Zone, Ethiopia' prepared, under my guidance by Moges Dessale. I recommend that ... University Library and is made available to borrowers under rules of the Library. I ... CRS. Constant Return to scale. DEA. Data Envelopment Analysis. DMU.
TECHNICAL EFFICIENCY IN TEFF PRODUCTION: THE CASE OF SMALLHOLDER FARMERS IN JAMMA DISTRICT, SOUTH WOLLO ZONE, ETHIOPIA
MSc THESIS
MOGES DESSALE
MAY 2017 HARAMAYA UNIVERSITY, HARAMAYA
Technical Efficiency in Teff Production: The Case of Smallholder Farmers in Jamma District, South Wollo Zone, Ethiopia
A Thesis Submitted to the School of Agricultural Economics and Agribusiness, Postgraduate Program Directorate HARAMAYA UNIVERSITY
In Partial Fulfillment of the Requirements for the Degree of MASTER OF SCIENCE IN AGRICULTURAL ECONOMICS
Moges Dessale
May 2017 Haramaya University, Haramaya
HARAMAYA UNIVERSITY POSTGRADUATE PROGRAM DIRECTORATE I hereby certify that I have read and evaluated this Thesis entitled ‘Technical Efficiency in Teff Production: The Case of Smallholder Farmers in Jamma District, south Wollo Zone, Ethiopia’ prepared, under my guidance by Moges Dessale. I recommend that it be submitted as fulfilling the thesis requirement. Bosena Tegegne (PhD)
_______________
Major Advisor Hassen Beshir (PhD)
Signature ________________
Co-advisor
Signature
______________ Date ______________ Date
As member of the Board of Examiners of the MSc Thesis Open Defense Examination, I certify that I have read and evaluated the Thesis prepared by Moges Dessale and examined the candidate. I recommended that the Thesis be accepted as fulfilling the Thesis requirement for the Degree of Master of Science in Agricultural Economics. _______________________
_______________
Chairperson _______________________
Signature _______________
Internal Examiner _______________________
Signature _______________
External Examiner
Signature
______________ Date ______________ Date ______________ Date
Final approval and acceptance of the Thesis is contingent up on the submission of its final copy to the Council of Graduate studies (CGS) through the candidate’s department or school graduate committee (DGC or PGPD).
ii
DEDICATION This work is dedicated to my lovely mother Mushira Semaw , my sister Zemed Dessale and my son Girma Moges.
iii
STATEMENT OF THE AUTHOR By my signature below, I declare and affirm that this Thesis is my own work. I have followed all ethical principles of scholarship in the preparation, data collection, data analysis and completion of this Thesis. All scholarly matter that is included in the thesis has been given recognition through citation. This Thesis is submitted in partial fulfillment of the requirements for MSc degree in Agricultural Economics at Haramaya University. The Thesis is deposited in Haramaya University Library and is made available to borrowers under rules of the Library. I solemnly declare that this Thesis has not been submitted to any other institution anywhere for the award of any academic degree, diploma, or certificate. Brief quotations from this Thesis may be made without special permission provided that accurate and complete acknowledgement of the source is made. Requests for permission for extended quotation from or reproduction of this Thesis in whole or in part may be granted by the head of the school or department when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author of the Thesis.
Name: Moges Dessale
Signature: ______________________
Date: May 2017 School: Agricultural Economics and Agribusiness
iv
BIOGRAPHICAL SKETCH The author was born in May 17, 1992 at Addis Alem Kebele, in Raya Kobo District of North Wollo Zone, Amhara National Regional State, Ethiopia. He attended his elementary education at Robbit Elementary School and followed his senior education at Robbit Secondary School and Kobo Secondary and preparatory school. After successful completion of his high school education, he joined Wollo University in November 2011 and graduated with Bachelor of Sciences Degree in Agricultural Economics in July 2014. Since his graduation he has served in Wollo University as Graduate Assistance. Then after, he joined Haramaya University in September 2015 to pursue his MSc degree in Agricultural Economics.
v
ACKNOWLEDGEMENTS I would like to forward the deepest of my appreciation and genuine gratitude to my major advisor Dr. Bosena Tegegne for her intensive guidance, invaluable comments, innumerable revisions and reorganizations at all stages of the work of this thesis. I am also grateful to my co-advisor, Dr. Hassen Beshir for his advice, guidance, constructive comments and suggestions during the planning and preparation of this thesis work. The work of this thesis would not have been possible without the financial sponsorship of Wollo University and Facilitation of Haramaya Universty. I would like to express my Special gratitude to them for their understanding and considerable financial support. Besides, I wish to convey my heartfelt thanks to Mengesha Dessale for multidimensional support and continuous encouragement in my endeavor. I extend my profound appreciation to the farmers of Jamma District for their hospitality who willingly participated in the survey and spent many hours explaining their livelihoods, which I never forget. Finally, I extend my gratitude to my sister, who was the source of special strength towards the successful completion of the study. Appreciation is also there to my sister Zemed Dessale for her constant encouragement and help in all my endeavors. Without her support, completion of the study would have not been possible.
vi
ACRONYMS AND ABBREVIATIONS ACSI
Amhara Credit and Saving Institution
AE
Allocative Efficiency
ATVET
Agricultural Technical and Vocational Education and Training
BCC
Banker Charnes Cooper
CCR
Charnes Cooper Rhodes
CSA
Central Statistical Agency
CRS
Constant Return to scale
DEA
Data Envelopment Analysis
DMU
Decision Making Units
DAs
Development Agents
DAP
Di Ammonium Phosphate
EE
Economic Efficiency
HHH
Household Head
GDP
Gross Domestic Production
KAs
Kebele Administrations
MD
Man-day
ML
Maximum Likelihood
MLE
Maximum Likelihood Estimation
MASL
Meter above Sea Level
MOA
Ministry of Agriculture
MOFED
Minister of Finance and Economic Development
NBE
National Bank of Ethiopia
OD
Oxen Day
OLS
Ordinary Least Square
Qt
Quintal
SFA
Stochastic Frontier Analysis
SPF
Stochastic Production Frontier
TE
Technical Efficiency
TLU
Total Livestock Unit
WOA
Woreda Office of Agriculture
UN
United Nations
vii
TABLE OF CONTENTS STATEMENT OF THE AUTHOR
iv
BIOGRAPHICAL SKETCH
v
ACKNOWLEDGEMENTS
vi
ACRONYMS AND ABBREVIATIONS
vii
LIST OF TABLES
x
LIST OF FIGURS
xi
LIST OF TABLES IN THE APPENDIX
xii
ABSTRACT
xiii
1. INTRODUCTION
1
1.1. Background of the Study
1
1.2. Statement of the Problem
2
1.3. Objectives of the Study
4
1.4. Significance of the Study
5
1.5. Scope and Limitation of the Study
5
1.6. Organization of the Thesis
6
2. LITERATURE REVIEW
7
2.1. Concept and Definition of Efficiency
7
2.2. Theoretical Framework for Efficiency of Production
9
2.2.1. Input-oriented efficiency measures 2.2.2. Output-oriented efficiency measures
9 11
2.3. Models of Efficiency Measurement
13
2.3.1. Non-Parametric Frontier Model
13
2.3.2. Parametric Frontier Models
14
2.4. Approaches to Identifying Efficiency Factors
16
2.5. Review of Empirical Studies on Technical Efficiency Abroad
17
2.6. Empirical Studies on Technical Efficiency in Ethiopia
20
2.7. Conceptual Framework of Technical Efficiency of Teff Production
24
3. RESEARCH METHODOLGY
27
3.1. Description of the Study Area
27
3.2. Data Types, Sources and Methods of Data Collection
28
3.3. Sampling Technique and Sample Size
29
Continues viii
3.4. Method of Data Analysis
30
3.5. Specification of the Empirical Model
30
3.6. Selection of the Functional Form
32
3.7. Definition of Input and Inefficiency Variables and their Hypotheses
35
3.7.1. Input Variables
35
3.7.2. Definition of Inefficiency Variables and Hypothesis
36
4. RESULTS AND DISCUSSION
41
4.1. Descriptive Results
41
4.1.1. Demographic and Socio-Economic Characteristics of Sample Households
41
4.1.2. Labor Availability and Gender Role
43
4.1.3. Resource Basis
44
4.1.4. Cropping System
47
4.1.5. Description of Production Function and Variables
47
4.1.6. Major Teff Production Constraints Faced by Sample Household Heads
48
4.1.7. Institutional Support
49
4.2. Econometric Results
50
4.2.1. Hypotheses Testing
51
4.2.2. Maximum Likelihood Estimation of Parameters
52
4.2.3. Variability of Output due to Technical Efficiency Differentials
54
4.2.4. Technical Efficiency of Farmers
54
4.2.5. Input Use and Technical Efficiency
55
4.2.6. Estimated Actual and Potential Level of Teff Output
56
4.2.7. Determinant of Technical Efficiency
57
5. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
62
5.1. Summary
62
5.2. Conclusions and Recommendations
64
6. REFERENCES
66
7. APPENDICES
74
ix
LIST OF TABLES Table
page
Table 1. Summary of definition, measurement and hypothesis of inefficiency variables
40
Table 2. Demographic and socio-economic characteristics of sample households
42
Table 3. Distribution of sample households by their marital status
43
Table 4. Level of pre-harvest labour in teff production by gender (in Man Equivalent)
43
Table 5. Distribution of sample households by their Land holding structure
44
Table 6. Number of plots and area coverage of major crops sown by sample farmers
45
Table 7. Number of livestock owned by sample household heads
45
Table 8. Distribution of sample household heads under various tropical local unit
46
Table 9. Distribution of sample household head by oxen and teff crop coverage
46
Table 10. Average yield obtained by sample farmers in 2015/2016 production year
47
Table 11. Descriptive statistics of variables used in production function estimation
48
Table 12. Major teff production constraints during production year of 2015/2016
49
Table 13. Institutional characteristics of the sample household
50
Table 14. Generalized likelihood-ratio test of hypotheses for model and parameters
52
Table 15. Maximum likelihood estimate for Cobb-Douglas production function
53
Table 16. Summary of technical efficiency differentials among sample household heads 55 Table 17. Utilization of production inputs and technical efficiency differentials
56
Table 18. Comparison of estimated actual yield and potential teff yield
57
Table 19. Maximum-likelihood estimates of technical inefficiency determinants
58
x
LIST OF FIGURS Figure
page
Figure 1: Input-oriented measures for technical, allocative and economic efficiencies
10
Figure 2: Output-oriented measures for technical, allocative and economic efficiencies
12
Figure 3. Conceptual framework of the study
26
Figure 4. Geographical location of the study area
28
xi
LIST OF TABLES IN THE APPENDIX Appendix Table
Page
1. Conversion factors used to compute Adult equivalent and man equivalent
75
2. Conversion factors used to estimate Tropical Livestock Unit (TLU)
75
3.Variance inflation factor for the input variables entered in to the SPF model
75
4. Variance inflation factor for the inefficiency variables
76
5. Contingency coefficient of dummy variables
76
6. The VIF for the continuous variables used in inefficiency variables
76
7. Educational status of the sampled farmers
77
8. Reasons of sample farmers for not using credit facilities
77
9.Technical efficiency estimate of sample farmers
77
10. General characteristics of sample household heads
78
11. The frequency distribution of farmers by technical efficiency scores
78
12. Distribution of sample farm household heads per Kebeles
78
13. Hetroskedasticity test
78
xii
Technical Efficiency in Teff Production: The Case of Smallholder Farmers in Jamma District, South Wollo Zone, Ethiopia ABSTRACT Teff is one of the dominant crops and its productivity is low in the study area. This means that it is possible to obtained additional output from existing inputs used, if resource are properly used and efficiently allocated. The aim of this study was to determine the level of technical efficiency of smallholder teff producers and identify factors affecting technical efficiency of smallholder farmers in teff production of Jamma district, South Wollo Zone, Ethiopia. A three-stage sampling technique was employed to select 149 sample farmers who were interviewed using a structured schedule to obtain data pertaining to teff production during 2015/2016 production year. A Cobb-Douglas stochastic frontier production analysis approach with the inefficiency effect model was used to estimate technical efficiency and identify the determinants of efficiency of teff producing farmers. The maximum likelihood parameter estimates showed that teff output was positively and significantly influenced by area, fertilizer, labor and number of oxen. This would mean that there is a room to increase teff output from the existing level if farmers are able to use these input variables in an efficient manner. The result further revealed that there were significant differences in technical efficiency among teff producers in the study area. The discrepancy ratio, which measures the relative deviation of output from the frontier level due to inefficiency, was about 75.6%. This implies that about 75.6% of the variation in teff output among the sample farmers was attributed to technical inefficiency effects. The estimated mean levels of technical efficiency of the sample farmers were about 78%. This shows that there exists a possibility to increase the level of teff output by 22% through efficiently utilizing the existing resources. The estimated stochastic production frontier model together with the inefficiency parameters showed that, age, education, improved seed, training and credit were found to have negative and significant effect on technical inefficiency while farm size was found to have positive and significant effect on technical inefficiency of teff production. Hence, local government should provide necessary supports such as formal as well as informal education, training, credit, improved seed and timely supply of fertilizer. Key words: Jamma district; stochastic frontier analysis; Technical efficiency; teff
xiii
1. INTRODUCTION 1.1. Background of the Study A large majority of Ethiopians and the poor living in rural areas are deriving their livelihood from agriculture. The proportion of the population of Ethiopia residing in rural areas in 2040 is predicted to be nearly 70%, when there will be 40 percent more rural residents (UN, 2014). Agriculture in Ethiopia is dominated by smallholder farming households, which cultivated 94% of the national cropped area in 2013/14 (CSA, 2014a). Growth in agriculture was one of the major drivers of the remarkable economic growth recorded in Ethiopia in the last decade (National Bank of Ethiopia, 2014). The major grain crops grown in Ethiopia are teff, wheat, maize, barley, sorghum and millet. Out of the total grain production, cereals account for roughly 60% of rural employment and 80% of total cultivated land (Abu and Quentin, 2013). In the crop production sub-sector, cereals are the dominant food grains. The major crops occupy over 8 million hectares of land with an estimated annual production of about 12 million tons (CSA, 2015). The potential to increase productivity of these crops is very high as it has been demonstrated and realized by recent extension activities in different parts of the country. However, population expansion, low productivity due to lack of technology transfer and decreasing availability of arable land are the major contributors to the current food shortage in Ethiopia (Hailemariam, 2015). According to CSA (2015), Ethiopian population will exceed 126 million by the year 2030. This increase in population will impose additional stress on the already depleted resources of land, water, food and energy. Teff is an important crop in terms of cultivated area, share of food expenditure, and contribution to gross domestic product. Despite the remarkable growth in teff production in the last decade, the drivers of this growth are not well understood. In particular, there is a lack of evidence on the contribution of improvements in productivity to this growth and the link between farm size and productivity. Moreover, doubts exist on whether it is possible to sustain such growth on landholdings that are declining in size (Fantu et al., 2015b). Teff accounted for about a fifth of the nationwide agricultural area and was cultivated by nearly half of smallholder farmers during the 2004/05-2013/14 period (CSA, 2014a).
2 During the same period teff output grew at average annual rate of 9.3% and yields grew at 5.2% (CSA, 2014b). Evidences indicate that part of the growth in teff output has been driven by increases in cultivated area (Dorosh et al., 2015). However, it is not well understood whether there were improvements in productivity and the contribution of such improvements for growth in output various staple crops dominate different parts of Ethiopia. However, teff is either the principal staple or among the most consumed crops in almost all parts of the country. Moreover, the demand for teff is elastic with respect to income. The share of spending on teff in food expenditure is highest in urban areas and increased by 3.4 % nationwide between 2005 and 2013, during which time real income increased considerably and the share of all other cereals declined (Worku et al., 2014). As it is one of the most common consumed cereals in Ethiopia, it has been historically neglected compared with other staple crops. Furthermore, approximately 6 million households grow teff and it is the dominant cereal crop in over 30 of the 83 high-potential agricultural districts. In terms of production, teff is the dominant cereal by area planted and second only to maize in production and consumption. However, yields are relatively low (around 1.4 ton/ha.) and high loss rates (25-30% both before and after harvest) reduce the quantity of grain available to consumers by up to 50% (CSA, 2014b). This holds true for the region in which this study was undertaken. Teff is the most widely adapted crop compared to any other cereal or pulse crop in the study area and is grown under wider agro-ecologies (variable rainfall, temperature and soil conditions) (WOA, 2015).
1.2. Statement of the Problem In Ethiopia agricultural production and productivity is very low and the growth in agricultural output has barely kept pace with the growth in population. The high potential areas of Ethiopia can produce enough grains to meet the needs of the people in the deficit areas. However, the inefficient agricultural systems and differences in efficiency of production discourage farmers to produce more (Knife et al., 2012). Gains in agricultural output through improvement of efficiency levels are becoming particularly important now a day. The opportunities to increase farm production by bringing additional forest land into cultivation or by increasing the utilization of the physical resources have been diminishing.
3 In addition, eliminating existing inefficiency among farmers can prove to be more cost effective than introducing new technologies as a means of increasing agricultural output and farm household income (Wondimu et al., 2014). The smallholder farmers in the north eastern Ethiopia are poor, individual land holding ranges between 0.5 and 2.5 hectares, large family sizes, land productivity is low and household food requirements are not fully met. The smallholder cereal-based farming systems have also remained traditional and non-commercial oriented. Thus, the system is unable to sustain the ever increasing population with food and energy demands. Therefore, an ever increasing population pressure and environmental degradation followed by declining productivity and expansion of marginal agricultural lands necessitates farmers either to use modern technologies or need to use resources efficiently in order to optimize outputs in the North Eastern Ethiopia (Mekonnen et al., 2015). According to previous researches in Ethiopia ( for example Getachew et al., 2014; Musa et al., 2014; Hassen, 2014; Wondimu and Hassen, 2014), there also exists a wide cereal yield gap among the farmers that might be attributed to many factors such as lack of knowledge and information on how to use new crop technologies, poor management, biotic, climate factors and more others (Mesay et al., 2013; Sisaye et al., 2015). Because of the scanty resources that are on ground, recently it is getting importance to use these resources at the optimum level which can be determined by efficiency searches (Gebregziabher et al., 2012). Thus, increasing crop production and productivity among smallholder producers requires a good knowledge of the current efficiency/inefficiency level inherent in the sector as well as factors responsible for this level of efficiency/inefficiency (Essa et al., 2012). Though there have been various empirical studies conducted to measure efficiency of agricultural production in Ethiopia, (for example, Hassen et al., 2011; Hassen et al., 2012a; Hassen et al., 2012b; Shumet, 2012; Hassen et al., 2013; Mesay et al., 2013; Beyan et al., 2013; Hassen, 2013; Tefera et al.,2014; Fantu et al., 2015a; Fantu et al., 2015b; Hailemaryam, 2015; Hassen, 2016; and Wudineh and Endrias, 2016), to the best of the author’s knowledge, there were no similar studies undertaken on technical efficiency of teff producing household in the study area. Moreover, since social development is dynamic, it is imperative to update the information based on the current productivity of farmers. However, the productivity of agricultural system in the study area is very low. The poor production and productivity of crop and livestock resulted in food insecurity.
4 The Jamma WOA (2015) report showed that, about 35,064 hectares land was covered by cereal crops. Out of which, teff is the leading one. However, there is no empirical study that showed whether the existing scare resources and technologies are utilized efficiently or not in the production of teff. The extent, causes and possible remedies of inefficiency of smallholders are not yet given attention. Therefore, this study was concerned with analysis of technical efficiency in teff production and provides valuable information so as to make an intervention in order to increase production and productivity of teff in Jamma district of Amhara region, Ethiopia. Despite its potential, Jamma District’s agricultural productivity is declining (CSA, 2012). Therefore, the need for the efficient allocation of productive resources cannot be overemphasized. However, in areas where there is inefficiency trying to introduce new technology may not bring the expected impact, unless factors associated with inefficiency among farmers are indentified and acted upon. The existence of inefficiency in production comes from inefficient use of scarce resources. The measurement of efficiency in agricultural production is important issue for agricultural development and it gives useful information for making relevant decision in the use of these scare resources and for reformulating agricultural policies. Thus, this study has attempted to generate information for policy implication by identifying factors that are associated with technical efficiency in teff production in Jamma district. Therefore, the study filled this information and knowledge gap at the study area.
1.3. Objectives of the Study The general objective of the study was to analyze technical efficiency in teff production of smallholder farmers in Jamma district. The specific objectives of the study were: 1. To estimate the level of technical efficiency in teff production of smallholder farmers in the study area, 2. To identify factors affecting the level of technical efficiency in teff production among farmers in the study area.
5
1.4. Significance of the Study The study was focus on the issue of technical efficiency in teff production and identifies factors associated with technical efficiency among farmers. It can play a significant role in providing useful information concerning technical inefficiencies in production and by identifying those factors, which were associated with inefficiencies that may exist. It can also indicate an entry point for further policy interventions to technical efficiency of smallholder farmers. Therefore, this study is expected to generate adequate understanding of the issues that might lead towards taking appropriate actions for improvement of efficiencies. Hence, the outcome of this piece of work can have important implications for the professionals and for the policy formulation purposes. Therefore, in the view of the above narrated importance of knowing the factors of inefficiency of production, the study will have significant important as follows; first, the result will provide useful information for the government and policy makers regarding the key factors affecting production. Thus, it will contribute to designing appropriate polices and strategies to increase teff production. Secondly, the study will also contribute to useful information for other grain crops that usually have similar production processes for farm households and helps in designing teff extension package in the context of the zone and region as well as the national level. Finally, it will serve as source for future empirical literature for scholars and students interested in the area of efficiency and in the field of agricultural economics and related fields.
1.5. Scope and Limitation of the Study This study focused on technical efficiency in teff production during meher season in one district, using cross sectional data of the 2015/16 production year collected from 149 teff producing smallholder farmers. The other limitation was related with the methodology used. The study does not show inter temporal differences in technical efficiency level of teff producing farmers. In addition, the study is limited to the analysis of technical efficiency of teff production without regard to other crops. Moreover, the study is limited to only Jamma district of South Wollo Zone, Amhara National Regional State, Ethiopia.
6
1.6. Organization of the Thesis The rest of the thesis is organized into the following. Chapter two contains the literature review part. Chapter three presents the methodologies adopted for this study together with brief description of the study area. Moreover, this section gives highlights about the physical and demographic features of study area, sampling procedure and sample size drawn for the study, methods of data collection and definition of variables and hypothesized effects of each determinant of efficiency. In the fourth chapter, both the descriptive and econometric results are presented and discussed in detail. The fifth chapter presents the summary, conclusions and recommendations of the study.
7
2. LITERATURE REVIEW 2.1. Concept and Definition of Efficiency Farrell (1957) defined efficiency as the ability of farm’s production to attain optimum level of output from a given bundle of input. Many scholars used productivity and efficiency interchangeably and consider both as the measure of performance of a given firm. However, these two interrelated terms are not precisely the same things (Coelli, 1995). In simple term, production frontier defines the current state of technology in an industry, firms in that industry would presently be operating either on that frontier, if they are perfectly efficient or beneath the frontier if they are not fully efficient. On the other hand, productivity improvements can be achieved in two ways. Once can either improve the state of the technology by inventing new ploughs, pesticides, rotation plans, and the like. This is commonly referred to as technological change and can be represented by an upward shift in the production frontier. Alternatively one can implement procedures, such as improved farmer education, to ensure farmers use of the existing technology more efficiently. This would be represented by the firms operating more closely to the existing frontier. It is thus evident that productivity growth may be achieved through either technological progress or efficiency improvement, and that the policies required to address these two issues are likely to be quite different. Production technology is commonly modeled by means of production function, which in the scalar output case specifies the maximum output obtainable from an input vector. The degree to which the actual output of a production unit approaches its maximum is the technical efficiency of production. Productivity is the quantity of a given output of a firm per unit of input. Technical efficiency (that part of efficiency which explains the physical performance of a firm) measures the relative ability of a firm to get the maximum possible output at given input or set of inputs. Technically efficient firms are those firms that are operating on the production frontier that represents the maximum output attainable from each input level (Coelli, 1995). The concept of efficiency is considered with the relative performance of processes used in transforming given inputs into output. Farrell (1957) identified at least two types of efficiency. These are technical and allocative efficiencies.
8 Technical and allocative efficiency (price efficiency) in production, which together comprises the economic efficiency are through the use of frontier production function. While technical efficiency relates the physical input with the optimum level of output that can be produced at a given level of technology, Allocative efficiency reflects the ability of a firm to use the inputs in optimal proportions, given their respective prices and the production technology. Economic efficiency is the multiplicative product of technical and allocative efficiencies. The simple and straight forward way of measuring efficiency of a farm could be yield per hectare. However, given output is a function of multiple inputs in the reality, this is very simplistic way of measurement in that it only considers a single input of production, land. The other technique is to use the conventional econometric analysis, which generally assumes that all producers always manage to optimize their production process. However, there are discrepancies between production amount and production values even if the enterprises have identical technological constraints. This depends upon different productive capabilities and less favorable utilization resources by some enterprises (Burhan et al., 2009). The traditional, least squares-based, regression techniques attribute all departures from the optimum exclusively to random statistical noise. However, producers do not always succeed in optimizing their production. Therefore, it is desirable to recast the analysis of production away from the traditional functions towards frontiers (Kumbhakar and Lovell, 2000). Thus production frontier characterizes the minimum input bundles required to produce a given level of output or the maximum possible level of production of output from a given level of inputs, commonly called technical efficiency. Even though there is some similarity between terms production efficiency and technical efficiency, however, they are not same. The simplest way to differentiate production and technical efficiency is to think of productive efficiency in terms of cost minimization by adjusting the mix of inputs, whereas TE is output maximization from a given mix of inputs (Palmer and Torgerson, 1999). According to Coelli (1995) in analyzing efficiency, fitting a frontier model performs better than Ordinary Least Square (OLS) regression. The two main benefits of estimating the frontier function, rather than average (OLS) functions are; first, estimation of an average function will provide a picture on the shape of technology of an average firm, while the
9 estimation of the frontier function will be most heavily influenced by the best performing firm and hence reflect the technology they are using. Secondly, the frontier function represents a best practice technology against which the efficiency of firms within the industry can be measured. It is this second use of frontiers, which leads to widely application of estimating frontier functions.
2.2. Theoretical Framework for Efficiency of Production The farmers output can be increased through increased inputs, increasing productivity of inputs and the combination of the two. Hence, efficiency is a central issue in production economics helping as a guide for allocation of resources (Farrell, 1957). Productivity improvements can be achieved in two ways. One can either improve the state of the technology by inventing new technology which leads to an upward shift in the production frontier or alternatively one can improve efficiency of the farmers to use the existing technology more efficiently. This would be represented by the firms operating more closely to the existing frontier. Therefore, it is evident that increase in productivity achieved through either technological progress or efficiency improvement so that the policies required to address these two issues are likely to be quite different (Coelli, 1995). Basically there are two approaches in measuring efficiency: input oriented and output oriented. The output oriented approach deals with the question “by how much output could be expanded from a given level of inputs?” Alternatively one could ask “by how much can input of quantities be proportionally reduced without changing the output quantity produced?” This is an input oriented measure of efficiency. However, both measures will coincide when the technology exhibits constant returns to scale, but are likely to vary otherwise (Coelli and Battese, 2005). 2.2.1. Input-oriented efficiency measures The concept of input-oriented measures of efficiency of a firm which uses two inputs x1 and x2 to produce a single output y, under the assumption of constant return to scale can be illustrated in Figure1. Two inputs x1 and x2 are represented on horizontal and vertical axes respectively. EE* represents an iso-quant of a fully efficient firm. All points on this isoquant represent technically efficient production. Assume a firm is producing at point A as shown in Figure 1; this firm produces the same level of output as is produced by the fully efficient firm.
10 To define the technical efficiency (TE) of this firm, a line is drawn from the origin to the point A. This line crosses the iso-quant at the point C. In the case of a fully efficient firm, y* amount of output (y) is produced using inputs (x1 and x2) at point C whereas in case of the observed firm, operating at A, additional inputs are used to produce y* amount of output (y). Therefore, observed firm, operating at A, does not use inputs efficiently. The technical efficiency of the observed firm can be defined as the ratio of the distance from the point C to the origin over the distance of the point A from the origin:
TE
OC OA
(1)
Figure 1: Input-oriented measures for technical, allocative and economic efficiencies Source: Reproduced from Coelli et al. (1998). The distance CA represents the technical inefficiency of the observed firm, which is the amount by which all inputs could be proportionally reduced without reduction in output. The value of TE lies between 0 and 1. A firm is technically efficient if it has TE equal to 1. If the value of TE is less than 1, the firm is technically inefficient. If input prices are given, allocative efficiency (AE) can also be calculated. A line DD* is drawn tangent to the isoquant EE* at the point C*. The line DD* represents an iso-cost line showing all possible quantities of the two inputs, given their relative market prices that would cost the same amount to the firm.
11 Slope of the iso-cost line represents the input price ratio. For output quantity produced at point C, the best use of inputs is at point C*, because it represents the minimum cost. The allocative efficiency of the firm is defined as:
AE
OB OC
(2)
At point C* a farm is both technically and allocatively efficient. Distance BC represents the reduction in production cost that would occur if production were to occur at allocatively and technically efficient point C*, instead of at technically efficient but allocatively inefficient point C. Value of allocative efficiency lies between 0 and 1. A value of 1 indicates that the firm is allocatively fully efficient while value less than 1 indicates that the firm is allocatively inefficient. The economic efficiency (EE) is defined as the product of technical and allocative efficiency. EE TE AE
(3)
EE
OC OB OA OC
(4)
EE
OB OA
(5)
Value of economic efficiency is bounded between 0 and 1. Value of 1 indicates that the firm is economically fully efficient while value less than 1 indicates that the firm is economically inefficient. 2.2.2. Output-oriented efficiency measures The output oriented measures of efficiency focuses on the changes in output of a firm that may be achieved when using the same quantity of inputs. The concept of output-oriented Measures of efficiency of a firm producing two outputs (y1 and y2) with one input can be illustrated using Figure 2. Two outputs y1 and y2 are represented on horizontal and vertical axes respectively. AA* is a production possibility curve showing different combinations of two outputs (y1 and y2) produced using a given level of input (x1). AA* production possibility curve represents a technically efficient practice. Any firm that is producing at this curve is said to be a technical efficient firm. A firm that is producing at point B is technically inefficient firm because it lies below the production possibility curve AA* that
12 represents the upper bound of production possibilities. To define the technical efficiency of the observed firm producing at point B, a line is drawn from the origin to the point B. This line crosses the production possibility curve at point C. The observed firm uses the same input level as is used by the fully efficient firm, operating at point C. The technical efficiency of the observed firm is defined by the ratio of the distance of the point B to the origin over the distance from the point C to the origin. TE = OB/OC The distance BC represents the level of technical inefficiency. It is the amount by which outputs could be increased without requiring extra inputs.
Figure 2: Output-oriented measures for technical, allocative and economic efficiencies Source: Reproduced from Coelli et al. (1998) If there is price information it is possible to calculate allocative efficiency. Line EE* represents an iso-revenue curve which is drawn tangent to the production possibility curve at F*. The line OB meets it at point D. The allocative efficiency of the observed firm is defined by the ratio of the distance of point C to the origin over the distance of point D to the origin.
AE
OC OD
(6)
The economic efficiency of the observed firm is defined as:
EE
OB OC OC OD
(7)
EE
OB OD
(8)
13
2.3. Models of Efficiency Measurement 2.3.1. Non-Parametric Frontier Model The non-parametric approach has been traditionally assimilated into Data Envelopment Analysis (DEA); a mathematical programming model applied to observed data that provides a way for the construction of production frontiers as well as for the calculation of efficiency scores relatives to those constructed frontiers. Data Envelopment Analysis (DEA) is a non-parametric method and can easily handle multiple input and output. Moreover, in DEA, application inputs and output can have very different units of measurement without requiring any a priori trade off or any input and output prices. An input oriented BCC/ Banker- charnes-cooper model/ suggested an extension of the CRS DEA model and the model is given below for N decision making unit (DMU), each producing M outputs by using K different inputs (Coelli et al., 1998). Min ᵩ λ Ф
(9)
Subject to -yi +Y λ 0 Фxi - X λ 0 NIλ=1, λ 0 Where Ф is a scalar, NI is convexity constraint and λ is N x 1 vector of constants. Y represents output matrix and X represents the input matrix. The value of Ф will be the efficiency score for the ith firm. This linear programming problem must be solved N times, once for each firm in the sample. A Ф value of 1 indicates that the firm is technically efficient according to the Farrell (1957) definition. Data Envelopment Analysis does not impose any assumptions about functional form; hence it is less prone to misspecification. Further, DEA does not take it in to account random error. It is not subject to the problems of assuming on underlying distribution about the error term. However, since DEA cannot take account of such statistical noise, the efficiency estimates may be biased if the production process is largely characterized by stochastic elements but this technique is not the matter of this study. Thus, it is only for the literature review purposes.
14 2.3.2. Parametric Frontier Models With respect to parametric approaches, these can be subdivided into deterministic and stochastic models. The first are also termed ‘full frontier’ models. They envelope all the observations, identifying the distance between the observed production and the maximum production, defined by the frontier and the available technology, as technical inefficiency. The deterministic model assumes that any deviation from the frontier is due to inefficiency, while the stochastic approach allows for statistical noise. A further classification of frontier models can be made according to the tools used to solve them, namely the distinction between mathematical programming and econometric approaches. The deterministic frontier functions can be solved either by using mathematical programming or by means of econometric techniques. The stochastic specifications are estimated by means of econometric techniques only. Coelli et al. (1998) recommended that stochastic frontier analysis is more appropriate than Data Envelopment Analysis and deterministic models in agricultural applications, especially in developing countries, where the data are heavily influenced by measurement errors, and the effect of weather, disease, and the like plays a significant role. 2.3.2.1. Deterministic models The parametric deterministic models used for measuring technical efficiency. We assume that production can be modeled as; yi = + !xi - ui
(10)
Where ui 0 represents inefficiency and all variables are specified in logarithms. In this case, DFi = exp (-u i)
(11)
It is the Debreu-Farrell measure of technical efficiency. It is not necessary to restrict the production function to Cobb-Douglas. This functional form is chosen to be consistent with Aigner and Chu (1968) for convenience. Alternatively, the flexible Translog production function, which is linear in the parameters, can be specified. This technique is considered deterministic because the stochastic component is completely generated by inefficiency and measurement error is assumed away. Following Greene (1980) the deterministic model can be estimated using OLS. In this case, the slope parameters are estimated consistently, but the intercept is biased.
15 Greene shows that a consistent estimate can be obtained by shifting the OLS line upward so that the largest adjusted residual is zero. If the true error term is composed of a normally distributed noise term and a non-negatively distributed inefficiency term, then OLS is not maximum likelihood but still produces unbiased and consistent estimates of the slope parameters. Hence, there will be minor differences between the estimated slope parameters from the stochastic frontier and OLS regressions. Correcting the intercept from an OLS regression is only one deterministic approach. Aligner and Chu (1968) developed linear and quadratic programming alternatives. The deterministic specification, therefore, assumes that all deviations from the efficient frontier are under the control of some circumstances out of the agent’s control that can also determine the suboptimal performance of units. Regulatory-competitive environments, weather, luck, socio-economic and demographic factors, uncertainty, etc., should not properly be considered as technical efficiency. The deterministic approach does so, however. Moreover, any specification problem is also considered as inefficiency from the point of view of deterministic techniques. On the contrary, stochastic frontier procedures model both specification failures and uncontrollable factors independently of the technical inefficiency component by introducing a double-sided random error into the specification of the frontier model. 2.3.2.2. Stochastic frontier model The stochastic frontier approach which was introduced by Meeusen and van den Broeck (1977) and Aigner et al. (1977), reversed the conventional belief that deviations from the production frontier are due to inefficiency of the producing units (that is, factors under the control of the producers, which may not be true). Hence, stochastic estimations of technical efficiency incorporate a measure of random error, which is one component of the composed error term of a stochastic production frontier. This model acknowledges the fact that factors, which are outside the farmers` control, can also affect the level of output. So it made possible to find out whether the deviations in production from the frontier output is due to firm specific factors or due to external random factors. The primary advantage of the stochastic frontier production function is that it enables one to estimate farm specific technical efficiencies. The measure of technical efficiency is equivalent to the production of the ith farm to the corresponding production value if the farm effect ui were zero.
16 However, the estimation of efficiency using stochastic method requires a prior specification of functional form and needs distributional assumptions (half-normal, gamma, truncated, etc.) for the estimation of Ui, which cannot be justified given the present state of knowledge (Coelli et al., 1998). The stochastic frontier production model incorporates a composed error structure with a two-sided symmetric term and a one-sided component. The one-sided component reflects inefficiency, while the two-sided error captures the random effects outside the control of the production unit including measurement errors and other statistical noise typical of empirical relationships. Hence, stochastic frontier models address the noise problem that characterized early deterministic frontiers. Stochastic frontiers also make it possible to estimate standard errors and to test hypotheses, which were problematic with deterministic frontiers because of their violation of certain maximum likelihood (ML) regularity conditions (Schmidt, 1976). In stochastic frontier method, technical efficiency is measured by estimating a production function. Different production functions such as Cobb-Douglas, Translog, Transcendental, and Quadratic etc. can be used to estimate the frontier. The Translog and Cobb-Douglas specifications are commonly used functional forms to estimate the frontier; but both have their merits and demerits. Therefore, the method avoids the imposition of unwarranted structures on both the frontier technology and the inefficiency component that might create distortion in the measurement of efficiency (Shafiq and Rehman, 2000). The choice is made on the basis of the variability of agricultural production, which is attributable to climatic hazards, and insect pests; Moreover, all information gathered on production is usually inaccurate since small farmers do not have updated data on their farm operations. In fact, the stochastic frontiers method makes it possible to estimate a frontier function that simultaneously takes into account the random error and the inefficiency component specific to teff producing farmers.
2.4. Approaches to Identifying Efficiency Factors The literature suggests two methodological approaches for analyzing the sources of technical efficiency based on stochastic production functions. Those are two-stage estimation approach and one-stage estimation approach. The first approach is the two-stage estimation procedure in which first the stochastic production function is estimated, from which efficiency scores are derived, then in the
17 second stage the derived efficiency scores are regressed on explanatory variables using ordinary least square methods or Tobit regression. This approach has been criticized on grounds that the firm’s knowledge of its level of technical inefficiency affects its input choices; hence inefficiency may be dependent of the explanatory variables. The inefficiency effects would only be identically distributed if the coefficients of the farm specific factors are simultaneously equal to zero. It is possible to overcome this problem by the use of a single stage maximum likelihood approach (Battese and Coelli, 1995). The second approach advocates a one stage simultaneous estimation approach as in Battese and Coelli (1995), in which the inefficiency effects are expressed as an explicit function of a vector of farm-specific variables. In the Battese and Coelli (1995) approach, the technical inefficiency effects are specified in the stochastic frontier model and assumed to be independently but not identically distributed non- negative random variables. The technical inefficiency effects are expressed as;
Uj = zj
(12)
Where for farm j, z is a vector of observable explanatory variables and δ is a vector of unknown parameters. Thus, the parameters of the frontier production function are simultaneously estimated with those of an inefficiency model, in which the technical inefficiency effects are specified as a function of other variables. The one stage simultaneous approach is also implemented in FRONTIER and in addition to the basic parameters the program also provides coefficients for the technical inefficiency model. Several
factors
including
socio-economic
and
demographic
factors,
plot-level
characteristics, environmental factors and non-physical factors are likely to affect the efficiency of smallholder farmers. This study has adopted to use the single stage estimation approach due to the measurement procedure of the overall technical efficiency.
2.5. Review of Empirical Studies on Technical Efficiency Abroad A number of efficiency analyses have been conducted by different researchers with the aim of identifying the sources of inefficiencies and 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 and Translog type of production function and commonly estimated parameters by using the MLE procedure and also used the SPF and DEA methodology.
18 In an analysis of technical efficiency in Northern Ghana by Luke et al. (2012) using bootstrap DEA, the average TE of crop production was found to be 77.26 percent. They indicated as nearly 23 percent production loss being due to technical inefficiency. The estimated scale efficiency was 94.21 percent. They used a two stage estimation method, which they found hired labor, geographical location of farms, gender and age of head of household significantly and positively affect TE. Ali and Khan (2014) applied stochastic frontier Cobb-Douglas production function to analysis technical efficiency of wheat production in district Peshawar of Khyber Pakhtunkhwa, Pakistan. Their results further showed that one percent increase in value of land under wheat crop, labor, chemical fertilizer and tractor plough would raise the wheat yield by 0.052, 0.566, 0.130 and 0.438 percent, respectively and were found statistically significant. Farmers’ education was found to be major determinant of technical efficiency/inefficiency. The estimated coefficient of farmers’ education was negative and statistically significant, implying that technical inefficiency decreases with the increase in farmers’ education. Their result also indicted that, the use of more labor and tractor plough hours would increase wheat production in the country. Musaba and Bwacha (2014) were applying SPF model that the average technical efficiency of maize production in Masaiti district, Zambia was found to be about 0.796. This indicated that farm level potential to increase maize production among smallholder farmers in the study area by 20.4% through efficient use of a given input and technology. Their results of the inefficiency model indicate that age, cooperative membership which implies access to fertilizer, and farm size were significantly and positively affect technical efficiency. The seed types used, rotation practices, and education level of the farmer were significantly and negatively affect technical efficiency. Ogada et al. (2014) used a two-stage nonparametric approach on household panel data to estimate the efficiency levels of the smallholders and establish the sources of its variation across Kenya’s smallholder food crop farmers. His result indicated that age, gender, education, size of household, credit, social capital, family labor, intensity of manure use, distance to the nearest road and distance to the nearest market are negatively and significantly affect technical inefficiency and plot size under crops, land under other activities, annual rainfall amount and wage rate to farm worker positively affect technical inefficiency in the study area.
19 Mburu et al. (2014) used SPF model to analysis of economic efficiency and farm size: A case study of wheat farmers in Nakuru District, Kenya. The study attempts to estimate the levels of technical, allocative, and economic efficiencies among the sampled 130 large and small scale wheat producers in Nakuru District. Their result indicated that the mean technical, allocative, and economic efficiency indices of small scale wheat farmers are 85%, 96%, and 84%, respectively. The corresponding figures for the large scale farmers are 91%, 94%, and 88%, respectively. Education, distance to extension advice, and the size of the household were significantly positively influence the efficiency levels. The relatively high levels of technical efficiency among the small scale farmers defy the notion that wheat can only be efficiently produced by the large scale farmers. Mwajombe and Mlozi (2015) used SPF to analyzed technical efficiency indices of Urban Agriculture (UA) farmers in Tanzanian urban wards of towns of Arusha, Dares Salaam and Dodoma. Their Study results revealed that a mean technical efficiency index (TEI) of 0.72 was achieved implying that output from urban agriculture production could be increased by 28% using available technologies. Despite of urban farmers having entrepreneurial acumen, they faced several challenges in resource allocation. Land size, total variable costs, and extension service charges negatively impacted on TEI. Ouedraogo`s (2015) paper was focused on technical and economic efficiency of rice producers in the Kou valley, located in the region of the high basins in the western part of Burkina Faso. The stochastic frontier approach was used to estimate the production function, from a Cobb-Douglas stochastic frontier function and its dual which allow the estimation of the technical, allocative and economic efficiencies. His result show that farm size, fertilizer used, labor availability, years of experience and education level were significant and positive effect on technical efficiency of rice production in the Kou valley. The TE, AE and EE of producers are, on average, 80.15%, 92.7% and 74.43% respectively. A 25% improvement of rice production is possible if producers optimize their economic efficiency. Biam et al. (2016) employed the Cobb-Douglas stochastic frontier cost function to measure the level of economic efficiency and its determinants in small scale soybean production in Central Agricultural Zone of Nigeria. Their result of the analysis showed that average economic efficiency was 52%. The study found age, farm size and household size to be negatively and significantly related to economic efficiency.
20 Education, farming experience, access to credit and fertilizer use were significantly and positively related to economic efficiency. No significant relationship was found between economic efficiency and extension contact and membership of farmers’ association.
2.6. Empirical Studies on Technical Efficiency in Ethiopia In Ethiopia, a number of researches are conducted on efficiency of farmers in different regions using different models and different variables in order to measure and identify the level and sources of technical inefficiency. Hassen et al. (2012) analyzed the efficiency of crop-livestock production and assessing their potential for improvement in North-East Ethiopian highlands. They had used crosssectional data to analyze the economic efficiency of mixed crop and livestock production system and identify its determinant factors. The parametric method stochastic frontier approach was employed to measure economic efficiency. Their results indicated that most farmers in the study area being not efficient with the mean technical, allocative and economic efficiencies of 62%, 51% and 29%, respectively. Their results also showed that improved agricultural technology adoption significantly improved production efficiency of households. Beyan et al. (2013) evaluated the technical efficiency of farm production of smallholder farmers in Girawa district. Cobb-Douglas production function was fitted using stochastic production frontier approach to estimate technical efficiency levels and to identify factors affecting efficiency levels of the sample farmers. His result showed that the mean technical efficiency of 81.5%. The discrepancy ratio (γ), which measures the relative deviation of output from the frontier level due to inefficiency, implied that about 75% of the variation in maize production was attributed to technical inefficiency effects. He also found that education, livestock holding, extension contact, farmer’s training, cultivated area and participation to irrigation were found to determine technical efficiencies of farmers positively while social status had negative relationship with technical efficiency. In Mesay et al. (2013), a Translog production function approach was used to investigate the source of technical inefficiency of smallholder wheat farmers in selected water logged areas of Ethiopia. Their result indicated that the mean technical efficiency of wheat farms of 0.55. Age, number of livestock, access to input and output market has a positive effect on efficient wheat production thereby integration of improved wheat production with the
21 input and output market plays a significant role in enhancing the technical efficiency of wheat producer farmers. Thus provision of input (improved seeds, fertilizer, pesticides herbicides and fungicides) and output market facilities raises farmers’ wheat production efficiency level. Wondimu and Hassen (2014) used SPF model to analyze technical efficiency in maize production of smallholder farmers in Dhidhessa district. From their result, the estimated gamma parameters indicated that 71% of the total variation in maize output was due to technical inefficiency. The average technical efficiency was 86% while return to scale (RTS) was 0.96 %. Based on the results, the existence of scope for increasing maize output by 14% through efficient use of existing resources was concluded from the study. Their result also indicated that area allocated under maize and chemical fertilizers appeared to be significantly influencing maize production and the marginal effect of inefficiency variables such as age, improved seed, labor availability and training were affect positively and significant. On the other hand number of livestock, market distance, and interaction of education and off farm income were affect negative and significant. Tefera et al. (2014) used the Cobb Douglas stochastic production frontier to analyze the technical efficiency in teff production in the Raya Alamata district. From his result Fertilizer application rate has contributed positively and significantly to teff production, indicating that there is a possibility to increase teff production by increasing fertilizer application rate. Education of the household has significant positive contribution to teff production indicating that there is scope for increasing teff production by improvement the education level of the farmers. The inefficiency in teff production was due to sowing of poor quality seed year after year and large operational farm size. Solomon (2014) used the SPF model together with the inefficiency parameters to identify factors affecting level of technical efficiency of teff and the study shown that age had a positive and significant effect on TE of teff production. The inefficiency effect analysis for major crop production shown that education, participation in soil and water conservation activities, poverty status and adoption of improved seed are the major determinants. Offfarm income of the household head was found to affect technical inefficiency in teff production positively, contrary to this age of household head, slop and numbers of livestock were found to affect negatively.
22 Musa et al. (2014) used a Cobb-Douglas stochastic frontier production analysis approach with the inefficiency effect model to analyze the technical efficiency in maize production of smallholder farmers in central rift valley of Ethiopia. Their result shows that the mean technical efficiency of the farmers in the production of maize as 88%. The estimated stochastic production frontier (SPF) model indicates that DAP fertilizer, Area, Labor, seed and oxen as significant determinants of maize production level. The estimated SPF model together with the inefficiency parameters shows that frequency of family size, extension contact, access to credit and number of weeding positively and significantly determining the technical efficiency level and farm size and distance to market negatively and significantly determined technical efficiency level of the farmers in maize production in the study area. Awol (2014) used SPF to Analysis Economic efficiency of rain-fed wheat producing farmers in north eastern Ethiopia: the case of Albuko district. His result indicated that the mean indices of TE, AE and EE varying widely, with an average of 60%, 42.7% and 31.65% respectively. The study found that sex of the household heads, land fragmentation, fertility status of land, slope, credit use, and training obtained and oxen numbers contributed significantly and positively affect TE, while it has inverse related with farm size. The allocative and economic efficiency of the farm household was positively and significantly affected by sex of the household heads, frequency of extension use, oxen number, family size, distance of wheat farm from residence, slope and training. On the contrary, age of the household heads and number of livestock unit have inverse related with allocative and economic efficiency level of the farmers in the area. Getachew and Bamlak (2014) used a stochastic frontier approach to analyzed technical efficiency of small holder maize growing farmers of Horo guduru wollega zone. Their result indicated inefficiency in the production of maize in the study area. The relative deviation from the frontier due to inefficiency was 85%. The average estimated technical efficiency for smallholder maize producers ranges from 0.06 to 0.92 with a mean technical efficiency of 0.66 (66%). The analysis also reveals that the educational level of the farmer, age of household head, land fragmentation, extension services and engagement in offfarm/non-farm activities were significantly and positively affect TE and total land holding of the farmer was significantly and negatively affects farmers’ technical efficiency of maize production.
23 Fantu et al. (2015b) conducted smallholder teff productivity and Efficiency analysis in High-Potential Districts of Ethiopia. They applied data envelopment analysis to measure smallholder teff producers' relative productivity and efficiency. Their result indicated that sex, education level, household size, area, tropical livestock unit and production information positively affected total factor productivity and efficiency and age, number of crops household cultivated, average area of plots household cultivates , average distance between plots and household participation were negatively affect total factor productivity and efficiency Hailemaraim (2015) used a Cobb-Douglas stochastic frontier production analysis approach with the inefficiency effect model to simultaneously estimate technical efficiency and identify the determinants of efficiency variations among teff producer farmers in Berehe District. From his result maximum likelihood parameter estimates showed that teff output was positively and significantly influenced by area, fertilizer, labor and number of oxen. The estimated mean level of technical efficiency of teff producers was about 72 percent. His result also indicated that Fertility status of the farm, off-farm occupation, education, credit service, and extension contact determining technical efficiency significantly and positively. However, age of the household head, family size, number of farm plot, and total farm size were found to reduce farmers’ technical efficiency. Wudineh and Endrias (2016) employed the stochastic frontier and translog functional form with a one-step approach to assess efficiency and factors affecting efficiency in wheat production. From their result the mean technical efficiency was found to be 57%. Factors such as sex, age and education level of the household head, livestock holding, group membership, farm size, fragmentation, tenure status and investment in inorganic fertilizers affected efficiency positively and distance to all weather roads negatively affected. The finding implies presence of an opportunity to improve technical efficiency among the farmers by 43% through gender-sensitive agricultural intervention, group approach extension, and attention to farmers’ education, scaling out of best farm practices. Hassen (2016) employed SPF to measure the level of technical efficiency and identify its determinants in wheat crop for smallholder farmers in south Wollo Zone, Ethiopia. His result showed that the average technical efficiency of wheat production in the study area was 79 percent indicating a good potential for increasing wheat output by 21 percent with the existing technology and levels of inputs. His econometric results of stochastic
24 production function indicated area, seed, fertilizer, man days and oxen days positively affecting the technical efficiency while off farm income was negatively affect technical efficiency. In summary, different studies used different models to analyze the efficiency of farmers and the influence of different agro-climatic and socio-economic conditions on farmers’ efficiency. Therefore, undertaking studies on farm households’ efficiencies in different localities help the policy makers and other development workers to design and implement an appropriate policy intervention. It was also indicated that a number of factors can affect the efficiency level of farmers, but these factors are not equally important and similar in all places at all time. A decisive factor in one place at certain time may not necessarily be a significant factor in other places or even in the same places after some time. Therefore, policy implications drawn from some of the above empirical works may not allow in designing area specific policies to be compatible with its socio-economic as well as agroecologic conditions. In case of Jamma district such type of research work has not been conducted and there is a need to know the level of technical efficiency of small scale farmers particularly with respect to teff production since teff is one of the important crops to the study areas as well as the nation. In a nutshell, what can be suggested from the literature is that at the current level of technology and factor endowment, there is a potential to increase agricultural production by improving the demographic, institutional, and environmental factors. Therefore, this study intends to fill this information and knowledge gaps.
2.7. Conceptual Framework of Technical Efficiency of Teff Production Conceptual framework is defined as a network or a plane of interlinked concepts that together provide a comprehensive understanding of a phenomenon. In other words, it is a visual or written product that explains either graphically or in a narrative form, the main things to be studied (key factors, concepts, variables and the presumed relationships among them) (Miles and Huberman,1994). The conceptual framework for this study is based on the institutional analysis and development approach of the new institutional economics. In the institutional analysis and development approach by Ostrom et al. (1994) it is assumed that an exogenous set of variables that influences situations of actors and the behavior of the actors in those situations leading to outcomes, which then feedback to modify both the exogenous variables and the actors and their situations.
25 The conceptual framework is shown in Figure 3 below, which represents how various factors inter-relate to influence teff productivity and hence the welfare of teff producers. The policy environment is characterized by agricultural policies, governance and existing political and economic trends in the country which have an influence on the farming system and indirectly determine teff productivity. However, within the farming system various sets of factors interrelate to determine teff productivity.
Production inputs such as amount of seed, fertilizers, area, oxen power and labor are used as input into teff production. The availability and distribution of these inputs may be influenced by policy framework in place which in turn determines teff productivity. It is expected that more inputs used by the farmers up to recommended level leads to higher teff productivity. In addition, teff productivity is also affected by technical efficiency because for a production to be effective, the way in which available inputs are utilized is crucial. However, technical efficiency of farmers is also influenced by farmer’s characteristics, cultivated land characteristics, crop specific factors, institutional and socio-economic characteristics of farmers. A farmer that is technically efficient is therefore expected to realize higher teff productivity compared to that of less efficient in teff production. Therefore, this has a positive spillover effect on the welfare of teff producer farmers. Improved welfare of farmers then provides a feedback effect in form of increased access to production inputs and relevant lessons to policy makers.
26
Policy factors Agricultural policies, Governance, Political and economic conditions
Production Inputs Cultivated area, Amount of seed, Fertilizer, Labor force Oxen power
Socioeconomic Factors Sex, Age of the household, family size, Education, Off/nonfarm income, Livestock holding,
Institutiona l factors
Farm Characteristics
Extension contact, Training, Credit
Farm size, Distance of plot from residence, Land fragmentation Improved seed
Teff productivity
Technical efficiency
Outcomes Increased output and increased household welfare, Sustainability of farming and competitiveness; and Poverty reduction and Food security, etc.
Figure 3. Conceptual framework of the study
27
3. RESEARCH METHODOLGY 3.1. Description of the Study Area The study was carried out in Jamma District. It is located in the North Eastern part of Amhara National Regional State, South Wollo Zone, Ethiopia, lying between 10° 23’ 0” and 10° 27’ 0” N latitude and between 39° 07’ 0” and 39° 24’ 0” E longitude. The district has an altitude that ranges from 1,600 to 2,776 m above sea level. The district is bordered on the southeast by Qechene River which separates it from North Shewa Zone, on the west by Kelala, on the north by Legahida, on the northeast by Wore Ilu, on the south by Mida, on the east by Gera Mider and Keya Gebrieal. The district capital town, Degolo is about 260 km away from Addis Ababa and 110 km away from the zonal city of South Wollo Zone, Dessie (CSA, 2012; WOA, 2012). According to CSA (2015), Jamma district has total human population of 146,168 of which 72,380 (49.9%) are male and 73,788 (50.1%) are female. Out of the total population, 4.8% are urban dwellers. Jamma district has a population density of 139.0 people per square kilometer, which is less than the zonal average of 174.7 per km2. The district has a total area of 1,051.93 km2. In the rural areas, farmers are organized in to 22 Kebele Administrations (KAs). The 22 KAs are organized under 22 Development Centers (DCs) for management purposes (CSA, 2015). According to WOA (2012), the agro climatic feature of the district described as 42.7%, 48.2% and 9.1% are Dega, Weyina dega and Kola respectively. Jamma district is characterized by relatively moderate rainfall with mean annual rainfall of 1,130 mm. The temperature varies from a minimum of 13.25°C to a maximum of 28.35°C annually. Compared to other districts of the zone, Jamma district has relatively moderate climate and it has mean annual temperature of 21°C. Agriculture is the main stay of the economy of Jamma district in which about 85% of the population is engaged. The farming system can generally be characterized as mixed, and includes the production of arable crops and the raising of livestock. The level of production for both sectors remains far below its potential, mainly because of adverse climatic conditions due to erratic rainfalls and long standing drought periods. Other reasons include, the relatively small land holdings; which range from 0.25 to 0.75 hectares, and insufficient application of basic agricultural inputs such as fertilizers and pest control techniques (WOA, 2015).
28 According to the information obtained from the district agricultural office (2016), Jamma's smallholder farmers own a substantial number of livestock. Cattle are the most important, in terms of contribution to the regional economy through the supply of milk, hides and skin and dung (cow droppings) used as household fuel. At times of drought, the cattle are sold to raise cash for the purchase of cereals and other essentials. Small stocks consisting of goats and sheep are reared primarily for their meat, and secondly as an investment as well as a source of cash in times of need. The total land area under cultivation in Jamma District is 37,280 hectares. Main types of crops grown during rainy season are teff, wheat and barley in their order of area coverage.
Figure 4. Geographical location of the study area
3.2. Data Types, Sources and Methods of Data Collection Both primary and secondary data as well as quantitative and qualitative data were employed for this study. The primary data were collected using structured questionnaire and focus group discussion. For this study structured questionnaire was designed and pretested. The feed-back from the pre-test was used to refine and modify the questionnaire. The process of primary data collection was done by enumerator, the Kebele development agents and the researcher.
29 The enumerators were trained on data collection procedures. In the study cross-sectional household data of 2015/2016 main harvest cropping season was used. Data for input (such as land, human labor, oxen labor, fertilizer, and seed amount) were used and output of teff production was collected from the specified period of time. Data on input use and outputs were collected in local units and converting into standard units. In addition, primary data were collected by interviewing the selected teff producing farmers and variables that cause variation in production efficiency like age, education, household size, extension contact, gender, and the like. In addition, socio-economic variables such as demographic data, credit access, livestock holding, wealth indicators and institutional data were collected. On the other hand, data related to teff production trend, input supply and extension services were collected to clarify and support analysis and interpretation of primary data. There was close supervision by the researcher during data collection so that errors, if any, could be corrected at earliest possible time. Besides to primary data, this study used secondary data from governmental and non-governmental institutions, published and unpublished documents, website and other relevant sources for analysis and descriptive purposes.
3.3. Sampling Technique and Sample Size In order to select sample households, three-stage sampling technique where combinations of purposive and simple random sampling techniques were used to select the district and sample household heads. Out of the 23 districts in South Wello Zone, Jamma district was purposively selected due to long year experience in teff production and extent of teff production in South Wollo Zone. This information is obtained from South Wello Zone Agricultural Office. In the first stage, out of the three agro-ecologies of the district weyina dega was selected purposively due to the major teff production part of the district. In the second stage, out of the total weyina dega kebeles (11), three kebeles were selected by simple random sampling. In the third stage, 149 sample teff producing farmers were selected using simple random sampling technique from each selected kebeles based on probability proportion to size sampling technique. The sample size for the study was determined based on Yamane (1967) since the population is homogenous in agro-ecology and production system. The simplified formula provided by Yamane is used to determine the required sample size at 91.89% confidence level and 8.11% level of precision. The simplified formula used to determine the sample size of the study was specified as follows.
30 n=
(
(13)
)
Where: n = sample size; N = total number of teff producing farmers in weyina dega kebeles (10,501); e = level of significance (8.11%). Based on the formula the total sample size of the study was 149 farmers.
3.4. Method of Data Analysis Descriptive statistics and inferential statistics along with econometric models were used to analyze the data. Descriptive statistics such as mean, standard deviation, frequency and percentage were employed to analyze the data collected on socio-economic, institutional and agro ecological characteristics of the sample households while inferential statistics such as t-test and chi-square (χ2) tests were used to undertake statistical tests. The econometric analyses were following the following processes. In the first step, the data were
checked
for
regression
model
assumption
including
multicollinearity,
heteroscedasticity and model specification test. Finally, the data were analyzed using stochastic frontier approach by using single stage estimation.
3.5. Specification of the Empirical Model Stochastic production frontier is the most appropriate technique for efficiency studies which have a probability of being affected by factors beyond control of decision making unit. This is because of the fact that this technique accounts for measuring inefficiency as a result of these factors and technical errors occurring during measurement and observation. Teff production at the study area is likely to be affected by natural hazards, unexpected weather conditions, pest and disease occurrence which are beyond the control of the farmers. In addition, measurement and observational errors could also occur during data collection. So as to capture effects of these errors, this study used stochastic frontier model. Stochastic frontier analysis was simultaneously introduced by Aigner et al. (1977) and Meeusen and Van der Broeck (1977). The stochastic frontier approach splits the deviation (error term) into two parts to accommodate factors which are purely random and are out of the control of the firm. One component is the technical inefficiency of a firm and the other component is random shocks (white noise) such as bad weather, measurement error, and omission of variables and so on.
31 The model is expressed as: ln Yi 0 i ln X
ij
exp
ei
(14)
Where: ln = denotes the natural logarithm; i = represent the ith farmer in the sample, Yi = represents yield of teff output of the ith farmer (Qt), Xij = refers to the farm inputs of the ith farmer ei = vi-ui which is the residual random term composed of two elements vi and ui. The vi is a symmetric component and permits a random variation in output due to factors such as weather, omitted variables and other exogenous shocks. Since OLS yields inconsistent estimate of
0
and it is impossible to decompose the
technical inefficiency from the white noise with OLS, a maximum likelihood estimation technique was employed. In addition, in obtaining a convenient method of moment’s estimator, Coelli (1996) stated that the two-stage MLE method leads to estimates that are less efficient than single stage estimates. Therefore, in this study a single-stage maximum likelihood estimation procedure was employed in obtaining the parameter coefficients of
i and the parameter coefficients of the in/efficiency effects. These parameters can be expressed in terms of the variance parameters of the stochastic frontier, sigma squared (σ2s) and the inefficiency effects gamma (γ), that is, the variance ratio given by, γ = σ2u /σ2s, where σ2s = σ2v + σ2u. This variance parameter (γ and σ2s), coefficients are the diagnostic statistics that indicate the relevance of the use of the stochastic frontier function and the correctness of the assumptions made on the distribution form of the composed error term (that is, for both Ui and Vi) respectively. The γ parameter measures technical inefficiency effect in teff production for the variation of observed output from the optimal one, and it has a value between zero and one as stated in Battese and Corra (1977). Since estimation of i and Ui by using MLE requires a prior imposition of distributional assumptions about Vi and Ui. The MLE techniques generally assume a normal distribution with mean zero and constant variance for Vi. On the other hand, there are different assumptions about distribution of Ui. This one sided term can follow such distribution as
32 half normal as indicated by (Aigner et al., 1977, and Jondrow et al., 1982), truncated normal (Stevenson, 1980) and exponential distribution (Aigner et al., 1977; Meeusen and van den Broeck, 1977) amongst the most frequently assumed distributions for Ui.
3.6. Selection of the Functional Form Another issue surrounding parametric frontiers relates to the choice of functional form. Among the possible algebraic forms, Cobb-Douglas and the translog functions have been the most widely used functional forms in most empirical production analysis studies. Each functional form has its own advantage and limitations. Some researchers argue that CobbDouglas functional form has advantages over the other functional forms in that it provides a comparison between adequate fit of the data and computational feasibility. It is also convenient in interpreting elasticity of production and it is very parsimonious with respect to degrees of freedom. So, it is widely used in the frontier production function studies as stated in Hazarika and Subramanian, (1999). In addition, due to its simplicity features, the Cobb-Douglas functional form has been commonly used in most empirical estimation of frontier models. This simplicity however, is associated with some restrictive features in that it assumes constant elasticity, constant return to scale for all firms/farms and elasticity of substitution are equal to one (Coelli et al. 1998). Moreover, several studies specifically farm efficiency studies, from both developing and developed countries, have used the Cobb-Douglas functional form despite its limitations (Battese and coelli, 1992). On the other hand, the translog functional form imposes no restrictions upon returns to scale or substitution possibilities. However, the problem of degrees of freedom and multicollinearity is a serious problem in translog production function (Coelli et al., 1998). As a result, taking the advantages and disadvantages of both functional forms into consideration, in this study, the Cobb-Douglas production frontier functional form was fitted to estimate the level of technical in/efficiency of farmer’s teff production in the study area. To summarize, the choice of this functional form is due to the fact that, first, the CobbDouglas has been the most commonly used production functional form in the specification and estimation of production frontiers in empirical studies in most efficiency especially farm efficiency studies in both developed and developing countries. Second, because of the logarithmic nature of the production function that makes econometric estimation of the parameters a very simple matter (in terms of analysis and interpretation).
33 In addition, the Cobb-Douglas functional form is also convenient in interpreting elasticity of production and it is very parsimonious with respect to degrees of freedom. Therefore, that is why Cobb-Douglas functional form was used in this study. The technical efficiency of teff production in Jamma district was measured by considering the output obtained per household head as the dependent variable. The output of teff from the 2015/16 production year was measured in quintals. The independent variables were the inputs (factors) of production used in the same production year. Accordingly the relevant inputs which were considered were as follow: Y = the total amount of teff produced in quintal by the ith farmer; X1= the total number of oxen-power used for teff production in oxen-days by the ith farmer; X2=the total labor (family and hired) in man-days used for teff production by the ith farmer; X3=the total quantity of teff seed in kilogram used for teff production by the ith farmer; X4 = the total amount of fertilizer in kilogram applied for teff production by the ith farmer; X5= the total area covered by teff in hectares of the ith farmer; The Cobb-Douglas form of stochastic frontier production is stated as follows; 5
lnY β 0 βjlnXij Vi Ui
(15)
j 1
Where: For ith farmer, Y is the total quantity of teff produced, x is the quantity of input j used in the production process including Oxen labor, Human labor, land, quantity of seed and quantity of fertilizer; Vj is the two-sided error term and Uj is the one-sided error term (technical inefficiency effects). The inefficiency model was estimated as the equation given below. 13
lnY = 0 +
nZni
(16)
n 1
Zi is the variable in the inefficiency model The technical inefficiency ( ) could be estimated by subtracting TE from unity. The function determining the technical inefficiency effect is defined in its general form as a linear function of socio-economic and management factors.
34 It can be defined in the following equation: 13
Ui = 0 + KZjk
(17)
k 1
Where, is the technical inefficiency effect, δk is the coefficient of explanatory variables, The Zi variables represent the socio-economic characteristics of the farm explaining inefficiency of the ith farmer. As a result the technical inefficiency was explained by the following determinants: Zi1 = Age of the household head (years); 2=
Sex of the household (a dummy variable. It takes a value of 1if male, 0 otherwise;
Zi3 = Household size (total numbers of family member who lives in one roof); Zi4 = Education (number of years of schooling of the farmer); Zi5 = Farm size measured by hectare; Zi6 = Land fragmentation (it include the number of locations of different plots); Zi7 = Distance to teff plot from residence measured in km; Zi8 = Number of livestock measured by TLU; Zi9 = Training (A dummy variable. It takes a value of 1 if yes, 0 otherwise); Zi10 = Extension contact (frequency of extension service during the farming season); Zi11 = off/nonfarm income (total amount of off/nonfarm income in birr); Zi12 = credit (total amount of credit received during the production season); Zi13 = Improved seed (A dummy variable. It takes a value of 1 if yes, 0 otherwise). The validity of the models used for the analysis and hypothesis test was investigated using the general Likelihood ratio test. Generalized Likelihood ratio computation was defined as; LR = −2 [lnLH0 – lnLH1]
(18)
Where, LR= Log likelihood ratio LH0 =Value of log likelihood of null hypothesis LH1= Value of log likelihood of alternate hypothesis m*=degree of freedom= number of restrictions= number of estimated inputs and inefficiency variables in the current model (alternate hypothesis) minus number of
35 estimated inputs and inefficiency variables in the preceding model (null hypothesis).The null hypothesis will be rejected when LR (calculated 2 m*) > tabulated 2 m*. If the null hypothesis will be true, the test statistic will approximately a 2 distribution or mixed 2 distributions with degrees of freedom equal to the difference between the number of parameters specified in the null hypothesis and alternative hypothesis. Moreover the LogLikelihood ratio will be used to test the null hypothesis that the inefficiency component of total error term is equal to zero ( = 0) against the alternate hypothesis that the inefficiency component is greater than zero ( > 0). Thus, the log likelihood ratio was calculated and compared with the critical value of 2 with one degree of freedom at 5% level of significance. Moreover from the stochastic model in equation (14), the actual output was given by: Yi : exp (Xi β+ Vi-Ui). From this equation, technical efficiency (exp-ui) is given as; (19)
TEi = Yi/Yi * Where TEi = technical efficiency of the ith household in teff production Yi* = the frontier output of the ith household in teff production, Yi = the actual output of the ith household in teff production. Then Yi* = Yi/TEi
(20)
3.7. Definition of Input and Inefficiency Variables and their Hypotheses Output (Y): This is the dependent variable of the production function. It is the total amount of teff produced expressed in terms of physical output of teff in quintal (Qt) of sample households. 3.7.1. Input Variables Oxen (X1): Primarily oxen are used as a source of drought power in teff production process in the study area. It refers to the number of drought power used for different activities in teff production and it was measured in oxen-days (ODs). Here, one oxen-day is equivalent to eight hours in order to aggregate the ODs used for different oxen-driven activities that include plough, and threshing.
36 Labor (X2): This represents the total labor (family, exchange and hired) utilized in various farm activities (plough, sowing and fertilizer application, weeding, harvesting and threshing). The record was done by the type of person participated on the given activity by categorizing as children, man and woman. Thus, labor inputs for major activities were converted into man-equivalent. The man-equivalent was computed by taking into account the age and sex of the laborer and using standard conversion factor reported by Strock et al. (1991) as indicated in Appendix Table 1. Then the total man-equivalent used by each sample farmer was included in the production function by converting into man-days. It was measured in man-days (MDs) where eight hours is equivalent to one man-day. Seed (X3): It refers to the quantity of teff seed used in kg by the household head. Farmers mainly use local and improved teff seed in production. Fertilizer (X4): Farmers commonly apply Di-Ammonia Phosphate (DAP) and UREA for teff production in the study area. Farmers used it in equal amount. The rate of application of this productivity enhancing input clearly matters to maximize teff yield. Its unit of measurement is kg. Area of land covered by teff (X5): This refers to the physical unit of land area under teff crop cultivation expressed in hectare (ha). It includes all plots (that is, plots of land owned and/or cultivated through different land use arrangements such as renting-in, leasing and/or sharecropping) of land under teff during 2015/16 production year for each household. 3.7.2. Definition of Inefficiency Variables and Hypothesis Based on previous studies and socio-economic conditions of the study area, the following factors were expected to determine efficiency differences among sample farmers. Age (Z1): Age of household head in years is hypothesized to reflect experience and physical strength of the farmer on efficiency. However the farmer become less efficient as he/she gets older and his ability to manage farming activities are expected to decrease. Younger farmers are tending to be more open and likely exposed to methods and techniques (Evaline et al., 2014; Bwala et al., 2015; Biam et al., 2016). Hence, in this study age was hypothesized to have negative effect on technical efficiency.
37 Sex of the household head (Z2): is a dummy variable representing the sex of household head taking a value of 1 for male headed households and 0 for female headed households. Even though women play a substantial role in agricultural activities, there are still tasks that are not used to be done by women. Besides, female headed households may also have to perform additional tasks such as taking care of children and therefore they may have to allocate their time between these tasks and actual farm activities (Fantu et al., 2011; Hazell, 2013). Hence, male headed households were expected to be more efficient than female headed households. Household size (Z3): It is total numbers of family member who lives in one roof (number of people living together and utilizing scarce resources) measured in adult equivalent. Empirical studies by Musa et al. (2014), Mburu (2014) and Kabir et al. (2015) showed that family size was positively and significantly affect the efficiency level of the farmer. Thus, family size was expected to have a positive influence on technical efficiency of farmers in the study area. Education of the household head (Z4): It is continues variable which is the year of schooling of the household head. This is used as a proxy variable for managerial ability of the decision making unit (household head). It is assumed that through education, the quality of labor is improved and he/she become active to adopt new technologies. Access to education together with increased experience could guide to better management of farm activities. The role of education toward improving farmers` efficiency is now widely accepted, in that it enables them to understand the socioeconomic conditions governing their farming activities and learn how to collect, retrieve, analyze and disseminate information. A lot of empirical studies showed that education is one of the most recognized factors in determining the efficiency level of the farmers in the world and resulted that education determined efficiency positively and significantly (Ali and Khan, 2014; Hailemariam, 2015; Ouedraogo, 2015). Hence, in this study education was expected to have positively related with teff production efficiency Farm size (Z5): This is a continuous variable which represents the total crop area in hectares managed by a farmer. It is important to evaluate whether relatively large farmers are more efficient or not than small ones. As the farm size of a farmer increases the managing ability of him will decrease given the level of technology. Empirical studies by Getachew and Bamlak (2014), Mwajombel and Mlozi (2015) and Wudineh and Endrias
38 (2016) showed that farm size is negatively and significantly affect technical efficiency of the farmer in the production of cereal crops. Thus, farm size was expected to have a negative influence on technical efficiency of farmers in the study area. Land fragmentation (Z6): It is the number of location under different plots which farmer managed during the production year. Farm plots in the area are fragmented and scattered over many places that would make it difficult to perform farming activities on time and effectively. Increased land fragmentation leads to inefficiency by creating shortage of family labor, costing time and other resources that should have been available at the same time (Abba, 2012; Mekonnen et al., 2015). Hence, in this study negative association between fragmentation and technical efficiency was expected. Distance of household's residence from teff farms (Z7): It is the distance of teff farms from household's residence measured by kilometer. Those farms near to homestead will save travel time and then get more management; whereas, those farms far away from household's residence will receive less management and the frequency of visits may reduce. As a result, it has negative effect on efficiency (Evaline et al., 2014; Fantu et al., 2015a). Hence, in this study distance of residence from teff farms was expected to have negatively related with teff production efficiency. Livestock size (Z8): It is the total number of livestock owned by farmers measured in tropical livestock unit. The variable used as proxy for wealth, farmers with more livestock units, which can readily be converted to money can be able to buy modern inputs than those that own fewer livestock units. Moreover, families with more animals are more likely to have larger protein intake than those with fewer animals, which helps improve their working efficiency (Mekonnen et al., 2015; Bwala et al, 2015). Therefore, in this study livestock number was expected to have positively relation with technical efficiency of teff production. Training (Z9): Training is an important tool in building the managerial capacity of the household head. Household heads that got training related with crop production and marketing or any related agricultural training were hypothesized to be more efficient than those who did not receive training. It is dummy variable having value equal to 1 if the household got training at least one day in the cropping season and zero otherwise. Many empirical studies showed that training found to be affect efficiency positively and
39 significantly (Beyan, 2013; Awol, 2014; Birhan, 2015). Hence, in this study training was expected to have positive effect on technical efficiency. Number of Extension contact (Z10): This is used as a proxy measure for access to extension services and defined as the number of time the extension agent visited the farmer during stated production season. Extension workers may play a central role in informing, motivating and educating farmers about available technologies. Hence, it may have a positive impact on efficiency levels of teff producers through improvement of their managerial ability and general agronomic practices (Abba, 2012; Krishnan and Patnam, 2014; Hailemariam, 2015; Chandrasekhar and Nikita, 2015). Hence, in this study number of extension contact was expected to have positively related with teff production efficiency Off/non-farm Income (Z11): It is continues variable which represent the total amount of income from off/nonfarm income during the production season. The amount of off/nonfarm income may have a systematic effect on the technical efficiency of farmers. This is because farmer may allocate more of his/her time to off /non-farm activities and thus may lag in agricultural activities. In the other hand, incomes from off/non-farm activities may be used as extra cash to buy agricultural inputs. Empirical studies by Solomon (2014) and Al-hassan (2012) indicated that the amount of off/non-farm income to have positive impact on technical efficiency. However, a study by Bwala et al. (2015) revealed that the amount of off/non-farm income had no effect on efficiency. Hence, in this study positive association between off/nonfarm income and technical efficiency was expected. Credit (Z12): This is continues variable that represents amount of credit received for farm related purposes by the farmer in the production year. Since farmers in developing countries have not sufficient working capital to run agricultural activities unlike developed countries, farmers need to have that potential to engage in such business. Hence, credit is an important source of financing the agricultural activities of smallholder farmers and this is supported by empirical studies conducted by Bifarin et al. (2010), Gebregziabher (2012) and Biam et al. (2016) amount of credit is positively and significantly related to level of technical efficiency of crop production. Hence, in this study credit was expected to have positively related with teff production efficiency. Improved seed (Z13): It is dummy variable which have the value of one if the farmer is used improved seed; and zero otherwise. This argues that, improved seed of teff is enhancing productivity and quality of the output. The empirical studies by Sultan and
40 Ahmed (2014) and Mukete et al. (2016) indicated that improved seed was important variable to maximize efficiency. Hence, in this study improved seed was expected to have positively related with technical efficiency. Table 1. Summary of definition, measurement and hypothesis of inefficiency variables Variable Definition Measur Expected sign ement with technical inefficiency Dependent Efficiency index Technical efficiency in teff production Independent Age Age of household head Year + Sex of the Sex of the household head, having a Dumm household value of 1 for male and 0 otherwise y Family size total numbers of family member who Adult lives in one roof equival ent Education number of years of schooling of the Years farmer Farm size Total cultivated farm size Hectare + Land fragmentation the total number of plots at different Numbe + locations r Distance from teff the distance of teff farms from Km + plot household's residence Livestock size Number of livestock TLU Training Farmers received training on related Dumm activities having a value of 1 if yes, 0 y otherwise Extension contact Frequency of extension service during Numbe the farming season r Credit Amount of credit received in the Birr production year Off/non farm Amount of off/non-farm income in Birr income the production year Improved seed It takes 1 if yes, 0 otherwise Dumm y Source: own computation, 2016
41
4. RESULTS AND DISCUSSION The chapter has been divided into two main sections. The first section deals with the results of descriptive analysis pertaining to socio-economics, demographic characteristics and various agricultural activities undertaken by sample household heads. In the second section, the econometric results related to level of technical efficiencies realized and factors affecting technical efficiency/inefficiency in teff production have been presented and discussed.
4.1. Descriptive Results The descriptive statistics presented in this section is comprised of various sub section. The discussion is included demographic and socio economic characteristics; institutional support; rate of input use, crop yields and description of variables used in SPF. 4.1.1. Demographic and Socio-Economic Characteristics of Sample Households Household size: Total numbers of individuals within the household determine the availability of labor power needed in the farm production. Family labor plays an important part in the success of a small- scale farming practices in that the farmer does not need to spend too much money on labor costs. In the study area, average household size for the sample farmers was about 4.31 adult equivalents per household. The largest household size was being 8.9 while the smallest size was 1.7 adult equivalents per household with standard deviation 1.49. Age of the household head: it is one of the important factors which determine the farming experience of the farmer. Diminution in the size of cultivated area and subdivision of holding are phenomena of long period. Age of household is important to study such a long period phenomenon, related with the change in farm size and extent of subdivision. All these contribute in determination of individual farm efficiency. The survey result showed that, the average age of the sample household heads was 44.43 years. Their age ranged from 23 to 80 years with standard deviations of 12.62. Education status of the household head: Education enhances the acquisition and utilization of information on improved technologies by farmers. Education together with increased experience could guide farmers to better manage their farm activities. Education upgrades the ability and changes the attitude of person in a given society.
42 Educated farmers were expected to adopt new agricultural technologies and had better managerial skill. An attempt was made to assess the educational status of the sample households who had informal and formal education (Lockheed et al., 1980). In the study area, the average years of formal schooling of sample farmers were found to be 3.50 years with standard deviations of 3.79. The maximum educational achievement for the sample farmers was grade 12. From the total sample household heads, 53.3% of the total sample household heads have attended formal level of schooling (Appendix Table 7). Sex of the household head: The survey result indicated that 16.11 percent of households are female-headed. It is understood that female-headed households face greater challenges in the agricultural production and marketing compared with their male-headed counterparts. This is due to the fact that female household heads in the rural Ethiopia hold various tasks including collecting of fire wood from the field, fetching water, childrearing and household management obligations. In addition, they have farm management tasks that increase the burden. Such multiple tasks combined with less resource accesses and ownership lead to more frequent and perhaps severe economic and social shocks particularly poverty and food insecurity. Table 2. Demographic and socio-economic characteristics of sample households Variable description Mean Std. Deviation Minimum Maximum Family size (adult equivalent) 4.31 1.49 1.7 8.9 Age of the HH 44.43 12.62 23 80 Education Level of the HH 3.50 3.79 0 12 Dummy variable Response Frequency Percent Sex of the household head Male 125 83.89 Female 24 16.11 Source: Own Survey, 2016 Marital status of the household head: Marital status of the household head with respect to teff producer was surveyed. Among the given sample households, 6.71% households were single. But, 80.54% households were married. The rest divorced and widowed household heads covered 4.7% and 8.05% respectively (Table 3). Females are head of households, when they were divorced or widowed, take responsibility and starting farming in addition to homemaking role.
43 Table 3. Distribution of sample households by their marital status Marital Status Number Married 120 Single 10 Divorced 7 Widowed 12 Total 149 Source: Own Survey, 2016
Percent 80.54 6.71 4.70 8.05 100
4.1.2. Labor Availability and Gender Role Family labor was the main source of labor for performing various farming activities for smallholder farmers. In the study area, it has been observed that there was a sort of labor division in various farm works between family members. Ploughing and planting were types of activities belonging to male whereas food preparation and childcare were left to female. In most of other cases than these both female and male worked together. Children participated in different farm and non-farm activities. In this specific study, labor availability of the sample household was calculated in man equivalent to examine the effect of variation in labor availability among the households. Because of differences in capacity and ability of performing a given activity between sex and different age groups labor force were standardized to a similar unit (man equivalent). The conversion factor used to standardize labor force has been given in appendix table 1. Majority of the sample households, 97.7% of the labor contribution was from family member and only 2.3% was supplied from hired labor. From total household labor supply for teff production, men supplied 66% labor while the children and women contribute 11% and 20.7%, respectively. In general, the average labor demanded per hectare of teff production was estimated to be 85.8 man days (Table 4). Table 4. Level of pre-harvest labour in teff production by gender (in Man Equivalent) Categorical description Mean Percent Male 56.63 66 Female 17.76 20.7 Child 9.44 11 Hired labor 1.97 2.3 Total 85.8 100 Source: Own Survey, 2016
44 4.1.3. Resource Basis 4.1.3.1. Average size of cultivated land holding by sample household heads An attempt was made to study the size of cultivated land holding by sample household heads of Jamma district. To mitigate the challenge of land shortage, young farmers usually shared land with their parents and relatives during marriage or obtained land use access through sharecropping and renting in land. The survey result indicates that 8.72% of the sample household had less than 0.5 hectare and 27.51% of household had more than 2 hectare of cultivated area. The analysis and pattern of cultivated land amongst sample households indicated that the average size of farm owned by the sample household heads were 1.48 ha (table 5). There were large variations in the distribution of the land holding among sample households. Above 40% of the households owned more than 1.48 ha of cultivated land. Table 5. Distribution of sample households by their Land holding structure Land holding Frequency Percent Less than 0.5 13 8.72 0.5-1.0 23 15.44 1.0-1.5 57 38.26 1.5-2.0 15 10.07 Above 2 41 27.51 Total 149 100.00 Average land holding 1.48 hectare Source: Own survey, 2016 The farming system of the district is mixed crop-livestock where crop plays the major role in the farmers’ income. Number of plots of major crops sown and their area coverage in the production year by sample farmers were presented in table 6. These listed crops were sown in different plots and highly fragmented and scattered over different places. A farmer on average has 2.9 plots covered with different annual crops. Teff, wheat and barley were the most frequently sown annual crops and cover 57.5%, 34.8% and 7.7% of the total plots respectively in 2015/16 production year. The average area covered by cereal crops was 1.48ha of which teff constituted 0.85 ha (divided 126.7 hectares of farm area under teff production by the total sampled household heads). The other areas were allocated for pulse crops, home garden, grazing land and other farm activities.
45 Table 6. Number of plots and area coverage of major crops sown by sample farmers Crop type Total area Total plots Ha % No. % Teff 126.7 57.5 257 59.5 Wheat 76.7 34.8 146 33.8 Barley 17.1 7.7 29 6.7 Total 220.5 100 432 100 Source: Own Survey, 2016 4.1.3.2. Ownership of livestock by sample household heads Livestock have diverse functions for the livelihood of farmers in mixed farming system. They provide food in the form of meat, milk, and non-food items such as draught power and manure as inputs into crop production. In addition, they were source of cash income and act as a store of wealth and play a determinant role in social status within the community and buffering risk. Cows were the largest class of livestock owned by sample farmers (Table 7). On average, the sample farmers owned 1.6 and 1.51 cows and oxen respectively. Donkey, mule and horse were employed for transport of farm inputs and outputs. Moreover, the sample respondent owned 0.93 sheep which were kept as a source of income and hedging against risk during the crop failure. Moreover, the survey result showed that the average tropical livestock unit per sample household was 6.1 per household. Table 7. Number of livestock owned by sample household heads Livestock Mean Std. Deviation Minimum Cows 1.60 1.34 0 Heifers 0.86 1.01 0 Oxen 1.51 1.16 0 Bulls 0.44 0.73 0 Calves 0.45 0.75 0 Sheep 0.93 1.47 0 Donkeys 0.12 0.35 0 Mules 0.09 0.36 0 Horses 0.10 0.43 0 TLU 6.10 3.52 0 Source: Own Survey, 2016
Maximum 10 5 6 4 4 11 2 3 3 18.9
The maximum livestock ownership in terms of TLU was 18.9. The survey result indicated that more than half of the sample households (60.4%) owned number of livestock above 6 TLU while 2.01% sample households had no livestock (Table 8). On the other hand 10.74% of the household heads had number of livestock within the range of 1 to 3 TLU and the remaining 26.85% of the household heads had within the range of 3 to 6 TLU.
46 Table 8. Distribution of sample household heads under various tropical local unit TLU Number of farmers Percent 0 3 2.01 0.1-3.0 16 10.74 3.0-6.0 40 26.85 Above 6 90 60.4 Total 149 100.0 Source: Own Survey, 2016 4.1.3.3. Level of oxen power utilization by sample household heads Oxen were the only sources of traction power in the area. Shortage of draught power limits the area that can be cultivated. Shortage of oxen power leads to poor land preparation and delayed completion of the operation. Poor land preparation leads to poor plant establishment, heavy weed infestation and low yields. The number of draught animals determines the amount of land that can be cropped, the types of crop grown, total crop production and yield (Mburu et al., 2014). Larger holding of oxen permit a greater area of land to be cultivated (Ogada et al., 2014). Oxen power was found as an important factor of production in the study area. Oxen power utilization by sample households was computed by assuming working of 8 hours by pair of oxen per day. Average oxen power used by the sample households in teff production was 22.71 oxen days with standard deviation of 12.26 (Table 12).Almost in all sample kebeles, farmers on average ploughed their land three to six times for production of teff. Usually the land preparation started from the first commencement of rain and they continued ploughing each month until sowing of the crop. Weed infestation was found to be a serious problem in the area due to the high rain fall from the month of June to August. It was also observed that the sample farmers in the study area gave more emphasis to ploughing as compared to weeding which is the major challenge for improving productivity. Given the above fact, 12.08% of the sample respondent had no oxen while 34.23% of them owned one pair of oxen. On the other hand, only 8.72% of the sample farmers owned more than two pair of oxen (Table 9). Table 9. Distribution of sample household head by number of oxen and teff crop coverage Number of oxen Average area covered Number of farmers Percent under teff 0 0.36 18 12.08 1 0.61 59 39.60 2 0.76 51 34.23 3 1.21 8 5.37 4 or more 1.33 13 8.72 Total 0.85 149 100 Source: Own Survey, 2016
47 4.1.3.4. Off/non-farm activities Farmers in the study area are engaged in various off/non-farm activities in parallel with the main farming activities during the farming season. Some of these activities are; grinding mills, handicraft, and selling of local drinks. The income they desperately need to obtain from such off/non-farm activities may substantiate the low income that is usually obtained from farming activities. In this study, the average amount of off/nonfarm income was birr 522.6 with standard deviation of 935.7 (Table 13). 4.1.4. Cropping System The dominant farming system of the district is mixed crop-livestock. Crop production of the district is limited to meher season and the major types of crop that are produced include teff, wheat, and barley from cereals, and beans and chick pea from pulses. Though modern input application especially fertilizer is there, the performance of major crops in terms of yield is not encouraging. The productivity level of major crops in 2015/16 production year is presented in table 10. The result indicated that on average sample farmers obtained 9.3 Qt of teff with minimum and maximum of 2 Qt and 25.25 Qt during a given production year, respectively. The result also indicated that sample households were obtaining average output of wheat and barley of 24.3 Qt and 14.25 Qt, respectively. On the hand the average output of bean and chickpea obtained by sample households in 2015/2016 production year were 6.65 Qt and 4.2 Qt, respectively (Table 10). Table 10. Average yield obtained by sample farmers in 2015/2016 production year Crop types Mean Minimum Maximum Teff 9.3 2 25.25 Wheat 24.3 4.5 34.5 Barley 14.25 5 19.7 Bean 6.65 1.25 8 Chickpea 4.2 1.85 13.9 Source: Own Survey, 2016 4.1.5. Description of Production Function and Variables This part present summary statistics results of production variables (both the physical inputs used in the production of teff output) used for analysis in the stochastic production frontier model. The result of analysis for output variable indicates that on average a household produced 9.3 Qt of teff output that ranges from 2 Qt to 25.25 Qt with standard deviation of 5.17 among the sample farmers in 2015/2016 production year (Table 11). This indicates the large variability of output among the farmers.
48 The average land area allocated to teff production (both owned and rented land) was approximately 0.85 ha and ranged from 0.25 ha to 2.3 ha with a standard deviation of 0.47 (Table 11). The mean land allocated to teff conforms to the fact that the farmers are smallscale and held family-managed and operated farm plots in the study area, which also confirms that, one of the characteristics of subsistence agriculture. The mean level of labor (both family and hired) used by teff growers in the study area was found to be 85.8 man-days, which was obtained by aggregating labor used for all teff production activities that include plowing, sowing, fertilizer application/top dressing, weeding, harvesting and threshing. The minimum and maximum level of labor (man-day) used were 20 and 235 man-days, respectively (Table 11). The average seed input (both improved and local seed) sown by teff producers in the study area during a given production year was 27.1 kg and which ranges from 3 to 54 minimum and maximum quantities of seed, respectively. Regarding fertilizer type, farmers in the study area commonly using DAP and Urea fertilizer. The summary result indicates the mean rate of fertilizer application of 265.2 Kg and which ranges from 50 Kg to 430 Kg minimum and maximum application rate, respectively. The use of oxen power in teff production activities like plowing and threshing in the study area is usual. The result indicated that the mean number of oxen power used was 22.7 oxen days, with the maximum and minimum of 66 and 6 pair of oxen days per season respectively (Table 11). Table 11. Descriptive statistics of variables used in production function estimation Variables Mean Std. Deviation Minimum Maximum Output (Qt) 9.3 5.17 2 25.25 Area (ha) 0.85 0.47 0.25 2.3 Labor (man days) 85.8 43.80 20 235 Seed (kg) 27.1 13.53 3 54 Chemical Fertilizer (kg) 265.2 75.04 50 430 Oxen power (pair oxen days) 22.71 12.26 6 66 Source: Own Survey, 2016 4.1.6. Major Teff Production Constraints Faced by Sample Household Heads The constraints that were found to operate in teff production in the study area were presented in table 12. The survey result indicated that Availability of quality improved seed, Sever weed infestation, Size of land holding, Damage of pests, Unfavorable weather condition and other constraints were observed as major constraint in teff production.
49 Table 12. Major teff production constraints during production year of 2015/2016 Type of problems Number of household heads Percent Availability of quality improved seed 57 46 Sever weed infestation 76 51 Size of land holding 109 73 Damage of pests 122 82 Unfavorable weather condition 138 93 Other 22 15 Source: Own Survey, 2016 4.1.7. Institutional Support Extension service: In order to give effective extension service to the farmers, the region assigned three DAs in each Kebele. The DAs are graduates of different ATVET colleges specializing in three agricultural streams such as, crop production, animal husbandry and natural resource management. In this study, 97% of the sample respondents reported that they have been receiving extension services out of which 90.6% received advice about teff production. The extension workers also visit farmers on different intervals. Some farmers are being visited more frequently while others have got no chance at all to be visited by extension workers. The survey result indicated that, the average frequency of extension contact of sample households was 3.68 times with standard deviation of 1.78 (Table 13). Credit: There exist both formal and informal lending institutions to provide credit. The formal sources of credit in the study area are Amhara Credit and Saving Institution (ACSI) and local cooperatives, whereas friends, relatives, traders, and the like are informal sources from which farmers could get credit. As far as the access to credit is concerned, on average birr 1890.5 from either source when they are in need of it provided that they fulfill the requirements set by the lending institutions (formal or informal) with standard deviation of 1969.9 (table 13). Nevertheless, the requirements and procedure to use credit from the formal institutions were not as easy as the local co-operatives and informal institutions. For instance, in the case of ACSI farmers were asked to form a group of ten to acquire credit. If any one of the group members was unable to pay back the amount he/she acquired, the remaining group members would be obliged to repay the total amount. Most of the time farmers face food shortage before the next new harvesting season. As a result 40% of the credit user farmers reported that they used the money to purchase food grains and medicines. On the other hand, 24% of credit users also reported that they used the credit to finance school expenses,
50 whereas about 36% of them used it to purchase farm inputs. Even if the constitution and other land laws prohibit to sale as well as use land as collateral, it is reported that land was asked as collateral by credit institutions (formal as well as informal). Livestock and other physical properties and crops were also used as collateral by most of the informal lenders. Training: An appropriate training given to the farmers may improve productivity by enhancing their management capacity. In the study area, farmers were getting training from surrounding research centers and other governmental and nongovernmental organizations. Among the sample farmers, 77.85% of farmers were trained on different teff related aspect and the rest had not received any training on the subject matter previously. This indicates that majority of the sample farmers were received (Table 13). Table 13. Institutional characteristics of the sample household Continues variable Mean Off/nonfarm incom 522.6 Number of extension contact 3.68 Credit 1890.5 Dummy variable Response Training Yes No Source: Own Survey, 2016
Std. Deviation 935.7 1.78 1969.9 Frequency Percent 116 77.85 33 22.15
4.2. Econometric Results This section presents the econometric results of the study. The results of technical efficiency level and factors affecting technical efficiency are discussed successively. Before running the econometric analysis, the data was tested against different econometric problems. Accordingly, the data was checked for hetroskedasticity using Breusch Pagan test, and the result showed absence of serious problem of hetroskedasticity (Appendix table 13). Multicollinearity test for both continuous and dummy variables at the same time was done using Variance Inflation Factor (VIF), to check multicolinearity problem among all variables entered in the model. In addition, multicolinearity test of continues and dummy variables were checked by using variance inflation factor and contingency coefficient, respectively. According to Gujarati (2004), value of VIF more than 10 is usually considered as an indicator of serious multicollinearity and should be excluded from the model. On the other hand, variables having variance inflation factor of less than 10 are believed to have no serious multicollinearity problem and able to included as explanatory variables in the model.
51 As a result, test for multicollinearity using both methods confirmed that there was no serious linear relation among explanatory variables (Appendix Table 3 and 4). 4.2.1. Hypotheses Testing The formulation and results of different hypotheses are presented in Table 14. All the hypotheses were tested by using generalized likelihood-ratio (LR). The first hypothesis related to the appropriateness of the Cobb-Douglas functional form in preference to translog model. The computed LR statistic was less than the critical value of chi-square at 5% probability level. The null hypothesis was accepted by indicating that the CobbDouglas functional form being a better representation of the data. These showed that the coefficients of the interaction terms and the square specifications of the input variables under the Translog specifications were not different from zero. Hence, Cobb-Douglas production function was the best to fit the data for estimation of technical efficiency for teff producing farm household in the study area. The second hypothesis was tested for the existence of the inefficiency component of the total error term of the stochastic production function. In other words, it was concluded whether the average production function (without considering the non-negative random error term) best fits the data. Hence, the second hypothesis stated that γ=0, was rejected at 5% level of probability confirming that inefficiencies existed and were indeed stochastic (LR statistic 14.64> chi-square =3.72). The coefficient for the parameter γ could be interpreted in such a way that about 71 percent of the variability in teff output in the study area was attributable to technical inefficiency effect, while the remaining about 29 percent variation in output was due to the effect of random noise. This implies that there was a scope for improving output of teff by first identifying those institutional, socioeconomic and farm specific factors causing this variation. The third hypothesis which stated the technical inefficiency effects were not related to the variables specified in the inefficiency effect model, was also rejected at 5% level of significance (LR statistic 25.43> critical chi-square =10.08). Thus the observed inefficiency among the teff farmers in the area could be attributed to the variables specified in the model and the variables exercised a significant role in explaining the observed inefficiency.
52 The fourth test conducted was, given such functional forms for the sample households; it was considered whether the technical efficiency levels were better estimated using a half normal or a truncated normal distribution of μi. The results indicated that the half normal distribution was appropriate for the sample households in the study area as the calculated LR value of 3.62 was less than the critical χ2 value of 4.24 at less than 5% probability level. Table 14. Generalized likelihood-ratio test of hypotheses for model and parameters Null hypothesis Degree of Calculate Critical Decision freedom d χ2 (LR) χ2-value 1. Production Function is CobbDouglas (H0 : βij = 0) 5 H1 :Translogproduction function 2. There is no inefficiency component (H0: γ = 0 ) H1 : there is difference in efficiency 1 3. The coefficients of determinants of inefficiency model equals zero (H0: δ1 =...= δ13 = 0) H1 : at least one of δ’s are not zero 4. H0: μ=0 distribution assumption H1: μ differ from zero
12.14
19.23
Accepted
14.64
3.72
Rejected
13
25.43
10.08
Rejected
1
3.62
4.24
Accepted
Source: Model Results, 2016 4.2.2. Maximum Likelihood Estimation of Parameters The maximum-likelihood estimates of parameters of the stochastic production frontier and inefficiency effect models as described with equations (15) and (16) were obtained after treating the dataset with STATA version 11.1. A stochastic production frontier model permits to consider production of teff in the study area with Cobb-Douglas stochastic production was tested and found to be best to fit the data. It was used to estimate efficiency of farmers and to identify factors determining the inefficiencies in teff producing farmers. Estimation of parameters was carried out with a one-stage procedure under the assumption of normal/half-normal distribution of the error terms. This approach leads us to the final estimates of parameters of the five explanatory variables of the frontier function; and thirteen explanatory variables which influence the mean efficiency of teff producing farmers.
53 The ML estimates of the parameters of the frontier production functions and inefficiency effects are presented in Table 15. The coefficients of the input variables were estimated under the full frontier production function (MLE). During the estimation, a single estimation procedure was applied using the Cobb-Douglas functional form. The computer program FRONTIER version 4.1 gave the value of the parameter estimations for the frontier model and the value of
2
. Moreover it gave the value of Log-likelihood function
for the stochastic production function. The Maximum Likelihood estimates of the parameter of SPF functions together with the inefficiency effects model are presented in Table 15 below. Out of the total five variables considered in the production function, four (land, labor, oxen power and fertilizer) had a significant effect in explaining the variation in teff production among farmers. The coefficients of production function variables were positive. The coefficients of land, oxen power and fertilizer were significant at 1% level of significance, and the coefficient of labor was significant at 10% level of significance. This informs that they were significantly different from zero and hence these variables were important in explaining teff production in the study area. The positive production elasticity with respect to land, fertilizer, oxen and labor imply that as each of these variables increase, teff output will increase. On average, as the farmer increases area allocated to teff, amount of chemical fertilizer application, labor and oxen power for the production of teff by 1% each, he/she can increase the level of teff output by 0.859, 0.668, 0.169 and 0.05 percent, respectively. Table 15. Maximum likelihood estimate for Cobb-Douglas production function Variable Maximum likelihood estimate Coefficient SE Z- value Intercept 1.971*** 0.288 6.84 lnOxen 0.050*** 0.019 2.63 lnLabor 0.169* 0.100 1.69 lnSeed 0 .013 0 .01 1.30 lnFertilizer 0.668*** 0.071 9.41 lnArea 0.859*** 0.188 4.57 Sigma-squared 0.14*** 0.024 5.83 Log likelihood -52.16 Gamma 0.756 Return to scale 1.74 Mean technical efficiency 0.78 Total sample size 149 ***, **, * represents significance at 1%, 5% and 10%probability levels, respectively Source: Own Survey, 2016
54 Summing the individual elasticity yields a scale elasticity of 1.74. This indicates that farmers are facing increasing returns to scale (Table 15) and depicts that there is potential for teff producers to increase their production. In other words, they are not efficient in allocation of resource this implies production is inefficient moreover there is a room to increase production with an increasing rate. 4.2.3. Variability of Output due to Technical Efficiency Differentials The Maximum Likelihood estimation of the frontier model gave the value for the parameter (γ), which is the ratio of the variance of the inefficiency component to the total error term (γ = σ2u /σ2s , where σ2 s = σ2v + σ2u ). The γ value indicated the relative variability of the one sided error term to the total error-term. In other words, it measured the extent of variability between observed and frontier output that is caused by the technical inefficiency. As a result the total variation in output from the maximum may not have necessarily caused efficiency differentials among the sample households. Hence, the disturbance term had also contributed in varying the output level. In this case, it was crucial in determining the relative contribution of both usual random noises and the inefficiency component in total variability. The TE analysis revealed that technical efficiency score of sample farmers varied from 15% to 95%, with the mean efficiency level of 78%. This variation was also confirmed by the value of gamma (γ) that was 0.756. The gamma value of 0.756 suggested that 75.6% variation in output was due to the differences in technical efficiencies of farm household in Jamma District while the remaining 24.4% was due to the effect of the disturbance term. 4.2.4. Technical Efficiency of Farmers One of the objectives of this study was to measure the technical efficiency levels of teff producing farmers in Jamma district. Given the chosen functional form used, estimation procedure implemented and the distributional assumptions made about the two error terms vi and ui, the technical efficiencies were estimated. The estimation result showed that the mean efficiency level of teff farmers were 78%, with the minimum and maximum efficiency level of about 15 and 95%, respectively (Appendix Table 9). This shows that there is a wide disparity among teff producer farmers in their level of technical efficiency which may in turn indicate that, there exists a room for improving the existing level of teff production through enhancing the level of farmers’ technical efficiency.
55 The mean level of technical efficiency further tells us that the level of teff output of the sample respondents can be increased on average by about 22% if appropriate measures are taken to improve the level of efficiency of teff growing farmers. In other words, there is a possibility to increase yield of teff by about 22% using the resources at their disposal in an efficient manner without introducing any other improved (external) inputs and practices. It also indicated that small farms in the study area, on average, can gain higher output growth at least by 17.1% (1-78/95) through the improvements in the technical efficiency. Moreover, from the total sample households, more than half scored above the mean TE score while almost half of sample respondent produces less than the mean TE score of farmers in their vicinity( Table 16). Table 16. Summary of technical efficiency differentials among sample household heads Efficiency Number Percent Mean Std. Deviation Minimum Maximum category of HHH 0.15-0.65 23 15.44 0.59 0.13 0.15 0.65 0.66-0.75 48 32.21 0.71 0.03 0.66 0.74 0.76-0.85 32 21.48 0.81 0.03 0.76 0.85 0.860-0.90 19 12.75 0.87 0.01 0.86 0.90 Above 0. 90 27 18.12 0.93 0.01 0.91 0.95 Total 149 100 0.78 0.13 0.15 0.95 Source: Own Survey, 2016 4.2.5. Input Use and Technical Efficiency Grouping of sample respondent based on their efficiency score was based on the relative performance of each sample households to the mean efficiency level. According to Stevenson (1980), grouping can also be done based on the relative performance of each sample household to the mean performance level. In this case, sample households were categorized into three groups based on the mean efficiency and the corresponding standard deviation. The sample households were considered as less efficient if they were operating at less than the mean minus standard deviation, high efficient if they were operating at more than the mean plus standard deviation. Hence, three sample farmers` categorical groups can be identified as the less efficient, moderately efficient and high efficient farmers based on their technical efficiency scores. In this respect, farmers are considered as moderately efficient if they were operating in the rage of mean efficiency plus or minus standard deviation, and less efficient or high efficient farmers if they used to operate below or above the average efficiency range, respectively.
56 Accordingly there were three groups. The efficiency percentage in input utilization and corresponding yield were obtained from the sample farm households in each groups (Table 17). Input use and yield varied across the three assumed efficiency group were summarized based on the technical efficiency score. Overall efficiency score was approximately 0.78 with standard deviation of 0.13. Producer`s having high efficiency score utilized more human labor, fertilizer, oxen days and seed than producers who were technically moderate and less efficient. The input utilization across various levels of technical efficiency score was also analyzed. The group possessing high TE used 27.1 oxen days and 123.8 MD in pre harvest agricultural operations. In addition, they used teff seed 54 kg. With regard to the seed utilization, it was observed that the less efficient group used less quantity compared with high and moderately efficient farmers. It implied that this group of farmers used very less quantity of improved and composites seed sold as compared to the most efficient farmers (Table 17). On the other hand, the quantity for inorganic fertilizers for more, moderately and less efficient farmers valued around 430 kg, 261.6 kg and 161.7 kg respectively Table 17. Utilization of production inputs and technical efficiency differentials Efficiency Group Grou SD TE Yield Oxen Labor Seed Group category p (%) score (Qt) (OD) (MD) (kg) High Above 0.91 17.0 0.03 0.94 16.1 27.1 123.8 54 efficient Moderately 0.65-0.95 67.5 0.04 0.79 9.1 24.2 87.3 33.8 efficient Less Below 0.65 15.5 0.03 0.59 4.5 22.4 47.0 16.5 efficient Overall 15-95 100 0.13 0.78 9.34 22.7 85.8 27.1 Source: Own Survey, 2016
Fertilize r (kg) 430 261.6 116.7 265.2
4.2.6. Estimated Actual and Potential Level of Teff Output Applying equation 19, the potential attainable level of teff yield was analyzed. The difference between the actual level and the frontier level of output was computed by estimating the individual and the mean level of frontier output. Using the values of the actual output obtained and the predicted technical efficiency indices, the potential output was estimated for each sample farm households. The mean levels of the actual and potential output during the production year were 9.34 Qt and 11.98 Qt, with the standard deviation of 5.16 and 6.57, respectively.
57 Moreover, paired sample t-test was used on the actual and potential yield to compare the difference in the amount of yield between two scenarios. There was a significant difference between potential yield and actual yield at 1% significance level. Potential yield was also calculated for each farm and the results were presented by range of technical efficiency group (table 18). In general, for the less efficient, moderately efficient and high efficient farm households, the recorded average actual yield was 4.43 Qt, 9.11 Qt and 16.06 Qt, respectively. Their corresponding less efficient, moderately efficient and high efficient group potential yield was 7.51 Qt, 11.53 Qt and 17.09 Qt, respectively. Table 18. Comparison of estimated actual yield and potential teff yield Potential yield Actual yield Efficiency category t-test Mean Std. Deviation Mean Std. Deviation 0.15-0.65 7.51 1.79 4.43 1.63 0.66-0.75 9.20 3.11 6.53 2.32 0.76-0.85 10.58 3.48 8.57 2.83 0.86-0.90 11.90 4.93 10.36 3.97 Above 0. 90 17.14 7.92 15.94 6.52 High efficient 17.09 8.14 16.06 6.69 Moderately efficient 11.53 4.65 9.11 3.65 Less efficient 7.51 1.79 4.43 1.63 Overall 11.98 6.57 9.34 5.16 3.48*** *** represent significance at 1% probability levels Source: Own Survey, 2016 4.2.7. Determinant of Technical Efficiency The focus of this analysis was to provide an empirical evidence of the determinant productivity variability/inefficiency gaps among smallholder teff farmers in the study area. Merely having knowledge that farmers were technically inefficient might not be useful unless the sources of the inefficiency are identified. Thus, in the second stage of this analysis, the study investigated farm and farmer-specific attributes that had impact on smallholders` technical efficiency. The driving force behind measuring farmer’s efficiency in teff 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 were estimated though MLE method using one-stage estimation procedure. The determinants of technical inefficiency/efficiency in a
58 given period vary considerably depending on the socio-economic conditions of the study area particularly pertaining to managerial characteristics and other related factors. Before discussing the significant factors which influencing inefficiency in teff production, it is important to see how efficiency and inefficiency are interpreted. The result can be presented in terms of efficiency or in terms of inefficiency. The result in the table 19 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 contribute 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 and vice versa. Accordingly, the negative and significant coefficients of age of the household head, education, improved seed, training and credit indicate that improving these factors contribute to reducing technical inefficiency. Whereas, the positive and significant variable such as farm size, affect the technical inefficiency positively that is increases in the magnitude of these factors aggravate the technical inefficiency level. The implications of significant variables on the technical inefficiency of the farmers in the study area were discussed here under. Table 19. Maximum-likelihood estimates of technical inefficiency determinants Variables Coefficients SE Z- value Constant 1.039*** 0.332 3.13 Sex -0.142 0.096 -1.48 Age -0.006** 0.003 -2.01 Household size -0 .061 0.041 -1.48 Education -0 .709*** 0.095 -7.46 Farm size 0.188*** 0.063 2.98 Improved seed -0.409*** 0.092 -4.44 Off/non farm income 0.002 0.005 0.40 Training -0.110* 0.060 -1.83 Credit -0 .523** 0.244 -2.14 Extension contact -0.008 0.019 -0.42 Land fragmentation 0.079 0.050 1.58 Distance to teff field 0.031 0.023 1.34 TLU 0 .0089 0.010 0.89 Log likelihood -52.16 Total sample size 149 *, **, *** represents significant at 10%, 5% and 1% probability level respectively Source: Own Survey, 2016
59 Age of farm household heads: The age of the household is the proxy for the experience of the household head in farming. The result indicated that age of the household heads influenced inefficiency negatively at 5% level of significance. This suggested that older farmers were more efficient than their young counterparts. The reason for this may probably be that the farmers become more skill full as they grow older due to cumulative farming experiences (Liu and Zhung, 2000). Moreover increase in farming experiences leads to a better assessment of the important and complexities of good farming decisionmaking including efficient use of input. This result was consistent with the arguments by Evaline et al. (2014, Mesay et al. (2013) and Ogada et al. (2014) they indicated that, since farming as any other professions needs accumulated knowledge, skill and physical capability, it is decisive in determining efficiency. The knowledge, the skills as well as the physical capability of farmers is likely to increase as their age increases. Education: Education enhances the acquisition and utilization of information on improved technology by the farmers. In this study, education measured in years of formal schooling, as expected, the sign of education was negative effect on technical inefficiency at 1% level of significance. This implying that less educated farmers are not technically efficient than those that have relatively more education. This could be because; educated farmers have the ability to use information from various sources and can apply the new information and technologies on their farm that would increase outputs of teff. In general, more educated farmers were able to perceive, interpret and respond to new information and adopt improved technologies such as fertilizers, pesticides and planting materials much faster than their counterparts. This result was in line with the findings of Tefera et al. (2014), Ali and Khan (2014), Hailemariam (2015), Fantu et al. (2015b), Ouedraogo (2015) and Michael and James(2017) who stated that an increase in human capital will augment the productivity of farmers. Farm size: It is measured as total land cultivated by the farmer including those rented and shared in. In this study, it was hypothesized that farm size affects inefficiency positively. As the farm size of a farmer increases the managing ability of him/her will decrease given the level of technology, this lead to reduce the efficiency of the farmer. Accordingly, the estimated result coincides with the expectation and that coefficients of this inefficiency variable found positive and statistically significant. That means total area cultivated by a household affected technical inefficiency level positively and significantly at 1% level of significance.
60 This shows that a household operating on large area is less efficient than a household with small land holding size. This might be because an existence of increased in area cultivated might entail that the farmer might not be able to carry out important crop husbandry practices that need to be done on time, given his limited access to resources. As a result, with increase farm holding size the technical inefficiency of the farmer might increase. This finding was in line with results obtained by Getachew and Bamlak (2014), Sultan and Ahmed (2014), Mwajombe1 and Mlozi (2015)and Kabir et al. (2015). Improved seed: Use of improved seeds negatively and significantly affected farmers` technical inefficiency in teff production at 1% level of significance. Thus, production of teff through the use of more of improved teff seeds was more efficient compared to using local seeds. This was in agreement with the findings of Sultan and Ahmed (2014) and Mukete et al. (2016). Moreover, the negative sign of the estimated coefficients had important implications on the technical inefficiency of the teff farmers in the study area. It means that the tendency for any teff farmers to increase his production depend on the type and quality of improved seed available at the right time of sowing. Training: Training is an important tool in building the managerial capacity of the household head. Household’s head that get training related with crop production and marketing or any related agricultural training are hypothesized to be more efficient than those who did not receive training. Training of farmers on teff crop was important because it could improve farmers` skill regarding production practices and related aspects. A number of farmers in the study areas received training on teff for few days mainly on production practices and importance of using improved package. The dummy coefficient of training was negative and significant in the technical inefficiency model of teff production at 10% level of significance. This implied that technical inefficiency effect decreases with farmers having training on teff. It may also be concluded that farmers with training on teff tended to have lower inefficiency effects than farmers without training. That is, farmers with training were technically more efficient than farmers without training. This result is in line with the arguments by Beyan et al. (2013), Getahun (2014), Birhan (2015) and Michael and James (2017) who indicated that training given outside locality relatively for longer period of time determined inefficiency negatively and significantly.
61 Credit: It is an important element in agricultural production systems. It allows producer to satisfy their cash needs induced by the production cycle. Amount of credit increases farmers’ efficiency because it temporarily solves shortage of liquidity/working capital. In this study, amount of credit was hypothesized in such a way that farmers who get more amount of credit at the given production season from either formal or informal sources were expected to be more efficient than those who get less amount of credit. In this study, amount of credit affected inefficiency of farmers negatively and significantly at 5% level of significance. This implies that credit availability shifts the cash constraint outwards and thus enables farmers to make timely purchases of inputs that they cannot afford otherwise from their own resources and enhances the use of agricultural inputs that leads to more efficiency. The empirical studies conducted by Gebregziabher et al. (2012), Musa et al. (2014) and Biam et al. (2016) found positive and significant relationship between credit and farmers` technical efficiency which was in line with this study.
62
5. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 5.1. Summary Productivity can be improved in two ways either by introducing new agricultural production technologies or improving the technical efficiency levels of farmers which is the possible strategies to increase the productivity of the agricultural sector in the country. Technical efficiency has remained an important subject of empirical investigation particularly in developing economies where majority of the farmers are resource poor. Alternatively, productivity growth may attribute to either technological progress or efficiency improvement. Improving technical efficiency of the farmer plays a great role in increasing productivity, given the current state of technology. The main objective of this study was measuring efficiency level of teff farmers and identifying those factors which affect technical efficiency of teff production in Jamma district of Amhara National Regional State. Data were collected for the 2015/16 production season by interviewing a total of 149 sample teff producing farmers using a structured questionnaire that encompasses question related to demographic characteristics, inputs and output, institutional and farm specific characteristics. Three- stage sampling technique was employed for selecting the respondents. Data were analyzed using both descriptive statistics and econometric model. The stochastic frontier production function of the CobbDouglas functional form was found to be best fitted the data to estimate the level of individual technical efficiency. Cobb-Douglas functional form constitutes five input variables in frontier function and thirteen explanatory variables in an inefficiency model. Direct or one stage estimation procedure was used to incorporate exogenous factors directly in the production frontier model. The study has also conducted a test of hypothesis which states a hundred percent efficiency. The hypothesis that technical efficiency effects are absent, given the specification of Cobb- Douglas stochastic frontier production function, was rejected based on the results of the econometric model. This shows that the technical inefficiency exists in the sample farmers considered and hence, the average response function that all farmers are fully technically efficient is not supported by the result obtained from statistical analysis of the data.
63 The estimated stochastic production frontier model indicated that area of teff, chemical fertilizers, labor and oxen power is significant determinants of teff output. The positive coefficient of these parameters indicates that increased use of these inputs will increase the production level to a higher extent. Hence, given that these inputs are used to their maximum potential, introduction and dissemination of these inputs will increase the production level of teff in the study area. The analysis also revealed that the sum of the partial output elasticity’s with the respective inputs is 1.74. This result indicates that production structure was characterized by increased returns to scale. The implication of this increased return to scale is that a proportional increase in all the factors of production leads to the high proportional increase in output. The value of the discrepancy ratio, γ, calculated from the Maximum Likelihood estimation of the frontier was, about 0.756. The estimated result of the Cobb-Douglas production frontier indicated that significant proportion of the variation in the stochastic frontier production function being due to technical inefficiency. This implies that presence of chance for improvement of farmers’ productivity through better technical efficiency. The mean technical efficiency level of farmers in teff production was 0.78 and it’s ranging from 0.15 to 0.95. The mean technical efficiency level of 78 percent indicates that production can be increased by 22 percent of the potential in those farmers who grow teff through better use of the available resources, given the current state of technologies. Moreover, there is a wide variability in the technical efficiency level of farmers, and only few farmers attained efficiency levels of more than 90 percent for teff production in the study area. The socio-economic variables that are important in determining farmers` level of technical efficiency were also identified. Accordingly, the results of technical efficiency effects model showed; age, education, farm size, improved seed, training and credit found to be the major determinants of efficiency level of the farmers in teff production. The negative coefficients of age, education, credit, training and improved seed in inefficiency model means that these factors positively affect efficiency of the farmers in the area where they are significant. While, the positive coefficients of farm size in inefficiency model indicated that these factors determine efficiency negatively.
64 In general, the SPF model showed that production can be improved by increasing the use of inputs. There is considerable room to improve the efficiency of farmers in teff production. The implication is that, there will be considerable gain in production level if introduction and distribution of agricultural technologies is joined with improving the existing level of efficiency. The significant inefficiency effect explanatory variables have important policy and development implications in an effort towards improving the technical efficiency of teff production in the study area. 5.2. Conclusions and Recommendations The implication of this study is that, technical efficiency of the farmers can be increased through better allocation of the available resources especially: land, oxen power, labor, and fertilizer. Thus, local government or other concerned bodies in the developmental activities working with the view to boost production efficiency of the farmers in the study area should work on improving productivity of farmers by giving especial emphasis for significant factors of production. Moreover, age should be considered in increasing resource use efficiency and agricultural productivity. This is because results showed that younger farmers are technically more inefficient than older ones. It implies that there should be policies to improve resource use efficiency of younger farmers and encourage them to be in farming activities by providing them incentives. Continues trainings on the agricultural business environment and follow up during agricultural operation for younger farmers should be provided. However, this should not be at the expense of older ones. Training determined technical efficiency positively and significantly in teff producing farmers. Provision of training for farmers to improve their skills in use of improved seed, resource management, post-harvest handling, and general farm management capabilities will increase their farm productivity. In addition to strengthening the practical training provided to farmers, efforts should be made to train farmers for relatively longer period of time using the already constructed farmers` training centers and agriculture research demonstration centers. The amount of credit received was found to positively and significantly influence household technical efficiency level. But Smallholder framers in the study area have financial constraints. This could imply that households needed external financial sources to
65 solve their own financial constraints. Therefore, Amhara Credit and Saving Institution (ACSI) have mandated to provide relatively high amount of credit for farmers should be encouraged and strengthen to deliver more than this and also harmonization loan delivery with the time input required and loan payment plans with harvesting seasons. In addition to this the regional government should intervene to strength the operation of rural saving and credit institutions at village level and creates awareness for those farmers to improve their saving habits so as to improve their asset formation. Total farm size was negatively and statistical significantly related with technical efficiency in teff production. This may be due to the nature of more time requirement of teff farming and demanding close supervision of the farm operator which share significant part of his/her time. Hence, smallholder farmers have a limitation of resources which are used for agriculture production on his/her available farm land in the given operation calendar, his/her production performance affected negatively. So, this gives a direction to use technologies like tractor, combiner for facilitating the farm operations work within the specified operational calendar. This in turn improves the production of the farmers due to using better technology which shifts the production frontier outward. Therefore, it would be better if the regional government or concerned body facilitate such machinery services either on credit bases or cooperative rendering rental service. Those farmers that are more educated are relatively more technically efficient than less educated ones, in the study area. This may be due to the fact that more educated farmers have access to information and better communication media that helps them to use modern teff production technologies. Education is fundamental in improving the technical efficiency of farmers. Therefore, the regional governments need to strengthen farmers’ access to education that could be implemented through expansion of farmers training center or expansion of formal and non- formal education in the area. Improved teff seed had a significant and negative effect on technical inefficiency of teff production. Even if, there are 33 improved teff seed and particularly 12 improved teff seed released for moisture stress areas, only Quncho was distributed to farmers during the 2015/16 production year. Hence, researchers and extension agent should improved the awareness of farmers to use improved teff seed and efforts should be made to access different types of high yielding teff seed before the starting of sowing date.
66
6. REFERENCES Abba, W. 2012. Technical efficiency of sorghum production in Hong local government area of Adamawa State, Nigeria. Russian Journal of Agricultural and Socio Economic Sciences, 6(6): 10-15. Abu Tefera and Quintin, G. 2013. Ethiopia grain and feed annual report. Report Number: ET1301. Addis Ababa, Ethiopia. Aigner, D.J. and Chu, S.F. 1968. Estimating the industry production function. American Journal of Economic Review, 58(4): 826-839. Aigner, D.J., Lovell, C.A. and Schmidt, P. 1977. Formulation and estimation of stochastic frontier production function models. Journal of Economics, 6(1): 21-37. Al-hassan, S. 2012. Technical efficiency in smallholder paddy farmers in Ghana: an analysis based on different farming systems and gender. Journal of Economics and Sustainable Development, 3(5): 2222-2855. Ali, S. and Khan, M. 2014. Technical efficiency of wheat production in district Peshawar, Khyber Pakhtunkhwa, Pakistan. Sarhad Journal of Agriculture, 30(4): 433-441. Awol Ahmed. 2014. Economic efficiency of rain-fed wheat producing farmers in North Eastern Ethiopia. MSc Thesis, Haramaya University, Haramaya, Ethiopia. Battese, G.E. and Corra, G.S. 1977. Estimation of a production frontier model: with application to the pastoral zone of Eastern Australia. Australia Journal of Agricultural Economics, 21(3): 169-179. Battese, G.E., Coelli, T.J. and Colby, T.C. 1989. Estimation of frontier production factors and the efficiencies of Indian farms using panel data from ICRISAT`S village level studies. Journal of Quantitative Economics, 5: 327-348. Battese, G.E. and Coelli, T.J. 1992. Frontier production functions, technical efficiency and panel data with application to paddy farmers in India. Journal of Productivity Analysis, 3(2): 153-169. Battese, G.E. and Coelli, T.J. 1993. A stochastic frontier production function incorporating a Model for technical inefficiency effects: working paper in econometrics and applied stastistics No. 69. Department of Economics, University of New England, Armidale, Australia. Battese, G.E. and Coelli, T.J. 1995. Model for technical efficiency effects in a stochastic frontier production function for panel data. Journal of Empirical Economics, 20(2): 325-332.
67 Beyan Ahmed, Jema Haji and Endrias Geta. 2013. Analysis of farm households' technical efficiency in production of smallholder farmers: the case of Girawa District, Ethiopia. Journal of Agriculture and Environmental Science, 13(12): 1615-1621. Biam, C. K., Okorie, A. and Nwibo, S.U. 2016. Economic efficiency of small scale soybean farmers in Central Agricultural Zone, Nigeria: A Cobb-Douglas stochastic frontier cost function approach. Journal of Development and Agricultural Economics, 8(3): 52-58. Birhan Tegegne. 2015. Determinants of technical, allocative and economic efficiencies among crop producing farmers in irrigated agriculture. African Journal of Agricultural Research, 10(20): 2180-2189. Burhan, O., Ceylan, R.F. and Hatice, K. 2009. Review of literature on productive efficiency in agricultural production. Journal of Applied Sciences Research, 5(7): 796-801. Bwala, M.A., Kokoye, H.S. and Yegbemey, R.N. 2015. Technical efficiency of cereal production in the north central Nigeria: A case for maize, rice and sorghum farmers. Journal of Agricultural Science and Environment, 15(1): 25-34. Chandrasekhar, V. and Nikita, G. 2015. Economic efficiency of ring seiners operated off Munambam Coast of Kerala using Data Envelopment Analysis. Journal of Agricultural Economics Research Review, 28 (1): 171-177. Coelli, T.J. 1995. Recent development in frontier modeling and efficiency measurement. Australian Journal of Agricultural Economics, 39(3): 219-245. Coelli, T.J. 1996. A guide to FRONTIER version 4.1: a computer program for frontier production function estimation. CEPA Working Paper 96/07, Department of Econometrics, University of New England, Armidale, Australia. Coelli, T.J., Rao, D.S. and Battese, G.E. 1998. Introduction to efficiency and production analysis. Kluwer Academic Publisher, USA. Coelli, T.J. and Battese, G. E. 2005. An Introduction to Efficiency and Productivity Analysis, 2nd Edition. Kluwer Academic Publishers, Boston. CSA (Central Statistical Agency). 2008. Summary and Statistical Report of the 2007 Population and Housing Census of Ethiopia. Addis Ababa, Ethiopia. CSA (Central Statistical Agency). 2012. Agricultural sample survey for the season of 2011/2012: Area and production of major crops. Volume I: Statistical Bulletin. Addis Ababa, Ethiopia. CSA (Central Statistical Agency). 2014a. Agricultural sample survey: 2004/5- 2013/14 Volume I: Report on Land Utilization (Private Peasant Holdings, Meher Season). Addis Ababa, Ethiopia.
68 CSA (Central Statistical Agency). 2014b. Agricultural sample survey: 2004/5- 2013/14 Volume IV: Report on Area and Production of Crops (Private Peasant Holdings, Meher Season). Addis Ababa, Ethiopia. CSA (Central Statistical Agency). 2015. Agricultural sample survey. Addis Ababa, Ethiopia. Dorosh, P., Robinson, S. and Thurlow, J. 2015. Implications of agricultural productivity growth for structural change and employment in Ethiopia. Paper presented at the EDRI-IFPRI Seminar, Addis Ababa, Ethiopia. Essa channe, Obare, A., Ayalneh Bogale and Simtowe P. 2012. Analysis of resource use efficiency in smallholder mixed crop-livestock agricultural systems: Empirical evidence from the central highlands of Ethiopia. Journal of Developing countries Study, 2(9): 3040. Evaline, C.E., Eric, k.B. and Kwana, K. 2014.Analysis of technical efficiency of sorghum production in lower eastern Kenya. Journals of Economics and sustainable Development, 5(4): 58-65. Fantu Nisrane, Guush Berhane, Sinafikeh Gerawork, Alemayahu Seyoum and John, H. 2011. Sources of inefficiency and grow in agricultural output in subsistence agriculture. ESSPІІ Working paper 019. Research Institute-Ethiopia Strategy Support Program ІІ. Addis Ababa, Ethiopia. Fantu Nisrane, Bethlehem Koru and Alemayehu Seyoum. 2015a. Cereals productivity and its drivers: The case of Ethiopia. IFPRI/ESSP II Working Paper 74. International Food Policy Research Institute, Addis Ababa, Ethiopia. Fantu Nisrane, Bethlehem Koru and Alemayehu Seyoum. 2015b. Smallholder teff productivity and efficiency: Evidence from High-Potential Districts of Ethiopia. A paper presented at the 29th International Conference of Agricultural Economists in Milan 8-14 August 2015. Farrell, M.J. 1957. The measurement of productive efficiency. Journal of Royal Statistical Society, 120(3): 253-290. Gebregziabher Gebrehaweria, Namara Regassa and Holden S. 2012. Technical efficiency of irrigated and rain-fed smallholder agriculture in Tigray, Ethiopia. Journal of International Agriculture, 51(3): 203-226. Gemeda Olani. 2011. Analysis of wheat production efficiency: the case of small-scale farmers in Gedeb-hasasa District, West Arsi Zone of Oromia Regional State. Msc Thesis, Haramaya University, Haramaya, Ethiopia.
69 Getahun Gemechu. 2014. Off-farm income and technical efficiency of smallholder farmers in Ethiopia. Journal of Agricultural Economics research review, 23(2): 167-183. Getachew Magnar and Bamlak Alamirew. 2014. Analysis of technical efficiency of smallholder maize growing farmers of Horo Guduru Wollega Zone, Ethiopia. Journal of Science, Technology and Arts Research Journal, 3(3): 204-212. Gong, B.H. and Sickles, R.C. 1992. Finite sample evidence on performance of stochastic frontier and data envelopment analysis using panel data. Journal of Econometrics, 51(1): 259284. Greene, W.H. 1980. Maximum Likelihood Estimation of Econometric Frontier Function. Journal of Econometrics, 13(1): 27-56. Greene, W.H. 1990. On the estimation of a flexible frontier production. Journal of Econometrics, 13(4): 101-115. Greene, W. H. 2003. Econometric Analysis, 5th edition. Macmillan, New York, pp 1083. Gujarati, D. N. 2004. Basic Econometrics, 4th Edition. McGraw Hill Companies, New York, pp 341-375 Hailemaraim Leggesse. 2015. Technical efficiency in teff production: the case of Bereh District of Ethiopia. MSc Thesis, Haramaya University, Haramaya, Ethiopia. Hassen Beshir, Bezabih Emana, Belay Kassa and Jema Haji. 2011. Factors affecting the adoption and intensity of use of improved wheat varieties in North East Ethiopia: a double hurdle approach. Ethiopian Journal of Agricultural Economics, 8 (1): 133-158. Hassen Beshir. 2011. Performance of mixed crop-livestock production system: the data envelopment approach. Published at Livestock Research for Rural Development. 23(9) Article #200 Retrieved, from http://www.lrrd.org/lrrd23/9/besh23200.htm. (Accessed October 2016) Hassen Beshir, Bezabih Emana, Belay Kassa and Jema Haji. 2012a. Determinants of chemical fertilizer technology adoption in north eastern highlands of Ethiopia: the double hurdle approach. Journal of Research in Economics and International Finance, 1(2): 39-49. Hassen Beshir, Bezabih Emana, Belay Kassa and Jema Haji. 2012b. Economic efficiency of mixed crop-livestock production system in the north eastern highlands of Ethiopia: the stochastic frontier approach. Journal of Agricultural Economics and Development, 1(1): 10-20. Hassen Beshir. 2013. Factors affecting the adoption and intensity of use of improved forages in North East Highlands of Ethiopia. American Journal of Experimental Agriculture, 4(1): 12-27.
70 Hassen Beshir. 2014. Economics of soil and water conservation: the case of smallholder farmers in North Eastern Highlands of Ethiopia. Published at proceeding of the 11th international conference of the Ethiopian economics Association, July 2014. Hassen Beshir. 2016. Technical efficiency measurement and their differential in wheat production: the case of smallholder farmers in South Wollo Zone, Ethiopia. International Journal of Economics, Business and Finance, 4(1): 1-16. Hazarika, C. and Subramanian, S.R. 1999. Technical efficiency of government supported horticultural crops in Oman. Indian Journal of Agricultural Economics, 54(2): 201-210. Hazell, P. 2013. Is Small farm led development still a relevant strategy for Africa and Asia? The Journal of Agricultural Sciences, 5(3): 2012-2019. Henderson, H. 2014. Structural transformation and smallholder agriculture: an informationtheoretic analysis of the Nicaraguan Case. Journal of Agricultural Economics, 45(4): 443-458. Kabir, H., Musharraf, M., Haque, M. and Khan, M. 2015. Technical efficiency of Boro rice production in Bangladesh. Journal of Bangladesh Agricultural University, 13 (1): 101108. Kalirajan. K.P. 1991. The importance of use in the adoption of technology: A macro panel data analysis. Journal of production analysis, 2(2): 113-126. Kinfe Aseyehegn, Chilot Yirga and Sundar, R. 2012. Effect of small scale irrigation on the income of rural farm households: The case of Laelay Maichew Distinct, central Tigray, Ethiopia. The Journal of Agricultural Sciences, 7(1): 43-57. Kopp, R.J. and Smith, V.K. 1980. Frontier production function estimates for steam electric generation: a comparative analysis. Journal of Southern Economic, 46(4): 1049-1059. Liu, Z. and Zhuang, J. 2000. Determinants of technical efficiency in post- collective Chinese agriculture: Evidence from Farm-Level Data. Journal of Comparative Economics, 28(3): 545-564. Lockheed, M.E., Jamison, D.T. and Lau, L.J. 1980. Farmer education and farm efficiency: A survey. Economic Development and Cultural Change,29(1): 37-76. Luke, N.A., Atakelty Hailu and Amin, W.M. 2012. Analysis of farm household technical efficiency in Northern Ghana using bootstrap DEA. Paper presented at the 56th Annual conference of the Australian Agricultural and Resource Economics Society. Western Australia, Australia.
71 Mburu, D., Chris, A.O. and Richard, M. 2014. Analysis of economic efficiency and farm size: a case study of wheat farmers in Nakuru District, Kenya. Hindawi Publishing Corporation Economics Research International. doi:10.1155/2014/802706. (Acceded June 2016) Meeusen, J. and
Broeck, V. 1977. Efficiency estimation from Cobb Douglas production
Functions with composed error. International Economic Review, 18(2): 435-444. Mekonnen Daniel, Spielman, D., Fonsah, E., and Dorfman, J. 2015. Innovation systems and technical efficiency in developing-country agriculture. Journal of Agricultural Economics, 46(5): 689-702. Mesay Yami, Tesafye Solomon, Bedada Begna, Fekadu Fufa, Tolesa Alemu and Dawit Alemu. 2013. Source of technical inefficiency of smallholder wheat farmers in selected waterlogged areas of Ethiopia. African Journal of Agricultural Research, 8(29): 39303940. Michael, L.M. and James, M.W. 2017. Ethiopian wheat yield and yield gap estimation: A spatial explicit small area integrated data approach. Field Crop Research, 201(2017): 60-74. Mukete, N., Zhu, J., Backline, M. and Gilbert, T. 2016. Analysis of the technical efficiency of smallholder cocoa farmers in southern west Cameroon. American Journal of Rural Development, 4(6): 129-133. Musa Ahmed, Lemma Zemedu and Endrias Geta. 2014. Technical efficiency of maize producing farmers in Arsi Negelle, Central rift valley of Ethiopia. Journal of Agriculture and Forestry, 60(1):157-167. Musab, E. and Bwacha, I. 2014. Technical efficiency of small scale maize production in Masaiti district, Zambia. Journal of Economics and sustainable development, 5(4): 2222-2855. Mwajombe, K. and Mlozi, S. 2015. Measuring farm-level technical efficiency of urban agriculture in Tanzanian Towns. World Journal of Social Science, 2(1):62 National Bank of Ethiopia. 2014. Annual Report 2013-2014. Addis Ababa, Ethiopia. Ouedraogom, S. 2015. Technical and economic efficiency of rice production in the Kou Valley (Burkinafaso). Asian Journal of Agriculture and Rural Development, 5(2): 53-63. Ogada, J., Dianah, M., Germano, M. and Mary, M. 2014. Technical efficiency of Kenya’s smallholder food crop farmers. Environment, Development and Sustainability, 16(5): 1065-1076. Palmer, S. and Torgerson, D.J. 1999. Definitions of efficiency. [Internet].318(7191): 1136. Available from doi: 10.1136/bmj.318.7191.1136. [Accessed 09 November 2016].
72 Ray, S.C. 1985. Measurement and test of efficiency of farms in linear programming model: a study of West Bengal farms. Oxford Bulletin of Economics and Statistics, 47(4): 371386. Schmidt, P. 1976. On the statistical estimation of parametric frontier production functions. Review of Econometrics and Statistics, 58(2): 238–239. Shafiq, M. and Rehman, T. 2000. The extent of resource use inefficiencies in cotton production in Pakistan Punjab: An Application of Data Envelopment Analysis. Agricultural Economics Journal, 22(3): 321-330. Shehu, J.F., Iyortyer, J.T., Mshelia, S.I. and Jongur, A.U. 2010. Determinants of yam production and technical efficiency among yam farmers in Benue State, Nigeria. Journal of Social Science, 24(2): 143-148. Shumet Assefa. 2012. Who is technically efficient in crop production in Tigray Region, Ethiopia: Stochastic frontier approach. Global Advanced Research Journal of Agriculture science, 1(7): 2315-5094. Sisay Debebe, Jema Haji, Degye Goshu and Abdi Endris. 2015. Technical, allocative, and economic efficiency among smallholder maize farmers in Southwestern Ethiopia: Parametric approach. Journal of Development and Agricultural Economics, 7(8): 282291. Solomon Bizuayehu. 2012. Economic efficiency of wheat seed production: the case of smallholders in Womberma Woreda of West Gojjam Zone. Msc Thesis, Haramaya University, Haramaya, Ethiopia. Solomon Bizuayehu. 2014. Technical efficiency of major crops in Ethiopia: Stochastic frontier model. Academic Journal of Agricultural Resource, 2(6): 147-153. Stevenson, R. E. 1980. Likelihood functions for generalized stochastic frontier estimation. Journal of Econometrics, 13(1): 57-66. Strock, H., Bezabih Emana, Berhanu Adnew, A. Borowiecki and Shimelis W/Hawariat. 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. Sultan, S. and Ahmed, S. 2014. Technical efficiency and farm size productivity: Micro level evidence from Jmmu and Kashmir. International Journal of Food and Agricultural Economics, 2(4):27-49.
73 Tefera Kebede, Gebremeskel Berhane and Menasbo Gebru. 2014. Technical efficiency in teff production by small scale farmers in Tigray. International journal of research, 4(10): 85-97 UN (United Nations). 2014. World Urbanization Prospects: The 2014 Revision, CD-ROM Edition. Department of Economic and Social Affairs, Washington D.C., USA. WOA (Woreda Office of Agriculture). 2012. Annual report of crop production in 2010/2011 cropping season. Unpublished. WOA (Woreda Office of Agriculture). 2015. Annual report of crop production in 2013/2014 cropping season. Unpublished. Wondimu Tesfaye and Hassen Beshir. 2014. Determinants of technical efficiency in maize production: the case of smallholder farmers in Dhidhessa District of Illuababora Zone, Ethiopia. Journal of Economics and Sustainable Development, 5(12): 274-284. Wondimu Tesfaye, Hassen Beshir and S.P.R.Chaurasia. 2014. Determinants of technical efficiency in maize production: the case of smallholder farmers in Dhidhessa District, Illubabor Zone, Ethiopia. Published at proceeding of the 11th international conference of the Ethiopian economics Association, July 2014. Worku Ibrahim, Minten Bart and Alemayehu Seyoum. 2014. Patterns and determinants of teff consumption. In The Economics of Teff Educations International Food Policy Research Institute/Ethiopia Strategy Support Program II. Addis Ababa, Ethiopia. Wudineh Getahun and Endrias Geta. 2016. Technical efficiency of smallholder wheat farmers: The case of Welmera District. Journal of Development and Agricultural Economics, 8(2): 39-51. Yamane, T. I. 1967. Statistics: An Introductory Analysis 2nd Edition. Harper and Row, New York, USA.
74
7. APPENDICES
75
7.1. Appendix I. List of Tables in the Appendices Appendix Table 1. Conversion factors used to compute Adult equivalent and man equivalent Age Sex ADE ME 0-7 Male 0.6 0.0 Female 0.6 0.0 8-10 Male 0.8 0.4 Female 0.8 0.4 11-15 Male 0.9 0.5 Female 0.9 0.5 16-24 Male 1.0 1.0 Female 0.8 0.8 25-50 Male 1.0 1.0 Female 0.8 0.8 51-65 Male 0.9 0.9 Female 0.7 0.7 >65 Male 0.8 0.7 Female 0.6 0.5 Source: Storck et al. (1991 as cited in Wondimu et al., 2014) Appendix Table 2. Conversion factors used to estimate Tropical Livestock Unit (TLU) Type of domestic Animal Conversion factor Cow/Ox 1.0 Young bull 0.80 Heifer 0.75 Calf 0.20 Weaned calf 0.34 Horse/Mule 1.10 Sheep/Goat adult 0.13 Sheep/Goat Young 0.06 Donkey-adult 0.70 Donkey-young 0.35 Chicken 0.013 Source: Stork et al. (1991 as cited in wondimu et al., 2014) Appendix Table 3.Variance inflation factor for the input variables entered in to the SPF model Variable 1/VIF VIF lnArea 0.301205 3.32 lnFertilizer 0.373134 2.68 lnOxen 0.826446 1.21 lnLabor 0.869565 1.15 lnSeed 0.884956 0.89 Mean VIF 1.90 Source: Own Survey, 2016
76 Appendix Table 4. Variance inflation factor for the inefficiency variables Variable 1/VIF VIF Land fragmentation 0.451364 2.22 Farm size 0.460933 2.17 Education status 0.508900 1.97 Credit 0.513816 1.95 Improved seed 0.548820 1.82 Distance from residence 0.508900 1.97 family size 0.802244 1.25 Off farm activity 0.821376 1.22 TLU 0.848080 1.18 Sex 0.861973 1.16 Age 0.897705 1.11 Extension service 0.927651 1.08 Training 0.954399 1.05 Mean VIF 1.50 Source: Owen Survey, 2016 Appendix Table 5. Contingency coefficient of dummy variables Sex of the Training Off farm Credit household activity Sex of the household 1.0000 Training -0.0318 Off farm activity -0.0053 Credit 0.1937 Improved seed -0.0389 Source: own survey, 2016
1.0000 0.0165 0.0865 0.067
1.0000 0.1717 0.2714
1.0000 0.4741
Improved seed
1.0000
Appendix Table 6. The VIF for the continuous variables used in inefficiency variables Variable 1/VIF VIF Farm size 0.505173 1.98 Land fragmentation 0.529438 1.89 Education 0.809177 1.24 Distance from residence 0.829124 1.21 Family size 0.860247 1.16 TLU Age of the household head Frequency of extension contact Mean VIF Source: own Survey 2016
0.861105 0.936023 0.975163
1.16 1.07 1.03 1.34
77 Appendix Table 7. Educational status of the sampled farmers Education level Frequency Illiterate 34 Basic education 40 Grade 1-8 54 Grade 9-12 21 Total 149 Source: Owen Survey, 2016
Percent 22.76 26.83 36.59 13.82 100
Appendix Table 8. Reasons of sample farmers for not using credit facilities Descriptions Frequency Percent High interest rate 14 35 Lack of collateral 10 25 Having enough working capital 7 17.5 Repayment schedule 5 12.5 Lack of lender 4 10 Total 40 100 Source: Owen Survey, 2016 Appendix Table 9.Technical efficiency estimate of sample farmers ID TE ID TE ID TE ID TE ID 1 0.59 26 0.87 51 0.68 76 0.93 101 2 0.47 27 0.73 52 0.62 77 0.53 102 3 0.76 28 0.74 53 0.77 78 0.69 103 4 0.67 29 0.72 54 0.89 79 0.90 104 5 0.62 30 0.94 55 0.82 80 0.74 105 6 0.84 31 0.83 56 0.72 81 0.92 106 7 0.68 32 0.72 57 0.77 82 0.86 107 8 0.60 33 0.61 58 0.73 83 0.72 108 9 0.74 34 0.74 59 0.94 84 0.62 109 10 0.89 35 0.70 60 0.74 85 0.81 110 11 0.76 36 0.74 61 0.69 86 0.94 111 12 0.68 37 0.95 62 0.85 87 0.93 112 13 0.86 38 0.72 63 0.15 88 0.82 113 14 0.91 39 0.50 64 0.62 89 0.67 114 15 0.73 40 0.57 65 0.92 90 0.79 115 16 0.74 41 0.84 66 0.36 91 0.67 116 17 0.93 42 0.85 67 0.78 92 0.95 117 18 0.87 43 0.79 68 0.63 93 0.86 118 19 0.67 44 0.78 69 0.85 94 0.82 119 20 0.81 45 0.63 70 0.94 95 0.92 120 21 0.71 46 0.86 71 0.58 96 0.70 121 22 0.85 47 0.87 72 0.72 97 0.95 122 23 0.79 48 0.64 73 0.73 98 0.75 123 24 0.68 49 0.69 74 0.74 99 0.87 124 25 0.92 50 0.68 75 0.85 100 0.93 125 Source: Owen Survey, 2016
TE 0.78 0.68 0.80 0.78 0.94 0.71 0.93 0.94 0.94 0.80 0.87 0.93 0.64 0.65 0.89 0.37 0.80 0.66 0.95 0.83 0.72 0.74 0.81 0.62 0.34
ID 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149
TE 0.67 0.62 0.84 0.68 0.60 0.74 0.89 0.74 0.92 0.86 0.73 0.94 0.74 0.69 0.85 0.15 0.62 0.73 0.94 0.74 0.69 0.85 0.15 0.63
78
Appendix Table 10. General characteristics of sample household heads Variables Number Minimum Maximum Mean Age 149 23 80 44.43 Total household size 149 2 10 5.39 Adult equivalent 149 1.7 8.9 4.31 Man equivalent 149 1.7 8.9 3.75 Academic grade 149 0 12 3.50
Std. Deviation 11.62 1.92 1.49 1.32 3.79
Source: Owen survey, 2016 Appendix Table 11. The frequency distribution of farmers by technical efficiency scores TE Range Frequency Percent 0.15-0.65 23 15.45 0.66-0.75 49 32.52 0.76-0.85 31 21.13 0.860-0.90 19 13.01 Above 0. 90 27 17.89 Source: Owen survey, 2016 Appendix Table 12. Distribution of sample farm household heads per Kebeles Name of KAs Total HH head Sample HH head No. of male No. of female Total Male Female Total Number Gomata (017) 632 117 749 33 6 39 Elshama (018) 847 176 1023 44 9 53 Yeddo (019) 923 169 1092 48 9 57 Total 2402 462 2864 125 24 149 Personnel communication with KAS, 2016
% 26.17 35.57 38.26 100
Appendix Table 13. Hetroskedasticity Test Breusch-Pagan or Cook Weisberg test for hetroskedasticity. Ho: Constant variance Variables: fitted values of lnOutput Chi2 (01) = 0.36 Prob>Chi2=0.5535 Decision is to Accept Ho.
7.2. Appendix II. Farm Households’ Interview Schedule Questionnaire developed for Farm Households survey Questionnaire for technical efficiency in Teff production on smallholder farmers in Jamma district. Name of the enumerator_____________________________Signature ________________ Date_______________ Respondent Identity Number_______________
79 PART I. General Information about Sample Farmers 1. Name of the Respondent: _______________________ 2.
Age of the respondent: _______ years
3.
Sex of the respondent:
1. Male
4. Education level of the respondent:
2. Female 1. Illiterate
2. Read and write If attend school,
grade____ 3. Certificate
4. Diploma 5. Degree
5. Marital status: 1. Married 2. Single 3. Divorce 4. Widowed 6. Respondent’s religion: 1. Muslim
2. Christianity:
7. Ethnic category: 1. Oromo 2. Amhara: 3. Tigray 8.
Main occupation 1. Farmer
(3): Other:___________ 4. Other (specify): ________
2. Non-farming: 3 Other: Specify:____________
9. Farming experience ___________years 10. What is the relative wealth position of the farmer? (categorized by peer groups) 1. Very rich
2. Rich
3. Medium
4. Poor
11. Type of the house? 1. Corrugated iron 2. Grass
5. Very poor
3. both
12. If corrugated iron, number of corrugated iron___________________ 13. Do you have your own transportation facilities? (√)
1. Yes
14. If yes, what type?
4. other
1. Mule
2. Horse 3. Camel
15. Household member’s characteristics : [___ ] Male [___] Female S.No
Name of the Age family member years
in Rela Sex1 tives
2. No
[____] Total
Level of Occupation 3 Education 2
Marital Status 4
1 2 3 4 5 6 7 8 9 10 11 12 1.Sex 1=male 2= female 2. Education 0= illiterate 1= read and write grade for the rest number of schooling, 3. Occupation 1= agriculture 2= student 3= non-agriculture, 4. Marital status 1=single 2=married 3=divorced 4=widowed 5. Health status 1=healthy 2= not healthy.
80 PART II. Economic Information 1.
Total area cultivated by your family ______timad__________ha
2. Number of plot owned by your family ________ 3. Number of location of your plot owned by your family_____________ No. of Distance from Total area of teff plot home (km) cultivated plot(ha)
Total yield of each plot (Qt)
Type of ownership1
Fertility status of the plot2
1 2 3 4 5 6 7 1. Ownership 1= owned 2= rented in 3= rented out 4= shared in 5= shared out 2. Fertility (farmer’s perception): 1= low 2= Medium 3= High fertile 3. Input use and their produce per crop category in 2015/16. S/N
No. of Area plots
Seed Variety name
Teff
Wheat Barely
Plot 1 Plot 2 Plot 3 Plot 4 Plot 5 Plot 1 Plot 2 Plot 1 Plot 2
Amoun t(kg)
Amount of fertilizer used (kg) UREA DAP
Total Yield in Amount sold Qt
81
4. Human labor and oxen power input used in teff production per plot in 2015/16 ctivities
Pair of Participants of family labor Oxen No Hr. Children Men Women . No. Hr. No. Hr. No. Hr.
Number of hired labor No.
Hr.
First Plowing Second plowing Third Plowing Fourth Plowing Fifth Plowing Sowing Top Dressing First weeding Second Weeding Third Weeding Harvesting Threshing Others 5. Livestock production 5.1. Do you own livestock?
1. Yes
2. No
5.2. If yes, what are the current number of livestock number and estimated price? No
Type of livestock Total number Cow Heifer Oxen Bull Calves Sheep Goat Donkey Mule Horse
Estimated price
unit Major Problem Remark *
1 2 3 4 5 6 7 8 9 10 11 *1. Feed shortage 2. Disease 3. Water shortage 4. Improved breed 6. Of/non farm income 6.1. Do you have off farm income source?
1. Yes
2. No
82 6.2. Which one of the following is the source of your income off farm or non farm in come 1. Off-farm activity Salary/wage
2. Pension payments
5. Rent from asset
3. Transfer payment
4.
6. Other specify
6.3. Annual income from off/nonfarm activity?_____________________ 6.4. Are some of your family engaged in off/non-farm activities? 1. Yes
2. No
If yes why? 1. Shortage of land
2.Excess family labor
3.Attractive income from off-farm
activities 4. Other, specify_______________________________________ PART III: General Information about the teff farming 1. Do you have own cultivated land for Teff?
1. Yes
2. No
2. If yes, Total land holding: ___________ timad __________ ha,
for teff only
_________timad __________ ha. (Note: 1 ha = 4 timad use local conversion rate if it is different and specify it) 3. Number of location of teff plots_________________________________ 4. At what time did you start land preparation? _______________________month. 5.
How many times do you plough your teff land? ____________________________
6.
At what time did you sow teff? ________________month.
7. How many pairs of oxen did you apply from land preparation up to planting? _________ 8.
Is weeding teff crop a common practice?
1. Yes
2. No
9.
If yes, when you start weeding teff? _____________Week of _________________month
10. How many times do you weed? ________________________times. 11. What method do you use for weeding? 1. Hand weeding
2. Hoeing 3. I use chemicals
4. Others, specify_____________________ 12. How long have you practiced production of teff products? __________________Years 13. When did you harvest your teff crop? ________________month 14. How many quintals of teff yield did you received this year? ___________________Qt 15. How is the trend of volume of teff crops production during the past 5 years? (√) Crop type Teff Other Cereals Others
Increasing
Decreasing
Same
16. If the production of teff increases, what are the reasons? _____________ 17. If the production of teff decreases, what are the reasons? ________________
83 18. Would you like to expand teff production?
1. Yes
2. No
19. What opportunities exist to expand teff production? ___________________ 20. What are the production constraints on your teff farm? You can select more than one. 1. Availability of quality improved seed 2. Sever weed infestation 3.
Size of land holding
4.
Damage of pests
5.
Unfavorable weather condition
6.
others
PART IV. Fertilizer and Improved seed
1. Do you use organic and chemical fertilizer in teff field?
1. Yes
2.
If yes, what type of fertilizer? 1. DAP 2. Urea
4. Other, specify_______
3.
When did you use fertilizer for the first time? ___________________Year.
4.
How many Kg of each per hectare do you use for crops?________________________ Crop type
3. Both
2. No
DAP in kg Urea in kg Manure in Kg Compost in Kg 2015 production 2015 production 2015 production 2015 production year year year year
Teff Wheat Barley Bean 5.
What is the application method of fertilizer? 1. Broadcasting
6.
Do you get fertilizer on time?
1. Yes
2. Basal application 2. No .
7. According to your view what are the advantages of fertilizer? Increase straw yield
1. Increase grain yield
3. Improve the quality of the crop
2.
4. Other (specify)
_____________________ 8. Do you think fertilizers have disadvantages? 9. If yes, what are the disadvantages?
1. Yes
2. No
1. Damages/burns the crop
2. Favors weed growth
3. Other (specify) _____________________________________ 10. If you do not use fertilizer, why? Not
timely
available
Other___________________
1. Too expensive
4.Not
available
5.
2. Inconvenient to transport Not
good
to
apply
on
3. 6.
84 11. What are the major local varieties of teff you are growing?_____________ 12. Do you use improved teff varieties in a year 2015?
1. Yes
2. No
13. If yes, name of variety and Source___________________________ 14. If no, why?
1. Too expensive
2. Not better than local varieties
3. It is dwarf
4. It is not easily accessible 5. Other __________________________________ 15. Do you get improved seed on time?
1. Yes
2. No
16. How far do you travel to buy improved seed? ____________Km __________hrs.
PART V: Extension service and training 1. Extension service 1.1.. Do you have contact with development workers?
1. Yes
2.No
1.2. When did you start getting extension service? ________year 1.3.Did you ever received any pieces of advises with regard teff Production? 1. Yes 2. No If yes, for ______________days or_______ months. 1.4.How many days did you visited the extension agent (starting from land preparation until harvesting of crop)? ______________________ Days 1.5.How often do the development workers visit your farm starting from land preparation until harvesting of crops?
1. Daily
5. They didn`t visit at all
2. Weekly
3. Monthly
4. Quarterly
6. Other, specify______________________________
2. Training
a. Have you ever attended a field day or demonstration trial on teff?
1. Yes
2. No
b. Have you ever attended training related to crop production in the last four years? 1. Yes
2. No
c. If yes, the type of training ________________________Duration_________________ PART VІ: Credit Service 1. Do you have credit access
1. Yes
.2. Have you borrowed money over the last two years
2. No 1. Yes
2. No
85 If yes state amount and purpose of credit for the last two years *Source Credit
of Amount credit
of **Purpose credit
of Distance of credit Proportion source Minute or repayment Km
of
* 1 .Amhara credit and saving institution 2. Local money lenders 3. Relatives 4. Cooperatives 5. Traders **Purpose of credit 1. DAP and UREA 2. Seed purchase 3. To send students to school 4. Home consumption 5. Others 3. Of the total amount you borrowed over the last two years have paid all as per the promise? 1. Yes
2. No, if it is not what the reason for delay is? ___________________
PART VІIІ: Focus group discussion 1. Major problems encountered in Teff farming 2. Trend of crop production system 3. Input utilization and input supply system 4. risk bearing ability of the household
Thank you for sharing your experience!!!
