1

A Paper Presented by

MULAT DEMEKE

Department of EconomicsAddis Ababa University

Presented to the Agricultural Transformation Policy Workshop

Nairobi, Kenya27-30 June 1999

AGRICULTURAL TECHNOLOGY, ECONOMICVIABILITY AND POVERTY ALLEVIATION IN

ETHIOPIA

2

ACRONYMS

ADLI Agricultural Development Led IndustrializationBOA Bureau of AgricultureCSA Central Statistical AuthorityCIMMYT International Maize and Wheat Improvement CenterEARO Ethiopian Agricultural Research OrganizationESE Ethiopian Seed EnterpriseFDRE Federal Democratic Republic of EthiopiaIAR Institute of Agricultural ResearchICU Input Coordination UnitILRI International Livestock Research InstituteMOA Ministry of AgricultureNARS National Agricultural Research SystemNFIA National Fertilizer Industry AgencyNVRC National Variety Release CommitteeNSIA National Seed Industry AgencyNSIC National Seed Industry CouncilPADETES Participatory Demonstration and Training Extension SystemSNNP South Nations and Nationalities Peoples

3

AGRICULTURAL TECHNOLOGY, ECONOMIC VIABILITY ANDPOVERTY ALLEVIATION IN ETHIOPIA1

I. BACKGROUND

1.1 The challenges

Agriculture dominates the Ethiopian economy. At least 85 percent of the country’spopulation live in the rural areas. The contribution of agriculture to GDP varied between46% and 58% between 1980/81 and 1997/97 (Table 1). About 85% of the foreignexchange is also derived from agriculture.

The performance of the agricultural sector in Ethiopia has remained weak and it isheavily influenced by weather conditions. Annual growth rates dropped by 13 and 21%during the drought years of 1982/83 and 1984/85, respectively. Favorable rains in1982/83 1986/87 and 1995/96, on the other hand, resulted in growth rates exceeding14% per annum. In general, agricultural growth rates averaged 2% between 1980/81 and1990/91 (under the former military government) (Table 1). In the mean time, populationgrew by nearly 3%, leading to a sharp decline in per capita agricultural production and arise in the incidence of poverty.

Civil war, recurrent drought and the command economic system have contributed to thepoor economic performance under the former military government. The end of the war in1991 gave rise to economic reform program, which took the form of StructuralAdjustment Program (SAP) under the auspices of the World Bank and IMF. The reformincluded the removal of substantial taxation of agriculture, market liberalization anddevaluation. The restrictions on grain movements, the quota system of grain delivery (tothe parastatal Agricultural Marketing Corporation) imposed on farmers and the system offixed pricing for farm produce was abolished. The fertilizer markets was liberalized,creating a multi-channel distribution system. The overriding objective of the governmentis to attain fast, broad-based economic development. A development strategy known asthe Agricultural Development-Led Industrialization (ADLI) has been formulated to focuson agriculture. The strategy views agriculture as the engine of growth, on account of itspotentially superior growth linkages, surplus generation, market creation, and provisionof raw materials and foreign exchange.

In accordance with its policy of decentralization, the government has brought many of thesmaller research centers under the control of regional governments, while those with 1 A number of people have helped in writing this paper. In particular, I would like to thank Dr. T.S. Jaynefor his invaluable comments and suggestions throughout the different stages of the research. I would alsolike to express my appreciation to Dr. Issac J. Mindae, Professor Eric Crowford, Dr. Wilfred Mwangi, Dr.Workneh Negatu, Ato Tesfaye Zegeye and Dr. Tekalign Mamo for their comments and/or response to myfrequent inquiries. Thanks are also due to Dr. Bedada, Dr. Hailu Tefera, Legesse Wolde, MengistuAlemayehu, Dr. Mengistu Hulluka, Dr. D.G. Tanner, Dr. T.S. Payne, Dr. S.Twumasi-Afriye, HabteMariam Abate, Regass Ensermu, Asefa Taa and Tarekegn Ferede for devoting their precious time tosupply valuable information and hold discussion with the author.

4

national significance (totaling 10 research centers) were kept under IAR management.Regional research centers are expected to address specific local and regional problems.IAR was also reorganized in 1997 and renamed as EARO (the Ethiopian AgriculturalResearch Organization). The National Seed Industry Agency (NSIA) and the NationalSeed Industry Council (NSIC) were established in 1993 to implement the new seedpolicy which envisages participation of both public and private enterprises in theproduction and marketing of improved seeds. The National Variety Release Committee(NVRC) has been reorganized and placed under NSIC. NSIA is responsible for qualitycontrol and certification of seeds.

Within the framework of ADLI, a new system of agricultural extension, known as theParticipatory Demonstration and Training Extension System (PADETES) was launchedin 1994/95. The approach attempts to merge the training and visit (T & V) system withthe technology diffusion experience of the Sasakawa/Global 2000 (SG 2000 for short).2

The principle of PADETES is to demonstrate to farmers the benefits of a package ofinputs, notably balanced and higher rates of fertilizer, improved seeds, pesticides andbetter cultural practices. Farmers who agree to participate in the new extension systemallocate ½ ha of land (in some cases ¼ ha or less) for the demonstration and make a 25 –50% down payment on the inputs used at the time of planting with the balance due afterharvest. The farmers manage the plots under a close supervision and advice of extensionagents. The number of such plots increased from 3,200 in 1994/95 to 350,000 in 1995/96and to 600,000 in 1996/97. Nearly 3 million demonstration plots were handled in1997/98. The operation is assisted by officials in the administration and othergovernment offices. Being at the center of the overall development strategy, PADETESenjoy the full support of the government.

Government officials closely follow and assist extension activities. Governmentauthorities coordinate input supply and distribution. The Input Coordination Unit (ICU),organized at local (peasant association), district (woreda), zonal (province), regional andcentral levels, consists of representatives from administration, banks, input suppliers andMinistry/ Bureau of Agriculture.3 The responsibility of credit disbursement andcollection has been transferred from the banks to regional governments since 1996/97.Regional states estimate loan amount and borrow directly from the banks. Farmers(through their cooperatives or peasant associations) receive input loans from the financeoffice of the district administration. The regional government collects repayment fromfarmers and pay back its loan, and it is liable as a guarantor in case of default (Mulat etal., 1998).

The recent policy reforms and extension strategies have positively influenced theperformance of agriculture. Ignoring 1991/92 (the year when the civil war to overthrowthe former military government reached its climax), agricultural production averaged2.7% per annum (compared to 2% under the previous government). Nonetheless,

2 SG 2000 is a non-governmental organization which started its operation in Ethiopia in 1993. Theextension approach of SG 2000 is based on farmers’ half-hectare demonstration plots, utilizing improvedseeds, improved management practices, and more balanced and higher rates of commercial fertilizers.3 The ICU is chaired by the chief administrator at district, zonal and regional levels.

5

favorable policy is not the only factor contributing to the growth. Good weatherconditions, especially during the three years of 1994/95 to 1996/97, have played a majorpart in the recovery of agricultural production. Bad weather in 1997/98 has led to a 7.6%fall in output, implying that fluctuations in agricultural output have yet to be arrested. Itshould also be noted that the average growth rate of population was greater than theagricultural growth rate in the post-reform years.

There is no evidence that poverty situation has improved in Ethiopia. Pervasive povertyis manifested in the highest incidence of malnutrition and food insecurity. It is estimatedthat 64 percent of children under the age of 5 are stunted, with little chance of ‘catch up’growth as they become older. Both chronic and transitory problems of food insecurityare severe in Ethiopia. High rates of unemployment in urban areas and shortage of landin rural areas, where one-third of the households farm less than 0.5 ha, are the majorcauses of chronic food insecurity. Given the average grain yield of a little over 1 ton perhectare, households with small plots seldom produce enough to meet their family’sconsumption requirements. Transitory food insecurity is mainly caused by drought. Theprobability of a drought occurring in Ethiopia is as high as 3 out of 10 years (FDRE,1996). The moisture stress zones cover 55% of the total area of the country and anestimated 25% of the population live in these areas.4

A few attempts have been made to determine the magnitude of poverty in Ethiopia. Usinghousehold expenditure survey in Addis Ababa and Mekele (1992) and a variety of othersources of data, a World Bank Mission concluded that about half of the population inEthiopia were poor in term of any reasonable criterion in 1992. A study by theTransitional Government of Ethiopia in 1992 also estimated that 51.7% of the ruralpopulation were poor (Getahun, 1998).

One of the recent estimates of poverty, based on the construction of poverty linesfollowing the cost of basic needs approach5, showed that approximately 41.5% of thetotal population in 1994 lived below the poverty line (Mekonen, et al., 1998). Theestimate for rural and urban households was given as 41.3% and 42.9%, respectively.The figure increased to 48% in 1995, mainly because of worsening inequality as opposedto declining or stagnating income. The study also concluded that the incidence ofpoverty was greater in ‘enset’ growing of the south than in cereal producing areas of thenorthern and central highlands. Applying the poverty line proposed by Ravillion andother in the World Bank, which suggests US$1 per person per day, Dercon and Kirshnan(1996) showed that 85% of the total population was below the poverty line at thebeginning of 1994.

4 Ethiopian Mapping Authority.1988. National Atlas of Ethiopia. Berhanena Selam Printing Press. AddisAbaba.

5 The poverty line was defined as the per capita monthly expenditure required to buy food (providing 2,200kgcal per person per day) and basic non-food items. Such approach, also known as the total consumptionpoverty line, gave 53.44 birr (or US 7.8) per month per person.

6

Table 1: Ethiopia: Sectoral Contribution of GDP (1980/81 - 90/91)

AGRICULTURE

YearGrowth rate ofReal GDP

As% of GDP Growth

rate

Growthrate of perCapitaGDP

1980/81 58.11981/82 0.51 55.7 -3.6 -2.41982/83 10.1 57.5 13.6 6.91983/84 -6.3 53.7 -12.5 -91984/85 -9.7 47.0 -20.9 -12.41985/86 9.9 49.6 16.0 6.41986/87 14.0 51.7 18.8 10.41987/88 -0.1 50.3 -2.8 -3.01988/89 0.3 50.6 1.0 -2.71989/90 4.1 51.2 5.3 0.91990/91 -4.2 56.3 5.2 -7.2Average 1.9 52.9 2.0 -1.21991/92 -3.7 56.8 -2.7 -6.71992/93 12.0 53.8 6.1 9.01993/94 1.6 51.0 -3.7 -1.61994/95 6.2 49.7 3.4 2.91995/96 10.7 51.5 14.7 7.41996/97 5.6 50.5 3.4 2.41997/98 0.5 46.4 -7.6 -2.6Average 4.7* 51.4 1.9 1.6

* The average growth rate of GDP for the years 1992/93 – 97/98 (i.e. excluding the year 1991/92 whichwas a year of political instability) is 6.1% and the corresponding per capita GDP growth rate is about 3%.Source: MEDaC, National Income Accounts, Revised Series,1998.

1.2 Objectives

The historical experience of the medium and high income countries of the world have thefollowing in common: substantial increases in agricultural productivity and eradicationof pervasive poverty were associated with a transformation from a predominantlyagrarian, semi-subsistence economy to one where a significant portion of the labor forceis engaged in non-agricultural activities. This demographic transformation has reachedits peak in the United States, where today only 2% of the population are farmers.Understanding why and how such a structural transformation has occurred along withincome growth and agricultural productivity growth in these countries may provideimportant insights for Africa – specifically for identifying how the continent may breakout of the low-input, low-productivity, semi-subsistence poverty trap that characterizesmuch of the continent.

7

Agricultural productivity growth and the associated agricultural transformation processalso has almost universally involved a transition from a raw resource-based agriculture toa science-based agriculture (involving new productive technologies in nutrient and watercontrol, new seed cultivars, and/or labor-saving techniques of land preparation andharvesting). This paper assesses three crop technology packages, namely maize, tef andwheat, and one livestock technology, dairy-traction, the current evidence of theireconomic viability, and major factors within the broader economic system that influence(and can potentially improve) economic viability of these technology packages inEthiopia. It looks at the role of improved technologies in poverty alleviation andagricultural transformation. The paper concludes by identifying areas where furtherpolicy attention is needed to increase the use of available crop technology.

1.3 Conceptual framework

On-farm trials attempt to test biophysical feasibility, farmer profitability of prototypesand acceptability through different trials, design and management. Although they seemto be quite fashionable, they do not promise to realistically enable the farmer to assess theprofitability of the technology as the trials are carried out using complementary inputsand management practices which are seldom feasible in actual practice. Moreover, thetrials hardly go beyond one season and a few selected sites.

Economic profitability of an agricultural technology can, for simplicity, be captured bythe following model:

Profit = (PyY – ΣPxiXi)

Where:Py = price of the physical output from the use of the technologyY = physical output response from the use of the technologyPxi = price of input Xi applied in the production of YXi = quantity of input Xi applied in the production of Y

The model shows that the profitability of technology depends on three main factors: (1)the response rate to technology application; (2) the price of output; and (3) the cost or theprice of the technology or inputs applied. The response rate depends on the quality of thetechnology, the natural growing conditions, management factors and availability ofcomplementary inputs. Superior technologies are less risky and generate large economicbenefit, and farmers adopt such innovations with enthusiasm. Good land preparation,timely sowing, and effective control against weed and insect are some the managementenvironment determining output response. The extension system plays a vital role inimproving the management capacity of small producers.

The price paid to farmers depends on the performance of the output marketing system.Physical infrastructure such as roads and market sites reduces transaction costs andensures more competitive price. Increase in farm wholesale prices resulting from more

8

competitive price formation act as production incentives. Good national harvest (due towidespread use of new technology) may depress grain prices but efficient grain marketmarketing system, together with increased productivity, can partially or wholly offset thefall in prices.

A number of variables influence the price paid by farmers to agricultural inputs. The costof improved seeds, for instance, includes the performance and efficiency of seedmultiplication, processing and marketing. A competitive wholesale and retail operationsensures access and reduces the cost of seeds. Deregulation and liberalization policies areexpected to increase the efficiency of both production and marketing of seeds. Accessand cost of credit also influence the cost of input.

This study is meant to provide some insights into the extent to which the recommendedtechnologies are profitable when all the different factors affecting the farming system areconsidered. More specifically, the study will attempt to answer questions such as:

- Is the recommended technology profitable? If yes, at what level, and to what extentin relation to other competing enterprises at the farm household setting?

- If it is not profitable, then what are the implications for policies related to agriculturalresearch agenda for the program leaders within NARS, for research administratorswithin NARS and to the Ministry of Agriculture?

- If the technology is profitable and yet adoption is low, especially among poorhouseholds, then, there is a need to look at policy areas that may be hindering theprocess of adoption and determine ways to mitigate them. Among these policy areascould be availability of agricultural inputs, transportation, storage of inputs andoutputs, and access to credit.

1.4 Methodology

The study is mainly based on a review of previous works on profitability of agriculturaltechnologies. The bulk of the information has come from recent studies carried out bythe Grain Market Research Project of the Ministry of Economic Development andCooperation, research reports of the Institute of Agricultural Research (now renamed asthe Ethiopian Agricultural Research Organization, EARO) and the InternationalLivestock Research Institute (ILRI). In addition, interviews and group discussion withresearchers at Kulumsa, Bako, Debre Zeit and Holleta research centers as well asCIMMYT have been used to update or bridge gaps in the information obtained fromsecondary sources. Information was also obtained from interviews and groupdiscussions with extension agents and farmers living around Kulumsa, Bako, Debre Zeitand Holleta research centers and Jimma. Discussions were held with experts of theMinistry of Agriculture (HQ), the Ethiopian Seed Enterprise (ESE), some majorimporters and distributors of fertilizer, the National Seed Industry Agency (NSIA) andthe National Fertilizer Industry Agency (NFIA).

9

II. SOME ISSUES IN TECHNOLOGY DEVELOPMENT ANDAGRICULTURAL TRANSFORMATION

Agricultural transformation is characterized by a process in which the share of agriculturein a country’s labor force and total output declines. The declining importance ofagriculture has been documented in nearly all developed countries. However, historicalevidence also confirm that rapid agricultural growth often accompanied or precededgeneral economic growth. The spectacular industrial revolution in England would havenot been possible without agricultural revolution that preceded it. Rising laborproductivity created opportunity to grow more food with less labor and to release ruralmanpower for capital construction and industrial development in England and otherEuropean counties and the United States. Agricultural growth was also a critical part ofthe process of the industrial revolution, because it reduced the price of food and made itpossible over the long run for an increasingly large portion of the population to engage innon-agricultural activities and be sustained by an increasingly small number of farmers.

At least two mechanisms account for the process of rapid agricultural growth andeconomic transformation reflected in the decline of the sector’s share in total output andlabor force in developed countries. First, a less-than-unitary income elasticity for theproducts of the agricultural sector guarantees that gross value of sales by farmers willgrow less rapidly that GDP. Second, rapid growth in agricultural productivity measuredby output per laborer or output per hectare, combined with slow growth in demand (dueto low-income elasticity), translates into declining long-run terms of trade or fallingprices for farm products. Lower prices for agricultural products put pressure onagricultural resources to move out of farming and into the more rapidly growing sectorsof the economy. The process of rapid technical change in agriculture in the 20th centuryfurther reduced prices and agricultural incomes. Nearly all rich countries responded byintroducing measures to protect the agricultural sector from international competition andproviding subsidies to counter the fall in agricultural income (Timmer, 1990).

A link between agricultural and overall economic growth has also been observed incurrently less developed countries. It has been concluded that GDP growth is positivelyand strongly correlated with agricultural growth. 6 Expanding agricultural productionthrough technological change and trade creates important demand for the outputs of othersectors, notably agricultural inputs, a wide range of consumer goods, transportation,commercial services and construction (World Bank, 1982). Nonetheless, poorperformance of agriculture has undermined economic growth in many developingcountries. The next section examines some of the problems of faced in sustaining rapidgrowth and transforming agriculture in Africa.

6 The only exception are some of the oil or mineral-based economies such as Algeria, Ecuador, Moroccoand Nigeria.

10

2.1 Technology development and poverty alleviation

The performance of agriculture and the economy in Sub-Saharan Africa has beendisappointing over the last three decades. Per capita food and cereal output has beendeclining in many parts of the region. Owing to lack of technological progress, the gapbetween domestic demand and domestic agricultural production has widened.

The contribution of agriculture to production and employment has been declining in Sub-Sahara African countries. The share of agriculture in the total GDP declined from 36% in1980 to 30% in 1995. The percentage of labor force in agriculture declined from 72 to68% over the same period (World Bank, 1997). However, these trends do not reflect anyfundamental transformation of Africa’s economy. For the region as a whole, new oilextraction simply made up for lagging agricultural production.

Productivity growth in agriculture is necessary to enable agricultural income to rise fasterthan rural population growth – i.e., to achieve growth in rural incomes. However, giventhat rural populations in Sub-Saharan Africa are still rising at a rate of 2-3% per year,there would need to be sustained agricultural growth of at least 5-6% per year toappreciably raise rural incomes unless rural population growth could be curbed. Ruralpopulation could be curbed either through declining birth rates and/or through rural-urbanmigration. Historically, sustained rural-urban migration has typically occurred through anon-farm engine of growth that was able to absorb rural labor off the farm, which inEurope manifested as the industrial revolution. Apart from technological changes, anincrease in the area cultivated per agricultural worker was one of the essential conditionsof an increase in productivity. By contrast, the non-farm sector in Sub-Saharan Africahas been in malaise. Modern industry remains mostly small, stagnating at around 12 -15% of GDP and 9% of employment between 1980 and 1995 (World Bank, 1997).Limited employment opportunities in the industrial sector has eroded the prospect ofrural-urban migration. In view of the population explosion, it is likely that cultivated areaper worker will further decline in Africa, implying that rapid technological change is thesole means of raising labor productivity.

Transforming agriculture and expanding its productive capacity is the prerequisite forsustained economic growth in Sub-Saharan Africa. It is impossible to stabilize themacroeconomy without stabilizing the food economy. Food prices are so important inthe overall welfare of consumers and producers that some reasonable degree of stabilityis essential (Timmer, 1993). Faster agricultural growth increases supply andstockbuilding of food, hence restraining and stabilizing local price of food staples. Themain beneficiaries of lower prices are poor households who spend 40 to 60 percent oftheir income on food items. Moreover, agriculture is often more labor intensive andfaster agricultural growth is expected to generate more employment for landless and near-landless groups. It has been reported that poor households in Africa have been mostseriously harmed by the negative growth in real farm output per person (Lipton, 1998).

It is widely thought that important lessons can be drawn for Africa from the GreenRevolution technology, which significantly increased production in many Asian and

11

Latin American countries over the last three to four decades. It gave rise to a dramaticincrease in yields of some major crops such as wheat and rice in parts of Mexico, thePhilippines, India, Pakistan and Thailand. The Green Revolution has increased theoverall economic growth rates of these countries, moderated the rate of inflation, reducedimports of food and thereby eased balance of payment problems, and raised supplies andstocks of food staples. For instance, the rice revolution in the Philippines resulted in thegrowth of rice output by 4.3% a year (1966-82), compared to 2.2% a year in the pre-green revolution period (1955-66). Per capita production of rice rose by about 1.5% ayear in the post-green revolution period. In addition to intensive plant breeding, majorgovernment investment in roads, irrigation and other infrastructure contributed to thesuccess. In India, the area devoted to high-yielding varieties of rice and wheat increasedrapidly between 1966 and 1997. As a result, the country has been able to reduce itsdependency on imported food and achieve self-sufficiency in grains (Griffin, 1989).

The Green Revolution technology was generally considered labor intensive with afavorable impact on rural poverty. A fall in the real price of food, following a dramaticincrease in production, was expected to benefit the rural as well as the urban poor.Nonetheless, the contribution of Green Revolution in reducing poverty has not been acomplete success. In some situations, Green-Revolution technologies have had adisappointing (though positive) total impact on the poor partly because of problems ofaccess to financial resources or delayed adoption by poor farmers. Owing to lateadoption, poor farmers often face low output prices and the benefit gained is inadequate,especially when compared to the various types risk involved. More seriously, farmemployers may reduce the effect of rising demand for labor on worker’s income by usingalternatives to labor, such as mechanical equipment and herbicides (Lipton, 1998).Distortions in output and input prices (e.g. subsidies on labor-displacing technologies)have often led towards mechanization well before labor becomes scarce. The proportionof the rural population below the official poverty line remains high in India after theGreen Revolution. There has been a decline in the real wage rate in most sectors of thePhilippines agricultural economy in the post-Green Revolution period (Griffin, 1989).Thus it would seem that any technological transformation for alleviation of povertyshould include a number of economic and institutional reforms that favor smallholders.

Limited technical progress in Africa has fostered an active debate over the performanceof technology generation and dissemination systems. For example, questions arefrequently posed whether viable food crop technology packages exist in Africa that arerelevant for smallholders (but are not being fully utilized for other reasons) or whetherthe available on-shelf food crop technologies simply aren’t viable enough to appreciablystimulate agricultural growth. Another question that may be posed is why, despite anempirical record indicating very high rates of return to agricultural research in Africa, hasfood crop production growth remained mostly stagnant in these same countries in whichthe empirical investigations were carried out.

Low adoption is generally perceived differently by those within and outside the nationalagricultural research system (NARS). For those within NARS, the main reasons forlimited adoption of new technologies are difficulties in the dissemination and adoption

12

process including weak extension system and inadequate input marketing systems. Thoseoutside NARS, on the other hand, emphasize that there was little or no economic benefitassociated with many of the technologies released, especially when one considers thehigh risks encountered by farmers. This paper argues that the real explanation for lack ofagricultural transformation in Sub-Saharan Africa is found in a combination of bothperspectives. The paper will further argue that the question of economic viability of anytechnology cannot be answered independently of the sub-sector of which it is a part,including the extension system, the input distribution system, the credit system, and theoutput marketing system.

Most observers outside NARS and a number of researchers within the system agree that anumber of constraints have undermined the effectiveness of agricultural research in Sub-Saharan Africa. Carr (1989), for instance, observed that ‘After forty years of breedingresearch on sorghum and millet at internationally- supported research station in westAfrica, less than 5% of the crops are planted to such material because it does not meetmost farmers’ needs’. Research has failed to provide appropriate technologies even inareas where physical and economic conditions are favorable. Farmers in many highpotential areas continue to rely on local seeds and traditional cultural methods.

One of the main reasons for the failure of natural agricultural research is lack ofgovernment commitment to research. Because of its long gestation period, agriculturalresearch does not have immediate potential political value. Research cannot develophighly focussed and effective research programs without the assurance of a coreoperating budget over an extended period of time (Collins, 1991). The absence ofcoherent agricultural research policies, lack of operational agricultural research resourceallocation criteria, inadequate and unstable research staff and poor management of NARSalso have been identified as the most important constraints (Idachaba, 1986). Researchmanagers lack technical capabilities to provide guidance and effective leadership as theyare usually appointed on the basis of political patronage rather than professionalexpertise. The institutional arrangements for administering research institutes, includingsupervising ministry, gets changed almost as frequently as there are changes in regime.Parastatals with statutory responsibility for research institutes are created and abolishedwith alarming frequency and ease, imparting a basic instability to national agriculturalresearch system. Research priorities and programs change and ongoing activities areterminated with changes in leadership and/or institutional setup. The absence ofmonitoring and evaluation or accountability has led to misuse and inefficient use ofresources. Competent researchers often migrate in protest against the uncertainty andautocratic leadership and poor incentives (Idachaba, 1986). Donors have oftenexacerbated the situation by bidding away competent researchers from their mandatedactivities in the NARS to engage in other less important activities (often administrativeand management activities).

There are few institutional arrangements allowing small farmers to participate in R & Ddecisions. Smallholders generally lack organization and political support to exert anyinfluence in setting research priorities. Funds are provided on the basis of block grantsand are controlled by bureaucrats instead of the stakeholders (Byerlee, 1998). Under the

13

circumstances, research projects are formulated and executed with little regard for thereal needs of farmers. For instance, chronic drought has threatened agriculturalproduction in many parts of Africa. Soil mining and soil depletion from poor culturalpractices have become widespread in the more densely populated high potential areas(Blackie, 1994). The price of commercial fertilizer has generally risen in the 1990sowing to currency devaluation and poor marketing infrastructure. Yet, breeding programsin sub-Saharan Africa have continued to focus on developing varieties adaptable tofavorable weather conditions and require high levels of inorganic fertilizers. Suchtechnologies perform very badly under conditions of moisture stress and low inputapplication. Agricultural researchers have emphasized fertilizer as a solution to decliningsoil fertility, at the expense of research on alternative means of maintaining soil fertility(Byerlee and Heisey, 1996). Applying fertilizers has become economically risky in drierareas as rains often fail at the critical stages of plant development. Breeding programsaimed at developing low phosphorous and nitrogen tolerant varieties and improvingcultural practices to increase soil moisture and crop nutrients have yet to attract theattention of many researchers.

It would be irrational to assume that the problem of agricultural transformation can beaddressed by improving the system of technology generation alone. Indeed, it hasincreasingly become clear to all those working within and outside NARS that theperformance of the extension system has remained weak even where appropriatetechnologies are available. Extension approaches and modalities have been changingfrequently with little concern for continuity and stability of the system. Resourcesallocated to extension are minimal and a good part of it is used to cater for the needs ofthe larger, more commercialized farmer rather than for the small, low-resource farmer.Extension staff operate with the assumption that they know what farmers need and oftenact autocratically. Often blanket recommendations are given disregarding theheterogeneity of the natural and human resources. More time is spent on communicatingrecommendations, rather than developing the skills of farmers to understand and adoptmore complex recommendations. In many cases, extension staff have to combine theireducational and training activities with monitoring and policing activities, such asenforcing soil conservation regulations, thus further limiting the time devoted to technicalproblems of farmers (Botha, 1995).

An extension officer may find it difficult to persuade a farmer to adopt new inputs unlessthere are efficient product and input markets that render the technology profitable (Fraser,1994). Adoption to technology is enhanced provided output prices are stable,input/output price ratios are favorable, and input delivery involves low transaction costsand risk sharing (Brossier, 1991). Governments in many developing countries createdlarge Parastatals to produce and distribute seeds. This was justified as a way toencourage the spread of new varieties by offering lower prices and reaching a wider area.However, inefficiency and subsidies imposed heavy burdens on development budgets andthe corporations found themselves with excess stock of seeds. Demand for commercialseeds also declined owing to sub-standard methods of seed multiplication and poorquality (Pray and Ruttan, 1990). Areas with poor infrastructure and limited access tomarket lacked adequate incentives to adopt new technology. Poor provision of credit and

14

facilities for product marketing and the purchase of inputs have further discouraged theuse of innovations. All of these systems influence the output and input prices as well asagronomic response rates of using a given technology, and hence are critical componentsof farm budget analysis used to determine economic profitability.

Any effort that is directed towards transforming agriculture cannot be effective without acoordinated effort among various actors in agricultural development. Poor linkagesbetween research, extension, seed companies, input and grain traders, developmentagents, farmers and farmers’ organizations have thwarted any move towards concertedand effective intervention in agriculture. Only effective linkages between the differentactors will ensure that “… research is relevant and innovative, that extension agentspresent appropriate and acceptable messages and technology, and that farmers play a keyrole in research and in designing and disseminating technology’ (Botha, 1995).

2.2 Technical change and agricultural productivity in Ethiopia7

Cereals are by far the most dominant among field crops in Ethiopia, accounting for88.3% and 83.2% of the total production and cultivated area, respectively, for the period1980 to 1996 (Annex I ). The share of pulses and oil seeds was 10.9 and 0.8% in thetotal production and 14.5 and 2.3% in the total area harvested.

Cereal output increased by 1.4% per annum during the period 1980-1996.8 This was dueto 0.9% growth in area under cultivation and 0.5% growth in yield. Expansion in areaunder cultivation was more important than increase in yield levels. In any case, theperformance of cereals remained poor, especially when compared to the rapid populationgrowth. Grain production improved after 1991, owing to improved policy environment,increased availability of inputs such as fertilizer and the relatively good weather.

Among cereal crops, tef, maize and wheat (the three crops chosen for this study)accounted for 31,17 and 13% of the total area, respectively, between 1980 and 1996(Annex II).

Tef is the most important grain crop in Ethiopia. It accounted for 31% of the area undercereals between 1980/81 and 1996/97. Its flour is favored above all other grain floursfor making fermented flat bread (enjera), a staple food item in most parts of the cerealgrowing areas. Although tef is found in almost all cereal growing areas, the major areasof production are the central and northern highlands. It is widely grown at altitudesranging from 1,700 to 2,800 masl. Tef is well adapted to the heavy, well-drained soils ofthe Ethiopian highlands (Caldwell, 1992), where most other cereal crops cannot be growneasily (due to the waterlogging conditions).

7 A more comprehensive discussion of productivity by major crop types is given in Mulat Demeke. TheStructure and Performance of Ethiopian Agriculture. Ethiopian Economic Association. Addis Ababa.November 1998.8 Annual growth rates were computed using the least squares method. The log-linear growth rate wasestimated by regressing the logarithmic values of output, area or yield against time.

15

Annual tef production averaged 1.3 million ton between 1980/81 and 1996/97 andregistered a growth rate of 1.1% per annum (Annex II). Since tef yield showed nostatistically significant increase during the period under consideration, the growth wasmainly due to expansion in area under cultivation. The average tef yield was also low,8.9 quintals per hectare. Tef yield was 26 to 28% lower than other small cereals such asbarley or wheat. Nonetheless, the market price of tef is often up to 100% higher,indicating that the return from tef could be superior (assuming differences in costs ofproduction are not significant).

Tef accounts for the highest share in the total fertilizer consumed by the peasant sector.In 1996/97, for instance, it absorbed 45.2% of the total fertilizer applied in the peasantsector or 49.3% of the fertilizer applied on cereals. About 54.2% of the tef area wasfertilized and the rate of application (taking all tef area) was 54 kg per ha. The rate ofapplication for fertilized area was 99 kg per hectare (Table 2). By contrast, only 1.7% ofthe tef area was covered with improved seeds in 1996/97.

Maize is mainly cultivated in the south, western and eastern highlands with altitudes of1000 to 2000 masl. Its share in the total cereal area was 17% between 1980/81 and1996/97. It is consumed as porridge, enjera (fermented flat bread) and bread. It is alsowidely used for making local drinks. An insignificant amount maize is used to feedlivestock.

Mean annual maize output which amounted to nearly 1.5 million between 1980/81 and1996/97 increased to 1.9 million ton, denoting the largest share (28.4%) among cereals,in 1992/93 - 1996/97 (Annex II). Maize output increased by nearly 1.9% per annumduring 1980/81 - 1996/97. This growth rate is explained mainly by expansion of area(which increased by 1.4% per annum). Area under maize rapidly expanded after thereform, increasing from 835,590 ha in 1993/94 to over 1.3 million in 1996/97. Theaverage yield of maize was higher than other crops (16.7 qt/ha), but there was nostatistically significant yield increase during the period 1980 to 1996.9

Maize production is not characterized by high rate of fertilizer application: only 33.8% ofthe maize area was fertilized and the average rate of application (for all maize fields) was24 kg per ha in 1996/97, well below the average for cereal (35 kg/ ha). However, thepercentage of maize area covered with improved seeds was the second largest, 4.6%,(next to wheat) in 1996/97 (Table 2). Improved seeds other than hybrid seeds areexpected to be multiplied by farmers and used as local seeds. According to CSA, maize isthe most important crop fertilized with organic matter (manure).10

The major wheat-producing areas lie between 1500 and 2500 masl. The major growingzones are Arsi, Bale, Kembata and Hadiya. Ethiopia was a major supplier wheat to theNear East during World War II and continuing around 1947. The county’s ability toexport wheat significantly declined thereafter as local demand increased and it becameimpossible to compete with better quality wheat produced by other countries. Ethiopia

9 The coefficient of time in a simple regression of yield against time is also significant.10 See for instance, CSA, Statistical Abstract, 1997.

16

has now come to depend on foreign aid and commercial imports of wheat to meet itslocal demand (Cladwell, 1992).

Average annual wheat production amounted to 0.75 million tons or 13% of the cerealarea between 1980/1 and 1996/97 (Annex II). Wheat production and area increased by1.6% and 0.7% per annum, respectively, during the period 1980/81 to 1996/97. On theaverage farmers achieved a yield of 12.2 quintals per hectare. Among cereal crops, astatistically significant yield increase (0.9% per annum) was observed only in the case ofwheat.

Wheat area covered with improved seeds in 1996/97 was the largest (6.5%) amongcereals (Table 2). The actual area is believed to be higher since farmers often refer toseeds saved from previous harvest or obtained from other farmers as own or local seedsalthough these were improved seeds originally. Wheat is also the most fertilized amongall crops. Nearly 58% of the wheat area was fertilized in 1996/97 and the rate ofapplication was 74 kg per ha. The rate of fertilizer application rose to 128 kg per ha forfertilized area.11 (Table 2). Pesticide was also more commonly applied on wheat than onany other crop.

One area in which there has been little change over the years in Ethiopian agriculture isthe technology of land preparation. Except for some small areas of the highlands wherehoe cultivation is practiced, all land preparation is carried out with oxen pulling thetraditional plough. It is believed that the Ethiopians inhabiting the highlands wereintroduced to animal traction between 1000 and 400 BC (Goe, 1987). The traditionalplough, which remained unchanged (perhaps since its introduction), require severalpasses to prepare land for planting. In most highland areas, some 5 or 6 months of theyear are spent on seed-bed preparation, which involves a pair of local zebu oxen pullingthe traditional plough. Of the estimated 31 million head of cattle, about 9 to 10 millionare used for draught purposes. Apart from taking most of the agricultural labor time andrequiring large number of draught animals, the existing technology of land preparation isof little use for turning the stubble and weeds into the soil (Abiye and Saleem, 1996).The result has been high level of weed infestation, one of the major factors contributingto low yield levels in the country.

In general, yields stagnated while area under cereals increased by about 1% per annumbetween 1980 and 1996. Since rural population increased by 2.5% per annum over thesame period, agricultural labor productivity or per capita agricultural production declinedby 1.5% per annum. It should be also be added that area expansion was achieved throughcultivation of hillsides and areas with steep slopes where erosion poses serious threat tosustainable farming. Land degradation was also exacerbated due to reduction orelimination of fallow land and conversion of woodland and forest areas into farmland.

11 These figures seem to be on the high side but the favorable position of wheat is unquestionable.

17

Table 2: Estimate of improved seed, irrigation, fertilizer applied area and quantity of commercial fertilizerapplied (1997/97)

Area (000'HA)

TOTALCROP

IMPROVED SEEDAPPLIED

IRRIGATED PESTICIDEAPPLIED

FERTILIZERAPPLIED

COMMERCIALFERTILIZER

RATE

CROP TYPE AREA AREA % AREA % AREA % AREA % (OOO'QUINTALS) QT./HACEREALS 6688.56 152.52 2.28 39.22 0.59 604.05 9.03 2501.98 37.41 2361.73 0.353 Tef Barely

2167.77697.67

37.241.01

1.720.14

4.79 0.22 318.5534.64

14.694.97

1175.65207.60

54.2329.76

1163.90177.30

0.5370.254

Wheat Maize Sorghum Millet Oats

772.231316.871399.95290.6643.41

50.2960.343.64

6.514.580.26

0.8013.6918.97

0.101.041.36

158.8323.4562.782.93

20.571.784.481.01

443.69445.52107.98102.3919.15

57.4633.837.7135.2344.11

568.39318.9471.0565.2110.35

0.7360.2420.0510.2240.238

PULSES 905.35 8.78 0.97 78.10 8.63 74.21 0.082OILSEEDS 461.21 33.04 7.16 32.25 0.070OTHER TEMP.CROPS 758.52 3.65 0.21 0.79 0.04 57.56 3.27 61.63 0.035

ALL TEMP.CROPS 8233.64 153.52 1.86 46.74 0.57 516.35 6.27 2670.68 32.44 2527.34 0.307ALL PERM.CROPS 591.76 11.86 2.00 21.47 3.63 7.83 1.32 174.18 29.43 63.74 0.108 Chat Coffee Enset Cotton Tobacco Fruits Other perm.

81.23191.6326016.212.0917.7422.86

11.74 6.134.536.11

3.874.61

5.583.19

21.8220.17

2.00 2.460.000.000.000.000.000.00

25.1427.92112.980.300.312.744.79

30.9514.5743.451.8514.8315.4520.95

48.658.912.990.000.000.001.03.

0.5990.0460.0120.000.000.000.045

ALL CROPS 8825.4 165.38 1.87 68.21 0.77 624.18 7.07 2844.86 32.23 2577.65 0.292* Includes manure

Source: CSA Statistical Abstract, 1997 Mainly refers to manure

18

III. TECHNOLOGY DEVELOPEMNT, PROFITABILITY AND ADOPTION

The formal beginnings of public agricultural research and extension in Ethiopia can betraced to the establishment of agricultural education establishments in the late 1940s and1950s. The staff of the Ambo and Jimma agricultural schools and the Alemya College ofAgriculture carried out simple experiments and extension activities. The Institute ofAgricultural Research (IAR) was established in 1966 with the formal mandate toformulate national agricultural research policy guidelines and undertake crops andlivestock research. About three-quarters of the total full-time researchers in the countryare believed to be placed in IAR, with the balance working mainly at the variousagricultural education institutions (Goshu, 1995). A major agricultural extension workbegan with the initiation of several package projects in the late 1960s and 1970s. It wasthought that concentrating resources on the most promising regions would yield betterresults than spreading resources thinly over a larger area. The package consisted ofmainly improved seeds, fertilizer and chemicals. The aim of this section is to examinethe impact of technology development and dissemination in Ethiopia with respect towheat, maize, tef and traction selected as case studies for the study.

3.1 Wheat

3.1.1 Research

Wheat improvement research in Ethiopia started in 1949. Testing of potential varietiesat multi-location under different seasons has been the major strategy used in releasingwheat varieties with high yield potential, good agronomic characters, wide adaptation anddisease/pest resistance (Amsal, et al., 1996).

Wheat research is conducted at Kulumsa, Holleta, Debre Zeit, Sinana, Adet and Mekeleresearch centers, representing different agroecological conditions and regionaladministrations. The National Wheat Coordination, located at Kulumsa ResearchCenter, coordinates varietal development and trials conducted at the various sites. Theprocess of releasing a variety often takes a long time and it is sometimes unnecessarilydelayed because of the stringent release mechanisms. Before a variety is recommendedfor release by the National Variety Release Committee (NVRC), it is evaluated fordisease resistance, yield performance, stability, and quality. Materials evaluatedpositively in the Advanced Observation Nursery (AON) are evaluated further (mainly foryield, disease resistance, and other desirable traits) in the Preliminary Yield Trial (PYT)and the Pre-National Variety Trial (PNVT) for two years. Promising genotypes areadvanced to the National Variety Trial (NVT) for further evaluation of two to three years.Some of the promising lines in the NVT may be included in the Cooperative VarietyAdaptation Trial (CVAT) and tested for three more years at more than 20 locations inmajor wheat-growing areas of the country. The materials with best performance at theNVT and CVAT are included in the Variety Verification Trial (VVT), conducted bothon-station and on-farm under recommended and farmers’ levels of management for oneto two years. The main objectives of the VVT are to evaluate the productivity andprofitability of the variety under real production system and to obtain farmers’ pre- and

19

post-harvest assessment of the variety. The results of the VVT are among the majorconsiderations of the National Variety Release Committee which is composed ofprofessionals from different research and seed user organizations such as the EthiopianSeed Enterprise and the Ministry of Agriculture. Farmers are not directly represented inthe Committee (Regassa, et al. 1998).

Wheat research in Ethiopia has produced over 30 bread and durum wheat varieties in thelast 40 to 50 years. However, the majority of these varieties are out of production due torust attack. A newly released, high yielding variety may be put out of production due toattack by a co-evolving new race or type of rust. Nearly all the major wheat growingareas are affected by different types of rusts, mainly stem, leaf and stripe rusts. Fieldtrials at hot-spot sites for specific diseases have reported a yield loss of up to 96% forstrip rust, 75% for leaf rust and 61% for stem rust (Amsal, 1996). Limited progress indeveloping and selecting bread wheat cultivars with adequate and durable resistance tothe different strains of rust has seriously constrained wheat production in Ethiopia((Asefa, et al., 1998 ).

Since the resistance of many popular bread wheat varieties to stripe and/or stem rust wasbroken down between 1988 and 1993, four new varieties (HAR 1685, HAR 1709, HAR604 and HAR 710) were released over the period 1993 to 1995. Of the four, the grainyield response of the semi-dwarf cultivar HAR 1685 (locally named as Qubsa) has beenvastly superior and variety was released on a national scale (along with HAR 1709). It isresistant to lodging and shattering and has yield potential of 50 to 70 qt/ ha (ESE, 1997).Agronomic study showed that HAR 1685 outperforms other varieties under conditions oflow (or no) and high fertilizer application. However, the variety is found highlysensitive to competition from grass weed (Asefa, et al., 1998). HAR 1685 is alsomoderately susceptible to stem rust.

3.1.2 Profitability of HAR 1685

A partial budget analysis showed that HAR 1685 was profitable under the recommendedmanagement.12 The return to land was 2,760.63 birr per ha and the return to labor was47.83 birr per labor day. Under farmers’ management, HAR 1685 gave a return of2,172.15 per ha (down 21.3%) and 43.45 birr per manday (down 9.2%). The marginalreturn and cost due to the recommended management (relative to farmer’ management)were 1,317.60 and 729.12 birr, respectively. Hence, the marginal rate of return was80.7%, lower than the recommended rate of 100% (Table 3).

12 Data for the profitability analysis came from two sources: (1) Mohammed Hassena, Regassa Ensermuand Kefyalew Girma. 1996. ‘On-farm verification of advanced bread wheat lines under recommended andfarmers’ management levels in Arsi Region’ and S. Gavian and Gemechu Degefa. 1996. The profitabilityof wheat production in Ethiopia: The case of Tiyo woreda in Arsi Region’ both in D. G. Tanner, op cit.Yield figures were from the first study and costs from the second. The recommended management requiredthe use of improved seeds along with the application of large amount of urea, two types of herbicides (onefor grass weeds). Farmers’ cultural practices were used in both trials.

20

An old variety (named Dashen), known to have high yield potential but susceptible torust, was used for the check/control trial. The net return to land of HAR 1685 under therecommended management gave better result than the old variety under farmers’ ownmanagement. However, the return to labor was slightly higher in the latter case (48.99 vs.47.83). Both the return to land and labor under check was better than that obtained usingHAR 1685 under farmers’ management. The superior performance of the Dashen wasdue to the low incidence of stripe in 1994/95.

High yield variability across sites within the same district was also observed in 1994/95.For instance, HAR 1685 yielded 45.73 quintal per ha at Gonde but gave only 15.57 atEtheya, only 15-20 kms away from Gonde. A similar wide variation was recorded acrossdistricts (Mohammed, et al., 1998).

21

Table 3: Partial budget analysis of wheat production atEtheya/Gonde - 1994/95

HAR 1685 under HAR 1685 under Check (Dashen) under farmerRecommended Management Farmer Management Management

Quantity Unit price Value Quantity Unit price Value Quantity Unit price ValueYIELD 4684.25 3366.65 3638.3 Grain 21.77- 30.65 qt/ha 30.65 145 4444.25 21.77 145 3156.65 23.54 145 3413.3 Straw 14 - 16 qt/ha 16 15 240 14 15 210 15 15 225

0 0 0INPUT COSTS 1135.62 510.5 460.8 Seed 1.75 -2.24 qt/ha 1.75 214 374.5 1.75 214 374.5 2.24 145 324.8 DAP 0.5 qt/ha 0.5 178 89 0.5 178 89 0.5 178 89

Urea 0 - 1.59 qt/ha 1.59 168 267.12 168 0 0 Puma Herbicide 0 - 1 lt/ha 1 300 300 300 0 0 Strane M. Herbicide 0 - 1 lt/ha 1 105 105 105 0 0 2-4, D Herbicide 47 0 1 47 47 1 47 47

Human labor (labor day) 58 - 66 (labor day) 66 6 396 58 6 348 58 6 348Animal labor (animal days) 60-72 animal days 70 5.6 392 60 5.6 336 60 5.6 336TOTAL COST 1923.62 1194.5 1144.8

Net return/ha (no land cost) 2760.63 2172.15 2493.5Return to labor (land cost) 47.82773 43.45086 48.99138

22

In 1998/99, wheat production was seriously affected by rust.13 Even new varietiesincluding HAR 1685 succumbed to rust. Researchers at Kulumsa estimated that the yieldof HAR 1685 declined by about 41% and older varieties such as Dashen by over 60%.The partial budget analysis was thus recomputed using adjusted yield figures and actualprices of input, labor and animal traction costs of 1998/99. The result showed that returnto land declined by 73.3% (only 737.56 birr per ha compared to 2,760.63 birr) underrecommended management. The return to land declined by 69.3% (667,5 birr per ha,compared to 2,172.15 birr per ha in 1994/95) under farmers’ own management. Forfarmers who used the old variety (Dashen), the return to land was a mere 210.0 birr perha (down 91.6%) (Annex III).

Many wheat producers in Arssi and Bale (the major wheat zones of the country) wereunable to repay their input loans in 1998/99, according to extension agents in the area.Significant losses were incurred in situations where land was rented, a common practicein these areas. The losses of those renting land and planting HAR 1685 were estimated at162.4 birr per ha under recommended management and 232.5 birr under farmers’ ownmanagement. The loss in the case of Dashen was the highest: 690 birr per ha. The returnto labor after deducting land cost became negative on the control plots.

A major alternative to growing wheat in the highlands of Arsi and Bale is barley. In thehigher altitudes, barley accounts for about 40% of the cultivated area, compared to 33%for wheat. The share of barley declines to 24% while wheat increases to 45% in thelower altitudes (Chilot, et al., 1992). But barley is largely grown as food crop and it isless profitable than wheat mainly because of its lower yields. The Ethiopian SeedEnterprises produces only a small amount of improved varieties of barley due lowerdemand and inferior yields relative to wheat (Annex IV).

3.1.3 Determinants of yield

Wheat researchers and farmers contacted around Kulumsa Research Center indicatedthat HAR 1685 performed better than many other varieties currently under production.Nonetheless, the available evidence suggest that the performance of the variety was neverclose to its full potential (50 – 70 qt/ ha) under farmers’ conditions. Two of the majorreasons are discussed below.

(a) Technological factors

HAR 1685 is resistant to lodging (semi-dwarf variety) and has a high tillering capacity.But it was only moderately resistant to stripe, leaf and stem rust even when it was firstreleased. It poorly competes with weeds and it requires early weed control. Farmersaround Kulumsa indicated that HAR 1685 is susceptible to shattering. Partly owing tothese technological limitations, the variety gave lower yield, especially under farmers’management, than older varieties at the on-farm verification trial of 1994/95 (Annex V).

13 Untimely rains after maturity created favorable weather for rust.

23

Ethiopia is a hot spot for rust diseases. The pathogens tend to mutate, making it difficultto achieve durable resistance to rust. It has become frustrating to researchers to see thatnew varieties that took several years to develop are attacked immediately after they arereleased to farmers. The disease poses an important threat to stable wheat production inthe country as research has yet to develop varieties with durable resistance to rust disease.Compounding the problem is the fact that it requires several years of trials to obtainapproval for release from the Variety Release Committee. A variety could loose much ofits resistance quality by the time the whole process of regional, pre-national, national andon-farm verification trial is completed and the material is recommended for release.Wheat researchers have suggested that the process need to be shorter and lessbureaucratic to overcome the threat posed by rust.14

(b) Management factors

Crop yield depends on a number of biotic and abiotic factors, all of which are dynamic inresponse to human interventions (Tanner, et al., 1998). Crop rotation, tillage practices,rate and method of fertilizer application, weed control and planting methods are amongthe most important management factors influencing wheat yield.

Weeds are among the major constraints to wheat production. In wheat dominated areas,the continuous production of wheat (rotation with barley) coupled with repeatedapplication of broad-leaf herbicides has increased the densities problematic grass weeds(Amanuel, et al., 1996). A review of yield loss assessment trials conducted on wheat inEthiopia indicated a mean yield reduction of 36.4% due to weed competition (Rezene,1985).

Grass weeds are difficult to control by hand weeding in a broadcast wheat crop as grassweeds are not easily distinguished from the crop at an early stage. Chemical control ofgrass weeds, on the other hand, is limited in Ethiopia because of unavailability and highcost of such herbicides. A better alternative could be row sowing of wheat in order tofacilitate hand or mechanical weeding. Field tests of animal-drawn four-row seeder(recently developed by the Nazareth Research Center of ERO) have shown that thetechnology is labor efficient, requiring less than 25% of the time required under farmers’conventional practice. It also facilitates placement of seed and fertilizer together in therow, enabling the crop to utilize fertilizer more effectively and limiting the nutrientavailable to weeds emerging between crop rows. The use of the row seeder increasedwheat grain yield by 28% in comparison with conventional broadcasting (Asefa, 1997).

Agronomic studies have examined the beneficial impact of break crops on wheatproduction. For instance, one study showed that a faba bean break crop increased wheatgrain yield by 1100 kg per ha or 69%, compared to the yield of second year continuouswheat (Hailu, et al. 1989).

14 Personal communication with wheat researchers of CIMMYT working in Ethiopia.

24

Cultural practices such as stubble management, tillage practices and cropping sequenceplay an important role in the growth and development of crops. According to Asefa, etal. (1996):

"... The retention of straw on the soil surface can be important for nutrient recycling.Furthermore, the presence of straw on or near the soil surface can reduce rain run-off and soil erosion, can enhance soil moisture levels, thereby indirectly enhancingthe activity of microorganisms involved in decomposing OM [organic matter] andchanging unavailable forms of nutrients to available forms. Minimum tillage canreduce soil compaction with the use of appropriate implements. Incorporating dicotcrops, especially legumes, in the production system can increase the level of availablenutrients to succeeding cereal crops."

Farmers in Ethiopia know the benefit of rotating cereals such as wheat with grainlegumes and oilseed crops. Crop rotation and fallowing were indeed the major means ofrestoring fertility before the introduction of fertilizer in the cereal producing areas. Croprotation helps in fixing nitrogen to the soil and breaking weed, disease and insect cycles.However, farming system surveys have shown that the area under cereals is expanding atthe expense of pulses and oilseeds (Berhanu, 1985;Chilot et al., 1992). Populationpressure and scarcity of land have meant that a higher proportion of the available land isallocated to cereals in order to meet subsistence requirements. Low yield due to rampantdiseases and lack of improved technologies has also discouraged the cultivation of fababean as part of the crop rotation system (Chilot et al., 1992).

Farmers realize that depleted soil fertility is a critical bottleneck in agriculturalproduction. Similarly, researchers have indicated that nitrogen deficiency is among themajor factors for low yield levels in wheat. Increased and split application of nitrogenwas found to have a significant impact on yield and economic benefit wheat productionin waterlogged areas (Tilahun, et al., 1996). But optimum fertilizer application isconstrained by high cost and unavailability of fertilizer (see section 4) as well as limitedawareness about soil nutrients on the part of farmers. Planting time is often delayed dueto late distribution of fertilizer.

Wheat straw is an important animal feed and it is rarely left on the farm. It is also widelyused for roof cover (thatched roof) in rural areas. Farmers realize the advantagesassociated with retention of straw on the field but they have critical forage problems. Soilmining is thus a critical bottleneck in increasing wheat yield. The problem is againrelated to the increasing population pressure.

Farmers expressed favorable views about row seeder. They observed that wheat in therow-seeded plots was superior to the broadcast crop (Asefa, et al., 1997). But farmershave no access to the technology since no manufacturing firm has shown interest incommercial production of the implement.

25

3. 2 Maize

3.2.1 Research

Maize research in Ethiopia was initiated at Jimma Junior Agricultural College in 1952and at Alemaya Agricultural College in the mid 1950s. The Bako Research Center, amajor center for maize research and the center for National Maize Coordination, wasestablished in 1966 under the Institute of Agricultural Research, IAR (recently renamedas the Ethiopian Research Organization, ERO) which was also set up in the same year.The National Coordinator is responsible for coordinating germplasm development andvarious trails at cooperating stations such as Nazareth, Alemaya, Awassa, Abobo andPawe. These stations/centers are expected to cover the four main agroecologicalenvironments of the country: Mid Altitude Sub-Humid Zone (lies between 1500 and1800 masl and has annual precipitation of 1000 – 1250 mm), Mid Altitude MoistureStress Zone (1000 – 1500 masl and less than 800 mm), High Altitude Sub-Humid Zone(1800 – 2400 masl and about 1500 mm of rainfall), and Low Altitude Sub-Humid Zone(below 1000 masl and about 1250 mm) (Kebede, et al., 1993).

All new research projects on maize are reviewed nationally and submitted to IAR/EROmanagement for funding, mainly from the government, through the National Coordinator.Since the mid 1980s, maize research at each research station has followed a teamapproach composed of breeders, agronomists/ physiologists, weed scientists, soilscientists, entomologists, pathologists, nutritionists, agricultural economists, farmingsystem researchers, and extension workers (Kebede, et al, 1993). The process ofreleasing a variety follows the same procedure as that of wheat.

The focus of research in the earlier years was on screening of introduced maizegermplasm for adaptation and resistance to diseases. The major constraints of maizeproduction were identified as low yield potential of local cultivars and attack by insectsand diseases. Stalk borers and common leaf diseases such as rust and blight wereconsidered as one of the major factors limiting maize production in Ethiopia. Varietiesthat were introduced from abroad and adapted to local conditions were thus selected andreleased as open pollinated improved seeds. Breeding work and systematic evaluation ofselected lines in hybrid combination was started in the 1980s. The first hybrid fromEthiopian breeding program was BH-140, which was released in 1988 from the BakoResearch Center (Kebede, et al, 1993).

BH 660 is one of the most successful hybrid varieties released in 1994 by the BakoResearch Center. It has a wider adaptability, growing at altitudes ranging from 1,650 to2200 masl with annul precipitation of 1000 to 1500 mm. It needs up to 170 days formaturing (a long maturing hybrid) and performs better under high rainfall, good soilconditions and high dose of fertilizer. Insufficient rains and poor soil fertility could resultin abortion. 15 The variety shows good resistance to important maize diseases. The yieldpotential of BH 660 is well over 100 qt/ha (Ethiopian Seed Enterprise, 1997). 15 Personal communication with Mr. Legesse Wolde, the National Maize Research Coordinator.

26

3.2.2 Profitability of BH – 660

The data for this and the next section (maize and tef technologies) came from the 1997/98survey, which looked at the performance of the new extension approach. 16 One of thesites for the study was Jimma, high rainfall area (1400 to 2000 mm of rain per annum),considered as ideal location for growing BH-660. Three different types of plots wereconsidered for the study: (i) BOA/SG plots which refer to plots planted using BH – 660and all complementary input and the farm work is closely supervised by extension agentsfrom the Bureaus of Agriculture which receives technical and financial assistance fromSG 2000; (ii) Traditional plots are managed using conventional practices and local seedsand are owned by the same group of farmers owning BOA/SG plots; (iii) Graduate plotsrefer to plots owned by farmers who previously participated in the BOA/SG program fortwo years and graduated to work on their own. However, all the ‘graduate’ farmers havedecided to remain under the program in order to gain access to the technology17.

Average yield of BH-660 under recommended management (BOA/SG plots) was veryhigh, 5.3 tons (53 qt.) per ha on the average (Table 4). It varied from 4.0 ton (40 qt.) perha for the lowest tercile to 6.6 ton (66 qt.) per ha for the highest tercile (Annex VI).Average yield was even higher, 6.5 ton (65 qt.)/ha on the average, for Graduate plots,indicating a ‘learning effect’18. By contrast, maize yield was only 2.7 ton (27 qt.) per haon plots planted with local seeds and managed using conventional practices (Traditionalplots).

The return (assuming no cost of land and guard labor) was 1,571.3 birr per ha forBOA/SG plots and 2,054.0 birr per ha for Graduate plots. The return to labor was 15.1and 18.2 birr per manday for BOA/SG and Graduate plots respectively. On the otherhand, the return to land for Traditional plots was less than half (707.8 birr per ha) of thatobtained on BOA/SG plots or Graduate plots. The marginal return of Graduate plots(relative to Traditional plots) was 2,077 Birr per ha, as opposed to the marginal cost of730.8 birr per ha. Hence, the maize technology is highly profitable with 184% marginalrate of return, a very attractive return by all standards.19

Cash rent of land is not a common practice in Jimma. But for the few who can afford, themaximum rate is 200 birr per ha. The return to land and labor declined in all cases andwas negative for the lowest tercile of the Traditional plots when land cost is taken intoaccount.

16 See Howrad J., Mulat Demeke, V. Kelly, M. Maredia, J. Stepanik. 1998. Can the Momentum besustained? An Economic Analysis of the Ministry of Agriculture/ Sasakawa Global 2000’s Experimentwith Improved Cereals Technology in Ethiopia, Department of Agricultural Economics, Michigan StateUniversity, E. Lansing, Michigan17 Non-program participants do not have access to improved seeds.18 Farmers have become more proficient in the application of improved technology and managementpractices over time.19 The marginal rate of return of BOA/SG plots (relative to Traditional) was ..

27

The most widely used arrangement for gaining access to land by farmers with little or noland of their own is sharecropping. Land owners (often female-headed households andold heads) who share input costs with their tenants and assist in farm work are entitled to50% of the harvest, while those not sharing input costs receive only one-third of the totaloutput. An important consideration under the sharecropping arrangement is the sharingof risk emanating from attack of maize fields by wild animals such as monkeys, wildpigs, wild hogs and porcupine.

Guarding against wild animals is by far the most difficult task in growing maize in areassuch as Jimma. It starts with the day of planting and ends with transporting the harvest tostore, lasting for some 165 days on the average. Children or women handle daytimeprotection (mainly from monkeys), while guarding at night is the sole task of adult males.Owners of adjacent plots form a circle around their common fields to guard against attackfrom all directions. It is impossible for a farmer to protect isolated field. The size of landor the number of plots cultivated by a given household depends largely on the number ofadult males available for nighttime guarding. Female-headed households with no adultlabor within the family hire labor from the neighborhood (often a farmer with adjacentplot of his own) for night-time guarding and the cost is about 4 quintals of unshelledmaize per plot (which is equivalent to 3 quintals shelled maize or birr 162 birr per plot atJan. 1998 price).20 Assuming a similar cost for the daytime protection, the overall cost ofguarding could be estimated at 324 birr. This reduced the return to land of BOA/SG,Traditional and Graduate plots by 17.6, 39.0 and 13.8%, respectively (not shown in theTable). The return to labor fell below the market wage rate for traditional plots when thelabor for guarding (assuming 165 mandays) is included as part of the labor required toproduce maize. The return to labor was marginally higher than the market wage rate inthe case of BOA/SG and Graduate plots.

20 Mengisty Buta. 1998. Field Report - Jimma. Addis Ababa, Ethiopia

28

Table 4: Partial Budget for Maize (Jimma) (1997/98 cropping season)

Budget Item MOA/SG Traditional GraduatePlots Plots Plots

n used in calculations 69 47 391. GRAIN YIELD (kg/ha) 1/ 5508 2814 67811.A. Adjusted grain yield (kg/ha) 1/ 5293 2704 65162. FARMGATE PRICE (birr/kg) 2/ 0.54 0.54 0.543. GROSS REVENUE 3/ 2883 1473 35504. TOTAL PACKAGE COSTS 4/ 640 279 6064.A. Treated seed (birr/ha) 129 40 1224.B. DAP 263 239 2494.C. Urea 248 0 2354.D. Herbicide 0.4 0 04.E. Insecticide 0 0 0.44.F. Fungicide 0 0 05. INTEREST 5/ 0 2.7 38.36. LABOR 6/6.A. Total family/mutual labor days 7/ 135 92 140 Total cost of family/mutual labor (labor days*3.5) 472.5 322 4906.B. Total wage labor (birr/ha) 8/ 62 36 717. ANIMAL TRACTION COST (birr/ha) 9/ 98 112 2138. HAND TOOLS AND SACKS (birr/ha) 39.2 13.5 77.78.A. Hand tools (birr) 10/ 2.8 2.9 5.58.B. Sacks (birr) 11/ 36.3 10.5 72.29. TOTAL COST 1311.7 765.2 149610. NET INCOME/HA 12/ (no land cost) 1571.3 707.8 205411. NET INCOME/FAMILY-MUTAL LABOR DAY 13/ 15.13926 11.19348 18.1714312. Net income/ha (less land cost) 1371.3 507.8 185413. Net income/family-mutual labor 13.65778 9.019565 16.7428614. Net income/ family-mutual labor (incl. Guard lab.) 6.146 3.228794 7.685246

1/Yield estimates based on crop cuts taken as part of the GMRP/MSU/MOA/SG2000 Survey. Assumes (a) no grain waslost during threshing; (b) maize was harvested in November; (c) storage losses were 1.98% per month, the average ofvarious estimates from Abraham et al. 1993.2/Source: EGTE price monitoring unit and GMRP/MSU/MOA/SG2000 Survey. Local market prices collected by EGTEwere adjusted to farmgate prices using survey data on prices reported by farmers. Average prices for maize during January1998 are used.3/ (1)* (2)

4/ 4.A. + 4.B. + 4.C. + 4.D. + 4.E. + 4.F. MOA/SG maize package consists of (quantities/ha) 25 kg seed, 100 kg DAP, 100kg urea.5/Source: GMRP/MSU/MOA/SG2000 Survey and rate information from MOA/SG2000. MOA/SG2000 programparticipants pay no interest; NEP program participants pay 10% interest annually. Assumes that period of loan is 10months.6/Source: GMRP/MSU/MOA/SG2000 Survey.

7/Includes shelling labor.

8/Valued at cash/in-kind payment rates provided by survey participants.

9/Sum of (a) rental costs reported by survey respondents and (b) for owned/borrowed oxen, maintenance + depreciatedvalue of animals and animal traction equipment multiplied by percentage of total farm represented by the program,traditional or graduate plot. Purchase price, life and salvage value of equipment based on supervisors’ field reports. Thismethod of estimating traction cost is different from the method used in Table 3. The former tends to underestimate tractioncosts.10/Depreciated value of 2 hoes, 2 axes, and 2 cutting knives. Purchase price, life and salvage value of equipment based onsupervisors’ field reports.

29

11/Depreciated value of sacks needed to transport maize marketed in 1998 season. Since sacks are retained by farmer andused for other purposes, cost is apportioned by multiplying depreciated sack value by percentage of total farm representedby program, traditional or graduate plot. Purchase price, life and salvage value based on supervisors’ field reports.12/(3) - (4 + 5 + 6B + 7 + 8)

12/(9)/(6.C.)

Source: Howard J., Mulat Demeke, V. Kelly, M. Maredia, J. Stepanik. 1998. Can the Momentum besustained? An Economic Analysis of the Ministry of Agriculture/ Sasakawa Global 2000’s Experimentwith Improved Cereals Technology in Ethiopia, Department of Agricultural Economics, Michigan StateUniversity, E. Lansing, Michigan

3.2.3 Constraints to a wider adoption of BH – 660

Investment in BH – 660 is undoubtedly profitable in areas with sufficient rainfall.BOA/SG participants in Wolliso (West Shoa) were able to get a yield of 5.3 ton (53 qt.)per ha in 1997/98 (Howard et al. 1998). The average yield of the variety was 6.1 tons (61qt.) /ha (1994/95) under on-farm verification trails of Bako Research Center.21 Fieldobservation in Jimma, West Shoa and Bako (East Wellega) showed that there is a hugeinterest in the technology. However, there are a number of impediments to expanded useof BH-660. First, shortage of seeds is undoubtedly the major constraint in growing BH –660. The total quantity of BH – 660 available for distribution for the 1997/98 season wasonly 5,894 quintals (Annex IV). This potentially covered an area of 18,646 to 23,576 haonly, given the recommended seed rate of 25 to 30 kg per hectare. Moreover, a goodproportion of the amount supplied was distributed to areas with less than the optimalamount of rainfall needed for growing the variety. Neither the supplier (the EthiopianSeed Enterprise) nor the regulatory agency (the National Seed Industry Agency) haseffective control over areas that ultimately receive the seeds through the bureaus ofagriculture.22 The available evidence suggest that the amount of seed that was effectivelyutilized was quite lower than the amount supplied.

Second, the maize package costs over 600 birr per ha (or 300 birr per 0.5 ha). Given thelow level of resources owned or income obtained,23 the cost is too much for many of thepoorer households. Indeed, the survey by Howard et al. (1998) showed that participantsin the extension program have a much better resource base than the average farmer. Astudy in Bako (one of the major maize growing areas) also indicated that SG2000participants have larger farm size and a greater number of oxen and farm labor than non-participants (Beyene, 1998).

Third, weather problems pose a major threat to the adoption of BH-660. Sufficient rain isone of the requirements for good results. The risk of the variety can be significant inareas where the rainfall is less reliable. Poor rains and hail in 1997 created serious

21 Personal communication with socio-economic researchers at Bako Researc Centre.22 The report of the Committee established by ESIA to investigate the complaints regarding germinationproblems of BH – 660 seeds during the 1998/99 cropping season showed that many of the zones thereported planting the seeds do not have the recommended rainfall of 1000 to 1500 ml. Among the zonesthat have a lower amount of rain and yet reported planting BH-660 were East Shoa, East Hararghe, Arsi,Borena and West Shoa in Oromiya region and Guraghe, Sidama, North Omo and Hadiya in SNNP region.23 As indicated above about one-third of the households farm less than 0.5 ha and about 42% of the ruralhouseholds earn less than 8 USD per month.

30

hardship for farmers in West Shoa (Wolliso). Some farmers were forced to sell theiroxen to take care of input loans. Risk-aversion compelled many farmers to dis-adopt thetechnology in 1998 (Howard, et al., 1998). In particular, the risk faced by resource-poorfarmers (with no asset or livestock) could be too high or unaffordable. There is no formalor informal institution to help legitimate loan-defaulter due to crop failure.

Fourth, low soil fertility is another environmental problem constraining the disseminationof high yielding varieties such as BH-660. Land degradation is expanding at an alarmingrate in Ethiopia owing to continuous cultivation with inadequate fertilization and farmingpractices that pay limited attention to conservation. Lack of tenure security and highprice of fertilizer have contributed to the worsening problem of low soil fertility.

Fifth, wild animals pose a major threat to adoption of the new maize technology in themore humid and high rainfall zones (e.g. Jimma, Illubabor and West Wellega). Densevegetation and patchy forest areas provide shelter to wild animals in these areas.Discussion with farmers revealed that ever since the Regulation on Wildlife Protectionwas strictly enforced (under the former government), the number of wild animalsattacking maize crops has increased enormously. The lapse of protection even for asingle day could result in a complete damage of the whole maize field. Households withno male labor of their own are less likely to adopt the new technology. Labor shortage isalso a critical constraint to all those who want to cultivate more land or exapnd.

Sixth, a number of insect pests and diseases attack maize in the field. Over 40 species ofinsects have been recorded on maize in the field, and of these, maize stalk borer is themost important. Termites are also prevalent in most western parts of the country. Leafblight and head smut are among the major diseases (Abraham, et al., 1993; Assefa andTewabech, 1993). Grey Leaf Spot (GLS) have also become important in the last twoyear. Most poor farmers are likely to find the cost of the chemicals required to controlmaize insects and diseases unaffordable.

Finally, storage and marketing problems are also among the most important bottlenecksof maize cultivation. The maize weevil is the most destructive and most common pest ofstored maize. The extent of damage due to weevils and other pests varies depending onmoisture and temperature levels.24 Farmers in Jimma reported losses of up to 100% injust five months. It was also reported that the new variety (BH – 660) is more easilyattacked than local varieties.25

The use of storage insecticide is very limited in Ethiopia. For instance, none of the surveyfarmers in West Shoa and Jimma used storage insecticide in 1997/98 (Howard, et al.,1998).

24 Losses in grain weight corrected for moisture changes ranged from 2.07 to 13.84% dueto weevils and from 0.15 to 17.87% due to grain moth larvae in the different sites visitedin the Bako area (Abrham, 1993).25 Mengistu Buta. 1998. Field Report – Jimma area. Addis Ababa, Ethiopia; Demeke Abate. 1998. FieldReport- West Shoa. Addis Ababa, Ethiopia

31

3.3 Tef

3.3.1 Research

Ethiopia is the origin and center of diversity for the tef plant. Tef is primarily grown toprepare ‘enjera’ (Ethiopian bread). It is the most widely adapted crop compared to anyother cereal or pulse crop in the country. It can be grown under various rainfall,temperature and soil regimes. Tef can withstand low moisture conditions better thanmaize and sorghum in lower altitude and drier areas. Farmers plant tef whenever theirmaize or sorghum crop fails due to low moisture. Tef can also withstand waterloggedconditions better than maize, wheat or sorghum. Apart from its high value in the marketplace, tef grain is not attacked by weevils or any other storage insect. Teff straw is alsothe most valuable animal feed particularly during the dry season (Seifu, 1993; Hailu, etal., 1995).

Scientific research to improve tef started in the late 1950s at Jimma Junior AgriculturalCollege. Until 1974, tef research focussed on mass selection of superior genotypes fromlandraces. Superior varieties were selected and recommended for production. Since thecultivation of tef has so far not been threatened seriously by plant pathogens or insectpests, most of these varieties26 still show good adaptation in many areas. Once thecrossing technique became known in 1974, tef research focussed on developing varietiesthrough trait recombination or crossbreeding. Some six varieties have been developedand released for adoption using this technique 27.

Lodging is identified as one of the major constraints to increase the yield of tef. The stemof tef plant has less fiber than wheat or barley. Heavy rains and strong wind before orafter heading can easily cause lodging in teff. Because the plant is susceptible tolodging, sufficient amount of nitrogen fertilizer cannot be applied to attain maximumyield increase. Lodging causes damage to the vegetative part of the plant due to rottingand fast spread of pests and diseases. The overall loss in grain yield due to lodging isestimated to be within the range of 11-22%, averaging 17% (Seifu, 1993).28 Lodgingmakes mechanized harvesting difficult and reduces the quality of the straw as fodder. Theattempt to develop stiff stem to hold the panicle upright has not been successful so far.

One of the most popular varieties released so far is DZ-01-196, locally known as ‘magna’or ‘qoledima’. It is widely adapted to different agroecological areas, growing at altitudesranging from 1800 to 2400 masl and with annual rainfall of 300 to 700 mm. It was firstrecommended for release around Debre Zeit in 1970. The average yield recorded in thenational yield trail was 21.5 qt/ha (varying between 18 and 25 qt/ ha) (Hailu, et al., 1995).Lodging on ‘magna’ fields can range from 90 to 94% (ESE, 1997). Because of its verywhite seed color and fairly large kernels that attract highest price on the market, there ishigh demand for ‘magna’ seeds in many tef growing areas of the country.

26 Some of old varieties that are still popular include DZ-01-354, DZ-01-196 AND DZ-01-99.27 These varieites include DZ-Cross-82, DZ-Cross-44, DZ-Cross-37, DZ-Cross-255, DZ-Cross-358, andDZ-Cross-112.28 Estimates of yield loss due to lodging are higher, 20-25%, according to the tef breeder Dr. Hailu Tefera.

32

3.3.2 Profitability of DZ-01-196

East Shoa has the one of the most favorable climatic and soil conditions for growing tef.Tef grown around Debre Zeit (Ada district, East Shoa) is rated as the best quality andfetches the highest price in the country. The data for the profitability assessment wascollected from this district in 1997/98.

Average tef yields were similar on all the three types of plot, about 1.4 tons (14 qt.) perha on BOA/SG and Traditional plots and 1.5 ton per ha on Graduate plots (Table 5)29.The latter used mainly their own seeds which they saved from previous harvest andapplied lower fertilizer rates (95 kg DAP and 59 kg urea per ha, in comparison with 104kg DAP and 104 kg urea applied on BOA/SG plots) (Howard, et al., 1998).

The lack of significance across plot types is due in large part to the use of similar types ofseeds and fertilizer. The Ethiopian Seed Enterprise has never been able to maintain thequality and purity of improved seeds of tef such as DZ-196 in its seed multiplicationprogram. Tef seeds are too small and are easily carried by water, wind, animals, humansand farm equipment from one field to another. Pure seeds get mixed with differentvarieties during seed multiplication and lose their quality (seed color) and yield potentialin just one or two years. The extension package for tef is therefore flexible allowingparticipant farmers to plant their own local seeds if they so desire and choose their ownlevel of fertilization. 30 According to tef researchers at Debre Zeit 31, it is not surprisingthat no significant yield difference was found between the local seeds of farmers (savedfrom previous harvest) and ‘improved’ seeds produced by the Ethiopian Seed Enterpriseand distributed through the extension system. The seeds are likely to be mixed in bothcases and have lost their original yield potential. As indicated above, uncontaminatedDZ-196 has the capacity to yield over 2 tons per hectare.

Tef production was found profitable on all the plots. But the average return per ha onTraditional and Graduate plots was about 14% higher (1,805 birr per ha) than BOA/SGplots (1,589 birr per ha) (Table 5). The tef package which includes ‘improved seeds’(for 31 of the 60 participants in the BOA/SG program) and the application of 104 kgDAP and 104 kg urea per ha gave a lower return than the conventional practice(Traditional and Graduate plots) of using seeds saved from previous harvest andapplication of roughly 100 kg DAP and 50 kg urea. The marginal rate of return onBOA/SG plots (relative to Traditional plots) was –78%. In other words, programparticipants incurred a marginal cost of birr 285.6 (mainly on fertilizer) and earned areturn of only 69 birr per ha. The yield response to the extra fertilizer applied (worth 272birr) was insignificant. Econometric analysis also showed that the plots that used farmerretained seeds obtained about 300 kg per ha more yields than their counterparts usingpurchased ‘improved’ seeds. The analysis also showed that both DAP and urea were notsignificant variables in explaining yield variations (Howard, et al., 1998).

29 See also Annex VII for details.30 Only 31 of the 60 BOA/SG plots were planted with ‘improved’ seeds bought from ESE or extension.Nine of the 60 Traditional plots and 3 of the Graduate plots were planted with purchased ‘improved’ seeds.31 Personal communication with Dr. Harilu Tefera , tef breeder.

33

The return per labor day (family and mutual labor) ranged between 28 and 37 birr,significantly higher than the market wage rate (6 birr per day). The Traditional plots thatrequired less labor yielded the highest return to labor.

The alternative to cultivation of tef is wheat in the study area. However, the yield ofwheat is generally lower32 mainly because the crop is less resistant (relative to tef) underthe water-logged conditions of the district. The widespread problem of rust disease alsomakes wheat a less preferred crop.

It is a common practice for tef growers around Debre Zeit to rent land (for cash).33 Thereturn to land declined by 55 to 63% when land rent is added to cost of production.Similarly, net income per family-mutual labor declined by 46 to 52%, 14 to 19 birr perday. This may not encourage landless farmers to grow tef on rented land.

32 For instance, Abate (1995) showed that the average yield of improved varieties of durum wheat (Booai)was less than less 2 tons per ha for the period 1991-93.

33 The rate for one season varies between 1000 bir per ha (for land planted cereals in the previous season)and 1400 birr per ha (planted legumes in the previous season). Since the former is more common than thelater, the cost of land was estimated as 1000 birr per ha.

34

Table 5: Partial Budget for Tef – Ada (Debre Zeit) ( East Shoa) (1997/98 croppingseason

BOA/SG Traditional Graduate

Budget Item Plots plots Plots

n used in calculations 60 60 601. YIELD 1/1.A. Grain Yield (kg/ha) 1389 1364 14551.B. Straw Yield (kg.ha) 2180 2025 20712. PRICE 2/2.A. Estimated farmgate price for grain (birr/kg) 2.04 2.04 2.042.B. Estimated farmgate price for straw (birr/kg) 0.11 0.11 0.113. GROSS REVENUE 3/ 3076 3007 31984. TOTAL PACKAGE COSTS 4/ 655 540 5714.A. Treated seed (birr/ha) 150 167 1904.B. DAP 251 214 2274.C. Urea 226 141 1294.D. Herbicide 28 18 254.E. Insecticide 0 0 04.F. Fungicide 0 0 05. INTEREST 5/ 28 20 19.76. LABOR 6/6.A. Total family/mutual labor days 63 57 78 Total cost of family/mutual labor (labor days*5) 315 285 3906.B. Total wage labor (birr/ha) 7/ 192 142 1837. ANIMAL TRACTION COST (birr/ha) 8/ 291 210 2248. HAND TOOLS AND SACKS (birr/ha) 6.5 4.9 5.78.A. hand tools (birr) 9/ 2.2 1.6 1.68.B. Sacks (birr) 10/ 4.3 3.3 4.19. TOTAL COST 1487.5 1201.9 1393.49. NET INCOME/HA 11/ (no land cost) 1588.5 1805.1 1804.610. NET INCOME/FAMILY-MUTUAL LABOR DAY 12/ 30.2 36.7 28.111. Net income/ha (less land rent) 588.5 805.1 804.612. Net income/family-mutual labor day 14.3 19.1 15.3

1/Yield estimates based on crop cuts taken as part of the GMRP/MSU/MOA/SG2000 Survey. Assumes no grain or straw was lostduring threshing.2/Source: EGTE price monitoring unit and GMRP/MSU/MOA/SG2000 Survey. Local market prices collected by EGTE wereadjusted to farmgate prices using survey data on prices reported by farmers. Average prices for white tef during January 1998 areused.3/(1.A*2.A) +(1.B. *2.B)

4/ 4.A + 4.B. + 4.C. + 4.D. + 4.E. + 4.F. MOA tef package consists of (quantities/ha) 35 kg seed, 100 kg DAP, 100 kg urea, U-46herbicide.5/Source: GMRP/MSU/MOA/SG2000 Survey and rate information from MOA/SG2000. MOA program participants pay 10%interest annually. Assumes that period of loan is 7 months.6/Source: GMRP/MSU/MOA/SG2000 Survey.

7/Valued at cash/in-kind payment rates provided by survey participants.

8/Sum of (a) rental costs reported by survey respondents and (b) for owned/borrowed oxen, maintenance + depreciated value ofanimals and animal traction equipment multiplied by percentage of total farm represented by the program, traditional or graduateplot. Purchase price, life and salvage value of equipment based on supervisors’ field reports.9/Depreciated value of 2 sickles, 2 hoes, and 2 spades. Purchase price, life and salvage value of equipment based on supervisors’field reports.

35

10/Depreciated value of sacks needed to transport tef marketed in 1998 season. Since sacks are retained by farmer and used forother purposes, cost is apportioned by multiplying depreciated sack value by percentage of total farm represented by program,traditional or graduate plot. Purchase price, life and salvage value based on supervisors’ field reports.11/(3) = (4 + 5 + 6B + 7 + 8)

12/(9)/(6.C.)

Source: Howrad J., Mulat Demeke, V. Kelly, M. Maredia, J. Stepanik. 1998. Can the Momentum besustained? An Economic Analysis of the Ministry of Agriculture/ Sasakawa Global 2000’s Experimentwith Improved Cereals Technology in Ethiopia, Department of Agricultural Economics, Michigan StateUniversity, E. Lansing, Michigan

3.3.3 Constraints to increasing tef yield

Unlike other crops, the cultivation of tef has not been threatened seriously by plantpathogens or insect pests so far (Hailu et al., 1995). Owing to high demand in the urbanareas, tef producers enjoy high and stable prices. However, tef yield has remained verylow (less than 1 ton per ha at national level. Several factors have contributed to the lowyield. But the most important bottleneck is lack of good quality improved seeds. Tefvarieties run out quickly due to mechanical contamination. A single plant of teff with itstillers produces a colossal number of seeds (up to 50,000), hence hastening thedeterioration of pure varieties from generation to generation even if a s single stray seedis mixed (Hailu, et al., 1995). Researchers and seed technologists have yet to develop asystem of production, which overcomes the problem of contamination.

Lodging is identified as the other major constraint to increasing tef yield. It causesdamage to the vegetative part of the plant due to rotting and fast spread of pests anddiseases. It makes hand harvesting more time consuming and tiresome and limits the useof mechanical harvester. The overall loss in grain yield due to lodging is estimated at17%. It also lowers 1000 seed weight by 35% and rate of germination by 44% (Seifu,1993). The quality of tef straw declines sharply due to lodging. Tef producers cannotapply adequate amount of nitrogen fertilizers (to increase yield) since vigorous vegetativegrowth enhances lodging. To date the attempt to develop lodge-resistant varieties has notbeen successful.

Imported germplasms (resulting from intensive breeding work) are the major sources fordeveloping improved seeds of maize, wheat and all other major grains. Tef breeders haveno such opportunity. They rely entirely on land races that are inherently low yielding. Inspite of its dominance in the total cultivated area, tef research attracts limited attentionfrom the government and the national research system. There were only 2 full-timeresearchers (one Ph. D and one Masters) working on tef in 1998/99. With little or nointernational interest, few individuals are prepared to work on tef research (for lack oftraining and visits abroad).

Poor cultural practices have also contributed to low yield of tef. Owing to shortage ofland, traditional cultural practices such as fallowing, crop rotation and retention of strawon the soil surface are abandoned in favor of intensive monocropping. Similar to wheatand other crops, the soil is losing both its mineral and organic matter content. Sustainableproduction is threatened by the existing management practices that have exacerbated the

36

problems of nutrient mining and weed infestation. High cost and delays in distributionhave constrained the application of balanced and optimal amount of chemical fertilizers.

3.4 Crossbred cows for milk and traction

3.4.1 Research

Because of increasing population and livestock pressure on the land, farmers in Ethiopiamay not be able to continue maintaining draft oxen specifically for work purposes.Many smallhoders are increasingly finding it difficult to keep cattle so as to ensure thatthere is always a pair of working oxen. Although oxen are currently used for less than113 days in a year, they have to be fed for the remaining 252 days as well (Alemu et. Al,1998). Oxen are also given priority over all other animals as much of the feed availableis reserved for them.

The use of dairy cows for traction implies that there is no need for keeping a separateherd of draught animals. This could benefit total farm output and incomes throughincreased milk production, while making it unnecessary to feed oxen year-round andmaintain a follower her to supply replacement oxen. In addition, using dairy cows fortraction would allow males to be fattened and sold younger and could also lead to greatersecurity of replacement. Dual-purpose dairy-draught animals could result in a reductionof stocking rates and overgrazing, thus contributing to the establishment of a moreproductive and sustainable farming system (Zerbini, et al., 1998).

The International Livestock Research Institute (ILRI) and the Ethiopian ResearchOrganization (ERO) have been researching on various aspects of using crossbred cowsfor traction since 1978. In order to demonstrate that the technology is working andbeneficial, various experiments were conducted with F1 crossbred cows (Holstein-Friesian X Boran (local breed)) to determine the technical trade-offs between milkproduction, calving interval and work output when cows are used as draught animals.On-station results indicated that milk production and reproduction was not significantlyaffected by work. Total milk yield between non-working and working cows was 2224.7kg and 1985.3 kg, respectively (Shapiro, et al., 1994). The work output of dairy cowswas equivalent or above that required by farmers for land cultivation (Zerbini, et al.,1998). On-farm research at Holleta showed that for all farm operations crossbred cowswere more time-efficient. It took crossbred cows less time to complete the first, secondand third plowing than it took local oxen. There was also no significant difference inmean milk yield between those using the crossbred cows for dairy-draught and thoseusing for milk only (Alemu, et al., 1998). However, availability of feed both in terms ofquantity and quality is a major factor determining the performance of dual-purpose cows.

37

3.4.2 Profitability of dual-purpose dairy cows

The profitability of using of crossbred cows for milk production and traction was testedby simulating the production parameters and investment returns over a three-year periodusing the ILRI bio-economic herd model (Shapiro, et al., 1994).34 In order to generatethe multi-year effects of working, on-station experimental results (covering three years)were used to simulate the herd production parameters over ten years with the ILRI model.The financial implications were then investigated using cost-benefit analysis.

(i) the multi-year production effects of using crossbred cows for both milkand traction, as compared to using crossbred cows for milk productionalone; (both working and non-working crossbred cows were supplementedor fed adequately) – SIMULATION 1

(ii) the multi-year production effects of introducing two pregnant crossbredcows for both milk and traction into a typical herd of local cattle overtime, as opposed to the traditional system of using local cows for milkproduction and local oxen for work; (the crossbred cows weresupplemented or fed adequately) – SIMULATION 2

The results of the first simulation (SIMULATION 1) showed that investment in workingcrossbred cows was much more profitable than non-working crossbred cows. The netpresent value of investment in working cows was 29,031 birr, compared to 10,274 birrfrom the non-working cows. The marginal or incremental benefit/cost ratio due toworking cows (relative to non-working cows) was 16 and the marginal internal rate ofreturn was 126% (Table 6). The value of work more than compensated for the smalldecline in milk production and longer reproductive cycle found in working cows.

The simulation on traditional practices versus working supplemented crossbred cowsindicated that investment in the latter alternative could yield a net present value of15,7121 birr, compared to only 4,086 from the first alternative. The marginal internal rateof return was 78% and the marginal benefit/cost ratio 3.5. Even though there wassubstantial additional investment associated with shifting from the traditional system tothe improved one, the net incremental returns on investment was 4. Inadequate feedingand low genetic potential were the major reasons for the low net present value. Theintroduction of working crossbred cows also resulted in a smaller average herd size, 8heads of cattle as opposed to 13 under the traditional practice (Table 6). A lowerstocking rate under the improved system would have a beneficial impact on farm land.

34 The discussion in thin section is based on the paper by Shapiro et al. (1994).

38

Table 6: Results of benefit/cost analysis of dairy traction trials

Simulation 1 Simulation 2SupplementedNon- WorkingCows

SupplementedWorking Cows

Traditional Practices- Local Animals

Supplemented WorkingCows

Average herd size, year 0 to year 10 19 17 13 8Average annual milk offtake per herd 12749 12741 1164 4280Annual undiscounted costs per cattle head 737 792 112 233Annual undiscounted revenues per Cow 5527 7906 3613 5496Annual undisc. Revenues per cattle head 957 1222 260 709Internal rate of return (%) 20.9 200.4 18.8 32.9Net present value 10274 29031 4086 15721Economic returns to the interventionInternal rate of return (%) - 125.7 - 78.4Net present value - 18757 - 11635Benefit cost ratio - 15.6 - 3.5Net Benefit investment ratio - 990.6 - 4.1

Source: Shapiro, B. I. Zerbini E. And Takele G. 1994. The returns to investment in the dual use of crossbred cows for milk productionand draft in the Ethiopian Highlands. Proceedings for he First Conference of the Animal Traction Network of Ethiopia, Addis Ababa,Ethiopia.

39

3.4.2 Constraints to adoption of dual-purpose crossbred cows

Cow traction is not an entirely new practice in Ethiopia. Local cows are used for tractionin some parts of the country, mainly Gojam and Wollo. The number of oxen seems tohave declined in theses areas largely due to scarcity of grazing land, thus forcing farmersto use other animals, including cows, donkeys and horses, for traction. Among theproblems of using local cows for traction are failure of such cows to calve (leading toreplacement problem), lower work performance and significantly reduced milk output.

Undoubtedly, dual-purpose crossbred cows are attractive innovations in areas whereincreasing demand for arable land has reduced the size of land available for naturalgrazing and browse. But the use of crossbred dairy cows for traction is unknown beyondthe few farmers who cooperated to participate in the on-farm research at Holleta. Severalproblems seem to undermine the adoption of the new technology. To start with lack ofadequate feed is a critical bottleneck under the present conditions of the country. Feedshortages and nutrient deficiencies are acute throughout the country. The overall poorquality of the available feed and particularly its deterioration during the dry season couldbe the major challenge to the adoption of working crossbred cows. Cultivation ofproductive fodder crops is largely unknown and commercial feed production is notdeveloped (Teferra, 1993).

Another major problem limiting adoption is poor veterinary service. There are a numberof infectious diseases as well as parasitic and viral diseases with high rates of mortality.Animal health service is underfunded and there are no sustainable and effective diseasecontrol programs. A large portion of the government veterinary budget is allocated tocover salary expenditures with very small amount used to cover operational costs (Addis,1991). Given the existing state of animal health service, most farmers would not riskinvesting in crossbred cows.

Ethiopia has one of the poorest transportation network in the world. Poor infrastructurehas hindered the movement of livestock inputs and outputs. Perishable products such asraw milk can hardly be transported over long distances to high demand urban centers.Commercial feed (concentrates) and veterinary services are either unavailable or cost toomuch in most parts of the country. Economic viability of dual-purpose dairy technologystrongly depends on good market for the milk produced. Milk must be sold at reasonableprice to recover investment costs and pay for concentrates and veterinary services.35

Farmers in distant areas (where interest in cow traction is high) would face a much lowerprice for milk but higher prices for concentrates and veterinary treatment.

Finally, the investment cost of dual-purpose dairy cows is unaffordable for most ruralhouseholds. Widespread adoption is feasible only if long-term loans (preferably insured

35 The profitability study discussed above assume a price of 1 birr per kg of raw milk. Concentrates wereassumed to cost 45 birr per quintal, mineral lick and veterinary care at 8.5 birr per treatement (Shaprio, etal., 1994). These prices are for Holleta which is less than 40 kms from the capital Addis Ababa. Distantand remote areas would face a much lower milk price and higher cost of inputs.

40

ones) are made available. Such type of credit services are non-existent in rural areas atpresent. A less expensive alternative could be sought in artificial insemination (AI) orbull services, instead of buying crossbred heifers. But here again poor farmers needfinancial resources to invest in local cows. Without adequate provision for long-termcredit, poor households are unlikely to benefit from the innovation.

In conclusion, agricultural transformation entails development of profitable technologiesand establishment of institutions for effective dissemination to farmers. It requiresinnovative approaches to improve the performance of research, agricultural extension andall support systems. The case studies discussed above have demonstrated that agriculturaltransformation cannot be an easy task in Ethiopia. There is no technology, which hasmade a widespread impact on agricultural productivity and poverty. With respect towheat, the new variety HAR - 1685 has succumbed to rust disease. It was badly affectedin 1998/99 and many farmers were unable to repay their input loans. The improvedvariety of tef (DZ-196) is superior to farmers’ own seed only on research plots. Failureto multiply the variety without contamination has prevented the technology fromreaching farmers in the case of tef. The technology of dairy traction has the potential totransform crop as well as livestock production in the country. Adopters could gain amarginal or incremental benefit several times higher than marginal cost. But the dairytraction technology has yet to spread beyond the few farmers contacted for on-farmresearch owing to feed constraints and other critical problems. Hybrid maize is perhapsthe only success story among the technologies under consideration. 36 The marginal rateof return for BH - 660 is well above (184%) the recommended rate of 100%. Asexpected there is high interest in the technology but most farmers do not have access toseeds of BH - 660. Weak capacity in producing and distributing the hybrid seed and highrisk associated with weather problems, among others, have limited its adoption.

36 It should be noted that hybred maize seeds have increased yield in Africa. Eicher and Byerelee (1997)described the situation as ‘Africa’s Emerging Maize Revolution: A Qulitified Success’ as a who

41

IV. KEY ISSUES IN TECNOLOGY DEVELOPMENT AND DISSEMINATION

The analysis in section 3 has indicated that lack of profitability and inadequateinstitutional capacity for technology transfer has hindered sustainable increase inagricultural productivity. This section focuses on key issues pertaining to research, seedindustry, agricultural extension and marketing.

(a) Strengthen agricultural research

Ethiopia’s highland exhibit wide differences in climate and soils, giving rise to diversegrowing conditions and farming systems. A total of 18 major agroecologies and 49 sub-agroecologies have been identified for research purposes. Addressing the technologicalneeds of each ecological situation is one of the major challenges facing researchers.Agricultural productivity would stay low and rural poverty would remain high in thecountry unless technologies appropriate for the different locations are developed.

Apart from the extreme diversity of soil, water and rainfall conditions, a number ofdifferent crops are grown in Ethiopia and each one has its own specific productionproblems. Focussing on one crop alone would not bring the desired transformation.Profitable technologies need to be developed for each crop type and agroecologicalconditions. In this regard, one of the main challenges facing wheat research isdeveloping a variety with durable resistance to rust for different agroecologicalconditions. Maize research would need to develop technologies that would performadequately under inadequate and erratic rainfall conditions where the bulk of the ruralpoor reside. Solutions to wild animal attacks may also sought in research aimed atdeveloping effective protection mechanisms. Producing varieties resistant to lodging andaddressing the problem of contamination in seed production are the major challengesfacing tef research. Livestock research needs to develop appropriate forage technologiesthat would make adoption of crossbreds feasible. The prevailing level of investment onlivestock research is considerably low, given the critical role of livestock in the ruraleconomy.

Agricultural research should be geared towards bringing about broad-based technicalchange. Research on soil fertility management must be steeped up to reverse the declinein soil fertility due to the breakdown in traditional soil fertility restoring techniques suchas fallowing and crop rotation especially in densely populated areas. Given the extent ofnutrient mining and land degradation, both organic and inorganic supply of nutrientsmust be expanded.37 Location specific agronomic research on land preparation, croprotation, level of fertilization and control of weed, insects and diseases could have asubstantial impact on land and labor productivity. The positive contribution of improvedfarm implements in raising agricultural productivity should also be appreciated. Thescope of socio-econmic research needs to be broadened, not just on-farm verificationtrials, to include policy-oriented research in the area of production, processing, marketingand consumption.

37 Eicher, C. K. and D. Byerlee. 1997. ‘Accelerating Maize Production: Synthesis’ in C. K Eicher and D.Byerelee (eds.), Africa’s Emerging Maize Revolution’, Lynne Rienner Publishers, Boulder.

42

Lack of adequate resources is among the major factors constraining broad-basedtechnical change in agriculture. Budgetary expenditure on technology development isinadequate especially when viewed against the diversity of agricultural production in thecountry and the recommended level of expenditure. For instance, agricultural researchexpenditure amounted to 26.6 million birr (or 5.3 million USD) in 1993/4, compared toagricultural GDP of roughly 14 billion. This would mean that research expenditure inEthiopia (only 0.2% of the agricultural GDP in 1993/94) falls far below therecommendation which calls for a target of 2% of agricultural GDP to be invested inagricultural research. 38 Lack of funds have meant that it is difficult to engage adequatenumber of scientists in agricultural research. 39 It has also become increasingly difficult toretain senior researchers as wages and working conditions (including schools forchildren) in the research centers are unattractive relative to jobs in the major cities oropportunities abroad. Since private research is non-existent in Ethiopia, largely due tothe absence of a strong commercial agriculture, individuals resigning from research oftenjoin a non-research profession.

Weak institutional base to provide adequate and predictable guidance and framework forgovernance, coordination, follow-up and evaluation system of the Ethiopian NARS isanother major impediment to effective research. The mode of research administration andpriorities have changed with government changes and policy shifts. The most recentchanges, as indicated above, reorganized IAR and brought many research centers underthe control of regional administration. New strategies and long-term plans for crops,livestock and natural resources are under preparation. It remains to be seen how the newarrangements and approaches influence research output. But it should be noted that ittakes 10 to 20 years for agricultural research to bear fruit in increased production.Ethiopian NARS has been reorganized several times since its establishment some 35years ago. The indications are that research programs and emphasis in Ethiopia havelacked stability and continuity.

(b) Use agricultural extension to transfer technical knowledge

Similar to research agricultural extension in Ethiopia is an entirely a public sector activityand suffers from lack of institutional stability and shortage of funds. The Ministry ofAgriculture, responsible for extension, has gone through a number of organizationalchanges in the last 20 to 30 years (Goshu, 1995). It was reorganized several times underthe former government in order to implement unpopular policies such as collectivizationand compulsory delivery quota. Since extension agents are the only public employeesdeployed very close to farmers, they have become indispensable in executing policies andadministrating affairs pertaining to rural areas. They were an important part of theadministrative machinery of the previous government.

38 ISNAR. 1988. Role of Research in Transforming Traditional Agriculture: An Emerging Perspective.The Hague, International Service for National Agricultural Research.39 For instance, there only three full-time researchers holding Ph.D degrees (two breeders and oneagronomist) currently working on wheat. According to the National Wheat Coordinator, the optimumnumber full-time Ph.D researchers is 16.

43

Under the new extension approach, a development agent is expected to serve about 1000households. The agent is responsible for introducing some seven or more differentpackages to farmers under his domain. 40 Every year, he/she receives instructions on thenumber of farmers that need to participate in the demonstration program. The number ofparticipating farmers is determined at regional levels based on some indications of thenational target given by MOA. At national level, the number of participating farmersincreased from 3,200 in 1994/95 to 2.9 million in 1997/98. Once the regional target forthe year is known, disaggregation at zonal and district levels is worked out by bureaus ofagriculture. Accordingly, an agent could be required to work with over 200 farmers.41 Inaddition, disbursement and collection of input loans is largely handled by the agents.Development workers are clearly operating under heavy work load. The problem iscompouned by lack of resources and means of transport to move from one household plotto the next. The perception that the role of agents is to facilitate input distribution hasundermined the main purpose of agricultural extension, i.e. assist in knowledge transfermainly through the process of diffusion.

There are no clear criteria for the selection of farmers participating in the extensionprogram, but willingness to allocate 0.5 ha or 0.25 ha of land and pay 25% of packagecost (in the form of down payment) is an important consideration. For profitabletechnologies such as maize, the number of applicants for participation is often high. Inview of the fact that the performance of an agents assessed based mainly on rates of loanrepayment in his/her respective area, better-off farmers with less risk of default are likelyto be given priority in the selection process. Willingness to participate in whet packagedepends on whether the seed offered is new and resistant to rust. There is very limitedinterest to participate in other technologies such as tef where there are no superior seeds.Participants are allowed to use their own seeds in the case tef packages, implying that thepackage is largely reduced to a single input – fertilizer - in this case. In 1997/98, manyparticipant farmers were supplied with fertilizer alone as improved seeds were either inshort supply or were not needed by farmers. The new system seems to take the form of aproduction campaign in recent years, instead of a proper extension activity.

Most development agents have only nine months of training after graduation from highschools. The training covers a wide range of agricultural activities, ranging from cropproduction to cooperative movement. Often they lack practical farming knowledge andspecific information about the packages that they try to promote. The existing approachdoes not include adaptive research, which used to be conducted by extension staff incollaboration with IAR/EARO. In the absence of local adaptive research, it has meantthat farmers are required to adopt technologies the local profitability and adaptability ofwhich is not always known.

40 There are packages for crop production (mainly wheat package), soil and water conservation, forestry,agroforestry, livestock, forage and horticultural crops.41 An extension agent contacted around Kulumsa Research Center worked with 227 participant farmers. Hewas assisted by 45 model farmers (farmers who previously participated) and three high school students.No payment or compensation was made to the farmers nor to the students. The students provided freeservice in return for support letters that would help them find employment or join training institutions.

44

The potential to increase yield through improved management practices is not given dueattention in Ethiopia. The effort directed towards enhancing technical skill andmanagement capacity of small farmers leaves much to be desired. The research systemhas yet to develop location-specific optimal fertilizer rates and management practices.Extension system operates based on recommendations that show little variation (blanketrecommendations) across different environments. Given the variability of growingconditions, extension agents have very little to offer in terms of profitable managementpractices. They also lack the capacity to chose ‘best” farm practice and provide advicebased on indigenous knowledge.

(c) Seed multiplication - balance public and private sector participation

The seed industry in Ethiopia is weak and of recent origin. Large-scale distribution ofimproved seeds to farmers started with the launching of the Chilalo AgriculturalDevelopment Unit in 1967. In 1978, the Ethiopian Seed Enterprise (ESE) (formerlyknown as the Ethiopian Seed Corporation) was established a government parastatalunder the now defunct Ministry of State Farms, Coffee and Tea Development. ESEdominates the production and distribution of seed in Ethiopia. Over 93% of the total seedsupply came from the Enterprise in 196/97 (Afri-Tech Consult, 1998). It handles seedsof some 16 different crops and several varieties of each crop types (Annex IV).

The quality of improved seeds in Ethiopia is low due to various reasons. It is adverselyaffected mainly by loss of genetic quality due to long period of repeated use and seedproduction under rainfed conditions. For instance, about 10% of the demonstration plotsin two regions (Oromia and SNNP) which were planted with BH - 660 faced germinationproblems during the 1998/99 cropping season. Similar problems were reported for wheat(HAR 1685) in all areas where the seed was distributed. Losses of up to 100% werereported on some plots.42 Seed quality can be adequately maintained when producedunder irrigation.

The multinational firm, Pioneer Hybrid, is the only other major actor in Ethiopia and it isinterested in hybrid maize only. It started operation in 1990/91 and produced 7000 and8725 quintals in 1996 and 1997, respectively. The company imports parent seed fromZimbabwe, where the parent company operates one of its two main research centers inAfrica (the other is located in South Africa). No other private seed producer has beenlicensed so far. A few private and state farms currently producing seeds operate undercontract from ESE. The legislation for licensing private seed producers is underpreparation. 43 Lack of access to land is expected to constrain private sector participationeven after the legislation is issued.44

42 Based on the report of the Committee organized in 1998 by the National Seed Industry Agency toinvestigate the problem of seed germination.43 Pioneer was licensed as foreign investor and lack of legislation did not apply in this case.44 Land for commercial purposes is obtained through lease from the government. Investors have access tounoccupied land only which is often located in remote or inaccessible areas. Such land may not beattractive to seed producers.

45

Seed producers generally prefer hybrid seeds to open-pollinated ones such as wheat andtef. The latter can be recycled by farmers (at least for two to three years), resulting inweaker demand relative to hybrid seeds which are not recycled normally. The EthiopianSeed Enterprise recovers its losses in wheat and other seeds from the profit it earns inhybrid maize.45 Private producers are expected to show limited interest in open-pollinatedseeds. Public support is needed to ensure adequate supply of non-hybrid seeds. Onlysubsidies and improved access to land would promote the production of open-pollinatedseeds.

(d) Improve the efficiency of input marketing

Liberalization of fertilizer market has removed price regulation and allowed privatesector participation. Nonetheless, fertilizer importers and distributors face considerableuncertainty. Fertilizer import in Ethiopia is financed mainly through foreign grants andloans that are obtained at unspecified times of the year and are often grated withrestrictive conditions such as source of supply. Unlike the weekly auction for all otherimported commodities, foreign exchange for fertilizer import is offered for tenderwhenever donations are received. The timing of the tender do not necessarily correspondto seasons of low fertilizer price in the world market. Moreover, the amount of foreignexchange available in any one tender is too small (equivalent to lots of 25,000 tons orless) to gain from economies of scale in import and shipment. These and other relatedproblems have raised the cost of fertilizer by at least 13% (Mulat et al., 1996). Sincefertilizer is by far the biggest financial cost component of production, reducing the cost offertilizer will enhance technology adoption.

Fertilizer distribution is marked by regional monopolies and lack of level-playing field.Contrary to the national goal of developing a free market, local authorities in Amhara andSNNP regions limit market share to importers and companies that have close links withthe government (owned by parties and local governments). The regional governmentscontrol fertilizer loan and limit credit sales to the preferred companies, making the marketless competitive and unpredictable. A survey in 1998 found that fertilizer prices wereabout 5% higher in regions that restricted competition (Stepanik, et al., 1998). Policyuncertainties have also raised the cost investing in the fertilizer sector and discouragednew entrants (Howard, et al., 1998). These problems, together with the devaluationmeasure and removal of subsidies, have further increased the price of fertilizer.

The majority of farmers in Ethiopia (over 80%) buy fertilizer on credit. But there is noeffective mechanism to enforce repayment. One commonly applied measure to enforcerepayment is to require all members of a service cooperative or peasant association repayall previous loans before a new loan for the current season is approved. Fertilizer salesmay be suspended even when the number of defaulters is small and when the reasons fordefault are legitimate (e.g. crop failure). Delays in fertilizer sales often result in delays ofplanting time. 45 Personal communication with Ato Tarekegn Ferede, ESE.

46

The seed market is less developed than the fertilizer market. Improved seeds are notalways available in the open market. Private sector participation in marketing improvedseeds is minimal. ESE’s sales to farmers are carried out largely by the bureaus ofagriculture, not by retailers of seed (seed-lots stockists) (Afri-Tech Consult, 1998). Non-participant farmers (in the extension program) have no access to BH - 660. Low demandfor ESE’s seeds (except for maize and newly released varieties of wheat) has alsodiscouraged private sector participation. Pioneer Hybrid appointed private dealers in1998 but sales are constrained by lack of credit for seed buyers (outside the extensionsystem), high prices and limited awareness about high quality hybrid seeds (Afri-TechConsult, 1998). Its prices are relatively higher (8.20 birr/kg in 1998, compared to 5.68birr/kg of ESE). Lack of coordination of seed supply, fertilizer distribution, credit andoutput marketing is one the major impediments to technical change in agriculture.

(e) Remove distortions in the land market

The average farm size in Ethiopia has declined to just one ha due to the rapidly growingpopulation. Over one-third (46%) of the rural holdings are less than 0.5 ha. Given the lowlevel of productivity, nearly all produce is devoted to home consumption for householdswith smaller plots. There is little surplus for investment and for input purchase.Empirical studies have also shown that the probability of adopting fertilizer and improvedseeds decreases with decline in farm size (Croppenstedt, et al., 1998; Mulat et al., 1998;Wolday, 1998). Sub-economic holdings operated by poverty-stricken farmers are notfavorable for widespread dissemination of new agricultural technology.

Apart from the population pressure, the land policy has significantly contributed to sub-economic holdings and tenure insecurity. Since the 1975 land reform which made allrural land public property, the possession of land plots has been conditional uponresidence in the village. The transfer of land through long-term lease or sales has beenforbidden and periodic redistribution has been carried out to give land to new entrants.Despite the reform aimed at creating a free market economy, the present governmenthas decided to maintain public ownership of all rural and urban land. The December 1994Constitution of the Federal Democratic Republic of Ethiopia proclaimed that ‘Land is acommon property of the nations, nationalities and peoples of Ethiopia and shall not besubject to sale or to other means of transfer’. In 19996, land redistribution was carriedout in the Amhara Regional State. Land was given to landless youth and returnee ex-soldiers by reducing the holding of farmers who were reportedly associated with previousgovernments. Communal grazing and wood land was allotted to new claimants in otherregions. Increasing population in the rural areas was thus absorbed in agriculture throughleveling down of holdings, rather than through alternative forms of employment.

Given the absence of any contractual or lease agreement with the government and thegeneral belief that the next round of land redistribution will take place any time, theincentive to invest on land improvement is minimal. This situation, coupled with the highlevel of poverty and subsistence pressure, has discouraged long-term investment andexacerbated the problem of land degradation. The soil in many areas have lost somebiological productivity and physical properties needed for optimal plant growth (FDRE,

47

1996). Removal of the distortions in the land market is a major prerequisite foragricultural transformation in Ethiopia.

(f) Output marketing

Output marketing in Ethiopia suffers from a number of constraints, including inadequatetransport network, limited number of large-interregional traders with adequate storageand working capital, high handling costs, inadequate market information system, weakbargaining power, and underdeveloped processing industrial sector. Transport andcommunication system leaves much to be desired in Ethiopia. The whole country isserved by a total of 4,109 km asphalt road, 9,287 gravel and 5,610 rural roads. Theaverage road density is only 21 km per 1,000 sq. Km or 0.44 km per 1,000 population,one of the lowest in Africa.46 An estimated 75% of farms are more than a half day walkfrom an all-weather road. Inadequate roads, both within Ethiopia and between Ethiopiaand neighboring countries, inflate transport costs and impede the viability of grain tradethat would otherwise moderate extreme price fluctuations. Lack of adequate transport(especially for milk marketing) is also a major constraint to adoption of more productivedairy breeds.

Large inter-regional traders are small in number and are in the process of establishmentfollowing the lifting of the restriction imposed on grain trade in the early 1990s. Smalltraders have limited capacity of handling large quantities for longer duration. Hence, thevolume of grain marketed falls sharply in years of poor harvest and prices riseconsiderably. Grain prices, on the other hand, are seriously depressed in good years andimmediately after harvest. Handling and transport costs are also high due to the smallquantities that farmers bring to market places (often small bags carried on head or on theback of pack animals) and the absence of grading and standards. Grain has to beunbagged and rebagged each time it changes hands, partly to inspect quality.

With large number of small traders in the market, it is difficult for the EthiopianStandards Committee to effectively inspect weighing scales or the bureaus of trade tomonitor trading activities. Traders in many village markets use containers of differentsizes, not proper weighing scales (Wolday, 1999). A good part of the margin that goes totraders comes from the inappropriate measuring instruments (under-weighing), not justprice differentials. Unfair trading practices, combined with lack of accurate and timelymarket information and high rate of illiteracy among farmers, have widened the gapbetween the price paid by consumers and the price received by farmers. Since there areno cooperatives or marketing societies, the bargaining power of farmers is weak relativeto traders.47 A more efficient marketing system calls for a more timely and widelydisseminated market information (Asfaw and Jayne, 1997) and establishment of aneffective monitoring system.

46 The figures refer to 1994 as reported by the Ethiopian Road Authority,47 Cooperative movement in Ethiopia has not recovered from the disastrous experience under the formersocialist government.

48

Storage problems coupled with the need to repay loans and meet other financialobligations (e.g. taxes) compel farmers to sell maize immediately after harvest. It isestimated that about 79% of farmers’ annual grain sales occur immediately after theharvest season (January to March) (Gebremeskel, et al., 1998). As a result, prices dropwhen most farmers are selling and rise later during the year when many poorer farmersrun out of stock and start buying from the market. Net income per hectare and per laborday would increase by an estimated 9-17% if storage losses in maize were reduced byhalf (Howard, et al., 1998). High seasonal price fluctuation is also likely to make surplusproducing farmers reluctant to make important investment in inputs such as fertilizers andimproved seeds.

49

REFERENCES

Abate Bekele. 1995. ‘On-farm analysis of improved production for durum wheat: a casestudy of three districts in Ethiopia’ in D. G. Tanner, T.S Payne and O.S. Abdalla (eds.) ..

Abiye Astatike and M. A. Mohammed Saleem. 1996. ‘Draught animal power for land-use intensification in the Ethiopian highlands’, WAR/RMZ 86

Abrham Tadesse, Firdu Azerefegn, Assefa Gebre Amlak and Adhanom Negassi. 1993.‘Research Highlights on Maize Insect Pests and their Management in Ethiopia’, in BentiTollesa and J. K. Ransom, ..

Addis Anteneh. 1991. The Financing and Staffing of Livestock Services in Sub-SaharanAfrica: A Cross-Country Analysis, ILCA, Addis Ababa, Ethiopia.

Afri-Tech Consult. 1998. Seed Marketing Study, Volume I, Main Report. National SeedIndustry Agency, Addis Ababa, Ethiopia.

Alemu Gebre Wold, Abebe Misgina and B. I. Shapiro. 1998. ‘Animal traction and thepotential for use of dairy cows in the Ethiopian Highlands’, in E. Zerbini, B. I. Shapiroand J.C Chirgwin (eds.), Technology Transfer: Multi-Purpose Cows for Milk, Meat andTraction in Smallholder Farming System, Proceedings of the an Expert Consultation heldat ILRI, Addis Ababa, Ethiopia, 11-14 Septemeber 1995.

Amanuel Gorfu, D. G. Tanner, Kefyalew Girma and Asefa Taa. 1996. ‘The influence ofcrop rotation and fertilizer level on weed density in bread wheat in southeastern Ethiopia’in in D.G. Tanner, T. S. Payne and O. S. Abdalla (eds.), The Ninth Regional WheatWorkshop – For Eastern, Central and Southern Africa, Addis Ababa, Ethiopia, Oct. 2- 6,1995.

Amsal Tarekegne, Getnet Gebeyehu, D.G. Tanner and Bekele Geleta. 1996. ‘Trends inBread Wheat Yield Improvement in Ethiopia from 1967 to 1994’, in D. G. Tanner, T. S.Payne and O. S. Abdalla. (eds.), The Ninth Regional Wheat Workshop for Eastern,Central and Southern Africa, October 2-6, 1995, Addis Ababa, Ethiopia.

Amsal Tarekegne. 1996. ‘Wheat Production Technology in Ethiopia’ in Aberra Deressaand Beyene Seboka (eds.). Research Achievements and Technology Transfer Attempts:Vignettes from Shoa. Proceedings of the First Technology Generation, Transfer and GapAnalysis Workshop.

Asefa Taa, et al. 1998. On-Farm Evaluation of Fertilizer Response of Two Recently-Released Bread Wheat Varieties in Ethiopia, IAR/CIMMYT.Regassa Ensermu, et al. 1998. Farmers’ Wheat Seed Sources and Seed Management inChilalo Awraja, Ethiopia, IAR/CIMMYT, Mexico.

50

Asefa Taa, D. G. Tanner, Kefyalew Girma, Amanuel Gorfu, and Duga Debele. 1996.‘Influence of Stubble Management, Tillage and Precursor Crop on Soil Physical-Chemical Properties in Southeastern Ethiopia’ in D.G. Tanner, T. S. Payne and O. S.Abdalla (eds.), The Ninth Regional Wheat Workshop – For Eastern, Central andSouthern Africa, Addis Ababa, Ethiopia, Oct. 2- 6, 1995.

Asefa Taa, D. G. Tanner, Melesse Temesgen, and Kefyalew Girma. 1997. ‘On-farmevaluation of animal drawn implement developed in Ethiopia for row placement of wheatseed and basal fertilizer, African Crop Science Journal, Vol. 5, No. 4, pp. 359 –369.

Assefa Tefferi and Tewabech Tilahun. 1992. ‘Review of maize research in Ethiopia’ inBenti Tollesa and J. K. Ransom, ..

Asfaw Negassa & T.S Jayne. 1997. The response of Ethiopian Grain Markets toLiberalization, Working Paper #6, GMRP, MADaC, Addis Ababa

Berhanu Kinfe. 1985. ‘Progress of weed management in wheat producltion in Ethiopia’,in Regional Wheat Workshop for eastern, Central and Southern Africa and the IndianOcean. Njoro, Kenya, Sept. 2 – 5, 1985, Nairobi: CYMMIT.

Beyene Tadesse. 1998. The Impact of Agricultural Extension on Maize Production: ACase study of Sasakawa Global 2000 Extension Project in Bako Area. Addis AbabaUniversity, Department of Economics, M.Sc thesis.

Blackie, M. J. (1994). ‘Maize Productivity for the 21st Century: the African Challenge’,in Proceedings of the Fourth Regional Eastern and Southern Africa Regional MaizeConference, 28 March - 1 April, CIMMYT, Harare.

Botha, CAJ. (1995). ‘Participation, Planning and Linkages: International Lessons forAgricultural Extension in South Africa’, Development Southern Africa, Vol. 12, No. 5,October.

Byerlee, D. (1996). ‘Modern Varieties, Productivity, and Sustainability: RecentExperience and Emerging Challenges’, World Development, Vol. 24, No. 4,

_________ (1998). ‘The Search for a New Paradigm for the Development of NationalAgricultural Research Systems’, World Development, Vol. 26, No. 6.

Brossier, J. 1991. ‘Agricultural Technology: The Priorities for Sub-Saharan Africa andSome Arguments in the Debate’ in in Suzanne Gnaegy, et al., Agricultural Technology inSub-Saharan Africa, World Bank Discussion Papers, Washington D.C.

Byerlee D. and P. W. Heisey. (1996). ‘Past and Potential Impacts of Maize Research inSub-Saharan Africa: A Critical Assessment’, Food Policy, Vol. 21, No. 3.

51

Chilot Yirga, Hailu Beyene, Lemma Zewdie and D. G. Tanner. 1992. ‘Farming Systemsof the Kulumsa Area’, in S. Franzel and H.van Houten (eds.), Research with Farmers –Lessons from Ethiopia, CAB International, Wallingford, U.K.

Cladwell, R.M. 1992. Ethiopia: A Country Profile for Famine Mitigation, Planning andImplementation. OFDA/USDA, Famine Mitigation Activity PASA.

Collins, C. 1991. ‘Technological Issues in World Bank Lending for Agriculture, inSuzanne Gnaegy, et al., Agricultural Technology in Sub-Saharan Africa, World BankDiscussion Papers, Washington D.C.

Croppenstedt, A., Mulat Demeke and M. Meschi. 1999. ‘An Empirical Analysis ofDemand for Fertilizer in Ethiopia’, Ethiopian Journal of Agricultural Ecnomics . Vol 3,No. 1.

Dercon, S. And P. Krishinan. 1998. Changes in Poverty in Rural Ethiopia 1989-1995:Measurement, Roboustness Tests and Decomposition. Center for the Study of AfricanEconomies, Oxford University and Katholieke Unversiteit Leuven (Unpublished Memo).

Ethiopian Seed Enterprise. 1997. Improved Crop Varieties Produced by Ethiopian SeedEnterprise, Addis Ababa

ESE. 1997. Improved Crop Varieties Produced by Ethiopian Seed Enterprise. AddisAbaba, Ethiopia.

FDRE. 1996. Food Security Strategy, Prepared for the Consultative Group Meeting ofDecember 10 - 12, Addis Ababa, Ethiopia.

Fraser, GCG. (1994). ‘Provision of Accessible Marketing and Extension Services inLess-Developed Agriculture’, (Research Note), Development Southern Africa, Vol. 11,No. 1, February.

Gebremeskel Dessalegn, T.S. Jayne and J.D. Shaffer. 1998. Market Structure, Conductand Performance: Constraints on Performance of Grain Markets. Working Paper No. 8,GMRP, MEDaC.

Getahun Tafesse. 1998. Poverty in Ethiopia. Ethiopian Economic Association. AddisAbaba.

Ghatak, S. (1987). ‘Agriculture and Economic Development’, in N. Gemmell (ed.),Surveys in Development Economics, Basil Blackwell, Oxford.

Goe, M. R. 1987. Animal traction on smallhodler farms in the Ethiopian highlands.Department of Animal Science, Cornell University, Ithaca, New York, (Ph.D.Dissertation)

52

Goshu Mekonen. 1995. ‘Agricultural Research and Extension in Ethiopia: An Overview’,in Dejene Aredo and Mulat Demeke (eds.), Ethiopian Agriculture: Problems ofTransformation, Proceedings of the Fourth Annual Conference on the EthiopianEconomy.

Griffin, K. 1989. Alternative Strategies for Economic Development. Macmillan inAssociation With OECD Development Center.

Hailu Tefera, Mulu Ayele and Kebebew Assefa. 1995. Improved Varieties of Tef(Eragrostis tef) in Ethiopia. Release of 1970- 1995. Research Bulletin No. 1, Debre ZeitAgricultural Research Center, Alemya University of Agriculture, Debre Zeit, Ethiopia.

Hailue Gebre, Amsal Tarekegne and Endale Asmare. 1989. ‘Beneficial Break Crops forWheat Production’, Ethiopian Journal of Agricultural Science, Vol. 11, pp. 15-24.

Howard J., Mulat Demeke, V. Kelly, M. Maredia, J. Stepanik. 1998. Can the Momentumbe sustained? An Economic Analysis of the Ministry of Agriculture/ Sasakawa Global2000’s Experiment with Improved Cereals Technology in Ethiopia, Department ofAgricultural Economics, Michigan State University, E. Lansing, Michigan

Idachaba, F. (1986). ‘The Evolution of National Agricultural Research Systems in Sub-Saharan Africa’, in A. Maunder and U. Renborg (eds.), Agriculture in a Turbulent WorldEconomy, Proceedings of the Nineteenth International Conference of AgriculturalEconomists, held at Malaga, Spain, 26 August - 4 September 1985.

Kebede Mulatu, et al. 1993. ‘Maize Production Trends and Research in Ethiopia’, inBenti Tolessa and J. K. Ransom (eds.), Proceedings of the First National MaizeWorkshop of Ethiopia, 5- 7 May 1992, Addis Ababa.

Lipton, M. (1998). Success in Anti-Poverty, ILO, Geneva.

Lipton, M with R. Longhurst. (1989). New Seeds and Poor People. London. As quotedin Lipton, M. (1998).

Mekonen Tadesse, Bereket Kebede and Abebe Shimelis. 1998. Economic Reform,Growth and Poverty in Ethiopia: Evidence from Household Panel Data. Paper presentedto the African Economic Research Consortium, Bi-annual Workshop, Narobi, Kenya.

Mohammed Hassena, Regassa Ensermu and Kefyalew Girma. 1996. ‘On-farmVerification of Advanced Bread Wheat Lines Under Recommended and Farmers’Management Levels in Arssi. in in D.G. Tanner, T. S. Payne and O. S. Abdalla (eds.),The Ninth Regional Wheat Workshop – For Eastern, Central and Southern Africa, AddisAbaba, Ethiopia, Oct. 2- 6, 1995.

53

Mohammed Hassen, Regassa Ensermu and Kefyalew Girma. 1998. On-farm Verificationof Advanced Bread Wheat Lines Under Recommended And Farmers’ ManagementLevels in Arssi Region. Kulumsa Research Center, Assela, Ethiopia. (UnpublishedMemo).

Monu, E. (1997). ‘Farmer Participation in Research: Implications for AgriculturalDevelopment’, Journal of Social Development in Africa, Vol. 12, No. 1.

Mulat Demeke, Ali Said and T.S. Jayne.1997. Promoting fertilizer use in Ethiopia: Theimplication of improving Grain market performance, input market efficiency and farmmanagement, working paper 5, GMRP MEADac, Addis Ababa.

Mulat Demeke, V. Kelley, T.S. Jayne, Ali Said, J.C.Le Vallee and H. Chen. 1998. Theresponse of Ethiopia Grain Market to liberalization, working paper 6, GMRPO,MEADaC, Addis Ababa.

Mulat Demeke. The Structure and Performance of Ethiopian Agriculture. EthiopianEconomic Association. Addis Ababa. November 1998.

Pray, C. E, and V. W. Ruttan. (1990). ‘Science and Technology Policy: Lessons from theAgricultural Sector in South and Southeast Asia’, in R. E. Evenson and G. Ranis (eds.),Science and Technology: Lessons for Development Policy, Intermediate TechnologyPublications, London.

Regassa Ensermu, W. Mwangi, H. Verkuijl, M. Hassen and Z. Alemayehu. 1998.Farmers’ Wheat Seed: Sources and Seed Management in Chilalo Awraja, Ethiopia.Mexico D.F.: IAR/ CIMMYT

Rezene Fesseha. 1985. ‘A review of wed science research activities on wheat and barleyin Ethiopia’, in Tsedeke Abate (ed.). A Review of Crop Production Research in Ethiopia.IAR. Addis Ababa, Ethiopia.Sanders, J. H., B. I. Shapiro and S. Ramaswamy. 1996. The Economics of AgriculturalTechnology in Semiarid Sub-Saharan Africa, The John Hopkins Univ. London.

Shapiro, B. I. Zerbini E. And Takele G. 1994. The returns to investment in the dual useof crossbred cows for milk production and draft in the Ethiopian Highlands. Proceedingsfot he First Conference of the Animal Traction Network of Ethiopia, Addis Ababa,Ethiopia.

Seifu Ketema. 1993. Tef (Eragrostis tef) – Breeding, Genetic Resources, Agronomy,Utilization and Role in Ethiopian Agriculture. Institute of Agricultural Research, AddisAbaba, Ethiopia.

Stepanek, J., V. Kelly and J. Howard. 1998. From a Sasakawa Global 2000 PilotProgram to Sustained Increases in Agricultural Productivity: The Critical Role ofGovernment Policy in Fostering the Ethiopian Transition. (Unpublishd Memo).

54

Tanner, D.G., H. Verkuijil and Asefa Taa. 1998. An Agronomic and Economic Analysisof Long-Term Wheat-Based Crop Rotation Trial in Ethiopia. CIMMYT / KulumsaResearch Center. (Unpublished Memo).

Tilahun Geleto, D. G. Tanner, Tekalign Mamo and Getinet Gebeyehu. 1996. ‘Responseof rainfed bread and durum wheat to source, level and timing of nitrogen fertilizer at twovertisol sites in Ethiopia’, in D.G. Tanner, T. S. Payne and O. S. Abdalla (eds.), TheNinth Regional Wheat Workshop – For Eastern, Central and Southern Africa, AddisAbaba, Ethiopia, Oct. 2- 6, 1995.

Timmer, P. 1990. ‘The Agricultural Transformation’, in C. K. Eicher and J. M. Staatz(eds.), Agricultural Development in the Third World,

_______ 1993. ‘Keynote Address: Setting the Stage’, EPAT/Winrock International,Agricultural Transformation in Africa, Arlington, USA.

von Braun, J. (1989). ‘Technology Utilization in the Agriculture of DevelopingCountries: Elements of Poverty-Oriented Policy and Programs’ in A. M. Kesseba (ed.),Technology Systems for Small Farmers: Issues and Options, Westview Press, London.

Wolday Amaha, 1999. ‘Improved Seed Marketing and Adoption in Ethiopia’, EthiopianJournal of Agricultural Economics, Vol. 3, No. 1,

________ . 1999. IGAD Grain Marketing Training Program: Country Report: Ethiopia.Addis Ababa, Ethiopia.

World Bank. 1982. World Development Report 1982. New York, Oxford UniversityPress.

World Bank. 1997. World Development Report.

Zerbin, E, Alemu Gebre Wold and B. I . Shapiro. 1998. ‘ Development of cow tractiontechnologies for adoption in the East African Highlands’ in E. Zerbini, B. I. Shapiro andJ.C Chirgwin (eds.), Technology Transfer: Multi-Purpose Cows for Milk, Meat andTraction in Smallholder Farming System, Proceedings of the an Expert Consultation heldat ILRI, Addis Ababa, Ethiopia, 11-14 Septemeber 1995.