over
consumers and consistent with household budget data (Deaton and Meulbauer, 1980;
Alston and Chalfant, 1993; Eales and Unnevehr, 1988). To be consistent with
microeconomic theory (consumer is a utility maximiser), certain restrictions were
imposed: (1) homogeneity of degree zero in prices and income (i.e., consumers have
no money illusion); (2) symmetrical cross elasticities, and; (3) additivity (all the
budget shares add up to 1). Since the errors of this system of equations tend to be
correlated as the samples drawn were almost identical, seemingly unrelated
regression estimation (SURE) model, proposed by Zellner (1962), was used to get the
efficient estimators of the model. The SURE model employs feasible generalised
least squares technique for estimating the model.
The expenditure elasticity for i-th food commodity with respect to total food
expenditure was estimated by:

Ci
ni = 1 + ….(8)
ϖi

Finally, the expenditure (income) elasticity of i-th food commodity was

calculated as a product of commodity group elasticity with respect to food
expenditure (from eqn 9) and food expenditure elasticity with respect to total
consumption expenditure (from eqn 8).

Food Demand Projection

The demand for the food commodities was projected for the year 2020 for each
region and both the sectors separately using the following expression;

Dt = d0 * Nt (1 + y * e)t …. (9)
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 541
 

where,
Dt = household demand of a commodity in year t,
d0 = per capita demand of the commodities in the base year,
y = growth in per capita income,
e = expenditure elasticity of demand for the commodity, and
Nt = projected population in year t.

To make demand projections for the consumption of food commodities, in the

simulation, two alternate scenarios of the income growth rate were used. First
scenario, i.e., Business as Usual (BAU) was built by estimating the growth rate in
gross domestic product at 1999-2000 prices from the year 2003-04 to 2007-08 for
different regions of the country. Alternatively, high growth rate (9 per cent per
annum) in income scenario was used. Growth rates in per capita income under
alternative scenarios were worked out by subtracting the population growth from
income growth. The projected population estimates were taken from the Registrar
General of India for the year 2020. State wise population estimates were aggregated
to represent population of the respective regions (Appendix 1).

III

RESULTS AND DISCUSSION

Income Category-wise Food Consumption Pattern in India

The consumption pattern of food commodities depends to a large extent on

income of the household. The share of food expenditure in total income was found to
be inversely related to the household income (Table 1). Very-poor households spent
65.95 per cent of their income (MPCE) on food as compared to 42.39 per cent by the
rich households in India in 2004-05. This trend was consistent in both rural as well as
urban sectors. Notwithstanding the proportion of MPCE spent on food was
comparatively higher for rural households than their urban counterparts in all the
income categories. The inverse relationship between income and food budget (share)
was in conformity with the Engel’s law which states that as income increases, the
proportion of income on food items decreases, though the actual expenditure may
increase.
TABLE 1. EXPENDITURE PATTERN ACROSS DIFFERENT INCOME CLASSES IN INDIA

Share of food in MPCE in 2004-05 Change between 1987-88 and 2009-10

Income class Rural Urban Combine Rural Urban Combine
(1) (2) (3) (4) (5) (6) (7)
Very-poor 67.53 63.16 65.95 -5.82 -7.97 -6.68
Poor 65.37 58.79 63.55 -6.93 -8.78 -7.47
Middle 62.22 53.29 59.85 -6.78 -9.91 -7.62
Rich 47.68 35.71 42.39 -7.94 -11.73 -10.24
All classes 54.36 41.01 49.25 -8.57 -13.51 -10.95
542 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

The real (at 2008 prices) MPCE increased by 11.27 and 23.77 per cent in rural
and urban sectors between 1987-88 and 2004-05, respectively. Consequently, the
share of food in total MPCE declined though at varying degree (6-12 per cent) for
different income categories during the same period. The decline in food share was
comparatively higher for urban and rich households than the rural and poor
households. Similar shift in consumption expenditure of Indian households has been
reported by several authors (Kumar, 1996; Meenakshi, 1996; Rao, 2000;
Radhakrishna, 2005). Thus, an improvement in income prompts consumers to
diversify their consumption expenditure away from food products towards non-food
items. However, the absolute consumption of food products increases with the
improvement in income.
Within the food basket, there existed contrasting variation in the consumption
pattern across different income categories. The per capita consumption of all food
groups increased with the increase in income, but their share in total food expenditure
showed mixed trend depending upon the food group (Table 2). In case of cereals,
pulses, edible oils and vegetables, the expenditure share reduced with the increase in
income, while for MMP, non-vegetarian products and fruits, it increased in the rural
as well as urban sectors. Decreasing share of essential food commodities (cereals,
pulses, edible oils and vegetables) and increasing share of high value agricultural
commodities (HVACs) (MMP, non-vegetarian products and fruits) with the rise in
income confirms empirically Bennett’s law of consumption which states that as
income increases, the consumers typically switch to a more expensive diet,
substituting quality for quantity. The shift in the dietary pattern away from cereal
consumption to more expensive milk, poultry and meat products is a consistent
change associated with economic growth (Meenakshi, 1996). Interestingly, the

TABLE 2. INCOME GROUP WISE CONSUMPTION PATTERN OF FOOD COMMODITIES

IN INDIA IN 2004-05
(share in food expenditure)
Very Poor Poor Middle Rich
Food commodity R U R U R U R U
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Cereals 48.74 38.62 43.19 33.14 36.22 28.59 27.21 20.76
(124) (116) (138) (122) (146) (122) (154) (118)
Pulses 6.35 6.35 6.38 6.35 6.14 6.04 5.58 5.08
(5) (6) (7) (8) (8) (9) (11) (12)
MMP 4.45 9.87 7.77 12.92 13.31 16.78 20.00 21.84
(8) (20) (18) (33) (39) (52) (86) (93)
Edible Oils 9.02 9.09 9.12 9.16 8.81 8.99 8.04 8.08
(3) (4) (4) (6) (5) (7) (8) (10)
Non-veg 3.74 5.33 4.74 6.60 5.46 6.50 7.11 6.73
(2) (3) (3) (5) (4) (6) (8) (8)
Vegetables 12.87 12.13 12.46 11.82 11.80 11.20 10.38 10.49
(41) (47) (52) (57) (61) (65) (75) (81)
Fruits 1.18 1.80 1.48 2.31 2.08 3.16 3.69 5.19
(3) (4) (4) (6) (6) (10) (14) (20)
R: Rural, U: Urban, MMP: milk and milk products.
Figures in parentheses are per capita annual consumption (kg/capita/annum) of respective food group.
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 543
 

growth in the consumption of HVACs with the rise in income was higher for rural
households than urban counterparts. This indicates that rural households have higher
propensity to consume HVACs than their urban counterparts. Thus, any income
generating opportunity in the rural areas would fuel the demand for HVACs more
than that in the urban areas.

Regional Variations in Food Consumption Pattern in India

A perusal of Table 3 reveals that food basket of the Indian households is

predominated by cereals followed by MMP. However, the food habits of the Indian
households varied across different geographical regions. The share of cereals in total
food budget varied from 21.81 per cent in urban West to 42.82 per cent in rural East
in 2004-05. The per capita consumption of cereals followed the similar trend. It is to
be noted that the consumption of cereals was comparatively higher in the rural sector
in all the regions. It might be due to easy and cheap availability as well as payment of
wages in kind form (grains) for most of the labour activities in rural India.

TABLE 3. REGION-WISE CONSUMPTION PATTERN OF FOOD COMMODITIES IN INDIA IN 2004-05

(share in food expenditure)
Food North West South East North-East India
group R U R U R U R U R U R U
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
Cereals 29.02 22.04 28.08 21.81 31.33 27.08 42.82 29.42 37.84 30.06 33.13 24.52
(149) (122) (135) (109) (133) (117) (159) (137) (156) (147) (145) (119)
Pulses 6.52 5.48 6.21 5.54 6.15 5.91 4.97 4.74 4.69 4.26 5.90 5.47
(10) (10) (9) (10) (9) (10) (7) (8) (7) (8) (9) (10)
MMP 23.45 26.27 21.48 21.70 10.42 14.73 8.14 11.42 6.26 9.27 15.58 19.30
(79) (89) (64) (69) (37) (58) (23) (38) (17) (26) (49) (66)
Edible oils 8.01 7.71 9.99 9.98 7.94 7.32 8.28 8.37 6.99 7.04 8.47 8.43
(6) (8) (7) (9) (6) (7) (5) (7) (5) (7) (6) (8)
Non-veg 2.71 3.38 2.79 4.10 9.32 8.77 7.95 12.23 16.50 17.92 6.12 6.60
(2) (3) (2) (3) (9) (9) (6) (11) (12) (15) (5) (6)
Vegetables 10.97 11.61 10.16 10.76 9.46 8.83 13.40 12.85 14.10 13.18 11.22 10.85
(69) (88) (45) (57) (45) (50) (81) (96) (84) (84) (62) (70)
Fruits 2.23 3.85 2.47 4.52 5.32 5.33 1.70 3.15 1.54 2.77 2.78 4.31
(8) (13) (7) (13) (15) (18) (6) (11) (7) (10) (8) (14)
R: Rural, U: Urban, MMP: Milk and milk products.
Figures in parentheses are per capita annual consumption (kg/capita/annum) of respective food group.

In case of MMP, per capita consumption was the least in North-Eastern region,
while it was the highest in Northern region. Conversely, the consumption of non-
vegetarian products was the highest in North-Eastern region, while it was the least in
Northern region. This indicated the sharp contrast in food preferences in different
regions as MMP and non-vegetarian products are preferred by different groups. The
regional variation in food consumption pattern might be due to diversity in cultural
tradition as well as relative availability of food products in the respective region.
Notwithstanding MMP is primarily produced in Northern and Western part of the
country and thus becomes a major constituent of food basket. The share of
commodities such as pulses, edible oils, vegetables and fruits in total food budget
544 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

ranged from 4-7 per cent, 7-10 per cent, 9-14 per cent and 2-5 per cent, respectively
across different regions depending upon the relative availability at affordable prices
in the respective region.

Temporal Changes in Food Consumption Pattern in India

The estimated divisia quantity index (DQ), which signifies the changes in
consumption of all the food commodities together, revealed that the average food
consumption increased in all the regions and both the sectors except rural North and
urban West between 1987-88 and 2004-05 (Table 4). Overall, the increase in food
consumption in urban India was almost double that of rural India. Among the regions,
Southern region witnessed highest increase of 8.08 and 10.32 per cent in the rural and
urban sectors, respectively. On the other hand, estimated divisia price index (DP)
revealed that though current market price (unit value) of the entire food basket
increased by more than 100 per cent in all the regions, the real price (at 2008 prices)
decreased between 1987-88 and 2004-05. The decrease in real prices was more
pronounced in the urban sector. The decrease in real prices might have contributed in
increased consumption of food commodities during the period under consideration.
However, the increase in quantity consumed was proportionately less than decrease
in real prices in all the regions and both the sectors except rural South, suggesting
inelastic demand for food commodities. Further, the estimated correlation coefficient
between divisia price and quantity index was negative which indicates inverse
relationship between prices and quantity, though with varying degree across different
regions.

TABLE 4. REGION-WISE DIVISIA QUANTITY (PRICE) INDEX FOR FOOD COMMODITIES AND
PRICE-QUANTITY CORRELATION COEFFICIENT

Divisia price index Divisia quantity Divisia price-Quantity

Region At 2008 price At market price index correlation coefficient
(1) (2) (3) (4) (5)
Rural
North -4.05 111.75 -2.24 -0.80
West -11.43 104.37 2.39 -0.67
South -7.24 108.57 8.08 -0.59
East -14.66 101.15 2.77 -0.16
North–East -7.82 107.98 5.74 -0.42
India -9.00 106.80 2.41 -0.54
Urban
North -13.28 110.86 2.22 -0.83
West -14.25 109.89 -0.84 -0.90
South -13.60 110.55 10.32 -0.75
East -18.25 105.89 9.86 -0.79
North–East -16.44 107.70 8.06 -0.79
India -14.66 109.48 5.10 -0.74
Note: Food groups include cereals, pulses, edible oil, MMP, non-veg, vegetables and fruits.
Indices were multiplied by100 to express in per cent terms.
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 545
 

Decomposition of DQ into components (food groups) and estimation of relative

quantity log changes revealed that change in quantity consumed was not uniform
across different food groups (Table 5). Within the food basket, the relative decline in
consumption of cereals, pulses and fruits was accompanied by relative increase in
consumption of MMP, non-vegetarian products, vegetables and edible oils. This
indicates the changing importance of food products and diversification within the
food basket. The consumption of the cereals and pulses reduced by 20 and 28 per
cent in rural India, and by 17 and 26 per cent in urban India, respectively. The
decrease in the relative quantity of cereal consumption might be because of change in
taste and preferences of households away from cereals (Radhakrishna, 1996;
Radhakrishna and Murty, 1999). Further, decline in cereals consumption was sharper
in the rural sector due to improvements in the rural infrastructure which made other
food items available in the rural areas (Rao, 2000). Dasgupta and Sirohi (2010)
stressed that the foodgrains consumption has declined largely on account of reduced
consumption of coarse cereals, while for superior cereals like, rice and wheat the
decline is marginal. Decline in the consumption of coarse grains is due to easy access
and increasing availability of rice and wheat resulting from favourable policy and
technology, through exposure to high value agricultural commodities and shifts in the
dietary pattern either due to a rise in income or fall in price (Chand and Kumar,
2002).

TABLE 5. RELATIVE QUANTITY-LOG CHANGES IN FOOD GROUPS BETWEEN 1987-88 AND 2004-05

Region Cereals Pulses MMP Edible oils Non-veg Vegetables Fruits

(1) (2) (3) (4) (5) (6) (7) (8)
Rural
North -0.163 -0.318 0.048 0.392 0.061 0.217 -0.432
West -0.251 -0.348 0.244 0.268 -0.071 0.261 -0.030
South -0.246 -0.114 0.201 0.379 0.219 0.296 0.246
East -0.180 -0.350 0.291 0.454 0.154 0.240 0.039
N -E -0.172 -0.235 -0.124 0.451 0.190 0.295 -0.285
India -0.200 -0.288 0.157 0.372 0.147 0.257 -0.058
Urban
North -0.128 -0.278 0.080 0.141 -0.148 0.217 -0.644
West -0.148 -0.288 0.109 0.122 -0.265 0.227 -0.237
South -0.194 -0.164 0.184 0.164 0.153 0.197 0.038
East -0.202 -0.328 0.058 0.206 0.199 0.152 -0.053
N -E -0.142 -0.355 -0.130 0.238 0.235 0.173 -0.252
India -0.170 -0.258 0.122 0.155 0.017 0.198 -0.247
MMP: milk and milk products.

The relative decline in per capita consumption of fruits, though not uniform
across different regions, might be due to low base value or increasing preference of
consumers towards processed fruit products and juices. The relative increase in the
consumption of MMP, non-vegetarian products, vegetables and edible oils indicated
improvement in the nutritional status of the households through increasing
consumption of these commodities. However, the implications of declining
546 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

foodgrains consumption and dietary diversification towards HVACs on food and

nutritional security depend on the net nutritional intake which seems to be
paradoxical. Increasing consumption of HVACs should outstrip the decline in the
nutritional intake due to decreasing foodgrains consumption.
Further, the increase in relative quantity of aforesaid commodities was not
uniform across the regions. For example, relative quantity of MMP in the North-
Eastern region and relative quantity of non-vegetarian products in the Northern
(urban sector) and Western (both rural and urban sector) regions declined between
1987-88 and 2004-05. The decline in MMP consumption in North-East and non-veg
consumption in Northern and Western regions might be because of less preference
towards these commodities in the respective region.

Estimated Expenditure Elasticity and Demand Projection of Food Commodities

Food, being a necessary item, exhibits inelastic demand. However, within the
food basket, different items respond differently with the change in income of
households. The estimated expenditure elasticity of cereals was positive but lowest
among the food commodities in all the regions indicating that increase in income of
households will lead only a marginal increase in cereals consumption (Table 6). On
the other hand, expenditure elasticity of HVACs such as MMP, non-vegetarian
products and fruits was comparatively higher than others, though with inter-regional
variations. Thus, HVACs are relatively income elastic and with the increase in
income (expenditure), demand for these commodities will be higher than staple
foods. Further, expenditure elasticity of food items was higher for rural households
than urban counterparts. This indicates that food demand will be comparatively
higher in rural areas with per unit rise in income. The estimated expenditure
elasticities and food projections were found to be comparable with other studies
(Appendix 2), though there exists wide variation in these estimations due to
differences in database used, methodology adopted and assumptions made in the
estimation procedure.

TABLE 6. ESTIMATED EXPENDITURE ELASTICITY OF FOOD COMMODITIES

North West South East North-East India

Food group R U R U R U R U R U R U
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
Cereals 0.15 0.10 0.19 0.13 0.23 0.22 0.31 0.16 0.32 0.17 0.24 0.15
Pulses 0.46 0.47 0.59 0.40 0.50 0.53 0.71 0.54 0.65 0.53 0.56 0.46
MMP 0.82 0.77 1.16 0.72 0.77 0.71 1.31 0.87 1.06 1.28 1.00 0.82
Edible oils 0.52 0.47 0.64 0.50 0.40 0.49 0.67 0.57 0.55 0.56 0.54 0.52
Non-veg 0.76 0.64 0.88 0.67 0.87 0.67 1.16 0.86 0.95 0.94 1.04 0.76
Vegetables 0.45 0.38 0.49 0.40 0.40 0.44 0.52 0.43 0.53 0.57 0.43 0.45
Fruits 1.09 1.13 1.04 1.11 1.22 0.99 1.09 1.04 0.76 1.48 1.31 1.09
R: rural, U: urban, MMP: Milk and milk products.
Estimated parameters from the model are given in Appendix 2.
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 547
 

Although, cereals exhibited inelastic demand, its total demand in future will be
high due to increase in population and high indirect demand (seed, animal feed,
industrial demand waste, etc). The total projected household demand for cereals in
India was estimated to be 226-237 million tonnes (mt) in 2020 (Table 7). With the
addition of indirect demand, total cereals demand will be even much higher. From the
food security point of view, many studies concluded that requirements of cereals in
the country will be adequately met by domestic supplies at least upto the year 2020
(Government of India, 2001) and India would be the net exporter of major cereals
crops (FAPRI, 2009). There are many studies on demand-supply mismatch on food
commodities with strikingly diverse conclusions for food security outlook in India
(Kumar et al., 1995; Bhalla et al., 1999; Chand, 2007; Mittal, 2006; Amarasinghe at
al., 2007; FAPRI, 2009). Household demand for the HVACs such as MMP, non-
vegetarian products, fruits and vegetables was projected as 179-216 mt, 17-20 mt, 40-
49 mt and 136-151 mt, respectively in 2020. Further, a perusal of Table 7 reveals that
there will be wide inter-regional variations in the household demand for food
commodities because of varying food habits, relative availability and other socio-
economic constraints. Thus, efforts should be made to match the demand and supply
at disaggregate level mainly for HVACs, which is causing spiralling inflation in the
recent past and improve the supply chain to even out the regional imbalance of food
availability.

TABLE 7. PROJECTED HOUSEHOLD DEMAND FOR FOOD COMMODITIES IN 2020

(million tonnes)
Zone Cereals Pulses Milk Edible oils Non-veg Vegetables Fruits
(1) (20 (3) (4) (5) (6) (7) (8)
Business as usual scenario
North 55.26 5.23 57.43 3.46 1.50 38.12 8.31
West 53.06 5.60 64.91 4.99 1.83 27.86 8.99
South 43.56 4.37 27.38 2.65 5.83 19.93 15.35
East 64.74 4.47 27.24 3.27 6.42 43.85 6.10
N –E 9.69 0.59 2.21 0.44 1.32 6.43 0.82
India 226.31 20.26 179.17 14.81 16.9 136.19 39.57
High growth scenario
North 57.43 5.92 71.15 3.96 1.81 42.87 11.10
West 54.70 6.12 76.97 5.52 2.09 30.12 10.74
South 45.12 4.72 30.56 2.83 6.56 21.22 18.09
East 68.57 5.10 34.35 3.71 7.90 48.46 7.49
N -E 10.90 0.76 3.39 0.55 1.90 7.96 1.21
India 236.72 22.62 216.42 16.57 20.26 150.63 48.63

IV

CONCLUSIONS AND POLICY IMPLICATIONS

Food consumption pattern of Indian households was found to be in conformity

with two most important law, i.e., Engel’s law and Bennett’s law of consumption.
The study revealed striking differences in consumption pattern across income
categories, geographical regions and rural and urban sectors. These differences might
548 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

be because of a complex set of factors such as differential taste and preferences, large
cultural diversity, relative affordability due to varying employment opportunities,
relative availability of food commodities, etc. Further, declining share of foodgrains
and simultaneous diversification of food basket over the years is a sign of consumers’
welfare. However, declining per capita consumption of foodgrains needs due
consideration by the policy makers from food and nutritional security point of view.
Therefore, every effort should be made to improve the per capita availability,
accessibility and affordability of foodgrains especially by the poor households who
spend a major portion of food budget on foodgrains. The inverse relationship between
food prices and consumption necessitates controlling food inflation to improve
affordability of food commodities by poor households. Further, estimated expenditure
elasticities confirmed that HVACs like MMP, non-vegetarian products, fruits, etc. are
more responsive to increase in income than stable food with wide inter-regional
variations. Although, the cereals exhibited inelastic demand, its total demand will be
high due to increase in population and high indirect demand (seed, animal feed,
industrial demand waste, etc.). Thus, there is a strong need to improve productivity,
control prices, and strengthen public distribution system and supply chain at
disaggregated level to fulfil the future food demand.

Received November 2011. Revision accepted December 2013.

NOTES

1. The conversion rates from numbers to kilograms used are: 1 orange= 0.125 kg, 1 green coconut = 0.25 kg,
1 banana = 0.1 kg, 1 lemon = 0.033 kg, 1 egg = 0.055 kg, 1 litre milk = 1.03kg.
2. Household expenditure is used as a proxy for household income because information on household income
is rarely available.

REFERENCES

Amarasinghe, U.A, T. Shah and O.P. Singh (2007), Changing Consumption Patterns: Implications on
Food and Water Demand in India, Research Report 119, Draft Paper, NRLP-IWMI, available on
http://nrlp.iwmi.org.
Alston, J.M. and J.A. Chalfant (1993), “The Silence of the Lambdas: A Test of the Almost Ideal and
Rotterdam Models”, American Journal of Agricultural Economics, Vol.75, pp.304–313.
Bhalla G.S., P. Hazell and K. Kerr (1999), Prospects for India’s Cereal Supply and Demand to 2020,
Food, Agriculture and the Environment Discussion Paper 29, International Food Policy Research
Institute, Washington, D.C., U.S.A.
Chand, Ramesh (2007), “Demand for Foodgrains”, Economic and Political Weekly, Vol. 42, No.52,
December 29, pp.10-13.
Chand, R. and P. Kumar (2002), “Long Term Changes in Coarse Cereals Consumption in India: Causes
and Implications”, Indian Journal of Agricultural Economics, Vol.57, No.3, July-September,
pp.316-325.
Dasgupta, P. and S. Sirohi (2010), “Indian Agricultural Scenario and Food Security Concerns in the
Context of Climate Change: A Review”, Munich Personal RePEc Archive (MPRA), Paper No.
24067, http://mpra.ub.uni-muenchen.de/24067.
 UNRAVELLING FOOD BASKET OF INDIAN HOUSEHOLDS 549
 

Dastagiri, M.B. (2004), Demand and Supply Projections for Livestock Products in India, Policy Paper
21, National Centre for Agricultural Economics and Policy Research (NCAP), New Delhi.
Deaton, A.S. and J. Muellbauer (1980), “An Almost Ideal Demand System”, American Economic
Review, Vol.70, pp.359-68.
Dey, M.M. (2000), “Analysis of Demand of Fish in Bangladesh”, Aquaculture Economics and
Management, Vol.4, pp.65-83.
Eales, J.S. and L.J. Unnevehr (1988), “Demand for Beef and Chicken Products: Separability and
Structural Change”, American Journal of Agricultural Economics, Vol.70, No.3, pp. 521-532.
Food and Agricultural Policy Research Institute (FAPRI) (2009), U.S. and World Agricultural Outlook,
FPRI Staff Report 09-FSR 1, Iowa State University of Missouri-Columbia.
Government of India (2001), Report of the Working Group on Public Distribution System and Food
Security for the Tenth Five Year Plan (2002-2007), Planning Commission, New Delhi.
Government of India (2007), Poverty Estimates for 2004-05, Press Information Bureau, Planning
Commission, New Delhi,
Heien, D. and R. Wessells (1990), “Demand System Estimation with Microdata: A Censored Regression
Approach”, Journal of Business and Economic Statistics, Vol.6, pp.313-325.
Keen, M. (1986), “Zero Expenditures and the Estimation of Engel Curves”, Journal of Applied
Econometrics, Vol.32, pp.560-568.
Keyzer, M.A.; M.D. Merbis, I.F.P.W. Pavel and C.F.A. Van Wesenbeeck (2005), “Diet Shifts Towards
Meat and the Effect on Cereal Use: Can We Feed the Animals in 2030”, Ecological Economics,
Vol.55, pp.187-202.
Kumar, P. and V.C. Mathur (1996), “Structural Changes in the Food Demand for India”, Indian Journal
of Agricultural Economics, Vol.51, No.4, October-December, pp.664-673.
Kumar, P. (1996), “Structural Changes in Consumption and Small Farm Diversification”, in T Haque
(Ed.), (1996), Small Farm Diversification: Problems and Prospects, National Centre for
Agricultural Economics and Policy Research, New Delhi.
Kumar, P., Mark Rosegrant and Peter Hazell (1995), Cereals Projections for India, Agricultural
Economics Policy Paper 98-01, Division of Agricultural Economics, Indian Agricultural Research
Institute, New Delhi.
Kumar, P.; Anjani Kumar, P. Shinoj and S.S. Raju (2011), “Estimation of Demand Elasticity for Food
Commodities in India”, Agricultural Economics Research Review, Vol.24, No.1, pp.1-14.
Meenakshi, J.V. (1996), “How Important are Changes in Taste: A State-level Analysis of Food
Demand”, Economic and Political Weekly, Vol. 31, No.50, December 14, pp.3265-3269.
Minhas, B.S. (1991), “On Estimating the Inadequacy of Energy Intakes: Revealed Food Consumption
Behaviour Verses Nutritional Norms (Nutritional Status of Indian People in 1983)”, The Journal of
Development Studies, Vol.28, No.1, pp.1-38.
Mittal, S. (2006), Structural Shift in Demand for Food: Projections for 2020, Working Paper No. 184,
Indian Council for Research on International Economic Relations, New Delhi.
M.S. Swaminathan Research Foundation (2008), Report on the Status of Food Insecurity in Rural India,
Chennai.
Mustapha, N., R.A. Ghaffar and D. Poerwono (1994), “An Almost Ideal Demand System Analysis of
Fresh Fish in Semarang, Indonesia”, Journal of International Food and Agribusiness Marketing,
Vol.6, pp.91-128.
Paroda, R.S. and P. Kumar (1999), “Food Production and Demand Situation in South Asia”, Paper
Submitted for the Study Week on Food Needs for the Developing World in the Early Twenty-First
Century, January 27-30, Vatican City, Rome.
Radhakrishna, R. (2005), “Food and Nutrition Security of the Poor”, Economic and Political Weekly,
Vol.40, No.18, April 30, pp.1817-1821.
Radhakrishna, R. (1996), “Food Trends, Public Distribution System and Food Security Concerns”,
Indian Journal of Agricultural Economics, Vol.51, No.1&2, January-June, pp.168-183.
Radhakrishna, R. and C. Ravi (1990), Food Demand Projections for India, Research Report, Centre for
Economics and Social Studies, Hyderabad.
550 INDIAN JOURNAL OF AGRICULTURAL ECONOMICS

Radhakrishna, R. and K.N. Murty (1999), “Econometrics of Complete Demand Systems with
Illustrations for India”, in K.L.Krishna (Ed.) (1999), Econometric Applications in India, Oxford
University Press, Delhi.
Rao, C.H. Hanumantha (2000), “Declining Demand for Foodgrains in Rural India: Causes and
Implications”, Economic and Political Weekly, Vol. 35, No.4, January 22, pp.201-206.
Selvanathan, S. and E.A. Selvanathan (2006), “A. Consumption Patterns of Food, Tobacco and
Beverages: A Cross Country Analysis”, Applied Economics, Vol.38, No.13, pp.1567- 1584.
Tiffin, A. and R. Tiffin (1999), “Estimates of Food Demand Elasticities for Great Britain: 1972-1994”,
Journal of Agricultural Economics, Vol.81, pp.140-147.
Zellner, A. (1962), “An Efficient Method of Estimating Seemingly Unrelated Regression Equations and
Tests of Aggregation Bias”, Journal of the American Statistical Association, Vol.57, pp.500-509.

APPENDIX 1
POPULATION AND INCOME GROWTH SCENARIO FOR DEMAND PROJECTION

Projected Income growth

population (000) Population growth (per cent)
Region Sector for 2020* (2004-2020) BAU HGS
(1) (2) (3) (4) (5) (6)
Rural 237468 1.33 7.25 9.00
North
Urban 110328 2.54 7.25 9.00
Rural 218211 1.07 8.02 9.00
South
Urban 134972 2.10 8.02 9.00
Rural 157920 0.25 7.71 9.00
East
Urban 107096 1.81 7.71 9.00
Rural 247977 1.04 7.89 9.00
West
Urban 63134 1.52 7.89 9.00
Rural 39392 0.93 6.43 9.00
North-East
Urban 9655 2.43 6.43 9.00
Rural 900973 0.97 7.70 9.00
India
Urban 425182 2.05 7.70 9.00
* Data source: Registrar General of India, BAU: Business as usual scenario, HGR: High growth scenario.
 APPENDIX 2
ESTIMATED PARAMETERS OF FIRST STAGE OF DEMAND SYSTEM

North West South East North-East India

Parameters Rural Urban Rural Urban Rural Urban Rural Urban Rural Urban Rural Urban
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
Intercept 1.283 1.519 1.119 1.616 1.042 1.086 1.034 1.598 1.506 1.460 1.237 1.514
(0.030) (0.037) (0.031) (0.033) (0.028) (0.040) (0.024) (0.038) (0.051) (0.072) (0.013) (0.018)
Food prices 0.352 0.369 0.233 0.325 0.507 0.438 0.272 0.206 0.307 0.270 0.272 0.273
(0.010) (0.012) (0.009) (0.010) (0.011) (0.016) (0.008) (0.013) (0.014) (0.021) (0.004) (0.006)
Income 0.581 0.532 0.639 0.523 0.541 0.540 0.647 0.582 0.569 0.588 0.611 0.561
(0.004) (0.005) (0.005) (0.005) (0.004) (0.006) (0.004) (0.006) (0.007) (0.012) (0.002) (0.003)
Family size -0.009 -0.012 -0.005 -0.013 -0.010 -0.006 -0.003 -0.006 -0.006 -0.014 -0.005 -0.008
(0.001) (0.001) (0.001) (0.001) (0.001) (0.002) (0.001) (0.001) (0.001) (0.003) (0.000) (0.001)
Age 0.001 0.001 0.001 0.001 0.000# 0.001 0.001 0.001 0.000# 0.000# 0.001 0.001
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Child proportion -0.128 -0.138 -0.070 -0.130 -0.148 -0.117 -0.071 -0.074 -0.053 -0.072 -0.082 -0.107
(0.010) (0.012) (0.010) (0.012) (0.010) (0.014) (0.008) (0.014) (0.013) (0.022) (0.005) (0.006)
R2 0.760 0.814 0.745 0.799 0.775 0.784 0.813 0.816 0.670 0.786 0.763 0.792
No. of
observations 14334 8330 13244 9302 14080 8382 15985 6400 5739 1967 63382 34381
Figures in parentheses are standard error of respective parameter, *Significant at 5 per cent level of significance.
# non-significant, and All other parameters at significant at 1 per cent level of significance.
Estimated parameters for 2nd stage of demand system could not be given due to paucity of space and can be obtained from authors ([email protected]) if
required.