Robust Poverty and Inequality Measurement in
Egypt: Correcting for Spatial-price Variation and
Sample Design Effects
Dean Jolliffe, Gaurav Datt, and Manohar Sharma*
Abstract
The paper estimates inequality and absolute poverty in Egypt for 1997 with measures that are robust to
sample design effects and corrected for spatial variation in price levels. Standard errors for inequality indices
are calculated using a bootstrap approach which replicates the sample design. Standard errors for poverty
indices are corrected for the design effects resulting from sample stratification and clustering. The authors
use data from the Egypt Integrated Household Survey and follow the cost-of-basic-needs methodology to
construct region-specific poverty lines. It is found that 15.7 million people were poor in Egypt in 1997, or
26.5% of the population. The estimates indicate a sharp sectoral difference with rural areas being signifi-
cantly poorer, but significant differences in poverty were not found between Upper and Lower Egypt. This
finding differs substantially from the conventional wisdom that Upper Egypt is poorer than Lower Egypt
and results from the correction for spatial-price variation.
1. Introduction
In 1990, the World Bank asserted that poverty reduction was the fundamental objec-
tive of economic development. This emphasis on poverty reduction continues, and the
World Bank (2000a) recently stated that fighting poverty is at the center of all work
that they do. Before policies can be designed and implemented for poverty reduction,
it is first necessary to identify who is poor. The aim of this paper is to identify the poor
and measure the level of inequality in Egypt in 1997. The source of data for this analy-
sis is the 1997 Egypt Integrated Household Survey (EIHS) and the key measure of
individual welfare is total consumption per person.
Poverty lines are constructed following a cost-of-basic-needs methodology for five
regions of Egypt, and the regional variation in these lines is used as a spatial-price
index. There are significant differences in the cost of living across regions of Egypt and
we illustrate the importance of correcting for spatial-price variation when measuring
poverty.An important characteristic of the EIHS, as with most nationally representa-
tive household surveys, is that the sample was not collected in a simple random draw
(or, in other words, a single-stage random draw as considered in most introductory
statistics textbooks), but rather in a multistage draw from the population. The impli-
cation of this is that conventional standard errors for poverty indices can significantly
Review of Development Economics, 8(4), 557–572, 2004
*Jolliffe: Economic Research Service, US Department of Agriculture, 1800 M Street NW, Washington,
DC 20036, USA.Tel: 202-694-5430; Fax: 202-694-5642; E-mail: jolliffe@ers.usda.gov. Datt:The World Bank,
Level 18, 14 Martin Pl., Sydney NSW 2000, Australia. Tel: (61-2) 9235-6533; Fax: (61-2) 9223-2533; E-mail:
gdatt@worldbank.org. Sharma: International Food Policy Research Institute, 2033 K Street NW,Washing-
ton, DC 20006, USA. Tel: 202-862-8142; Fax: 202-467-4439; E-mail: m.sharma@cgiar.org. We wish to thank
Akhter Ahmed, Dowoud Al-Husseini, Heba El-Laithy, Fatma El-Zanaty, Lehman Fletcher, Lawrence
Haddad, Akila Hamza, Amina Hegazy, Nelly Shaheen, Ayman Tawfik, and Maged El Wakkad. The authors
are solely responsible for any errors. Work for this paper was partially funded by USAID grant 263-G-
00-96-00030-00. The views and opinions expressed in this paper do not necessarily reflect the views of the
Economic Research Service of the US Department of Agriculture.
© Blackwell Publishing Ltd 2004, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA