Technical Efficiency Analysis of Malt Barley Production: The Case of

Smallholder Farmers in Debark Woreda, North Gondar Zone of the

Amhara National Regional State

M.Sc. Thesis

Endalkachew Yehun

January 2012

Haramaya University
 SCHOOL OF GRADUATE STUDIES
HARAMAYA UNIVERSITY

A Thesis Submitted to the Department of Agricultural Economics,

Schoolof Graduate Studies
Haramaya University

In Partial Fulfillment for the Degree of

MASTER OF SCIENCE IN AGRICULTURAL ECONOMICS

By
Endalkachew Yehun

January 2012
Haramaya University
 APPROVAL SHEET OF THE THESIS

SCHOOL OF GRADUATE STUDIES

HARAMAYA UNIVERSITY

As members of the Examining Board of the Final MSc Open Defense, we certify that wehave
read and evaluated the thesis prepared by Endalkachew Yehun entitled: Technical Efficiency
Analysis of Malt Barley Production: The Case of Smallholder Farmers in Debark Woreda,
North Gondar Zone of the Amahara Regional State and recommend that it be accepted as
fulfilling the thesis requirementfor the degree of: Master of Science in Agricultural
Economics.

---------------------------------- --------------------------------- -----------------------

Name of Chairman Signature Date

----------------------------------- --------------------------------- -----------------------

Name of Advisor SignatureDate

......................................... …………………………. ………………….

Name of Co-advisor Signature Date

---------------------------- ------------------------- --------------------

Name of Internal ExaminerSignature Date

----------------------------------- --------------------------------- -----------------------

Name of External Examiner Signature Date

Final approval and acceptance of the thesis is contingent upon the submission of the finalcopy
of the thesis to the Council of Graduate Studies (CGS) through the DepartmentalGraduate
Committee (DGC) of the candidate’s major department.

I hereby certify that this thesis prepared under my direction and recommend that it beaccepted
as fulfilling the thesis requirement.

--------------------------- ------------------------- ----------------------

Name of Thesis Advisor Signature Date

ii
 DEDICATION

To the late Dejen Alemu, who was always seeking my educational advancement. May
Almighty God rest his soul in heaven

iii
 STATEMENT OF THE AUTHOR

First, I declare that this thesis is my real work that all sources of materials used for this
thesishave been duly acknowledged. This thesis has been submitted in partial fulfillment of
therequirement for an advanced MSc degree at the Haramaya University and is deposited at
theUniversity Library to be made available to borrowers under the rules of the Library.
Isolemnly declare that this thesis is notsubmitted to any other institution anywhere for
theaward of any academic degree, diploma or certificate.

Brief quotations from this thesis are allowable without special permission provided
thataccurate acknowledgment of source is made. Requests for permission for extended
quotationfrom or reproduction of this manuscript in whole or in part may be granted by the
head of themajor department or the Dean of School of Graduate Studies when in his or her
judgment theproposed use of the material is in the interests of scholarship. In all other
instances, however,permission must be obtained from the author.

Name: Endalkachew Yehun Signature _______________

Place: Haramaya University, Haramaya
Date of Submission: _______________

iv
 ACRONYMS AND ABBREVIATIONS

AE Allocative Efficiency
AME Adult Man Equivalent
ANRS Amhara National Regional State
ARARI Amhara Regional Agricultural Research Institute
BoFED Bureau of Finance and Economic Development
BoARD Bureau of Agricultural and Rural Development
COMESA Common Market for East and South Africa
CSA Central Statistical Agency
DAP Di Ammonium Phosphate
DEA Data Envelopment Analysis
DWAO Debark Woreda Agricultural Office
ECOWAS Economic Community of West African States
EE Economic Efficiency
EPCC Ethiopian Population Census Commission
FAO Food and Agricultural Organization
FDRE Federal Democratic Republic of Ethiopia
FSDPO Food Security and Disaster Prevention Office
GDP Gross Domestic Product
GTP Growth and Transformation Plan
HYV High Yielding Variety
IGAD Inter-Governmental Group on African Development
Iid Independently and identically distributed
KA Kebele administration
Ln Natural logarithm
m.a.s.l Meter above Sea Level
MLE Maximum Likelihood Estimate
MoFED Ministry of Finance and Economic Development
MoARD Ministry of Agricultural and Rural Development
OLS Ordinary Least Squares

v
 ACRONYMS AND ABBREVIATIONS (CONTINUED)

ORDA Organization for Rehabilitation and Development in Amhara

PSNP Productive Safety Net Program
SADCC South African Development Coordination Conference
SFPF Stochastic Frontier Production Function
STD Standard Deviation,
TE Technical Efficiency
TLU Tropical Livestock Unit
WFP World Food Program
VIF Variance inflation Factor

vi
 BIOGRAPHY

The author was born at Burie town, West Gojjam zone of Amahara, in 1967. He attended his
primary and junior secondary education at Burie Ras Bitewoded Mengesha Atikem
elementary and junior school. He completed high school education in the nearby Finote Selam
town at Damot Comprehensive Secondary School in 1984. He joined the then Alemaya
College of Agriculture (changed in the course of study to Alemaya University of Agriculture,
and now Haramaya University) in September 1985 and graduated with B.Sc. degree in
Agricultural Economics in July 1989.

Upon graduation, he was employed in the former Arsi Administrative Zone under the
Ministry of Agriculture and served for 4 years. Then,he worked for different Governmental,
Bilateral, and Private Organizations at various capacities in different parts of the country. In
2008, he joined the Amhara Food Security and Disaster Prevention Office as an expert until
he joined Alemaya University for his postgraduate study in September 2009.The author is
married and with three children: A son and two daughters

vii
 ACKNOWLEDGEMENT

No one is like you. All the ups and downs in the journey of my study turned in to success by
your mercy and grace. Thank youthe Almighty God, as this is the only word you seek from
mefor all your priceless gifts.

Special thanks are due to my major advisor Dr. Admasu Shibru for his intellectual
stimulation, helpful comments and brother-hood affection. I am also highly indebted to Mr
Fekadu Gelaw for both his friendship and co-advisory role. His contribution goes beyond rich
professional support and devotion of tight time in the research. His efforts in awakening me
for learning andfor my enrolment under the graduate study program are immense.He always
has great understanding and good solutions for all my problems. My study could not have
been materialized without his critical expertise inputs, which has of great value to my work.
I am thankful, to the ANRS for allowing me to pursue MSc study, salary grant during the
study period, and limited research support. My thanks also go to food security and DWAO
staffs for their support all the way through my survey work. I am also grateful to the
respondent farmers andenumerators for their involvement and patience in the challenging
household survey, data collection process.

I am very much indebted to Ato Yeshiwas Menesha, Mengesha Jatu,Dr Biresaw Mahitot,
Amare Kinde, Amare Birhanu, Biazin Atinafu, Tilahun Mequanint, Wondimagegn Muchae,
Fentahun Ameshe, and other friends who are too many to list all their names here. I have
benefited a lot from their different material, moral, and expertise supports to bridge the gaps
in my study and family life so as not to stop what I have started.I remain thankful, to Emuhay
Fentaye Yehun, W/ro Enu Lemlemu, Zufan, Fetlework, Silas,andAbebech Dejen who are
always anxious to see my success, and share with me sorrows and happiness.

Finally, my indebtedness and heart felt thanks are due to my beloved wife, Alegntaye Dejen
for her strength and shouldering of all family responsibilities in the difficult days of my
absence. Our darling kids; Fikir Edalkachew,Yonatan Endalkachew,and Efrata Endalkachew

viii
also merit special thanks for paying the most expensive price of alienation from their father
hood affection during the study period, for the betterment of their father.

ix
 TABLE OF CONTENT

STATEMENT OF THE AUTHOR IV

ACRONYMS AND ABBREVIATIONS V
BIOGRAPHY VII
ACKNOWLEDGEMENT VIII
LIST OF TABLES XI
LIST OF FIGURES XII
LIST OF TABLES IN THE APPENDICES XIII
ABSTRACT XIV
1. INTRODUCTION 1
1.1. Background 1
1.2. Statement of the Problem 4
1.3 Objectives of the Study 6
1.4. Scope and limitations of the Study 6
1.5. Significance of the Study 7
1.6. Organization of the thesis 8
2. LITERATURE REVIEW 9
2.1. Concepts of Efficiency in Production 9
2.2. Models for Measuring Technical Efficiency 12
2.2.1. Non parametric frontier estimation methods 13
2.2.2. Parametric frontier estimation methods 14
2.3. Empirical Studies of Efficiency Estimation and Its Determinants in Ethiopia 19
2.3.1. Empirical studies on efficiency level 20
2.3.2. Empirical studies on determinant of technical efficiency 23
3. METHODOLOGY 26
3.1. Description of the Study Area 26
3.2. Methods of Data Collection 28
3.2.1 Data types and sources 28
3.2.2.Sampling procedure and sample size 28
3.2.3. Data collection 29
3.3. Variable Selection, Measurement Units and Hypotheses 30

x
 TABLE OF CONTENT (CONTINUED)

3.3.1. Input and output variables 30

3.3.2. Inefficiency Variables 32
3.4.Methods of Data Analysis 35
4. RESULTS AND DISCUSSIONS 42
4.1. Descriptive Results 42
4.1.1. Demographic characteristics and Education of sample households 42
4.1.2. Institutional /Infrastructural access and services 45
4.1.2.1. Institutional and Infrastructural access 45
4.1.3. Holidays 48
4.1.4. Resource bases 48
4.1.5. Cropping system and production 54
4.1.6. Major Production Constraints 62
4.2. Results of Econometric Analysis 62
4.2.1. Hypotheses testing 62
4.2.2. Estimation of parameters for Stochastic Frontier Production Function (SFPF) 64
4.2.3. Technical efficiency scores 67
4.2.4. Actual and potential levels of output 69
4.2.5. Variability of output from the frontier 69
4.2.6. Determinants of technical inefficiency 70
5. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 80
6. REFERENCES 84
7. APPENDICES 90

xi
 LIST OF TABLES
Tables
page

1. Sample kebele administrations and sample size selected from each kebele........................29
2. Summary of input and technical inefficiency variables and their expected signs...............41
3. Demographic characteristics and education of household heads.........................................43
4. Demographic conditions and Education of the family members.........................................44
5. Households’ proximity to institutional services and all-weather roads in walking minutes45
6. Fertilizer use level in malt barley production by sample households’ in 2010...................48
7. Cultivated land of sample households in 2010 by sources..................................................49
8. Average distance of Malt barley plots from homesteads of the sample households...........51
9. Livestock ownership of sample households by TLU and numbers of the livestock type....52
10. Horse ownership of the sample households.......................................................................53
11. Household horse holdings and additional land access.......................................................53
12. Size and plots of annual cropland of sample farmers’ in 2010..........................................55
13. Recommended & actual fertilizer use in malt barley and wheat by households’ in2010..56
14. Crop productivity loss estimates (qt/ha) of sample households in 2010...........................57
15. Labor availability and use by malt barley production activities and sources (in MAE)....60
16. Farmers’ performances and potential of malt barley productivity in different seasons....61
17. Tests of hypotheses for model and parameters of the SPF................................................64
18. ML estimates for parameters of Stochastic Frontier Production Function........................65
19. Technical efficiency score of sample households..............................................................67
20. Frequency distribution of sample households by efficiency group...................................68
21. Maximum likelihood estimates of the inefficiency effect model......................................71
22. Mean technical efficiency (TE) and exchange labor use (ELU) by family size................73
23. Mean technical efficiency level of sample households by age groups..............................75
24. Mean technical efficiency by educational levels and mean schooling of households.......75
25. Mean technical efficiency level by experience group........................................................76
26. Mean technical efficiency between fertile and infertile land operators.............................77
27. Technical efficiency and plot distance relationship...........................................................79
xii
 LIST OF FIGURES
Figure
Page

1: Illustrations of TE, allocative and economic efficiencies, adapted from Farrell 1957.........11
2 Potential Malt Barley Growing Woredas in Amhara Region................................................27
3 Actual and potential malt barley production of sample farmers...........................................69

xiii
 LIST OF TABLES IN THE APPENDICES

1. Participant KAs and Farmers of the Woreda in 2010 malt barley production program......91
2. Man Adult Equivalent conversion factor for labor use........................................................91
3. Conversion factors used to estimate Tropical Livestock Unit (TLU)..................................91
4. Individual technical efficiency estimates of Sample farmers...............................................92
5. Distribution of sample farmers by technical efficiency levels.............................................93
6 . Actual vis-a-vis potential malt barley productivity............................................................93
7. Variance Inflation Factors of Dependant Variables of the study..........................................94

xiv
 TECHNICAL EFFICIENCY ANALYSIS OF MALT BARLEY
PRODUCTION: THE CASE OF SMALLHOLDER FEARMERS IN
DEBARK DISTRRICT, NORTH GONDAR ZONE OF THE AMHARA
REGGIONAL STATE

ABSTRACT
Ethiopia currently imports about 69% of its malt barley requirement. The volume of import
and related hard currency demand increase with the capacity expansion of the existing 6
breweries and beer production operation of the new breweries which are under
establishment. Recent effort has been started in the country to substitute the increasing barley
importation through smallholder farmers’ domestic production. This study analysed the
technical efficiency of smallholder farmers growing malt barley in Debark Woreda using
cross sectional data collected from 120 sample households in 2010 cropping season. Cobb-
Douglas functional form with maximum likelihood estimation method was used in a single
estimation procedure to estimate the technical efficiency. The result shows mean efficiency of
0.805 and thus the existence of about 19% technical inefficiency in the production of malt
barley. The significant result of the, γ (gamma) ratio in the analysis also indicates the
importance of studying technical efficiency in the Debark Woreda malt barley production
system. The implication of the result is that efficiency improvement efforts could lead to up to
19% productivity gain with the same level of resource and technology use. The results also
revealed that malt barleyarea; horsepower in plowing, value of credit inputs and number of
plowing are the determinants of the production level. Furthermore, the sum of these
coefficients is greater than one (i.e. 1.16) showing some increasing returns to scale in malt
barley production. The MLE result has also indicated determinants of efficiency/inefficiency.
The significant negative coefficients of age, education, malt barley experience, soil fertility,
and livestock holding implied that efficiency improves with increase use of these inputs. On
the other hand, family size, age square and plot distance appeared with positive coefficients,
denoting the increment in these factors lead to diminishing technical efficiency. In conclusion,
two important policy implications can be drawn from the study: The significant determinant
variables identified, constitute instruments that can be manipulated to improve technical
efficiency of smallholder malt barley producers. Besides, malt barley farmers can also take
the advantage of scale economies; linked to more than proportional increase in output to a
proportional increase of inputs, in addition to pure technical efficiency improvement efforts in
promoting the started malt barley production in the region.

xv
 1. INTRODUCTION
1.1. Background

Ethiopia, a country of 73,918,505 population with a growth rate of 2.6 percent per annum
(EPCC, 2008), and an area of 1.12 million square kilometers is endowed with high
environmental diversity. Its location and geographical formation has made her to enjoy both
the temperate and tropical climates which have become a base for its high agricultural
potential and wide range of life forms, both flora and fauna with significant diversity and
many endemic lives (ARARI, 2006)

Consequently, agriculture is the main stay of the economy and a livelihood base for about 85
% of the entire population. In 2006, the sector accounted for 44% of the GDP, 76.9% the
export earning, 80% of the labor force employment and about 70% of the raw materials for
domestic industries (CSA, 2007). Although the contribution of agriculture has started
declining slowly, still it is the predominant sector of the national economy.

Ethiopia’s agriculture, despite its significance and the country’s rich and diversified climate,
altitude, soil and water resources, remains backward and undeveloped. It is characterized by
low productivity, smallholder subsistence farming, and instability of production mainly
related with rain fed farming and traditional and primitive production system. As a result the
country has been a scene of poverty and persistent food insecurity.

Rural poverty is further compounded by extremely small and continuously declining size of
landholding. In the highlands, the average landholding has fallen from 0.5 ha in the 1960s to
only 0.2 ha by 2005 and accompanied by a lower marginal productivity of labor that is
estimated to be close to zero (World Bank, 2007).

Towards addressing the agricultural economic development problems, the Ethiopian

government has embarked on the implementationof ’’The Agricultural and Rural
Development Policy”. The main aim of the policy is to achieve a rapid and sustainable
economic development that would benefit the people at large and relieve the nation from food
aid as well.

The policy states that rapid and sustainable economic development would be ensured through
agricultural led and rural centered development. The trade and industry sectors will grow
faster in complementary with the agricultural sector. Progress in the agricultural sector
stimulates the growth in trade and industry by supplying raw materials, creating opportunities
for capital accumulation, and enhancing domestic markets (FDRE-MoARD, 2002).

Recently, the national economy led by the above development policy is moving from being
primarily agriculture-based towards more reliance on non-agricultural sectors. And there is a
steady annual growth rate in the economy of the country, which makes Ethiopia one of the
best performing countries in Sub-Saharan Africa. However, income disparities among the
population groups are still pronounced (FAO-WFP, 2008). According to the current national
five years (2010-2014), Growth and Transformation Plan (FDRE,2010) still about 7.8 million
smallholder rural households living in drought prone and marginal areas have been identified
as chronically food insecure and to be beneficiaries of the national Productive Safety Net
Program (PSNP) program in the following five years.

In view of these challenges, and opportunities for smallholder farming in development, the
Ethiopian government has continued its emphasis and support on the smallholder farmers as a
central focus of development policies and strategies. In a situation where many are food
insecure, the agricultural and rural development goals remain to bound unattainable and the
success of development and transformation of Ethiopian economy heavily depends upon the
speed with which agricultural growth is achieved (FDRE, 2010).

Developing countries account for about 18% of the total barley production and 25% of the
total harvested area in the world (HSTconsulting, 2007).Ethiopia, the homeland of many
cereals, is also assumed to be the origin for barley and due to this fact there are many
indigenous barley cultivars which are adaptive to ecological situations of the country.
Ethiopia’s barley accounts for 13% of its total cereals production and appreciated for its
2
resilience to adverse weather conditions and to its relatively stable and higher yield in the
highland environments (CSA, 2005). Though barley in Ethiopia is primarily a food grain, it is
also used for brewing beers and sometimes as livestock feeds. It accounts for over 60% of
total food crops in many parts of the highlands (ORDA-Oxfam, 2008).

In the Amhara regional state barley is the fifth most important cereal crop next to teff,
sorghum/corn, wheat and maize in terms of the volume of production. It is produced by the
highland and mid altitude smallholder farmers mainly for consumption. In the year 2010/11
Meher (main) production season, the region cultivated 328269 ha of land and produced
4307416 qt barley. This covers 31.3%of the country’s barley land and 25.3% of the total
barley production, which constitutes 8% and,6.4% of the region’s total cultivated land and
production respectively (CSA, 2010/2011).

Particularly in the Amhara region, malt barley which is a special variety used in brewery
industry can be grown on smallholder basis in many food barley growing areas (HST
consulting, 2007). An estimated cultivated area of about 0.38 million hectares where the
annual rainfall is between 500-700 mm and altitude ranging from 2300-2800 was identified as
potential areas for the crop (CSA, 2004).

There is generally a growing demand for malt barley in the world (ORDA-Oxfam,2008).
Following the establishment of Dashen Brewery plant in Gondar town, a new initiative to
introduce malt barley production by smallholder food barley producers started since 2007.
The production was piloted in 2005 and2006 with selected farmers in 13 selected Woredas in
the region. The aim of the initiative is to substitute imports by encouraging farmers to produce
such high value malt barley. Among these Woredas, Debark is identified as the first potential
Woredafor malt barley production. Accordingly, in 2010 production season, about 885
farmers from 8KAs, which is the largest of all, have engaged in the production of the crop in
theWoreda (DWAO, 2010; ANRS-BoFED, 2010). Volunteer farmers participated in the
production of the crop by entering into a non-binding contractual agreement with Dashen
Brewery factory. Malt barley productionis therefore a new practicein the region as well as
tothe Woredasmallholder farmers.

3
Hence, farmersneed close and integrated support to adopt this new intervention successfully.
To enhance their benefit from the business and make it sustainable,apart from adopting input
technologies, improving their internal efficiency would have paramount importance.
Identifying and promoting proper efficiency improvement interventions calls for determining
the existing efficiency level and identifying the sources of inefficiency of the smallholders in
the production of malt barley.

To this end, this study will try to measure the technical efficiency level and identify the major
sources of inefficiency of smallholder farmers who participated in the production of malt
barley in DebarkWoreda. This could provide findings that are relevant towards strengthening
the newly started malt barley production efforts and achieving sustainable production system.

1.2. Statement of the Problem

Ethiopia and food insecurity have been tied up together until they seem theinseparable sides
of a coin. Several national efforts have been made to achieve food security at household level.
Nevertheless, the issue is still a priority development agenda for the country and many rural
households.

Ethiopia is one of the largest food aid recipientcountries annually receiving about 0.7 million
metric tons over the past ten years (Breylee et al., 2007). Although the Amhara National
Regional State (ANRS) contributes a large amount of the nation’s domestic agricultural
products, it is also one of the regions getting a large amount of food aid. Debark, is one of the
64 chronically food insecure Woredas of the region.

Malt barley accounts for up to 30% of the internationally traded barley in the world (ORDA,
2008). Despite its high production, Ethiopia’s significance as an important barley producing
nation in Sub Saharan Africa is not reflected in its international barley markets. No attempt
has ever been made to enhance the share of Ethiopian barley at the international markets.
Ethiopia can be a significant regional supplier of malt barley to African Beer industries.
Opportunities in those regional trading blocs with large beer markets, such as many member

4
countries in IGAD, ECOWAS, COMESA, and SADCC, where favorable trading terms have
already been negotiated, should be closely researched and monitored (ORDA-Oxfam, 2008).

However, much emphasis has not been given to produce malt barley in the country until
recently. Hence, although Ethiopia could potentially export malt barley, currently 69% of its
demand is being met through import, while only the remaining 31% is supplied by the Assela
Malt factory. In many malt barley potential highland areas of the county, almost the entire
smallholder farmers produce food barley as the domestic market for malt barley has not been
well developed (ORDA-Oxfam 2008).

Despite the fact that malt barley can be grown on about 0.38 million hectares of land in the
ANRS, so far only 7% of the households in the suitable area have grown the crop.The average
productivity is also only10.6 quintal per hectare which is even below the national average of
20 qt/ha (ARARI, 2006). Moreover, there are periods at which smallholder malt barley
producer farmers do not fully sell their marketablesurplus. It was only 69% of the total malt
barley produced in 2009 that was supplied to the market (Fentahun, 2010). In this regard,
among the researchable questions, analysis of factors that determine smallholder farmers’
efficiency is found to be important.

Agricultural output can be increased either through introduction of modern technologies or

byimproving the efficiency of inputs such as labor and management given the existing level
of technology. In other words, productivity can be increased through dissemination of
improved technologies suchas fertilizer and high yielding varieties (HYV) and/or by
improving the productivecapacity of the already employed resources. These two are not
mutually exclusive because the introduction of modern technology could not bring the
expected shift of production frontier, ifthe existing level of efficiency is already low. This
implies the need for the integration of modern technologies with improved level of efficiency.

Most Ethiopian farmers are smallholders and land is a limiting factor of production besides
many other scarce resources. According to the World Bank (2007), in highlands of Ethiopia,
the demand for land has been increasing significantly in the last three decades. Available

5
evidences show that over the years, the total land holding per household is becoming smaller
and smaller. Given the rapidly growing population and consequent degradation of natural
resources, the opportunity to increase smallholders’ productivity through area expansion is
limited. Hence, especially in developing agricultural economies where resources and
technologies are scarce, measuring efficiency is highly important to improve production and
productivity with a given level of resources. The presence of inefficiency not only limits the
gains from the existing resources, it also hinders the benefits that could arise from the use of
improved technological inputs.

However, a study by the World Bank (2006) as cited in Byrelee et al., (2007) indicates that
farmers in Ethiopia are generally operating at low level of technical efficiency, producing
only about 60% of their potential output. More over, malt barley production in the ANRS is a
recent development intervention, and there has been no previous technical efficiency study on
the crop.

1.3 Objectives of the Study

The overall objective of this study is to analyze the technical efficiency of smallholdermalt
barley producers in Debark Woreda of North Gondar Zone.

The specific objectives are:

1. To measure the level of technical efficiency of malt barley producers in the study area;
and
2. To identify factors that causes efficiency differentials among farmers.

1.4. Scope and limitations of the Study

The study focused only on technical efficiency of malt barley production in one Woreda,
using cross sectional data of the 2010 production year collected from 120 sample households.
Some drawbacks of this study arose due to shortage of budget and time. Neither the allocative
and economic efficiencies in malt barley production nor the technical efficiency of

6
households in other crops production was studied.Besides, the data collection was alsoheld
late on March after the end of the cropping year on December. This might have some impact
on the memory of farmers and hence on the quality of data although all the possible care was
taken to overcome the problem during data collection. The other limitation is related with the
methodology used.The efficiency scores of the frontier method are only relative to the best
farms in the sample. Thus, efficiency score might be reduced if more efficient farmers from
other area were included in the study. Likewise, it could increase if less efficient farmers were
included (Coelli et al., 1998).

Hence, the results that will be drawn from the study denotes only about technical efficiency of
malt barley crop. The result may not be even useful to malt barley producers outside the study
area, which may not have similar production process, and for other crops production of the
sample households. Furthermore, the use of cross-sectional data in technical efficiency
analysis reflects farmers’ circumstances on the cropping season. This may be affected by the
specific climate of the year, as agriculture in the country is much vulnerable to externalities.
As a result, the use of cross-sectional data may affect cross year use of the findings sinceit
doesn’t show the inter-temporal differences in efficiency levels of households, although it can
portray useful generalizable relationships pertaining to the malt barley production system.

1.5. Significance of the Study

Attainment of economic growth and food security are widely perceived as the two
fundamental contributions of agriculture to the population of developing countries. Increase in
productivity per unit area is becoming a more important part of agricultural growth as putting
additional lands under cultivation becomes more and more difficult.

An empirical investigation of farm specific efficiency is still very important not only because
of some factors that influences efficiencies of a given farm could be peculiar to the farm , but
also the extent of influences of these factors change over time. Thus examining the efficiency
level of a farm in a given period helps in identifying and taking the timely corrective
measures to benefit more from efficiency improvement efforts, which could vary through

7
time. In other words, specific farm efficiency study helps to determine the level to which
farms are using the existing technologies efficiently; the potential for raising output with the
existing technology; and eventually the possibility to raise productivity by improving both
efficiency and technology adoption. The proper utilization of the existing technologies will
make the development efforts fruitful. Hence, improving farmers’ efficiency is an alternative
source of growth of the agricultural sector.

In view of this, studying the efficiency of the emerging smallholder malt barley producers that
are switching from food barley production in Debark Woredawith the aim to supply malt
barley to the maltfactory that is under establishment in Gondar town primarily to feed Dashen
brewery plant will have three practical importances for the local farmers and the country.
1. It enables to identifying and promoting the necessary actions to improve the technical
efficiency of smallholder farmers engaged in malt barley production and thereby
increase the production and supply to the factory
2. Growing malt barley production domestically will also help substitute the current
huge malt barley importation and save hard currency to the nation, and
3. Improve/raise smallholder participant farmers’ income and livelihoods.

Therefore, this study will be helpful for the participant farmers, development organizations
and policy makers intending to achieve the above-mentioned benefits. The findings of this
study will also be applicable in similar situation in other parts of region. Moreover, it may
help as a benchmark for further studies on similar problems.

1.6. Organization of the thesis

This research paper is organized in five chapters. Chapter two contains the literature review
part. Chapter three deal with the methodology used in the study. Chapter four presents the
results and discussion of the research. Finally, chapter five summarizes and concludes the
study by forwardingrecommendations and policy implications.

8
 2. LITERATURE REVIEW

2.1. Concepts of Efficiency in Production

The main production decision in economics centers on questions what to produce (i.e., which
product or combination of products), how much to produce (the levels of output) and how to
produce (the combination of inputs to use). The decision making unit is the firm which is
defined as “distinct agent specialized in the conversion of inputs into desired goods as
outputs”

In micro economic theory, a production function represents the maximum output attainable
from a given set of inputs as it assumes technology as given. In reality, however, for various
reasons some firms fail to produce the maximum possible output and hence there are
inefficiency differentials among firms. A production frontier represents the maximum
possible output for any given set of inputs setting a limit or frontier on the observed values of
dependent variable, in the sense that no observed value of output is expected to lie above the
production function (Hamed and Mohmoud., 2008). But some firms produce below the
frontier and hence they are said to be technically inefficient while some others produce on the
frontier line and they are said to be efficient. The extent of its inefficiency of a farmer is
judged by the level of output the farmer produced vis-à-vis the output of those farmers that
used equal inputs but still have obtained the maximum output possible at that level of inputs.
In other words, their level of inefficiency is measured by the distance they are far from
frontier relative to the total distance of the frontier from the origin. The ratio of actual output
of a farmer under consideration to frontier output shows the relative efficiency of that
farmer.

The area of efficiency analysis has become the central issue of performance analysis since the
groundbreaking works of Farrell in1957. Technical efficiency (TE) can be defined as the
ability of a decision-making unit (e.g. a farm) to produce the maximum possible output at a
given set of inputs and technology. In other words, given the technology, TE is the ability of a
household to produce on production frontier. Any feasible points below the frontier line are

9
all technically inefficient. However, technical efficiency accounts for only physical inputs and
outputs and does not account for price of inputs and outputs, which deals with allocative
efficiency. Allocative efficiency (AE) is the ability of a firm to produce a given level of
output using cost minimizing input ratios. The product of technical efficiency and allocative
efficiency is simply an economic efficiency.

The firm’s output can be increased through one of the three factors: Through increased
quantity of inputs, through increasing productivity of inputs and the combination of changes
in quantity of inputs and productivity. Therefore, efficiency is a central issue in production
economics helping as a guide for allocation of resources (Farrell, 1957). Productivity
improvement is achieved in two ways; one can either improve the state of technology by
investing on new seed or fertilizer or plough which can shift production frontier upward; or
use of the existing technology more efficiently which enables the firms to operate closely to
the existing frontier (Coelli, 1995).

Many scholars use productivity and efficiency interchangeably and both are considered as the
measure of performance of a given firm. However, these two interrelated terms are not
precisely the same (Coelli et al., 1998). In simple terms,productivity is the quantity of a given
output of a firm per unit of input. Technical efficiency is that part of efficiency which explains
the physical performance of a firm that measures the relative ability of a firm to get the
maximum possible output at a given input or set of inputs.

As 1957, Farell was the first who studied the productive efficiency empirically. He
decomposes the firms efficiency as technical efficiency which reflects the ability of a firm to
obtain maximum output from a given set of input, and alloctive efficiency which reflects the
ability of a firm to use the inputs optimum proportion with their given respective prices. The
two measures of efficiency combined to provide a measure of total economic efficiency
(Coelli, 1995).

Farrell illustrated his idea using a simple example involving firms, which use two inputs
(X1and X2) to produce a single output (Y) under an assumption of constant returns to

10
scale(CRS). The assumption of CRS permits all the relevant information to be presented in
asimple isoquant diagram. The isoquant SS’ in the following Figure 1 represents the
variouscombinations of two inputs X1 and X2 that a perfectly efficient firm might use to
produce unitoutput.

X2

S ●P

A ●Q
R
●
Q/
● S/
A/
O X1
Figure 1: Illustrations of Technical, allocative and economic efficiencies, adapted from
Farrell, 1957

If a given firm uses quantities of inputs, defined by the point P, to produce a unit of output,
the technical efficiency of that firm is defined by the ratio OQ/OP, which is the
proportionalreduction in all inputs (QP/OP) that could theoretically be achieved without any
reduction inoutput. The point Q represents an efficient firm using the two inputs in the same
ratio as P. Itproduces the same output as P using only a fraction OQ/OP as much of each
input. The pointQ’ is also technically efficient because it lies on the efficient isoquant SS’.
Allocative efficiency also be determined if the input prices are given which are represented by
the slope of the line AA’ (equal to the ratio of the price of the two inputs) in the above figure.
The allocative efficiency of the firm operating at P is defined to be the ratio OR/OQ, sincethe
distance RQ represents the reduction in production costs that would occur if productionwere
to occur at the point where it is allocatively (and technically) efficient. Although pointQ /and Q
are technically efficient, point Q’ is allocatively efficient than point Q. Since thecosts of

11
production at Q’ is a fraction of OR/OQ, firm Q has to reduce its cost by the fractionof
RQ/OQ to be allocatively efficient, keeping its technically efficient constant.

The total economic efficiency is defined to be the ratio of OR/OP, where the distance RP
canalso be interpreted in terms of a cost reduction. The product of technical and
allocativeefficiency provides the overall efficiency i.e. economic efficiency = OQ/OP *
OR/OQ =OR/OP.

2.2. Models for Measuring Technical Efficiency

Earlier, various techniques of measuring productive efficiency were being carried out with
average production function. But the modern empirical measuring of efficiency was started,
and there have been many developments and refinements in the methodology of efficiency
measurements since the groundbreaking work of Farrell in1957 (Aigner et al., 1976).

Technical efficiency measurements are carried out using frontier methodologies, which shift
the average response functions to the maximum output or to the efficient firm. Many
empirical studies of efficiency were devoted in analysing what impact a given model
specification has on the efficiency measurements. Various issues pertaining to model
specification are still debatable. The selection of specific frontier model depends upon many
considerations such as the type of data, cross-sectional or panel data, the underlying
behavioral assumptions of firms, the relevance to consider and extent of noise in the data and
the objective of the study (Battese and Broca, 1996).

The frontier methodologies have been widely used in applied production analysis. This
popularity is evidenced by the proliferation of methodological and empirical frontier studies
over the last two decades. Despite these wide arrays of applied work, the extent to which
empirical measures of efficiency are sensitive to the choice of methodology remains a matter
of controversy (Thiam et al., 2001).

These frontier methodologies are broadly categorized as parametric and non-parametric

frontier models. The parametric frontier model may further be classified into deterministic

12
frontier model and stochastic frontier models. The parametric models are basically estimated
based on econometric methods and the non-parametric technical efficiency model, often
referred to as data envelopment analysis (DEA), involves the use of linear programming
method to construct a non-parametric 'piece-wise' surface (or frontier) over the data (Coelli et
al., 1998). The following review focuses mainly on these two broad categories of frontier
models.

2.2.1. Non parametric frontier estimation methods

One of the methods of measuring efficiency in agricultural production is the non parametric
approach of the Data Envelopment Analysis (DEA). It was first introduced by Fare et al.
(1985) based on the Farrell’s approach. He relaxed the restrictive assumptions of constant
returns to scale and the strong disposability of inputs, which were the major criticism of the
method (Llewelyn and Williams, 1996). Minimum assumptions of monotonicity and
convexity of the efficient frontierare required for this DEA frontier methodology (Banker et
al., 1984).

DEA is an evaluation method particularly adapted to comprise a set of multiple indicators into
overall performance. It enables frontier estimation with the use of non-parametric
programming models leading to a ranking of all unit observations on the basis of efficiency
scores. The focus is not on the estimation of an average technology production function used
by all units analyzed, but to identify the best practicing units. The best-practice production
frontier is constructed and all units of analysis are related to this frontier. DEA is based on the
simple notion that an organization that employs less input than another to produce the same
amount of output can be considered as more efficient. The efficiency frontier is constructed of
linear segments that join up those observations with the highest ratio of output to input. The
resulting frontier thus ‘envelops’ all the other observations (Jema,2008; after Frank and
Thanda,1999).

The main advantages of the DEA approach are that it can handle multiple inputs and multiple
outputs and that it avoids the parametric specification of technology as well as the

13
distributional assumption for the inefficiency term. However, because DEA is deterministic
and attributes all the deviations from the frontier to inefficiencies, a frontier estimated by
DEA is likely to be sensitive to measurement errors and other noise in the data (Coelli, 1995).
In addition nonparametric frontier methodology may overstate inefficiencies and hence
outliers may have profound effect on the magnitude of inefficiency (Ibid, 1995). The other
disadvantage of DEA is that it is not possible to test the hypotheses regarding the existence of
inefficiency and the structure of the production technology that were possible in stochastic
production frontier analysis.

2.2.2. Parametric frontier estimation methods

Parametric frontier model can further be classified into deterministic and stochastic frontier
methods. Typically, both models use econometric techniques to estimate the parameters of pre
specified functional forms. However the deterministic model assumes that any deviation from
the frontier is due to inefficiency, while the stochastic approach allows for statistical noise
(such as measurement error, weather, industrial action, etc…)which are beyond the control of
the decision making unit(in this case the household head).

2.2.2.1. Deterministic frontier model

The first deterministic frontier function was estimated by Aigner and Chu in1968 by assuming
a function giving maximum possible output as a function of certain inputs (Aigner, et al.,
1976). Coelli (1995) stated that they consider the estimation of parametric frontier production
function using a Cobb-Douglas production function (in log form) for a sample of n
households as:

ln ( y i ) =F ( x i ; β ) −ui i = 1, 2 . . . n (1)
Where i – is the number of sample households in the study
ln ( y i ) – is the natural log of (scalar) output of the ithousehold;
x i – is a vector of input quantities used by the ithhousehold;
β – is a vector of unknown parameters to be estimated;

14
 F(.) – is an appropriate production function; and
ui - is a non-negative random variable associated with technical inefficiency in
production of firms in the industry involved, and

where
n

∑ u iis minimized, subject to the constraint ui > = 0, i=1, 2… n

i=1

The ratio of observed output of the i th firm, relative to the potential output defined by the
estimated frontier, given the input vector X i was suggested as an estimate of the technical
efficiency of the ith firm

yi
TE= =exp ⁡(u i) (2)
exp ( F ( x i ; β ) )

The inefficiency term, exp (-ui) is bounded between the values of 0 and 1, thus restricting the
output of the ith firm be less than, or equal to the predicted value of the deterministic frontier
, F ( x i ; β ).

The main criticism of the deterministic frontier model is that it does not account for possible
influence of measurement error and other noise upon the shape and positioning of the
estimated frontier (Coelli et al., 1998). All observed deviations from the estimated frontier are
thus, assumed to be the result of technical inefficiency. Therefore, the method sums up all the
effects of exogenous shocks together with measurement errors and inefficiency.

As a solution for such problem (associated to outliers in the above deterministic approach),
Timmer (1971) developed the probabilistic frontier. He imposed Cobb-Douglas (CD) type
production function and estimated the parameters using linear programming by discarding the
outliers until the parameter values stabilize. Therefore, the presence of high random errors on
a data leads to exaggeration of the inefficiency estimates in deterministic model as compared
to other models, which takes random errors in to account.

15
Neff et al. (1993) and Apezteguia and Garate (1997) in their comparative studies of frontier
models found that deterministic frontier model generate higher technical inefficiency indices
than stochastic frontier model. This is because the observed input and output quantities may
not only deviate from some imaginary frontier through inefficient management (as in a
deterministic interpretation) but also through noise.

Thiam et al. (2001), on the other hand, found that models which used stochastic frontier did
not significantly generate different technical efficiency indicators than that of deterministic
models. Here they indicated that their findings contradicted with their a priori expectations
that inefficiency scores for the firms would be high for deterministic models than stochastic
frontiers. They argued that the proportion of deviation might be very small as compared to the
deviation due to inefficiency.

However, although the theoretical and empirical findings revealed that the smaller the random
noise present on the data in question, the closer will be inefficiency estimates between the
deterministic and the stochastic frontier models, mostly, empirical, production efficiency
studies in agriculture are treated by using stochastic frontier model. This is due to the fact the
very nature of the agricultural outputs are easily susceptible to natural hazards, climatic
conditions, and measurement errors that could contribute to the existence of noises in the data
(Coelli et al., 1998).For similar reasonsas assessed by Fekadu (2004) other researchers
including Mohammad and Premachandra (2003), Thiam et al. (2001), Ajibefun (2002), Parikh
and Shan (1996) also agree that stochastic frontiers are likely to be more appropriate than
DEA in agricultural applications of 21efficiency analysis.

2.2.2.2. Stochastic frontier model

In order to overcome the problem associated with random error in the deterministic approach
an alternative estimation method called Stochastic Production Frontier Approach was
independently developed by Aigneret al.(1977) and Van den Broek (1977) (Coelli, 1995). The
application of the model for efficiency analysis was first indicated in the work of Aigner et al.

16
(1977) to the US agricultural data and by Battese andCorra(1977) to the pastoral zone of
Australia. Similarly Bravo-Ureta Pinheiro (1993) offered a comprehensive review of the
application of the model inmeasuring efficiency of agricultural producers in developing
countries. The stochastic Production Frontier was developed by adding a symmetric error
term (vi) to the non negative error term of the equation in (1) as:

ln ( y i ) =F ( x i ; β ) + v i−ui i = 1, 2, . .., n (3)

In this equation the s’ are assumed to be independent and identically distributed random
errors following a normal distribution with zero mean and variance σ v2. The random error
accounts for measurement error, and other external factors such as climatic changes in

production process which is out of the control of the producer; whereas the s’ are the
technically inefficiency terms which are associated with the technical inefficiency of the
firms.

Estimation of the technical efficiency of individual farmers usually requires specifying the
stochastic production frontier. The parameter of this model is preferably estimated by the help
of an econometric procedure known as the Maximum Likelihood Estimation (MLE)
approach, given suitable distributional assumptions for the error terms As it was suggested in
similar way by Afrait (1972) the corrected ordinary least squares(COLS) method could also
serve as parameters estimation technique for stochastic model. However, the former is
asymptotically more efficient than the later.

The maximum likelihood estimates of the production function parameters and the technical
efficiencies of the sample firms are computed using LIMDEP, FRONTIER4.1version, and
STATA software packages(Coelli,1995).

By assuming half-normal distribution for the technical inefficiency effects (u i’s), Aigner et al.
(1977) proposed the log likelihood function for the model on the equation (3). They used λ
parameterization to express the likelihood function, where λ is the ratio of the standard errors

17
of the non-symmetric to symmetric error terms (i.e.
λ =σ u / σ v ). However, due to the
reason that λ could be any non-negative value while γ ranges from zero to one where it better
measures the distance between the frontier output and observed output and basically separates
the effects of noises from technical inefficiency. Given a symmetric normal distribution for v i
and half normal distribution for ui, Battese and Corra (1977) proposed that the γ
2
u
γ =σ /( σ +σ )
parameterization, where v2 u2 , be used instead of λ.

Accordingly, the maximum likelihood estimates of the parameters of the frontier model,
following the works of Battese and Corra (1977) and Battese and Coelli (1992), are estimated
from the log-likelihood function expressed in terms of γ parameterization as follows:

[[] ] [ [ √ ]]
εi √ γ
N N
−N π γ 1
+ ln σ + ∑ ln 1−ф 2∑ i
2 2
ln ( L )= ln 2
− ε (4)
2 2 i=1 σ 1−γ 2 σ i=1

Where; ε i=lnY −X i β is the residual of (3);

N is the number of observations;
Φ (.) ,is the standard normal distribution;
σ2= σ2v + σ2u and γ = σ2u/ σ2v + σ2u are the variance parameters of the model. Here, the γ
parameter has a value between “0” and “1”. A value of zero indicates deviations from the
frontier are entirely due to noise, while a value of one would indicate that all deviations are
due to technical inefficiency.
.
Minimization of the function on equation (4) with respect to the parameters β, σ 2, and γ and
solving simultaneously all the first partial derivatives of the function by equating to zero
produces the efficient maximum likelihood estimates of β, σ2, and γ.

Once the parameters of the model are estimated, the next step involves the computation of
firm-specific technical efficiency scores. The process, however, needs decomposing the total

disturbance term (
ε i ) into the random error (v ) and the inefficiency component (u ). Jondrow
i i

et al. (1982) and Kalirajan and Shand (1986) suggested that u i’s can be predicted by
18
conditional expectation of ui given the value of total disturbance term;
ε i=v i−ui .They
recommended that the technical efficiency of the i th firm can be estimated by using
1−E [ u i /ε i ]
. The basic distinction between the works Jondrow et al. (1982) and Kalirajan and
Flinn (1983) are the earlier used the λ parameterization, while the latter applied γ
parameterization.

On the other hand firms could produce below the frontier level of output for a number of
reasons. Many empirical studies showed that a range of firm- specific characteristics,
institutional and environmental factors cause technical inefficiencies. The significance of
these factors can be analyzed using the following technical inefficiency effect model (Battese
andCoelli, 1995);
ui=z i δ+ wi (5)
Where;

is the technical inefficiency of the i th firm/farmer and assumed to be independently

distributed random variable.

stands for vector of firm specific variables which may affect the efficiency of the i th firm;
δ denotes a vector of parameters to be estimated; and

s’ are unobserved random variables and assumed to be independently distributed and

obtained by truncation of the normal distribution with zero mean and constant variance.

Coelli (1995) indicated the stochastic model has the advantage of addressing the problem of
noises: permitted the estimation of standard errors and tests of the hypothesis which the
deterministic model lacks. However it has also some limitations; there is no priori justification
for the selection of any particular distributional form for the u i. But most of the empirical
studies used the half normal distribution.

2.3. Empirical Studies of Efficiency Estimation and Its Determinants in Ethiopia

A number of efficiency analysis have been conducted by different researchers both from
abroad and in the country with the aim of identifyingthe sources of inefficiencies and their

19
policy implications so as to improve the future development endeavors through enhancing the
prevailing technical efficiencies. Most of the studies have specified the Cobb-Douglas type
production function and commonly estimated parameters by using the MLE procedure and
also used the SPF methodology. The survey by Coelli (1995) on the application of frontier
model in the world indicated that there has been vast number of agricultural studies that used
frontier methodologies. Amongst the 38 studies he selected without exhaustive search, 3 of
the studies employed DEA, 7 estimated with deterministic, 24 used stochastic frontier and 4
estimated both deterministic and stochastic frontiers. The survey indicated that stochastic
frontier modeling is the most dominant; and most of them used the estimation of single
equation using cross sectional data.

2.3.1. Empirical studies on efficiency level

To measure technical efficiency of fertilized farm employing stochastic frontier model, Abrar
(1995), used cross sectional data from three villages in Ethiopia. He found that expanding the
output of the average farmer by more than 40 percent was possible, if appropriate measures
are employed to improve internal efficiencies. The result of his investigation enabled him to
reach two basic conclusions. First, the traditional Cobb-Douglas production function is not a
suitable model to explain the production behavior of the farms under consideration. This
means that technical inefficiency is one of the main characteristics of agricultural production
for these farmers. Second, technical inefficiency does not only vary between villages but there
is also a high variation across farmers within a village.

Croppenstedt and Abbi (1996) studied TE of cereal producing farmers in Ethiopia using five
PAs from three regions (Oromia, Amhara and SNNR) without intending to know the causes
of variation in technical efficiencies among farmers. Farmers were found to be on average
72% technically efficient with the range of 32 – 92%.

Seyoum et al. (1998) compared the efficiency level of Sasakawa-Global 2000 (SG2000)
project participants and non- project participant farmers in maize production at Kersa and
Kombolcha Woredas of eastern Harerge. A separate stochastic Cobb-Douglas production

20
functions were specified and 20 sample farmers were considered from each group. The result
has shown that the average technical efficiency for the SG2000 project participants and non
project maize producer farmers were 93.7% and 79.4% respectively.
Mohamed (1999) has also analysed technical efficiency of wheat, barley and overall crop
production in Gedeb-Hasasa Woreda of Oromya region. He used stochastic frontier
estimation techniques with both translog and Cobb-Douglas functional forms and found that
the Cobb-Douglas functional form has been appropriate for his data in the wheat, barley and
overall efficiency measures. The researcher has found average technical efficiency results of
51% for wheat, 57% for barley and 53% for overall crop production.

Temesgen (2003) analyzed technical efficiency of Awi Zone farmers in the production of
irrigated potato in the Amhara National Regional State. He applied the Translog stochastic
frontier production function, using cross-sectional data collected from randomly selected 80
farmers in four Woredas of the zone. Technical efficiency of farmers was estimated
independently for the farms under modern irrigation schemes and under traditional irrigation
schemes. The mean level of technical efficiency was 77 and 97 percent, respectively for
modern and traditional schemes. The findings indicate that improving the level of efficiency
could raise productivity under modern schemes, while improving productivity under
traditional schemes need introduction of new technology.

Fekadu (2004) also researched the technical efficiency of farmers in the production of wheat
at Machakel Woredain east Gojjam zone of Amahara regional state of Ethiopia. Cross-
sectional data collected in 2003 production year, and the Cobb-Douglastype of production
function were employed to estimate the efficiency of smallholder wheat producers in the area.

The results indicated that there is inefficiency in the production of wheat in the study area.
The estimated Cobb-Douglass stochastic production frontier with inefficiency variables has
shown that the mean technical efficiency of farmers in the production of wheat was71%
implying further productivity improvement by 29% with the given level of inputs and
technology. The relative deviation from the frontier due to inefficiency is 92% .There is also
high efficiency difference among smallholder plots producing wheat. Hence, if inputs are

21
used to their maximum potential there will be considerable gains from improvement in
technical efficiency.

Haileselassie (2005) undertaken study to know the technical efficiency of smallholder farmers
in sorghum production in Raya-Azebo Woreda,south Tigrayof theTigray national regional
state. It was based on cross-sectional data of 120 sorghum producing farmers collected during
2004/2005 production season. Individual levels of technical efficiency were estimated using
Cobb-Douglass production functional form. He also found the Cobb-Douglas type of
functional form to best fit to the data.

Kinde (2005) analysed technical efficiency in maize production using Trans log stochastic
frontier model in Asosa, Benshangul State .The survey research has shown that the technical
efficiency score varies from 3% to 97% and 67% mean level of efficiency. The technical
efficiency differences revealed in most of the farmers has arisen not so much due to chance
factors, which is beyond the control of farmers but rather from differences in the farmers’
ability in using the best practices.

Maize production technical efficiency analysisin Mecha Woredaof West Gojjam Zone of
Amahara regional state of Ethiopia was also undertaken by Aynalem (2006). A single stage
estimation technique for parameters of stochastic production function and the technical
inefficiency function together with Cobb-Douglas production functional specification were
used to analyze the cross-sectional data collected from 120 sample farmers in the
productionseason of 2004/2005.The results revealed that technical efficiency score
differential among farmers vary from33%to95% with the mean technical efficiency level
of78% signifying the room for 22% possible improvement of the then existing efficiency
level of the given input and technology .

Jema (2008) studied the economic efficiencies of smallholders farmers in mixed crop and
market driven vegetable production systems in two Woredas of eastern Ethiopia employing
both DEA and SPF methods. The mean technical, allocative and economic efficiencies were
found to be 91, 60 and 56 percents, respectively. This indicates the existence of substantial

22
allocative and economic inefficiencies of production in the study areas. An econometric
analyses based on Tobit model indicate that asset, farm size, extension visits and family size
were the significant determinants of technical efficiency, whereas asset, crop diversification,
consumption expenditures and farm size have significant impact on allocative and economic
efficiencies. He used the Tobit analysis to see the inefficiency differential and found that the
inefficiency in the sample is on average 44%, mainly as a result of allocative inefficiency,
which is attributed to low asset ownership and farm size, high consumer spending, crop
diversification and barriers to the flow of labor between farm and off/non-farm activities.

2.3.2. Empirical studies on determinant of technical efficiency

Most empirical studies of technical efficiencies are targeted at identifying the sources of
inefficiencies that exist behind certain level of prevailing average efficiency of farmers. As
explained by Fekadu (2004) measurement of efficiency per se has no policy implication and
practical development relevance unless the real causes of efficiency differences are identified.

From the study of Mohammed (1999) analysis of efficiency determinants revealed that
livestock holdings significantly the technical inefficiency of the farmers supporting the
theoretical complements between livestock production and general crop cultivation facilitates
productivity increases. On the other hand, contact with extension agents had unexpected sign.
That is farmers who frequently communicate with extension workers had lower efficiency of
49% while farmers without any extension contact had 56% level of technical efficiency.

From Fekadu(2004) the estimated SPF model together with the inefficiency parameters shows
that fertility status of the plot, age, age square, land ownership status, training, number of
plots (fragmentation) family size and livestock significantly determine efficiency level of
farmers in wheat production in the study area. The negative coefficient of fertility status of the
plot, age, type of landownership, training and extension means these factors are important in
determining the existing efficiency of farmers positively and significantly. While the positive
coefficients ofage square, family size, fragmentation of farmland and livestock indicate that
the increments in these factors increase inefficiency.

23
Haileselassie (2005) in his study of the technical efficiency of sorghum production of the
Raya-Azebo Woreda smallholder farmers has come up with the negative impacts of off-farm
employment on the efficiency level. Farmers engaged in off-farm activities were operating
under lower level of technical efficiency. Those farmers who engaged in off-farm activities
tend to operate their farm activities relatively late because of the competitive nature of the off-
farm employment type with the crop production activity

Jema (2008) revealed the variables such as access to capital markets, assets, level of consumer
spending, and family size, and crop diversification are determinants of economic efficiency.
The lower economic efficiency scores for market oriented vegetable production results of the
study was probably to be associated with limited access to capital markets, high consumer
spending and large family size.

Amadu (2007) conducted a study with a view to identifying factors determining the technical
efficiency of Arabica coffee producers in Cameron. His study disclosed the existence of a 10
percent technical inefficiency. Access to credit and education levelof farmers werefound to be
the only significant determinants of technical efficiency out of the ten hypothesized efficiency
determinant variables.

Generally, the above assessed empirical studies in single, multiple or overall production cases
have shown the presence of technical inefficiency of farmers. The levels of inefficiency vary
not only among farmers in a given study but also from place to place and through time
depending on the localities, agro-climatic, socio-economic conditions and the various
managerial inputs put in production practices. Thus knowing the current level of efficiency
and causes existing behind for low efficiency of a given research area in a given period of
time is crucial. It will help drawing relevant and timely development interventions that could
have paramount importance in realizing fostered and sustainable development of a given
locality. Malt barley production in the Amhara regional state is a new initiative that has been
started recently through switching smallholder food barley farmers in to malt barley producers
with the purpose of substituting the current ever increasing huge malt barley imports of the

24
country while improving the income of participant smallholder farmers. In the malt barley
production program there is often acontention between producers and contract provider
(Dashen brewery) on the level ofprice set for malt barley purchase. Such a price setting will
have a negative impact on the ongoing malt barley development,if it does not account the
opportunity cost forgone by the households while producing malt barley. Technical efficiency
study in malt barley productionwill thus help mitigating price complaints by the producers
through improving the productivities of the given resources and thereby reducing cost of
production.

25
 3. METHODOLOGY

3.1. Description of the Study Area

The study area is located at 850 km north of Addis Ababa in Debark Woreda in North Gondar
Administrative zone of the Amhara National Regional State (ANRS). The region has an
estimated area of 152.6 thousand sq km (ANRS-BoFED; 2007) with a population of about 17.21
millions (EPCC, 2008) and it is structured into 10 administrative zones including North Gondar.

North Gondar zone covers an area of 44,556 sq km and accounts for about 29% of the total area
of the region. North Gondar has 529 rural and 48 urban Kebeles, and the total population of the
zone is about 2,921,470, of which about 1,481,728 are male and 1439,744 are female(EPPC,
2008). Most of the zone is found in northern part of ANRS. Gondar is administrative center of
the zone, which is found at 175 Km from Bahir Dar and 740 Km from Addis Ababa (ANRS-
BoFED, 2008)

Debark Woreda is one of the 21 Woredas found in North Gondar zone. Debark town (the capital
town of the Woreda) is located at about 100 Kms North of Gondar town. The Woreda has an
area of 1,822 sq km with a population of 160,130 living in 25 rural and 8 urban Kebeles (ORDA-
oxfam, 2008; EPCC, 2008).

The Woreda has an undulated landscape with high altitudinal variations which ranges from an
altitude of 4620 of Ras DashenMountain- the highest peak of the nation- to lowland areas as low
as 500. Dega is more dominant agro-ecology than Woyna-dega and Kola which both comprises
about 46% of the total Woreda.The Woreda is the habitat of different wild lives including Walya
Ibex, which is endemic to Ethiopia. The mean annual rainfall and temperature in the Woreda
ranges between 400-1200mm and between 18-290C, respectively.

26
Mixed crop-livestock farming system is the main livelihood base of the population in the
Woreda. The four major crops grown in the area are malt barley, food barley, wheat and pulses.
Related to this, frost, low crop bio-diversity, poor infrastructure, and low institutional supports
are thought to be the major problems of crop production. The Woredais one of the 9 chronically
food insecure Woredas of the zone that are included in the PSNP national program.

As part of the income diversification scheme, the region designed the introduction and
dissemination of malt barley production technology in 13 selected potential Woredas as shown in
Figure 1 below. Accordingly, Debark woreda is selected an area with huge potential. Of the total
25 rural kebelesof the woreda, 18 Kebelesare identified for theirpotential .The program started as
a pilot in 2005 on 3kebeles. In the 2006/07 production year, out of the total cultivated land of
19033 ha, malt barley covered only 143 ha (about 0.75%) (ANRS-BoFED, 2007).In 2010
production year, 885 households selected from 8 kebeles participated in the malt barley
production program.

27
Figure 2 Potential Malt Barley Growing Woredas in Amhara Region
Source: http://www.mapliberary.org/status/Africa/Ethiopia/Amhara/N.Gondar/index.php

3.2. Methods of Data Collection

3.2.1 Data types and sources

The study usedboth primary and secondary data. The different inputs used in the production of
malt barley in 2010 production year and the corresponding output as well as other agronomic
data were collected from selected sample households using a structured questionnaire. In
addition,key informant discussions held with community leaders and administratorsalso served
as the primary sources of generating data to augmentthe sample household information. Primary
data were also supplemented with secondary data.Information on demographic, agro-ecological
conditions, institutional, social and economic information wereobtained from published and
unpublished sources. Relevant documents of the development centers, Woreda agricultural

28
office, Dashenbrewery, BoA, BoFED, CSA and other relevant sourceswere consulted as
secondary information sources.

3.2.2.Sampling procedure and sample size

The study was conducted in DebarkWoreda, whichis one of the 13 potentialWoredasidentified

for malt barley productioninthe AmharaNational Regional State.Theworeda waspurposively
selected due toits huge potential in malt barley production. About 885 recruited households from
8 highland kebelesof theworeda engaged in the 2010 malt barley production program.
Participant’s interest, availability of suitable land, proper and timely land preparation arethe
major criteria taken in to consideration for households’ enrolment in the program. Female
participants meeting mentioned criteria are given priority although only 41 female-headed
households participatedin the 2010production year. Taking in to account the more or less
homogeneity of the population with respect toagro ecology and wealth conditions 120 samples
that is about 14% of the total malt barley producer households were selectedfrom 3 kebele
administrations for the study.

In order to select the 120 sample households, a three stage sampling was used. At the first
stage,out of the total 8 kebeles participated in malt barley production in 2010 cropping season 5
kebeles that have sufficient number of participants wereidentified as potential sample kebeles
(see Appendix 1for number of participants in each kebele). Then, from the identified 5 kebeles, 3
were selected randomly.In the second stage,a complete enumeration of participant farmers
wasused as a sampling frame (Table 1).In the next step,the numbers of farmers to be selected
was decided on the basis of the proportion of malt barley grower householdsfrom each kebelein
the total number of growers. Finally, 120 sample households were selected using simple random
sampling technique.

Table 1.Sample kebele administrations and sample size selected from each kebele

Malt barley growing households

Name ofkebele Sample size selected
Participant households Percent
Miligebsa 306 56 67

29
Gomia 115 21 25
Kino 125 23 28
Total 546 100 120
Source: Own survey (2011)

3.2.3. Data collection

A designed structured questionnaire was employed to collect primary data from the sample
households. The questionnaire included information on the demographic, institutional services,
resource availability, the different input use and related malt barley output of the sample
households (Appendix part B). Questionnaire pretest was conducted for completeness, relevance,
non-ambiguity, coherence of questions, and unexpected responses that could arise from loss of
memory. Necessary modifications were made based on the feedbacksfrom the pretest.The data
collection process was administered by 5 enumerators who have been recruited based on their
educational background. The enumerators recruited are a minimum of diploma holders.
Additional criteria of having good knowledge about the local farming system and motivation for
the work were also considered in the recruitment.A three days intensive training was given to the
enumerators about methods of data collection, how to approach the farmers and the content of
the questionnaire.
Questionnaire was translated in to local language to eliminate the errors associated with
translation by enumerators that has a consequence on the quality of the data to be gathered.
Moreover, local measurement scales customarily used by farmers such as Timad (to measure
thearea); Madiga, and Chan (to measure quantity) were measured to convert into their respective
standard units. This helps to minimize measurement errors that could arise from variability of
local units.There was close supervision by the researcher so that errors, if any, could be corrected
at earliest possible time.In addition to the structured questionnaire, personal observations,
informal and key informant discussionswere also used.Secondary data were collected from
relevant sources such as books, journals, and reports and other documents of relevant offices.

3.3. Variable Selection, Measurement Units and Hypotheses

30
Based on the empirical evidences reviewed in the area, theoretical ground and lessons learnt
during the survey, various input and inefficiency variables were considered. Since malt growers
were made, to use inputs of DAP and urea fertilizers and improved seed in the form of input-
packageby the malt barley production program, the total costs of the input-package was used as
one input variable. On the other hand, frequency of agronomic practices of plowing and weeding
are strongly recommended practices to increase yield of malt barley. These two inputs were
introduced in the analysis to see the impact of the variables on malt barley production.However,
pesticidesuse was presumed to be an important input variable in the production of malt barley,
the survey revealed thatnone of the sample farmer did use this input and hence it was omitted
from the analysis. Accordingly, input-package, number of plowing, number of weeding, malt
barley land size, labor and horse-power were used as input variables.

3.3.1. Input and output variables

Output:the quantity of malt barley output produced by the sample household during the 2010
production year measured in quintals per unit area was used as the dependent variable.

The following input variables were used as explanatory variables:

Area of land under malt barley cultivation:a continuous variable defined as the size of land
used for malt barley production by each farmer measured in hectares irrespective of their
ownership type.

Labor:It includes human labor used for plowing, seeding, weeding, fertilizer application and
other agronomic practices.The various categories of labor such as child labor, adult men and
women, and aged men and women labor were recorded separately and converted into man-
equivalent labor using standard conversion factor as reported by Storck et al. (1991), (seeTable
2 of Appendix A).

Costs of technological input-package: DAP, urea and seed are given in the form of package for
malt barley producing farmers on credit basis. Though these inputs were provided to farmers in
the form of package, the survey revealed that there were some differences in the actual use of the

31
inputs from the recommended quantity. Thus, the actual cost of inputs was used to aggregate the
three inputs.

Draught power: It is a continuous variable measured by the number of pair of horse-days

employed by the sample farmer in malt barley cultivation. About 8 hours of working period for a
pair of horses was taken as one horse-day. Here horse-day is used as a unit instead of oxen day as
the draught power used in the area is generally horses.

Number of plowing: Represented by the frequency of plowing performed on malt barley plot of
the farmer. The assumption is that as the number of plowing increases, it increases efficiency by
improving agronomic practices.

Number of weeding:Represented by the frequency of weeding performed on malt barley plot of

the farmer.

3.3.2. Inefficiency Variables

After measuring the magnitude of inefficiency of each household, identifying factors causing
inefficiency differentials is an important step for designing better development interventions.
Thus, based on theoretical backgrounds and empirical evidences the following socioeconomic,
institutional and agro-ecological characteristics of the households were hypothesized to explain
efficiency differences observed among farmers:

Family Size: is measured by the total number of family members in the household. The family
size is hypothesized to affect technical efficiency level of the farmers positively. This is because,
as labor is the main input in crop production, a farmer that has large family size could carryout
the required agronomic practices evenin peak production periods. Family size was found to relate
positively to technical efficiency (Kinde, 2005). However, family size was found to relate
negatively to technical efficiency (Mohammed, 1999; Fekadu, 2004).

32
Age of household head and age square: The continuous variable, age of the household head
measured in number of yearsis used as a proxy measure to indicate the general farming
experience, of the sample household.The hypothesis is that as the age of the farmer increases, the
farmer becomes more proficient in the method of production and optimal resource allocation.
However,after certain age limit as farmers get older and older they start to be more conservative
and less willing to adopt technologies as a result of which their technical efficiency decline. To
see the diminishing effect of age on efficiency age square is also considered together to provide
the variable age a quadratic functional form.Accordingly, a positive coefficient for age and a
negative coefficient for age square are hypothesized to capture diminishing efficiency of old age.
If this hypothesis is true we can conclude that middle age farmers are more efficient than others.

Education: This is used as a proxy variable for managerial input.It is measured in years of
formal schooling of the household head.It is hypothesized in this study that education determines
technical efficiency level of the farmer positively. This is because, farmers are expected to
acquire the skill of better management of crops through formal education as they have the
opportunity to process and use information from various sources for making decision to increase
output through efficient input use. Formal education received most of the attention in the frontier
efficiency literatureand itwas found to relate positively to technical efficiency in several studies
(Getu, 1998; Seyoumet al., 1998; Mohammed, 1999; Fekadu, 2004; Kinde, 2005).

Malt barley production experience:Malt barley production is a relatively new practice in the
area.It is a discrete variable intended therefore to capture the contributions of specific
experiences acquired by farmers as they stay longer in the production of malt barley. This could
be particularly important when most of the recommended technologies and practices are less
refined and when the contribution of adaptive learning of farmers through repeated participation
could have positive effect on efficiency. In addition, the longer the farmers participate in malt
production the higher the knowledge and skill they could get through learning by doing and
hence the more they will be efficient. Thus, it is hypothesized that mart barley production
experience affects efficiency positively.

33
Extension contact:This variable serves, as a proxy measure for access to extension services
since all the extension user households, may not get the services every time they demand. It is a
discrete variable measured by the number of visits made by development agents in relation to
malt barley production in the cropping year. It is hypothesized that household contact with DA
will be positively related to technical efficiency. This is because, farmers that have frequent
contact with extension workers will have better access to information and new technology that
would be productively used on their farm. Extension service was found to relate positively to
technical efficiency (Seyoum et al., 1998; Fekadu, 2004 and Kinde, 2005). However, Extension
service was found to be related negatively to technical efficiency(Mohammed, 1999),and
insignificant relationship(Aynalem,2006 and Wondimaegn,2010).

Farm size: This is a continuous variable measured as the total crop area in hectares under the
management of the samplehousehold including own, rented-in, and cultivated through
sharecropping arrangement. It is hypothesized that as the farm size increases the technical
efficiency of the farmer will increase. This is becauseas the farm size of a farmer increases it
could enable him/her to enjoy the economies of scale. In the previous studies different results
were obtained. As an instance cultivated land was found to be related positively with technical
efficiency in Fekadu’s study(2004) and Wondimagegn(2010) while it was found to be related
negatively with technical efficiency by Mohammed(1999) and Kinde(2005).

Perception of Fertility of the plot: This is a dummy variable that takes a value of 1 if a
household perceives his plots as fertile and 0, otherwise. For households who possess more than
one malt barley plot with different fertility conditions, fertility situation of the larger size of the
plot/s is or are considered. The variable is hypothesized to determine technical efficiency
positively. This is because fertile plotsare more productive than less fertile plots. Fertility of the
plot was found to be related negatively to technical efficiency in Getu (1998) but was found to
berelated positively in Fekadu (2004).

Proximity: This is a continuous variable defined as the distance of the malt barley plot from
homestead measured in walking minutes.In case a household has more than one plot with
different plot sizes and distances, the plots size were disaggregated in to 0.25 ha bases to
calculate the average plot distance. It is hypothesized that distance of plots will be negatively
related to technical efficiency. This is because plots that are far away from homestead will

34
receive less management attention by the farmer. Proximity of plot to home was found to be
related negatively with technical efficiency in some studies (Getu, 1998; Mohammed, 1999;
Kinde, 2005).

Size of livestock holding: It is a continuous variable measured in tropical livestock units (TLU).
This variable enters the inefficiency model as a proxy variable for the wealth of the farmers.
Theoretically, livestock can support crop production in many ways: cash from livestock sale can
improve crop production, supply draft power for many farming-related activities, and they also
produce manure that will be used to maintain soil fertility.In this case they will have
complementary relationship with crop production implying positive relationship with efficiency.

However, livestock production activities can exhibit competitive relationshipif both are
competing for the same resources. It may compete with malt barley production for different
resources such as land, labor and management. In this case we can argue that size of livestock
has negative association with efficiency. Therefore, it is difficult to hypothesizea priorithe effect
of TLU on efficiency. Earlier studies also indicated conflicting results. For instance in some
studies (Mohammed, 1999; Fekadu, 2004; Aynalem, 2006)livestock holding was found to be
positively related with technical efficiency while it was found to be related negatively with
technical efficiency in the work reported byKinde (2005) (The list of explanatory variables used
in the model and their expected signs are summarized in Table 2).
Multi-Co linearity Test

One of the serious problems with the identification of variables to be included in the model is the
existence of multicollinearity relationship.Multicollinearity refers to a situation where it becomes
difficult to disentangle the separate effects of independent variables on the dependent variable
because of strong relationships among them (Maddalla, 1977). The existence of this situation in
this study is tested using the methods of variance inflation factor and contingency coefficients.
As a rule of thumb, if the VIF of a continuous variable exceeds 10 (this will happen if R i2
exceeds 0.90), and a contingency coefficient of 0.75and above for discrete variables are
indications of serious multicollinearity relationship between variables (Gujarati, 1995). Under
this situation one has to take the better remedial measure. How ever, currently against the older
assumption of no relationship exists between continuous and discrete variables VIF can be

35
computed for both continuous and discrete variablesusing STATA soft ware to see if there are
severe multicollinearity problems among explanatory variables.

Hence, before using the proposed variables for analysis VIF wascalculated for bothcontinuous
and discrete explanatoryvariables of the two models(Annex Table 7)

3.4.Methods of Data Analysis

Based on the objectives of the study, both descriptive statistics and econometric models were
used to analyze the data. Descriptive statistics like mean, standard deviation and frequency
distribution, t-test and chi square testwere employed todescribe demographic, agro-ecological,
resource and other related data.
The econometric analysis usesStochastic Production Frontier(SPF) model that integrates the
frontier andthe technical inefficiency effect regression models. The general stochastic production
frontier is specified as (Equation 6);

i = 1, 2 . . . n (6)

Where:i – is the number ofhousehold in the study

ln(yi) - is the natural log of (scalar) output of the ithsample household;
xi – is a vector of input quantities used by the ithsample household;
β – is a vector of unknown parameters to be estimated;
εi – is composed error (vi –ui)
F(.) – is an appropriate (Cobb-Douglas or translog) production function; and
vi - is a random error term which are assumed to be independent and identically
distributed (iid) with N(0, σv2) and independent of ui.
ui – is a non-negative random variable associated with technical inefficiency of the i th
sample household and assumed to be iid withN(μ, σu2).

The stochastic frontier approach has been preferably applied in many agricultural
economicresearches (Coelli, 1995). This may be because of the difficulties to accept the DEA’s
and the deterministic parametric methods assumption of that all the deviations from the frontier

36
are caused by inefficiency. Risks in agriculture like weather, pests, diseases, etc cause the
inherent variability of agricultural production. In addition, the fact that record keeping is not a
priority in the many small family operated farms also result in measurement errors. The effect of
these noises and measurement errors are separately considered under the stochastic frontier
approach unlike the DEA method.

Following Greene, (2003) the most appropriate functional form that better fits the sample datais
selected after testing the two functional forms (Cobb-Douglas or Translog) using the Log-
likelihood ratio test (LR) result as in Equation (7):

LR = -2 {ln[L(Ho) ] –ln [L (H1)] }

(7)

Where:L(Ho) and L(H1) are the likelihood function values under the null (H 0) and alternative
hypothesis(H1).

In most cases, this function has an asymptotic chi-square distribution or mixed chi-square
distribution with degrees of freedom equal to the difference between the number of parameters of
H0 and H1. If H0 is true, meaning the Cobb-Douglas functional form of the stochastic frontier
production function adequately represents the collected data. On the other hand ifthe log
likelihood ratio (LR) testvalue exceeds the upper critical value from the chi-squared tables it will
be difficult to accept the null hypothesis that states the data can be adequately represented by the
Cobb-Douglas functional form in which case the resort of the analysis will be the trans log
frontier function.

The presence of technical inefficiency in the first place is tested with log-likelihood ratio test in
which the null hypothesis (H0: γ = 0) and the alternative hypothesis (H1: γ≠0) using the log-
likelihood values of the OLS and the MLE. Besides,under this study, the SFPF has been
specified in Cobb-Douglas functional form based on the test result for an appropriate model to
estimate the technical efficiency level, and from the SFPF the technical inefficiency effect model
for identifying the causes of inefficiency in the smallholder farmers’ malt barley production.

To achieve the objective of the study, it is necessary to estimate separately the statistical noise
viand the inefficiency uiwhich represent variables under the control of the farmer to extract the

37
error term εi for each producer. This requires the choice of distributional assumptions for ui,
which is important because the Maximum Likelihood Estimates (MLE) depend on it in a
fundamental way. Different studies assumed different distributional form like half-normal,
normal, gamma, exponential, but most commonly used is the half- normal. Following Coelli
(1995), the half normal distributional assumption is considered for this study.The log likelihood
function specified for a sample of n farmers by Equation 4 is used to determine the parameters
(β, γ, σ2) by employing the first order condition (differentiation).
Again considering the above production frontier model given by Equation (6) and taking the ε i =
vi – ui term; the vi is a random error term which are assumed to be independent and identically
distributed (iid) N (0,σ v 2) and ui; is a non-negative random variable associated with technical

inefficiency in production of malt barley i.i.d. N +( σu2).Maximum likelihood estimation of the

above frontier model yields consistent estimator for β, γ and σ2.
Where: β is vector of unknown parameters

Zi =
( ln y i−x i β )
σ
2
√ γ
1−γ
(8)

2 2
σu σu
γ= 2
= 2 2 , and σ 2=σ 2u + σ 2v ,the γ parameter has a value between 0 and 1. A value of γ is zero
σ σu+ σv
indicates that the deviations from the frontier are entirely due to noise, while a value of one
would indicate that all deviations are due to technical inefficiency.

Once the parameters are estimated by Equation 4,the individual farmers technical inefficiency
can be obtained from the conditional distribution of ui given εi. As Battesse and Corra 1985 cited
in Coelli (1998), showed that if ui ~N+(0,σ2), the conditional distribution of ui given εiby equation
(8)

(
1−Φ σ A +
γ εi
)
E [ exp (−u i|ε i ) ] =
σA
(
exp γ ε i +
σ 2A
) (9)

( )
γ εi 2
1−Φ
σA

Where: Φ (.)is the standard normal distribution and the density function evaluated.
ε i=ln ( y i ) −X i β

38
σ A= √ γ (1−γ )σ 2
γ and ui are as explained above.

The technical efficiency effect model (Battesse and Coelli; 1995) in which both the stochastic
frontier and factors affecting inefficiency are estimated simultaneously (one stage estimation
approach) is specified as follows.

TE = (10)
Where: i = 1, 2 …n sample households; k = 0, 1, 2….t ; and j = 0, 1, 2…m
TE: technical efficiency of the ithhousehold
Xk+1: are the explanatory variables (inputs) used inmalt barleyproduction by the
ithhousehold

: Parameters that estimated for the production function

Zj+1: are socioeconomic variables that are expected to determine inefficiency
δj: are unknown parameters to be estimated
vi: random error in measuring the level of technical efficiency of malt barley production
The stochastic frontier production function, like any other function, describes the physical
relationship between malt barley outputs with inputs used by each selected sample household
during 2010 production season. Whereas the technical inefficiency effects model relates the
estimated technical inefficiencies of farmers with various hypothetical efficiency determinants,
supposed to be sources of inefficiency.
This study, technical efficiency analysis in malt barley production, considered six input variables
and ten inefficiency variables in the production of malt barley. The analysis was made based on
the samplehousehold level. The Cobb-Douglas stochastic frontier model was tested to be the
chosen functional formas indicated in Table17 in the result part.Consequently, the
econometrically specified Cobb-Douglas Stochastic frontier production function indicated in
equation (11) belowwas estimated with a one-stage estimation procedure.

ln (OUTP)i =β 0i + β 1 ln( AREA)i + β 2 ln( LAB)i+ β 3 ln (DRFT POWR)i + β 4 ln(CRDT INPT )i+ β 5 ln( FRQ PLW )i + β 6
(11)
Where:
ln__The natural logarithm

39
OUTP__Total output of malt barley produced by the ithsample household, measuredin qt;
AREA__Size of land on which malt barleywas grown by theithsample householdmeasuredin ha;
LAB__Amount of man-equivalent labour used in malt barley productionbythe i thsample
household
DRFT POWER__Amount of Pair of horsepower used for 8 hoursin malt barley production by
the ithsample household
CRDTINPT__Cost of DAP, Urea, and malt barley seedtechnological packagecredit, used by the
ithsample householdfor malt barley production

FRQ PLW__Frequencyof plowingson a plot or average plowings done on plots ofmalt barley by
the ithsample household

FRQ WDG__Frequency of weedings on a plot or average weedings done on plots of malt barley
by the ithsample household

Inefficiency variables are:

FAM SZ__Size of family members including the household head living in the ith sample
household
AGE__Age of the household head in years,
AGE2__Square of the household head age
EDUC__Schooling of the household head in years
MB EXP__The household’s years of participation in malt barley production
EXT CTCT__ Extension visit daysmade by the development agent to the household in the
cropping season
FRM SZ__The total land sizeofthe sample household used for cultivation of different crops
measured in ha,
FERT__Perceived fertility status ofa plot or the larger size of the plots of maltbarley of thei th
sample household. It is a binary 1 if fertile and 0 otherwise

40
DIST__Average malt barley plot distancefrom the homesteadof the household in walking minutes
calculated on bases of 0.25 ha to account the size differences of malt barley plots of
theith sample household
LV STCK__Number of livestock owned by the farmer and measured in TLU
vi__The disturbance error term, independently and identically distributed as N( 0 , σ v )intended to
2

capture events beyond the control of the household; and

ui__A non-negative random variable, independently and identically distributed asN(0,σ 2u)
intended to capture technical inefficiency effects in the production of malt barleymeasured as the
ratio of observed output to maximum feasible output ofthe i thhousehold. ui isassumed to follow
half-normal distribution with mean μi and variance σ2

Data on the technical and socioeconomic factors that are targeted to measure the efficiency of the
farmer and those determinants were collected principally based on theabove chosen variables.

41
Table 2.Summary of input and technical inefficiency variables and their expected signs

Variables Coefficient Expected sign in the

model
Input Variables
AREA β1 +
LAB β2 +
DRFT POWER β3 +
NMBR WDG β4 +
NMBR PLW β5 +
CRDTINPT β6 +
Inefficiency Variables
FAM SZ σ1 -
AGE σ2 -
AGE2 σ3 +
EDUC σ4 -
MB EXP σ5 -

EXT CTCT σ6 -
FRM SZ σ7 -
DIST σ8 -
FERT σ9 +
LV STCK σ10 -/+

Source: Different previous researches (2011)

42
 4. RESULTS AND DISCUSSIONS

This chapter is presented in two sections: descriptive and econometrics sections. The descriptive
part explains the demographic, institutional, infrastructural, and socio-economic variables
specificto the sample households. The econometrics part deals with the analysis of the input
output variables and the variables thought to cause inefficiency of sample households in the
production of malt barley in the study area.

4.1. Descriptive Results

4.1.1. Demographic characteristics and Education of sample households

The age of the sample household heads ranges between 20 and 81 years with a mean of 45.7
years and standard deviation of 11.7 (Table 3).Based on Strock et.al (1991), about 69% of the
household heads are in the most economically active age group of 17-50 years while the
remaining are above this age limit.

Regarding education the average schooling including the only read and write household(which is
estimated to be at least grade 2 level of education by the Woreda education office)is 1.5years.
The detailed collected data also indicate that outof the total sample households 55% are illiterate,
22 % somehow able only to read and write, and the remaining 23% are formally educated at
different levels, dominantly primary education. Only 3 farmers (i.e.2.5 percent) have achieved
above grade 6 level of education in the total sample households.

One of the typical features of the Ethiopian rural farm economy is its high dependence on family
labor. The survey result shows that the family size ranges from 1to 13persons (including the
household head) with the mean of 6.5 persons in the household(Table 3).The average family size
is greater than both the country’s and regional average of 4.7 and 4.3 respectively (CSA, 2007).
In terms of adult man equivalent, the family size ranges from 1 to 6.2 with mean of 2.64The
variation between the family size count and its adult equivalent reflects the sizes of economically
dependent and less labor contributing family members.In the study area, female household heads

43
constitute only 7.5% in the total sample. Their number is too small to make technical efficiency
comparison in malt barley production between the two sexes.

Table 3. Demographic characteristics and education of household heads

Description n Mean STD

Age (years) 120 45.7 11.7
Family size(number) 120 6.46 2.2
Family size(AE) 120 2.64 1.2
Education(year) 120 1.5 2.06
n Percent
Sex
Male 111 92.5
Female 9 7.5
Household size
1-5 39 31
>5(national average) 83 69
Education
Illiterate 66 55
Literate 54 45
Age
17-50 83 69.2
51-64 26 21.7
>64 11 9.1
Source: Own survey data (2011)

Family size and its structure apart from its great importance in the performance of the rural
economy it also determines the resources availability, consumption and other socio-economic
behaviors of the farm community. The survey of the detailed age structures indicates that from
the total of 655 family members of households’ 336 are males and 319 are females. Further, it
reveals that 275 family members,42% of the totalare found in the most active farm labor supply

44
age group of 17-50 years, while 167 family members or 25% are below the farm labor supply
age group below11 years old(Table 4).

About 55% of the family members can be considered as literate of whom about 5% achieved
grade level of 9 and above. On the other hand,33%, are illiterate and about 14% are under school
age of less than seven.

Table 4.Demographic conditions and Education of the family members

Description N %
Total 655 100
Male 336 51
Female 319 49
Age* 655 100
<11 167 25
11_13 88 13
14_16 86 13
17_50 275 42
>50 39 6
Education 655 100
Preschool(0- 6 years) 90 14
Illiterate 204 31
Read and write only (informal) 109 16
Grade1-6 182 28
Grade 7_8 42 6
Grade 9_10 24 4
Above10 4 1
Source: own sample survey (2011)
* Age classification is based on Man adult labour contribution as reported in Strock et al (1991)

45
4.1.2. Institutional /Infrastructural access and services

4.1.2.1. Institutional and Infrastructural access

All the farmers engaged in malt barley production in 2010 crop season in the region obtained
agricultural credit and product specific extension services, as there is a strong desire by the
government to enhance malt barley production in the region. These and other more institutional
services and infrastructural access such as health, educational services, and road facility affect
directly or indirectly the productive capacity of households. Yet, institutional services, which are
supposed to boost agricultural production, are poorly developed in many rural areas. Besides,
farmers scattered settlement in the study area causes access differences and difficulties in the use
of such services.

Table 5.Sample households’ proximity to institutional services and all-weather roads in walking
minutes

Institutions/infrastructure Mean Max Min STD

All-weather road 24.4 90 0 21.4

Nearest market 100.7 180 35 35.0
Cooperative center 28.2 120 2 18.6
Farmer training center 25 120 5 15.6
Health post/center 22.2 120 2 15.7
Primary school 18.5 120 2 14.6
Development agent office 22.1 120 0 15.5
Source: Own sample survey data (2011)

Table 5 above shows that farmers and their families are compelled to travel more than 3 hours of
a round trip if they have to get market access for transactions. Based on the farmers’ opinions,
the alternative car transport to market may not be a time saving option in many cases due to poor
public transportation facility in the area and also due to the distant location of their residence to
the nearby car transport road. Consequently, mostly farmers do opt for on foot traveldue tolack

46
of regularly scheduled public transport services in the area and remoteness of their home from
the nearest road.

Cognizant of the influence of distances of the different services from the scattered farmers’
home,currently the government is making an effort to establish the different services in one area
that centers the communitywho use the services. Presently farmers’ training centers, health
post/center, primary schools, development agent offices, multipurpose cooperative in the sample
kebelesare resettled in one central area with the aim of providing efficient institutional services.

4.1.2.2. Extension and credit services

Extension service: Extension work in the region focuses on the provision of general advisory
services on major agronomic practices (such as proper land preparation, application of fertilizer,
timely and quality operation of weeding and harvesting), post harvest handling, cooperative
organization, soil and water conservation practices and irrigation development activities.

Since 2005, an additional effort has been undertaken to introduce malt barley production in the
region.. To this end, a product specific extension program formalt barley has been designed and
implemented. Development centers, responsible for the provision of the required agricultural
extension service have been established on all the 25 rural kebele administrations found in the
Woreda. Three development agents each specialized in one of the disciplines of natural resource,
crop production, and livestock development are assigned in each development center supposing
these development agents will be giving extension services in their respective field of
specializations. However, development agents are currently providing unspecialized general
extension services in the study area. It was also observed during the survey period that there are
farmers who do not give much attention to extension advices provided by the development
agents. This may pose question on the relevance of the service in addressing the farmers’
practical problems and highlight the need for consistent revision of the extension program
depending on the specific farming conditions.

47
In the 2010 cropping season885 farmers participated in malt barley production programin the
Woreda. Duringthe period, farmersreceived production supports from the specific malt barley
extension service designed and provided by the regional agricultural bureau. The extension
program mainly entails matters on; appropriate land selection, season and frequency of land
preparation and weeding, rates and periods of sowing and fertilizer application, and time of
harvesting. Although recommendations on fertilizer and seed rates were developed by Amhara
National Regional State through pilot researches, they are blanket recommendations supposed to
serve across the region.Accordingly,125kg of seed, 100kg DAP and 110 kg Urea are
recommended for a hectare of malt barley production. Regarding the variety of malt barley seed,
mainly one-type variety called ‘Holker’ is introducedin the program.

Credit service: agricultural credit in the area is limited and mainly goes for the purchase of
fertilizer, improved seed and small ruminant animal production. In 2010 production year all malt
barley program participants took DAP, Urea and improved seed package credit.The level of
credit is similar as their malt barley size has no much variation, and no additional credits were
given for other activities. The sources of credit are mostly farmers’ service cooperatives and the
Amhara Credit and Saving Institution (ACSI). Although fertilizer credit constitutes the largest
proportion, still the credit supply falls short of the demand for it and hence a large portion of
farmers land is sown without fertilizer. For instance, out of the total area of 241 ha
oflandcultivatedby the sample householdsin 2010 cropping season, only about88 ha(36.5%) of
the total land were sown with fertilizer. All malt barley plots and about 46% of their wheat plot
areas were sown with fertilizer.Malt barley growers are given priority for fertilizer and improved
seed credits based on the assessment of their land preparation and the recommendation rates of
applications. As indicated inTable 6 sample farmers applied 39 kg of DAP and 42 kg of Urea on
their mean malt barley holding of 0.47 ha.

48
Table 6.Fertilizer use levelin malt barley production by sample households’ in2010 production
season

Description Mean Max Min STD

Total area (ha) 2 15 0.45 1.6

Malt barley sown with fertilizer 0.47 1.5 0.25 0.2

Fertilizer used(kg)
DAP 39 51 24 3.2
Urea 42 59 15 5.4
Source: Own survey (2010)

4.1.3. Holidays

All sample households in the study area are Orthodox Christian. In relation to their
religion,farmers celebrate many monthly holidays without involving on major farm operations
such as plowing, weeding, and harvesting. The survey result reveals that all the households’
synonymously celebrate Saturday and Sunday. In addition, on average sample households
celebrate otherreligious holidays of about 7.73 days per month with a range between
8to11days,and standard deviation of1.6 days. Assuming about the 2 other holidays overlapping
with Saturday and Sunday holidays of the month, about 14 days (i.e. at least 8 weekend holidays
and 5.73 non-weekend holidays) are celebrated. Hence, the religious holidays are many with
little variation among households toanalyze their impacts on technical efficiency in malt barley
production. Generally, they spare about 46% of the working period of the month to celebrate
religious holidays without undertaking major farming operations.
4.1.4. Resource bases

Under this part land holdings and its proximity, livestock, and off-farm activities which are the
major resources in the rural area affecting farmers decision in crop farming are discussed as
follows;

49
Land holding

Land is the indispensible livelihood means or resource base for the rural community. Farmland
in the study area is rather scarce. The survey indicates the average landholding is only 1.04ha
(Table7).This holding size is even less by 17% thanthe national average landholding of 1.24ha
(CSA, 2006).This coupled with poor fertility status of the soil and large family sizes of
households in the area aggravate the dearth of farmland.

When possible, farmland in the study area is also acquired by the able bodied and betteroff
farmers through rent and share cropping arrangements. Group discussions made with farmers
reveals thatwhen farmers leave their location for some periods, theyrent or sharecrop their land
to others. Some others who are disabled or lacking the working capital (hoses, human labor and
cash) also rent or sharecrop their land to others. On the other hand, others acquire lands if they
have more money, more labor to operate and other resources like horsepower.

The detail of collected data also indicates that own land, which is possessed by 99.2% of the
respondents’ accounts 51.6%, of the total cultivated lands.Similarly, 38.6%, and 9.8%.of the total
crop land in the cropping year was cultivated by 74.2%sharecropping and 10.8% rented in
operators of the sample households respectively.

Table 7.Cultivated land of samplehouseholds in2010 by sources

Land source Mean Min Max SD

Own land (n=119) 1.04 0 4 0.5
Sharecropping land (n=89) 1.04 0 5 1
Rented in land (n=13) 1.8 0 10 1
Total land (n=120) 2 0.45 15 1.6
Source: Survey data result (2011)
Note: Cases by land sources add up more than 120 because of land acquisition of a household
from more than one source

50
Again, when malt barley producers distribution is seen by sources of land, 119 farmers possess
own farmland, and 89 farmers operated with share cropping arrangements, 13farmers cultivated
rented in land in addition to their own land in 2010cropping year/season (Table7).The value for
rented land in the cropping season generally ranged between 2000 and 3500 birr per hectare
where variations mainly dependon the fertility, slope conditions, and proximity of the plot to the
farmer residence.

As also shown in the same Table sharecropping, is the second important means of acquiring land
for cultivation next to own land both in terms of sizes of land it contributes and proportion of
malt barley farmers engagement. Large participation of malt barley producers in sharecropping
compared with land renting is justified by the farmers as the relatively fertile land requirement of
malt barley crop and sharecropping preference of owners of the land. Sharecropping preference
of land owners to renting may be due to high price expectation for malt barley cropand there by
more benefit than the land rent income. Sharecropping agreements are mostly effected in two
options in malt barley production: annual agreement or agreement exceeding a year. Under the
former condition, the owner provides the cultivable land to the operator only for one year, and
the owner shares fertilizer and seed costs with the operator equally. Whereas in the later case the
land is given to the operator for 2-5 years in which case all costs will be covered by the
operators. Generally, the choice of the option and the decision of the contract period
sharecropping, highly depend on the willingness of the owner, as both parties believe longer
period agreements will favor the land operator.

Proximity of malt barley plots to homestead

A relatively fertile, less sloppy, and high altitudinal preference of malt barley and the
smallholding feature of the farmers in the area could make the malt barley plots scatter at
different places and distances from the farmers’ residence. These in turn have implications on
efficient use of time and other resources. About 27 % of the farmers’ plots are situated withinless
than 15 minutes of walkingdistances (Table 8).On the other hand nearly11% of the farmers ought
to travel more than an hour to find their malt barley plots while the remaining 62% of the farmers
walk between15 to 60 minutes

51
Table 8.Average distance of Malt barley plots from homesteads of the sample households

Walking time in minutes n %

<15 32 26.7
15-30 45 37.5
31-45 14 11.7
46-60 16 13.3
>60 13 10.8
Total 120 100
Source: Own survey result (2010)

Livestock holding

Livestock being an integral part in the farming system have many uses forsmallholder
farmers.They are sources of draught power, food, manure, cash income and are also means of
wealth acquisition or prestige to the farmer.

The types of livestock kept by farmers’ arecattle, horses, sheep, goats, poultry and
beehives(Table9).Average livestock holding of farmers in the study area is 4.99 TLU (Tropical
Livestock Unit) in the range of zeroto 13.92 TLU per household.

52
Table 9.Livestock ownership of sample households by TLU and numbers of the livestock type

Standard
Mean Minimum Maximum
deviation
TLU(n=119) 4.99 3.31 0 13.92
Cow(n=63) 1.62 1.50 0 8
Ox(n=44) 1.1 0.35 0 2
Steer(n=21) 1.1 0.30 0 2
Heifer(n=27) 1.1 0.32 0 2
Calf(n=54) 1.15 0.45 0 3
Horse(n=113) 2.04 0.92 0 5
Mule(n=19) 1.37 0.68 0 3
Donkey(n=15) 1.07 0.26 0 2
Sheep(n=93) 5.2 3.81 0 20
Goat(n=37) 2..19 1.21 0 6
Poultry(n=106) 4.7 2.7 0 16
Beehive(n=42) 2.05 1.5 0 8
Source: Own survey result (2010)

In the study area plowing is exclusively carried out by horses except in some occasions when an
ox is used as a pair much of a horse in plowing. Horses in the area are also transportation means
of the family for both goods and humans. Horse holdings vary in the households
(Table10).About 6% of the sample farmers have no horse, and 28% and 43%of the farmers have
one and two horses respectively while the remaining 38% percent of sample farmers possess 3 to
5horses.

53
Table 10.Horse ownership of the sample households

Horse number n %
0 7 5.8
1 34 28.3
2 51 42.5
3 28 23.3
Total 120 100
Source: Survey result (2010)

Those farmers owning 3 or more horses cultivate more lands through sharecropping or hiring
other’s land (Table11).Besides, they also rent horsepower for other farmers during peak season.
Number of horses under this situation explained the complementarily of the crop and livestock
production. As indicated in the same Tablebelow, farmers who possess 3 or more horses were
able to cultivate 132%more lands than those farmers having less than 3 horses,through land
renting and sharecropping.

Table 11.Household horse holdings and additional land access

Mean sharecropped Standard

Horse number N t-value
and rented land deviation
>=3 28 1.72 1.3 3.4***
<3 92 0.74 0.8
Source: Own survey result (2010)

Non/off farm activities

Although a study carried out by Oxfam GB, (Office memo, 2007) on the nation’s potentials for
off farm opportunities shows that ANRS to have high untapped off farm job opportunity, there is
very limited engagement in off farm practices by the households in the study area. The detail
data on this respect indicates only 8 households (7%)engaged in off farm activities either by
household head or through the family members. Among the 8 off farm participants, only 3 of

54
them fetch a relatively better income that reaches4800 to 6000 birr per annum through mule
renting for tourist. On the other hand, the off farm work of the remaining participants is mainly
tea making through the female family members of households. In addition, based on the
discussion held with respondents the average net income from such business does not exceed birr
5 a day in most cases. Hence, currently the impact (i.e. either positively or negatively) of the off
farm activities on malt barley production could be inconsiderable.

4.1.5. Cropping system and production

Cropping system

Highland crop-livestock farming characterizes the agricultural production system of the area.
Food barley, wheat, pulses such as beans, peas, lentil, and flax and niger seed from oil crops are
crops commonly grown in the area. As indicated in Table 12farmers harvest their annual crops
from 594 plots i.e.an average of 4.95 plots per farmer. There are no as such, perennial as well as
vegetable cropping except some backyard efforts by few farmers where limited production of
crops like eucalyptus, hop (Gesho), potato, and cabbage are undertaken. Inputs of land,
horsepower for plowing, labor and mostly local seeds (except for malt barley and wheat) are the
locally originating inputs of the cropping system in the area. Commercial fertilizer was applied
only for malt barley and to a lesser extent wheat crops in the 2010 cropping season. Other
chemical inputs such as herbicides and pesticide are not totally parts of the crop production
system so far. Crop production in the area generally seems not diversified, probably due to the
different production problems and lack of focus in this regard by the development support
organizations. Farmers tell that deteriorating soil fertility, small average land holding, little use
of improved inputs, and the different production calamities or natural risks prevailing in the area
are the major production problems.

55
Table 12.Size and plots of annual cropland of sample farmers’ in2010

Area(ha) % Number of plots %

Malt barley 56.98 23.2 131 22
Food barley 42.38 17.3 96 16.2
Wheat 67.42 27.4 165 27.8
Pulses 69.62 28.3 177 29.8
Others 4.25 1.7 12 2.0
Fallow land 5 2 13 2.2
Total 245.65 100 594 100
Source: Own survey result (2011)

Soil fertility management

In tackling the land fertility problems, different measures that have an immediate and long run
effects are undertaken. The efforts include soil and water conservation, use of commercial
fertilizer, compost, manure, and crop rotation.

The tendency of chemical fertilizer application in the area is very low. Farmers apply DAP and
Urea fertilizers only for malt barley and partially for wheat crops. In 2010 cropping season all of
the 56.98 ha malt barley plots and 31 ha (74%)of their wheat plots were fertilized using DAP
and Urea fertilizer (Table15). The rates of application were 16% and 18% lower than the
recommended rate of 100kg DAP and 110kg Urea per hectare in malt barley production,
respectively. Similarly, rate of application for wheat is less by 55% and 56% of the blanket
recommendation rate of100kg DAP and 100kg urea per hectare. The sample households mention
financial scarcity, lack of adequate credit access, and untimely fertilizer supplyas constraints
foradequate fertilizeruse in wheat production. On the other hand, although many farmers know
the recommended rate, they do not stick much init even when they do not haveshortage of
fertilizer. This could be probablyeither, stillthey are benefitingfrom the sub recommendation
practical application ratesalthough not optimal or due to their differences in the recommended
rate.

56
Table 13.Recommended and actual fertilizer use in malt barley and wheat production by
households’ in2010

Recommended rate Actual rate used Difference (%)

Crop DAP Urea DAP Urea DAP Urea
Malt barley 100 110 84 90 16 18
Wheat 100 100 45 44 55 56
Source: Own survey result (2011)

Compost production and use is currently highly promoted by the government. Farmers explained
the problem in this regard, and thus little use of the technology in the area. The cool environment
that hinders easy decomposition for compost formation, the undulated terrain, poor
infrastructure, shortage of biomass, and fragmented plots make difficult the wider application of
the technology. Besides, although shortage of land is a serious problem in the area, occasionally
fallowing is also taken as a strategy of fertility replenishment when production of pulses or oil
crops in the exhausted land is not found feasible.

Moreover, there are other methods of soil fertility management including crop rotation andsoil
and water conservation structures practiced in the study area. Particularly, the significance of
crop rotation in soil fertility maintenance is highly appreciated in the farmers cropping plan.Even
in the case of land renting and sharecropping agreements, the past crop cover of the land is
considered by the prospective land operator as the base for decision. Crop rotation practice of the
area involves growing leguminous crops every other year after cereals. The importance of
legumes and pulses in improving soil fertility is well recognized by the farmers in the
community. In 2010 cropping season, exceptionally 8 farmers out of 120 sample households
grew wheat or food barleyon their malt barley plots in the previous production year. With regard
to soil and water conservation in addition to the communal works in the upper catchment areas,
13 sample respondents (11%) undertook different soil and water conservation works such as cut
off drain, soil and stone bund andfarmyard fencings.

57
58
Crop production

In 2010 cropping season sample farmers could able to produce on average 10.7quintals per
hectare (Table 14).All crops except, malt barley were produced by the sample farmers for family
consumption. The purpose of malt barley production by all the farmers was originally only for
sale. However,about 38% of the households utilized it also for home consumptions to meet their
family food demand. From the total crop produced in the 2010 cropping season, wheat
accountsfor 31%, food barley19%, malt barley 33%, pulses16%, and other crops1%.Table 14also
compares the2010 cropping season productivities of major crops of farmers withthe respective
averageproductivitiesof the Woreda, based on the information of the woreda agricultural office
that the sample Kebeles has similar productivity with the woreda for the crops grown. The
reduction in yield in most of the crops in the cropping season probably could be partly due to
frost damage as reported by 107 sample households.

Table 14.Crop productivity loss estimates (qt/ha) of sample households in 2010

2010 Woreda average Productivity difference

productivity productivity (%)
Malt barley 15.2 18 -15.6
Food barley 11.1 16 -30.6
Wheat 12.1 18 -32.8
Pulses 6.2 9 -31.1
Others 5 5 nil
Total 10.7

Source: Own survey result (2011)

Malt barley production practices and major inputs

Malt barley takes 9 months of operations from land preparation to threshing. Briefly, 4 plowings
from march to June, sowing in mid June,3 weedings from July to early September,and harvesting
and threshing around end of November are the recommended major activities and seasons of
operation in malt barley production. In 2010 cropping season, all the sample farmers used two

59
external inputs of chemical fertilizer and improved seeds in malt barley cropping. In addition,
labor of own and exchange sources were intensively used in relation to the recommended
activities of malt barley production.

Seed

The major maltbarley variety in Ethiopia are the Beka,the Holker,the HB-52,the HB-120,and the
HB-1533.But only two varieties vis Beka and the Holker are introduced in the study area. These
two varieties have been selected for their high yield in the area, good market demand and their
straw that is used for animal feeding. However, only the Beka variety was supplied to farmers as
a seed source in the woreda for the production year.Malt barley seed is given to farmers in terms
of kind credit.Moreover,it was found by this study that farmers’ seeding rate of malt barley in the
cropping season complied with the given seeding rate recommendation of 125kg.This may be
due to appropriateness of the recommendation rate,orthefarmers perception of the new practice
requirement of malt barley crop which is new in the area, or because of limited alternative use in
other types of crop production unlike fertilizer, or still due to the crop’s specific land and
cultivation requirement.

Chemical fertilizer

Fertilizer is the major input in malt barley production. Malt barley producers are advised to apply
on average 100kg of DAP and 110kg of Urea per hectare. Fertilizer inputs are provided for malt
barely producers in credit on priority basis, Consequently, all malt barley participants of the
Woreda in general and the samplefarmers in particular used fertilizer for malt barley production
in the cropping season. As indicated in Table15above in the soil fertility management part,
therate of DAP and Urea usein the cropping year were84and 90 kg per hectare respectively.
Although the rates of fertilizer actually used in malt barley production are lower than the
recommended rates both for DAP and Urea it is still higher than the rate used in wheat
production. The relative higher input use in malt barley production may indicate relative
importance of the enterprises.

60
Labor

The conscious effort of labor, acts on other factors of production in the output realization
process. High labor use is the indispensible element of malt barley production process in the
area, as man executes all activities of production, with some help of draught power mainly in
plowing. Neither herbicide for weed controlling, nor insecticides are used in the production
ofmalt barley to protect the crop from the damage of insects and weeds. None of the sample
farmer used these chemicals in the production year. Farmers control weed infestation only by
means of hand weeding. In the cropping season sample, farmers employed 26MAE labour(Table
15)for the production of malt barley on their average plot of 0.47 ha. This is equivalent to the
labor use of 55MAE per hectare .As shown in the same Table, out of the total labor input
28%was used for plowing, 52 % for weeding, and 20%for harvesting. Family being the main
source of labor in the study area it contributed 66% of the total labor requirement. While labor
outside the family, which is mainly obtained through exchange constitutes 34% of the
contribution in malt barley production. It interests to appraise how much of the available family
labor in the production period is utilized in malt barley production and crop production in
general. The daily available 2.6MAEL in the family make up702 MAEL in the 9 months of malt
barley production period. This implies only 3.7% of the family labor available in the production
period is put in malt barley production had only family labor been used. Assuming, other crop
lands of the sample household are also labor intensive like malt barley farming, the total labor
requirement of crop production for the average 1.04 ha farmer’s holding will be about 57MAEL
which is still about 8%of the family labor available only in the malt barley production period.
Hence, crop production that is believed to take the larger labor share of the farmers whole
occupation consumes very little of the available large family labor. Further, this highlights the
room for further labor-intensive interventions in the study area.

61
Table 15.Labor availability and use by malt barley production activities and sources (in MAE)

%of the total

Mean STD
requirement
Daily available family labor (MAE) 2.6 1.3
Labor used per ha of malt barley 25.9 1.6 100
plowing 7.2 4.2 28
weeding 13.5 1.1 52
harvesting 5.2 3.8 20

Family source contribution 17.2 1.1 66

Exchange/hired source contribution 8.7 1.1 34

Source: Own survey result (2010)

Malt barley yield

The survey reveals sample farmers malt barley productivity in the cropping year was 15.2
quintals per hectare, with standard deviation of 6.4.The maximum and the minimum productivity
were also 4 and 48 quintals per hectare respectively (Table 18). Many sample farmers explain
the productivity level achieved in the cropping season is lower than their average yield of the
good season. Frost and wind at the blooming stage, flooding, and late time moisture stress are
mentioned to be causes of yield loss for malt barley in the cropping season. Assessing the
productivity loss estimate by these and other similar externalities will have policy relevance. To
this end, farmers were asked how much they could have obtained from their malt barley plots in
the cropping season had there been no hazard to their malt barley cultivation. Accordingly, malt
barley productivity in the cropping season would have been19.3quintals per hectare if production
were not affected by the above mentioned natural calamities. This implies about 21% yield loss
occurred due to reasons largely, beyond the control of the farmer. However, the damages could
be meaningfully, mitigated through vigorous government log term interventions.

62
Table 16.Farmers’performances and potentialof malt barley productivityin different seasons

Average
n STD Minimum Maximum
yield/ha
Cropping season(2010) 120 15.2 6.4 4 48
Expected potential of 116 19.3 8.7 10 32
(2010)
Good year (perceived) 80 19.9 9.7 4 60
Bad year (perceived) 78 9.3 6.4 0 48
Source: Own survey result (2010)

Moreover, as indicated in Table 16 productivity comparison was also made between the farmers’
good season harvest and bad season harvest for those farmers who have participated more than
once in malt barley production. Farmers obtained malt barley average highest yield of 19.9
quintals per hectare in the good season while yield fall to its lowest average of 9.3 quintals per
hectare in bad season. This shows the adverse effect of natural calamitiesis causing about 53 %
yield loss. On the other hand, the cropping year potential malt barley productivity of 19.3 qt/ha is
close to the good season’s yield level of 19.9 qt/ha.

As indicated in the above Table large differences between, the actual and the expected farmers’
average yield as well as the ‘good’ season and ‘bad’ season yield is an obstacle for malt barley
development initiative in the region. One important factor determining farmer’s success and their
stay in the business is the productivity and stability of malt barley production. Generally, farmers
would have incentive to keep on adopting the new technology as long as the net incremental
benefit from the new enterprise is positive. In other words, farmers will continue to participate in
malt barley production as long as the net benefits from malt is greater than the forgone income
that could be obtained from the second best alternative which is food barley.

63
4.1.6. Major Production Constraints

Farmers in the area confront numerous challenges in malt barleyproduction ranging from farmer
specific problems to the more general community-wide ones. Erratic rainfall, drought, poor soil
fertility, frost, flooding, inadequate and poor quality extension service, limited access to credit,
market and input technologies, poor infrastructure, delayed payment for their output are
constraints mentioned by producers.

Among the natural hazards frost, rainfall distribution problem and associated drought, and flood
were mentioned by 88%, 83%, and 43% of the respondents, respectively, as the serious
production problems they faced in the cropping season. Moreover, almost all sample households
reported that there is delay in payment for their malt barley credit sale to Dashen brewery.
Payment was made six months after the output was delivered without any interest compensation
for the delay,while the input taken by farmers on credit for the production of the output were
continually bearing interests until it was paid by the cash earned from output sales.. In addition,
malt barley producers who originally perceived the agreed price of Birr 618 as a good price
during the contract agreement eventually found it to be unattractive when the price of food
barley rises dramatically well above the price of malt barley. Hence farmers were forced to buy
one quintal of food barley for birr 1200 which was available just for birr 550.

4.2. Results of Econometric Analysis

4.2.1. Hypotheses testing

The fact telling ability of the research and its contribution in the enhancement of the efficiency of
the given production system depend among other things on the selection of the type of the model
and on the appropriateness of the inefficiency effects variables considered in the study. It is
therefore, a worthy exercise to test the appropriateness of the chosen study model, the existence
of inefficiency among malt barley farmers in Debark Woreda, and whether these inefficiency
differentials are adequately explained by the inefficiency variables. These tests of interest can be
conducted using the LR ratio statistics (Coelli and Battesse, 1996).

64
The Cobb-Douglas and the Translog functional forms are the most commonly used stochastic
frontier functions in the analysis of technical efficiency in production. The translog frontier
function turns into the Cobb-Douglas when all the square and interaction terms in the translog
are zero. In order tochoose between the two alternative functional forms that can better fit to the
survey data collected, the null hypothesis that all the interaction and square terms are all equal to
zero(H0 = ij = 0), i.e.Cobb-Douglas frontier function specification, is tested against the
alternative hypothesis that these coefficients are different from zero.The test is made based on
the value of likelihood ratio (LR) statistics, which can be computed from the log likelihood
values obtained from estimation of Cobb-Douglas and Translog functional specifications using
Equation 4.7. Then, this computed value is compared with the upper 5% critical value of the 2 at
the degree of freedom equals to the difference between the numbers of explanatory variables
used in the two functional forms (in this case df = 21).

The estimated log likelihood values of the Cobb-Douglas and Translog production functions
were 11.9and 26.4, respectively. The computed value of 29 is lower than the upper 5% critical
value of the 2(Table 17).Thus, the null hypothesis that all coefficients of the square and
interaction terms in Translog specification are equal to zero was not rejected. This implies that
the Cobb-Douglas functioType equation here .nal form adequately represents the data under
consideration. Hence, the Cobb-Douglas functional form is used to estimate the technical
efficiency of the sample households in the study area.

The next testof hypothesis is to examine for the existence of inefficiency in the malt barley
production of the sample households. This is carried out through testing the null hypothesis of
‘no inefficiencies’ (i.e. Ho: = 0)and therefore the data can be sufficiently represented by the
average response function (OLS specification) that does not consider the non-negative random
error term against the alternative hypothesis that SPF represent the data i.e. there is inefficiency
in the production of malt barley.

The likelihood ratio (LR) computed from the log likelihood functions of both the null average
response function and the alternative stochastic frontier production function (applying equation
4.7) is compared with the upper 5% critical value. Accordingly, the computed log likelihood

65
ratio was 57.46.The value is much higher than critical value at the degree of freedom equal to
one. Hence, the rejection of the null hypothesis indicates the existence of inefficiency or one-
sided error component in the model. Hence, the hypothesis that malt barley producers in the area
are on average efficient is rejected. As a result, production of malt barley in the study area is
characterized with technical inefficiency and thus the data can be better represented by the
stochastic production function.

The last null hypothesis to be tested specifies that all the coefficients of the explanatory variables
in the inefficiency model are equal to zero (i.e. H o: 1 = 2 … = 10 = 0). It is to mean that the
explanatory variables considered in the inefficiency effect model do not contribute significantly
to the explanation of the technical inefficiency variation for the malt barley-producing farmers.
Thecomputed log likelihood ratio (LR) obtained through similar calculation is 47.82 and it is
greater than the tabulated critical valueat10 degrees of freedom (which is equal to the number of
inefficiency variables included in the unrestricted model) of18.31. As a result, the null
hypothesis is rejected in favor of the alternative hypothesis that the coefficients of the
inefficiency variables are jointlydifferent from zero. It implies that the inefficiency variables
jointly explain the technical inefficiency deferential among farmers

Table 17.Tests of hypotheses for model and parameters of the SPF

Null hypothesis LR value Critical value Decision

H0 : ij = 0 29 32.671 Accept
H0:  = 0 57.46 2.71 Reject H0
H0: 1 =… = 10 = 0 47.82 18.31 Reject H0
Source: Own computation result (2011)

4.2.2. Estimation of parameters for Stochastic Frontier Production Function (SFPF)

As indicated in the data analysis of the methodological part, the specified Cobb-Douglas
functional form of the stochastic frontier model with half-normal distributional assumption of the
error terms is considered to estimate the model or parameters of the model. The Parameters were
estimated simultaneously with those involved in the model for the inefficiency effects. The

66
estimation is carried out using FRONTIER version 4.1 computer program in a single stage-
estimation procedure which has an advantage of consistency over the two stage estimation
approach ( Coelli and Battese,1996). As shown in equation 12, the model includes 6 explanatory
variables of the frontier function and 10 explanatory variables for the efficiency differentials.

Table 18.Maximum likelihood estimates for parameters of Stochastic Frontier Production

Function

Standard-
Coefficient t-ratio
error
Constanta 1.29 0.25 5.24***
Land under MB 0.45 0.11 4.27***
Labor -0.02 0.08 -0.25
Draught Power 0.37 0.12 3.07***
Frequency of plowing 0.11 0.06 1.82*
Frequency of weeding 0.03 0.07 0.42
Value of inputs(DAP,Urea&Seed) 0.23 0.07 3.19***
2
Sigma squared(σ ) 0.08 0.02 4.08***
Gamma(γ) 0.22 0.1 2.14**
Returns to scale 1.16
Log likelihood function -11.3
Source: Own computation result (2011)
Note: a, ***, **,* indicate the natural log of the constant term and significance at 1%, 5% and10
percent levels, respectively.

The above table results reveal the estimated coefficients of the stochastic frontier have signs that
generally conform with the expectations except the labor power. Area put under malt barley
cultivation, draught power, and cost ofinputs (DAP,Urea&Seed) used are statistically significant
at 1% level while the frequency of plowings performed on the differentmalt barley plots is
significant at 10 % level. On the other hand, frequency of weeding, and human laborused in malt
barley production do not significantly determine production. Besides, labor power has
unexpected result contrary to what one would expect although it is not statistically different from
zero.

67
One of the appealing features of the Cobb-Douglas functional form is the direct interpretation of
its parametric coefficients as a partial elasticity of production with respect to the input used. This
attribute allows one to evaluate the potential effects of changes inthe amount of each input on the
output

As shown on table 20 the parametric coefficients or partial elasticity of significant input

variables are 0.45 for area,0.37 for draught power,0.23 for material inputs,and0.11 for number of
plowing. These values indicate the relative importance of each factor in malt barley production.
Otherwise, a one percent increase in the use of land, draught power, material inputs, and number
of plowings will result at about 0.45, 0.37, 0.23, and 0.11%increase in the efficiency level of
malt barley output, respectively. Consequently, land is the most important factor of production,
followed by draught power, values of material inputs and frequency of plowings in the order.

Coefficient parameters are also summed up to know about the returns to scale. The value of 1.16
scale coefficient in this case could indicate that malt barley production in the area exhibit
increasing returns to scale. The implication of such a result is that proportional and
simultaneousincrease of all production inputs could lead to a more than proportional increase in
output. This result further reveals that, if the input markets are competitive, malt barley
producers could enjoy economies of scale.

The maximum-likelihood parameter estimates of the model besides telling the relative
importance of individual inputs and their aggregate significance in the production process
theyare alsoused to predict the level of individual farmer’s technical efficiency.

68
4.2.3. Technical efficiency scores

One of the main objectives of this study was to measure the technical efficiency levels of malt
barley producing farmers in Debark Woreda.Given the chosen functional form and the
distributional assumptions made about the two error terms vi and u1, the technical efficiencies
were estimated. The estimation result shows that the mean efficiency level of malt barley
producers is 0.81 and their efficiency ranges from the most inefficient level of 0.38 to the highest
level of 0.99 (Table 21).

Accordingly, farmers producing malt barley in the area are not efficient and as a result, on
average 19% of the malt barley output is lost due to inefficiency of producers. This suggests that
on average farmers can increase their current level of output by 19 % without increasing the
existing levels of inputs they are using. Alternatively, the result indicates that farmers on the
average could decrease their current levels of inputs by 19% to obtain the level ofmalt barley
output they are just obtainingcurrently.

Table 19.Technical efficiency score of sample households

Efficiency Score estimates

Mean 0.81
Maximum 0.99
Minimum 0.38
STD 0.15
Source: Own computation results (2011)

In addition to the above descriptive results, grouping the levels ofindividual technical efficiency
scores into certain classescan give better picture about the distribution of individual efficiency
scores. One way of looking at frequency distribution of the individual efficiency values is taking
the mean efficiency as a milestone. According to Stevenson (1980), grouping can be done based

69
on the relative performance of each sample farmer in relation to the mean performance level and
the corresponding standard deviation.
Hence, three sample farmers’ categorical groups could be identified as the less efficient, average
and more efficient farmers based on their technical efficiency scores. In this respect, farmers are
considered as averagely efficient if they were operating in the range of mean efficiency plus or
minus one standard deviation, and less efficient or more efficient farmers if they used to operate
below or above the average efficiency range, respectively. Table 22shows the frequency
distribution of the three different groups.

Table 20.Frequency distribution of sample households by efficiency group

Efficiency group Score Frequency Percent

Less efficient <0.66 25 20.8
Averagely efficient 0.66-0.96 81 67.5
More efficient >0.96 14 11.7
Total mean 0.81 120 100
Source: own survey data (2011)

The result shows that about 21, 67, and 12% of the farmers are less efficient, average and more
efficient malt barley producers, respectively. Out of the total farmers, 8 operate just near around
the mean efficiency level of 0.81. And about 52 of the sample farmers(the less efficient farmers
and about 33% of the averagely efficient farmers) are operating below the mean efficiency level
of 0.81.

A further and separate investigation of efficiency scoresof these 60 sample households who are
operating below and around the mean efficiency levelrevealsthat they operate with mean
efficiency of 0.66. The mean efficiency level for the less efficient households who account for
21% of the sample households is even lower than this figure and is only 0.56. These findings
indicate the existence of substantial inefficiencies, the elimination ofwhich could lead to the
larger efficiency gains. Up to 34% average productivity increase in malt barley production could

70
be achieved through technical efficiency improvement of 50% of the sample households who are
currently operating at or below the overall mean efficiency score of 81

4.2.4. Actual and potential levels of output

The knowledge of the individual farmer efficiency level and their corresponding actual output
enables to determine how much yield is lost because of efficiency problems in the current
production practice. Similarly, it is possible to find out the potential level of production that
could have been produced by the farmer had there been efficient use of the existing resources.
From the relationship of technical efficiency in a given period of time as the ratio of the actual
output (Yi: exp (xiβ +vi-ui)) to the potential output (Yi*: exp (xiβ+vi)) indicated in equation 12,
the potential malt barley production of each individual farmer is calculated as follows:

¿ ¿
TEi=Y i i/Y i =exp(−u i)which gives Y i =Y i /TE

Hence, the mean level of both the actual and the potential malt barley yield in the cropping
season was thus 15.18 and 18.59 qt per ha with the standard deviation of 6.77 and 6.54
respectively.The mean difference between the actual and potential production indicates that there
is a room to increase the production level on average by 3.41 qt per ha with the existing level of
input use. Figure 3 depicts the actual and potential production levels of sample farmers in the
production of malt barley in the study area.

120
Actual and potential output
pot.out
100
act.out/ha

80
yield/ha

60

40

20

0
1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103109115
71 sample households
Figure 3Actual and potential malt barley production of sample farmers
4.2.5. Variability of output from the frontier

It is shown inTable18 that both σ2 and γ are statistically significant, respectively showing the
existence of significant variation from the frontier function and the importance of the technical
inefficiency effects in studying the malt barley production system in the Woreda. Deviation of
the observed output from the frontier is associated to two sources of error terms: one to the
farmer’s inefficiency problem, which is under his/her control, and the other one is to the random
variation/or usual noise component, which is out of the control of the farmer.In view of this, it is
necessary to determine the variability of the output in malt barley production in the study area
attributed to each error components. However, although the significant γ coefficient shows the
importance of the technical inefficiency effect in the given production system it is not directly
interpreted as the proportion of the variance of the inefficiency effect relative to the sum of
variances of the inefficiency effects and the random variation/noise(Coelli and Battese,1996).
Thus, the relative contribution of the inefficiency effect to the total variance term γ*is given by
γ* = γ/[γ+(1- γ)π/(π - 2)]. This is because the variance of u i is equal to [(π - 2)/π] σ2 not σ2
(Coelli et al., 1998).The more this ratio is close to 1, the more the output variability is affected by
technical inefficiency than the usual random noise.

4.2.6. Determinants of technical inefficiency

The driving force behind measuring farmer’s efficiency in malt barley production is the
identification of important variables/determinants with which to work for development in order
to improve the existing level of efficiency. The parameters of the various hypothesized variables
in the technical inefficiency effect model that are expected to determine efficiency differences
among farmers are estimated through MLE method using a one-stage estimation procedure.

The determinants of technical inefficiency/efficiency in a given period vary considerably

depending on the socio economic conditions of the study area particularly pertaining to
managerial characteristics and other related factors.

72
The inefficiency variables in this study are classified under three categories. These are the
demographic and educational factors (family size, age, age square, education, and farm
experience), resource related factors (farm size, soil fertility status, distance of the plot, and
livestock holdings), and the extension contact as an institutional factor. A negative sign in the
parameter of coefficients indicate that any improvement or increase in the variable use reduces
the technical inefficiency (i.e. improvement of efficiency), and the positive sign is interpreted
inversely.

Table 23 below shows the results of the technical inefficiency model estimates. Among the 10
explanatory variables entered in the analysis 3 variables namely, family size, extension contact
and farm size have appeared with unexpected signs, of which only family size is statistically
significant. The results of the remaining7 inefficiency variables conform to the priori
expectations, in their signs and significance level as well in determining inefficiency /efficiency
of malt barley production in the study area

Table 21.Maximum likelihood estimates of the inefficiency effect model

Standard
Variable Coefficient t-ratio
error
Constant 2.4 0.87 2.77***
Family size 0.14 0.07 1.95**
Age -0.1 0.04 -2.45***
Age squared 0.001 0.0004 2.60***
Education -1.22 0.5 -2.44***
Farm experience 0.08 0.05 -1.58*
Extension contact 0.12 0.14 0.86
Farm size 0.03 0.04 0.96
Fertility of soil -0.41 0.17 -2.50***
Plot distance 0.46 0.19 2.38***
Livestock holding -0.05 0.03 -1.57*
Sigma squared(σ2) 0.08 0.02 4.08***
Gamma(γ) 0.22 0.1 2.24***
Returns to scale 1.16
Log likelihood function -11.3
Source: own survey (2010)

73
*, **, and *** shows 10%, 5% and 1% level of significance respectively.

Before discussing the significant determinants of inefficiency in malt barley production it is

important to see how efficiency and inefficiency are interrelated. The result can be presented in
terms of efficiency or in terms of inefficiency. The above result is presented in terms of
inefficiency and hence the negative sign shows the increase in the value of the variable attached
to the coefficient means the variable negatively contributes to inefficiency level or conversely it
contributes positively to efficiency levels. Thus any negative coefficient happens to reduce
inefficiency which implies its positive effect in increasing or improving the efficiency of the firm
viceversa.

Accordingly, the negative and significant coefficients of age, education, experience on malt
barley production, fertility status of the soil, and livestock holdings indicate that improving these
factors contribute to reducing technical inefficiency (Table 21). Whereas, the positive and
significant variables of family size, age square and plot distance, affect the technical inefficiency
positively that increase in the magnitude of these factors aggravate the technical inefficiency
levels.

The implications of significant variables on the technical efficiency of the farmers in the study
area are discussed here under.

4.2.6.1. Demographic factors and education

Family size

Taking the indispensible role family labor has in the rural economy as a rule of thumb, the family
size was hypothesized to reduce technical inefficiency. However, the coefficient of family size
on the inefficiency effect model isestimated to be positive and significant (Table 21). The result
is contrary to expectation. This may be explained by the financial limitations and existence of
abundant labor, which could be often associated with larger family sizes. With increase in family
size expenditures on consumption and other expenses will increase and as a result it creates

74
financial scarcity for timely and adequately purchasing ofall the necessary agricultural inputs
which in turn limits the production capacity. Family size mean efficiency comparison, taking the
national average family size of 4.7(EPCC, 2008)as the cutting point, indicate larger family sizes
are less efficient than the smaller family sizes with mean technical efficiency levels of 0.79 and
0.83 respectively (Table22).

Related to the abundance of labor, one another possible reason for positive correlation between
inefficiency and family size is probably the higher exchange labor use (ELU) arrangements
observed with larger family size households (Table22). The man adult equivalent exchange labor
use in farming practices, which tend to increase with family size may contribute to higher
inefficiency through lesser farming efforts. Similar studies about family and technical efficiency
show different results. Mohammad et al. (2000) and Ajibefun (2002) among others have found
negative but insignificant relationship between technical inefficiency and family size while the
research findings of Coelli et al. (2002), Fekadu (2004), Jema(2006), Khuda
(2007),Abebe(2009),and Wondimagegn(2010) have come up with positive and significant
results.

Table 22.Mean technical efficiency (TE) and exchange labor use (ELU) by family size of sample
households

Household n Mean STD t-value

family size
TE
>=5 83 0.79 0.16
1.4*
<5 37 0.83 0.13

ELU

>=5 83 8.6 11.9

1.8**

75
<5 37 5.7 5.8

Source: Own survey data (2011)

Age

Age changes the general farming experience and physical capacity of farmers, which in turn
bring differences in abilities of decision-making and laborious farming practices. These efforts,
again could create efficiency differential across the different age groups. To appraise the
existence of technical efficiency differences among different age groups, sample farmers are
arbitrarily classified in to three as young, middle and old age groups with age span of less than
30, 31_55 and greater than 55 years of age respectively. It is estimated that the age has a negative
effect on the technical inefficiency effects with the first two age groups until it starts to relate
positively in the third age group(Table 23). The explanation for this result could be that old age
group lack the capacity to work energetically and they also become conservative to adopt new
technologies even if they have good experience in farming. Farmers below the age of 30 also
lack experience and motivation towards farming since their mean technical efficiency was found
lower than the succeeding age group. On the other hand, farmers of the middle age group are
relatively more energetic and adopt technologies compared to old farmers and have relatively
more experience and motivation than the youngsters have.

The result indicates that the coefficients of age and age square are statistically significant in the
study area and have negative and positive signs respectively. It implies that the middle age
groups of sample household heads are more efficient and efficiency starts to decline after certain
age level as one gets older. As shown in Table 23, the mean technical efficiency of middle age
group (0.84) is greater than that of the young (0.73) and old age (0.70) groups mean efficiency.
Similar results were obtained in different studies Getu et al., (1997) Ahmad (2002), Fekadu
(2004), Haileselassie (2005), kinde (2005), Kampruzzaman and Muhammad (2008).

76
Table 23.Mean technical efficiencylevel of sample households by age groups

Proportion of respondents
Age group Mean TE STD
Number percentage
<30 12 10 0.73 0.15
31-55 87 72.5 0.84 0.13
>55 21 17.5 0.70 0.18
Source: Own computation result (2011)

Education

Education is important to increase the managerial capacity of the farmers indecision making. It is
measured in years of schooling and hypothesized to determineinefficiency negatively. As
expected, education affects the technical inefficiency effect of malt barley production
significantly and negatively.The negative sign implies that farmers that are more educated tend
to be less inefficient or more efficient in agricultural production than the less educated ones.
Table 24indicates 45% of the farmers are educated with mean schooling years of 2.07 years, and
operate with mean efficiency level of 0.84, which is above the overall mean efficiency of 0.81.
This could be because, educated farmers have the ability to use information from various sources
and undertake more informed decisions that could improve the farm management and hence,
their efficiency on malt barley production. This result is in line with the findings of Coelli and
Battese (1996) Getu (1998) and Mohammed (1999).

Table 24.Mean technical efficiency by educational levels and mean schooling of sample
households

n Mean STD

Efficiency: literate 54 0.84 0.14

Illiterate 66 0.77 0.15
Schooling years of literates 120 2.07 1.7
Source: Own computation result (2011)

77
Experience in farming

Farmers who are currently engaged in malt barley production although they are not new for
general farming practices, they have no previously acquired age experience 5 years ago. Malt
barley productionstarted even in the region only since 5 years. Learning by doing, and exchange
of experiences with similar colleagues therefore enable farmers build up relevant knowledge on
the specific malt barley production system. Consequently,this variable entered in the analysis
with an interest of knowing whether efficiency improves with the number of years spend in malt
barley production. The result reveals specific experience in malt barley production significantly
and negatively affects the technical inefficiency effects. The sign is consistent with prior
expectations. In making mean efficiency comparison between the experienced and non-
experienced farmers, farmers who have produced malt barley 3 times or more are taken as
experienced framers based on the discussion made with the sample households.

Table 25.Mean technical efficiency level by experience group

N Mean TE STD
>= 3 years 74 0.83 0.14
< 3 years 45 0.77 0.17
Source: Own survey data (2011)

As shown in the above table the mean technical efficiency for the experienced framers in malt
barley production is 0.83 while for the nonexprienced farmers is 0.77.The two mean technical
efficiencies are significantly different at 10% significance level. This denotes experienced
farmers are 6 % more efficient than the nonexprienced farmers are. Drawing important lessons
from experienced malt barley producers particularly for the benefit of non-experienced farmers
will therefore help contributing its part to enhancing the technical efficiency in malt barley
production. This may have an important implication for the malt barley extension program,
which is being implemented but according to the result of this study is not significantly
contributing to technical efficiency improvement in the production.

78
4.2.6.2. Resource factors

Soil fertility

Soil fertility loss is a serious crop production problem in the study area. The steep and undulated
landscape, long years of continuous land cultivation, lack of awareness and related little soil
conservation efforts by the community are the major reasons of soil fertility deterioration in the
area(Debark Woreda agricultural office,2010). Short run efforts are made by farmers to replenish
the soil fertility during the production season through the application of different kinds of
fertilizers. Besides, long run measures such as crop rotation, soil and water conservation, wise
land cultivations, and other similar efforts have important contribution in maintaining and
improving the soil fertility status, which normally helps increasing the crop yield.

A pilot research conducted (ARARI, 2005) before the commencement of the ANRS regional
malt barley production program reveals the relatively fertile land requirement of malt barley in
the context of the general poor fertility condition of the study area.

Hence, in this study the soil fertility variable entered the model to capture the impact of fertility
differences on the technical inefficiency effect of malt barley production. It was hypothesized
that the fertility status of the plot could have a negative effect on the technical inefficiency of
farmers in the production of malt barley. As expected, the result indicates that the coefficient of
fertility was negative and significant .This implies that fertility of land is an important factor in
influencing the level of efficiency in the production of malt barley.

Table 26.Mean technical efficiency between fertile and infertile land operators

N Mean TE STD t-value

Fertile land 24 0.86 0.16
1.74**
Infertile land 96 0.80 0.10
Source: own survey (2011)

79
Table26shows the mean comparison of efficiency between fertile and infertile land. Farmers who
are operating on fertile land are 6% more efficient than those farmers operating on infertile land.
Nevertheless, only 20% of the total malt barley farmers are able to operate on relatively fertile
lands. Therefore, expansion and sustainable development programs in improving and
maintaining the fertility of land will have positive impact in raising efficiency. The result on this
study variable is in conformity with earlier studies conducted in the region by Fekadu (2004) and
Wondimagegn(2010).

Plot distance

Farmers whose malt barley plots are located away from their home not only spend extra time to
reach their plots and back to home but also would get damage or loss to their inputs while
transporting to the plot area. In addition, labor that will be involved in agricultural operations
may also be fatigue by the time it arrives at the plot that consequences on the labor efficiency.
The ragged topography and existing underdeveloped infrastructural conditions of the study area
also worsen the effect of the farm plot distance. Generally, the aforementioned distance related
problems exert influences on the volume, quality, frequency and timeliness of the different
agricultural operations of malt barley production.

Thus, the plot distance variable was hypothesized to have a positive effect on the technical
inefficiency of the farmers. The result is in line with the expectations.. As indicated in Table 28
farmers who have their plots at the walking distance of 15 minute or less are more efficient than
those farmers who travel more than 15 minutes to get their plots. It implies that any viable plot
distance reduction measures will have an effect on lessening technical inefficiency or improving
the technical efficiency of malt barley producers in the area. Similarly, proximity of plot to home
was found to relate positively with technical inefficiency in study of technical efficiency
ofsmallholder farmers wheat production in Machakel woreda by Fekadu (2004).

80
Table27. Technical efficiency and plot distance relationship

N Mean TE STD t- value

>15 88 0.80 0.15
1.79*
<=15 32 0.86 0.14
Source: Own computation result (2011)

Livestock

Livestockin a mixed farming system have importance in the supply of animal power for plowing
and threshing, provide draught power, manure, fire energy, and they are sources of income and
food for the family. It can also be a proxy variable for the wealth status of the farmer.

As livestock could have both competitive and complementary relationship with crop production,
the direction of influence on production and efficiency depends on which form of the
relationship outweighs under the considered study area. It was therefore, hypothesized that
livestock affects the efficiency of the farmers positively or negatively depending on the sign and
magnitude of the coefficient.

The estimated technical inefficiency model indicates that the livestock coefficient has a negative
and statistically significant coefficient. It means that farmers who have more livestock are more
efficient than those having less livestock in malt barley production. The possible explanation for
the negative relationship between livestock owned and technical inefficiency could be that farm
households that own more livestock might have enjoyed more from the complementary
relationship. Such benefits for improving technical efficiency includes, manure, cash for
purchase of malt barley production inputs, animal power use for timely and adequately plowing
of malt barley plots ,and transportation of inputs. This result is also in complete agreement with
the findings of most of the studies reviewed (Mulat, 1989; Assefa, 1995; Mohammed, 1999;
Aynalem,2006; and Wondimagegn ,2010).

81
 5. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS

The objectives of this study were to measure the technical efficiency level and study the
determinants of malt barley production in Debark Woreda. The analysis was based on the cross
sectional data of 120 samples chosen from 885 sample households engaged in the Woreda 2010
malt barley program.The sample households were drawn in 3 stage sampling technique, in which
5 of the 8 kebeles that have sufficient participants were selected in the first stage. And then the 3
study kebelesusing simple random sampling method in the second stage, and lastly household
samples from the 3 kebeles using simple random sampling method proportionate to sizein the
third stage.

To achieve the objective of the study, both descriptive and econometric analyses were
undertaken using the EXCEL, SPSS and Frontier Version4.1 computer softwares. The existence
of inefficiency to opt for sfpf, the choice of appropriate stochastic functional form, and the joint
explanatory power of the considered inefficiency effect variables were formally tested with
respect to the survey data in order to proceed further with the analysis and connotations of
parameter estimates in the study being undertaken.

Accordingly, the stochastic frontier production function of the Cobb-Douglas functional form
was found to best fit the data. Parameters of the stochastic frontier function (from which
efficiency scores have to be measured) and inefficiency effects model were thus, estimated by
the maximum likelihood methods in a single stage estimation procedure using Frontier Version
4.1-computer program.

The estimated stochastic production frontier model indicates that area of the malt barley plot,
horsepower in plowing, number of plowing, and credit of other inputs (fertilizer and seeds) were
found to be the significant determinants of malt barley output level. The parameters of these
variables are interpreted as the partial output elasticity coefficients of their respective inputs

82
The analysis also reveals that the sum of the partial output elasticities with the respective input
is 1.16. This result indicates an increasing return to scale in malt barley production. The
implication of such a result is that a proportional increase in all the factors of productionleads to
a more than proportional increase in output. The result further reveals thatmalt barley farmers
can benefit from economies of scale linked to increasing returns to boost production.

The efficiency score results obtained from the parameter estimates of the frontier function ranged
from 0.38 and 0.99 with a mean technical efficiency level of 0.81. In addition, more than 43 % of
the farmers operated below the mean technical efficiency level. These results together show not
only the existence of 19% technical inefficiencies the elimination of which will lead to 19%
increase in malt barley yield with out any increase in the level of resources used, but also the
large portion of farmers who will be benefiting from the efficiency gain aspect of development.

The value of the discrepancy ratio, γ, obtained from the Maximum Likelihood estimation of the
frontier was 0.22 and is significantly different from zero. The significant result thus, indicates the
random component of the inefficiency effects does make a significant contribution in the
analysis of malt barley production in the study area. The corrected discrepancy ratio, γ*,
calculated from γ, which can be directly interpreted as the proportion of the inefficiency effect
variation relative to the sum of the variations of the inefficiency component and the random
noise was computed to be 0.09. This coefficient implies about9% of the variability of malt
barley output of farmers in the study area in the year 2010 was attributed to technical
inefficiency effect, while the remaining about 90% variation in output is due to the effect of
random noise. This might happened due to the serious natural hazards damage occurred in the
cropping season related to the areas frequent vulnerability for natural production calamities.

The socio-economic variables that are important in determining farmers' level of technical
efficiency were also identified. Accordingly, the result of the technical inefficiency effects model
shows; family size, age and its square, education, experience on malt barley production , plot
distance, fertility status, and livestock holdings are found to be the major determinants of
efficiency level of farmers in malt barley production. The negative coefficients of age, education,
and experience on malt barley production, fertility status, and livestock holdings mean these

83
factors positively affect efficiency of farmers. While the positive coefficients of age square,
family size and plot distance indicate that these factors determine efficiency negatively. In
general, the significant inefficiency effect explanatory variables have important policy and
development implications in an effort towards improving the technical efficiency of malt barley
production in the study area.

Finally, the following important recommendationsare given in the light of the findings of this
study.

1. Promote experience sharing on malt barley production

As shown in the analysis of the different socio economic inefficiency variables, the general
farming experience or knowledge acquired through age and the specific experiences on malt
barley production positively influenced the technical efficiency in malt barley production.
Hence, lessons from the elder households withlonger general farm experience and from
farmers who have more experience specific to malt barley production could be an important
input to improve the technical efficiency of the relatively less efficient malt barley farmers.
These experiences can be surveyed and incorporated in the malt barley extension program
and reach the needy malt farmers either through field visit or different training arrangements.
This may also help to improve the extension program, which currently has no significant role
in determining efficiency of malt barley production.

2. Improve farmers’ access and involvement in education

The results of the inefficiency model showed that those farmers that are educated are
relatively more technically efficient than uneducated ones, in the study area. This is likely;
educated farmers have better access and use of information and communication media that
helps them to use technologies more efficiently. Nevertheless, about 55% of the farmers in
the study area are illiterate that are not ableeven to read and write. The existence of high rate
of illiteracy behind the significant role of education in improving the farmers’ technical
efficiency shows that much more remains to work about educating farmers. Therefore, the

84
 regional government should strengthen the current adult and vocational educational programs
in the area.

3. Promote fertility improvement and maintenance measures

Fertile lands perform more efficiently hypothesis, was accepted by this study. However, the
current fertility situation of the area is generally poor despite the several soil and water
conservation efforts by the government and the community since long years. According to
this study only 20% of the sample households undertake their malt barley farming on
relatively fertile soil. The future technical efficiency improvement of malt barley production
in the area, there fore, calls for fertility improvements as infertile lands are currently
subtracting from potential malt barley yield. In this regard the government, besides
facilitating to replenish fertility with commercial fertilizer should also worry and device
appropriate strategies on how to make effective the soil conservation and other measures
intended to maintain and rehabilitate the fertility status of the area.

4. Exploit economies of scale

The sum of the partial elasticities with respect to all significant inputs is 1.16 in the frontier
function indicates that malt barley production of farmers in the area is also characterized by
increasing returns to scale. This in turn reveals apart from some productivity gains linked to
improvements in technical efficiency through the manipulation of the explained significant
inefficiency variables, farmers can also take the advantage of scale economies linked to
increasing returns to increase their output.

85
 6. REFERENCES

Abebe Dagnew, 2009.Technical Efficiency of Onion Production Under Irrigation: The case of
Kalu woreda of south Wollo,Ethiopia. An MSc Thesis Presented to the School of Graduate
Studies of Haramaya University

Abrar Suleiman, 1995. Technical efficiency of fertilized farms in Ethiopia: An application of

the stochastic frontier approach. Ethiopia. J. Development Research, 18 (2):1-30

Afriat, S.N., 1972. Efficiency estimation of production functions. International Economic

Review 13: 568-598

Aigner D., K. Lovell, and P. Schmidt, 1976. Formulation and Estimation of Stochastic
Frontier Production Function Models. Santa Monica, California p5649

Aigne, D. K., Lovell, and P. Schmidt, 1977. Formulation and Estimation ofStochastic Frontier
Production Function Models, Jornal of Econometrics ( 6): 21-37

Ajibefun, I.A., 2002. Analysis of Policy Issues in the Technical Efficiency of Small Scale
Farmers Using the Stochastic Frontier Production: With Application to Nigerian Farmers. Paper
Prepared for Presentation at the International Farm Management Association Congress,
Wageningen, The Netherlands.

Amadou,N.,2007.Analysis of Factors Affecting Technical Effficiency of Arebica Coffee

Producers in Cameron,AERC Research Paper,163,Nirobi

ANRS-BoFED, 2007.Annual Budget Year Statistical Bulletin, Amharic and English versions.

ANRS-BoFED,2010. Annual Budget Year Statistical Bulletin, Amharic and English versions.

Apezteguia B.I. and M.P. Garate, 1997.Technical Efficiency in the Spanish Agro Food Industry.
J. Agricultural Economics, 17, 179 – 189

ARARI, 2006.Amahara Regional Malt Barley Development Project. Amhara Regional State
Agricultural Research Institute, Bahir Dar, Ethiopia

Assefa Admassie, 1995. Analysis of production efficiency and the use of modern technology
in crop production: A study of smallholders in the central highlands of Ethiopia. Arbeiten Zur
Agrawirtschaft in Entwicklungslander. Wissenschaftsverlag Vauk Kiel KG

Aynalem Gezahegn,2006. Technical Efficiency in Maize Production: A Case of Mecha

Woreda. An M.Sc. Thesis Submitted to the School of Graduate Studies of Haramaya
University.
Banker, R., A. Charnes and W. Cooper, 1984. Some Models for Estimating Technical and
86
Scale Inefficiencies in Data Envelopment Analysis. Management Science, 30(9):1078 – 1092.

Battese G.E. and S.S. Broca, 1996. Functional Forms of Stochastic Production Functions &
Model for Technical Inefficiency Effects: A Comparable Study of Wheat Farmers in Pakistan.
CEPA Working Papers, No. 4 Department of Econometrics,University of New England,
Armidale

Battese, G.E. and G.S. Corra, 1977.Estimation of production frontier model: with application to
the pastoral zone of eastern Australia. Australian J. of Agricultural Economics 21:169-179

Battese, G.E. and T.J. Coelli, 1992. Frontier production functions: technical efficiency and
panel data: with application to paddy farmers in India, J. of Productivity Analysis 3: 153-169

Battese G.E, and T.J. Coelli, 1995. A Model for Technical Inefficiency Effects in a Stochastic
Frontier Production Function for Panel Data.Empirical Economics, 20, 325 – 332

Bravo – Ureta, B. E. and E. A. Pinheiro, 1993. Efficiency analysis of developing country

agriculture: A Review of the Frontier Function Literature. Agricultural and Resource
Economics Review 22(1): 88 – 101

Byerlee D., J. Davud, Dawit Alemu, and G. Mautam, 2007.Policies to Promote Cereal
Intensification in Ethiopia: A Review of Evidence and Experience, Development Strategy and
Governance Division.IFPRI Discussion Paper 00707 June 2007

Coelli T.J, 1995.Recent Development in Frontier Modeling and Efficiency

Measurement.Australian Journal of Agricultural Economics vol. 39, No.3, pp 119-245

Coelli T., 1996. A Guide to FRONTIER Version 4.1: A Computer Program for Stochastic
Frontier Production and Cost Function Estimation. CEPA Working Paper 96/07, University of
New England, Armidale, NSW, 2351, Austria.

Coelli T., and G. Battese, 1996. Identification of Factors which Influence the Technical
Inefficiency of Indian Farmers. Australian Journal of Agriculture, 40(2): 103-128

Coelli T.J, P. Rao and G. Battese, 1998. An Introduction to Efficiency and Productivity
Analysis. Kluwer Academic Publishers, Boston, Dordrecht/London, P, 134 – 249

Coelli T., R. Sandura and T. Colin, 2002. ‘Technical, Allocative, Cost and Scale Efficiencies in
Bangladesh Rice Production: A Non-Parametric Approach. Journal of Agricultural
Economics, 53: pp 607–626.

Croppenstedt A. Abbi Mammo, 1996. An Analysis of the Extent and Cause of Technical
Efficiency of Farmers Growing Cereals in Ethiopia: Evidence from Three Regions.Resource
Use. Ethiopian Journal of Agricultural Economics, 5 (2), 39 – 62

CSA (Central Statistics Authority),2004, Area and production forecasts of major crops:
Agricultural sample enumeration survey, Addis Ababa, Ethiopia.
87
CSA, 2005. Area and production forecasts of major crops: Agricultural sample enumeration
survey, Addis Ababa, Ethiopia.

CSA, Report of Federal Democratic Republic of Ethiopia, 2006. Statistical Report on Socio-
Economic Characteristics of the Population in Agricultural Households, Land Use Area and
Production of Crops. Addis Ababa: CSA, Pp.193-211.

CSA,2007. Area and production forecasts of major crops: Agricultural sample enumeration
survey, Addis Ababa, Ethiopia.

CSA,2008. Area and production forecasts of major crops: Agricultural sample enumeration
survey, Addis Ababa, Ethiopia.

CSA, 2010/11. Agricultural Sample Survey: Area and Production of Major Crops,Private
Peasant Households,Meher Season,.Addis Ababa, Ethiopia.

Debark Woreda Agricultural Office, 2010.Crop Production Assessment,Annual woreda

Agricultural Report(Amahric Version)

EPCC (Ethiopian Population Census Comission),2008. A summary and statistical report of the
2007 population and housing census. Addis Ababa, Ethiopia.

Fantahun Ameshe, 2010. Malt Barley Market Chain Analysis: The Case of Wogera Woreda. An
MSc Thesis Presented to the School of Graduate Studies of Haramaya University.

FAO/WFP, 2008. Crop and Food Security Assessment Mission (Phase 1). Special Report, FAO
–UN, January, 2008 ROME. pp13-26

Fare, R.S.Grosskopf and C.A.K. Lovell, 1985.Technical Efficiency of Philippine

Agriculture.Applied Economics,17:205-214.

Farrell, M.J., 1957. The measurement of productive efficiency.Journal of Royal Statistical

Society Series A., 120: 253-290

FDRE- Food Security Office and Oxfam-GB, 2007. Off-farm Opportunity in Food Insecure
Areas (Office Memo)

FDRE-MoFED, 2010. Ethiopia’s Growth and Transformation Plan, Addis Ababa, Ethiopia

FDRE-MoARD,2002.Agricultural and Rural Development Policies, Strategies, and Tactics,

Amaharic version, Birhanna Selam, Addis Ababa, Ethiopia

Fekadu Gelaw, 2004. Analysis of Technical Efficiency of Wheat Production: A Study in

Machakel Woreda. An MSc Thesis Submitted to the School of Graduate Studies of Alemaya
University.

88
Getu Hailu, 1997. Analysis of performance variation: The case of small holder farming in
Eastern Haraghe region. M.Sc. Thesis Presented to the School of Graduates Studies of Alemaya
University.

Getu Hailu, Vera Bitsch, Belay Kassa and Harmen Strock, 1998. Technical efficiency of
smallholder annual crop production in moisture stress area of Eastern Oromia of Ethiopia. A
stochastic frontier analysis.Ethiop. J Agricultural Economics, 2 (2): 91-115

Gujarati, D. N. 1995. Basic Econometrics. Second Edition. New York, MacGraw Hill, Inc.

Haileselassie, A., 2005. Analysis of technical efficiency in sorghum production: The case of
small holder farmers in Raya-Azebo Woreda. Thesis presented to School of Graduate Studies at
Haramaya University.

Hamed, G, R, Mahmoud, 2008. Estimation of Technical Efficiency of Wheat Farms: A Case

Study in Kurdistan Province, Iran. Department of Agricultural Economics, University of
Kurdistan, Sanadaj, Iran pp104-109

HST Consulting, 2007.Project feasibility study for the establishment of Dashen brewery barley
malt producing factory. Addis Ababa, Ethiopia.
.
Jema Haji, 2008. Economic Efficiency and Marketing Performance of Vegetable Production in
the Eastern and Central Parts of Ethiopia; Doctoral Thesis,No.2008:17, Swedish University of
Agricultural Sciences, Uppsala, Sweden.

Jondrow, J., K.C.A. Lovell, I. S. Materov and P.S. Schmidt, 1982.On the estimation of technical
inefficiency in the stochastic frontier model.Journal of Econometrics 19: 233-238

Khuda B., 2007. An Analysis of Technical Efficiency and Productivity of Growing Potato,
Carrot, Radish and Bitter Ground: A Case Study of Pakistan Punjab. Doctorial thesis submitted
to the faculty of Agricultural Economics and Rural Sociology of Agricultural University of
Faisalabad, 2007, pp 105-138.

Kinde Teshome, 2005. Technical efficiency of Maize Production: A Case of Smallholder

Farmers in Assosa Woreda. MSc thesis Presented to the School of Graduate Studies of Alemaya
University, April 2005

Kamruzzama and Mohammad, 2008. Technical Efficiency of Wheat Growers in some

Selected Sites of Dinajpur District. Bangladesh Journal of Agricultural Research, 33(3): 363-
373.
Kalirajan, K.P. and J.C. Flinn, 1983. The Measurement of Farm-Specific Technical Efficiency,
Pakistan Journal of Applied Economics, 2:167-180.
Kalirajan, P. and R.Shand,1986. Estimating location specific and firm specific technical
efficiency: An analysis of Malaysian Agriculture. J. Economics Development, 11: 147-160.

89
Llewelyn R.V and J. Williams, 1996. Nonparametric analysis of technical, pure technical and
scale efficiencies for food production in East Java, Indonesia. J. of Agricultural Economics, 15:
113-126

Maddala, G. S. 1977. Econometrics. Singapore, McGrow-Hill Book Co.

Mohammad Hassena, 1999. Factors Influencing Technical Efficiency of Crop Production in

Assassa Woreda of Southern Ethiopia. An MSc Thesis presented to the School of Graduate
Studies of Alemaya University.

Mohammad J and E. Premachandra, 2003. Sensitivity of Technical Efficiency Estimates to

Estimation Approaches: An Investigation Using New Zealand Dairy Industry Data. University
of Otago, Economics Discussion Papers No. 0

Mulat Demeke, 1989. The production efficiency of food grains producer in Ethiopia: The case
of small farmers in Aris Province. PhD. Dissertation, Department of Economics, University of
Strathcyde.UK

Neff D.L., P. Garcia & C.H. Nelson, 1993. Technical Efficiency: A Comparison of
Production Frontier Methods. J. Agricultural Economics, 3, 479 -489

ORDA (Organization for Rehabilitation and Development in Amhara)- Oxfam, 2008. Baseline
survey of six Malt Barley potential Woredas of north and south Gonder zones of the Amhara
Region final report, Ethiopia.

ORDA,2010. Second Quarter Report on Malt barley production in Amhara region

(unpublished),Bahir Dar.

Parikh A. and M.K. Shan, 1996.Various Approaches to Measurement of Technical Efficiency in

North-West Frontier Province of Pakistan. Pakistan Journal of Applied Econometrics, Vol. XII
No. 1, 31-65.

Seyoum, E.T., G.E. Battese and E.M. Fleming, 1998. Technical efficiency and productivity of
maize producers in eastern Ethiopia: a study of farmers within and outside Sasakawa-Global
2000 Project. J. of Agricultural Economics 19: 341 – 348

Stevenson, R. E., 1980. Likelihood Functions for Generalized Stochastic Frontier

Estimation.Journal of Economics, 13:57-66.

Strock H, Bezabih Emana, Berhanu Adenew, A. Borowiecki and Shimelis W/Harwariate,

1991.Farming systems and farm management practices of smallholders in the Hararghe
Highlands. Farming system in practices of resource economics in the tropics, Vol. 11,
Wissenchaftsverlog, Vauk, Keil KG, Germany
Temesgen Bogale, 2003. Efficiency of resource use in the production of irrigated potato: A
study in Awi zone. M.Sc. These Presented to the School of Graduates of Alemaya University.

90
Thiam A., B.E. Bravo-Ureta and T.E. Rivas, 2001. Technical Efficiency in Developing
Country Agriculture: A Meta-Analysis. J. Agricultural Economics, 25, 235 - 243

Timmer, C. P., 1971. Using a probabilistic frontier production function to measure technical
efficiency.Journal of Political Economy 79(4): 776-794

Wondimagegn Muche,2010. Analysis of Technical Efficiency of Wheat Production on

Vertisol, The case of Enebsie Sar Mieir woreda in East Gojjam, . An MSc Thesis Presented to
the School of Graduate Studies of Haramaya University.

World Bank, 2007.Opportunities and Constraints in Peasant Farming. World Development

Report, Working Paper New York. U.S.A

91
 7. APPENDICES

Appendix A:Different Appendix Tables

92
Appendix Table1. Participant Kebele Administrations and Farmers of the Woredain 2010 malt
barley production program.

Participant KAs Number of participant farmers

Milligebsa 306
Keyno 125
Abergina 145
Gomia 115
Adigagira 10
Abrham 9
Adagat 14
Total 885
Source: Debark Woreda Agricultural office (2011)

Appendix Table 2.Man Adult Equivalent conversion factor for labor use

Age group in years Male Female

<=10 0 0
11-13 0.2 0.2
14-16 0.5 0.4
17-50 1 0.8
>50 0.7 0.5
Source: Storck, et al. (1991)

Appendix Table 3.Conversion factors used to estimate Tropical Livestock Unit (TLU)

Animal category TLU equivalence

Calf 0.25
Weaned Calf 0.34
Heifer 0.75
Sheep and Goat 0.13
Cow and Ox 1
Sheep and Goat(young) 0.06
Horse/Mule 1.1
Chicken 0.013
Donkey 0.7
Source: Storck, et al. (1991)

Appendix Table 4.Individual technical efficiency estimates of Sample farmers

93
Farmer’s TE Farmer’s TE Farmer’s TE Farmer’s TE
code code code code

94
1 0.97 31 0.98 61 0.86 91 0.89
2 0.93 32 0.93 62 0.8 92 0.76
3 0.98 33 0.8 63 0.82 93 0.76
4 0.65 34 0.89 64 0.8 94 0.81
5 0.57 35 0.65 65 0.81 95 0.78
6 0.59 36 0.95 66 0.65 96 0.89
7 0.44 37 0.96 67 0.9 97 0.53
8 0.65 38 0.71 68 0.92 98 0.94
9 0.87 39 0.95 69 0.46 99 0.89
10 0.67 40 0.93 70 0.88 100 0.92
11 0.65 41 0.72 71 0.97 101 0.55
12 0.76 42 0.88 72 0.54 102 0.51
13 0.83 43 0.91 73 0.95 103 0.73
14 0.38 44 0.8 74 0.96 104 0.94
15 0.76 45 0.96 75 0.83 105 0.98
16 0.87 46 0.81 76 0.96 106 0.72
17 0.95 47 0.63 77 0.63 107 0.97
18 0.7 48 0.96 78 0.61 108 0.98
19 0.73 49 0.73 79 0.97 109 0.97
20 0.87 50 0.72 80 0.4 110 0.96
21 0.88 51 0.52 81 0.89 111 0.98
22 0.97 52 0.68 82 0.63 112 0.81
23 0.65 53 0.95 83 0.99 113 0.87
24 0.97 54 0.78 84 0.62 114 0.93
25 0.95 55 0.98 85 0.88 115 0.8
26 0.95 56 0.73 86 0.7 116 0.96
27 0.93 57 0.78 87 0.92 117 0.72
28 0.53 58 0.95 88 0.86 118 0.9
29 0.59 59 0.68 89 0.9 119 0.74
30 0.41 60 0.93 90 0.95 120 0.87
Source: Own survey (2011)

Appendix Table5.Distribution of sample farmers by technical efficiency levels

TE range Frequency Percentage Cumulative

percentage
0.3-0.39 1 0.8 0.8

95
0.4-0.49 4 3.3 4.1
0.5-0.59 9 7.5 11.6
0.6-0.69 14 11.7 23.3
0.7-0.79 19 15.8 39.1
0.8-0.89 28 23.4 62.4
0.9-1.00 45 37.5 100
Total 120 100
At mean TE 4 3.3 3.3
Below mean TE 52 43.3 46.6
Above meanTE 64 53.4 100
Source: Own survey (2011)

Appendix Table 6.Actual vis-a-vis potential malt barley productivity

Mean STD Min Max

Actual 15.18 6.77 3 48
productivity(qt/ha)
Potential 18.59 6.54 5.5 49
productivity(qt/ha)
Source: Own survey (2011)

Appendix Table7. Variance Inflation Factors ofDependant Variables of the study

Explanatory VIF 1/VIF R2

96
Variable

MB AREA 1.19 0.839 0.161

LAB 1.30 0.768 0.232
DRFT POWER 1.98 0.540 0.496
CRDTINPT 1.76 0.568 0.432
FRQ PLW 2.36 0.424 0.576
FRQ WDG 1.09 0.918 0.082

FAM SZ 2.68 0.373 0.627

AGE 1.75 0.570 0.430
LV STCK 1.61 0.620 0.380
DIST 1.50 0.665 0.335
FERT 1.31 0.765 0.235
FRM SZ 1.26 0.791 0.209
EXT CTCT 1.17 0.854 0.145
MB EXP 1.45 0.690 0.310
EDUC 1.73 0.578 0.421

Source: Own computation result (2011)

97
 Appendix B. Questionnaire used in the survey

HARAMAY UNIVERSITY
SCHOOL OF GRADUATE STUDIES
DEPARTMENT OF AGRICULTURAL ECONOMICS

Questionnaire for Undertaking Technical Efficiency Study in Malt Barley

Production: The Case of Smallholders in Debark Woreda of North Gondar Zone
of ANRS.

Reminders for Enumerators:

Ourcountry’s future development above all depends on how accurate information you
provide !
Introduce yourself and tell the purpose of the study before starting the interview
Put (X) mark for closed questions where appropriate and use the space provided for open-
ended questions

Part 1 . Farm Level Information

1. General
Kebele _____________ Name of household head ___________Code of the household
_________Farm experience______ years
Name of enumerator_________________________ Date_______________
Signature _______________Telephone________________________

2. Demographic& Education

Family members ‘P’for

Relationshi Responsibilit
No (including the household Sex Age Educ presence‘N’
p y
head) fornot in2010
1 :Hhh
2
3
4
5
6

Education: Illiterate=1;Only R & W=2; literate= grade completed

Sex: M = male, F= female

98
Responsibility: Farming (ploughing, weeding, sowing harvesting)=1,Petty trading=2, Look after
livestock=3 ; More than one of the above=4;Other (specify)=5;

Relationship: 1= Father; 2= Mother; 3= Daughter; 4= Son; 5= Other kinships

Hhh: Household head

3. Infrastructural: Home distance to Infrastructure

Type of infrastructure Distance in walking

minutes
All-weather road
Nearest market
Cooperative office
Farmer training center
Health center
School
DA office/FTC

4. Land use and possession in 2010

 Total Area of land ………… (ha) : own( excluding rented out land)
____rented in____ obtained through share cropping____
 Land use(ha): Homestead_____ Annual cultivated ______Perennial_____
Fallow_________ Grazing_____ Others_____
 Number of plots ( total ____) :
Malt barley ___ other annual crops ____ Perennial crops ___
Fallow_______
 If you have rented outland,______ ha ______ plot/s

99
5. Livestock holding

Purpose for maintaining A: saleB: Remark

Class of livestock Number consumption C: Draught D: Others
( wealth ,culture, …)
Cows
Oxen
Heifers
Bulls
Calves
Sheep
Goat
Donkey
Horse
Mule
Poultry
Beehives

6. Institutional services

6.1. Credit use

6.1.1. Credit taken and repayment situations of the farmer in the past 3 years

Purpose
Credit
Othe Remark
Pro & Plant
Other r
N d. repayme Protecti
DA Ure See agricultu non
o yea nt on
P a d ral farm
r conditio chemica
credits credi
ns ls
ts
200 Amount
8 taken(Bir
r)
Timely
repaid
Remaine
d unpaid
Amount Dap,Urea,Seed,Chem
taken(kg) icals
200 Amount
9 taken
Timely

100
 repaid
Remaine
d unpaid
Amount Dap,Urea,Seed,Chem
taken(kg) icals
201 Amount
0 taken
Timely
repaid
Remaine
d unpaid
Amount Dap,Urea,Seed,Chem
taken(kg) icals

6.1.2. If there is a deliance or default in full credit repayment, what are the reasons/?

A. Production failure B. Market problem C. Loose credit follow up & poor credit
repayment arrangements of the lender D. Expectation of debt cancellation by the
farmer E. Others(specify)

6.1.3. How much did you borrow for malt barley production in2010 (Birr)?

1. fertilizer--------- 2. Seed------------- -3.Herbicide----- 4. Pesticides----------------

6.1.4. Which of these is/are the source(s) of your credit:

A. Cooperative B. Amhara Credit and Saving Institution (ACSI) C. Agricultural and Rural
development office D. Commercial banks E. Private banks F. individual Moneylenders G.
Others (specify)

6.1.5. Major problems in credit use

A. inadequate amount of credit B. high interest rate C. untimely credit supply E. Lack of
adequate kind credit F. Other (specify)

6.2. Extension services

6.2.1. Do you get agricultural extension services? Yes___ No____If yes, since when have
you started? _______

6.2.2. If yes, average frequency of contact with DA in 2010?_______ times in______month/s.

6.2.3. Is there farmer-training center in your area? 1. Yes 2. No

6.2.4. When have you started malt barley production? ________and how many trainings have
you taken so far? ____

101
6.2.5. How much malt barley you would have produced in 2010 had there no been an extension
service?
A. the same as I produced B. a little lower than I produced C. much lower than I produced D.
totally impossible to produce without the extension service E. I couldn’t determine this way

6.2.6. What are the major agricultural extension service problems in the area? ______________
____________________________________________________________________________

6.3. Marketing conditions

6.3.1. How many quintals of malt barley did you produce in 2010? _________

6.3.2. Purpose of production

A. all for sale B. all for consumption C. Partly for consumption and partly for sale
6.3.3. What was the average selling price? ______ birr/qt

6.3.4. Do you have unsold malt barley waiting for good market season? Yes___ No___ .
6.3.5. If yes, how many? ______ quintals. In which month/s do you expect to
sell_________________________________ & at what price? _________ birr/qt.
6.3.6. If yes, what marketing problems are you facing?

A. Low price B. High seasonal price fluctuation C. Inadequate demand D. High marketing
cost E. Lack of market information F. Credit sale

6.3.7. Again, if yes what do you think will be the solution? __________________________

7. Off-farm opportunity & wealth conditions

7.1.Does the household participate in off-farm activities? 1) Yes 2.No

7.2. If yes, what type of activity you are doing?

A. handcrafting B. Petty trading C. Service rendering D. processing and resource extraction

E. wage employment

7.3. Who participate in the off-farm activities? 1. Me 2.Wife 3.Children 4. Other

specify---------------------------------- , and how many members of the family? ______
persons.

7.4.On average,how many days, each family member spends? 1. One day in a week 2. Two
days in a week 3.three days & above in a week

7.5. What is the relative wealth position of the farmer? (categorized by peer groups)
A. Very rich B. Rich C. Medium D. Poor E. Very poor

102
7.6. Which of the following assets are owned in the household? A. Corrugated Iron roof house
B. Mobile C. Television D. Radio E. House in town E. Cash saving

103
7.7. Income, expenditure, and saving conditions of the household in 2010

Income Expenditure& Saving Remark

Details Amount( Details Amount(
birr) birr)
Off-farm incomes Food
 Non-farm Clothes
 Employment Health
 Unearned(food Education
aid,remittance,
pension etc.)
Farm incomes Social obligations
 Annual crop Others
 Perennial crop Total expenditure
 Livestock Saving

8. Holidays

How many holidays per month do you consistently celebrate being out of main farming
activities like ploughing, weeding?. Saturday and Sunday ----------days, Others------- days

104
 Part 2. Plot Level Information
1. Plot characteristics, types of crops grown and amounts of production in 2010
Ownership Production(qt)
Soil type
1=own Soil Slope Crops
Plot Plot 1.red
Plot 2=rented in (price/ha) fertility 1=plain grown on
size distance 2.black Main By-
No. 3=share cropped 1=good 2=moderate the plot in
(ha) (walking minutes) 3.brown prod. product
(amount of share to the 2=low 3= steep 2010
4.other specify
operator)

2. Production Inputs used in a plot in 2010

Fertilizer Manure/ Pesticide Farming practice
Pl Plot Herbi
(kg) compost Seed 1.good
ot size cide
1= used (kg) 2.medium
No (ha) Dap Urea (Lt) Lt Kg
0=not used 3.low

105
 3. Labor and oxen use summary

Pair of Family labor(Number) Hired and exchange labor(Number) Plot no

horse Men Women Children Men Women Children & size
Activities
power(n 14-16 14-16
17-50 >50 17-50 >50 11-13 17-50 >50 17-50 >50 11-13
o.) M F M F
Plowing
Sowing
Weeding Plot 1__
Chemical application ha
Harvesting
Threshing
Plowing
Sowing
Weeding Plot 2
Chemical application ___ ha
Harvesting
Threshing
Plowing
Sowing
Weeding Plot 3
Chemical application ___ ha
Harvesting
Threshing
Plowing
Sowing
Weeding Plot 4
Chemical application ___ ha
Harvesting
Threshing

106
 4. Optimality of major agricultural operations in malt barley production

Frequencies of
Timeliness of operation (period) Plot no &
Activities operation(number)
Optimal Ac Optimal (from__ to __) Actual (from__ to __) size
Plowing
Sowing p
Weeding Plot- 1
Chemical appl. ___ha
Harvesting
Threshing
Plowing
Sowing Plot- 1
Weeding ___ha
Chemical appl.
Harvesting
Threshing

5. Production

5. 1. How many quintals of malt barley did you ever obtain?

Plot-1 Plot-2
Highest _________ ________
Medium _________ ________
Lowest _________ ________

5.2. What were the causes for getting this lowest yield?_________________ .

6. Natural and Artificial Hazards

6.1. What hazard you faced in 2010 malt barley production if there is any?
A. Storm B. Frost C. Flood D. Drought E. Animal damage F. Disease G. High weed
infestation

6.2. How much output you would have got, had there been no hazard?
Plot 1_____qt, Plot 2____ qt

6.3. What is the best crop rotation from your experience?

1. Malt barley _________2. ________3.________ 4._________ 5.________

7. What is your future plan about malt barley production area?

A. increase B. decrease C. No change D. depends on the conditions

107