AFREPREN/FWD - Energy, Environment and Development Network for Africa Website

Home l Site Map l Contact Us l

Working Paper 295


About Us

Focal Areas




Energy News

Energy Events


Related Links

Picture Gallery

Contact Us

No. of Website Visitors:

Affordability of Modern Fuels and  Patterns of Energy Demand in  Urban Ethiopia


Bereket Kebede, Almaz Bekele and Elias Kedir


The report starts by examining costs of modern energy in the different urban areas. Cost estimates for kerosene, butane gas and electricity in different urban areas of Ethiopia indicate significant variations. After looking at variations in costs of energy, a comparison of costs with purchasing power is done. Kerosene seems to be affordable for all urban areas and for per capita expenditure deciles. While the upper half of the expenditure distribution seems to have sufficient purchasing power to use butane gas, the position of households in the lower half is doubtful. Households near but below the poverty line may have the capacity to use it. Electricity is too expensive even for the relatively well to do.

Households expend around a tenth of their income on energy. The expenditure on firewood is the most important (38% of the households' energy budget) followed by electricity (20%) and kerosene (19%). Charcoal accounts for 17% and dung cakes for 5% of the household energy budget. Butane gas and sawdust are the least important. The combined percentage share of firewood, charcoal and dung cakes is 60% implying heavy depends on biomass fuels. The mean budget share of poor households is 10% while that for non-poor is around 7% (but the non-poor expend more in absolute terms). Energy budget share consistently decreases with per capita expenditures if we classify households into deciles (from 11% to 5% from the lowest to the highest deciles). The dependence of households on biomass fuels significantly decreases with income. The lowest decile expends 70% of its energy budget on biomass fuels but the corresponding figure for the richest decile is only 42%.

The dependence on biomass fuels in the urban centres varies widely. The shares of biomass fuels range from 26% (Addis Ababa) to 85% (Gondar). Addis Ababa is the only urban centre without heavy dependence on biomass fuels and is on top of the list for all three modern forms of energy.

Even though the budget share of total energy expenditure falls with income when considering the whole sample, this trend is consistently observed only for five of the urban centres out of the eleven. There are many instances where total energy budget shares increase in the middle deciles. Similarly, energy demand patterns in different urban areas significantly vary. In some towns, firewood is the most important fuel for all income groups (Jimma and Harer). In others, charcoal plays a similar role except for the lowest deciles that depend on firewood (Gondar and Nazret). In some cases, while firewood is the most important, charcoal takes over for the richest (Mekele and Bahr Dar). But in other instances higher income groups seem to shift from firewood to kerosene (Harer and Dire Dawa), while in the case of Addis Ababa the most important fuel changes from kerosene to electricity when increasing income.

The observed consistent increase in the importance of modern fuels as income of households rise on the aggregate level does not appear to hold for all individual urban areas. Econometric tests indicate that this pattern does not exist at least for some urban areas. The explanatory power of the 'energy ladder' hypothesis is weakened on lower levels of aggregation.

An increase in the income of households is expected to increase the demand for all types of fuels. Positive budget elasticity for traditional fuels is symptomatic of a lower level on the 'energy transition'. Income elasticities indicate that while firewood and kerosene are basic fuels, households tend to use more charcoal and electricity when their income increases. This is not a neat transition as predicted by the 'energy ladder' hypothesis.

Own-price elasticities of firewood and charcoal are high implying that the demand for them will significantly decline with a rise in their prices. But the demand for electricity is price inelastic. In light of the recent increase in electricity tariff towards its long run marginal cost, the second result is comforting; the rise in electricity price is not expected to drastically decrease the demand for electricity.

The computed cross-price elasticities indicate that while dung cakes/electricity and firewood/electricity are substitutes for each other, kerosene/dung cakes and kerosene/electricity are complements. The substitution between electricity and dung cakes on the one hand and electricity and firewood on the other is mainly explained by fuel used for injera preparation. Due to the recent expansion of baking injera with an electric utensil, electricity has become a substitute for dung cakes and firewood. Kerosene is a complement to electricity and dung cakes because they are used for different purposes. Apart from lighting, electricity is mainly used for injera baking (and so are dung cakes); but kerosene is used for cooking and preparing tea/coffee and not for injera preparation.

An increase in the price of traditional fuels is usually considered as one important reason pushing households towards the use of modern fuels. Our cross-price elasticities indicate that is true in the case of electricity; increases in the prices of firewood and dung cakes increase the demand for electricity while that of charcoal does not affect it. But the effect of an increase in the price of traditional fuels on the demand for kerosene is ambiguous; an increase in the prices of firewood and dung cakes decreases the demand for kerosene. On the other hand, an increase in the price of charcoal increases kerosene expenditure. This result tallies with the end use pattern of households since kerosene is mainly a substitute to charcoal.

With a unit increase in income, the highest rises in energy expenditures are on firewood and electricity for the poor and non-poor respectively. These results give us an idea of how different types of economic growth affect energy demand. If economic growth is pro-poor (that is increasing the income of the poor relatively more than that of the non-poor) we expect a significant increase in the demand for firewood. On the other hand, if it is not pro-poor we expect a significant increase in the demand for electricity. In terms of satisfying the energy needs of the urban poor in the short run, policy issues related to biomass fuels are more important.

More options exist in urban areas even for the poor as compared to their rural counterparts. For instance, while the budget share of electricity is second for the non-poor after firewood, charcoal and kerosene occupy that position for poor households. As a more efficient source of heat compared to firewood, policy may have to take charcoal as an important energy source for the urban poor in the short run. In addition, that poor households expend the same amount of their energy budget on kerosene as on charcoal underscores the possibility that kerosene is a realistic substitute for biomass fuels even for the poor in urban areas.

The poverty reduction strategy of the government is not given sufficient emphasis in the energy policy document. Finer distinctions in the energy demand of poor and non-poor households in urban (and rural) areas and their policy implications were not clarified in the document. To address these issues the policy document must be based on background information provided from empirical data and energy studies.

When examining government budget, a pattern that is immediately observed is that even though absolute amounts were fluctuating over the years, the percentage shares of current expenditure on mining and energy from total government expenditure consistently declined over the period 1992/93-1998/99. Both the absolute and relative amounts expended on electric power consistently increased between 1992/93 and 1996/97. With the exception of 1997/98 government expenditures seem to consistently reflect the emphasis given to the development of hydropower in the energy policy document.

The fluctuation in government budget items related to urban development without any apparent pattern and the consistent increase (or protection) of agricultural, educational and health expenditures probably are indicators that the former are not considered as 'core' budget items. If the budget constraint starts to bite harder, most likely expenditures on urban development will be slashed before affecting the 'core' budget items on agriculture, education and health. Probably few will argue against this priority. But it is also an indication that even within the limited resources the government can mobilise, very limited amount can be devoted to urban development in general and the urban poor in particular. The limited expenditures by the government underscore the need to mobilise resources from non-governmental sources; the active involvement of non-governmental organisations in pursuing projects for poor urban households must be encouraged in this light.

Public interest on the issue of energy to the urban poor is very limited. This is illustrated with the small (almost zero) number of articles that appear in newspapers on energy and the urban poor.

Return to top

This paper is available on an exchange basis. If you find it to be useful, we encourage you to send us any relevant publications from your organization. To request for the full paper, please fill in the publications request form

View the listing of Other Working Papers



wordpress visitor counter