Consumption patterns of male and female headed households in Pakistan: evidence from PSLM 2007-08.
Khan, Ashfaque H. ; Khalid, Umer
I. INTRODUCTION
Recent years have witnessed growing interest in analysing the
welfare outcomes of female headed households (FHHs) in the developing
world. The theoretical argument for examining female headship and family
welfare is underpinned by two important considerations. The first
concerns households' access to resources, while the second pertains
to control over the allocation of resources within the household
[DeGraff and Bilsborrow (1993)]. A priori female headed households are
expected to have access to a lower level of resources than the
conventional male-headed households for a variety of reasons. (1)
However, this lower resource envelop experienced by female headed
households may be partially offset by the way resources are allocated
within such households. Several studies have revealed that resources
under the control of women are more likely to be allocated for
productive purposes that promote family welfare as compared to resource
allocation under the control of men. In the context of Pakistan, the
present paper aims to explore how resource allocation within female
headed households differs from male headed households by examining the
consumption patterns of both female and male headed households in the
country.
The study will make use of the Engel curve framework, which shows
the relationship between a household's expenditure on a particular
good and total household income, holding prices constant. The Engel
curve framework has been used in a large strand of empirical literature
examining household consumption behaviour, for both the developed and
developing countries, including Pakistan. In case of Pakistan, a large
number of studies have examined household consumption patterns for
Pakistan as a whole and / or by its urban-rural regions. (2) More
recently, household consumption behaviour across the four provinces of
the country has also been investigated [Khan and Khalid (2011)]. To our
knowledge, no study has so far examined separately the consumption
behaviour of female and male headed households in Pakistan.
An analysis of household budgets by gender of the household head is
necessary as research from other developing countries shows differential
patterns of consumption for households headed by females [Blumberg
(1988); Handa (1996); Panda (1997); Seebens (2009); Yabut-Bemardino
(2011); Michael, et al. (2010) and Dokah and Amikuzuno (2011)]. It is
observed that female headed households tend to expend a proportionally
higher share of their spending on more productive avenues, like human
capital development (education, health) compared to the conventional
male headed households. The present study will attempt to empirically
test for the heterogeneity of household consumption patterns across
female and male headed households across Pakistan, as well as by the
urban rural divide within the country.
The paper will estimate average and marginal expenditure shares and
expenditure elasticities at the national level and by urban/rural areas
as well as by expenditure quintiles for these two sets of households,
using data from a recent round of nationally representative household
survey--the Pakistan Social and Living Standards Measurement Survey
(PSLM) 2007-08. The study would make use of the 12 broad commodity
groupings employed by Khan and Khalid (2010) to examine the consumption
behaviour and the analysis by consumption quintiles will illustrate how
the consumption behaviour differs by economic status for these two sets
of household.
At the onset, it would be important to keep in mind that the
analysis of consumption behaviour of female headed households being
presented is preliminary and stylistic, as women defined as household
heads in traditional societies like Pakistan while being socially
empowered (3) may lack economic empowerment. Moreover, empirical
evidence from developing countries clearly shows that female headed
households are a heterogeneous group in terms of size, age structure,
marital status and employment opportunities. An important distinction in
this regard is whether the female headship of a household is de facto or
de jure [Lampiettii and Stalker (2000)]. In de facto female-headed
households, the husband is not present in the community at the time of
the survey, but may still play an active role in supporting the
household through remittances. On the other hand, in de jure
female-headed households there is no adult male present with widows,
divorced women, or unmarried women being the household heads. These
caveats need to be kept in mind while drawing any policy inferences from
the findings of this paper.
The layout of the paper is as follows: Section II presents the
theoretical framework and methodology used in the analysis, while
Section III discusses the data. Results are reported and discussed in
Section IV. The analysis of consumption behaviour of MHHs and FHHs by
expenditure quintiles is presented in Section V. The final section
presents the concluding remarks.
II. THEORETICAL FRAMEWORK
The Engel curve is a demand function derived from a constrained
utility maximisation problem, which can be expressed mathematically as;
(4)
[E.sub.i] = [[alpha].sub.i] + [[beta].sub.i] Y + [u.sub.i] (1)
Where [E.sub.i], = Expenditure on commodity i = [p.sub.i]
[x.sub.i];
Y = Total income
[u.sub.i] = Random error term
This framework of analysis is underpinned by two important
assumptions; (i) all households face the same prices for each commodity,
and (ii) all households have the same underlying utility function. These
assumptions are, however, quite strong and are likely to introduce bias
in the estimated parameters. The limitations of these assumptions are
discussed in detail in Khan and Khalid (2010).
Being derived from constrained maximisation, the Engel curve has to
satisfy the general restrictions of demand theory. With prices assumed
to be the same across households when using cross section data, the
restrictions in terms of price derivatives such as homogeneity, symmetry
and negativity of own price effect cannot be tested. Only the
'adding up' condition is left to be tested, which in terms of
the parameters of Equation (1) implies that [[alpha].sub.i] and
[[beta].sub.i], sum to zero and unity, respectively; i.e., [summation]
[[alpha].sub.i], = 0 and [summation] [[beta].sub.i], = 1.
For the estimation of Engel Curve as outlined in Equation (1), the
choice of an appropriate functional form is also an important issue
which has been the subject of many empirical studies. Various functional
forms, which includes linear; semi-logarithmic, double logarithmic, etc.
have been used in the literature, but consensus on the most appropriate
form has not been developed. In this study, we make use of the linear
and double-logarithmic forms which have also been used by Khan and
Khalid (2010), to ensure direct comparability of results with the
earlier study.
The household income and/or consumption expenditure has generally
been employed as explanatory variables in empirical work examining
household consumption patterns. As household consumption is a preferred
welfare indicator over household income in developing countries [Deaton
(1997); Glewwe, et al. (2001)], this study would use household
consumption expenditure as the explanatory variable in estimation of the
Engel Curve equation. In addition, we would also use the household size
as an explanatory variable to capture the effect of economies of scale
in consumption in large households, which Houthakker (1957) has referred
to as a combination of two effects--the specific effect and the income
effect.
In this study, we use the household consumption expenditure and
household size as explanatory variables to estimate the following linear
and double log functional form of the Engel curve, respectively;
[E.sup.q.sub.ij] = [[alpha].sup.q.sub.ij] + [[beta].sup.q.sub.ij] +
[r.sup.q.sub.ij] [HS.sup.q.sub.j] (2)
ln[E.sub.ij] = [[theta].sub.ij] + [y.sub.ij] ln[E.sub.j] +
[s.sub.ij]ln[HS.sub.j] (3)
where j = 1,2, ... k households
q = male, female sex of household head
i = 1,2, ... n commodity groups
[E.sup.q.sub.ij] = Expenditure of jth household on ith commodity in
qth headed household
[E.sup.q.sub.j] = Total consumption expenditure of jth household in
qth headed household
[HS.sup.q.sub.j] = Household size of jth household in qth headed
household
[[beta].sup.q.sub.ij] = Expenditure share of ith commodity in total
household expenditure of households with qth head
[r.sup.q.sub.ij] = Change in expenditure on ith commodity of
households with qth head
[y.sub.ij] = Expenditure elasticity of ith commodity with respect
to total expenditure of households with qth head.
[s.sub.ij] = Expenditure elasticity of ith commodity with respect
to household size for households with qth head.
III. DATA
The study is based on the micro data tapes of the Pakistan Social
and Living Standards Measurement Survey (PSLM) 2007-08 conducted by the
Federal Bureau of Statistics. This nationally representative survey
consists of data on a sample of 15,512 households. Out of this sample,
observations for 4 households having household size greater than 34 were
dropped from analysis. Thus, the analysis carried out in this paper is
based on a sample of 15,508 households across the urban and rural areas
of Pakistan, out of which 14,275 households are male headed, while 1233
households are female headed (Table 1). Overall, the female headed
households represent around 8 percent of the sample, with their share
being slightly higher in urban areas at 8.4 percent, compared to the
rural areas at 7.7 percent.
In the second part of the paper where the analysis of consumption
behaviour of both male and female headed households is carried out by
expenditure quintiles, the distribution of the sample by both set of
households is given in Table 2.
The examination of consumption patterns of male and female headed
households is carried out for the 12 commodity groupings used by Khan
and Khalid (2010). These consumption categories include food and drinks,
clothing and footwear, fuel and lighting, housing, transport and
communications, household effects, personal effects, health care,
education, entertainment, durables and miscellaneous items. The details
of commodities covered within each of the 12 groups are given in
Appendix 1.
The survey data contains information on both the amount spent on
purchase of a particular commodity as well as its imputed value in case
it is self-produced and / or received as gift in kind. For the purpose
of this study, we group together both these two sets of information to
get the total expenditure on each commodity, which is the amount spent
on buying that commodity plus its imputed value.
The average expenditure shares of the 12 groups of commodities for
both male headed and female headed households are reported in Table 4a
for overall Pakistan and in Table 4b for male and female headed
households across both the urban and rural sectors, respectively. In
both the tables, the results of the two sample t-test with equal
variance are also presented to test for the significance of difference
between the budget shares of MHHs and FHHs at the national level as well
as by the urban and rural areas, respectively. Overall, expenditures on
food and drinks, followed by housing account for the highest share of
total household consumption expenditures for both male and female headed
households at the national level. However, there exist statistically
significant differences between the expenditure shares of male and
female headed households in some expenditure categories, as shown by the
two sample t-test. These include the higher budget shares of female
headed households for education (3.72 percent vs. 2.51 percent), housing
(14.85 percent vs. 12.87 percent), fuel and lighting (9.21 percent vs.
8.02 percent), clothing and footwear (5.78 vs. 5.46 percent) and
household effects (0.72 percent vs. 0.6 percent) and lower average
expenditures on food and drinks (44.25 percent vs. 48.34 percent) and
transport and communications (4.28 percent vs. 5.72 percent), compared
to their male headed counterparts. Female headed households are seen to
be spending slightly more on healthcare also but the result is only
significant at the lower 10 percent level of significance.
The urban-rural disaggregation of consumption patterns of male and
female headed households reveals broadly similar patterns as observed at
the national level (Table 4b). Compared to their male headed
counterparts, female headed households spend significantly more, on
average, on education, housing and fuel and lighting across both the
urban and rural areas of the country, while female headed households
residing in rural (urban) areas spend significantly more on clothing and
footwear and household effects (entertainment). Average expenditure
shares of female headed households on food and drinks and transport and
communication are significantly lower than those of their male headed
counterparts across both the urban and rural sectors.
IV. RESULTS
The results of the empirical analysis of household consumption
patterns for male and female headed households, at the national level as
well as by urban and rural areas, are presented and discussed in this
section. The Engel curves have been estimated using both the linear and
double log functional forms, employing the Ordinary Least Squares (OLS)
method. The estimated marginal expenditure shares for the 12 commodity
groups are reported in Table 5 for both set of households for overall
Pakistan as well as by the urban and rural sectors.
It is seen that the marginal expenditure shares are the highest for
housing, durables and food and drinks for MHHs, while in case of FHHs
they are highest for durables, followed by housing and food and drinks.
One noteworthy finding is the higher marginal expenditures by FHHs on
education and durables in comparison to their male counterparts, with
this result being consistent in the urban-rural areas as well.
The regional breakup reveals that in urban areas, both MHHs and
FHHs have the highest levels of marginal expenditures on housing,
followed by food and drinks for MHHs and durables for FHHs. In the rural
sector, both MHHs and FHHs have the highest spending at the margin on
durables, although the magnitude for FHHs is much higher--they spend Rs
0.43 on this category for every one rupee of increase in total
consumption expenditure, compared to Rs 0.3 spent by their male
counterparts. MHHs in the rural sector spend considerably more at the
margin on food and drinks as compared to their female counterparts.
The analysis of expenditure elasticities, reported in Table 6 shows
that for both set of households being investigated; education, transport
and communications, household effects, durables, entertainment and
miscellaneous items can be considered as luxury goods--having
expenditure elasticities in excess of unity, at the national level as
well as by the urban-rural divide. Housing is a luxury good for both
MHHs and FHHs at national and regional level except for MHHs residing in
rural areas of the country, while entertainment is observed to a luxury
for MHHs at the national level and for both MHHs and FHHs in the rural
sector. The remaining commodity groups are necessitates, with healthcare
being a luxury only for MHHs in the rural sector.
V. CONSUMPTION PATTERNS BY EXPENDITURE QUINTILES
The preceding analysis presented an overview of the consumption
patterns of both male headed and female headed households, at the
national level as well as its disaggregation by urban-rural sectors.
This analysis, however, is likely to mask considerable variation in the
consumption behaviour as the economic status of both set of households
changes. In order to examine the consumption patterns by different
economic status of households, we analyse the consumption behaviour of
MHHs and FHHs by consumption expenditure quintiles. The consumption
expenditures of male and female headed households by the five
expenditure quintiles given in Table 7 show that in the first three
expenditure quintiles, the consumption expenditures of male headed
households are higher than those of their female counterparts. However,
the expenditure gap between male and female headed households narrows
down as one moves up the quintiles, from 92 percent of MHHs'
expenditures in first quintile to 99.8 percent in the third. In the
highest two expenditure quintiles, the consumption expenditures of
female headed households are, on average, slightly higher than those of
male headed households.
The average expenditure shares by expenditure quintiles with
respect to the 12 commodity groups being examined for MHHs and FHHs are
given in Table 8. (5) At a first glance, it can be seen that the
Engel's Law is being validated, as the share of household
expenditures allocated to food and drinks declines progressively with
improvement in the economic status (as proxied by the expenditure
quintiles) of both MHHs and FHHs. It can also be seen that for each
expenditure quintile, the proportionate spending of MHHs on food and
drinks is higher as compared to the FHHs.
The analysis further shows that the average budget shares of both
MHHs and FHHs rise across the expenditure quintiles for housing,
transport and communications, household effects, education,
entertainment and miscellaneous items; with the average spending of FHHs
being higher in all expenditure quintiles for housing, household
effects, education and miscellaneous items and lower in all quintiles
for transport and communications. In case of clothing and footwear, fuel
and lighting and personal effects; the average expenditure shares for
both MHHs and FHHs decline with each expenditure quintile. The average
expenditures on durables rise progressively with each expenditure
quintile, in case of MHHs; while for FHHs proportionate spending
declines between the first and second quintile but increases
subsequently in the remaining three quintiles.
The marginal expenditure shares of MHHs and FHHs by expenditure
quintiles are presented in Table 9, which show mixed and varying trends
and patterns with respect to the gender of household head and
expenditure quintiles. Among the MHHs, households in the bottom and
middle expenditure quintile have highest spending at the margin on food
and drinks, followed by housing and miscellaneous items, while the
households in the top quintile have highest marginal spending on
durables, housing and miscellaneous goods. The marginal expenditures of
MHHs on food and drinks decline progressively with each expenditure
quintile, while in case of FHHs they decline in the first two
expenditure quintiles, peak in the third quintile and then again fall in
the 4th and 5th quintiles.
In case of education, it is seen that marginal expenditure shares
of MHHs rise successively from the first to fourth quintile and fall in
the highest quintile. For FHHs, marginal spending on education rises in
the first two expenditure quintiles, while they decline from the fourth
to fifth quintile. However, it is worth pointing out that the marginal
expenditures of FHHs on education in all quintiles are higher than those
of MHHs in the corresponding quintiles with the exception of the third
expenditure quintile. This is an important finding which shows that
controlling for household economic status; FHHs tend to spend more on
education both proportionately as well as at the margin, compared to
their male headed counterparts.
VI. CONCLUDING REMARKS
The purpose of this study has been to empirically test for the
homogeneity of household consumption patterns of male headed and female
headed across Pakistan as well as explore the urban-rural variations
within each set of households. The paper estimated average expenditure
shares, marginal expenditure shares and expenditure elasticities for
both MHHs and FHHs at the national level as well as by urban/rural
sectors within each province, using household level micro data for the
year 2007-08. In addition, the average expenditure shares, marginal
expenditure shares and expenditure elasticities for both MHHs and FHHs
are also estimated by consumption quintiles to analyse how consumption
behaviour differs by economic status for these two set of household.
The results of this preliminary analysis show that household
consumption patterns are not homogeneous across the male and female
headed households at the national level and also exhibit variations
across the urban/ rural divide within the country. Overall, expenditures
on food and drinks, followed by housing account for the highest share of
total household consumption expenditures for both male and female headed
households at the national level. FHHs are seen to have higher budget
shares for education, housing, fuel and lighting, clothing and footwear
and household effects and lower average expenditures on food and
drinks and transport and communications compared to their male headed
counterparts. The findings with regards to the expenditure shares of
FHHs on education, housing, food and drinks and communication and
transport; also hold across the urban-rural divide.
Our findings thus lend support to the existing body of evidence
from developing world that female headed households allocate a greater
share of financial resources to activities that promote human capital
formation and development. In terms of the marginal expenditure shares,
it is seen that MHHs spend the highest at the margin on housing,
durables and food and drinks, while FHHs' marginal expenditures are
highest for durables, followed by housing and food and drinks. It is
observed that FHHs have a higher marginal spending on education and
durables in comparison to their male counterparts, with this result
being consistent in the urban-rural areas as well. In the rural sector,
both MHHs and FHHs have the highest spending at the margin on durables,
although the magnitude for FHHs is much higher--they spend Rs 0.43 on
this category for every one rupee of increase in total consumption
expenditure, compared to Rs 0.3 spent by their male counterparts. MHHs
in the rural sector spend considerably more at the margin on food and
drinks as compared to their female counterparts.
The study also examined the consumption behaviour of MHHs and FHHs
by expenditure quintiles, to determine how patterns of consumption
varied with the economic status of both set of households. The analysis
of average expenditure shares of both set of households by expenditure
quintiles clearly shows that the Engel's Law is validated as the
share of household expenditures allocated to food and drinks declines
with improvement in the economic status (as proxied by the expenditure
quintiles) of both MHHs and FHHs. It can also be seen that for each
expenditure quintile, the proportionate spending of MHHs on food and
drinks is higher as compared to the FHHs. In general, the analysis shows
that the patterns and trends seen for the MHHs and FHHs at national
level previously also hold broadly in the disaggregated analysis by
expenditure levels across both set of households. In this regard, an
important finding is the comparatively higher budgetary shares of FHHs
on education and healthcare across all expenditure quintiles compared to
the MHHs.
The exploratory analysis of the consumption behaviour by gender of
the household head presented in this study offers useful insights into
how consumption patterns differ across male and female headed households
in the country. In terms of providing guidelines for the targeting of
anti-poverty programmes and interventions in the country, the findings
of this study tend to support the notion that targeting of female headed
households at the bottom of the expenditure distribution can be a more
effective mechanism for eliminating inter-generational poverty traps.
However, this implication needs to be offset by findings of previous
research on socio-economic determinants of female headed households
living below the poverty line in Pakistan, which show that more than 70
percent of female headed households in the country were receiving
domestic and/or foreign remittances during the period 2000-04. (6)
In order to draw more conclusive and concrete policy inferences,
this analysis needs to be supplemented by more in-depth research into
the socio-economic determinants of female headed households in the
country, which is beyond the scope of this study and can be taken up as
topic for future research. This research should focus on examining among
other things, the overall composition of female headed households
(proportion that is de facto and de jure), the educational status, the
occupational grouping and the sources of income.
Appendix 1
Details of Commodity Groups
1. Food and Drinks Milk and milk products, meat
poultry and Fish, fresh
fruits, dry fruits and nuts,
cereals, pulses, edible oils
and fats, tea and coffee,
baked and fried products,
miscellaneous food items,
2. Clothing and Clothing, clothing material
Footwear and services, footwear and
3. Fuel and Lighting repair charges, other expenses
on tire, tube, spare parts,
repairs of vehicle etc. and
service charges. Gas,
electricity, fire-wood,
kerosene oil, other household
effects (bulbs, tubes,
switches, battery cells, lamp
shades etc.)
4. Personal Effects Personal care articles,
personal care services,
household laundry, cleaning
and paper articles, personal
durable effects (wrist/pocket
watches, sun glasses, etc.),
laundry and cleaning equipment
(washer/dryer, vacuum cleaner,
iron, iron board, etc.)
5. Housing House rent and housing
expenses, house and property
tax etc.
6. Transport and Personal transport and
Communications travelling, petrol charges,
repairing of wheel puncture,
annual driving license fee,
expenses on traveling by road
by train and by air, vehicle
registration fee, etc.
7. Household Effects Readymade pillow covers, bed
sheets, blankets, curtains,
mosquito nets etc., purchase
of cloth (for pillow covers,
bed sheets quilts etc.) and
purchase of cotton (for
quilts, pillows, etc.),
carding and other stitching
charges on household textile,
chinaware, silverware and
kitchen equipment, furniture,
fixture and furnishing, other
household effects,
8. Healthcare Purchase of medicine,
hospitalisation expenses,
medical fees, laboratory and
physician's charges.
9. Education School/college fees and
private tuition fees, books
and exercise note
books/copies, stationary etc.
other education expenses
(bags, professional society
membership, transportation
etc.), hostel expenses,
calculators, personal
computers, mobiles etc.,
10. Entertainment Recreation and reading,
expenditure on hobbies, cable
installation recreational
membership fee, toys, games,
photography, lodging charges
etc., radio and musical
instruments (tape recorder,
gramophone, TV, VCR, VCP,
cassettes), recreational
equipment (cameras, projector,
shot gun, angling kit, bats,
balls etc.)
11. Durables Electric/ oil fans (table,
pedestal, ceiling, exhaust),
Air conditioners, air coolers,
refrigerators, freezers etc.
heater, boiler, geyser
(electric, gas, oil), table
lamp, sewing machine, knitting
machine (electric / hand),
Other (trunks, suitcase etc.),
Wall / table clock, water
pipes (rubber, nylon,
plastic), thermos bottle etc.,
service and repair charges of
household effects, etc.
mentioned above, transport and
travelling vehicles (bicycle,
Motorcycle, scooter, car,
horses, camels, tongas etc.)
12. Miscellaneous Stationery supplies such as
pen, pencils, stapling
machine, pin etc. (other than
education purpose), crockery
and cutlery for daily use,
taxes and fines and all other
miscellaneous expenditure,
personal effects and service
and repair charges
Appendix 2
Expenditure Elasticities for Different Commodity Groups, by Gender of
Household Head and Expenditure Quintiles
Male
Commodity Groups Q1 Q2 Q3 Q4 Q5
Food and Drinks 0.919 0.835 0.672 0.707 0.509
Clothing and
Footwear 0.808 0.804 0.826 0.702 0.620
Fuel and Lighting 0.S62 0.693 0.961 0.654 0.647
Housing 1.250 1.182 1.074 1.438 1.300
Transport and
Communications 1.388 1.314 0.954 1.709 1.188
Household Effects 1.448 0.605 * 1.126 1.545 1.063
Personal Effects 0.819 0.877 0.873 1.060 0.799
Healthcare 0.882 0.951 1.329 0.666 0.620
Education 1.242 2.074 2.613 2.070 1.394
Entertainment 2.122 1.018 * 1.479 1.097 0.867
Durables -0.055 * 2.268 1.433 * 0.376 * 1.534
Miscellaneous 1.714 1.424 1.445 1.490 1.206
Female
Commodity Groups Ql Q2 Q3 Q4 Q5
Food and Drinks 0.748 0.816 1.112 0.788 0.513
Clothing and
Footwear 0.694 0.789 0.545 * 0.243 * 0.555
Fuel and Lighting 0.773 0.674 * 1.051 0.478 * 0.651
Housing 1.258 0.899 * 0.399 * 0.712 * 1.145
Transport and
Communications 1.250 1.370 0.784 * 1.229 1.325
Household Effects 0.900 0.216 * 0.317 * 0.324 * 0.568
Personal Effects 0.863 0.799 1.521 0.833 0.760
Healthcare 0.884 1.791 0.679 * 0.462 * 0.587
Education 2.096 1.986 -0.402 * 2.073 0.974
Entertainment 0.734 * -0.968 * -1.836 * 2.332 1.004
Durables 0.525 * 5.719 2.192 * 2.240 * 1.226
Miscellaneous 1.901 2.158 2.159 1.810 1.155
* Not significant at 5 percent level of significance.
REFERENCES
Blumberg, R. (1988) Income under Female versus Male Control.
Journal of Family Issues 9:1.
Bruce, J. and D. Dwyer (1988) A Home Divided: Women and Income in
the Third World. Stanford: Stanford University Press.
Buvinic, M., M. A. Lycette, and W. P. McGreevey (1983) Women and
Poverty in the Third World. Baltimore: Johns Hopkins University Press.
Chernichovsky, D. (1985) Socioeconomic and Demographic Aspects of
School Enrolment and Attendance in Rural Botswana. Economic Development
and Cultural Change 33:2, 319-332.
Deaton, Angus S. (1988) Quality, Quantity, and Spatial Variation of
Price. American Economic Review 78:3, 418-430.
Deaton, Angus S. (1997) The Analysis of Household Surveys: A
Micro-econometric Approach to Development Policy. Johns Hopkins
University Press.
DeGraff, D. S. and R. E. Bilsborrow (1993) Female-headed Households
and Family Welfare in Rural Ecuador. Journal of Population Economics 6,
317-336.
Donkoh, S. A. and J. A. Amikuzuno (2011) The Determinants of
Household Education Expenditure in Ghana. Educational Research and
Reviews 6:8, 570-579.
Glewwe, P. P., M. M. Gragnolati, and H. H. Zaman (2002) Who Gained
from Vietnam's Boom in the 1990s? Economic Development and Cultural
Change 50:4.
Handa, S. (1996) Expenditure Behaviour and Children's Welfare:
An Analysis of Female Headed Households in Jamaica. Journal of
Development Economics 50, 165-187. Houthakker, H. S. (1957) An
International Comparison of Household Expenditure Pattern, Commemorating
the Century of Engel's Law. Econometrica 25:1-4, 532-551.
Khalid, U. and S. Akhtar (2011) Poverty Dynamics of Female Headed
Households in Pakistan: Evidence from PIHS 2000-01 and PSLM 2004-05.
Pakistan Institute of Development Economics, Islamabad. (Working Paper
No. 2011:80).
Khan, A. and U. Khalid (2010) Household Consumption Patterns in
Pakistan: Evidence from Household Data. NUST Journal of Business and
Economics 3:1, 1-14.
Khan, A. and U. Khalid (2011) Is Consumption Pattern Homogeneous in
Pakistan? Evidence from PSLM 2007-08. The Pakistan Development Review
50:4.
Kossoudji, S. and E. Mueller (1983) The Economic and Demographic
Status of Female-Headed Households in Rural Botswana. Economic
Development and Cultural Change 31, 831-859.
Lampietti, J. A. and L. Stalker (2000) Consumption Expenditure and
Female Poverty: A Review of Evidence. Policy Research Report on Gender
and Development. The World Bank Development Research Group. (Working
Paper Series No. 11).
Levin, C. E., D. G. Maxwell, M. Armar-Klemesu, M. T. Ruel, S. S.
Morris, and C. Ahiadeke (1999) Working Women in An Urban Setting:
Traders, Vendors, and Food Security in Accra. (FCND Discussion Paper No.
66).
Munisamy, T. M., S. A. Haron, and B. C. Yin-Fah (2010) Human
Capital Investment Expenditures of Women of Female-Headed Household in
Peninsular Malaysia. Asian Social Science 6:4, 31-38.
Panda, P. K. (1997) Female Headship, Poverty and Child Welfare: A
Study of Rural Orissa, India. Thiruvananthapuram: Center for Development
Studies.
Rosenhouse, S. (1988) Identifying the Poor: Is Headship a Useful
Concept? Paper prepared for ICRW and Population Council Seminar Series
on the Determinants and Consequences of Female-Headed Households.
Seebens, H. (2009) Child Welfare and Old-Age Security in Female
Headed Households in Tanzania. Bonn. (IZA Discussion Paper No. 3929).
Yabut-Bemardino, N. (2011) Implications of Female Household
Headship on Consumption Expenditure for Children's Welfare:
Philippine Case.
(1) For a detailed exposition, see Buvinic, et al. (1983);
Kossoudji and Mueller (1983); Chemichovsky (1985); Bruce and Dwyer
(1988) and Rosenhouse (1988).
(2) For an overview of the empirical research for Pakistan, see
Khan and Khalid (2010)
(3) An elder woman may be declared as the head of household out of
respect by her family.
(4) The derivation of the Engel Curve is discussed in Khan and
Khalid (2010).
(5) Due to the small sample size of FHHs by each expenditure
quintile, it was not possible to apply the two sample t-test to check
for the statistical significance of the difference between the average
expenditure shares of MHHs and FHHs.
(6) The study, Khalid and Akhtar (2011) used data from the Pakistan
Integrated Household Survey 200001 and the Pakistan Social and Living
Standards Measurement Survey 2004-05.
Ashfaque H. Khan <akhan@nbs.edu.pk> is Principal and Dean,
NUST Business School, National University of Sciences and Technology,
Islamabad. Umer Khalid <umerkhalid@hotmail.com> is Industrial
Policy Advisor, Ministry of Finance, Islamabad.
Authors' Note: They are extremely grateful to Dr Lubna Shahnaz
for her valuable comments on an earlier draft of the paper. The authors
alone are responsible for any errors and omissions.
Table 1
Distribution of Sample Size, by Gender and Region
Total Urban Rural
Overall 15508 6253 9255
Male 14275 5730 8545
Female 1233 523 710
Table 2
Distribution of Sample Size, by Gender of Household
Head and Expenditure Quintile
Male Female Total
Quintile 1 2736 366 3102
Quintile 2 2866 236 3102
Quintile 3 2896 205 3101
Quintile 4 2898 204 3102
Quintile 5 2879 222 3101
Total 14275 1233 15508
Table 3
Average Monthly Household Consumption Expenditure (Rs),
By Gender of Household Head and Region
Total Urban Rural
Male 13652.01 16908.95 11468.01
Female 12586.2 15976.07 10089.16
Table 4a
Average Expenditure Shares for Different Commodity Groups,
by Gender of Household Head (Overall)
Male- Female-
Commodity Groups headed headed t-test
Food and Drinks 48.34 44.25 11.17 *
Clothing and Footwear 5.46 5.78 -4.63 *
Fuel and Lighting 8.02 9.21 -10.02 *
Housing 12.87 14.85 -6.89 *
Transport and Communications 5.72 4.28 9.98 *
Household Effects 0.60 0.72 -3.05 *
Personal Effects 3.55 3.57 0.44
Healthcare 3.50 3.73 -1.89
Education 2.51 3.72 -9.19 *
Entertainment 0.54 0.55 -0.36
Durables 0.71 0.67 0.04
Miscellaneous 8.17 8.67 -2.84 *
* Significant at 5 percent level of significance.
Table 4b
Average Expenditure Shares for Different Commodity Groups,
by Region (Male-Female)
Urban
Male Female T-test
Food and Drinks 42.31 38.79 6.61 *
Clothing and Footwear 5.04 5.24 -1.91
Fuel and Lighting 7.47 8.01 -3.15 *
Housing 18.55 20.96 -4.95 *
Transport and
Communications 5.87 4.68 5.28 *
Household Effects .59 .69 -1.81
Personal Effects 3.68 3.68 -0.02
Healthcare 2.99 3.27 -1.53
Education 3.80 4.88 -4.39 *
Entertainment .93 .81 1.86
Durables .72 .88 -0.84
Miscellaneous 8.05 8.10 -0.25
Rural
Male Female T-test
Food and Drinks 52.39 48.26 9.40 *
Clothing and Footwear 5.74 6.18 -4.81 *
Fuel and Lighting 8.39 10.09 -10.63 *
Housing 9.06 10.35 -5.00 *
Transport and
Communications 5.62 3.98 8.70 *
Household Effects .60 .74 -2.47 *
Personal Effects 3.46 3.49 -0.44
Healthcare 3.84 4.08 -1.40
Education 1.65 2.86 -9.29 *
Entertainment .28 .36 -2.26 *
Durables .71 .52 1.27
Miscellaneous 8.25 9.08 -3.34 *
* Significant at 5 percent level of significance.
Table 5
Marginal Expenditure Shares for Different Commodity Groups,
By Gender and Region
Overall Urban
Commodity Groups Male Female Male Female
Food and Drinks .198 .145 .195 .145
Clothing and Footwear .035 .022 .029 .020
Fuel and Lighting .047 .045 .048 .049
Housing .226 .235 .285 .263
Transport and .092 .100 .093 .087
Communications
Household Effects .024 .010 .014 .009
Personal Effects .029 .038 .032 .044
Healthcare .026 .015 .024 .013
Education .081 .108 .092 .113
Entertainment .012 .014 .013 .014
Durables .218 .237 .145 .192
Miscellaneous .158 .135 .143 .141
Rural
Commodity Groups Male Female
Food and Drinks .212 .147
Clothing and Footwear .047 .026
Fuel and Lighting .043 .032
Housing .073 .065
Transport and .091 .164
Communications
Household Effects .043 .014
Personal Effects .021 .014
Healthcare .032 .026
Education .051 .083
Entertainment .010 .016
Durables .301 .434
Miscellaneous .201 .120
Table 6
Expenditure Elasticities for Different Commodity
Groups, by Gender and Region
Commodity Overall Urban
Groups
Male Female Male Female
Food and Drinks .636 .591 .638 .570
Clothing and
Footwear .766 .683 .737 .664
Fuel and Lighting .730 .626 .690 .637
Housing 1.342 * 1.357 * 1.247 * 1.262 *
Transport and
Communications 1.373 * 1.273 * 1.424 * 1.245 *
Household Effects 1.223 * 1.015 * 1.247 * .994
Personal Effects .836 .805 .859 .800
Healthcare .825 .725 .824 .699
Education 1.788 * 1.644 * 1.6202 * 1.383 *
Entertainment 1.046 * .991 .828 .798
Durables 1.470 * 1.247 * 1.424 * 1.289 *
Miscellaneous 1.425 * 1.454 * 1.437 * 1.570 *
Commodity Rural
Groups
Male Female
Food and Drinks .706 .686
Clothing and
Footwear .845 .757
Fuel and Lighting .804 .674
Housing .943 1.005 *
Transport and
Communications 1.392 * 1.332 *
Household Effects 1.311 * 1.178 *
Personal Effects .717 .727
Healthcare 1.009 * .946
Education 1.688 * 1.755 *
Entertainment 1.133 * 1.231 *
Durables 1.767 * 1.318 *
Miscellaneous 1.529 * 1.517 *
* Luxury Good.
Table 7
Average Monthly Household Consumption Expenditure (Rs),
by Gender of Household Head and Expenditure Quintile
Female
Expenditures
Male Female as % of Male
Quintile 1 5,502.22 5,072.42 92.2
Quintile 2 8,204.67 8,175.85 99.6
Quintile 3 10,675.78 10,649.65 99.8
Quintile 4 14,216.92 14,374.73 101.1
Quintile 5 29,244.9 29,807.02 101.9
Table 8
Average Expenditure Shares for Different Commodity Groups,
by Gender of Household Head and Expenditure Quintile
Male
Commodity Groups Q1 Q2 Q3 Q4 Q5
Food and Drinks 54.61 52.47 50.29 46.68 38.01
Clothing and Footwear 5.74 5.66 5.55 5.40 4.94
Fuel and Lighting 9.23 8.52 8.13 7.79 6.51
Housing 11.45 11.56 12.15 13.05 16.08
Transport and
Communications 4.38 5.13 5.30 6.20 7.53
Household Effects 0.37 0.45 0.53 0.64 0.99
Personal Effects 3.91 3.69 3.57 3.42 3.18
Healthcare 3.47 3.48 3.49 3.51 3.53
Education 0.66 1.38 2.01 3.13 5.28
Entertainment 0.21 0.38 0.54 0.67 0.88
Durables 0.19 0.23 0.45 0.83 1.85
Miscellaneous 5.77 7.05 7.98 8.69 11.21
Female
Commodity Groups Q1 Q2 Q3 Q4 Q5
Food and Drinks 50.05 47.82 46.28 42.11 30.96
Clothing and Footwear 5.90 6.22 6.12 5.85 4.73
Fuel and Lighting 10.89 10.16 8.94 8.19 6.62
Housing 13.40 12.90 13.18 14.96 20.76
Transport and
Communications 3.49 3.72 3.98 4.39 6.34
Household Effects 0.48 0.55 0.69 0.92 1.14
Personal Effects 3.88 3.75 3.42 3.48 3.10
Healthcare 3.84 3.52 3.90 3.70 3.66
Education 1.13 3.01 3.44 4.83 7.98
Entertainment 0.35 0.49 0.49 0.73 0.84
Durables 0.24 0.16 0.71 0.90 1.69
Miscellaneous 6.34 7.72 8.84 9.96 12.18
Table 9
Marginal Expenditure Shares for Different Commodity Groups,
by Gender of Household Head and Expenditure Quintile
Male
Commodity
Groups Q1 Q2 Q3 Q4 Q5
Food and Drinks 0.482 0.440 0.349 0.322 0.155
Clothing and
Footwear 0.046 0.049 0.047 0.038 0.031
Fuel and Lighting 0.075 0.062 0.079 0.049 0.043
Housing 0.150 0.149 0.159 0.181 0.227
Transport and 0.070 0.069 0.058 0.105 0.088
Communications
Household Effects 0.010 0.010 0.013 0.018 0.029
Personal Effects 0.032 0.033 0.031 0.037 0.028
Healthcare 0.032 0.039 0.053 0.026 0.019
Education 0.028 0.046 0.082 0.090 0.073
Entertainment 0.027 0.005 * 0.021 0.016 0.012
Durables 0.001 * 0.087 0.057 * 0.049 * 0.263
Miscellaneous 0.078 0.089 0.123 0.123 0.174
Female
Commodity
Groups Q1 Q2 Q3 Q4 Q5
Food and Drinks 0.399 0.373 0.537 0.336 0.113
Clothing and
Footwear 0.042 0.046 0.039 * 0.015 * 0.015
Fuel and Lighting 0.085 0.063 0.065 0.035 * 0.044
Housing 0.174 0.126 * 0.010 * 0.142 0.204
Transport and 0.073 0.066 0.059 0.066 0.114
Communications
Household Effects 0.010 * 0.020 * 0.010 * 0.021 * 0.009
Personal Effects 0.035 0030 0.052 0.030 0.044
Healthcare 0.042 0.074 0.024 * 0.030 * 0.004 *
Education 0.063 0.071 * -0.029 * 0.123 0.097
Entertainment -0.005 * -0.017 * -0.027 * 0.044 0.014
Durables -0.004 * 0.040 0.018 * 0.117 * 0.284
Miscellaneous 0.104 0.127 0.166 0.187 0.131
* Not significant at 5 percent level of significance.
