1
Prices and poverty in urban Ethiopia
Abstract
Poverty is an ongoing issue in Ethiopia. The identification of policy options to address the
problem primarily requires that poverty be measured accurately. One of the most important
ingredients in the measurement of poverty are prices. The magnitude of poverty is affected by
how cost of living differences across time and regions are adjusted. This paper derives a set of
price indices for Urban Ethiopia using data from four urban household surveys conducted in
1994, 1995, 1997, and 2000. The results show that the cities of Dire Dawa and Mekelle are the
two most expensive cities, while Jimma and Bahir Dar are the least expensive. The findings
also confirm that poverty is indeed high in urban Ethiopia with poverty head count of over 40
percent. Poverty estimates derived using country level consumer price indexes, which do not
adjust for spatial cost of living differences, are misleading. But using poverty lines as deflators
to account for price differences does not affect the poverty estimates obtained.
Keywords: Poverty; Urban Ethiopia; Price indexes
By: Tesfaye Gebremedhin
The University of Sydney
E-Mail: tgeb0150@mail.usyd.edu.au
Telephone: (61-2) 9950 9884
Stephen Whelan
The University of Sydney
E-Mail: S.Whelan@econ.usyd.edu.au
Telephone: (61-2) 9036 9251
Address for written correspondence:
Discipline of Economics
University of Sydney NSW 2006
Australia
2
Introduction:
Ethiopia is one of the world’s poorest countries by any standard. According to a recent World
Development Report, the country has the lowest GNP per head in the world, and its purchasing
power parity adjusted GNP is ranked 200
th
out of 206 countries (World Bank 2000). Human
development indicators of the United Nations Development Program (UNDP) also attest to the
seriousness and extent of poverty in the country. For instance, the Human Development Index
(HDI) of Ethiopia is the sixth lowest out of 175 countries in the world. Similarly, the Human
Poverty Index (HPI) ranks Ethiopia 91
st
out of 94 developing countries.
The poverty experienced by many Ethiopians is reflected in a range of well being measures of
the population. For example, the life expectancy at birth in the country is approximately 46
years, which is substantially lower than the average 77 and 67 years recorded for countries
with high and medium human development indices respectively. Moreover, three quarters of
the population do not have access to an adequate water source, a figure that is amongst the
highest for countries experiencing a low measure of human development. The percentage of
population with access to suitable sanitation, which stands at 12 per cent is significantly lower
that the 53 per cent average for the sub-Saharan Africa (UNDP 2003, pp. 237-257). On the
other hand, the adult illiteracy rate at around 60 percent is significantly higher than the average
for sub-Saharan Africa and other developing countries.
Previous analyses of poverty in Ethiopia have generally focused on rural rather than urban
areas (see Dercon and Krishnan (1998), Dercon and Krishnan (2000), and Dercon (2001)).
This is understandable in light of the fact that around 85 per cent of the population lives in
rural areas. Unfavourable weather fluctuations usually take a heavy toll on the lives of rural
farmers and bring them to the brink of starvation. It is the plight of urban Ethiopians, however,
that is the focus of the analysis in this paper. Although urban Ethiopians generally enjoy a
higher standard of living when compared to their rural counterparts, poverty remains a problem
in urban areas (Tadesse 1999).
To understand the extent of the problem and develop viable policy options for its alleviation
requires primarily that poverty be measured accurately. One of the most important ingredients
3
in the measurement of poverty is price. The effect of prices on poverty measurement has
received little attention in the wider poverty literature. Prices are used to calculate the rate of
inflation so that nominal measures of welfare can be compared across time. Inaccuracy in
measuring the rate of inflation will lead to erroneous poverty estimates. Overestimation of a
price increase, for instance, would lead to an overestimation of the level of poverty.
Consequently, any reduction in poverty would be underestimated (Deaton and Tarozzi, 2000).
Besides measuring inflation, price indexes are also required to compare cost of living
differences between different regions of a country. It is imperative that nominal measures of
welfare also be adjusted for spatial cost of living differences because the overall magnitude and
geographic dimension of poverty could be very sensitive to how and whether such adjustments
are made (Hentschel and Lanjouw, 1996). The need to have a robust and accurate regional
poverty profile is evident as it would influence decisions regarding the transfer of resources
designed to alleviate poverty.
In Ethiopia, the price index that had been widely used for over three decades for measurement
purposes was the Addis Ababa Retail Price Index (RPI) of the Central Statistical Authority
(CSA). The RPI was a Laspeyres price index with weights derived from a survey of 600
households taken from the Addis Ababa Household Consumption and Expenditure Survey of
1962/63. This index was, however, inadequate to capture the true picture of inflation in
Ethiopia due to several limitations. Firstly, the coverage of the survey was very small as the
survey had not been designed to cover all the urban enumeration areas and take sufficient
number of sample households (CSA, 1996). As a result, the expenditure shares used to weight
the basket of goods were distorted. Secondly, the ability of the index to reflect true inflation
was undermined because the classification of household goods and services in the final basket
was flawed. For instance, items like transport and communication had been lumped with items
on personal care and effects. The index, thus, used the relative price of dissimilar items making
the identification of price movement of certain groups of items less transparent (CSA, 1996).
Another caveat of the Addis Ababa RPI was that the basket of goods and expenditure shares in
1963 were used until 1995/96. The items included in the original basket become increasingly
4
unrepresentative as the availability of goods and services in the market changed over time. In
turn, the weights used to calculate prices changes had become long out of date. As Deaton and
Tarozzi (2000) point out, “… whether or not the price indexes are seriously affected is
ultimately an empirical question, though it is often supposed (for example in the comparable
debate over the CPI in the United States) that Laspeyres indexes will increasingly overstate
inflation as the base period recedes into the past, a tendency that will be exacerbated by the
failure to pick up new goods (whose prices are often falling rapidly) and discard old ones
(whose prices may be stagnant or ever rising)”.
The above limitations prompted the CSA to construct a new consumer price index (CPI) from
the 1995/96 Household Income, Consumption and Expenditure Survey. This survey is greater
in scope and covers the whole country, unlike its 1963 counterpart. One advantage of this CPI
is the possibility of calculating urban and rural consumer price movements. Although this has
proven to be a major improvement over the Addis Ababa RPI, the CSA still does not issue
inflation figures further disaggregated by urban and rural centres. Moreover, they don’t issue
figures showing spatial cost of living differences, which usually cannot be obtained from
published sources. As a result, previous studies of poverty in Ethiopia used poverty lines as
cost of living deflators to circumvent this problem (see for instance Tadesse (1999), Dercon
and Krishnan (1998)). No study, other than Kedir (2003)
1
, has attempted to make explicit
calculation of price indexes for poverty analysis.
A general misgiving that could be expressed against the CSA prices concerns quality of the
data and collection process. As Kedir (2003) notes, “…the enumerators of the CSA price
surveys are often recruited from a pool of high school drop outs with little knowledge of the
use of the price information they were collecting”. There is also some anecdotal evidence
suggesting that enumerators may have avoided visiting markets to collect price information
(Kedir, 2003). Together, these issues present a strong case for deriving prices from other
surveys for use in the calculation of price indexes.
1
Kedir (2003) derived price indexes for urban Ethiopia using the first round of the Ethiopian Urban Household
socioeconomic survey (EUHS).
5
The contribution of this paper is to derive a set of price indices for Urban Ethiopia using data
from four urban Household Surveys conducted in 1994, 1995, 1997 and 1999. We provide
estimates of the rates of inflation over the six year period for seven urban centres of the
country. Estimates of price levels between the seven urban centres are also provided for each
of the four periods. The price indices developed are used to calculate poverty rates between
1994 and 2000. The implications on poverty of using alternative sources of consumer price
indices are discussed. The benefit of such an approach is to provide additional insight into the
nature and patterns of poverty in urban Ethiopia during the study period. Moreover, it
highlights the role played by price measures on the evaluation of poverty.
This paper is organized as follows. In the next section, the data set used in the study is briefly
described. Subsequently, the different techniques and methodologies employed to do the
analysis will be laid down. Specifically, issues related to the construction of the measure of
welfare, the poverty line and the price indexes will be discussed. Next, summaries of the
results will be reported and discussed. The final section will set out some concluding remarks.
Data
This study uses the 1994, 1995, 1997 and 2000 household survey data obtained from the
Ethiopian Urban Socio-economic Surveys (EUHS) conducted by the Department of
Economics, Addis Ababa University. The1994 round was collected in collaboration with the
Department of Economics, Goteborg University.
The surveys were conducted over a period of four successive weeks during a month considered
to represent average conditions so that seasonal factors affecting the patterns of household
consumption and prices will be least operative. Seven cities and towns thought to reflect the
major socio-economic characteristics of the urban population were selected for the survey.
These were Addis Ababa (the capital city), Awassa, Bahir Dar, Dessie, Dire Dawa, Jimma, and
Mekele. A total sample size of 1,500 households were allotted in proportion to the size of the
population residing in the selected urban centres in 1992. Thus, 900 households were drawn
from Addis Ababa, 126 from Dire Dawa, 73 from Awassa, 101 from Dessie, and 100 from
each of the remaining three cities (Mekonnen, 1999). Proportional samples were then taken
