Exploring the Potentials of Water Mills in the Grain-Milling Industry in Ethiopia

Dejene Aredo

Source: Technology Policy in Africa edited by Osita M. Ogbu, Banji O. Oyeinka, and Hasa M. Mlawa . IDRC 1995, 396pp. ISBN 0-88936-790


Technology transfer in Ethiopia has meant the importation of largely labour-savinginnovations to replace obsolete equipment. Water mills, which used to be the mostwidespread rural technology, lost their importance when diesel mills entered the rural areasafter the Italian occupation of Ethiopia (1936–41). In large urban centres, water millswere wholly replaced by electric mills. Today, water mills are restricted to virtuallyinaccessible rural areas. Neglected by policy makers and rural-development practitioners,the technology of water milling survived for decades without access to spare parts orcomponents from the modern sector. Government policy in the 1950s and 1960sencouraged the spread of diesel engine mills by providing cheap, imported fuel. Today,however, water mills are showing some signs of revival in rural areas, following the risein the prices of fuel and spare parts for diesel mills. It is not yet known how this type oftechnology has survived the period of massive importation; its immense potential hasremained hidden from researchers and rural-development practitioners.

The demand for improved water mills will likely increase when diesel millsbecome too expensive for the rural poor, whose real per capita income has been decliningin recent years. The water mill's social value is expected to rise when imported fuelsbecome more expensive.

The purpose of this study is to explore the hidden potentials of a microenterpriseby identifying and analyzing the complementary roles of water mills and modern mills,with a view to instigating further research in the rural, hydro-based grain-processingindustry. The specific objectives of the study were (1) to characterize the grain-millingindustry in Ethiopia; (2) to identify and explain the relative advantages of water mills; (3)to identify and explain the major constraints on the expansion of water mills; and (4) topropose that further research be done on alternative designs for water mills for theconsideration of promoters of rural technology in Ethiopia.

The data in this study were obtained mainly in northern Shewa. The region wasappropriate for this study because the earliest mills are still found there, as well as manypartly abandoned water mills. I captured a general picture of water mills by undertakinga survey of 12 woredas (subdistricts) in northern Shewa. A detailed study of the millsfocused on two villages. In one of the two villages, Gedilge, a household survey (n = 21)characterized the users and nonusers of the mills. The interviewees were all women. Othersources of information included mill owners, local officials from the Ministry ofAgriculture, leaders of peasant associations, and village elders. A mill in the other village,Chaka, was selected to illustrate the mechanics of milling.

Theoretical perspectives

Bagachwa (1991) succinctly reviewed the approaches to technology choice. These werethe neoclassical approach, the fixed-factor-proportions approach, and the appropriate-technology approach. The neoclassical approach is based on the model of pure and perfectcompetition. For developing countries, this implies (1) the adoption of the labour-intensivetechniques of production, arising from the assumption that labour is the most importantresource in these countries; and (2) the correction of distortions in market prices.

The fixed-factor-proportions approach, which is based on the Leontief–Harrod–Dommar assumption of constant-input coefficients, questions the plausibility ofthe neoclassical assumption of a near-infinite range of available technologies. Thisapproach rules out the possibility of factor substitution. It draws much of its support fromthe observation that almost all technological innovations take place in developed countries,where the direction of technological change is toward labour-saving innovations. It isassumed that newer capital-intensive technologies supersede older ones as they becomeobsolete and unproductive (Romijn and De Wilde 1991, p. 103):

Once modern western technologies are brought into traditional society, they manage to superimpose themselves and compete successfully with local production processes to such an extent that the latter find it difficult to survive.
As a result, developing countries may not have efficient technology alternatives, other thanthose with the high capital–labour ratios found in developed countries. The efficient-factorcombination is considered fixed in the proportions found in developed countries (Eckaus1955; White 1978; Bagachwa 1991). This type of technology choice creates a state ofdependence in which factor proportions in developing countries are determined by patternsof resource endowments in developed countries.

The appropriate-technology approach combines aspects of both the neoclassicaland fixed-factor-proportions approaches. The centrepiece of this approach is theassumption that (1) there is a lack of technology tailored to or adapted to the conditionsof developing countries and (2) there is a need to develop technologies consistent with thepatterns of resource endowment in these countries. An appropriate technology can, ingeneral, be characterized as follows:

  1. Appropriate technology should be technically efficient, not wasteful. It should beeconomically efficient, making the best use of available resources. It should beinexpensive and small scale so that poor people can afford it, leading to a moreequitable distribution of incomes and assets.
  2. Appropriate technology should be socially and culturally compatible, enhance thequality of life, be satisfying (creativity of work), involve machines that aresubordinate to people, use communal rather than individual goods and services,foster social participation, and facilitate deconcentration of power.
  3. Appropriate technology should be environmentally sound. It should preferably userenewable rather than nonrenewable energy and raw materials. It should producedurable goods that can be recycled or reused, cause minimal pollution and wastes,and blend into local ecosystems. It should be compatible with the rational,sustained use of the environment.
Hydropower provides a developing economy with opportunities to developappropriate technologies. It has been noted that "one of the first things a country can dois to assess its opportunities for developing alternative energy sources. In hydropower, thehydrologic studies are basic to the entire process" (NRECA 1980, p. 23).

Hydropower in developing countries (NRECA 1980) has nine distinguishingcharacteristics: sustainability, dependence on local resources, cost effectiveness, durability,flexibility, simplicity, ability to fit into existing systems, accessibility to isolated ruralcommunities, and ability to meet multiple purposes.

The sustainability of hydropower arises from the fact that it uses a renewablesource of energy — water. It is essentially nonpolluting. It is environmentally sound andacceptable. Hydropower makes maximum use of local resources and, thus, compared withthermal-power, is usually much more appropriate for conditions in many developingcountries, which face shortages of the foreign exchange required to import fuel oil.Hydropower is largely cost effective and is, to some extent, insulated from inflation. Nofuel is required and heat is not involved, so operating costs are low. Approximately 650kW·h production by a hydropower plant will reduce the requirement for oil (or its fuelequivalent) by 1 bbl (1 bbl = about 0.16 m³). Because of this and the durability of thefacilities, a hydropower installation is to some extent inflation proof. Because no heat isinvolved, the equipment has a long life, and malfunctioning is uncommon. Dams andcontrol works can perform for decades, and limited maintenance is required. Hydropower'sreliability and flexibility of operation, including fast start-up and shutdown times inresponse to rapid changes in demand, makes it an especially valuable part of a large powersystem of a developing country. The relative simplicity of a small-scale, hydro-basedenterprise makes energy instantly available. Small-scale hydropower fits nicely into theenergy balance of a country. It can contribute to interregional equity by meeting the needsof isolated rural communities. It can be made available in small installations and withrelative ease in remote areas of developing countries. A small-scale hydropower facilitycan generate enough power for grain milling, sorghum dehusking, and village-levelelectrification. Hydropower, of course, presupposes the availability of water. Therefore,it is difficult to reach every part of a country with small-scale facilities.

The total energy of Ethiopia is largely obtained from traditional biomass fuels. Itis estimated that biomass fuels account for 95% of the total energy consumption, with only5% coming from modern energy sources. Deforestation is so pervasive that today less than4% of the total land area of the country is covered by natural forests, compared with 40%just a century ago.

Ethiopia's potential for hydroelectric power is considerable. The gross hydroenergypotential is estimated at 650 TW/year, which is roughly 8% of Africa's potential. However,the installed capacity of the five major hydroelectric plants is only about 360 MW/year.Ethiopia's per capita electricity consumption, at about 25 kW/year, is among the lowest inthe world.

Large-scale use of imported fuel has been precluded by the ever growing shortagesof foreign exchange. Today, fuel accounts for about one fifth of the value of total importmerchandise. Therefore, it is high time to explore the economic potential of small-scalehydropower facilities in rural industrialization. A study conducted by Tebicke and Gebre-Mariam (1990) clearly indicates that where the resource is available near the locality,small-scale hydroelectricity offers considerable advantages over both the grid-extensionand diesel-electric sources. Small-scale hydroelectricity sources offer considerable scopefor indigenous technical capacity, contributing to lower investment and supply costs,especially in foreign exchange. The same cannot be expected from the alternatives becauseof the high level of technical sophistication of the diesel and grid equipment components(Tebicke and Gebre-Mariam 1990).

The grain-milling industry in Ethiopia

In Ethiopia, on-farm consumption accounts for as much as 80% of the total output ofgrain. Quite a substantial proportion of rural households still hand-grind grains, using astone grinder, or pound the grain into flour, using a pound and pestle. However, innorthern Shewa, grain mills are widely used. An important characteristic of the food-processing industry in Ethiopia is the scarcity of commercial milling. Custom milling,which is done by private or cooperative mills in exchange for payment of milling fees, isstill the dominant form of food processing in the country. An Ethiopian woman rarely buysflour from shops or mills.

Four alternative types of technology are available in the food-processing industry:hand grinding (or pounding), water mills, diesel-engine-powered mills, and electric-motor-powered mills. Flour for the bakeries is produced largely by state-owned mills. The stategets the grain from imports or from the agricultural sector. In the past, state-owned millsobtained grain through a parastatal, the Agricultural Marketing Corporation. Many ruralhouseholds are net purchasers of food. Urban dwellers occasionally buy bread (made fromwheat) from bakeries. Otherwise, they buy grain from the market and pay to have it groundinto flour. In recent years, the private sector and the market system have played anincreasing role in the distribution and processing of food grains.

In Ethiopia, industry plays a limited role in the economy. In 1993, the share ofmanufacturing output in the gross domestic product was only 11%; that of small-scaleindustry was 4% (Mulat 1994).

Commercial milling is little practiced. Most of the flour required by householdsis processed by women using the traditional stone grinder, which is backbreaking and timeconsuming, or by small-scale custom mills. A foreign traveller, observing the grinding ofgrain in traditional Ethiopia, described it like this:

Women spent much time, and effort... in grinding. This was often carried out on hand-mills which consisted of a large fat stone of cellular lava, two feet long and one foot broad, raised upon a rude pedestal of stones and mud, about one foot and half from the ground. The rough surface of this stone sloped gradually forwards into a basin-like cavity, into which the flour fell as it was ground. A second stone, which weighed about three pounds, would be grasped in the hand of a grinding-woman who would move it up and down the inclined stone, thereby crushing the grain and gradually converting it into coarse flour.
Commercial milling is limited to 17 state-owned, large-scale mills (CSA 1992),which produce flour for the urban bakeries. These mills produce mainly wheat flour. Onesurvey reported that 88% of the grain used by state mills was wheat, and the rest wasmaize (CSA 1992).

Agricultural processing in Ethiopia, which has forward-production linkages, isdone in small-scale establishments for two reasons: (1) crops are bulky and heavy and areoften perishable, and transport costs can be greatly reduced if agricultural processing isdone close to the source of supply; and (2) the highly dispersed pattern of settlementrequires dispersed milling establishments. Grain milling is the most widespread power-driven small-scale industry in Ethiopia, in both urban and rural areas. A survey of 11 townsin the country reported that grain mills accounted for 55% of all small-scale industrialenterprises (wood works accounted for 9%) (HSSIDA 1979). In a similar survey,conducted later, this was found to be 64% (HSSIDA 1980). In predominantly rural areasor remote places, grain mills may account for 100% of power-driven enterprises. On theother hand, this proportion falls with the size of urban centres. For example, one surveyreported that in Addis Ababa, the largest city in Ethiopia, the proportion of grain mills inthe total number of establishments was only 34%, compared with 55% for all the towns(HSSIDA 1979).

The number of people employed at grain mills is considerable, though theworker–mill ratio is quite small. A survey of 963 small-scale industrial establishments inEthiopia reported that grain mills provided jobs for 1823 people; all the establishments,including mills, employed 9695 people. In other words, employment in grain millsaccounted for 19% of the total employment in industry (HSSIDA 1979). In another survey,grain mills accounted for 51% of the total employment in privately owned small-scaleindustries (HSSIDA 1985). But this proportion tends to fall with growth in urbanization.For example, in Addis Ababa, where there are many other industries, grain mills accountedfor only 9% of the total employment in private industries (Ministry of Industry 1992).

A recent comprehensive survey of small-scale industries in Addis Ababa providedthe following information about private grain mills: the average worker–mill ratio was 2.9(1–12 paid workers); the average capital per mill was 19826 birr; and the capital perworker was 6771 birr (in 1995, 6.3 Ethiopian birr = 1 United States dollar [USD]).Cooperative mills, however, were found to be quite large: the worker–mill ratio was 14;the average capital per mill was 154518 birr; and the capital per worker was 11109 birr.But cooperative mills, most of which are likely owned by urban dwellers and theirassociations, accounted for only 2.4% of the total number of mills in Addis Ababa (Region14 Administration 1994).

The number of workers per mill is quite small compared with that in other small-scale industries, as evident from many surveys. In one survey, the average number ofworkers per mill was 3.4, compared with 10 per establishment for all types of industries(HSSIDA 1979). Another survey suggested that employment in the milling industryaveraged 3.3 persons, compared with 6.3 persons per establishment for all types ofenterprises (Ministry of Industry 1992). On the other hand, this ratio has been found to behigh for commercial mills, which are largely state owned. A survey of enterprisesemploying more than 10 workers reported that there were 222 people per establishment(CSA 1992).

Wages in the grain-milling industry are small. One survey reported that 86% of theworkers employed in the milling industry earned less than 100 birr per month, whereas inthe food industry, as a whole, 72% of the workers earned less than 100 birr per month(HSSIDA 1985).

Women's rate of participation in the milling industry is lower than that of men. Onesurvey reported that women accounted for 20% of the total employment in the industry(Ministry of Industry 1992). Also, it appears that women earn less than men. A survey oflarge-scale mills indicated that women's wages were 82% of men's (CSA 1992).

The contribution of grain mills to the gross value of output of the small-scaleindustries is quite small, compared with their relative size within the small-scale industrialsector. According to one survey, grain mills accounted for only 6% of the value of the totaloutput of small-scale industries but for 55% of the total number of establishments in theindustry (HSSIDA 1979). In another survey, the value of the services provided annuallyby grain mills amounted to an average of 19665 birr per mill (Ministry of Industry 1992).The gross value added in the milling industry is also low, compared with that of othersmall-scale industries. For example, one survey reported that the gross value added in thisindustry was only 20% of that of coffee- and grain-clearing enterprises (HSSIDA 1979).

Operating surplus is the difference between value added in national accountconcept at factor cost and total wages, salaries, and benefits (Ministry of Agriculture1992). The operating surplus of the milling industry was estimated at 48% of that of thefood and beverage industry (Ministry of Industry 1992). In other words, profit perestablishment is very likely to be lower in the milling industry than in other types of small-scale industry.

Small grain mills are privately owned. Public ownership is restricted to large-scalecommercial mills. This is an area where the private sector played a very important roleduring the socialization drive of the military regime. In a survey of 11 towns in the country,it was estimated that 86% of the milling establishments were owned by individuals; 10%,by partners; and 4%, by cooperatives (HSSIDA 1979). In another survey, it was estimatedthat 82% of them were owned by individuals; 8%, by partners; 5%, by cooperatives; and5%, by training institutions, etc. (HSSIDA 1980). Among cooperatives, peasant servicecooperatives play a very important role. Funds for the establishment of grain mills comemainly from the informal sector. Owners of mills make little use of the banking systembecause banks are not available in rural areas, where 85% of the population lives. Inaddition, the banks require borrowers to present their books of account to get credit forexpansion or new investment; however, 79% of the small-scale industries in 1978/79 didnot keep books of accounts. Most of the funds for the milling industry come from theinformal financial sector. One survey reported that 97% of the total investment funds comefrom the owners of the mills (HSSIDA 1980).

Grain mills seem to need small investments. In one survey, grain mills, representing 64% of small-scale establishments, accounted for only 21% of the value of fixed assets(HSSIDA 1980). Working capital requirements are also small. According to one survey,the ratio of working capital to fixed assets in the privately owned industries was 0.12 forthe milling industry and 0.35 for the food and beverage industry (Ministry of Industry1992).

The cost of running a mill is much lower than the cost of running other small-scaleindustrial enterprises. According to one survey, "industrial" and "nonindustrial" costs ofrunning an average mill were 13 069 birr and 28 975 birr for the whole of the food industry(HSSIDA 1985). Industrial costs, in particular, were found to be very small. Industrialcosts included cost of energy, water consumption, repair and maintenance, rent, wages andsalaries, benefits, and raw materials consumed. Nonindustrial costs included postage,telecommunications, and advertisements. The same survey reported that industrial costsper establishment were only one third of that for the food industry as a whole. In contrast,nonindustrial costs were higher for grain mills than for the food industry, amounting to anaverage of 5014 birr for the grain industry and 4723 birr for the food industry (HSSIDA1985). The high nonindustrial costs of running grain mills could be largely attributed togovernment policy, which makes the mills pay high taxes. The various types of taxes themills paid in 1984/85 amounted to 84% of their total nonindustrial costs (HSSIDA). (It is,however, possible that mill owners, like other taxpayers, deliberately overstate the amountof tax they pay when they are interviewed.) The major cost component in the grain millindustry is fuel. According to one survey, about 49% of the total industrial costs of millingestablishments is for electricity and diesel fuels (HSSIDA 1985). In urban areas, electricityis used as a major source of power for grain mills. A survey of private industries in AddisAbaba indicated that expenditures on electricity accounted for 71% of the total industrialcosts of milling, with diesel fuels accounting for 5% (Ministry of Industry 1992). In largeurban centres, diesel fuel is little used in grain milling. Electricity consumption alsoincreases with the size of the enterprise. One report indicated that 55% of the totalexpenditure of the large mills was for electricity, 23% was for wood and charcoal, and22% was for other fuels (CSA 1992).

On the other hand, diesel fuel is an important source of power for mills operatingin rural areas where electricity is not available. However, the cost of fuel has been steadilyrising since the 1970s. Large mills try to overcome this problem by switching to electricpower. Nevertheless, the proportion of the total industrial cost of large mills given toenergy steadily increased from 5.6% in 1977 to 8.7% in 1981 (CSA 1992).

The milling industry encounters a lot of problems (Mulat 1994), with the resultthat enterprises operate much below capacity. One survey indicated that grain mills operateat about 40% below capacity (HSSIDA 1980). According to a detailed study of mills inthree areas in Ethiopia, actual capacity as a proportion of theoretical capacity was 46%(Lirenso and Aredo 1988). The major problems encountered by the industry can beclassified as supply-side problems or demand-side problems. The socialist-orientedmilitary regime, which ruled Ethiopia from 1974 to 1991, discouraged the expansion ofsmall-scale industries. Private mills encountered shortages of spare parts and components.The demand for milling was constrained by shortages of grain and by limitations inhousehold incomes.

A closer picture of the milling industry can be captured by considering thedistribution of different types of mills in northern Shewa. In a survey of 122 "peasants'associations," it was found that the average size was about 185 households. A peasants'association was usually established in an area of 800 ha. A group of three to sevenpeasants' associations formed a service cooperative, often with its own grain mill.However, many of these mills were destroyed at the downfall of the military regime in1991. There were no peasants' associations without at least one grain mill. The mostcommon type of mill was the diesel-engine mill (1.4 diesel mills per peasants' association),which accounted for 66% of the mills covered by the survey. But most of these mills wereinstalled in small towns and market places, areas accessible by vehicles. Next to dieselmills, water mills were the dominant type of technology, accounting for 29% of the mills.The corresponding proportion in southwestern Ethiopia was 25% (ONCCP 1980).However, the distribution of water mills among woredas was uneven, depending on theavailability of water and accessibility. Most of the water mills were found in two relativelyinaccessible woredas, Hagere-Mariam and Mafound. Of the 23 water mills found inMafound woreda, 15 belonged to a single peasants' association, Gedilgie. The averagedistance from a water mill to the main town was estimated to be a 3 h walk. Electric mills,which accounted for only 5% of the establishments, were limited to areas located nearhighways. Further details of the milling technology in Ethiopia are given in Aredo (1987),Lirenso and Aredo (1988, 1989), and Aredo and Abebe (1991).

The advantages of water mills

The origins of water mills in Ethiopia can be traced to the mid-19th century, when KingSahle-Selassie of Shewa installed a mill along the Airara River, with the assistance offoreigners. However, its use was prohibited by the clergy of the local religion, whoconsidered the innovation the work of a demon. Water mills had their heyday in the firsthalf of this century, when water mills were the most widespread power-driven industry inEthiopia. They were also one of the important sources of tax revenue. One testimony tothe past importance of water mills is the exceedingly large numbers of abandoned mills inmany locations in central Ethiopia. For example, at the village of Gedilge, some 15 kmfrom the town of Debre Sina, six partly abandoned mills were found at a single site alonga stream. Today, only one of them functions for commercial purposes.

The importance of water mills declined with the introduction of diesel mills afterWorld War II. Their importance further declined as hydroelectric power stations madeelectric mills possible in urban areas. However, water mills are far from a dying industry.Recent years have seen their revival in some inaccessible areas. This is, perhaps, becauseof the sharp increases in the price of diesel fuel and spare parts for diesel engines and alsoan increase in electricity tariff rates.

Table 1 compares the production capacity, costs, income, number of workers,import dependence, profitability, capacity utilization, and working time of the three typesof flour mills (i.e., water mills, diesel mills, and electric mills). Water mills have the lowestcapacity; they produce about 9 quintals of flour in a day; diesel and electric mills produce25 and 45 quintals, respectively. This is based on the assumption that the mills operate atfull capacity. Water mills operate relatively slowly. However, the waiting time at a watermill is usually nil because customers tend to leave the grain with the mill owners andcollect the flour at a convenient time. Strong personal relations exist between customersand mill owners. In the case of modern mills (diesel and electric), users often come fromdistant places or from urban centres, where the density of population limits personalrelations with owners. The travel time saved by users of water mills is considerable. In thestudy area, the average number of daily visitors to water mills was 9, whereas that to dieselmills and electric mills was 60 and 210, respectively.

Table 1. Comparative performances of three types of mills.
VariableWater millsDiesel millsElectric mills
Throughput (quintals/day)a
Book value of equipment (birr)b
Average number of clients (persons/day)
Service charge (birr/quintal)
Daily income (birr/working day)
Number of mill operators
Working hours (h/day)
Waiting time at mill site (min)
Running cost (birr/year)c
The degree of capital use (%)d
Rate of return (%)e
Ratio of net income to gross income
Import component (%)f
20 000
12 400
35 000
10 371
a Throughput is the quantity of grain that would be processed into flour daily if the mill wereoperating at full capacity. The working day is assumed to be 8 h.
b Book values are estimated for different years. The water mills were purchased more than 60 yearsago, whereas the diesel- and electric-engine mills were installed very recently.
c Running costs are recurrent costs, such as wages, taxes, and costs of fuel, electricity, andlubricants.
d The degree of capital use was estimated by dividing the average amount of flour actuallyprocessed by the potential output (throughput) for each type of mill.
e The rate of return was estimated by dividing net income by the value of fixed capital. In the caseof water mills, current value of a mill was taken. The value of the shelter was excluded from theestimate of fixed capital.
f The import component is the ratio of the value of imported materials to the total recurrentexpenditure incurred in a year.

Water mills cater to the needs of the very poor rural households, as evident fromthe very low service charges these mills demand. In the study area, owners of water millscharged about 2 birr per quintal for processing grain into flour, whereas owners of dieselmills and electric mills charged about 4 and 5 birr per quintal, respectively. The slow speedof water mills could be offset by the low service charge. Moreover, water mills can serviceinaccessible regions.

Investment outlays on water mills can be within the reach of the better-off peasantfarmers. Table 1 shows that the book value of water mills is very low. Moreover, watermills present an investment opportunity in rural areas, in contrast to modern mills, whichtend to be in urban centres. The cost of installing a new water mill is much lower than thecost of installing a modern mill. In one of the study villages, 300 birr was required to installa water mill. On the other hand, to install a diesel mill in the same area would requireabout 1100 birr. Another advantage of water mills is their very long life span. Most of thecurrently operational water mills are 60 or more years old. The recurrent cost ofmaintaining them is very low. The annual recurrent expenditure for water mills averages120 birr, whereas that for diesel mills and electric mills averages 12 400 and 10 371 birr,respectively. Diesel mills, in particular, are very costly to maintain.

A big advantage of water mills is their greater ability to rely on local resources,making use of almost no direct imports. On the other hand, diesel mills heavily depend onimported fuel and other imported inputs. According to the case studies, the ratio of thevalue of imported inputs to the total annual recurrent expenditure was nil for water mills.But in the case of diesel mills, imported inputs accounted for 79% of the total recurrentexpenditure. This is because of the high cost of imported fuel. Water mills also have highsocial value because of their use of local materials.

Simple ratios suggest that water mills are profitable. For example, the rate ofreturn to fixed capital for water mills was estimated at 23%, whereas it was 16% for dieselmills and 37% for electric mills. However, the net income from operating a water mill istoo small to attract urban-based investors.

Capacity underutilization is common for all types of mills. Diesel mills, inparticular, operate much below capacity, mainly as a result of frequent breakdowns andshortages of fuel and spare parts. This study found that diesel mills operate, on average,at 28% of full capacity. Water mills, although they face relatively low demand, performbetter (about 60% of full capacity) because they depend on local materials for their spareparts and because their parts and components are much more durable. The shaft, forexample, lasts for about 38 years. Its current price is only about 150 birr. The grinder(which is made of a special type of stone) costs about 60 birr and may last up to 6 years.

A water mill can be an important source of income for the farmer. The typicalEthiopian farmer is subsistence oriented and has little cash for purchasing modern inputsand consumer goods or for meeting other types of outlay. A daily income of 20 birr fromoperating water mills (see Table 1) is vital for the Ethiopian peasant. Income generated bysmall-scale rural enterprises, such as water mills, can contribute to increased demand forproducts from the agricultural and other sectors.

Water mills can create off-the-farm employment opportunities for some farmhouseholds. In the study area, an average of two people were needed to operate a watermill. In the case of a diesel mill or an electric mill, on average, three people were needed.Fixed schedules are rarely used by owners of water mills because of the often irregulardemand they face. The mill owner is assisted by family members or neighbours when thereis a peak in the demand for flour. Cash payments are rarely offered to assistants in a societywhere farm households are little integrated into the market. Modern mills, on the otherhand, often pay cash to mill operators.

Water mills are almost invariably located near streams, rivers, or springs because,obviously, they require water as a source of power. Today, they are largely restricted toinaccessible areas. An additional factor in the location of water mills is population density:a reasonable population density is needed to make a mill financially feasible.

The opportunity cost of the land used for a water mill is small. The site is oftenunsuitable for cultivation or for grazing because of the terrain, which is often very steep.The actual area used as a mill site measures about 50 m2. The water discharged from themill is often used for small-scale irrigation. The area around the mill is typically woody andluxuriant.

The relative advantages of water mills can also be analyzed from the point of viewof the customers canvassed in the household survey. The major reasons for using watermills (in order of importance) are (1) personal relations with the mill owner, (2) proximity,and (3) low service charges. Users and mill owners are often neighbours or relatives withstrong personal and social ties. Owner–customer relations sometimes involve reciprocityand similar nonfinancial dealings. A woman may frequent a particular water mill simplybecause she feels that it is her obligation to do so. She may not want to harm the feelingsof the owner, who may process her grain for free when she runs short of money.

The catchment area of a water mill is often restricted to its neighbourhood. Thelack of transportation to distant places may preclude the use of modern mills, which arelocated in towns or market places. Poorer households cannot afford the pack animals theywould need to transport the grain to town.

Water mills are also attractive to poorer households because the service chargesdemanded by owners of water mills are lower than those demanded by the owners of themodern mills. It is likely that the demand for cheaper milling technology will increase asthe decline in real per capita income in rural Ethiopia continues. Several studies havesuggested that modern mills operate much below capacity because of the rural people'sshortages of cash (Aredo 1987; Lirenso and Aredo 1988; Aredo and Abebe 1991).

In addition to the regular users, other people visit water mills only occasionally,mainly when a diesel or electric mill is malfunctioning because of a breakdown or shortageof power, especially fuel. The demand for water mills peaks when a diesel or electric millstops work. In this way, water mills complement modern mills. Water mills are morereliable and flexible than modern mills.

Modern mills are preferred for their high speed. About 42% of the samplehouseholds reported that they had frequented electric mills for this reason. People oftencombine their visits to modern mills with other tasks they are undertaking. About 21% ofthe women visited mills on their way to the market.

How do we characterize the regular users of water mills? According to thehousehold survey, people who frequent water mills are younger than those who frequentelectric or diesel mills. The average age of those who frequent water mills (of thehousehold head) was 37, whereas that of the users of electric and diesel mills was 41. Thenumber of people in the households that used the water mills averaged five, and that of thehouseholds that used the electric and diesel mills averaged six. The average size of landholdings per household of the frequent users of water mills was 0.68 ha, and that of thefrequent users of modern mills was 0.95 ha. Clients of water mills live a few minutes' walkfrom the mill site. In general, it seems that regular users of water mills are poorer andyounger households, residing near the mill.

Water mills, however, have their disadvantages:

  1. They are very slow to operate. The long waiting time may discourage householdsfrom using water mills. One way of overcoming this problem is to leave the grainwith the mill owner and collect the flour at a convenient time. There is mutualtrust between mill owners and clients.

  2. They are subject to water problems. During the rainy reasons, their sites could beflooded, which may cause work interruptions. In the extreme case, the entirestructure could be destroyed and carried away by floods, which happened to a millin the village of Chaka recently. During dry seasons, on the other hand, there couldbe too little water to run the mill. It is also at this time that people demand morewater for irrigation. Conflict between mill owners and neighbours is not unheardof. In short, irregular supply of water is a major technical problem faced by millowners. A topic for further research is, therefore, a way to ensure regular suppliesof water for grain processing, as well for irrigation. So far, there has been noattempt to address this problem. In fact, there are cases where detrimentalmeasures were taken; for example, water in the village of Gedilge was diverted tothe town of Debre Sina by a small dam in the very place where there were manywater mills.

  3. They operate very little on cloudy days, especially in the rainy seasons, becausethere is not enough heat from the sun to dry the grain brought for milling. Watermills process only dry grains.

  4. Their wide is precluded by the fact that their location depends on the availabilityof water. However, they could be promoted in the southern part of Ethiopia,where there are numerous streams, rivers, and springs.
The relative advantages of the three types of mills are summarized in Table 2.Water mills rank first for all the desirable characteristics of an appropriate technology,except for waiting time, product quality, and location flexibility. One major weakness ofa water mill is that it is location specific: its uses are restricted to places where water poweris available. Electric mills, admittedly, are restricted to places where electric power isavailable, but diesel mills can be established anywhere there is sufficient population densityand reasonable transportation facilities. Of all the characteristics listed in the table, thehighest weight should be attached to reliance on local resources.

Table 2. Ranking of milling technologies according to their relative advantages.
AdvantageWater millsDiesel millsElectric mills
Dependence on local resources
Fit with local farming system
Capacity utilization
Location flexibility
Customer waiting time
Accessibility to the poor
Contribution to interregional equity
Product quality
Working conditions
Contribution to environmental protection

Water mills fit into the local farming system by (1) making water available forsmall-scale irrigation, (2) using the spare labour of the farmer, (3) making use of the skillof local artisans (such as blacksmiths, who repair and improvise components), and (4)making use of materials available within the locality. The lower capacity of water mills isin harmony with the capacity of the local economy, which has characteristically low-leveloutput and limited cash income. Modern mills operate with high excess capacity becauseof shortages of grain and the limited ability of customers to pay service charges. Butmodern mills require less waiting time at the mill site. Water mills are accessible to poorerhouseholds and to people living in remote areas. Although consumers prefer the textureof the flour from modern mills, there are those who say that these mills "burn" the flour,meaning, perhaps, that the strong heat released by these mills tends to shorten the shelf lifeof the flour. Working conditions at water mills are appreciated because of the cool,noiseless, fresh environment. Also, water mills contribute to environmental protection,using a renewable source of energy and recycling water for irrigation.

Water mills could, therefore, complement modern mills if their designs wereimproved and policy makers appreciated their importance. They could be of immense usein relatively inaccessible areas with sufficient hydropower.

The case of Chaka

The village of Chaka is located in one of the most inaccessible regions of Ankober woreda,some 42 km from Debre Birhan, the capital city of northern Shewa. After less than anhour's drive from Debre Birhan to the town of Gorebella, one has to walk (and sometimescrawl and roll down) along a steep gorge and then cross the Airara River to reach thevillage of Chaka. The people of Chaka grow wheat, barley, horse beans, and other crops.There are about 400 households in the village. Numerous streams flow from the chains ofmountains overlooking the Airara Valley.

It was along these streams that water mills were established many years ago, withina few kilometres of each other. The village of Chaka, itself, is located a few kilometresaway from the historic town of Ankober, the seat of kings of Shewa. Minilik II moved hiscapital city from Ankober to Addis Ababa. Some of these mills were established by foreignresidents (Greeks and Armenians). For example, the oldest mill (and yet the most powerfulone in the village) was installed by a certain Mr George, some 75 years ago. In those days,hand grinding (using a stone grinder) was the most effective technique for processinggrain, and slave labour was available. Households sought milling services at the water millonly on important occasions, such as a wedding, an annual holiday, or a grand feast, whena lot of flour was needed to prepare the food. Payment for milling services was made inkind (e.g., eggs and grain). Gradually, water mills gained in popularity among the localpeople. At peak times, mills operated 24 h a day. Customers waited in line for as long as8 h at mill sites. However, those water mills gradually lost their market to the diesel millthat was established in the nearby town of Gorebella. Foreign residents switched to otheractivities as water mills became relatively unprofitable. Mr George, the owner of the oldestmill in Chaka, sold his mill to a local farmer and left the area. But the mill is stilloperational .

The mill was bought for a few hundred birr by the present owner, Mr H, who wasa part-time mill operator for the original owner. Mr H. established a workshop andundertakes all the repairs and maintenance of the equipment. He has made a number ofinnovations, including manufacturing from local materials the iron block on which theshaft is mounted . The only skill he doesn't possess for his business is the skillof manufacturing the grinders, for which he pays 600 birr every 4–6 years. The twogrinders are made of a special type of stone by local crafts people. As a by-product of hismilling business, his services as a blacksmith are provided to the village people. Both his workshop and his house are located near the Airara River. Upthe hill, he grows barley, wheat, horse beans, and other crops, and he grows vegetables,enset, and hopper, using the water discharged from the mill. He gets a substantial incomefrom the sale of these products, and he has planted eucalyptus trees along the river. Fromhis old mill, he earns about 1800 birr annually. He demands a service charge of 1–1.50 birrper 50 kg of grain. Sometimes, he provides free milling services to close relatives andneighbours. These people often help him a great deal when he has difficulties. One of hisrelatives helps him with mill operations. Mr H says that his business has been constrainedby a lack of market and some technical problems, such as shortages of bolts and barrelsfor the mill. His mill does not function from June to September, as this is the time whenthe sky is cloudy and it is difficult to get sun-dried grain for processing. Some of thereasons why he lost his market to diesel mills were that (1) his mill was slow to operate,(2) the bran was not ground into powder, and (3) better-off households considered his millan "inferior" form of technology and the diesel mill a status symbol. On the other hand, hismill has many loyal clients, especially neighbours and relatives. Peak demand for his millcoincides with the frequent interruptions in the operation of the diesel mill in Gorebella.He knows each one of his regular customers and has close personal relationships withthem. Many of them are poor, so he rarely uses the scale for weighing grain brought formilling. Weight is determined roughly: he just looks at the amount of grain. He will startthe mill no matter what the size of the load of grain or the number of customers. A loyalcustomer is not turned back simply because there is not enough business that day. He mayinterrupt farm work to start mill operations.

The working principles of a typical water mill in Ethiopia are as follows (see Fig. 1). The water jet, coming through the nozzle, causes the turbine to rotate, which drives thegrain mill. The grain enters the mill through the hopper, and the flour is delivered at theflour exit. The quality of the flour can be adjusted by varying the clearance between thegrinding stones.

The maximum output of the turbine, shown in Fig. 1, can be estimated at about6.3 HP (4.7 kW). If he installed a more efficient turbine, then an output of 15–20 HP(11–15 kW) could be obtained.

Water mills tend to have mechanical problems, which could be avoided throughsimple improvements. Mr H. identified these problems:

When a properly designed turbine is used, there are no major difficulties duringoperation. Some modern turbines were installed by the Evangelical Church of Ethiopia,and so far no major problems have been observed. The cross-flow turbine has provenespecially durable.

There are several possibilities for developing modified designs of water mills,using local resources. In the Ethiopian context, there are three possibilities: (1) improvingthe common water mills that already exist, (2) developing new water-propelled mills, and(3) making other improvements. The multipurpose mill is another possibility.


This study attempted to throw light on a neglected postharvest technology and the role itcould play in responding to the rising costs of imported diesel fuel and the growingshortages of cash incomes in rural areas. The strengths of water mills are that they makeuse of locally available materials and are accessible to poor households in remote andinaccessible areas. Water mills provide a striking but a rare case of a foreign technologythat has been almost fully "indigenized" in rural Ethiopia. The technology fits nicely intothe local farming system.

By exploring the economic and technical feasibility of water mills in selected ruralareas, this study has suggested the possibility of raising the efficiency of the water mill byabout 20–25% and tripling its horsepower through design improvements, using localmaterials. Researchers and promoters of rural technologies can develop the alternativedesigns proposed in this study. An engineering study, in particular, is highly recommendedto further investigate and develop alternative designs and the other proposals of this study.

Policymakers and rural-development practitioners may appreciate the immensepotential of hydropower-based technology, water mills in particular. The Science andTechnology Commission and the Rural Technology Promotion Department of the Ministryof Agriculture may encourage the expansion of improved water mills in selected areas.Appropriate policy instruments should be designed to encourage the expansion of watermills in areas where water is available. Some of the measures that could be taken are (1)removing the taxes imposed on water mills, (2) establishing a water-mills promotionproject within the Rural Technology Promotion Department of the Ministry of Agriculture,and (3) commissioning feasibility studies.