Inequality and welfare by food expenditure components.
Haq, Rashida ; Arshid, Nabeela
This paper examines inequality and welfare in food consumption
expenditure in Pakistan. The indices are derived on the basis of
Sen's welfare function by employing HIES, 2005-06. The empirical
results show that inequality and welfare varied widely in urban and
rural areas. It can be seen that the concentration index of non
essential/luxury food components are unevenly distributed over the total
expenditure in favour of rich families. The price elasticity of welfare
indicates that rising prices of food components have negative effect on
the aggregate welfare. The present analysis provides a basis for
determining the magnitude of inequality and welfare that can help in
designing appropriate policies.
JEL classification: D12, D3, D31, I31
Keywords: Food Consumption, Inequality, Welfare, Elasticity
I. INTRODUCTION
The effect of rising global food crises had a potential impact on
food consumption in Pakistan and also its implications for inequality
and welfare in urban and rural areas. In Pakistan agriculture is the
predominant sector of the economy accounting 46.2 percent of the gross
domestic product in 2008-09 which provides livelihood for 60.94 percent
of its rural population where its 27 percent poor reside [Pakistan
(2009)]. Due to food inflation of 26.61 percent in 2008-09, food
expenditure are increasingly dominating household budget, the poor are
consuming even less than before and the quality of their diet has
deteriorated further indicating a 30 percent undernourished population
[UNDP (2003)]. The Household Integrated Economic Survey (HIES), 2005-06
demonstrates that, out of total expenditure on average 43.05 percent
goes for food mainly for essential food, i.e., cereals, pluses, milk,
oil, vegetables and sugar where it constitutes 81 percent of food budget
share for bottom 20 percent poor population. This food purchase
decisions is also based on a budget that also must cover expenses for
clothing, housing, and other goods and services. The overall budget
available for food depends on the amount of total household budget spent
on other goods and services. This phenomenon indicates that low-income
countries spend a greater portion of their budget on necessities such as
food than wealthier countries do. Therefore, an in-depth study on food
components distribution requires an understanding of the complete
welfare function of consumers.
Effective evaluation of food policy issues requires accurate
information on distribution of food consumption patterns. This
distribution of household food expenditure also represents a very
important aspect of the distribution of economic welfare. Its role
becomes even more important when the distribution of economic welfare is
observed using micro data. In this case income data is considered the
best proxy to the level of the household's economic welfare is
usually not reliable and total consumption is taken as the second best
solution. However, micro level data do not present consumption
statistics and expenditure is used instead as a proxy. One important
problem associated with the use of total expenditure is that many
households account zero expenditure for various goods while their
consumption is not zero, simply because they do not purchase those goods
during the survey period. This becomes more important in the case of
long-life durable goods or goods purchased only occasionally [Lazaridis
(2000)]. Second, measures of consumption inequality are a useful
complement of even alternative to income or earnings inequality because
households do take some steps to smooth consumption. Therefore,
consumption inequality is probably the better measure of inequality in
welfare. For all these reasons, the distribution of total expenditure on
food, despite its limitations, is considered a good indicator of the
distribution of economic welfare.
However, food expenditure consists of several food components with
different shares and distributions which can be classified in essential
and non-essential or luxury items. In terms of foods, each society
determines which foods are social necessities and which are luxuries,
and this distinction reflects the 'social grammar' of a
society. Luxury usually denotes foods that are desirable or hard to
obtain but not essential to human nutrition. If the number of people who
have access to a luxury increases, the status of these goods changes;
they turn into commonplace goods and may ultimately become necessities
[Veen (2003)]. In Pakistan meat, fruits, soft drinks and other readymade
food products have little access to majority of the masses, thus
categorised as luxury food for the present analysis of inequality and
welfare. The principal purpose of this paper is to examine the
contribution of each food item to overall food expenditure inequality
and welfare. This analysis enables the policy-maker to reduce inequality
through taxes or subsidies in the most efficient way.
This paper examines the impact of eleven food components including
essential and luxury on the inequality of total food expenditures in
Pakistan. It also evaluates the effect of price changes on the aggregate
welfare. The study is also concerned with such questions as: what is the
extent of the inequality in food expenditure components? How do the
various expenditure components affect total welfare? The paper is
organised as follows: Section 2 discusses briefly, some evidence in
distribution of food consumption patterns in Pakistan. Section 3
develops techniques intended to analyse inequality and welfare in terms
of per capita expenditure components while Section 4 deals with results.
Finally, Section 5 is devoted to some concluding remarks.
2. SOME EVIDENCE ON FOOD CONSUMPTION PATTERNS
Recent shifts in food trade reveal dramatic changes in consumption
patterns for food around the globe. These changes are likely to continue
well into the future. A number of driving forces are working behind
these changing consumption patterns, but growth in income is perhaps the
most important of them. Growth of income during the past few decades has
resulted in increased purchasing power of the consumers around the world
that in turn has caused a shift of consumption to more expensive food
items. The overall budget available for food depends on the amount of
total household budget spent on other goods and services. Although the
food budget may grow at a relatively slow rate among high-income
consumers, global food consumption patterns are rapidly changing with
growing demands for quality, variety, and convenience. Globalisation,
improved transportation, and increased purchasing power have generally
increased the demand for higher value food products such as fruit and
vegetables, meats, and processed food products across all countries
[Seale, Anita, and Jason (2003)].
Despite considerable progress in recent decades, the goal of
adequate food and nutrition for all is still elusive. Eradicating hunger
and malnutrition is a high priority of aid agencies and governments
throughout the world. Several policies have been enacted to fight
malnutrition. One of them has focused on raising the incomes of the
poorest people [Grimard (1996)]. The distribution of budget shares for
the eight food subcategories across three income groups countries are
presented in Table 1 which demonstrates that cereals, fats and oils, and
fruits and vegetables account for a larger share of the total food
budget in low-income countries compared with high-income countries.
Low-value staples, such as cereals, account for a larger share of the
food budget in poorer countries, while higher-valued food items of
luxury foods, such as dairy and meat, and beverages and tobacco are a
larger share of the food budget in richer countries. In terms of food,
luxury usually denotes foods that are desirable of hard to obtain but
not essential to human nutrition. For a good to be a luxury it needs to
be desired by many but attained by few. As the true arena of luxury
goods, by definition, outside the reach of mass consumption; using
Berry's (1994) words, it is not possible to 'democratise'
luxuries. The conclusion that can be drawn from this is that luxury
goods will occur only in societies with strong social stratification,
where elites require goods in order to display and maintain their
status.
Table 2 shows inequality in budget shares for food sub-categories
for poorest quintile and richest quintile over time in Pakistan which is
largely a reflection of poverty: people do not have income for food.
There are some distinct differences in the expenditure patterns of poor
and rich households. Poor households spend on average 55 percent of
their budgets on food, whereas rich household's average per capita
food expenditure is 33 percent in 2005-06. The consumption pattern of
poor and rich households also shows that the richest 20 percent allocate
a large share of expenditure on high valued food items such as meat and
fruits. Given the slow income growth that is likely for the poorest
people in the foreseeable future, large numbers will remain malnourished
for decades to come.
In this scenario, Pakistan ranks at 61st position in Global Hunger
Index-2008 with 21.7 points out of 88 countries surveyed indicating
alarming level of hunger. The Index ranks countries on a 100-point scale
with zero being the best score (no hunger) and 100 being the worst in
three indicators which include child malnutrition, rates of child
mortality and the number of people who are calorie deficient [IFPRI
(2008)].
The food consumption patterns of households have been analysed in a
number of studies (1) in Pakistan. These studies differ not only in
their scope but also by the period of the data which are discussed here.
Rise in food prices have a big impact on the poorest households in
rural and urban areas in Pakistan which is explored by Naqvi and Akbar
(2000). They revealed that Pakistani households they have shifted
expenditures toward basic needs and away from non-essential items over
the 1987-88 to 1996-97, a period associated with rapidly increase in
price levels and with very slowly rising real incomes. High own- price
elasticity of demand for food combined with high expenditure allocations
on food indicates that the poorest households in Pakistan are facing a
distressful situation due to food price hikes. Currently, the nominal
prices of many staple food commodities are at their highest levels in
nearly 50 years leading to food riots in some developing countries. Haq,
Hina, and Meilke (2008) had estimated the impacts of rising world food
prices on poverty in rural and urban areas of Pakistan. As compared to
2004-2005 the unexpected food price changes resulting from the food
crisis increased poverty by 8.2 percentage points (34.8 percent),
severely affecting the urban areas where poverty doubled. The effect of
inflation are highly non-egalitarian and in fact, contributed more to
increase the existing inequalities in food expenditure than in non-food
expenditure especially in lower income brackets [Afridi, et al. (1984)].
The rise in inequality was more pronounced in urban areas than in rural
areas where both the poorest and middle income groups lost their
consumption share while the richest top two deciles gained in their
consumption share implying that inequality in Pakistan increased at the
expense of the poor and the middle income groups [Anwar (2009)].
The conventional wisdom proposed that income increases should allow
individuals to increase their food intake and nutrient consumption,
which in turn should improve their nutritional status.
Taking calories as a proxy for all nutrients Grimard (1996)
revealed that rural households in Pakistan spend 69 percent of their
food budget on cereals, dairy, edible oils and sugar thus translated
into 77 percent of daily calories consumption share per household. As
households get richer, they would substitute away from the lower quality
toward the better quality food items of a given category. Cheema and
Malik (1985) also showed that redistribution of income from rich to the
poor households will raise the consumption demand for basic necessities
like, wheat, pulses, edible oils, clothing and footwear, etc; while the
demand for personal effects: meat, fish and poultry, furniture would
decreased.
Sen's welfare index that determined the level of disparities
in different section of the society is used by Haq (1998) and Zakir and
Idrees (2009). The result confirmed that disparity in food consumption
is not high as compared to non-food consumption. As far as economic
welfare is concerned its level increased over time and its magnitude is
greater in the urban sector. It is also evaluated that 'food'
and 'housing' components of expenditure are the major
contributors in total welfare. The price elasticity of welfare has shown
that 'food' expenditure is more responsive in price changes in
all areas [Haq (1998)]. Trends in inequality and welfare during 1963-64
to 2004-05 revealed that income and consumption inequalities were more
severe in urban areas than in rural areas [Zakir and Idrees (2009)].
Since the consumption patterns are likely to change considerably
since the last two decades, there is a need to analyse the per capita
households' consumption patterns on the basis of more recent micro
data.
3. THEORETICAL FRAMEWORK AND THE DATA
It is important to employ a single measure of welfare to acquire a
complete welfare ranking of distribution. This type of measure can be
derived by giving different weights to individuals with different income
levels. It can be assumed that in a country there are n persons who are
arranged in ascending order of their income as: [X.sub.1] [less than or
equal to] [X.sub.2] ... ... ... [less than or equal to] [X.sub.n], then
a welfare measure may be described as a unique function of [X.sub.1]
[X.sub.2] ... ... ... ..., [X.sub.n] (1974) illustrated a welfare
function as follow:
W = [n.summation over (i=1)] [x.sub.i] [v.sub.i] ... ... ... ...
... ... ... (1)
Where [v.sub.i] is the weight given to the person with income
[x.sub.i]. It is evident that if [v.sub.i] = 1/n for
all i, then W is equal to average income of individuals. To make W
sensitive to inequality in the distribution, a higher (lower) weight
should be assigned to the individual with lower (higher) incomes. Sen
(1974) proposed that [v.sub.i] should be proportional to the number of
persons who are at least as well off as i. This proposition leads Sen to
write the welfare function as:
W = x(1 - G) ... ... ... ... ... ... ... (2)
Where x the mean income of the society and G is the Gini index
which is a widely used measure of the dispersion in welfare, with values
closer to unity associated with higher inequality. When G=I, all income
is received by one individual (perfect inequality) and G=0 indicates
absolute equality. The method of estimating Gini coefficient is due to
Lorenz's (1905) defined as:
G = 1 - [n.summation over (i=1)]([Z.sub.i] [Z.sub.i-1] ([x.sub.i] +
[x.sub.i-1]) ... ... ... ... ... (3)
Where Z is the cumulative proportion of income receivers and
[x.sub.i] is the corresponding cumulative proportion of income received.
Although there are several alternative welfare measures such as Atkinson
(1970), this paper uses W as a basis for analysing welfare in Pakistan
because it can be considered as an appropriate welfare measure which
takes into account both size and distribution of income.
(a) Inequality and Welfare by Food Expenditure Components
In measuring inequality, consumption expenditure is used as an
appropriate indicator of economic welfare because utility is derived
from the consumption of goods and services. It is also a better
indicator measure of long term standard of living than income. Since the
individual expenditure is the sum of several expenditure components, it
will be useful to analyse total inequality and welfare in terms of
individual total food expenditure components as proposed by Sen (1974).
Suppose there are k expenditure components whose prices are
[p.sub.1], [p.sub.2,.] .... [p.sub.k] and [q.sub.1], [q.sub.2], ...
[q.sub.k] are the quantities consumed of the k food expenditure
components then [u.sub.j] is taken the mean of jth expenditure
component. Then it is obvious that:
W = 1 - [k.summation over (j=1)] [[mu].sub.j] ([1 - [C.sub.j]) ...
... ... ... ... ... (4)
Here the disaggregation of Gini index in terms of expenditure
components can be written as [Kakwani (1980)]:
W = 1 / [mu] [k.summation over (j=1)] [[mu].sub.j] [C.sub.j] ...
... ... ... ... ... ... (5)
Where [C.sub.j] is considered as the concentration index of the jth
expenditure component. The concentration index [C.sub.j] is similar to
the Gini index except that the ranking of individual is by the total
expenditure and not the jth expenditure component. The concentration
index of expenditure component measures how evenly or unevenly that
expenditure component is distributed over the aggregate per capita
expenditure. If [C.sub.j] is greater (smaller) than G, it implies that
the jth expenditure component is distributed over the aggregate
expenditure in favour of rich (poor) individuals. Thus combining Sen
welfare index (2) with (4) and (5) gives:
W = [k.summation over (j=1)] [[mu].sub.j] (1 - [C.sub.j]) ... ...
... ... ... ... (6)
Which shows how the total welfare can be decomposed in terms of
individual expenditure components; [[mu].sub.j](1-[C.sub.j]) being the
contribution of the jth expenditure component to total welfare.
(b) The Price Elasticity of Aggregate Welfare
To evaluate the effect of price change on the total welfare, then
following equation expresses the price elasticity of aggregate welfare
as Kakwani (1980):
[n.sub.j] = - [[mu].sub.j] (1 - [C.sub.j])/[mu](1 - G) ... ... ...
... ... ... (7)
Which implies that if the price of jth commodity increases by 1
percent, then the aggregate welfare changes by [n.sub.j] percent,
[n.sub.j] will always be negative. Thus the magnitude of [n.sub.j] can
be used to evaluate the effects of price changes on the aggregate
welfare.
(c) Progressivity Index
This progressivity index is expressed as the ratio of the
inequality component to aggregate expenditure:
[P.sub.j] = (G - [C.sub.j]) / (1 - G) ... ... ... ... ... ... ...
(8)
A positive value of [P.sub.j] implies the jth expenditure component
to be progressive and the negative value implies the jth component to be
regressive. Thus, the magnitude of [P.sub.j] indicates whether the
increase in the jth component favours the poor or the rich. If the jth
component is distributed in proportion to total expenditure, [C.sub.j]
will be equal to G which gives [P.sub.j] to be equal to zero. In this
case, the effect of an increase in the jth component favours neither the
poor nor the rich. [P.sub.j] provides a quantitative basis for
maximising the country's total welfare with minimum cost.
The above measure should be interpreted with care since it is based
on the assumption that per capita total expenditure on food is
considered a good indicator of the distribution of welfare. This is an
approximate measure of individual welfare because it assumes that the
basic needs of every household member are the same.
(d) The Data
All computations are performed on the complete set of data
collected in Household Integrated Economic Survey (HIES), 2005-06
conducted by Federal Bureau of Statistics (FBS). The universe of HIES
consists of all urban and rural areas of all four provinces as defined
by the Provincial Governments. Military restricted areas have been
excluded from the scope of the survey. Separate sampling frames have
been used in the survey for urban areas and rural areas. FBS has
developed its own urban area frame. All urban areas known as
cities/towns of the urban domain of the sampling frame have been divided
into small compact areas known as enumeration blocks. Each enumeration
block comprises about 200-250 households. Each enumeration block has
been divided into low, middle and high income group, keeping in view the
status of the living of majority of households. With regard to the rural
areas, the lists of villages/mouzas/dehs according to population census,
1998 have been used as sampling frame.
A sample size of 15,453 households was considered appropriate to
provide reliable estimates of key characteristics at the
National/provincial level with urban/rural breakdown. The entire sample
of households has been drawn from 1109 Primary Sampling Units (PSUs) out
of which 531 are urban and 578 are rural. Households within each Primary
Sampling Units have been considered as secondary sampling units. 16 and
12 households have been selected from each sample village and
enumeration block respectively by random systematic sampling technique
with a random start [Pakistan (2005-06)]. It collects data on household
characters, consumption patterns, households' income by
sources/occupation/sectors and social indicators that make it possible
to estimate inequality and welfare across various sections of the
society. In this paper weights are assigned to each household in order
to insure the representativeness of the sample, and to obtain unbiased
sample population parameter estimates.
For the analysis of inequality and welfare, total food expenditure
is decomposed into essential and non essential food groups which are
further disaggregated into its food components as presented by Naqvi and
Akbar (2000).
[FIGURE 1 OMITTED]
4. ANALYSIS
In this section per capita food expenditure inequality and welfare
is analysed for Pakistan. The basic sample weights for sampled
households are used that reduces biases due to imperfections in the
sample related to non-coverage and non- response households. Inequality
measures such as quintile shares, Gini index and decile dispersion ratio
(DDR) in Table 3 are based on per capita household food consumption
expenditure across region in Pakistan. The DDR is defined as the share
of the bottom 20 percent in relation to the share of the top 20 percent.
For an in depth analysis of inequality, total food expenditure is
decomposed in essential and non essential food components. It is
observed that disparity in essential food expenditure is less as compare
to non essential food components; approximately fifty percent of non
essential food expenditure is incurred by top fifth quintile of the
population. Inequality parameters indicate that distribution in
expenditure is more pronounced in urban area as compare to rural area.
The decile dispersion ratio also indicates disparity in non essential
food expenditure is quite high within urban households.
Total per capita food expenditures inequality by its components is
presented in Table 4. Food consumption patterns show that a typical
household spends a larger share of food budget on essential food item,
i.e. on average cereals and pluses 22.1 percent, dairy products 23.34
percent, edible oils 9.26 percent, vegetables 8.6 percent, spice 3.28
percent and sugar 8.13 percent constituting 77.37 percent of total food
budget. Cereals and pluses, a basic foodstuff, sees its relative
importance in food budget share in urban and rural areas. In Pakistan it
is a major source of daily calories intake which constitute
approximately 62 percent of total calories intake [Grimard (1996)]. All
other essential food items represent a consistent consumption pattern in
urban and rural areas.
The concentration index of food expenditure component measures how
evenly of unevenly that component is distributed over the per capita
total expenditure. It can be seen that the concentration index of dairy
products, meat, poultry and fish, fruits (fresh and dried) and all
readymade food products are higher than overall Gini coefficient which
implies that expenditures on these commodities are unevenly distributed
over the total expenditure in favour of rich families. Wheat and rice
which is included in cereals and pluses category had high inequality in
rural area as compare to urban area. Concentration of all essential food
items is less as compare to total food expenditure in all regions.
Overall non essential food expenditure is more unevenly distributed as
compare to total food inequality. The last column shows the contribution
of food components inequality in total inequality. Dairy products and
meat, poultry and fish had major share in total inequality in all
regions. All essential food items which had 77.37 percent budget share
contribute 62.77 percent in total inequality.
To make statements about the magnitude of welfare, it will be
necessary to compute single measures of welfare. Statistics given in
Table 5 is based on Sen Welfare function (1974) given by Equation (1)
which takes into account both size and distribution of expenditure. It
is evident that cereals and pulses and dairy products are the major
source of welfare in urban, rural and overall Pakistan. Tea and tobacco
contribute least magnitude of welfare in Pakistan indicating that these
items cannot be categorised as social necessities.
While prices play an important role in our lives, inflation
undeniably affects the welfare of the poorest in society. Rise in food
prices have a big impact on the poorest households in rural and urban
areas. If the prices of essential food items increases faster than those
of luxuries, the poorest in first quintile will be hurt more than the
top 20 percent. If the prime concern is to protect the poor, it is
important to know how changes in prices affect their welfare. To
evaluate the effect of price changes on the aggregate welfare, price
elasticity of welfare is computed which indicates that if the price of
the jth commodity increases by 1 percent, then the aggregate welfare
changes by [n.sub.j] percent as in Equation (5). The price elasticity
welfare for cereals and pulses is 0.269 in Pakistan; it means if the
price of cereals and pulses increases by 1 percent aggregate welfare
decreases by 0.269 percent. It can also be analysed that essential food
item such as cereals and pulses, dairy products and edible oils have 55
percent budget share; if the price increases by one percent the
aggregate welfare will decrease by 0.58 percent. This phenomenon is
observed in recent food inflation in Pakistan where bottom 20 percent
population hurt proportionately more than the rich as they have greater
share of these essential items in the food budget. A welfare comparison
of the two components cannot be made without taking into account the
cost involved in increasing welfare. A progressivity index is also
computed to make this cost adjustment in the last column of Table 5. It
is the ratio of the jth inequality component to the aggregate
expenditure inequality implies that if its value is positive the jth
component is to be progressive and the negative value implies the jth
component to be regressive. It is observed that essential food
components except dairy products are the most progressive expenditure as
the poor spend a greater proportion of their income on food. If the
government decides to give subsidy on cereals and pluses the poor will
advantaged more than the rich. The magnitude of progressivity index
indicates whether the tax/subsidy on the jth commodity have an adverse
impact on the poor or the non poor or how prices changes affect
aggregate inequality in a society.
5. CONCLUSIONS
A central area of inquiry in stratification research concerns
income or expenditure inequality. Per capita inequality and welfare in
the distribution of food consumption expenditures across consumer units
has been focus of this research. Using the Kakwani (1980) disaggregation
of Gini index and Sen (1974) welfare function, inequality and welfare in
food expenditure components have been examined by employing full sample
of 'Household Integrated Economic Survey' 2005-06.
The total food expenditure is decomposed in essential and non
essential food components, having budget share of 77.37 percent and
22.63 percent respectively. Cereals and dairy products; the basic
foodstuff, find its relative importance in food budget share in urban
and rural areas. Inequality parameters indicate that distribution in
food expenditure is more pronounced in urban area as compare to rural
area. The decile dispersion ratio indicates that disparity in non
essential food expenditure is quite high within urban households. It can
be seen that the concentration index of dairy products, meat, fruits and
all readymade food products are higher than overall Gini coefficient
which implies that expenditures on these commodities are unevenly
distributed over the total expenditure in favour of rich families. It is
evident that cereals and dairy products are the major source of welfare
in urban, rural and overall Pakistan. The study also present results on
price elasticity of welfare and the progressivity index for each
component. The price elasticity welfare for cereals is 0.269 in Pakistan
which indicates that aggregate welfare is very responsive to price
changes. It can also be analysed that essential food items such as,
dairy products and edible oils having 32 percent food budget share are
fairly sensitive to price changes. This phenomenon is observed in recent
food inflation in Pakistan where bottom 20 percent population hurt
proportionately more than the rich as they have greater share of these
essential items in the food budget. It is also observed that essential
food components except dairy products are the most progressive
expenditure as the poor spend a greater proportion of their income on
food. If the government decides to give subsidy on cereals and pluses
the poor will advantaged more than the rich.
The present analysis provides a basis for determining the magnitude
of inequality and welfare that can be helped in designing appropriate
policies. As the changes in prices have generally affected the welfare
of the poor more adversely than the non- poor, safety nets will be
required for the most vulnerable population as they will not be able to
sustain their livelihoods.
Comments
This is an important paper in the areas of distribution and
welfare. Indeed, effective evaluation of food policy issues will be
possible through analysing distribution of food consumption patterns
which is the objective of this paper. First of all, I would like to
comment on page 5 where authors says that inequality in budget shares
for food subcategories for poorest quintile and richest quintile over
time in Pakistan is largely a reflection of poverty as people do not
have income for food. But I would comment that most of the poor do have
income for food as their food share is higher compared with the
non-poor. Therefore, most of the poverty is mainly due to insufficient
income to spend on the non-food need.
Secondly, I would like to point out that authors say at page 11
that consumption is also a better indicator measure of long term
standard of living than income. While current consumption expenditure
may be a good indicator on different account, one of the drawbacks of
consumption is that it gives a short run status of household resources.
It may be possible for the poor to finance their consumption payments by
dissaving of by borrowing--a process which cannot be sustained in the
long run. It is, therefore, recommended to examine welfare using both
income and consumption and draw inference from the comparison.
Thirdly, I would like to say that authors have used the per capita
consumption that ignored the differences needs and economies of scale in
household consumption. Different individuals of different age and sex in
a household have different requirement for food and non-food. For
example adult calorie needs are higher than children and similarly male
adult needs more calories than female adult. Since the paper focuses on
inequality on consumption in particular food, it is therefore, important
to use adult equivalent household consumption rather than per capita
household. While analysis is mostly focused on presenting the
statistical results, more discussion is required on policy implications
in the concluding section.
Talat Anwar
Canadian International Development Agency (CIDA), Programme Support
Unit,
Islamabad.
REFERENCES
Afridi, U., A. Qadir, and J. Zaki (1984) Effect of Dual Sector
Inflation across income Levels in Pakistan. The Pakistan Development
Review 23:2, 381-390.
Anwar, T. (2009) Measuring Inequality of Consumption and
Opportunities in Pakistan, 2001-02 and 2004-05. Pakistan Economic and
Social Review 47:2, 157-181.
Atkinson, A. B. (1970) On the Measurement of Inequality. Journal of
Economic Theory 2, 244-263.
Berry, C. J. (1994) The Idea of Luxury: A Conceptual and Historical
Investigation. Cambridg: Cambridge University Press.
Burney, N. A. and A. H. Khan (1991) Household Consumption Patterns
in Pakistan: An Urban-Rural Comparison Using Micro Data. The Pakistan
Development Review 30:2, 145-172.
Cheema, A. A. and M. H. Malik (1985) Changes in Consumption
Patterns and Employment under Alternative Income Distribution in
Pakistan. The Pakistan Development Review 24:1, 1-22.
Grimard, F. (1996) Does the Poor's Consumption of Calories
Respond to Changes in Income? Evidence from Pakistan. The Pakistan
Development Review 35:3, 257-283.
Haq, R. (1998) Trends in Inequality and Welfare in Consumption
Expenditure: The Case of Pakistan. The Pakistan Development Review 37:4,
257-283.
Haq, Z., H. Nazli, and K. Meilke (2008) Implications of High Food
Prices for Poverty in Pakistan. Agricultural Economics (39) Supplement
477-484. Department of Food, Agricultural and Resource Economics,
University of Guelph, Ontario, Canada.
Idrees, M. (2006) An Analysis of Income and Consumption
Inequalities in Pakistan. Unpublished PhD Dissertation, Department of
Economics, Quaid-i-Azam University, Islamabad.
International Food Policy Research Institute (2008) Global Hunger
Index. Retrieve: http://www.ifpri.org/pressrelease/new_global_hunger_index
Kakwani, N. (1980) Income Inequality and Poverty: Methods of
Estimation and Policy Applications. New York: Oxford University Press.
Kakwani, N. (1980) On the Class of Poverty Measures. Econometrica
48, 437-446.
Lazaridis, P. (2000) Decomposition of Food Expenditure Inequality:
An Application of the Extended Gini Coefficient to Greek Micro-data.
Social Indicators Research 52, 179-193.
Lorenz, M. O. (1905) Methods of Measuring Concentration of Wealth.
Journal of the American Statistical Association 9.
Naqvi, Z. F. and M. Akmal (2000) How Do the Poor Respond to Rising
Prices? The Pakistan Development Review 39:4, 827-842.
Pakistan, Government of (2002) Household Integrated Economic
Survey. Islamabad: Federal Bureau of Statistics.
Pakistan, Government of (2005) Household Integrated Economic
Survey. Islamabad: Federal Bureau of Statistics.
Pakistan, Government of (2006) Household Integrated Economic
Survey. Islamabad: Federal Bureau of Statistics.
Pakistan, Government of (2009) Pakistan Economic Survey 2008-09.
Islamabad: Finance Division, Economic Adviser's Wing.
Seale, J., R. Ahita, and B. Jason (2003) International Evidence on
Food Consumption Patterns. United States Department of Agriculture.
Retrieved from www.ers.usda.gov.
Sen, A. K. (1974) Information Base of Alternative Welfare Approach:
Aggregation and Income Distribution. Journal of Public Economics 44:1,
219-31.
UNDP (2003) Human Development Report 2000. New York: Oxford
University Press.
Veen, M. (2003) When Is Food a Luxury? World Archaeology 34:3,
Luxury Foods. pp. 405-427. Taylor & Francis, Ltd. Stable URL:
http://www.jstor.org/stable/3560194
Zakir, N. and M. Idrees (2009) Trends in Inequality, Welfare and
Growth in Pakistan, 1963-64 and 2004-2005. Pakistan Institute of
Development Economics, Islamabad. (PIDE Working Paper 2009:53).
(1) See, for example Burney and Khan (1991), Haq. Nazli and Meilke
(2008), Naqvi and Akbar (2000). Grimard (1996), Haq (1998), Anwar,
(2009) and Zakir and Idrees (2009).
Rashida Haq<rashida_haq@hotmail.com> and Nabeela Arshid
<nabeela@pide.org.pk> are Senior Research Economist and Senior
Systems Analyst at the Pakistan Institute of Development Economics,
Islamabad.
Authors' Note: We would like to thank Dr Attiya Javid, Dr
Najam us Saqib and Mr Masood Ishfaq Ahmed for their useful comments and
suggestions on an earlier draft of the paper.
Table 1
Percentage Distribution of Budget Shares for Food Sub-categories
Across Countries
Low Middle High
Income Income Income
Food Sub-categories Countries Countries Countries
Cereals/breads 26.97 17.98 11.83
Meat and Fish 19.65 22.48 23.41
Dairy 7.89 10.78 10.03
Fruits and Vegetables 20.34 18.24 14.62
Beverages and Tobacco 9.25 14.22 25.92
Fats and Oils 5.90 4.36 2.78
Other Foods 10.0 11.94 11.41
Food Expenditure Share as
% of Total Expenditure 52.58 34.69 16.97
Source: www.ers.usda.gov.
Table 2
Inequality in Monthly Budget Shares for Essential and Luxury Food
Items: Pakistan
2000-01 2004-05
Poorest Richest Poorest Richest
Food Categories 20% 20% 20% 20%
Cereals/Pulses 32.72 16.03 31.54 15.06
Dairy 15.02 24.50 18.51 25.03
Vegetables 9.50 8.16 10.02 7.61
Fats and Oils 8.44 6.69 10.47 8.11
Condiments and Spices 3.56 2.97 3.51 3.04
Gur and Sugar 8.94 6.99 7.14 5.30
Tea 3.38 2.56 2.77 2.08
Meat, Poultry and Fish 4.84 12.87 7.18 15.68
Fruits 1.67 5.12 2.28 5.58
Tobacco 4.36 3.13 2.69 2.79
Readymade Food 7.57 10.99 3.87 9.70
Food Expenditure as % of Total
Expenditure 63.44 49.33 59.22 40.57
2005-06
Poorest Richest
Food Categories 20% 20%
Cereals/Pulses 31.14 15.18
Dairy 17.44 26.19
Vegetables 10.33 7.36
Fats and Oils 9.49 7.72
Condiments and Spices 3.41 3.04
Gur and Sugar 9.79 7.12
Tea 3.04 2.04
Meat, Poultry and Fish 6.77 14.70
Fruits 1.90 5.55
Tobacco 3.72 2.46
Readymade Food 2.98 8.62
Food Expenditure as % of Total
Expenditure 55.61 33.1
Source. 'Household Integrated Economic Survey' for respective years.
Table 3
Inequality of Per capita Household Food Consumption
Expenditure: 2005-06
Share of Total Food
Expenditure (%)
Quintiles Pakistan Urban Rural
1st 9.7 7.3 11.4
2nd 13.1 10.6 14.7
3rd 16.8 14.9 17.9
4th 22.5 21.6 23.0
5th 37.8 45.8 32.9
Gini Index 26.27 27.10 25.34
DDR 25.66 15.94 34.65
Share of Total Food
Expenditure (%)
Quintiles Pakistan Urban Rural
1st 11.1 8.6 12.4
2nd 14.4 12.0 15.7
3rd 17.8 16.4 18.6
4th 23.1 22.7 23.3
5th 33.7 40.3 30.0
Gini Index 21.0 20.52 21.80
DDR 32.93 21.34 41.33
Share of Total Food
Expenditure (%)
Quintiles Pakistan Urban Rural
1st 5.7 3.7 7.4
2nd 9.0 6.8 10.9
3rd 13.1 10.8 15.1
4th 20.4 18.8 21.9
5th 51.9 60.0 45.7
Gini Index 43.0 44.37 39.9
DDR 10.98 8.33 16.19
Source: Computations are based on 'Household Integrated Economic
Survey (HIES), 2005-06'.
Table 4
Food Expenditure Inequality by It's Components: 2005-06
% Share of Food
Expenditure
Food Expenditure Components Pakistan Urban Rural
Essential Food Components
Cereals and Pluses 22.15 18.54 24.36
Dairy Products 23.34 23.92 22.98
Edible Oils and Fats 9.26 8.60 9.66
Vegetables 8.60 8.15 8.88
Condiments and Spices 3.28 3.51 3.13
Gur and Sugar 8.13 6.97 8.84
Tea 2.61 2.60 2.61
Non Essential Food Components
Meat, Poultry and Fish 11.26 14.49 9.29
Fruits (Fresh and Dried) 4.03 5.25 3.28
Tobacco 3.29 3.00 3.47
Readymade Food Products 4.05 4.96 3.50
Total Food Expenditure 100.00 100.00 100.00
All Essential Food Expenditure 77.37 72.30 80.47
Non Essential Food Expenditure 22.63 27.70 19.53
Concentration
Index (%)
Food Expenditure Components Pakistan Urban Rural
Essential Food Components
Cereals and Pluses 11.2 9.60 12.73
Dairy Products 33.0 30.30 34.03
Edible Oils and Fats 21.0 18.80 22.26
Vegetables 18.66 18.91 18.49
Condiments and Spices 21.6 21.0 20.84
Gur and Sugar 19.47 20.19 19.81
Tea 17.67 19.25 16.19
Non Essential Food Components
Meat, Poultry and Fish 40.46 41.45 36.61
Fruits (Fresh and Dried) 41.69 41.87 38.34
Tobacco 23.8 26.18 22.78
Readymade Food Products 67.9 66.54 67.50
Total Food Expenditure 26.27 27.10 25.34
All Essential Food Expenditure 21.0 20.52 21.8
Non Essential Food Expenditure 43.0 44.37 39.97
% Contribution to total
Inequality
Food Expenditure Components Pakistan Urban Rural
Essential Food Components
Cereals and Pluses 9.44 6.64 12.25
Dairy Products 29.32 26.57 30.83
Edible Oils and Fats 7.40 5.90 8.30
Vegetables 6.11 5.54 6.32
Condiments and Spices 2.70 2.58 2.77
Gur and Sugar 6.03 5.17 7.11
Tea 1.76 1.85 1.58
Non Essential Food Components
Meat, Poultry and Fish 17.34 22.14 13.44
Fruits (Fresh and Dried) 6.40 8.12 5.14
Tobacco 2.99 2.95 3.16
Readymade Food Products 10.48 12.18 9.49
Total Food Expenditure 100 100 100
All Essential Food Expenditure 62.77 54.61 69.17
Non Essential Food Expenditure 37.23 45.39 30.83
Computations are based on 'Household Integrated Economic Survey
(HIES), 2005-06'.
Table 5
Welfare by Food Expenditure Components: 2005-06
Food Expenditure Components % Contribution to Total
Welfare
Pakistan Urban Rural
Cereals and Pulses 26.90 23.00 28.48
Dairy Products 21.33 22.88 20.31
Edible Oils and Fats 9.98 9.59 10.06
Vegetables 9.55 9.07 9.69
Condiments and Spices 3.50 3.80 3.32
Gur and Sugar 8.94 7.63 9.50
Tea 2.93 2.88 2.93
Meat, Poultry and Fish 9.15 11.64 7.89
Fruits (Fresh and Dried) 3.20 4.19 2.71
Tobacco 1.44 3.03 3.59
Readymade Food Products 4.21 2.28 1.52
Food Expenditure 100 100 100
All Essential Food Expenditure 83.39 78.35 84.28
Non Essential Food Expenditure 17.60 21.65 15.72
Food Expenditure Components Price Elasticity of
Welfare by Components
Pakistan Urban Rural
Cereals and Pulses 0.269 0.23 0.28
Dairy Products 0.213 0.23 0.20
Edible Oils and Fats 0.100 0.10 0.10
Vegetables 0.095 0.09 0.10
Condiments and Spices 0.035 0.04 0.03
Gur and Sugar 0.089 0.08 0.10
Tea 0.029 0.03 0.03
Meat, Poultry and Fish 0.091 0.12 0.08
Fruits (Fresh and Dried) 0.032 0.04 0.03
Tobacco 0.014 0.03 0.04
Readymade Food Products 0.042 0.02 0.02
Food Expenditure 1 1 1
All Essential Food Expenditure 0.834 0.783 0.843
Non Essential Food Expenditure 0.177 0.216 0.157
Food Expenditure Components Progressivity Index
by Components
Pakistan Urban Rural
Cereals and Pulses 0.214 0.240 0.169
Dairy Products -0.086 -0.044 -0.116
Edible Oils and Fats 0.078 0.114 0.041
Vegetables 0.110 0.112 0.092
Condiments and Spices 0.070 0.084 0.060
Gur and Sugar 0.099 0.095 0.074
Tea 0.124 0.108 0.123
Meat, Poultry and Fish -0.188 -0.197 -0.151
Fruits (Fresh and Dried) -0.205 -0.203 -0.174
Tobacco 0.040 0.013 0.034
Readymade Food Products -0.562 -0.541 -0.565
Food Expenditure 0 0 0
All Essential Food Expenditure 0.078 0.090 0.047
Non Essential Food Expenditure -0.222 -0.237 -0.196
Source: Computations are based on "Household Integrated Economic
Survey (HIES), 2005-06".
