Economic Analysis of Sustainable Coffee Production in Vietnam
Thong Quoc Ho
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
School of Economics and Finance
Faculty of Business
Queensland University of Technology
Principal supervisor: Dr Viet-Ngu (Vincent) Hoang
Associate supervisor: Prof. Clevo Wilson
2018
Statement of original authorship
The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.
Signature: QUT verified signature
Date: June 2018
______Economic Analysis of Sustainable Coffee Production in Vietnam i
Keywords
Coffee, cost, data envelopment analysis (DEA), efficiency, environment, farming, irrigation water, material balance principles (MBP), meta-frontier, meta-technology ratio
(MTR), nutrient, stochastic frontier analysis (SFA), Vietnam.
______Economic Analysis of Sustainable Coffee Production in Vietnam ii
Abstract
Vietnam is the second largest coffee producer in the world. The little literature reports that coffee farming in Vietnam has been increasingly economically and environmentally unsustainable. More specially, few studies documented that the consumption of chemical fertilisers and irrigation in Vietnam coffee farms are much greater than the average levels in other major coffee producing countries. Environmental regulations in Vietnam appear to be not sufficient while voluntary participations in sustainability certification programs become increasing popular. However, there is little empirical research that examines both economic and environmental performance of coffee farms in Vietnam. This thesis aims to enhance the understanding of the economic and environmental sustainability of Vietnam’s coffee farming sector.
The author has conducted surveys to collect data from 1,063 farms. Apart from coffee industry background and socio-economic profile of coffee farmers participated in the main survey, this thesis includes four empirical studies that assess a variety of the economic and environmental performance of the Vietnam’s coffee farming sector using the collected data. The first study investigates the productive efficiency of different farming systems, using a multiple output input stochastic distance function. The major findings show that combination of coffee and other industrial crops deliver the highest efficiency levels. The second study examines eco-efficiency and its variation across sustainability certified and non-certified farms. The empirical results of the second study indicate great potential improvement in the eco-efficiency level across coffee farms.
Results also show that although certified farms are likely to perform better than non-
______Economic Analysis of Sustainable Coffee Production in Vietnam iii
certified counterparts, the difference between these two groups appears to converge overtime. The third study provides an examination of potential trade-off between cost and nutrient efficiency. An elaborative material balance principles (MBP) based efficiency is proposed to decompose the production technology set into four distinct groups of sampled farms. Each group of farms should pursue differing strategies to improve cost and environmental performance. The findings also suggest that not all farms necessarily are faced trade-offs between cost and nutrient efficiency. Certified farms are generally more environmentally and cost-efficient than non-certified ones and this difference disappears overtime. The last study focuses on irrigation water efficiency using a sub-vector DEA meta-frontier technique. The findings suggest a significant potential to improve water efficiency and, advanced technologies are likely to have a much larger impact on the irrigation water efficiency than voluntary participation in sustainability certification schemes. Policy implications drawn from these studies are discussed in the thesis.
______Economic Analysis of Sustainable Coffee Production in Vietnam iv
Table of Contents
Statement of original authorship ...... i Keywords ...... ii Abstract ...... iii Table of Contents ...... v List of tables ...... x List of figures ...... xi List of appendices ...... xi List of abbriviations ...... xii List of publications and presentations ...... xiii Acknowledgements ...... xiv
Introduction ...... 1
1.1 Background ...... 1 1.2 Research problems: unsustainable coffee production in Vietnam ...... 3 1.3 Research questions ...... 6 1.4 Specific research objectives ...... 11 1.5 Methodologies ...... 11 1.6 Contributions ...... 12 1.7 Thesis outline ...... 15
Industry background ...... 20
2.1 Introduction ...... 20 2.2 Coffee production in Vietnam ...... 20
2.2.1 Overview of coffee production in Vietnam ...... 20 2.2.2 Primary Vietnamese coffee farming characteristics ...... 21
2.3 Sustainability certification programs in Vietnam ...... 22 2.4 Policy context ...... 24 2.5 Overview of studies on coffee production ...... 26
______Economic Analysis of Sustainable Coffee Production in Vietnam v
Study design ...... 29
3.1 Introduction ...... 29 3.2 Hypothesis and overall objective ...... 29 3.3 Research aims ...... 30 3.4 Study site, data and ethics ...... 31
3.4.1 Study site ...... 31 3.4.2 Data and ethics ...... 32
3.5 Methodological reviews ...... 33
3.5.1 Overview of efficiency analysis ...... 33 3.5.2 Methodological review of environmental efficiency ...... 34 3.5.3 Indicator approach ...... 35 3.5.4 Frontier- based approach ...... 37 3.5.5 Materials balance-based approach ...... 38 3.5.6 Nutrient efficiency ...... 39 3.5.7 Irrigation water efficiency ...... 40
Socio-economic industry profile ...... 43
4.1 Introduction ...... 43 4.2 Demographic characteristics ...... 44 4.3 Economic profile of coffee farming ...... 48 4.4 Conclusion ...... 60
Diversification and Productive Efficiency ...... 62
5.1 Introduction ...... 62 5.2 Literature review ...... 65 5.3 Input distance function ...... 69 5.4 Empirical model specifications ...... 72 5.5 Farming systems and data collection ...... 75 5.6 Empirical results ...... 78
5.6.1 Maximum likelihood estimates ...... 78 5.6.2 Inefficiency model ...... 80 ______Economic Analysis of Sustainable Coffee Production in Vietnam vi
5.6.3 Relative efficiency levels of differing farming systems ...... 81
5.7 Discussions ...... 84
5.7.1 Evidence of agronomic benefits between coffee and other industrial crops...... 84 5.7.2 Why coffee farmers still choose rice ...... 86
5.8 Conclusions ...... 87
Eco-efficiency analysis of coffee farming ...... 89
6.1 Introduction ...... 89 6.2 Literature review ...... 93 6.3 Methodology ...... 97
6.3.1 First stage: Frontier eco-efficiency ...... 97 6.3.2 Second stage: Fractional regression models (FRMs) ...... 100
6.4 Empirical strategies ...... 102
6.4.1 Data descriptions ...... 102 6.4.2 Variable selection ...... 102 6.4.3 Weighted number of coffee trees ...... 105
6.5 Results ...... 106
6.5.1 Comparison of eco-efficiency between certified farms and non-certified farms ...... 106 6.5.2 Eco-inefficiency variance analysis ...... 112
6.6 Conclusions ...... 116
Trade-off Analysis between cost and nutrient efficiency ...... 119
7.1 Introduction ...... 119 7.2 The MBP approach to cost and environmental efficiency measure ...... 122 7.3 Existing approach to analysing a trade-off between cost and nutrient efficiency ...... 123 7.4 Elaborative approach of MBP cost and environmental efficiency ...... 127 7.5 Empirical strategies ...... 129
7.5.1 Study site and data ...... 129 7.5.2 Model specifications ...... 131 ______Economic Analysis of Sustainable Coffee Production in Vietnam vii
7.5.3 Efficiency results ...... 134 7.5.4 Analysis of cost savings and reduction in nutrient consumption ...... 136 7.5.5 Trade-off between economic and environmental performance analysis ...... 139 7.5.6 The role of the sustainability certification program ...... 143
7.6 Conclusions and policy implications ...... 146
Irrigation water efficiency ...... 149
8.1 Introduction ...... 149 8.2 Irrigation water and certification in Vietnam coffee farming ...... 153 8.3 Empirical strategies ...... 155
8.3.1 Sub-vector DEA meta-frontier models ...... 155 8.3.2 Irrigation water-oriented meta-technology ratio (MTR) ...... 158 8.3.3 Efficiency econometric models ...... 159 8.3.4 Data and variables ...... 160
8.4 Result and discussion ...... 163
8.4.1 Efficiency and irrigation meta-technology ratio...... 163 8.4.2 Variation in irrigation water efficiency and sustainability certification...... 168
8.5 Conclusions and policy implications ...... 173
Syntheses, conclusion and policy options ...... 176
9.1 Introduction ...... 176 9.2 Key findings ...... 177
9.2.1 Efficient farming systems ...... 177 9.2.2 Eco-efficiency of Vietnam’s coffee farming sector ...... 178 9.2.3 The nexus between cost and environmental efficiency ...... 179 9.2.4 Irrigation efficiency ...... 181
9.3 Management and policy implications ...... 181
9.3.1 The choice of farming systems by coffee farmers ...... 182 9.3.2 Improving quality rather than rapid expansion of certification schemes ...... 183 9.3.3 Promoting efficient use of material inputs ...... 184 ______Economic Analysis of Sustainable Coffee Production in Vietnam viii
9.3.4 Promoting the use of advanced irrigation technologies ...... 185
9.4 Limitations of the thesis...... 186 9.5 Further research ...... 187 9.6 Conclusion ...... 188
Bibliography ...... 191 Appendix ...... 211 Appendix 1: Derivation of output complementary effect ...... 211 Appendix 2: Weighted number of coffee trees ...... 213 Appendix 3: Mean efficiency of farms dropped out and joined in certification programs ...... 213 Appendix 4: Ethics approval and survey ...... 214
______Economic Analysis of Sustainable Coffee Production in Vietnam ix
List of tables
Table 4.1: Demographic characteristics of surveyed coffee farmers ...... 45 Table 4.2: Descriptive statistics of economic profile (measured in million VND per ha) ...... 49 Table 4.3: Economic profile of coffee production over crop years ...... 51 Table 4.4: Economic profile of non-certified vs certified farms ...... 57 Table 5.1: Descriptive statistics of variables ...... 77 Table 5.2: MLE for the stochastic input distance function ...... 79 Table 5.3: Inefficiency model ...... 81 Table 5.4: Complementary effects and diversification efficiencies ...... 82 Table 5.5: Characteristics of rice versus non-rice farms and coffee-rice versus other farms ...... 86 Table 6.1: Recent studies on the effects of certification in coffee production ...... 93 Table 6.2: Non-certified and certified farm categories ...... 102 Table 6.3: Descriptive statistics ...... 103 Table 6.4: Mean eco-efficiency levels ...... 107 Table 6.5: Tests for equality of eco-efficiency distribution ...... 110 Table 6.6: Factors affecting eco-efficiency ...... 113 Table 7.1: Descriptive statistics of variables ...... 130 Table 7.2: Cost and environmental efficiency measures ...... 134 Table 7.3: Cost and environmental performance ...... 137 Table 7.4: Changes in cost and nutrient usage by different groups of farms ...... 140 Table 7.5: Efficiency between certified and non-certified groups ...... 144 Table 7.6: Efficiency between certified and non-certified farms over time ...... 145 Table 8.1: Descriptive statistics of production factors per weighted tree ...... 162 Table 8.2: Distribution of efficiency scores and irrigation MTR ...... 165 Table 8.3: Irrigation water efficiency random effect OLS models ...... 170
______Economic Analysis of Sustainable Coffee Production in Vietnam x
List of figures
Figure 1.1 Coffee production by major producing countries (ICO, 2017a) ...... 1 Figure 1.2: Thesis outline ...... 19 Figure 4.1: Distribution of household head ages ...... 46 Figure 4.2: Family size distribution ...... 47 Figure 4.3: Education levels of the household heads ...... 48 Figure 4.4: Direct production cost per ha in million VND (excluding family labour cost) ...... 53 Figure 4.5: Irrigation water per ha (measured in m3) ...... 59 Figure 5.1 Three distinct farming systems...... 75 Figure 6.1: Kernel estimated densities of farm eco-efficiency scores ...... 109 Figure 8.1 Irrigation water meta-frontier efficiency ...... 157 Figure 8.2: Irrigation MTR distribution: sprinkle vs micro-basin ...... 168 Figure 8.3: Meta-technology ratio to the meta-frontier ...... 172
List of appendices
Appendix 1: Derivation of output complementary effect ...... 211 Appendix 2: Weighted number of coffee trees ...... 213 Appendix 3: Mean efficiency of farms dropped out and joined in certification programs ...... 213 Appendix 4: Ethics approval and survey ...... 214
______Economic Analysis of Sustainable Coffee Production in Vietnam xi
List of abbriviations
CAE Cost Allocative Efficiency CE Cost Efficiency CRS Constant Return to Scale DGP Data Generating Process DMU Decision Making Unit FEM Frontier-based Eco-efficiency Model FRM Fractional Regression Model IWE Irrigation Water Efficiency MBM material balance-based model MTR Meta-technology Ratio NAE Nutrient Allocative Efficiency NE Nutrient Efficiency OLS Ordinary Least Squares QUT Queensland University of Technology SD Standard Deviation TE Technical Efficiency VRS Variable Return to Scale
______Economic Analysis of Sustainable Coffee Production in Vietnam xii
List of publications and presentations
Publications:
Ho T.Q., Hoang V., Wilson C., Eco-efficiency analysis of sustainability-certified coffee production: Evidence from Vietnam. Journal of Cleaner Production (2018), https://doi.org/10.1016/j.jclepro.2018.02.147. (Chapter 6).
Ho T.Q., Hoang V., Wilson C., Nguyen T., Which farming systems are efficient for Vietnamese coffee farmers? Economic Analysis and Policy (2017), http://dx.doi.org/10.1016/j.eap.2017.09.002 (Chapter 5).
Ho T. Q., Measuring Environmental Sustainability of Coffee Production Using Econometric and Frontier-Based Models: Vietnam as a Case Study. EEPSEA Research Report No. 2017-RR6. Economy and Environment Program for Southeast Asia, Laguna, Philippines. (Chapters 6, 7 and 8).
Presentations:
Ho T.Q., Hoang V., Wilson C., The nexus between cost and environmental performance of coffee production in Vietnam. 14th Western Economic Association International (WEAI) (2018). (Chapter 7).
Ho T.Q., Hoang V., Wilson C., Eco-efficiency analysis of sustainability-certified coffee production: Evidence from Vietnam. 61st Annual Conference of the Australian Agricultural and Resource Economics Society (2017). (Chapter 6).
Ho T.Q., Hoang V., Wilson C., The nexus between cost and environmental performance of coffee production in Vietnam. School of Economics and Finance, Queensland University of Technology (Brownbag Seminar) (2017). (Chapter 7).
______Economic Analysis of Sustainable Coffee Production in Vietnam xiii
Acknowledgements
Since I enrolled in my PhD program at the Queensland University of Technology
(QUT), I have received great assistance and guidance from individuals and institutions.
First, I would like to express sincere thanks to my supervisors, Dr. Viet-Ngu Hoang and
Prof. Clevo Wilson, who have continuously guided and encouraged me throughout my
PhD journey. I am very grateful to them for the opportunity to learn from their experiences and for helping me to develop my research skills.
I would like to express my great gratitude to Prof. Sean Pascoe and Dr. Son
Nghiem who served as my confirmation panel members for their helpful comments and suggestions. Their invaluable advice on my proposal kept track of the studies presented in this thesis.
The school of Economics and Finance (QUT) has provided great academic support and excellent learning and working environment for which I am very appreciative.
I gratefully acknowledge the Australia Awards Scholarship for providing financial support for my PhD program. I am also thankful for the financial support I received from the Economy and Environment Program for Southeast Asia (EEPSEA) under the research grant number PCO15-0724-011, for field research and data collection.
Much appreciation is extended to data collection team members including Hoang
Anh Vu, Lang Van Trinh, Nguyen Trong Su, Huynh Trung Cao, Mai Thi Thoa, Dinh
Thi Anh Nguyet and Doan Thi Hong Ha.
______Economic Analysis of Sustainable Coffee Production in Vietnam xiv
I am most thankful to the coffee farmers in Dak Lak, Lam Dong and Gia Lai provinces for participating in a survey that conducted in 2015, for sharing their production data and insight knowledge about coffee farming with me.
I am grateful to my family and friends. I am especially thankful to my wife, Thu
Hoang who have exhibited the supreme virtue of patience and understanding. I am most grateful to my son Quan Ho, and my little daughter, Linh-Dan Ho, born during my PhD journey, for offering me joys and great happiness every day.
Special thanks go to my dearest Australian family, Kim Rivera, for their support, caring and friendship that they have silently offered during my journey.
I am indebted to more people than I can name. Thank you all for being present in this unforgettable stage of my life.
______Economic Analysis of Sustainable Coffee Production in Vietnam xv
Introduction
1.1 Background
There are two primary types of coffee, Arabica (C. arabica) and Robusta (C. canephora), both with a long development history. It is believed that Arabica coffee was first discovered in the highland forests of Ethiopia (Murthy and Madhava Naidu, 2012).
The first semi-cultivation of Robusta was on the islands of Lake Victoria and in the north of Kagera, Tanzania (Waller, Bigger, & Hillocks, 2007, p. 6). Today, coffee is cultivated in more than 60 tropical countries most of which are developing countries and mainly carried out by smallholder farmers. In contrast, the consumption of coffee beverage is largely in developed countries. Coffee production is a billion-dollar industry for the major developing country producers Brazil, Vietnam, Columbia and Indonesia.
Figure 1.1 Coffee production by major producing countries (ICO, 2017a)
______Chapter 1: Introduction
In recent years, coffee has played an important role for many developing economies. Since 2012 annual total global coffee production has been approximately
150 million bags (60-kilograms). Approximately 78% was exported (ICO, 2017a) and was valued around US $20 billion. In the past the value of exports over the period 1997-
2005 has ranged from about US $5 to $12 billion, (Mohan, 2007). As shown in Figure
1.1, overall, coffee production has dramatically increased globally over the past decade.
In the current crop year, 2017/18, it is forecast that world production will reach approximately 159 million bags, of which 111 million are estimated to be exported.
Overall, coffee is one of the biggest trading commodities in the developing world.
There are four major coffee-producing countries which account for around 70% of total global production. Brazil is the world’s leading coffee producer with an output of over 50 million bags annually. 75% is Arabica and the rest Robusta. About 65% of the country’s production is for export. The second largest coffee producing country is
Vietnam, which is the world’s largest Robusta producer. As indicated in Figure 1.1, since 1999, coffee production in Vietnam has increased dramatically. Between 25 and
30 million bags annually have been produced in recent years over 80% of which has been exported. The next major coffee producing and exporting countries are Colombia and Indonesia. Each has produced over 10 million bags annually in recent years, with about 70% of the national output exported. The rest of the coffee producing world includes Ethiopia, a number of Central American countries, Mexico and India (ICO,
2017a).
______Chapter 1: Introduction
The focus of the study presented in this thesis is on Vietnam’s coffee farming sector, which globally has exhibited the most rapid growth in production. This crop is an important part of the Vietnamese agricultural sector, as discussed in Section 2.1.
1.2 Research problems: unsustainable coffee production in Vietnam
Although coffee production has played an increasingly important role in
Vietnam’s Central Highland economy and in the global coffee market, recent literature indicates that production has been undertaken in a highly unsustainable way in terms of degradation of environmental resources and economic outcomes which affect the welfare of farmers. In particular, the literature has highlighted a number of resultant environmental, economic and social challenges for Vietnam’s coffee production (Dzung et al., 2011, and Marsh, 2007). There is mounting evidence that more chemical fertilisers and water have been used in coffee production than in other crops in Vietnam (Dzung,, et al., 2011; Kuit et al., 2013). In terms of chemical fertiliser consumption, in 2010
Vietnamese farmers applied on average 156 kilograms of nitrogen and phosphate fertilisers per hectare (ha) of arable and crop permanent area (FAO, 2014), while coffee plants requires up to 800 kg of nitrogen per ha in order to be able to achieve sufficient output quantity (Bruno et al., 2011). Comparing internationally, coffee production, in order to be profitable, is believed to require more fertilisers in Vietnam than in other countries. As noted by Marsh (2007), Vietnamese coffee farmers can only be profitable if coffee trees are grown intensively with large amounts of fertilisers (approximately 2 tonnes of chemical fertilisers per ha per year), compared with 63, 102, and 76 kilograms per ha for agricultural production in the Philippine, Thailand, and Indonesia respectively
______Chapter 1: Introduction
(FAO, 2014). The surplus nitrogen and phosphate content of inorganic fertilisers used in agricultural production injected into the water system is cited as being the cause of serious environmental problems (Hoang and Nguyen, 2013).
Irrigation water is also an environmental resource that is over-applied in coffee production in Vietnam1 (D’haeze et al., 2003; Dzung et al., 2011) . Although coffee farmers extract a large amount of water for coffee plantations, they currently do not pay user fees. In fact, D’haeze et al. (2003) have argued that lack of regulations in the use of water is one of the reasons leading to over-irrigation in coffee production and which may accordingly degrade groundwater sources. If there is no policy change in relation to water use, it could be expected that some farms will over-irrigate, creating water shortages for other farms including coffee and non-coffee farms. In a broader context, continued overuse of water in coffee production will negatively affect the entire agricultural production in Vietnam’s Central Highlands in the future.
Vietnamese coffee farmers are exposed to a number of risks that may lead to economic losses. Price volatility in general and reductions in prices in particular, have been the major challenges for coffee farmers. For example, during the coffee price crisis of 2001 prices declined to below production cost. A number of studies argue that one of the primary causes of this crisis was the significant expansion in coffee production
(Marsh, 2007), which was seen as a natural adaptation of farmers to high prices in the
1 Coffee farms in Vietnam use mostly underground water, meaning use of such water is tightly linked between neighbouring farms by being a common resource. Nitrogen and phosphorus compounds in chemical fertilizers can contaminate the soil and/or be transmitted into underground water systems polluting neighbouring farms.
______Chapter 1: Introduction
previous season (coffee prices peaked at historically high levels in the period, 1993-1995
(ICO, 2014).
Coffee farmers have also faced serious challenges including unexpected weather conditions, pests and diseases and the risk of other unforeseen events (Valkila, 2014).
However farmers currently have limited access to insurance services which could mitigate these risks (Gemech et al., 2011). This is an important deficiency given that investment in coffee farming is long-term with many uncertainties and risks associated with price fluctuations and unexpected climatic changes. In response to these risks, it has been argued that, on the one hand, coffee farmers could diversify their farming activities. On the other hand, there is some empirical evidence to show that specialization in coffee production could enhance productive efficiency and act as a mitigation strategy against these risks.
Given increasing environmental concerns from foreign and local consumers, coffee farmers have taken part in several internationally recognised sustainability- certified production practices. Several studies have proposed the expansion of these certification programs to deal with environmental concerns in Vietnam
(TECHNOSERVE, 2013, p.6). However, there lacks an informed analysis of benefits associated with these sustainability schemes2 with some studies showing that, in comparison to conventional farming, the economic benefits of certified coffee production for Vietnamese farmers are unclear (Kuit et. al. 2013). The literature shows
2 Over the past few years the total amount of Utz certified coffee has increased from 17,925 MT in
2009 to 38,669 MT in 2012, accounting for 21% globally of Utz certified coffee (Utz-certified, 2014).
______Chapter 1: Introduction
that small-scale coffee farmers in countries such as Ethiopia did not gain significantly from certification programs due to low productivity, an insignificant price premium and limitations in provision of credit and information (Jena et al., 2012). Coffee producers in
Nicaragua have experienced similar difficulties. Several authors including Valkila
(2009), have also argued that sustainable development in coffee production may lead farmers into a poverty trap due to low yields and an insignificant price premium compared to conventional growers. To date, none of these studies have examined the relationship between environmental efficiency and these certification programs.
1.3 Research questions
Although there are a number of empirical studies measuring efficiency levels and its variations across farms in agricultural production, few assess environmental performance in conjunction with economic benefits for coffee production, particularly in
Vietnam. Technical and cost efficiency varies from farm to farm due to different technologies and socio-economic factors. In addition, it is unclear whether crop diversification is an appropriate strategy to enhance economic efficiency due to a synergy effect, i.e., agronomic benefit. Environmental performance is a critical factor in sustainable coffee development, which may ensure economic benefits to farmers in the long-run. However, the literature (see, for example, Bacon, 2005; Wollni & Zeller
2007; Basu & Hicks, 2008; Wollni & Brümmer 2012), indicates that economic and environmental efficiency can vary considerably. However, there is a lack of empirical evidence of the dynamic trade-off. Since coffee is a primary crop in the region being researched, the role of economic and environmental efficiency enhancement can be
______Chapter 1: Introduction
significant in improving farmers’ welfare. In fact, many sustainable coffee production schemes aimed at improving economic and environmental dimensions have been adopted. Thus, empirical evidence of this trend is likely to be of interest to policymakers and other stakeholders involved in the coffee industry.
Therefore, the overall objective of this study is to provide a broad picture of both economic and environmental performance of coffee production in Vietnam. Specifically, it investigates the following research questions.
RQ1: What is the relationship between efficiency and levels of diversification/specialization?
As the literature review shows, empirical evidence is mixed on the question of whether coffee farms should pursue diversification or specialization strategies. This study will examine the relationship between economic and environmental efficiency and levels of diversification and specialization. Empirical evidence from the study is expected to provide information on whether coffee farmers in Vietnam should diversify or specialize to achieve further improvement in terms of economic performance due to a synergy effect, i.e., agronomic benefits. The concept of economies of scope and scale are used to identify whether diversification is beneficial for the efficiency of coffee production.
RQ2: Are there significant differences in the economic and environmental performance between sustainability-certified and non-certified farms?
______Chapter 1: Introduction
Using the same framework of production frontiers, the thesis will examine if eco- efficiency varies across two groups of farms - conventional farms and farms taking part in sustainability related certification programs.
Sustainable certification in the coffee industry has become increasingly popular.
However, the literature shows different outcomes for the costs and benefits associated with certification schemes and indicates a lack of empirical evidence on the level of environmental efficiency of certified coffee production. This study, therefore, provides an empirical comparison of both economic and environmental efficiency between certified and conventional coffee production.
Economic benefit is one of the most important incentives to attract coffee farmers to participate in certified production schemes. Accordingly, certified coffee production cannot be viable if farmers are unable to receive higher economic benefits than uncertified farmers. More importantly, given the limited market demand for certified coffee, certified coffee farmers with a low level of efficiency need to be encouraged to diversify out of certified coffee and choose alternative crops or conventional coffee production. This may provide better economic prospects for such farmers.
Additionally, there is an argument that coffee farmers should receive higher premiums for certified coffee if their production is environmentally friendlier (i.e., significantly higher in environmental efficiency). However, if there is no significant difference in environmental efficiency between certified and conventional coffee production, both farmers and certifiers should seek adjustments which would produce better environmental responsibility. The information needed for such adjustments which
______Chapter 1: Introduction
enhance environmental responsibility in coffee production and sustainable coffee production in the long-run is useful for all stakeholders, including coffee farmers, consumers, certifiers and policy makers.
RQ3: What are the determinants of variations in environmental and economic performance across coffee farms in Vietnam?
In order to propose appropriate policies for more sustainable coffee production, determinants of economic and environmental performance of Vietnamese coffee farms are essential to identify. This thesis will firstly examine levels of economic and environmental performance of farms. For measuring economic performance in terms of the technical, cost and cost allocative efficiency of coffee farmers, a production frontier analytical framework will be applied. In order to measure the environmental efficiency of farm this study will incorporate a wide range of environmental efficiency measurement techniques, i.e., an eco-efficiency approach and the Materials Balance
Principle in the production frontier framework. In the second stage, efficiency scores are regressed against socio-economic characteristics to identify sources of inefficiency.
RQ4: What are the trade-offs between cost and environmental efficiency and scenarios to pursue cost efficiency and/or environmental efficiency?
By incorporating both an economic and environmental efficiency assessment, this study will investigate whether there is a trade-off for coffee farmers between environmental and economic efficiency. Trade-off analysis provides an analysis of the current relationship between economic and environmental efficiency levels in farming.
This creates a basis for identifying the long-term total economic benefit that farmers ______Chapter 1: Introduction
may achieve in comparison to what they could have sacrificed to improve the current level of environmental efficiency in farming. Under the assumption of weak disposability3 of pollution, a negative economic-environmental trade-off is implied. In contrast, under the assumption of strong disposability4 of pollution, a farmer utilizes inputs more efficiently leading to a reduction of pollution. Thus, partly replacing chemical fertilisers with organic fertilisers may improve environmental efficiency given that coffee plants normally take a longer time to utilize organic than chemical fertilisers.
Hence, better environmental performance in the past may lead to higher economic efficiency in the future. This dynamic trade-off between environmental efficiency and cost efficiency is therefore examined. This is expected to provide empirical evidence on the relationship between economic benefit and environmental performance.
RQ5: What are the irrigation water efficiency levels? Is there a difference in water efficiency between certified and non-certified farms?
In addition to chemical fertilisers, irrigation water is a crucial input in coffee production. As identified in the literature, coffee cultivation heavily depends on irrigation water. There are serious environmental problems associated with water resources in coffee production: i.e., drought, over-extraction and regional shortage of
3 Weak disposability of pollution describes the situation where a farm can only reduce its pollution levels by decreasing outputs and pollution proportionally. In other words, reduction of pollution is costly
(Chung et al., 1997).
4 Strong disposability of pollution represents the situation where pollution can be reduced without incurring a cost in reduced outputs. This means that there is no private cost of pollution (Chung et al.,
1997). Also, see Macpherson et al. (2010) for a detailed explanation of strong and weak disposability.
______Chapter 1: Introduction
water. Thus, improving irrigation water efficiency is a strategic objective that the coffee farming sector could pursue to ensure better environmental outcomes.
1.4 Specific research objectives
Specific objectives are:
1. To examine the relationship between economies of diversification and efficiency
levels.
2. To identify differences in economic and environmental performance between
certified and conventional coffee production in Vietnam, and inefficiency
determinants.
3. To estimate economic efficiency (i.e., technical, cost and cost), environmental
efficiency (i.e., eco-efficiency, nutrient and water efficiency), and evaluate the
relationship between environmental and cost efficiency.
4. To examine the efficiency of irrigation systems which have different irrigation
technologies and water efficiencies between certified and non-certified farms.
1.5 Methodologies
The thesis employs a range of methodologies which can be categorised in two main groups. For the economic performance analysis the studies employ frontier frameworks. This included both parametric and non-parametric approaches. In chapter 5, a stochastic multiple output/input distance methodology is used. Chapter 6 uses a Data
Envelopment Analysis (DEA) eco-efficiency technique to estimate eco-efficiency scores. In the second stage, fractional regression and OLS are used to examine the
______Chapter 1: Introduction
drivers of eco-efficiency variation. In chapter 7, a DEA material principle based nutrient frontier and cost frontier are used to estimate cost and nutrient efficiency. Chapter 8 uses a sub-vector DEA meta-frontier technique to estimate irrigation water efficiency under different technologies. A standard OLS approach is used to examine the variation of irrigation efficiency under different irrigation technologies and the meta-technology.
More details about these methods are described in each of the studies.
1.6 Contributions
The literature indicates that there is a knowledge gap in the area of sustainability measurement of coffee production. This study is aimed at making an empirical contribution in the case of coffee production in Vietnam. The results are original and enhance the economic and environmental understanding of coffee production in
Vietnam. To our knowledge, then, this thesis provides several significant contributions.
First, this thesis is the first study examining the diversification efficiency of coffee production in Vietnam. Among different diversified and specialised farming enterprises, it is indicated that there may be agronomic benefits; i.e., a synergy effect from the joint cultivation of coffee and perennial crops. This indicates the transition of agronomic benefits into economic performance. It is noted in this context that in a developing country such as Vietnam, farmers still chose inefficient staple crops, such as rice, as a food security strategy.
Second, one of the primary purposes of sustainability certification schemes is to improve both economic and environmental performance of coffee cultivation. This
______Chapter 1: Introduction
thesis produces the first empirical examination of the economic and environmental performances of sustainability certification, i.e., certified farms versus non-certified farms. Certification schemes are likely to have a positive effect on eco-efficiency.
However, the level of eco-efficiency between certified and non-certified farms appears to converge over time. This convergence may be due to positive externalities of certification, less compliance with certification standards or a combination of these effects. This is useful information which can be used to inform various stakeholders about the gap between certification objectives and the real situation and how to bridge this gap.
Third, using the Material Balance Principles-based efficiency measure, this study presented in Chapter 7 proposes a new approach to classifying farms into four distinct groups. In terms of methodological contribution, it helps examining both efficient and inefficient farms, while previous studies only focused on technically efficient farms.
Furthermore, each group of farms faces with different problems (i.e., using unnecessary nutrients) and they should pursue different strategies to enhance better cost or environmental performance. This provides different scenarios in which coffee farms may choose to be both cost and environmentally efficient, more cost efficiency oriented or more environmentally friendly. This provides policy considerations on how farms could achieve better cost and environmental outcomes. For instance, the reconciliation between iso-cost and iso-nutrient lines may be helpful. This is valuable information for the development of tax and subsidy policies or water-pricing policies in terms of
______Chapter 1: Introduction
ensuring that more environmentally harmful inputs are costlier and irrigation water is priced using the nutrient shadow price.
Fourth, given irrigation water is a key factor in coffee production, further investigation is conducted into irrigation water efficiency under two distinct irrigation technologies of certified and non-certified coffee farming practices. This study confirms that modern irrigation technology, i.e., overhead sprinklers, is more productive than conventional technology. That is, modern technology has a significantly higher meta- technology ratio. This is important as over-irrigation is the most serious environmental challenges in coffee production. Note that one of the most important aims of certification schemes is to improve water efficiency. The empirical results suggest that irrigation technologies are better at enhancing water efficiency than by being sustainably certified. This information is useful for consideration as to whether certification schemes should incentivise coffee farmers to switch to modern irrigation technology.
Most importantly, this thesis provides a broader picture of the current economic and environmental performance of coffee production in Vietnam. That is, it seeks to go beyond focussing on improved economic benefits which are the “business as usual” aims of coffee producers. Primarily examined is economic viability in relation to the most serious environmental issues in coffee production. Hence, this dissertation offers a foundation for creating policies and management schemes that can support sustainable coffee production.
______Chapter 1: Introduction
1.7 Thesis outline
This thesis consists of nine chapters. Chapter 1 provides an overall summary of the information contained in the thesis. The background to coffee production in Vietnam and an overview of studies on coffee production are presented in Chapter 2. Chapter 3 provides information about the data sets used, the nature of ethics clearances and an overview of methodologies used. Chapter 4 provides a socio-economic profile of the
Vietnamese coffee farming industry. Chapter 5 to 8 detail the major studies of this thesis on the nature of coffee production in Vietnam as it relates to eco and productive efficiency. Chapter 9 summarise main findings of the empirical studies presented in this thesis and it offers policy options for the coffee farming sector in Vietnam and across the globe where production and management settings are similar.
Chapter 5 provides a comparative assessment of productive efficiency of three common coffee growing systems in Vietnam: mono-cropping, synchronization and segregation. Results from an input distance function approach deliver several important findings. First, the average inefficiency level is estimated to be approximately 18% and which vary significantly among the three farming systems. Second, the synchronised system of growing coffee and the other industry crops is found to be the most efficient.
Third, coffee mono-cropping is found to be less efficient than synchronised systems due to the presence of economies of scope between coffee and industrial crops. Fourth, the least efficient system is shown to be where cultivation of coffee and rice are only organisationally integrated. Other than efficiency, food insecurity could be a primary reason why coffee farmers diversify into rice. These findings also provide empirical
______Chapter 1: Introduction
evidence for the presence of agronomic benefits of synchronised systems which are translated into productive efficiency. Policy options aimed at promoting synchronised farming systems may therefore enhance both economic and agronomic benefits.
Chapter 6 aims at examining the eco-efficiency of sustainability certified and non- certified coffee farms. There is a belief that sustainability-certified coffee production helps increase economic benefits to farmers and reduces negative environmental impacts. However, international empirical evidence is not conclusive. This paper then, provides the first empirical examination of the differences in eco-efficiency between conventional and sustainability-certified coffee-growing farms in Vietnam.
Environmental pressures measured by the level of consumption of nitrogen, phosphorus, irrigation water, pesticides, herbicides, fungicides and land are investigated in relation to the value-added factor of coffee production. Empirical results show that, in each crop year, coffee farms could reduce environmental negative effects by more than 50% while holding the value-added of outputs constant. On average, sustainability-certified farms are found to be more eco-efficient than conventional farms but efficiency differences appear to converge overtime. This convergence may be due to positive externalities of certifications, less compliance to certification standards or a combination of these effects. Higher eco-efficiency levels are also correlated with farms located in higher elevations, having wind-break trees, and using sprinkler irrigation techniques.
In Chapter 7, cost and environmental efficiency are examined. The literature documents significant trade-offs between economic and environmental performance in many areas of agricultural production. However, there lacks research about such trade-
______Chapter 1: Introduction
offs in coffee farming, especially in Vietnam where significant overuse of nutrients and increasing voluntary participation in sustainability certification schemes are observed.
Also, past empirical studies focus only on the trade-off for technically efficient farms, while in the literature differing types of trade-offs exist. The present paper addresses these two main gaps in the literature by analysing different types of trade-offs between nutrient efficiency and cost efficiency for Vietnamese coffee farms in three crop years using the Material Balance Principles (MBP) efficiency frontier framework. Our study has several important findings. First, three distinct groups face different types of trade- offs with one group having no trade-off between production cost and nutrient consumption. Second, sustainability certified farms show higher cost and nutrient efficiency levels and have higher technical efficiency than non-certified farms.
Empirical results therefore suggest the need for both a more heterogeneous but also a more integrated policy approach to improve both cost and environmental efficiency of coffee farmers in Vietnam.
Chapter 8 focuses on an analysis of irrigation water efficiency under different irrigation technologies. As documented in the literature, the unsustainable use of irrigation water in coffee production is one of the most serious environmental concerns.
Many sustainability certification schemes aim to promote more sustainable water use; however, there is a lack of empirical evidence on the effects of those scheme on water efficiency. This study measures water efficiency and examines the effect of certification on water. Due to distinct differences between the two predominant irrigation methods - overhead sprinkler and micro-basin irrigation technologies - the meta-frontier
______Chapter 1: Introduction
framework is used to estimate water efficiency of certified and non-certified farms. To the best of our knowledge, this is the first study using the sub-vector DEA meta-frontier approaches to examine irrigation water efficiency. Results confirm that the sprinkler system is more advanced than the micro-basin irrigation technology while average water efficiency scores with respect to these technologies are less than 0.5, implying a large potential for farms to reduce water use. The effect of sustainability certification on irrigation water efficiency under different irrigation methods is examined. The results show that there is no significant impact of certification schemes on water efficiency.
Persuading farmers to switch to better technologies such as use of sprinklers is therefore recommended to reduce consumption of irrigation water.
Chapter 9 synthesizes the main findings from the major studies presented in this thesis. As well, a set of integrated policy implications and recommendations are developed. This is designed to assist the coffee farming industry in Vietnam to become more economically and environmentally viable.
An outline of the thesis is presented in Figure 1.
______Chapter 1: Introduction
Research problems:
Unsustainable coffee production in Vietnam: (1) environmental problems (overuse of chemical fertilisers and unsustainable extraction of irrigation water) and (2) economic inefficiency through excessive use of inputs
Research Aims: Economic performance Environmental performance Economic and analysis analysis environmental trade-off
Research Objectives:
Diversification and Eco-efficiency analysis: Trade-offs between cost Irrigation water productive efficiency sustainability certified and nutrient efficiency efficiency analysis versus non-certified di Approaches to Empirical Analysis:
Multiple outputs input Eco-efficiency modelling MBP-based nutrient Sub-vector DEA meta- distance function DEA and Fractional efficiency model frontier Stochastic frontier analysis regression DEA
Outcome: Provide empirical evidence to inform policy to achieve sustainable coffee production
______Chapter 1: Introduction 19
Industry background
2.1 Introduction
Vietnam’s coffee farming is one of the most important farming sectors in terms of its contribution to national export earnings as well as to the livelihood of a large number of households. To understand the current economic and environmental characteristics of this sector, information about historic and current production and relevant policies is important.
The purpose of this chapter is to provide a brief overview of coffee production in
Vietnam, the primary characteristics of which make coffee cultivation different to other countries, and to underline the importance of the industry for Vietnam’s economy and the global coffee market. This chapter also offers a detailed overview of the policy context within which Vietnam’s coffee farming industry operates and provides a summary of studies on coffee production.
2.2 Coffee production in Vietnam
2.2.1 Overview of coffee production in Vietnam
As cited in many previous studies, i.e., Dang and Shively (2008), coffee was first introduced by the French in the 1850s but remained an insignificant crop until the 1980s.
In the 1990s, Vietnam coffee production increased sharply with around an annual 30% growth rate. By the end of the decade coffee accounted for about 10% of the country’s annual export earnings. Production which was 1.31 million bags in 1990 increased to
______Chapter 2: Industry background
25.5 million bags in 2016 (Figure 1.1) when the total planting area reached 670 thousand hectares. In recent years, Vietnam’s share of global output has risen from 17% to 19%
(ICO, 2017b). Also, coffee production in recent times has earned Vietnam approximately US $3.2 billion in export earnings becoming one of the leading export earners for the country.
In Vietnam, out of 64 provinces, the five Central Highlands provinces of Dak Lak,
Lam Dong, Dak Nong, Gia Lai and Kon Tum produce about 90% of the country’s coffee output. This is equivalent to roughly 15% of global coffee production (Tran and Smith,
2017). According to the Sustainable Coffee Program (2013), coffee production currently provides the major livelihood for over 500 thousands households, who mostly reside in the Central Highlands. While coffee production and exports are clearly important to
Vietnam’s economy, to the Central Highlands people’s livelihoods and the global coffee market, its sustainability has been questioned.
2.2.2 Primary Vietnamese coffee farming characteristics
There are several characteristics making Vietnam coffee farming different to other major coffee producing countries. First, small-scale production is the typical production mode in Vietnam accounting for more 90% of coffee output. Each farm is on average between one to two hectares (Luong and Tauer, 2006). The average size of Brazilian coffee farms is around nine hectares (Waller et al., 2007, p.23). Second, Vietnam has, globally, the highest yield for coffee output being on average 2.5 tons per hectare (Tran and Smith, 2017). The figure for Brazil is approximately 1.6 (Rubio and Barros, 2017).
In a number of crop years, some Vietnamese coffee farmers have achieved yields of
______Chapter 2: Industry background
3.458 tons per hectare. This was the average yield of 644 farms surveyed between 2005 to 2009 (Amarasinghe et al., 2015). For Indonesia and other remaining coffee producing countries yields have been considerably lower (USDA, 2017).
As documented, while high-intensity Robusta cultivation brings high yields for
Vietnamese coffee farmers, the coffee sector has also faced many challenges. A dense monoculture of coffee trees with intensive use of chemical fertilisers and irrigation water is a typical coffee farming practice in Vietnam. As reported by Amarasinghe et al.,
(2015), the cost of fertiliser accounts for almost half production costs. Although irrigation water has no usage fee, labour cost for harvesting and irrigation accounted for about 34% of the production cost. It is noted that the variation in fertiliser application across farms and years is significant. This indicates that there is an opportunity to improve efficiency of fertiliser use, hence enhancing both economic and environmental benefits.
2.3 Sustainability certification programs in Vietnam
There are several certifiers actively providing certification services for coffee production. These certification schemes include Organic, Fair Trade certified, Rainforest
Alliance, Bird Friendly, Utz Certified, and 4C. A comparison matrix indicating each scheme’s main objectives and standards is available at the Specialty Coffee
Association’s website (Kline, 2009). In Vietnam, the major certifiers are 4C and Utz-
______Chapter 2: Industry background
certified (Utz): details of their standards can be found on their websites5. Overall, these certifications are aimed at improving economic, environmental and social sustainability through better farming practices. They cover both the issues of chemical fertilisers and irrigation water in focussing on the economic and environmental dimensions. Although both certifiers indicate the importance of the efficient use of these inputs in coffee production, neither provides appropriate doses of these inputs nor advocate particularly types of advanced technology, i.e., particular irrigation techniques.
Voluntary sustainability certification in coffee production is a market oriented strategy and this trend in certification has increased in Vietnam over recent years. Rising consumer concern about environmental and social sustainability has meant an increasing number of coffee roasters have committed to provide coffee that complies with sustainability certification standards. Being certified is largely based on cooperation between roasters or exporters and individual farmers or farmer groups. There are a number of roasters and exporters operating in the Vietnamese coffee market. But there is a lack of official statistics which detail certification programs. Relevant statistics are collected from reports of different stakeholders.
The certification trend among Vietnamese coffee farmers has expanded rapidly in recent years with approximately 33% of the coffee output in 2012 being certified by either
4C (20%), Utz (11%) or Rainforest Alliance (1%) (as cited in Boselie, 2016). According
5 Details of the indicators and standards of these certification schemes subsequently updated are accessible at their websites (4C: http://www.globalcoffeeplatform.org/resources/gcp- baseline-common-code; and Utz-certified: https://utz.org/?attachment_id=3623).
______Chapter 2: Industry background
to a presentation by a deputy chair of the Vietnam Coffee Coordination Board, the current certified area of coffee cultivation accounts for about 60% of the total coffee cultivated area. Another source estimated that give the level of increase in certification up to 2013, over 80% of Vietnam’s coffee supply would be certified by 2016 and thereby making certified coffee by far the most common type supplied (The Sustainable Coffee Program,
2013).
2.4 Policy context
Historically, there have been several particular policies contributing to the rapid growth of Vietnam’s coffee production. First, after the country’s reunion in 1975, the government of Vietnam implemented the “new economic zone (Doi Moi)” policy promoting state-sponsored migration to the Central Highlands region. This is also known as a period of privatisation and economic liberalisation (Dang and Shively,
2008). Second, the adoption of market-oriented policies by the country in the late 1980s and the collapse of the International Coffee Agreement (ICA) in 1989 (Waller et al.,
2007), enabled Vietnamese coffee producers to compete freely in the global coffee market (Luong & Tauer, 2006). These influences are accepted as being the main drivers promoting the significant growth of coffee production in Vietnam.
Vietnamese policy makers have recognised the many challenges facing development of the coffee industry, and in doing so have attempted to support its development in a sustainable manner. As cited in a recent report (IDH 2014, 4–5), a
Deputy Director of Vietnam’s Coffee Sector Coordinating Board, in the Ministry of
Agriculture and Rural Development (MARD) remarked: ______Chapter 2: Industry background
“To date, Vietnam’s coffee sector has made outstanding progress.
However, it faces many sustainability challenges. These include “hot”
development of cultivation area; exhausted soil; abuse of fertilizers;
inefficient water use, which leads to exhausted water sources in the
context of the global climate change; quality and food safety control;
outdated technologies; adding value to export products; and benefit-
sharing mechanisms between coffee farmers and companies to ensure
long-term business relationships.”
In 2013, the Extension Program for Sustainable Coffee Production in Vietnam was approved by MARD. This program has provided technical training for coffee growers.
Another program for coffee rejuvenation has provided coffee farmers with seedling support, technical guidelines for land preparation, root treatment and crop rotation and credit/loans for coffee production. In addition, the government has also implemented a program to restructure the agriculture and rural sectors and other rural development programs.
In 2014, the Ministry of Agricultural and Rural Development created a sustainable coffee development plan which encompasses the years through to 2020. The primary objectives are to (1) stabilise the coffee planting area of around 600 thousand hectares
(currently about 670 thousand), (2) maintain 80% under sustainability certified production, (3) maintain average yields of 2.7 tons per hectare, (4) sustain exports at approximately 1.6 million tons annually, and (5) derive annual export revenues of between US $3.8 to $4.2 billion.
______Chapter 2: Industry background
To date, there is no program that evaluates sustainable coffee production in terms of interdisciplinary aspects of its cultivation. Thus, this study is designed to produce policy options based on empirical evidence to enhance future sustainable coffee production.
2.5 Overview of studies on coffee production
Because of its importance for many developing economies, coffee farming has been the subject of numerous studies (see, for example, Giuliani et al., 2017; Chiputwa et al., 2015; Luong and Tauer, 2006). Although many have been centred on the technical aspects, e.g., genetic (Goldfarb et al., 2005), the majority have focussed on coffee farmers’ benefits and the effect of sustainability certification schemes ( see, for example,
Barham and Weber, 2012; Kilian et al., 2006; Méndez et al., 2010; Ruben and Fort,
2012; Wollni and Zeller, 2007). A review of 46 studies on the effect of certification on agricultural production by Blackman and Rivera (2011), showed there were 22 studies devoted to coffee farming. Since then, the number of studies on coffee farming has increased (for example, Bray and Neilson, 2017; Elder et al., 2013; Hardt et al., 2015;
Ibanez and Blackman, 2016).
There are several studies that examined cost and technical efficiency and inefficiency variations, although there are none investigating envrionmental efficiency.
Rios & Shively (2006) using constant return to scale DEA found that coffee farmers in
Vietnam’s Dak Lak province were technically inefficient, with an average score of 0.78.
Using the same technique, the authors also found that cost efficiency was relatively low with an average score of 0.39. This meant farmers would have been able to save 61% of ______Chapter 2: Industry background
production costs while maintaining the same output level. Another study also use the
DEA technique found a relatively low technical efficiency level for Vietnamese coffee farmers of 0.70 (Garcia and Shively, 2011). In addition, Ho et al. (2014) using the stochastic frontier analysis (SFA) technique found that the existence of inefficiency was statistically significant at an average level of 0.70 for coffee farmers in Dak Lak,
Vietnam. These imply a posibility of efficiency improvement in regard to less intensive use of input factors, since economic theory would suggest some benefits of specialisation.
In creating an inefficiency model for the seond stage, a series of socio-economic and geographical variables were incorporated. For example credit loan, formal education of the household head, farming experience and extension services were seen as key factors in efficiency improvement (Ho et al., 2014). Participation in a farming association was also shown to help coffee farmers be more efficient in Cote d’Ivoire
(Binam et al., 2003) and similarly in Costa Rica (Wollni and Brümmer, 2012). In addition, farming experience, other income sources rather than coffee, and age of the household head have been shown to be important efficiency determinants for speciality coffee production (Wollni and Brümmer, 2012). Other socio-economic factors affecting coffee production efficiency have also been identified in studies by Binam et al. ( 2003),
Rios & Shively (2006), Vedenov et al. (2007) and Wollni & Brümmer (2012).
Therefore, it is aruged that economic efficiency analysis using a panel data set for coffee produciton is necessary in order to produce robust empirical evidence given previous studies used only cross sectional data.
______Chapter 2: Industry background
Despite the many international studies on coffee farming, there are only a few devoted to coffee farming in Vietnam none of which examine economic and environmental sustainability. These previous studies have indicated coffee farming was being carried out at an unsustainable level (for example, Dzung et al., 2011;
Amarasinghe et al., 2015). A detailed review of studies on coffee farming examining economic and environmental performance are presented in Sections 4.2, 5.2, 6.1 and 7.1.
In short, none focus on a comprehensive assessment of both economic and environmental aspects of Vietnam’s coffee farming sector. In addition, the economic and environmental effect of sustainability certification schemes on Vietnam’s coffee production remains unclear.
______Chapter 2: Industry background
Study design
3.1 Introduction
From the overview of the literature in Chapter 2, it is clear there is only limited economic and environmental information available on the coffee farming sector in
Vietnam. However, in-depth information about the current level of economic and environmental performance is of significance for the development of evidence-based policy options promoting the future sustainable development of the coffee sector.
The purpose of this chapter is to provide information about the way in which the economic and environmental performance of Vietnam’s coffee farming sector is measured in this research paper. The research design includes a hypothesis, an overall research objective and research aims. This chapter also presents information on the scope and nature of data used, ethics and a review of methodologies used in the studies presented in this dissertation.
3.2 Hypothesis and overall objective
The literature indicates that Vietnam’s coffee farming sector has become economically and environmentally unsustainable given the exceptionally rapid expansion in both non-certified and certified coffee production in recent years. This motivates a comprehensive assessment of economic and environmental performance of the industry. Therefore, this thesis has focused on the following hypothesis: There is a potential to improve both economic and environmental viability of Vietnam’s coffee farming sector. ______Chapter 3: Study design
The overall objective of the research presented in this thesis is to enhance the understanding of the economic and environmental status and potential sustainable viability of the Vietnam’s coffee farming sector.
3.3 Research aims
Based on the hypothesis and the research objectives, the thesis pursues the following research aims:
Aim 1: Determine the level of productive efficiency of Vietnam’s coffee farming;
identify efficient farming systems and explain the reasons why current farming
systems exist.
Aim 2: Measure economic and environmental performance of the Vietnam’s coffee
farming via the eco-efficiency indicator, identify the drivers for eco-efficiency
variation and examine the difference in eco-efficiency between certified farms
and non-certified farms.
Aim 3: Measure economic and environmental performance of the Vietnam’s coffee
farming via two indicators, namely cost efficiency and nutrient efficiency, and
identify possible scenarios that farms could achieve cost efficiency, nutrient
efficiency and trade-offs between cost and environmentally efficient options.
Aim 4: Examine environmental performance via irrigation water efficiency under
different irrigation technologies and compare the irrigation water efficiency
between certified and non-certified farms.
______Chapter 3: Study design
Aim 5: Discuss policy implications that support a sustainable future development of the
Vietnam’s coffee farming sector.
3.4 Study site, data and ethics
3.4.1 Study site
The research site of this study is the Central Highlands, the dominantly largest coffee-producing area in Vietnam. The Central Highlands region includes five provinces: Dak Lak, Lam Dong, Dak Nong, Gia Lai, and Kon Tum. In these provinces, soil and climatic conditions are especially favourable for industrial crop cultivation (i.e., coffee, pepper, and avocado). In 2015, the Central Highlands accounted for approximately 88 percent of the coffee planting area and coffee production in Vietnam
(Dries et al., 2015). Over 85 percent of coffee plantations are operated by smallholders, with a small production scale of 1–2 hectares (Luong and Tauer, 2006). In the region, certified coffee production has increasingly expanded in recent years. This trend is recognized as coffee farmers and other value chain partners adhere to predefined sustainable standards including social, environmental, and technical aspects. In the coffee sector in Vietnam, certified production has been mostly operated under cooperation between coffee farmers and roasters, and local or international exporters.
Common certification schemes that have actively operated in the region are UTZ, 4C, and Rainforest Alliance.
______Chapter 3: Study design
3.4.2 Data and ethics
There are two datasets used to achieve the above outlined research aims which are largely based on primary data. Specifically, the dataset used in Chapter 4 was provided by the Department of Economics, Tay Nguyen University, Vietnam. More details about the survey are described in Section 4.5. This dataset includes aspects relevant to coffee production such as coffee output, cultivating area, information on input use and relevant farm characteristics. The second dataset used in the analysis in Chapters, 6, 7 and 8, covers various aspects of coffee production in terms of economic, environmental and social variables relevant to coffee production. The collection of this dataset was approved by the QUT Human Research Ethics Committee (approval number:
1500000663). A copy of the questionnaire is provided in Appendix A.
A mixed cluster and stratified random survey of 896 coffee farmers was conducted using a face-to-face interview technique in three provinces (328 farms in Dak Lak, 348 farms in Lam Dong and 220 farms in Gia Lai provinces). Data on three crop years
2012/13 to 2014/156 was collected for each farm. The selection of districts and communes where farms were located was based on several criteria including the importance of coffee as the key livelihood for farmers, geographical and ecological representation as well as the popularity of certified coffee production. Consultation with the local offices of the Department of Agriculture and Rural Development was carried
6 The crop year of coffee farming in Vietnam is similar to the calendar year. It starts in January and finishes in January of the following year. Therefore, hereafter the crop years 2012/13, 2013/14 and
2014/15 are used as years 2012, 2013 and 2014 respectively.
______Chapter 3: Study design
out to reduce potential biases due to variations in economic conditions, production scale, as well as social dimensions across the three provinces. Each local administrative unit is considered as a cluster.
A structured questionnaire set was used in the survey. Prior to the main survey, focus group discussions and pre-tests were conducted to finalise the survey instrument.
The field survey was undertaken from September 2015 to January 2016. This was based on the fact that most coffee farmers use a log book or a diary to record their farming activities. Only in cases where they did not have a farming log book, were farmers required to recall information by memory. We acknowledge that recalled data could pose limitations on information accuracy and for this reason we asked for data for only three historical years. Some observations have missing data; hence and were accordingly removed.
3.5 Methodological reviews
3.5.1 Overview of efficiency analysis
In recent years, there has been an abundant addition to the literature on empirical analysis of farming activities. This literature focuses on the technical, cost and environmental efficiency of farmers. The term, ‘technical efficiency of an individual producer’, can be described as his/her ability to minimize input use in the production of a given output vector, or the ability to obtain maximum output from a given input vector.
The measurement of technical, allocative and economic efficiency was initially proposed
______Chapter 3: Study design
by Farrell (1957) and has since attracted considerable attention by researchers who have attempted to narrow the gap between theory and empirical application.
Efficiency and productivity measurement is a longstanding area of study in economics widely used to gauge economic performance of a firm, industry or economy.
There have been two primary trends in efficiency analysis - econometric and mathematical. Since Aigner et al. (1977) proposed the SFA - known as the parametric approach - many studies have successfully been developed and applied to measure efficiency and analyse variations in efficiency (for example: Pitt & Lee 1981;
Reifschneider & Stevenson 1991; Kumbhakar et al. 1991; Huang & Liu 1994; Battese &
Coelli 1995; Alvarez & Arias 2004; Illukpitiya & Yanagida 2010; Yu-Ying Lin, Chen,
& Chen, 2013; Mutoko et al., 2014). Another technique known as the non-parametric approach was proposed by Charnes et al. (1978) based on mathematically linear programming to estimate efficiency scores as the first stage of Data Envelopment
Analysis (DEA). In the second stage the efficiency indexes are regressed by a vector of variables to examine the variation in efficiency. The DEA approach has been widely employed in many areas of economics e.g. agricultural economics and the measurement of industry and environmental efficiency (also see: Chambers et al. 1998; Coelli 1998;
Ara 2002; Amirteimoori et al. 2006; Lilienfeld & Asmild 2007; Munksgaard et al.
2007).
3.5.2 Methodological review of environmental efficiency
In agricultural production, environmental aspects have become an increasingly important benefit for farmers as well as for the global economy in the longer run.
______Chapter 3: Study design
Pittman (1983) produced one of the first studies attempting to incorporate undesirable outputs in agricultural production into conventional efficiency indexes. He reveals that incorporation of undesirable outputs significantly affects the rankings. Using efficiency and productivity measurements based on production frontiers, many studies have attempted to integrate environmental performance of producers into technical and economic efficiency measures. See, for example, Färe et al. (1989), Reinhard et al.
(1999); Reinhard et al. (2000); Wossink & Denaux (2006); Coelli et al. (2007);
Lilienfeld & Asmild (2007); Lauwers (2009); Frija et al. (2009); Hoang & Nguyen
(2013).
Generally, there are two primary approaches to measuring environmental performance in agricultural production. The first focuses on the development of various indicators which can be used to describe differences across different systems (Wu &
Wu, 2012; Bell & Morse, 2008). The second approach employs production frontiers to derive efficiency and productivity measures and uses both parametric and non- parametric techniques to estimate relative efficiency. Both approaches examine relative performance as well as providing a temporal analysis of performance changes. The relative performance comparison and assessment are useful in evaluating the performance of individual firms or systems (hereinafter also called a decision-making unit (DMU)) relative to all other DMUs.
3.5.3 Indicator approach
Various indicators representing environmental performance are constructed by using the flows of both monetary values and physical values such as the content of
______Chapter 3: Study design
materials or energy involved in the production process. Common physical indicators are energy balance, nutrient efficiency and thermodynamic energy efficiency. These indicators are also called agroecosystem performance indicators (APIs) (Tellarini and
Caporali, 2000) or agri-environmental indicators (AEIs) (Piorr, 2003). They describe the relationship between material (such as nutrients) and energy inputs and outputs. The balance represents the difference between energy or material (nutrient) inputs and those of the outputs. Thus efficiency reflects the ratio of energy or nutrients in inputs to those in outputs (Hoang, 2011, p. 4,5). In regard to ecological sustainability, the quality of energy and mass, defined as exergy (Wall, 1977), has been increasingly used to construct environmental or ecological indicators. This includes estimating cumulative exergy balance and exergy-based system transformity (Chen & Chen, 2007; Hoang &
Alauddin, 2010). Some studies have attempted to include all the flows of monetary value, energy, materials and exergy in order to analyse economic and environmental performance of production systems (see, for example, Tellarini & Caporali, 2000;
Hoang & Alauddin, 2010).
The indicator approach is simple to calculate and useful in providing information on the economic and environmental performance of individual DMUs. This approach does not, however, help to explain changes in the economic benefits and environmental performance of a network or an industry where relative efficiency levels among DMUs are taken into account (Hoang, 2011, p. 5). The drawbacks of the indicator approach can be overcome by the frontier-based methodology where efficiency and productivity are measured, based on production frontiers.
______Chapter 3: Study design
3.5.4 Frontier- based approach
The frontier-based approach has been traditionally used to examine the relationship between a set of inputs and corresponding outputs of DMUs associated with production technology. This relationship represents the economic performance of individual DMUs compared with other DMUs. These measures can also be decomposed into two meaningful economic components - technical efficiency and allocative (price) efficiency (Farrell, 1957). A number of studies have developed and employed this approach to measure, not only technical efficiency, but also allocative, cost, revenue and profit efficiency using information on prices. In addition, it allows investigation of factors affecting variations in efficiency levels across DMUs. Information about key determinants of efficiency variation provides useful guidelines on how to affect these determinants in order to improve the efficiency of DMUs.
In terms of environmental efficiency measurement, both parametric and non- parametric techniques have been employed. Lauwers (2009) summarised two types of frontier-based eco-efficiency models and the advantage of the material balance based model (MBM). The first type, the environmentally adjusted production efficiency
(EAPE) model, considers pollution as either environmentally detrimental inputs or undesirable outputs, which are to be minimised. Therefore, technical efficiency can be estimated using an input/output oriented approach, or using hyperbolic or directional distance functions (Färe et al., 1989; Chung et al., 1997; Reinhard et al., 2000; Wossink
& Denaux, 2006)). However, these EAPE models are in conflict with fundamental thermodynamic laws, such as mass conservation (see: Coelli et al. 2007, Hoang & Coelli
______Chapter 3: Study design
2011; Hoang & Nguyen 2013). The second frontier-based eco-efficiency model (FEM), examines relationships between economic and ecological outcomes, using the frontier framework. The primary use of the FEM is to assess relative environmental performance among DMUs. This is important given there are many types of environmental pressures caused by production and consumption activities.
3.5.5 Materials balance-based approach
The principle of materials balance explains that materials in inputs are transformed into outputs, including desirable and undesirable outputs (e.g., emissions) that can cause pollution (Ayres, 1995). Since the materials balance principle hold true in any cultural system, it has received increasing attention in efficiency literature. Reinhard &
Thijssen's (2000) study is one of the first attempts to address the materials balance issue.
The study used the materials balance principle to calculate the nitrogen pollution variable which is treated as an input variable in DEA models. Environmental efficiency is measured by contracting the pollution input variable, while holding the output and the conventional inputs constant. In another study Reinhard et al. (1999) treat nitrogen surplus as the most serious environmentally detrimental input for Dutch dairy farms and which is shown to have been brought about by the application of manure and chemical fertilisers. In this case, nitrogen efficiency (which is known as environmental efficiency), is estimated to be 44% on average. This means that 66% of nitrogen can be reduced without affecting output. However, this treatment of nutrient balance as an input in the production frontier framework causes mathematical violation of the DEA solution
(Coelli et al. 2007).
______Chapter 3: Study design
Coelli et al. (2007) proposed the use of the material balance principle based on standard cost minimization. This study focuses on both cost minimization and nitrogen or nutrient minimization, in which the nitrogen or nutrients are contained in inputs.
Unlike the previous approaches, this method does not require incorporating additional inputs or outputs in production frontiers. However some shortcomings - such as the lack of universal acceptance of weights for various materials and the ambiguous treatment of immaterial inputs – have been identified. These two limitations can be overcome by the use of cumulative exergy content (Hoang and Rao, 2010).
Regarding measurement of nutrient and environmental efficiencies, an innovative approach is proposed by Coelli et al. (2007). This is based on the material balance-based model (MBM). Lauwers (2009) indicates that the MBM bridges the gap between the conventional concepts of production efficiency and eco-efficiency. Compared to the environmentally adjusted production efficiency models, the MBM is not in conflict with fundamental thermodynamic laws. In comparison to the frontier eco-efficiency models, the MBM makes its economic and environmental outcomes equally explicit (Lauwers,
2009). Additionally, Hoang & Nguyen (2013) employed both DEA and SFA techniques with the application of the MBM to examine environmental efficiency and its determinants for rice production .
3.5.6 Nutrient efficiency
Nutrient efficiency is one direct application of the materials-based efficiency models in which only several nutrients such as nitrogen and phosphorous are considered.
This is in response to the fact that the excessive use of these two nutrients are known as
______Chapter 3: Study design
two major causes of eutrophication (Smith et al., 2006). Using the application of the materials balance principle, Hoang & Nguyen (2013) propose that because of the aggregation of nitrogen and phosphorus contaminated in the environment due to rice production they should be key elements for examining materials-based environmental efficiency. In this study, nutrient efficiency is defined as the potential reduction in the amount of nutrients that each DMU can achieve without reducing its output.
Interestingly, the nutrient efficiency can also be decomposed into two terms: technical efficiency and nutrient allocative efficiency. Technical efficiency refers to potential reductions in the use of conventional input factors such as land, labour, and chemicals.
The nutrient allocative efficiency refers to the potential for reducing the use of nutrients contained in different compounds contained in chemical fertilisers, organic fertilisers, and even in the top layer of soil. This type of analysis is particularly useful in analysing the trade-off between cost efficiency and environmental efficiency as both measures of efficiency are based on the farm cost frontier framework. A search of the literature indicates that there is no empirical investigation into the environmental efficiency of
Vietnamese coffee production, notwithstanding the fact that the overuse of chemical fertilisers has been identified as a critically important issue (Dzung et al., 2011).
3.5.7 Irrigation water efficiency
Irrigation water efficiency is measured as the minimum possible amount of irrigation water used to produce a given level of outputs. Where a potential reduction in irrigation water can be achieved without adversely affecting yields, this indicates excessive use of irrigation water (Lilienfeld & Asmild, 2007; Speelman et al., 2008;
______Chapter 3: Study design
Varghese et al., 2013). Over-irrigation of coffee producing land has been demonstrated by a number of studies (D’haeze et al. 2003; Kuit et al. 2013). This study finds that in
Vietnam, larger farms and farms with younger operators are more likely to be characterised by efficient use of water. Notably, irrigation water efficiency is low in tomato, melon and pepper production, in spite of the government’s subsidies for water- saving technologies. It is clear that production technology and irrigation systems can help farmers improve the level of irrigation water efficiency (Wang, 2010). For example it has been shown that farmers who received greater exposure to extension services (i.e., technical training, water conservation, and the use of fertigation) tended to achieve a higher level of irrigation water use efficiency (Frija et al., 2009).
In this study of Vietnamese coffee producers the calculation of irrigation water efficiency is by way of the DEA technique rather than the SFA. Lilienfeld & Asmild
(2007) provide an example of relative irrigation water inefficiency in agricultural production employing the sub-vector DEA measurement approach (also see: Speelman et al., 2008; Frija et al., 2009; Varghese et al., 2013). These studies examine the reduction potential of irrigation water alone, based on the input-oriented efficiency model. In addition, sub-vector DEA efficiency is also employed to estimate irrigation water use efficiency in horticultural greenhouses. Variations of the irrigation water use efficiency can be analysed using a Tobit regression model. Wang (2010) used the same approach to investigate the irrigation water efficiency of farmers in north-western China.
Overall the DEA is preferred to SFA due to its flexibility in calculating sub-vector efficiencies (Speelman et al., 2008; Varghese et al., 2013).
______Chapter 3: Study design
______Chapter 3: Study design
Socio-economic industry profile
4.1 Introduction
Economic literature has indicated that socio-economic characteristics play an important role for the development of a wide range of industries. In small scale coffee farming, factors including ethnicity, education, age, and gender have been shown to have a significant effect on management behaviour of households (Ho et al., 2014;
Wollni and Brümmer, 2012). Investigating the socio-economic characteristics of the coffee farming sector can provide useful information on the issues facing the sector and industry characteristics. This is important for both researchers and practitioners when examining practical issues faced by in coffee farmers and building up their management expertise.
Despite the significant contribution to the Vietnamese economy, people’s livelihood and the global coffee market, the coffee farming sector faces many challenges. However, socio-economic information on coffee farmers and their families is lacking in the literature. Clearly, in the development of farming programs and policies, socio-economic profiles of the target groups can provide essential background.
Previous studies have identified a number of serious problems facing the coffee farming sector and in particular heavy dependence on chemical fertilisers and overuse of irrigation water (Amarasinghe et al., 2015; D’haeze et al., 2005b). Moreover there is little information regarding the current situation in terms of to what extent there has been a change in the overuse of these inputs.
______Chapter 4: Socio-economic industry profile
The aim of this Chapter is to develop a current socio-economic profile of the
Vietnamese coffee farming sector and an up to date analysis of production, costs, profitability and income of coffee growers. This study seeks to answer following research questions: What are the key elements of a socio-demographic profile of
Vietnamese coffee farmers? What are the main factors influencing coffee production in
Vietnam? Do economic and environmental factors, such as chemical fertilisers and irrigation water, have an important influence on Vietnam’s coffee production?
The data used in this study comes from a recent coffee farming survey, conducted from September 2015 to January 20167. To our best knowledge, this is one of the largest surveys carried out of Vietnamese coffee farming households in the Central Highlands providing a wide range of information. The main method used in this Chapter is the descriptive approach which provides the key features of the socio-economic profile and main factors affecting production.
4.2 Demographic characteristics
The demographic characteristics of the 896 surveyed coffee farmers were categorised according to the three-major coffee growing provinces in Vietnam, namely
7 There are two datasets used in this thesis, although this socio-economic profile and economics of coffee farming was developed using the recent and largest dataset, as described in Section 3.4.2. The survey was undertaken among coffee farmer from three largest coffee growing provinces in Vietnam, namely Dak Lak, Lam Dong and Gia Lai.
______Chapter 4: Socio-economic industry profile
Dak Lak, Lam Dong and Gia Lai. Information includes age of the household head, family size and ethnicity.
Table 4.1: Demographic characteristics of surveyed coffee farmers Demographic characteristics Dak Lak Lam Dong Gia Lai Total Number of farms 328 348 220 896 Age of the household head (years) Mean 49.91 46.87 48.34 48.34 St. dev. 9.51 11.86 9.53 10.56 Min 20 20 23 20 Max 79 80 74 80 Family size (persons) Mean 4.71 4.43 4.45 4.54 St. dev. 1.52 1.55 1.37 1.50 Min 2 1 1 1 Max 11 11 10 11 Ethnicity Kinh 72.87% 92.73% 84.20% 82.14% Others 27.13% 7.27% 15.80% 17.86%
The age of coffee growers ranged from 20 to 80 years with an average of 48.34.
There is no significant difference in the age of coffee farmers across the three provinces.
Coffee farmers in Dak Lak were slightly older than the farmers in the other provinces, while farms from Lam Dong were the youngest on average - 46.87 years (Table 4.1).
Across the three provinces, the majority of coffee farmers were aged between 40 to 50 years. The second largest group of household heads were aged between 50 to 60 years.
______Chapter 4: Socio-economic industry profile
Nearly 30% of the total sample were aged under 30, while about 70 household heads
(8% of the sampled coffee growers) were aged over 60 (Figure 4.4).
Figure 4.1: Distribution of household head ages
In developing countries such as Vietnam, the size of a family is usually associated with the size of its available labour force. Thus in the coffee farming sector the family is the main source of labour and provides a useful insight into the labour availability for harvesting and irrigating. The household composition of all sampled coffee growers indicated they were composed of an average of 4.54 people which was roughly similar across the three provinces. Dak Lak had the largest average family size of 4.71 people, while Gia Lai and Lam Dong have slightly smaller family sizes of 4.43 and 4.45 people.
The majority of households - 60.2% - had a family size ranked between 3 and 5 people.
4% had families greater than 7 people, while almost 20% had small families less than or equal to 3 people
(Figure 4.2).
______Chapter 4: Socio-economic industry profile
Figure 4.2: Family size distribution
In terms of ethnic diversity, the surveyed area of the Central Highlands is known as the country’s most diverse region. The dominant group of coffee farmers is the Kinh group accounting for 82.14%. The rest of the surveyed participants belong to the Ede,
Tay, Nung, K’ho or M’nong groups. 27.13% of the farmers in Dak Lak were ethnic minority people as were 15.80% and 7.27% in in Gia Lai and Lam Dong provinces respectively (Table 4.1).
Figure 4.3 shows the educational attainment levels of household heads. There were eight education categories8 comprising of no education attainment, primary school not finished, primary school finished, secondary school not finished, secondary school finished, high school not finished, high school finished, and more advanced education attainment. In general, the sample had quite low education backgrounds with a majority not finishing high school. Lam Dong had the lowest education level, with the highest proportion not finishing secondary school. In Dak Lak and Gia Lai, the largest group of coffee farmers were those who had not progressed past finishing the secondary school
(Figure 4.3).
8 The current education system in Vietnam defines that primary school level is from year 1 to year 5, secondary school level is from year 6 to year 9, high school level is from year 10 to year 12 and more advanced level includes vocational trainings, higher education and graduate levels. ______Chapter 4: Socio-economic industry profile
Figur e 4.3: Educ ation level s of the hous ehol d heads
4.3 Economic profile of coffee farming
Table 4.2 presents an economic profile of coffee farming derived from 1,934 survey observations. This includes production factors, financial performance measures and direct production costs. Of the sampled farmers 52% participated in sustainability certification schemes, i.e., Utz-certified and 4C (also see Section 2.3). Regarding production factors, the average coffee yield was 3.07 tons of dried coffee beans per ha.
Other physical inputs are described and summarised in Table 4.2. It is important to note that coffee farmers do not pay usage fees for extracting water for irrigation purposes.
Thus, irrigation costs are only composed of the associated labour, fuel, electricity and irrigation facilities. These are therefore captured in family labour, hired labour and irrigation costs.
If family labour, land and coffee tree depreciation costs are excluded, the survey shows that coffee growers generally made a profit from their farms which averaged 1.5 hectares in size. The net income (the difference between gross income and direct ______Chapter 4: Socio-economic industry profile
production costs) averaged 80.11 million VND per hectare. The total direct production cost was 39.83 million VND per hectare. This includes the costs of hired labour, NPK fertilisers, other chemical fertilisers, organic fertilisers, irrigation (i.e., fuel and electricity costs), pesticides, herbicides, fungicides and machinery.
Table 4.2: Descriptive statistics of economic profile (measured in million VND per ha) St. Statistic Measure Mean dev. Min Max Certification % 0.52 0.50 0.00 1.00 Physical production factors Coffee yield ton per ha 3.07 0.99 0.50 7.00 Area per farm ha 1.59 1.18 0.20 13.00 Family labour man-days 213.54 90.91 0.00 450.00 Hired labour man-days 49.17 49.30 0.00 300.00 Chemical fertilisers tons 2.43 1.18 0.00 11.40 Irrigation water m3 1.16 0.50 0.03 5.28 Financial performance measures million VND per Gross income ha 117.97 41.03 0.00 250.00 million VND per Net income ha 79.10 38.04 -67.90 170.24 Production costs million VND per Direct production cost ha 38.87 17.47 5.05 148.98 million VND per Hired labour cost ha 7.35 7.56 0.00 60.00 million VND per Chemical fertiliser cost ha 22.25 10.50 0.00 86.67
______Chapter 4: Socio-economic industry profile
million VND per Organic fertiliser cost ha 4.25 6.41 0.00 92.19 million VND per Irrigation cost ha 3.39 3.39 0.04 44.00 million VND per Pesticide cost ha 0.81 0.90 0.00 9.19 million VND per Machinery cost ha 0.83 1.27 0.00 23.00
Among the production cost components, that of chemical fertiliser was the predominant one. On average, the cost was 22.25 million VND per ha, accounting for about 56% of total direct costs. The second largest cost component was hired labour equating to an average of 7.76 million VND per ha. As noted, the irrigation cost only includes fuel and electricity costs. Irrigation activities also require a large amount of labour which is included in the labour input components (i.e., family labour and hired labour).
Table 4.3 presents production factors, income and production cost breakdowns of average coffee production per hectare for the three provinces and three crop years. In terms of production scale, the average coffee farms in Lam Dong were the largest followed by those in Gia Lai and Dak Lak. However, farmers in Gia Lai and Dak Lak achieved markedly higher yields than those in Lam Dong. The use of key inputs also varied significantly between provinces but not over crop years. In particular, coffee farmers in Gia Lai tended to use more labour and chemical fertilisers than those in the other two provinces on average.
______Chapter 4: Socio-economic industry profile
Province Dak Lak Gia Lai Lam Dong Year 2012 2013 2014 2012 2013 2014 2012 2013 2014 Physical production factors Coffee yield (ton/ha) 3.14 3.18 3.03 3.74 3.71 3.69 2.62 2.67 2.74 Area per farm (ha) 1.16 1.17 1.18 1.62 1.73 1.76 2.06 2.03 2.04 Family labour (man-days) 209.23 209.38 208.05 250.40 247.13 243.00 198.68 202.69 202.36 Hired labour (man-days) 35.82 36.79 38.32 82.11 82.06 79.43 44.76 45.81 49.41 Chemical fertilisers (tons) 2.09 2.20 2.15 2.83 2.84 2.80 2.45 2.58 2.63 Irrigation water (m3) 1.31 1.30 1.32 1.24 1.21 1.24 0.96 0.93 0.94 Financial performance measures (million VND per hectare) Gross income 120.89 122.71 113.48 144.65 144.72 139.88 102.36 101.07 106.12 Net income 87.63 88.59 78.43 95.12 94.51 87.51 65.54 63.39 66.08 Production cost breakdowns (million VND per hectare) Total direct production cost 33.25 34.13 35.05 49.54 50.21 52.37 36.82 37.68 40.04 100% 100% 100% 100% 100% 100% 100% 100% 100% Hired labour cost 4.96 5.34 5.71 12.10 12.32 11.91 6.58 7.08 7.96 15% 16% 16% 24% 25% 23% 18% 19% 20% Chemical fertiliser cost 19.59 20.49 20.30 24.83 25.45 25.97 22.27 23.10 23.96 59% 60% 58% 50% 51% 50% 60% 61% 60% Organic fertiliser cost 3.62 3.13 3.85 7.76 7.56 9.10 2.93 2.71 3.38 11% 9% 11% 16% 15% 17% 8% 7% 8% Irrigation cost 3.64 3.63 3.67 3.49 3.34 3.85 3.16 2.93 2.83 11% 11% 10% 7% 7% 7% 9% 8% 7% Pesticide cost 0.61 0.61 0.67 0.86 0.89 1.01 0.93 0.93 1.01 2% 2% 2% 2% 2% 2% 3% 2% 3% Machinery cost 0.84 0.93 0.85 0.49 0.65 0.53 0.93 0.93 0.89 3% 3% 2% 1% 1% 1% 3% 2% 2% Table 4.3: Economic profile of coffee production over crop years
______Chapter 4: Socio-economic industry profile
An examination of the total direct costs of coffee production in the three surveyed crop years in the three provinces indicates a modest variation over time and over different geographical locations. On average, the total direct cost of coffee farming slightly increased over the three crop years in all provinces. Figure 4.4 also shows that farmers in Gia Lai were likely to shoulder the highest direct production cost level, although those in Dak Lak and Lam Dong faced has only slightly lower direct production costs. The variation in this cost between certified and non-certified farms in
Gia Lai and Lam
Dong
provinces was larger than that in Dak Lak (Figure 4.4).
______Chapter 4: Socio-economic industry profile
Figure 4.4: Direct production cost per ha in million VND (excluding family labour cost)
In terms of irrigation water, the sampled coffee farmers in Dak Lak and Gia Lai clearly consume more water than those in Lam Dong (also see Figure 4.5). Given irrigation demands large inputs of labour energy, the amount of irrigation water is positively associated with labour and irrigation cost. Thus coffee farmers in Lam Dong used less labour and spent less on irrigation. On the other hand, water extraction may negatively affect environment and soil quality. Thus, the irrigation factor involves in both economic and environmental issues.
In terms of financial performance, both gross average income and average net income tended to decrease over time for Dak Lak and Gia Lai, contrasting with a slight increase for Lam Dong. Average direct production cost per hectare increased in all provinces over the three crop years which explains the decrease in net income for coffee farmers over time for Dak Lak and Gai Lai.
Table 4.3 also sets out the production cost structure and changes in cost components over the three crop years in the three provinces. Overall the structure of coffee production cost predictably changed little over the limited time frame of three crop years. There were nevertheless some differences in cost components among the three provinces. The average coffee farmer in Dak Lak had a lower proportion of hired labour cost than those in Lam Dong and Gia Lai. Overall, coffee farmers demanded slightly more hired labour over time which on average accounted for about 15%, 25% and 20% of total direct costs for farmers in Dak Lak, Gia Lai and Lam Dong respectively.
______Chapter 4: Socio-economic industry profile
It is clear that coffee farmers in the three provinces used different proportions of inorganic and organic fertilisers. Note that the predominant production input, chemical fertilisers, stayed stable over time within a province, although it varied across different provinces. The sampled coffee farmers in Dak Lak and Lam Dong used a higher proportion of chemical fertilisers (i.e., about 60% of total direct costs on average) than those in Gia Lai (i.e., about 50% of total direct cost on average). In contrast, the sampled coffee farmers in Gia Lai applied more organic fertilisers in terms of both absolute values and proportion (the proportion varying over time between 15% and 17%) than those in the other two provinces (where the proportion over time varied from 7% to 9% in Lam Dong and from 10% to 12% in Dak Lak). Generally, over the three crop years, coffee farmers tended to increase the proportion of organic fertilisers in their cost structure.
The components making up the rest of the costs which included irrigation, pesticides, and machinery, remained stable over the three crop years within each province. The irrigation cost, which is composed of fuel and electricity for water- pumping, was the highest for farmers in Dak Lak (11% of total average direct costs) and averaged only 7% for the other two provinces. The other two cost items, pesticide and machinery, accounted for only between 3% and 5% of the direct production costs in all provinces and across the three crop years.
Table 4.4 provides a snapshot comparison between the sampled certified and non- certified farms with respect to coffee production factors, financial performance indicators and production cost breakdowns. As indicated in the last two columns of
Table 4.4, overall, the sampled certified coffee farms had a larger production scale than
______Chapter 4: Socio-economic industry profile
the non-certified farms. The results indicate that sustainability certified farms achieved higher coffee yield and were more intensive in the use of inputs (i.e., land, fertilisers, irrigation water) than the non-certified farms. As a result, certified coffee farms achieved a better financial outcome (i.e., gross income and net income from coffee production) than those of non-certified farms.
It is apparent from the survey that coffee production in Vietnam is heavily dependent on chemical fertilisers. For both non-certified and certified farms chemical fertiliser costs accounted for 63% and 52% of total direct production costs respectively.
Further examination of the data indicates that in all provinces use of chemical fertilisers in certified farms accounted a significantly lower proportion of total costs that non- certified farms. In monetary terms certified farmers in all three all three provinces spent slightly less on chemical fertilisers than their non-certified counterparts. However, certified farms in Lam Dong and Dak Lak used a marginally greater volume of chemical fertilisers than non-certified farms while in Gia Lay slightly less was used. Overall then, although there was no significant difference in the volume of chemical fertilisers used between the certified and non-certified farms, in terms of percentage of overall costs certified farms were measurably less dependent on chemical fertilisers than the non- certified farms.
Regarding the use of organic fertilisers, the variety of organic fertilisers used by surveyed farmers differed considerably in terms of weight, nutrient content and unit cost. Therefore, we used the organic fertiliser cost as an aggregated organic fertiliser index, instead of a physical factor of production. Generally, certified coffee farms
______Chapter 4: Socio-economic industry profile
consumed more organic fertilisers (i.e., had a higher organic fertiliser cost) than non- certified farms, except for those in Lam Dong.
______Chapter 4: Socio-economic industry profile
Table 4.4: Economic profile of non-certified vs certified farms
Dak Lak Gia Lai Lam Dong Total Non- Non- Non- Non- Certified Certified Certified Certified certified certified certified certified Physical production factors Coffee yield (ton/ha) 3.12 3.11 3.21 3.84 *** 2.76 2.52 *** 2.95 3.19 *** Area per farm (ha) 1.05 1.28 *** 1.37 1.79 *** 1.79 2.52 *** 1.44 1.74 *** Family labour (man-days) 224.07 195.52 *** 295.36 234.18 *** 215.47 174.59 *** 225.8 202.06 *** Hired labour (man-days) 28.43 44.51 *** 74.39 82.93 35.28 68.1 *** 35.62 61.85 *** Chemical fertiliser (tons) 2.14 2.15 3.07 2.76 2.47 2.72 ** 2.38 2.47 * Irrigation water (m3) 1.21 1.40 *** 1.20 1.24 0.99 0.86 *** 1.10 1.22 *** Financial performance measures (million VND per hectare) Gross income 117.49 120.28 125.14 147.68 *** 104.03 101.67 111.54 123.99 *** Net income 86.19 83.62 80.27 95.43 *** 66.51 62.21 76.05 81.95 *** Production cost breakdowns (million VND per hectare) Total direct production cost 31.31 36.65 *** 44.87 52.25 *** 37.52 39.46 35.48 42.04 *** 100% 100% 100% 100% 100% 100% 100% 100% Hired labour cost 4.07 6.45 *** 11.25 12.33 5.42 10.58 *** 5.33 9.23 *** 13% 18% 25% 24% 14% 27% 15% 22%
Chemical fertiliser cost 20.69 19.63 *** 25.71 25.36 23.36 22.67 22.42 22.10 66% 54% 57% 48% 62% 58% 63% 52% Organic fertiliser cost 2.58 4.38 *** 3.91 9.25 *** 3.36 2.36 *** 3.07 5.35 *** 8% 12% 9% 18% 9% 6% 9% 13% Irrigation cost 2.7 4.48 *** 2.86 3.75 ** 3.31 2.34 *** 3.01 3.74 *** 9%12%6%7%9%6%8%9% Pesticide cost 0.51 0.74 *** 0.89 0.93 1.10 0.71 *** 0.83 0.79 2% 2% 2% 2% 3% 2% 2% 2% Machinery cost 0.76 0.97 *** 0.25 0.64 *** 0.98 0.80 * 0.82 0.83 2% 3% 1% 1% 3% 2% 2% 2% Pair wise t-tests were performed. ***, ** and * refer the 99%, 95% and 90% significance level respectively.
______Chapter 4: Socio-economic industry profile 57
______Chapter 4: Socio-economic industry profile 58
In terms of the use of irrigation water and its cost, the differential effect between certified and non-certified farming is unclear (see Figure 4.5) If only this single indicator is used to examine the effect of certification schemes, the result should be interpreted with caution. There was considerable variation from province to province in terms of both water consumption and cost. Certified farms in Dak Lak tended to consume more irrigation water than the non-certified farms while in Lam Dong, certified farms used less than non-certified farms and Gia Lai the difference was not significant. Figure 4.5 also indicates that the difference in the volume of irrigation water used between crop years in a province was not particularly marked, although location of farms within a province may be an important factor affecting irrigation water consumption.
Figu re 4.5: Irrig atio n wate r per ha (measured in m3) In terms of overall cost, it is clear that certified farms have a higher level that non- certified farms. That reflects generally higher costs for labour (i.e. greater use of family labour), irrigation and organic fertiliser costs (the exception being for non-certified farmers from Lam Dong who, on average, had lower costs than their certified
______Chapter 4: Socio-economic industry profile
counterparts for both organic fertiliser and irrigation). The cost of pesticides and machinery accounted for only a small proportion of total production cost, although some of these are statistically different between the certified farms and non-certified farms.
4.4 Conclusion
The aim of this chapter was to develop a socio-economic profile of the Vietnamese coffee farming industry. We used a descriptive approach to investigate data from one of the largest surveys made of coffee farming in Vietnam. The result provides a broad array of primary information on the socio-economic characteristics of Vietnamese coffee farmers.
This socio-demographic profile offers information on age, family size, ethnicity and education of the household heads. The findings indicate that ages of the coffee farmers are distributed between 20 and 80 years. Most are aged between 40 and 50 years, with an average of 48 years. The family size of households averaged 4.54 people while the majority of households - 60% - were between 3 and 5 people. 82% were from the Kinh group, the majority ethnic group in Vietnam. The rest belonged to variety of other ethnic groups such as Ede, K’ho, Tay, Nung and M’nong. Household heads’ education usually did not reach secondary school completion.
The economic profile of coffee farming in Vietnam provides information on farming scale, main production factors and the resulting issues facing the farming industry. Coffee farming in Vietnam is typically small-scale heavily dependent on chemical fertilisers. Given coffee is an irrigation water intensive crop a number of issues arise relating to indirect economic cost burdens and in particular those relating to
______Chapter 4: Socio-economic industry profile
environmental degradation. In dealing with these concerns, some half of the sampled coffee farms participated in sustainability certification schemes.
______Chapter 4: Socio-economic industry profile
Diversification and Productive Efficiency9
Chapter 5 offers answers for following questions, (1) What are technical efficiency level of coffee production; (2) Given different farming systems, monoculture, segregation and synchronisation, which farming systems are efficient? and (3) What are possible reasons explaining the existence of efficient systems, i.e., presence of agronomic benefits, and inefficient systems, i.e., food security?
5.1 Introduction
Coffee production is one of the primary economic sectors in the Central Highlands region of Vietnam with nearly 96% of Vietnam’s export of coffee coming from this region. Due to significant price increases in the early 1990s, the area used for coffee cultivation increased by approximately 400% from 1999 to 2000. This expansion appears to be a natural adaptation of farmers in response to past increases in prices.
However in subsequent periods the resulting increases in market supply caused prices to drop to a level which, by 2001, was lower than the production cost (Marsh, 2007). This forced many coffee farmers into bankruptcy (Wollni & Zeller 2007) and is seen as one of the reasons motivating coffee farmers to diversify their business and for the presence of several distinct farming systems in Vietnam.
In this study, we consider three typical coffee growing systems in Vietnam: mono- cropping, synchronization and segregation. The nature of specialization and
9 This chapter has been published in Economic Analysis and Policy, which can be found at http://dx.doi.org/10.1016/j.eap.2017.09.002. Only minor changes/ edits have been done.
______Chapter 5: Diversification and Productive Efficiency
diversification vary significantly across these three distinct systems. Mono-cropping farms have only one land plot and grows only coffee. Segregated farming systems have more than one plot of land with each plot growing one primary type of crop. For example, where farms have two plots, one plot grows coffee and another plot grows rice.
Synchronized farming systems grow coffee together with other industrial crops in one plot and rice in a separate plot(s). There is an obvious need to know which farming system are most efficient for coffee farmers in Vietnam - an issue which the existing literature provides no empirical evidence. This literature gap is, therefore, the primary motivation of the present chapter.
It is noted that diversified systems, particularly through crops diversification, may obtain higher yields and/or cause less environmental damage (Letourneau et al., 2011).
This is known as complementary or synergy effects among crops sharing the same environment. However, there may be diseconomies of scope or negative effects of synergy as empirically observed in, for example, Coelli and Fleming (2004). Since appropriate crop diversification strategies can deliver positive effects of synthesis, it could be expected that coffee farmers would be motivated to diversify by growing industrial crops such as pepper or durian which may also mitigate market risk. However, it becomes less clear why coffee farmers have chosen the segregated system in which rice and coffee are grown in separate land plots given it does not deliver positive synthesis effects and could produce diseconomies of scope (Villano, et al., 2010). More particularly, some studies have hypothesized that Vietnamese coffee farmers diversify to rice because of the insecurity created by low incomes and volatile market conditions
(i.e., Dang 2003). If this is true, segregated farms face a trade-off between productive
______Chapter 5: Diversification and Productive Efficiency
efficiency and income or poverty risks. In this chapter, we aim to provide empirical evidence on this trade-off hypothesis. Our empirical results are therefore designed to indicate the rationale for Vietnamese farmers’ decisions over which crops to grow rather than accepting that they may be made on ad hoc or irrational basis (Dang &
Shively, 2008). As such, this study can provide a useful guide for policy makers in raising the productivity of Vietnamese coffee farmers.
We utilize several techniques to examine differences in the level of productive efficiency among the three farming systems using a dataset of 167 farms surveyed in five Central Highlands communes in 2012. The input distance function is used to estimate efficiency scores for each farm. Parametric and non-parametric tests are then applied to assess if these differences are statistically different across the farming systems of the three districts. Additionally, the input distance function allows us to examine economies of diversification which is based on the concept of economies of scope in diversified farms (Baumol et al., 1988; Willig, 1979). While there are several approaches to measuring scope economies (Chavas and Di Falco, 2012; Chavas, 2011;
Ofori-Bah and Asafu-Adjaye, 2011; Chavas and Kim, 2010; Hajargasht et al., 2006), we use Coelli and Fleming's (2004) model as it does not require price information and provides a more straightforward interpretation of both efficiency results and diversification economies of each pair of crops.
The remaining part of this chapter is set out in eight sections. Section 4.2 provides a literature review. Section 4.3 provides a measure of economies of diversification using the distance function. Section 4.4 sets out the empirical models, data sources and the use of relevant variables. Survey and descriptive statistics are presented in Section 4.5.
______Chapter 5: Diversification and Productive Efficiency
Section 4.6 provides the empirical results. Section 4.7 discusses the presence of agronomic benefits and the way in which they are translated into efficiency improvements and provides an explanation of why coffee farmers still choose rice.
Section 8 sets out the conclusions, policy recommendations and avenues for further study.
5.2 Literature review
The various dimensions of farming management practices are well captured in the literature (Bell and Moore 2012). In particular the farming system in which crops use the same resources, i.e., water and nutrients simultaneously, is known as an intercropping system or synchronization (van Asten et al., 2011). Another common farming integration system is crop rotation: however, it is not applicable to perennials such as coffee and other industrial crops. In addition, segregated systems are known as integration of spatially separated crops. This farming practice is found to be attractive to smallholder farmers cultivating both subsistence crops and cash crops (Solís et al.,
2009). For example, in the Central Highlands of Vietnam, coffee is a dominant crop and farming is mostly small scale (Luong and Tauer, 2006) mixed with some diversified subsistence crops, i.e., rice (Doutriaux et al., 2008). Therefore, by examining the economic benefits of different farming practices, i.e., crop specialization (not integrated organizationally), segregation (only integrated organizationally) and synchronization
(integrated organizationally and spatially and temporally) (Bell and Moore, 2012), it is intended to make a useful contribution to the farming management literature.
______Chapter 5: Diversification and Productive Efficiency
Crop diversification in synchronized systems has, on the one hand, the potential to deliver agronomic and ecological benefits; however realizing these potential benefits depends on the characteristics of ecosystems and the choice of crops (Bacon 2005; Dang and Shively 2008). On the other hand, there are little or no agronomic benefits from crop segregation, although this type of farming system may have other desirable outcomes in terms of food security and allocation of inputs (Bell and Moore, 2012).
There is a rich literature on various synchronized systems of crop diversification
(i.e., Rahman 2009; Kim et al. 2012), but only a few studies examine coffee farming
(i.e., van Asten et al. 2011) and no study compares the productive efficiency between synchronized and segregated systems. For synchronized systems, it can be expected there will be a direct transformation of agronomic and ecological benefits into economic benefits through reductions in consumption of inputs without sacrificing output levels or through increasing output levels without requiring more input consumption. For example, the agronomic literature has identified crops such as avocados and fruit trees as being suitable for cultivation with coffee (Borkhataria et al., 2012) joint-production of which can result in less fertilizer being required. It is noted that lower fertilizer consumption delivers both a cost reduction and a reduction in negative environmental impacts. But synchronized systems may require greater management attention (Bell and
Moore, 2012). However, growing different crops in different land plots in segregated systems may not deliver benefits of synchronization and in many situations segregation exposes farms to a higher risk of productive inefficiency. This comes about through misallocation of resources as farmers maybe inefficient in allocating limited resources among different plots growing different types of crops (Bell and Moore, 2012).
______Chapter 5: Diversification and Productive Efficiency
Although benefits of diversification are often highlighted in the literature, evaluating these benefits in economic terms is not straightforward. One reason is that a relationship between agronomic or ecological benefits and monetary benefits is not straightforward because of the diversity of impacts of input and output prices on monetary outcomes accrued to farmers, consumers and society as a whole. However, it can reasonably be expected that there is a direct transformation of agronomic benefits into improvements in productive efficiency. Therefore, we focus on this direct transformation and use these efficiency changes to quantify the impacts of diversification in coffee growing.
Coelli & Fleming (2004) suggest that diversification economies derived from an extra unit of one output can be measured by increases the marginal efficiency level of producing an extra unit of another output, holding other variables constant. This increase in marginal efficiency is due to deflating the amount of inputs to the extent that it puts the observed farm closer to the production frontier. The existence of diversification economies, therefore, implies that joint production of two outputs can improve the productive efficiency in comparison to the situation where there is separate production of two specialized outputs. An empirical study employing the input distance function by
Coelli & Fleming (2004) is based on a small panel data set of 18 coffee smallholders collected in 1992 and 1993 in six villages in Papua New Guinea. They examine diversification economies of pairs of crops, i.e., coffee and subsistence food, coffee and cash food and subsistence and cash food production. The authors report weak empirical evidence of diversification economies between coffee and subsistence food production.
______Chapter 5: Diversification and Productive Efficiency
As well, diseconomies of diversification from the combination of coffee and cash food production are identified, implying a negative impact of diversification on efficiency.
Vedenov et al. (2007) uses a data set of 24 coffee-producing districts in Mexico covering the period 1997 to 2002 employing the input distance function approach to estimate technical efficiency of the districts. The authors use Coelli & Fleming’s (2004) method of measurement to examine economies of diversification for three different pairs of crops (coffee-corn, coffee-other cash crops and corn-other cash crops). However, their calculations of diversification (dis)economies are not reliable given inappropriate use of the derivation formulae (see discussion in Section 3). Our literature review finds two empirical studies - that of Coelli & Fleming (2004) and Vedenov et al. (2007) - which focus on measuring diversification economies in coffee production. However, they do not offer a comparative analysis of differing farming systems.
Empirical analysis comparing efficiency of differing farming systems is particularly important in the context of coffee production in Vietnam. Coffee growing is the primary economic activity in the Highlands region of Vietnam, where most coffee is produced. In response to favorable coffee prices in the early 1990s, farmers increased cultivation of coffee leading to increased supply and downward pressure on coffee prices. Coffee farmers, then appear to have to diversify into rice cultivation and other industrial crops. To-date, there are three distinct farming systems for small farmers in the Highland regions of Vietnam: mono-cropping, synchronization and segregation.
However there are few empirical studies on the productive efficiency of coffee farming in Vietnam (Garcia & Shively, 2011; Cheesman & Bennett, 2008) none of which compare efficiency across the different types of farming systems.
______Chapter 5: Diversification and Productive Efficiency
In summary, we identify two important gaps in the empirical literature on coffee farming. First, there is no empirical study examining the efficiency effects brought by different diversification strategies of coffee farmers (i.e. different crop mixtures such as segregation versus synchronization). Second, there is no previous efficiency study focusing on the three typical farming systems for coffee cultivation in Vietnam - the world’s second largest coffee producing country. Our study aims to fill these two gaps by empirically examining the efficiency benefits of diversification in the framework of input distance functions. In doing so we aim to provide both farmers and local agricultural extension officers new information which can lead to greater efficiency in agricultural practices.
5.3 Input distance function
The measure of diversification proposed by Coelli & Fleming (2004) in the framework of input distance function has several advantages. First, it is applicable to cases where data on input prices are not available. Second, this measure allows the derivation of complementary effects directly from the distance function. Third, an inefficiency model can also be estimated to examine factors that can explain variations in the technical efficiency across farms. Following Coelli & Fleming (2004), the input distance function is defined as:
x d(,x yD )=∈ : Ly (), 5.1 D
where L(y) refers to the set of all input vectors x that can produce the output vector y. The distance function, d(x,y), is non-decreasing in x, and non-increasing in y, and
______Chapter 5: Diversification and Productive Efficiency
linearly homogeneous and concave in x. If x belongs to the input set of y (i.e., x ∈ L(y)) then the value of the distance function is equal to or greater than one (d(,xy )≥ 1). The distance is equal to unity if x belongs to the isoquant of y. That is, the firm is said to be technically efficient or inefficient if the value of the distance equals one or exceeds one respectively. Note that the value of the distance equals the inverse of the traditional input-oriented technical efficiency score proposed by Farrell (1957).
Coelli & Fleming (2004) suggest that diversification economies produced from an extra unit of output i increases the marginal effici ency level of producing an extra unit of output j, holding other variables constant. That is, the first partial derivative of the distance function with respect to the ith output is the marginal distance of producing an extra unit of the ith output. The cross derivative of the distance with respect to the ith and jth output represents the effect of change in one additional unit of the jth output on the marginal distance of producing ith output10. The existence of diversification economies implies joint production of two outputs – the ith and the jth - delivering higher input-orientated efficiency. That is, there is less input consumption with the given level
10 This is also similar to the utility theory which states the cross partial derivative of the utility function with respect to good X and good Y is positive if these two goods are complements in
∂∂∂2 ==UU > consumption (U XY 0 ). This means that consuming more of good Y increases ∂∂XY ∂ X ∂ Y the marginal utility of good X.
______Chapter 5: Diversification and Productive Efficiency
of output than that of producing two outputs separately. Economies of diversification11 between two outputs is therefore observed if:
∂∂∂<=2 Dyyijq(x, y) /ij 0, , 1,...., 5.2
Note that a positive cross derivative with respect to the ith and jth output implies diseconomies of diversification12.
We also examine the absolute values of the cross derivative to infer the relative magnitudes of the diversification economies. This provides an indication of the relative effects of diversification economies when there are more than two outputs (i.e. more than one pair of two outputs i and j). More specifically, if economies of diversification are observed, the most favourable combination (i.e. largest marginal efficiency gain due to joint production) is quantified as:
11 There are both similarities and differences between the term ‘diversification economies’ used in this study and the term ‘output complementarities’ used in other studies. Villano et al. (2010) define output complementarities as a marginal increase of one output as the result of increasing the quantity of another output. However, the cross derivative of the input distance function in expression (2) with respect to the ith and jth output does not match this definition of output complementarities.
12 ∂∂∂≠∂∂∂=22 β Note that DyyD(x, y) /ij ln ( x, y) / ln yy i ln j ij. In Vedenov et al. (2007) only used interaction between only two outputs (βij) in calculating this measures while interaction term (βi βj+
βij) should be used. More details are provided upon an email request.
______Chapter 5: Diversification and Productive Efficiency
∂ 2 D Max{}, i≠ j and ij,= 1.... q for the largest efficiency gain due to joint ∂∂ YYij production. 5.3
If diseconomies of diversification are observed, the combinations that have minimum negative effect on the efficiency loss due to joint production is:
∂ 2 D Min{}, i≠ j and ij,= 1.... q for the least efficiency loss due to joint ∂∂ YYij production. 5.4 Since inputs and outputs are normalized at their means, for more meaningful interpretations of the empirical results of the average farm, the expression (5.2) can be represented as13:
∂2 DD 1 Δ==D Δ= ()0ββ +< β 5.5 YYij|1∂∂ i j ij YYij YY ij 2
Theoretically, this distance D is always positive; therefore, from expression (5.5)
1 we see that ()0ββ+< β implies evidence of the diversification economies. ij2 ij
5.4 Empirical model specifications
A translog input distance function is used in this study to characterize the production technology as it is more flexible than the Cobb-Douglass functional form and does not require restrictions of rectangular hyperbola isoquants (Morrison-Paul et al.,
2000). The fully specified translog input distance function is specified as:
13 See Appendix 1 for the derivation of the cross derivative of the input distance function with respect to output i and output j.
______Chapter 5: Diversification and Productive Efficiency
qqqp pp I =+αβ +11 β + α + α lnDyikkkkkkmmmmmm0'''' ln ln yyxxx ln ln ln ln kkkmmm====122 1 '1 1 == 1 '1 pq + 1 τ kmlnxy m ln k 2 mk==11 5.6 where i = 1 …. N refers to the number of firms, q and p represent the number of outputs and inputs.
According to O’Donnell & Coelli (2005), the homogeneity of degree +1 in inputs implies:
pp p αα==τ = mmm1,' 0, and km 0 5.7 mm==11 m=1
Plugging the constraints in Equation (7) into Equation (6), we have:
qqqp I =+αβ +1 β + α ln(Dxikkkkkkmm /10 ) ln y ' ln y ln y ' ln( xx / 1 ) kkkm====11'122 pp pq 5.8 ++11ατ mm'1'1ln(x m /xxx )ln( m / ) mk ln( xxy m / 1 )ln k 22mm==2'2 m == 2 k 1 ββ= αα= The symmetry requires that kk'' k k and mm'' m m .
I i i Following Coelli & Perelman (1999), the “-ln( D i )” term is set to be v - u , and re-arranging Equation 2.8 gives an input distance function as:
qqqp −=+αβ +1 β + α lnxy10kk ln kkkk ' ln yyxx ln ' m ln( m / 1 ) kkkm====11'122 pp pq 5.9 +++−11ατ mm'1'1ln(/)ln(/)x mxxx m mk ln(/)ln xxyvu m 1 k i i 22mm==2'2 m == 2 k 1
where y refers to output; vi, captures the effects of statistical noise which is
assumed to be independently and identically distributed as N (0, ). ui are non- negative random variables associated with technical inefficiency in production and ______Chapter 5: Diversification and Productive Efficiency
which are assumed to be independently and identically distributed such that u is defined
by the truncation at zero of the normal distribution with unknown variance and
unknown mean, μ, [ui ~ (|N(0, )|)].
We follow Battese & Greg (1977) in calculating and explaining the variance
σ 2 parameters, and as γ = u and σσσ222=+ for the model. The parameter σ 22+σ svu uv
γ is ranged between zero and one, where γ = 1, there is no random noise and where
γ = 0, technical inefficiency is not present. The input distance is presented as Di =
E[exp(ui)|ei], where ei = vi – ui.
The inefficiency model is specified as:
L μδ=+ δ 0 llz 5.10 l=1
where zs are exogenous variables associated with the production environment. This explains how the observed output level can be produced using the minimum level of inputs. However, inefficient firms utilize the actual level of inputs which is due to the technical inefficiency effect. It is assumed that these factors indirectly affect the production technology so that the deterministic frontier does not capture these variables.
Instead, they are treated as factors explaining inefficiency variation of the stochastic input distance function. These independent variables are known as factors characterized by the production technology and relate to input and output production factors.
______Chapter 5: Diversification and Productive Efficiency
5.5 Farming systems and data collection
Figure 5.1 Three distinct farming systems As depicted in Figure 5.1, the sample is categorized into coffee specialized farms and diversified farms with more than one farming enterprise. There are two main types of diversified farms. The first are farms where different crops share or co-locate in the same plot of land. This is called a synchronized farming system, i.e., coffee – other industrial crops. The second type are farms are characterised by different crops being planted in separated plots of land. This is called a segregated farming system (also see
Bell and Moore 2012).
This survey was conducted in five communes in the Cu M’gar District, Dak Lak
Province of the Central Highlands, in 2012 for the crop year 2010/11. In recent years,
Vietnam produces about 20% of global coffee production of which the Central
Highlands region contributes about 85%. Of the five provinces, Dak Lak is the largest coffee-growing region in terms of both cultivated area and production with about 50% of total national coffee production. The Cu M’gar District where our data is derived from is known as a key coffee planting area with diversified farm models (also see: Dang &
Shively, 2008). In addition to coffee, other cash crops (avocado, pepper and fruit) and short gestation industrial and food crops (i.e., rice) are also cultivated. ______Chapter 5: Diversification and Productive Efficiency
The data collection procedure involved a two-stage random sampling technique.
Firstly, five out of 13 communes in the district were randomly identified. Given the distribution of coffee farmers in the five selected communes was relatively equal, around
35 households in each commune were randomly selected. This procedure was used to ensure geographical representation of farmers with different production conditions across the district. In addition, this procedure also allowed the survey to be conducted to meet cost and time constraints. In total, 167 farmers were interviewed using the face-to- face technique with a structured questionnaire set. The questionnaire sought information on demographics, household characteristics, input and output data, and socio-economic and geographical profiles relating to agricultural production. Prior to the main survey, a pre-test was conducted for the purpose of evaluation and refinement of the instrument.
Descriptive statistics of all variables used in Equations (4.6) to (4.10) are set out in Table
5.1.
In the Central Highlands there are representatives of most Vietnamese ethnic groups working in the agricultural production sector. The Kinh group is in the majority with other groups classified as being local or migrated such as Ede, Mnong, and Tay. In the present study, a dummy for ethnicity is used as we expect there are differences in both farming practice and other socio-economic conditions between and the Kinh households and others minorities (Dang & Shively, 2008; van de Walle &
Gunewardena, 2001).
______Chapter 5: Diversification and Productive Efficiency
Std. Variable Variable definition Obs Mean Dev. Min Max Outputs: y1 Coffee production in tons 167 1.27 1.34 0.00 8.00 y2 Rice production in tons 167 0.95 0.92 0.00 4.60 Income from the other y3 industrial crops in million 167 10.25 15.23 0.00 75.00 VND Inputs: x1 Cultivating area in ha 167 0.69 0.48 0.10 3.50 Chemical fertilizers applied in x2 167 1.04 1.15 0.00 7.37 tons Irrigation water in thousand x3 167 1.78 1.93 0.00 12.00 cubic meters x4 Total labour used in man-days 167 147.37 110.19 0.00 660.00 Other production cost in x5 167 2.02 3.50 0.00 25.00 million VND Socio-economic characteristics: Ethnicity of the household z1 head, 1 if Kinh majority and 0 167 0.43 0.50 0.00 1.00 for otherwise Number of years the z2 household head completed 167 5.75 4.05 0.00 12.00 formal education Amount of credit loans to the z3 167 21.34 20.91 0.00 115.00 household in million VND Farming experience of the z4 household head measured in 167 8.44 5.46 0.00 27.00 years Family size measured in z5 167 3.96 1.70 1.00 9.00 persons Crop diversification (number Diversity 167 of farms) d=0 Coffee mono crop farms 20 d=1 Coffee and rice mixed farms 41 Coffee and the other industrial d=2 23 crops mixed farms Rice and the other industrial d=3 12 crops mixed farms d=4 Other combinations of crops 71 Table 5.1: Descriptive statistics of variables
______Chapter 5: Diversification and Productive Efficiency
Household characteristics are often included in the technical inefficiency model in empirical studies of smallholders farming. The common independent variables include: formal education level of household heads (Picazo-Tadeo et al., 2011; Illukpitiya &
Yanagida, 2010), formal credit loans (Binam et al., 2004), crop diversification (Chavas
& Kim, 2010; Illukpitiya & Yanagida, 2010), farming experience and age of household head (Cinemre et al., 2006; Ofori-Bah and Asafu-Adjaye 2011); family size in number of persons (Bozoğlu & Ceyhan, 2007; Haji, 2006).
In terms of crop diversification, dummy coding is used for ‘Diversity’ to examine if the efficiency level varies between coffee mono crop farms (the base group) and farms having different multiple crops. If those dummy variables are statistically significant, we can interpret the extent of differences in the average efficiency levels of various groups of farms.
5.6 Empirical results
5.6.1 Maximum likelihood estimates
The result’s maximum likelihood estimation (MLE) of the stochastic input distance function is presented in Table 5.2. The statistically significant negative signs of the first order coefficients of outputs imply an inverse relationship between the input distance and output quantities; that is, more outputs result in a smaller distance or a higher efficiency level. The estimated coefficients, β1= -0.4956 and β2= -0.2875, indicate that an increase of coffee and rice production of 10% will decrease the distance to the frontier by about 4.956% and 2.875% respectively, ceteris paribus. These results indicate the greater capacity of coffee production relative to rice production in
______Chapter 5: Diversification and Productive Efficiency
enhancing the overall input-orientated efficiency level. The estimated coefficients associated with coffee and rice also predict the same finding. Table 5.2: MLE for the stochastic input distance function
Variable Parameter Coef. Std. err. Constant α0 0.2325 *** 0.0355 lny1 (coffee) β1 -0.4956 *** 0.0354 lny2 (rice) β2 -0.2875 *** 0.0298 lny3 (other industrial crops) β3 0.0508 ** 0.0245 2 0.5(lny1) β11 -0.0125 *** 0.0012 0.5lny1 * lny2 β12 0.0061 *** 0.0014 0.5lny1*lny3 β13 -0.0034 *** 0.0010 2 0.5(lny2) β22 -0.0095 *** 0.0010 0.5lny2 * lny3 β23 0.0000 0.0001 2 0.5(lny3) β33 0.0017 ** 0.0008 ln(x2/x1) (chemical fertilizers/ area) α2 0.0836 * 0.0483 ln(x3/x1) (Irrigation water/ area) α3 0.0856 *** 0.0311 ln(x4/x1) (Labour/ area) α4 0.1316 * 0.0772 ln(x5/x1) (Other cost/ area) α5 0.0620 *** 0.0175 2 0.5[ln(x2/x1)] α22 0.0074 *** 0.0021 ln(x2/x1)* ln(x3/x1) α23 0.0077 0.0234 ln(x2/x1) *ln(x4/x1) α24 -0.0552 *** 0.0177 ln(x2/x1) * ln(x5/x1) α25 -0.0034 0.0035 2 0.5[ln(x3/x1)] α33 -0.0027 ** 0.0012 0.5ln(x3/x1) * ln(x4/x1) α34 0.0055 * 0.0033 0.5ln(x3/x1) * ln(x5/x1) α35 0.0161 *** 0.0055 2 0.5[ln(x4/x1)] α44 -0.0029 0.0024 0.5ln(x4/x1) * ln(x5/x1) α45 -0.0350 *** 0.0069 2 0.5[ln(x5/x1)] α55 0.0023 *** 0.0006 ln(x2/x1) * lny1 τ21 -0.0098 0.0230 ln(x2/x1) * lny2 τ22 0.0033 * 0.0019 ln(x2/x1) * lny3 τ23 -0.0057 *** 0.0015 ln(x3/x1) * lny1 τ31 0.0029 * 0.0016 ln(x3/x1) * lny2 τ32 0.0027 * 0.0014 ln(x3/x1) * lny3 τ33 0.0035 *** 0.0010 ln(x4/x1) * lny1 τ41 0.0018 0.0053 ln(x4/x1) * lny2 τ42 0.0017 0.0033 ln(x4/x1) * lny3 τ43 0.0071 *** 0.0026 ln(x5/x1) * lny1 τ51 -0.0148 *** 0.0048 ln(x5/x1) * lny2 τ52 0.0015 ** 0.0007
______Chapter 5: Diversification and Productive Efficiency
ln(x5/x1) * lny3 τ53 -0.0002 0.0002 Note: Three, two and one asterisks indicate significance at the 1%, 5% and 10% levels, P respectively. ositive and statistically significant coefficients for inputs represent the complementarity of these inputs with the land input. Larger farms are likely to invest more in all input factors on average. In addition, most of the coefficients of the interaction-variables are significantly different from zero. This implies non-linearities in the production technology and hence justifies the use of the flexible translog specification.
2 The Chi-squared, χ = 1948.19 represents the effect of technical inefficiency
σ = which is statistically significant at the 99% confidence level (H0: u 0, was rejected at
1% level of significance). Since all production variables are mean-corrected prior to the estimation, the inverse of the sum of the output coefficients provides a measure of radial return to scale elasticity (Coelli & Fleming, 2004; Hajargasht et al., 2008). The empirical result, -1/(β1+β2+ β3) = 1.37 suggests the existence of scale economies.
5.6.2 Inefficiency model
The empirical results suggest an average level of input-orientated technical efficiency of around 82%. This implies that farmers could reduce inputs by 18% (i.e. 1 –
0.82) with the output levels remaining unchanged. The estimated results for the inefficiency model (Equation 10) are reported in Table 5.3.
______Chapter 5: Diversification and Productive Efficiency
Table 5.3: Inefficiency model
Variable Parameter coef. Std. err. Constant δ0 -6.4899 *** 1.2358 z1 – Ethnicity δ1 -0.7900 * 0.4750 z2 – Education δ2 -0.1237 ** 0.0517 z3 – Credit δ3 0.0003 0.0106 z4 - Farming experience δ4 0.0125 0.0347 z5 - Family size δ5 1.2460 *** 0.1869 Farming systems (coffee mono is based) Coffee and rice crops -0.4275 0.7237 Coffee and the other industrial crops - 4.1516 *** 1.3734 Rice and other crops 1.5352 1.0054 Coffee, rice and other industrial crops -1.2549 * 0.7125
Inefficiency was negatively associated with education and ethnicity and positively correlated with larger family size. This means that in comparison to the major ethnic group the Kinh people, other ethnic farmers tend to be less efficient. In addition, farms with household heads who had achieved a higher number of year of schooling obtained higher efficiency levels than their counterparts. Coffee households with smaller family sizes also appeared to reach higher efficiency levels than larger family size households.
These findings are consistent with efficiency literature (Bozoğlu and Ceyhan, 2007).
5.6.3 Relative efficiency levels of differing farming systems
As shown in Table 5.4, both economies and diseconomies of scope are detected in differing farming systems. By definition, scope economies may lead to efficiency improvements; thus, the result of the inefficiency model also provides robustness to the result of diversification economies. ______Chapter 5: Diversification and Productive Efficiency
Table 5.4: Complementary effects and diversification efficiencies
Sampl Mean Economies of Relationship with e size efficiency diversification inefficiency in score using Equation (5) inefficiency model Farming system a (i.e. Equation (10) b Coffee mono crop 20 0.79 N/A Base Coffee and rice 41 0.80 -0.4275 crops No (0.1875**) Coffee and the other industrial crops 23 0.94 Yes (- 0.0360**) - 4.1516*** Coffee, rice and other industrial crops 71 0.82 N/A -1.2549* a A negative (positive) sign refers economies (diseconomies) of scope. Two asterisks imply the significance level at 95%. b : A negative (positive) sign implies a higher (lower) efficiency level compared with a coffee mono crop system. Two asterisks and one asterisk represent the significance level at 95% and 90% respectively.
Table 5.4 provides summaries of efficiency scores across the three farming systems with three main findings. First, the synchronized farming system –farms growing coffee and other industrial crops - have the highest average efficiency scores
(0.94 compared with the benchmark mono-cropping’s average score of 0.80). Both parametric (t-test) and nonparametric (Kolmogorov-Smirnov) tests also confirmed these results are statistically significant14. Also, the combination of coffee crop and other industrial crops exhibit economies of diversification. The cross derivative of the distance function with respect to these two outputs, defined in Equation 5, was calculated to be -0.0360, which is statistically different from zero. This implies that producing an extra unit of other industrial crops other than coffee crop could reduce the
14 p-value of the t-test equals 0.0004 and this number of the Kolmogorov-Smirnov test was 0.001. ______Chapter 5: Diversification and Productive Efficiency
marginal distance by 0.036. A reduction in marginal distance means an improvement in the productive efficiency level in the production of both coffee and the other industrial crops. Note that the results from the inefficiency model also produce the similar implication that farms jointly cultivating coffee and other industrial crops are found to be more efficient than the other farming systems.
Second, there is statistically significant evidence of the diseconomies of diversification between coffee and rice in the segregated farming system. The cross derivative measure of 0.1875 for coffee-rice suggests that an extra unit increase in rice production volume leads to a 0.1875 increase in the marginal distance of producing coffee or vice-versa. An increase in the value of the marginal distance means a reduction in the productive efficiency level. Note that the inefficiency model reveals a negative sign for the dummy of the coffee-rice farming group but this variable is not statistically significant while there is no statistical difference in the average values of efficiency scores between the two systems (i.e. mono coffee and coffee-rice). Since coffee and rice are cultivated in spatially separated plots of land, there does not exist agronomic and/or ecological symbiosis between these two crops. Our findings show that this segregated farming system does not provide any efficiency benefits. This raises the question as to why this farming strategy is still common in Vietnam’s coffee growing community.
Section 4.7 provides further discussion on this issue.
Third, the coffee mono crop system and the coffee-rice segregated system are found to be less efficient than farming systems growing coffee together with other industrial crops and/or rice. It could be inferred that these results are due to the presence of negative complementary effects between coffee and rice and positive complementary
______Chapter 5: Diversification and Productive Efficiency
effects between coffee and other industrial crops. Section 4.7 also provides further discussions on this point.
5.7 Discussions
5.7.1 Evidence of agronomic benefits between coffee and other industrial
crops
The results from both the diversification measure and the inefficiency model consistently confirm that growing other industrial crops in the same land plot of coffee provide efficiency benefits. It should be noted that our efficiency measures are not impacted by prices of inputs, and we argue that this observation can be due to at least three possibilities. First, the positive complementary effects of growing other industrial crops in the coffee plot is due to lower overall consumption of inputs compared to using inputs separately for coffee and other industrial crops. This is important for policy- makers as crop cultivation has been dependent on the heavy use of chemical inputs
(Athukorala et al., 2015, p.334). Reduction in the use of chemicals or more environmentally friendly practices to cope with environmental issues while increasing agricultural productivity to meet food security purposes was identified as an urgent need in many Asian countries (Ahmed, 2012),
Second, as our output includes the value of the other industrial crops, efficiency maybe affected by the prices of other industrial crops being more favorable than that of coffee. Third, there exists agronomic and/or ecological benefits which are translated into efficiency improvements. As data does not allow us to investigate the first two possibilities further, we argue that it is important for both farmers and local
______Chapter 5: Diversification and Productive Efficiency
governmental agencies to pay more attention to the agronomic or ecological benefits of synchronization of coffee and other industrial crops. Interestingly, Ogundari (2013) also reported that combining coffee with other industrial crops such as bananas, provides agronomic and social benefits to farmers. This is in line with the India studies which indicate that crop diversification could both increase economic gain and environmental benefits (Mandal, 2014).
Industrial crops such as pepper and avocado when spatially synchronized with coffee therefore promise agronomic benefits to farmers as well as ecological benefits for the surrounding natural environment. Given that this type of crop diversification was not chosen by many Vietnamese coffee farmers, we also hypothesize that there could exist another type of trade-off. That is, there is the observed divergence between short-run loss and the longer-run gains - in the form of ecological benefits of inter-cropping coffee with other industrial crops such as pepper, avocados and bananas. It can be assumed that given the potential to translate the agronomic and ecological benefits into direct economic gains (for example via savings in production costs or additional flows of capital) is not directly observable by farmers, they tend to underestimate the values of these gains. This leads to underproduction of agronomic and ecological benefits. As a manifestation of market failure this shows there is a positive externality in appropriate crop synchronization (Chavas, 2008). Therefore, policy interventions can be designed so as to internalize these externalities. For example, government can provide incentives in the form of subsidies or cash rewards to coffee farms to encourage them to synchronize rather than segregate - particularly for crops that deliver agronomic values.
______Chapter 5: Diversification and Productive Efficiency
5.7.2 Why coffee farmers still choose rice
Table 5.5 provides a snapshot of socio-economic variations between different categories of farms in terms of crop diversification. Rice is a fairly common crop chosen by coffee farmers in the area covered by this research. One notable characteristic is that most farms cultivated rice for their own consumption. In other words, these farmers appear to have cultivated rice as a form of subsistence for food security15 purposes as it is the major staple food in Vietnam. In addition, compared to other industrial crops (i.e., coffee or pepper) rice cultivation requires much lower capital investment (Amarasinghe et al., 2015). Therefore, it is attractive to farmers who have little capital or have difficult access to capital.16 This is also shown in Table 5.5 with coffee-rice farms having lower income levels and less access to credit loans than other farms.
Table 5.5: Characteristics of rice versus non-rice farms and coffee-rice versus other farms
Non- Rice and Other Coffee- rice others farms rice farms farms farms (n=43) (n=124) (n=126) (n=41) Technical efficiency 0.87 0.81 * 0.83 0.80 Total annual income in 103.37 69.33 ** 87.10 50.41 *** million VND z1 – Ethnicity 0.63 0.35 *** 0.51 0.17 *** z2 – Education 6.75 5.41 * 6.00 4.98 z3 – Credit 26.24 19.65 * 23.38 15.09 ** z4 - Farming experience 10.05 7.89 ** 7.95 9.95 ** z5 - Family size 4.19 3.88 4.13 3.41 **
15 Farming diversification in which rice is an option also delivers food security, i.e., rice-fish integration was found a viable diversification option in Bangladesh (Ahmed et al., 2011).
16 We note that low capital investment means lower financial risk; therefore it can be argued that farmers might be risk averse. However, we do not have data on which to base further analysis.
______Chapter 5: Diversification and Productive Efficiency
The significance levels of the t-test are indicated as: * p<0.10, ** p<0.05, *** p<0.010
In fact, empirical results also allow us to hypothesize that there is sup-optimal
investment decision making. Coffee farmers, given resource constraints, cultivate rice as
rice cropping requires less investment, has less risk and increases food security. But
cultivating rice reduces their overall productive efficiency which causes higher
production costs and lower total farm income. The literature has shown that subsistence
production is commonly considered as less efficient than commercial crops (Kostov and
Lingard, 2004) and that to enhance poverty alleviation, shifting from subsistence to cash
crop production may be a desirable strategy for subsistence farmers (Solís et al., 2009).
As it is desirable for policies to promote more efficient farming diversification, these
results suggest several policy options which could be adopted to encourage farmers to
diversify crops with a greater focus on efficiency gains. For example, if rice farmers
have difficulties in accessing capital to cultivate non-rice crops, policies should target
facilitating access to credit for these farmers. Similarly, if rice farmers have concerns
over food security, policies should be designed to address these concerns.
5.8 Conclusions
This chapter adds to the literature on crop farming diversification by coffee
farmers by providing additional empirical evidence derived from estimations of the input
distance function based on a new dataset of Vietnam’s coffee farmers. Empirical results
provide several important findings. First, the average inefficiency level is estimated to
be 18%. Second, there is evidence of positive economic benefits in the form of
economies of scope and diversification efficiencies flowing from the synchronized
______Chapter 5: Diversification and Productive Efficiency
farming system, i.e., between coffee and the other industrial crops. We hypothesized that these benefits could come from agronomic values and more efficient consumption of inputs. In addition, segregated farming systems cultivating coffee and rice in differing plots of land
While inter-cropping between coffee with other industrial crops - such as avocados and pepper - could provide some biological benefits to prevent further soil erosion in the
Highlands of Vietnam as well as other related bio-diversification benefits (see: Singh,
2000; Ogundari, 2013), there is a tendency for farmers not to do so. Rather, coffee farmers still choose to cultivate rice as a segregated crop integration system. We argue that if farmers in long-run can be encouraged to move from rice towards other industrial crops they can increase the overall efficiency level of the coffee industry. It is therefore desirable to investigate further the nature of the market failure to achieve a socially optimal level of production of ecological benefits in the context of crop diversity.
However, to achieve this optimal policy result, interventions is needed to remove or reduce barriers farmers face in terms of capital investment, to change risk-averse attitudes and to improve rice food security.
Further research could add robustness to some of this study’s conclusions by examining larger datasets and panel data to better address the issue of heterogeneity in coffee farming. Further analysis of specific types of crops that deliver both efficiency gains, economies of scope and higher profitability to coffee farmers in Vietnam could also complement this study. In addition, it is desirable to investigate further the nature of the market failure in order to achieve a socially optimal level of production of ecological benefits from the diversification of synchronized crops.
______Chapter 5: Diversification and Productive Efficiency
Eco-efficiency analysis of coffee farming 17
In this chapter, we aimed at answering following questions, (1) What are eco- efficiency level across coffee farms? (2) Are there significant differences in the eco- efficiency level between sustainability certified farms and non-certified farms? and (3)
What eco-efficiency determinants drive the variations in the eco-efficiency level across farms?
6.1 Introduction
Past studies showed that attributes related to sustainability certification are important for coffee consumers (Van Loo et al., 2015) and consumers are willing to pay more for sustainability certified coffee products. For example, consumers in the United
States and Germany are willing to pay higher premiums for shade grown or Fairtrade certified coffee than organic certified coffee (Basu and Hicks, 2008; Loureiro and
Lotade, 2005). If certified coffee receives higher prices, one could expect that these price premiums would generate higher revenue to farmers. Additionally, as
17 This chapter has been published in Journal of Cleaner Production, which can be found at https://doi.org/10.1016/j.jclepro.2018.02.147. Only minor changes/ edits have been done.
______Chapter 6: Eco-efficiency analysis of coffee farming
sustainability certification programs educate farmers about environmental ethics, sustainability certified production has the potential to improve environmental performance.
However, empirical evidence of the economic and environmental effects of sustainability certification is mixed. Some studies showed that certification delivered benefits to farmers through lower production cost (Kilian et al., 2006; Lyngbaek et al.,
2001), increased revenue (Bolwig et al., 2009) and reduction of farmers’ livelihood vulnerability (Bacon, 2005). In addition, certified production was shown to have provided positive environmental outcomes through increased use of organic fertiliser application (Ibanez and Blackman, 2016). Several studies also linked sustainability certified production with better biodiversity (Philpott et al., 2007) and greater biodiversity conservation (Hardt et al., 2015).
Yet, other empirical studies claimed that the economic effects of sustainability certification were not significant (Valkila, 2009; Chiputwa et al., 2015; Ibanez and
Blackman, 2016). In particular, Lyngbaek et al. (2001) indicated that sustainability certified coffee farms obtained lower yield than conventional coffee farms. Moreover,
Castro-Tanzi et al. (2012) reported that little evidence was found for the positive impacts of certified practices on the environment.
Vietnam is the second largest coffee producing country globally, with an average contribution of more than 19% of global production volume. Coffee farming in Vietnam is largely characterised by small household farms (Ho et al., 2017). Recent literature indicates that coffee production in Vietnam has been undertaken in a highly unsustainable manner. Two of the most serious observations are related to excessive use ______Chapter 6: Eco-efficiency analysis of coffee farming
of chemical inputs and irrigation water (Luong and Tauer, 2006; Dzung et al., 2011).
The average amount of nitrogen, phosphate and potassium applied to coffee production in Vietnam is about 1,115 kilograms per hectare (Dang and Shively, 2005), while coffee plants require only up to 800 kg nitrogen per ha in order to be able to provide an adequate yield (Bruno et al., 2011). Also, a recent study suggested that groundwater pumping for coffee farming is excessive and could be reduced by a significant amount
(Amarasinghe et al., 2015).
Meyfroidt et al. (2013) reported that sustainability certification has become increasingly popular in Vietnam in recent years even though the excessive application of irrigation and fertilizers have made it difficult for farmers to conform to most certification standards and programs. Interestingly, Basu and Hicks (2008) argued that
‘easy’ labels showing environmental performance costs much less than certifying with international agencies. If true, it might be predicted that Vietnamese farmers would opt for the cheaper labeling scheme and move away from international certification.
However, if international certification, in fact, improves environmental performance, moving away from these certification programs could hinder the potential environmental benefits. Therefore, empirical evidence on the economic and environmental effects of certified coffee production in Vietnam is needed as such empirical evidence would be helpful for farmers, certification organisations as well as governmental agencies.
The primary aim of the present paper is to investigate differences in both environmental and economic performance between conventional and certified coffee farms. As interest is placed on both aspects of the potential benefits of sustainability certifications, i.e. enhancing farmers’ revenue and reducing the potential of making
______Chapter 6: Eco-efficiency analysis of coffee farming
environmental impacts, we use a frontier eco-efficiency (FEE) model as often used in the literature on ecological economics (see Lauwers, 2009 for more details). In this model, production outcomes in terms of value-added and environmental pressures contained in environmentally detrimental inputs (i.e. chemical fertilisers, chemical, water, and arable land) are optimised into a simple framework of Data Envelopment
Analysis (DEA). Our literature review shows that the present paper is one of the first empirical applications of the eco-efficiency model in the context of sustainability certification in coffee production.
This paper also investigates which factors explain variations in eco-inefficiency across farms by using OLS and fractional regression models. This second stage of the empirical analysis serves two purposes. First, it assesses econometrically if participation in sustainability certification is one of the drivers of efficiency variations. Second, it identifies other important factors that farmers and policy makers can look at in order to learn how to improve efficiency. The fractional regression model is used as it is a more natural way of estimating fractional data using the generalised method of moments
(GMM) estimator.
The remaining parts of the paper are organized as follows. Section 6.2 presents a review of the literature. Sections 6.3 and 6.4 outline our methodology and empirical strategies. Section 6.5 presents and discusses results while Section 6.6 provides our conclusions.
______Chapter 6: Eco-efficiency analysis of coffee farming
6.2 Literature review
The economic and environmental effects of sustainability certification in coffee production have attracted much attention. 22 out of 46 studies in the meta-analysis of
Blackman and Rivera (2011) which investigated the effects of sustainability certification for various products and services, focused on coffee. Among these studies, many provide quantitative measures of economic benefits while only four studies examined environmental benefits and two focused on both economic and environmental benefits. Following the review by Blackman and Rivera (2011), a snapshot of the focus of recent literature are set out in Table 6.1.
Table 6.1: Recent studies on the effects of certification in coffee production With focus on economic With focus on With focus on both benefit environmental or economic and ecological benefits environmental benefits Barham et al., 2011 Blackman and Ibanez and Blackman, Barham and Weber, 2012 Naranjo, 2012 2016 Ruben and Fort, 2012 Hardt et al., 2015 Haggar et al., 2017 Jena et al., 2012 Bose et al., 2016 Vellema et al., 2015 Giuliani et al., 2017 Chiputwa et al. 2015 Latynskiy and Berger, 2017
Several studies found positive economic effects from sustainability certification, but these examinations focused on individual economic indicators separately. Using a data set of 845 coffee farms (including 513 certified farms collected in 2005-06 in
Southern Mexican), Barham et al. (2011) reported that certified growers obtained higher land and labour returns than their conventional counterparts. This was mainly due to
______Chapter 6: Eco-efficiency analysis of coffee farming
higher yields rather than price premiums. Similarly, in a number of other studies certification was shown to improve yields, hence enhancing higher gross revenue - see for example Barham and Weber (2012). Using multivariate probit and tobit models from a data set of 263 certified and 247 noncertified coffee farms in Colombia, Vellema et al.
(2015) found that coffee certification motivated coffee specialisation and was associated with coffee income improvement, but not in total household income. On the contrary,
Jena et al. (2012) reported that sustainability certification had a very little impact on income and poverty reduction among 166 certified and 83 noncertified farms in
Ethiopia. Indeed, this study detected low productivity, an insignificant price premium, and heterogeneity across the certified cooperatives of which the coffee farmers were members. Similarly, data from Peru showed insignificant effects of certification (Ruben and Fort, 2012).
Some studies examined the ecological and environmental effects of sustainability certification. Hardt et al. (2015) evaluated the ecological effects before and after sustainability certification was adopted by coffee farmers in Cerrado, Brazil. Results of several statistical tests using a small sample size showed no significant difference between certified and non-certified farms. Interestingly, Bose et al. (2016), focusing on environmental and social impacts of a certification scheme, showed that certified farms did not change their farm management practices. On the other hand, Blackman and
Naranjo (2012) examined data from 2,603 farms of which only 36 were certified organic farms and which reported a better environmental performance due to the consumption of less chemical input and of more organic fertiliser.
______Chapter 6: Eco-efficiency analysis of coffee farming
Other studies investigated the performance of certified farms along both economic and environmental dimensions. Ibanez and Blackman (2016), in a survey data set of 56 certified and 168 non-certified Columbian farms between 1997 and 2007, examined the effects of organic coffee certification using propensity score matching and difference-in- differences techniques. Their results showed that certification had no significant effect on income or net return but that there are positive environmental outcomes in terms of a decrease in the use of nitrogen and phosphorus and an increase in the use of organic fertilizers. Since propensity matching was used to examine a wide range of different indicators separately, the net effects in cases where all indicators are considered were not available.
One major limitation of analysing various single indicators separately is that analysts cannot draw inferences on the net effects when all indicators are considered.
Hence, an integrated indicator may better reflect the economic performance of farms.
For example, where the economic benefits of certification are concerned, using net income or value-added is better than using single indicators of labour return, yields, or price premiums. Similarly, separate examination of the use of different environmentally detrimental inputs such as chemical fertilisers, pesticides, or water cannot provide information on the total effects. In order to address these shortcomings, the present paper employs DEA benchmarking to evaluate both economic and environmental performance simultaneously through the use of an eco-efficiency frontier approach.
It is noted that the eco-efficiency frontier approach is one of the three types of environmental efficiency analysis which is widely used in the efficiency literature
(Lauwers, 2009). It is built on the concepts of cost-benefit analysis and the productive
______Chapter 6: Eco-efficiency analysis of coffee farming
efficiency principles (Callens and Tyteca, 1999). One typical application of the eco- efficiency frontier approach is the study of Kuosmanen and Kortelainen (2005) who proposed an eco-efficiency measure to evaluate the relationship between value-added and environmental pressures, using the DEA technique. Since then, there have been many empirical studies using this approach to examine the economic and environmental performance of agricultural production (see, for example, Picazo-Tadeo et al., 2011;
Zhou et al., 2018).
Picazo-Tadeo et al. (2011) also calculated eco-efficiency scores for 3,960 Spanish farms. Economic value-added was treated as an output and five environmental pressures
(a specialisation indicator, nitrogen, phosphorus balance, pesticide risk, and energy ratio) were treated as inputs. The radial eco-efficiency (the ratio of output to aggregate input) produced a mean of 0.56 which indicates that an average farm could achieve its value-added with only 56% of its current level of environmental pressures. In other words, a 44% potential proportional reduction of all environmental pressures could be attained without reducing the current level of value-added. In this paper, we follow this type of modeling approach.
Another advantage of using the eco-efficiency frontier approach is that we can investigate empirically the relationship between eco-efficiency levels and various factors that are believed to explain the variation in the eco-efficiency level across farms. For this purpose, several approaches have been used to examine the contextual factors.
McDonald (2009) advocated using the ordinal least squares (OLS), while Banker and
Natarajan (2008) argued that OLS, Tobit and maximum likelihood are all appropriate.
Simar and Wilson (2007) proposed the double bootstrap truncated regression. However,
______Chapter 6: Eco-efficiency analysis of coffee farming
the bootstrap approach requires several restrictive assumptions (see assumptions 1-8 in
Simar and Wilson, 2007). Ramalho et al. (2015), departing from Papke and Wooldridge
(1996)’s work, proposed logit and complementary log-log fractional regression models for panel data. The fractional response model is shown to have some advantages over
OLS, particularly in terms of asymptotical efficiency (McDonald, 2009; Ramalho et al.,
2015). For these reasons, we propose to use both OLS and the fractional regression models to provide the robustness check of our result.
6.3 Methodology
There are two stages involved in this analysis. In the first stage, eco-efficiency scores are obtained by solving a linear optimization problem. Value-added is treated as the output, while environmental pressures are treated as inputs in the DEA models
(Kuosmanen and Kortelainen, 2005). In the second stage, the eco-inefficiency scores are then regressed on socio-economic covariates, as environment, where the entity is operating, may consist of favourable or unfavourable characteristics (Emrouznejad and
De Witte, 2010).
6.3.1 First stage: Frontier eco-efficiency
In order to measure the performance of decision-making units (DMUs), DEA – a mathematical approach – has been in a wide variety of studies from 1978 to 2016 (see, for example, Emrouznejad and Yang, 2017). In terms of convexity constraints, there are two branches of the DEA approach used (Cook and Seiford, 2009), namely the Charnes,
Cooper and Rhodes (CCR) model (Charnes et al., 1978) which exhibits the constant
______Chapter 6: Eco-efficiency analysis of coffee farming
return to scale (CRS). The other branch is the Banker, Charnes and Copper (BCC) model (Banker et al., 1984), which incorporates the variable return to scale (VRS).
In terms of sustainability analysis, eco-efficiency measurement is the key route followed when applying DEA (Zhou et al., 2018). Many previous studies on eco- efficiency (i.e., Kuosmanen and Kortelainen, 2005; Arabi et al., 2017) employed CRS models. This was based on a number of perceived advantages. First, its use as a tool for benchmarking allowed comparison of the performance of a DMU with its peers. Given
DEA models do not require functional forms, characterizing production technologies is one of its most important advantages. It is therefore important to note that eco-efficiency does not reflect the effect of production technologies, but rather focuses on the potential reduction of environmental pressures, given a certain level of economic outputs. This is not the case under VRS (Emrouznejad and De Witte, 2010). Second, from an ecological standpoint, what really matters are the total environmental pressures exerted and not land allocation into different farms (Picazo-Tadeo et al., 2011). In addition, the choice of convexity constraints is not clear-cut in measuring eco-efficiency (Picazo-Tadeo et al.,
2012). Therefore, in this paper, we assumed that the relationship between economic outputs and environmental pressures exhibits constant returns to scale.
Following the standard DEA framework proposed and developed by Charnes, et al., (1978), eco-efficiency measurement can be expressed as the following optimisation problem (also see Kuosmanen and Kortelainen, 2005; Picazo-Tadeo et al., 2011):
______Chapter 6: Eco-efficiency analysis of coffee farming
μV max m μ , w wp j=1 jj 6.1 V ≤≥≥μ subject to :m 1;w 0; 0. wp j=1 jj
where V is the vector of output measuring the value-added of coffee farms; pj represents the level of environmental pressure of j (J = 1, …, m); m is the number of environmental pressures; wj denotes the weighting factor of the environmental pressure;
μ is the weight associated with the value-added. For example, technically efficient inputs are the weighted sum of the inputs of other firms in the dataset, including itself.
However, this model still has the limitation of yielding an infinite number of solutions.
Consequently, the constraint of virtual pressures is set to be unity to identify a unique solution. The above problem can be transformed to a linear form:
max μV μ,w m st.. w p ≤ 1; j=1 jj m − wp ≤−1; 6.2 j=1 jj m Vwp−≤0; j=1 jj w ≥≥0;μ 0.
The dual form for this problem can be expressed as:
min Φ λ,Φ n s.t. λV −≥V 0 0; i=1 ii n Φ−P0 λ p ≥0; 6.3 i=1 ii λ ≥ i 0; i =1...n
______Chapter 6: Eco-efficiency analysis of coffee farming
where λ is a vector of weights associated with m environmental pressures, V0 and
P0 represent the value-added and environmental pressures of individual farms respectively and Φ is a scalar shrinking factor.
By solving the above linear programming problem, the DEA eco-efficiency score for each individual farm is obtained. This score ranges between zero and one, where higher values denote a higher eco-efficiency level which means higher total revenue per total environmental pressures. As we use input-orientated optimisation, the eco- efficiency level indicates that given the level of value-added, the potential decrease in all environmental pressures is greater, proportionally, that the farm could have obtained otherwise.
6.3.2 Second stage: Fractional regression models (FRMs)
Following Ramalho et al. (2015), FRMs that require the data generating process
(DGP) can be expressed as:
=++θα yGxit1(), it i v it 6.4
where vit denotes time-varying unobserved heterogeneity, αi represents time- invariant unobserved heterogeneity and assuming that G( ) is specified by a logit,
= exp( ) G() , or cloglog, G ( )=− 1 exp[ − exp( )]form. The DGP (4) can be 1+ exp( ) transformed into an exponential regression form where G( ) has a logit or cloglog specification as:
αθα=++ E(yxit |,,) it i v it G1 [exp( x it i v it )]. 6.5
______Chapter 6: Eco-efficiency analysis of coffee farming
-1 Let H1( ) = G1( ) . Then an exponential model in which the dependent variable is transformed can be obtained as:
=++θα H1(yxvit ) exp( it i it ), 6.6
yit where Hy()= (logit model) or H ()yy=− ln(1) − (cloglog model) and yit is 1 it − 1 it it 1 yit set to the interval [0,1). These model specifications can deal with zero value of the dependent variable, which is a known advantage when compared to the linear-fractional model.
The choice of estimators depends on the structure of data. For example, the generalized method of moments with a pooled random effects estimator is only consistent under the random effects assumption. If the assumption holds, the marginal effects of time-invariant covariates can only be identified in random-effects models, and then the assumption that xit and αi are independently distributed is reliable. This suffices for consistent estimation of θ. Therefore, two components, xit and αi can be treated as a single error term, and thus equation 6 can be expressed as:
Hy() 1 it −=1 exp(α +vv ) −≡ 1pre , 6.7 θ iit it exp(xit )
Given the lack of a theoretical analysis which allows us to choose among different functional forms (logit or complementary log-log), we employ both models with the goal of providing more robustness of the results.
______Chapter 6: Eco-efficiency analysis of coffee farming
6.4 Empirical strategies
6.4.1 Data descriptions
The dataset used for this study is based on the surveyed data as described in
Section 3.4. The final sample for this study is thus described as in Table 6.2 as below.
Table 6.2: Non-certified and certified farm categories Certification schemes Crop year Total 2012 2013 2014 Non-certified farms 364 327 329 1,017 Certified farms 348 399 394 1,144 Utz- certified farms 124 130 140 394 4C – certified farms 144 193 179 516 Other schemes certified farms 80 76 74 231 Total 712 726 723 2,161
6.4.2 Variable selection
The radial DEA Eco-efficiency measure which is specified by Kuosmanen and
Kortelainen (2005) is employed in this paper. This model is particularly useful in benchmarking peers without specifying a production technology. It describes a relationship between economic value-added and environmental pressures caused by the inputs used by decision-making units (also see, Picazo-Tadeo et al., 2011). This model incorporates both input and output factors of production and environmental pressures into a single eco-efficiency index (Kuosmanen and Kortelainen, 2005). In the context of coffee production considered in this paper, the output of production (i.e. coffee production volume), and the inputs of production (i.e. organic fertilizers, labor and other expenses) are captured implicitly in the calculation of the economic value-added. ______Chapter 6: Eco-efficiency analysis of coffee farming
Explicitly, five environmental pressures, namely nitrogen, phosphorus, water use, pesticides and land area, are accounted for in the eco-efficiency index.
Table 6.3: Descriptive statistics
Variable Definition Mean Std. Output and input variables Total value added per weighted tree measured 92.799 36.52 VApwt (V) in 1000 VND18 Amount of nitrogen applied per weighted tree 0.412 0.23 nitropwt (P1) measured in kilograms Amount of phosphorus applied per weighted 0.225 0.15 phospwt (P2) tree19 measured in kilograms Amount of irrigation water applied per 1.210 0.54 waterpwt (P3) weighted tree measured in m3 Amount of pesticides, herbicides and 0.006 0.01 phfpwt (P4) fungicides applied per weighted tree measured in kilograms areapwt (P5) Density per weighted tree measured in hectares 0.001 0.00 Contextual variables elev (z1) Elevation of farm from sea level measured in 807.051 319.62
18 The exchange rate at the time of data (2015) was 1 USD = 22,550 VND.
19 Also see Appendix A1 for the weighted number of coffee trees.
______Chapter 6: Eco-efficiency analysis of coffee farming
meters familisize (z2) Family size measured in number of people 4.538 1.50 Household head’s coffee farming experience 19.036 7.13 T expi (z3) measured in years Access to public extension services, 1 for yes 0.381 0.49 wo Extensv (z4) and 0 for otherwise Availability of shade trees in coffee plantation, 0.546 0.50 prim shadetree (z5) 1 for yes and 0 for otherwise Availability of windbreak trees around coffee 0.525 0.50 ary windtree (z6) plantation, 1 for yes and 0 for otherwise Irrigation types, 0 for micro-basin and 1 for 0.569 0.50 objec irrtype (z7) otherwise (sprinkle) Sustainability certification status, 1 for yes and 0.541 0.50 tives cert (z8) 0 for otherwise of cfsize (z9) Coffee farm size measured in hectares 1.645 1.45 Tree density measured in number of coffee tree 1,081 113.53 dens (z10) this per hectare paper are to measure both economic and environmental performance of coffee farms and to identify factors affecting their performance. In the eco-efficiency literature, value- added has been widely used to characterise the economic benefit of farming (see, Ullah et al., 2016; Picazo-Tadeo et al., 2011). In terms of environmental impacts, many empirical studies considered the consumption level of nitrogen as an environmental pressure (see, for example, Picazo-Tadeo et al., 2011; Masuda, 2016). The use of phosphorus, water, and pesticides was also used in other studies (Ullah et al., 2016;
Ibanez and Blackman, 2016). Since the conversion of forest into coffee plantations is also a serious environmental issue in Vietnam (D’haeze et al., 2005a) and other countries (Marchand, 2012), we propose to use the size of cultivated land for which data is available, as an environmental pressure.
Explanatory variables used in the second stage include farm characteristics and factors associated with sustainable practices. The selection of these variables mainly follows the empirical literature. Farm average elevation from sea level was found to be
______Chapter 6: Eco-efficiency analysis of coffee farming
significantly and negatively associated with the probability of being organically certified
(Blackman and Naranjo, 2012). This geophysical characteristic was also tested as a factor affecting the net revenue of coffee production in Peru (Barham and Weber, 2012).
Family size, farming experience, participation in public extension services and coffee farm size are treated as farm characteristics and which are commonly used in the inefficiency studies (see, for example, Rahman, 2009; Wollni and Brümmer, 2012).
Shade trees and wind-break trees were used as indicators of environmental benefits to assess eco-certification in coffee production (Blackman and Naranjo, 2012). In addition, the irrigation type variable captures different irrigation technologies used in Vietnam’s coffee growing sector (e.g., flooding irrigation as a control group refers to conventional practices versus more advanced irrigation technologies such as drip and spray). We also hypothesize that actual tree density, representing the level of coffee intensification, influences eco-efficiency variation.
6.4.3 Weighted number of coffee trees
If the age of perennial trees is biologically associated with various doses of variable inputs, fertility and yield production, tree ages should be captured to avoid estimation bias. The literature has highlighted the need to adjust for different age groups of perennial trees when estimating production frontiers (Hasnah et al., 2004; Ofori-Bah and Asafu-Adjaye, 2011). For coffee trees, it takes about three years for newly planted trees to start bearing fruit while the life of a coffee tree is usually about 30 years, depending on varieties. Trees aged under eight years show strong yield production growth and reach their highest yield potentials between 16 to 20 years of age, from which point yields start declining. For this reason, we used a nonlinear least-squares ______Chapter 6: Eco-efficiency analysis of coffee farming
estimation to obtain the weights associated with different groups of ages. The equation is specified as:
wct=+ w T w−− T + w−− T + w − T − + w++ T g g( it )8 8 (it ) 9 15 9 15( it ) 16 20 16 20( it ) 21 21 ( it ) 6.8
where, wct denotes the weighted number of coffee trees; Tg(it) is the number of grafted coffee trees; T8- denotes the number of trees aged under 8 years; T9-15 represents the number of trees aged 9 to 15 years; T16-20 denotes the number of trees aged 16 to 20
+ years; T21 refers the number of trees aged greater than 21 years and w’s are the weights to be estimated. Trees between 16 to 20 years, their full production period, are assigned a weight of w16-20 = 1. A mean of the remaining w’s is derived from the data using the average weight of dried coffee beans per tree for each tree age group. The results from a nonlinear regression model are presented in Appendix 1. Then the weighted number of the coffee tree for each farm in a particular year is calculated as
5 β j wct=+ T Tj ,4 j ≠. it16− 20( it ) it β 4 j=1
6.5 Results
6.5.1 Comparison of eco-efficiency between certified farms and non-certified farms
Three year Eco-efficiency 2012 2013 2014 pooled Certified farming Mean 0.5425 0.4951 0.5013 0.5117 Std. 0.2216 0.2188 0.2022 0.2149 Obs. 348 399 394 1,141 Conventional farming Mean 0.4537 0.4440 0.4621 0.4533 Std. 0.1931 0.1875 0.1899 0.1903 Obs. 364 327 329 1,020
______Chapter 6: Eco-efficiency analysis of coffee farming
All farms Mean 0.4971 0.4721 0.4835 0.481 Table Std. 0.2121 0.2067 0.1975 0.2057 6.4: Obs. 712 726 723 2,161 Mean eco- efficiency levels
______Chapter 6: Eco-efficiency analysis of coffee farming
We used three DEA models for three crop years separately to avoid the effect of possible technology changes on the frontier. Each model included both certified farms and conventional farms. Table 6.4 shows the summary statistics of average eco-efficiency levels for the entire sample over the three crop years. The mean values of eco-efficiency levels are 49.7%, 47.2% and 48.3%20 respectively for the three crop years. Thus, an average farm can reduce all environmental pressures proportionally by more than 50% without a reduction in the total value-added of coffee production. Note that the five dimensions of environmental pressure considered capture the level of consumption of various inputs. Hence by reducing consumption of inputs, farms can reduce both production costs as well as the negative pressures of coffee production on the surrounding environment. Such a low level of eco-efficiency provides a strong signal to relevant stakeholders in the industry that there are large opportunities for coffee farmers to increase
20 In agricultural production, previous studies also indicated low eco-efficiency levels (see, for example, Picazo-Tadeo et al., 2011; Ullah et al., 2016).
______Chapter 6: Eco-efficiency analysis of coffee farming
pure productive efficiency (i.e. reducing input consumption). There is no trade-off between cost considerations and environmental impacts of input consumption in this context as farms need only to reduce input consumption to achieve a reduction of production costs and environmental pressures. Equally, by reducing the consumption of environmentally detrimental inputs, both cost-saving benefits to farms and environment- improving benefits to the community can be derived.
Figure 6.1: Kernel estimated densities of farm eco- efficiency scores
Figure 6.1 illustrates the estimated Kernel densities of eco-efficiency scores over the three crop years of conventional and certified farm groups. The shapes of Kernel densities over three crop years are intuitively similar to the normal distribution; hence it
______Chapter 6: Eco-efficiency analysis of coffee farming
is reasonable to use parametric and non-parametric tests to assess a hypothesis if there is difference in eco-efficiency levels between non-certified and certified farms.
Table 6.5: Tests for equality of eco-efficiency distribution
Simar- Kolmogorov- Mann- Null hypothesis Zelenyuk Li21 Smirnov (KS) Whitney (MW) Li statistic D statistic W statistic 4.5859 0.1494 46030 f(Cert Eff2012) = f(Conv Eff2012) (0.0000) (0.0011) (0.0000) 3.3257 0.1299 48932 f(Cert Eff2013) = f(Conv Eff2013) (0.0010) (0.0076) (0.0142) 0.3070 0.0838 51299 f(Cert Eff2014) = f(Conv Eff2014) (0.7000) (0.1922) (0.0659) p-values are in parentheses
Following the efficiency literature, the Simar-Zelenyuk Li and other non- parametric tests for equality of eco-efficiency distribution were performed and are presented in Table 6.5. The Li-test and KS test show that in the first two crop years, the difference in eco-efficiency levels between conventional and certified coffee farms is statistically significant, but not the last year of the sample. The MW test provides similar results except that the difference in 2014 was also statistically significant.22
21 Taking into account the normality problems, we used the bootstrapped Li test (Li, 1996) to compare distributions of DEA scores of the two groups. The Algorithm II for comparing distribution of the estimated DEA efficiency scores can be found in Simar and Zelenyuk (2006).
22 Note that this paper does not attempt to quantify the effects of sustainability certification on eco- efficiency. Such assessment can be conducted in a more direct way, for example by using other econometric techniques such as propensity matching in the setting of “before and after” and/or “control
______Chapter 6: Eco-efficiency analysis of coffee farming
More interestingly, over the three-year period the power of Li- test decreases and the magnitude of the difference in efficiency level appears to reduce. This poses a hypothesis of convergence in the eco-efficiency level between two groups of farms.
Note that this goes beyond the scope of our empirical analysis and we only attempt to hypothesize that there are at least three possibilities explaining the eco-efficiency convergence: non-certified farms may become more eco-efficient over time, certified farms become less eco-efficient over the time and a combination of these two former possibilities. In general, the eco-efficiency measure has two aspects (i.e. value-added and environmental pressures); hence there could be many forces driving the convergence of the efficiency gap. There are several key examples of these forces which may be relevant to the operation of farms. One is the premium in the price of certified coffee.
There is the possibility that the price difference between certified and non-certified coffee could reduce over time, leading to a convergence of the value-added. However, our data shows that the price premium remains unchanged over the three crop years.
Another factor could be the presence of positive externalities derived certification programs in the form of learning and skills transferred from certified to non-certified farms. Due to lack of data, we are not able to assess whether this is the case.
Another important driving force of the convergence could be reductions in the effort level of certified farms over time. In the case of Vietnam’s coffee sector, there is
and treatment” framework. For the sake of completeness of our analysis, Appendix 2 provides some summary statistics of farmers who dropped out of the certification program and those who joined in the program over the three crop years. T-tests show no significant evidence of changes in the efficiency levels of those farms over time.
______Chapter 6: Eco-efficiency analysis of coffee farming
no direct monetary cost for farmers to participate in certification programs. Farmers, however, need to spend several extra working days to fulfill the compliance tasks of certification documentation, procedures, and standards. Our data reveals that the average number of these extra working days has declined over time (10.24 days in 2012, 9.16 days in 2013, and 8.30 days in 2014). Common t-tests were performed and the results confirmed that the time certified farms spent on certification had declined over time
(except between 2013 and 2014). Therefore, we suspect that the reduction in the efforts or compliance with certification standards of farmers could be one of the main driving forces behind the convergence.
6.5.2 Eco-inefficiency variance analysis
Another objective of this paper is to examine contextual factors affecting eco- efficiency levels. All farms with the eco-efficiency score of unity (i.e. 51 out of 2,161 observations) are excluded in the second stage in order to focus on the evaluation of the relationship between contextual factors and inefficiency variance. As the distribution of efficiency scores is intuitively normal, OLS and even truncated regression would provide similar results (see, for example, Thiam et al., 2001; Besstremyannaya, 2013).
In addition to OLS models, FRMs are also estimated as they appear to be more appropriate for modeling a proportional and bounded response variable. We note that given the sample size is relatively large the issue of small-sample bias is believed to be minor in the present study.
The results of the Breusch-Pagan test indicate that the data set should be treated as a panel over the three years rather than a pooled data set. The structure of data is a three- year panel, hence random effects models are expected to be more meaningful than fixed ______Chapter 6: Eco-efficiency analysis of coffee farming
effects. The random effects model using OLS and pooled random effects models using
FRMs were both reported in Table 6.6. Overall, the OLS models are statistically significant at the 99% confidence level with a Wald chi-squared of 128.21. The fractional cloglog and logit forms were able to obtain statistically feasible solutions. In general, the results are relatively consistent across the three models (both between and within farms) in terms of the significance level and the signs of estimated coefficients.
Table 6.6: Factors affecting eco-efficiency
Model 1: OLS Fractional models Variable (random Model 2 : Model 3: logit
______Chapter 6: Eco-efficiency analysis of coffee farming
T effects model) cloglog Elevation z1 0.00006** 0.00027*** 0.00077*** he (0.0000) (0.0000) (0.0002) Family size z2 -0.00231 -0.01527*** -0.04731 result (0.0040) (0.0039) (0.0486) Coffee farming experience z3 0.00053 0.00097 -0.0066 s (0.0010) (0.0010) (0.0096) Public extension service z4 -0.02824** -0.05806*** -0.25985* prese (0.0131) (0.0156) (0.1473) Shade trees z5 -0.02249* -0.05044*** -0.13562 nt (0.0121) (0.0158) (0.1171) Wind-break trees z6 0.06271*** 0.18961*** 0.51998*** some (0.0156) (0.0170) (0.1336) Irrigation types z7 -0.08334*** -0.25048*** -0.60773*** impo (0.0134) (0.0184) (0.1421) Certification status z8 0.02371*** 0.07275*** 0.28614*** rtant (0.0089) (0.0152) (0.1058) Coffee farm size z9 -0.00330 -0.03207*** -0.07702*** findi (0.0032) (0.0012) (0.0208) Coffee tree density z10 0.00004 0.00030*** 0.00111** ngs. (0.0001) (0.0001) (0.0005) Year = 2013 d1 -0.02909*** -0.10708*** -0.59173*** First, (0.0050) (0.0182) (0.2239) Year = 2014 d2 -0.02123*** -0.06330*** -0.60900*** the (0.0061) (0.0182) (0.2179) Constant 0.43828*** -0.63582*** -0.26726 eco- (0.0883) (0.0852) (0.4921) Note: Standard errors in parentheses; * p < 0.1, ** p < 0.05, *** p < 0.01 effici ency level is positively correlated with sustainability certification (as a dummy variable) and negatively correlated with year dummies. This result confirms our tests for the difference in the efficiency levels between two groups of farms and the convergence hypothesis.
The coefficient associated with elevation is positive and statistically significant in all models. This implies that coffee farms with higher elevation tend to achieve higher eco-efficiency levels than farms with lower elevation. In addition, the coefficient of windbreak trees is positive and statistically significant, indicating that the availability of
______Chapter 6: Eco-efficiency analysis of coffee farming
these types of trees grown to surround coffee plantations are useful for improving eco- efficiency. This finding favours a crop diversification approach to coffee farming. For example, coffee farmers could grow windbreak trees such as avocado and durian to generate additional cash income. Irrigation types are negatively correlated with eco- efficiency, suggesting that the conventional irrigation system (i.e. flood irrigation) is not as good as modern irrigation systems as, for example, drip and spray irrigation. In addition, the coefficient associated with coffee tree density is positive in all models and statistically significant in the fractional models. This reveals that many farms could increase the density of coffee trees to enhance eco-efficiency. However, the corresponding coefficient of shade trees is statistically negative, indicating that eco- efficiency of coffee trees may be negatively affected by sharing land area and nutrients with shade trees.
Another important finding is that extension services are negative and statistically significant in all models. Agricultural extension services mainly focus on conventional issues, i.e., pruning techniques, to promote growth in yields (Barham and Weber, 2012), rather than reducing environmental pressures. This poses a question for future research in regard to the effectiveness of the extension services provided by local government agencies in terms of agricultural sustainability or long-run economic benefits. Other control variables such as family size and coffee farm size also seem to be negatively associated with the eco-efficiency in coffee production.
______Chapter 6: Eco-efficiency analysis of coffee farming
6.6 Conclusions
This paper uses the frontier-based eco-efficiency model to assess whether coffee farmers with sustainability certification are more eco-efficient than non-certified farms.
Our empirical study provides one of the first such assessments of Vietnamese coffee farmers in which the standard DEA technique was used to calculate eco-efficiency scores and a variety of parametric and non-parametric tests were conducted. OLS and the fractional regression models are also used to examine the relationship between the efficiency variance and relevant contextual or environmental variables. The results raise several important issues.
First, eco-efficiency scores are very low and exhibit wide variation across farms.
Coffee farms have an average eco-efficiency score of less than 50% and this indicates that farms could reduce environmental pressures by more than 50% without any reduction in the value-added of coffee production. The magnitude of this score is unexpected and shows that there is a very large potential for coffee farmers to improve both economic performance and environmental performance.
Second, certified coffee farmers have higher eco-efficiency levels than non- certified farmers. Moreover, participation in the certification schemes reduces the use of irrigation water and other environmental pressures. However, a key issue is that the price premium derived from certification does not pay off in terms of the extra effort required and therefore could lead to market failure of the scheme. Consequently, some form of policy intervention is desirable to ensure the uptake of certification programs.
______Chapter 6: Eco-efficiency analysis of coffee farming
Third, there appears to be a convergence in the level of eco-efficiency between non-certified and certified farms after several years. Given limited data, we are unable to assess the effect of a number of possible reasons. However, we suspect that this could be caused by reductions in the eco-efficiency levels of certified farms over time while uncertified farms are showing evidence of catch-up. We suggest that this issue is worth further investigation. However, the quality of sustainability certification schemes is no less important than the rapid expansion of these programs. Hence further efforts are needed to sustain the positive effects of certification schemes if such certification programs are to be considered an effective alternative way of making coffee growing more sustainable.
Fourth, spray and drip irrigation systems are shown to be more eco-efficient flood irrigation systems. As well, other natural conditions such as the level of elevation and the availability of windbreak trees are positively correlated with eco-efficiency, although it is noted that shade trees may have a negative impact on eco-efficiency. Unexpectedly, the availability of extension services provided by local governments is negatively related to efficiency levels. These findings can, therefore, provide the government the basis for a thoroughgoing review of the extension service scheme which fully embraces the important objectives of sustainability of Vietnam’s coffee growing industry.
This study’s findings provide several directions for further research. First, other ecological and environmental dimensions such as the number of biological species could be included in the analysis to provide a more comprehensive perspective. The current study found a negative effect of shade trees on eco-efficiency improvement. However, this shaded coffee may be highly correlated with the conservation of biodiversity.
______Chapter 6: Eco-efficiency analysis of coffee farming
Worthy of study is the extent to which this could provide ecological benefits
(Borkhataria et al., 2012). Second, other socio-economic aspects relevant to the welfare of coffee households, i.e., income and poverty, should be incorporated. Last, stochastic analysis of the relationship between economic benefits and ecological pressure could provide a useful alternative modeling approach where a stochastic production environment is more in evidence.
______Chapter 6: Eco-efficiency analysis of coffee farming
Trade-off analysis between cost and nutrient efficiency
This chapter particularly attempted answering following research questions: (1) what are the cost efficiency (cost efficiency (CE), cost allocative efficiency (CAE) and technical efficiency (TE)), and environmental efficiency (nutrient efficiency (NE), nutrient allocative efficiency (NAE) and technical efficiency (TE)) level of the
Vietnam’s coffee farming sector? (2) Is there a trade-off between cost and environmental efficiency?
7.1 Introduction
Over use of a variety of nutrients in the form of chemical fertilisers and other nutrient materials in agricultural production has been identified as a main cause of environmental problems in many countries and regions including OECD countries
(Hoang & Coelli 2011) and other European and Asian countries (Vlontzos et al., 2014;
Hou et al., 2012; Berg, 2002; Nguyen et al., 2012). In the context of coffee farming, the literature similarly provides observations to the effect that intensive use of nutrients in the world’s top producing countries including Brazil, Costa Rica (Castro-Tanzi et al.,
2012), Mexico (Eakin et al., 2009) and Vietnam (Amarasinghe et al., 2015), has caused significant environmental damage. Obviously, lower consumption of nutrients would help reduce pollution although this requires farmers to undertake better nutrient management techniques. However, several studies show that farmers, especially smallholders, are reluctant to reduce consumption of fertilisers given this could reduce yields, increase costs or lower profits (Jena et al., 2012; Ranjan et al., 2016). This means that farmers may well perceive the existence of a trade-off between ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
environmental and economic outcomes. There is no previous empirical examination of such a trade-off in coffee production. This study aims to fill this gap.
A further literature gap is related to the analysis of trade-offs in the existing literature, especially those studies that are built on Coelli et al.'s (2007) model, using the materials balance principle approach. Nguyen et al., (2012), using the same approach, investigated the cost and environmental efficiency of rice production in Korea. However the authors focused on the trade-off between cost and nutrient efficiency of technically efficient farms only. In contrast to these studies, Aldanondo-Ochoa et al. (2017) used cross-constrained measure to extend the MBP approach to provide more information on the trade-off between cost and environmental efficiency. In the context of green-house horticultural emissions in Spain, farms were categorized into distinct groups. This was based on the cost and nutrient constraints at the level of farm technically efficient when minimising nutrient consumption and production cost respectively (Aldanondo-Ochoa et al., 2017). The authors calculated the improvement to cost efficiency and environmental efficiency separately.
While trade-off analysis for technically efficient farms is useful, a focus on technically inefficient farms is missing in the literature. Previous studies only examined the difference between cost and environmental efficiency and the movement of technically efficient farms to either cost or environmnetally efficient operation (Coelli et al., 2007; Nguyen et al., 2012; Welch and Barnum, 2009). However, we hypothesize that farms, including both the technically efficient and technically inefficient, may have different degrees of trade-offs while some technically inefficient farms may not have a trade-off at all. Based on this, the current study proposes to categorise the sample farms
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
into four distinct production feasibility subsets a - categorisation based on the interaction between changes in cost and nutrient amount if farms are to move to the cost or the nutrient minimising point. We focus the analysis on differing situations in which farms in each production possibility subset have distinct characteristics and differing options for improving their performance. In our empirical application, this type of analysis is shown to provide more practical management information on how farms belonging to each group tend to improve performance in either the environmental dimension or the cost dimension or both.
This study is designed to make several important contributions. First, trade-off analysis allows decision-makers to forecast the potential improvement in terms of reduction of the production cost or reduction in nutrient consumption if farms improve both cost and environmental performance. Second, we quantify the cost of becoming environmentally efficient and the environmental harm from becoming cost-efficient for both technically efficient and inefficient farms. Third, our study presents the first empirical study examining both cost and nutrient efficiency of sustainability certified and non-certified farms in Vietnam.
The rest of this study is organized as follows. Section 7.2 briefly describes the
MBP environmental and efficiency measure. Section 7.3 presents a review of existing literature of the MBP based trade-off analysis. Section 7.4 provides extra decompositions of trade-offs between cost and environmental performance. Section 7.5 present the empirical study in the context of Vietnam. Section 7.6 provides conclusions and policy recommendations.
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
7.2 The MBP approach to cost and environmental efficiency measure
Coelli et al. (2007) in proposing the MBP models to measure environmental efficiency defined the concept of nutrient-oriented environmental efficiency (NE) as the ratio of the minimum nutrient amount to the observed nutrient amount. In the input- oriented framework, the MBP model of Coelli et al. (2007) is designed to solve the following optimization problem:
NC(ya , ) =min { ax ' | xy ,∈ T} 7.1 x where, the feasible production set23, T, is defined as:
T = {(yx , ) : x can produce y } 7.2
NC is the total amount of nutrient in vector inputs (x) and NC = a’x in which vector a denotes non-negative nutrient content of each input in the input vector x. Note that some inputs such as labour or capital may not contain nutrients including nitrogen or phosphorus. According to the material balance principle, nutrients containing the inputs of fertilisers, land or water will be apportioned to the desirable outputs (i.e. coffee beans) and the balance of the nutrients apportioned between inputs and outputs that form run-offs into the environment. The balance of nutrients has the potential to cause pollution; therefore, it is desirable to minimise it. Given this output vector, fixed farms and especially small farms, can aim to reduce the nutrients in the input vector to obtain higher environmental outcomes.
The nutrient efficiency (i.e. environmental efficiency) is defined as
23 Rødseth (2016) extended this model by accounting for pollutant control activities adopted by farmers. However, in our empirical context, the new technique is not applicable because pollutant control activities were not documented and the information on the nutrient change due to such activities was not available. ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
NCax ' NE =NE = NE 7.3 NCax '
where NCNE is a solution to Equation 1 and xNE is a vector of inputs in which the nutrient amount is minimised. That is, NCNE = a’xNE.
Similarly, input-oriented technical efficiency (TE) is defined as:
ax'NC TE =θ =TE = TE 7.4 ax'NC NE can be decomposed as follows into two components - TE and the input- oriented nutrient allocative efficiency, NAE:
NCax 'ax' ax ' NE =NE = NE =TE × NE = TE× NAE NCax ' ax ' ax ' TE 7.5 This way of decomposing NE into TE and NAE is identical to the decomposition of cost efficiency into TE and cost allocative efficiency (CAE) as in the literature of cost efficiency:
wx''wx' wx CE =CE=×=×TE CE TE CAE 7.6 wx''' wx wxTE
TE and CE can be estimated using a standard input-oriented approach (i.e., DEA) whereas NE is similar to CE in term of estimation procedure in which the vector of nutrient contents of the inputs a, is used instead of input prices, w.
7.3 Existing approach to analysing a trade-off between cost and nutrient efficiency
Conventionally, a trade-off between environmental and cost performance exists if farms, aiming to increase their environmental performance, incur a higher level of production costs (and vice-versa). In the framework of MBP based efficiency discussed
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
above, the existing literature attempts to analyse such a trade-off by examining how farms are supposed to move to relevant frontiers (i.e. cost efficient or environmental
efficient frontiers) as shown in Figure 7.1.
Figure 7.1: Cost and nutrient efficiency
In Figure 7.1, a set of iso-nutrient lines (i.e. N1 and N2) is drawn using information of the nutrient content of two inputs for the observed farms. This is done in a similar way to that of the specification of price information of inputs used to construct iso-cost lines (C1 and C2). All efficiency terms, namely TE, CE, CAE, NE, and NAE, in equations (3), (4),
(5) and (6) are illustrated in Figure 7.1. It is easy to see that farms may pursue different operational objectives, although in efficiency literature farms are often benchmarked at the cost-efficient point (i.e., point C in Figure 7.1) or the environmentally efficiency point (i.e., point N in Figure 7.1). From a typical inefficient farm such as point A, wishing to achieve the minimum production cost (being cost efficient), may improve or decrease its environmental efficiency level. Similarly, opting to be more ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
environmentally efficient may result in changes in its cost efficiency. So, a trade-off may exist for farm A. Thus, for farms at the cost-efficient point C, in order for C to be nutrient efficient, their level of nutrient consumption will be at a level higher than the level of nutrient consumption at the nutrient point N. In other words, a movement from points C to N presents a trade-off. Similarly, a movement from point B to N presents a trade-off. Practical operational strategies discussed in previous literature focused on these two types of trade-offs (Coelli et al., 2007; Nguyen et al., 2012).
Coelli et al. (2007) demonstrated a typical trade-off between points C and N for 183
Belgian pig farms from 1996 to 1997. Movement from point C to point N in Figure 7.1 could reduce 5.3% of nutrient inputs with a shadow cost of 27 euros per kg. This study also focused on the trade-off between cost and nutrient consumption of technically efficient pig farms (i.e. a trade-off between moving from point B to either point C or N).
Similar analysis is used in a study of rice production in Korea by Nguyen et al. (2012) and for electricity generation in American power plants by Welch and Barnum (2009).
Aldanondo-Ochoa (2017) presents an interesting extension of trade-off analysis in which the cost and nutrient consumption levels of technically efficient farms are used as the constraint for all other farms. The authors argue that when minimising total cost, total nutrient consumption is constrained to be less than or equal to the nutrient level of the technically efficient farm. Similarly, in minimising total nutrient consumption total production cost is constrained to be less than or equal to the cost level of the technically efficient farm. This approach allows authors to have different ways of categorising farms according to each constrain type. In fact, from a managerial perspective this approach has the advantage of providing two ways for farms to improve performance. If farms are
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
supposed to reduce their costs (or their nutrient consumption), increases in the nutrient consumption level (or cost level) can be used to capture the trade-off. In the empirical application, 44% of farms are shown to use more nutrient than that of the technically efficient farms (i.e., located on the right of nutrient line N2 in Figure 7.1). The rest of the farms (56%) fall within the nutrient effient range (i.e., located on the left of nutrient line N2 in Figure 7.1). Using the cost constraint, the results indicated that only 7.3% of farms had higher costs than that of the technically efficient farms (i.e., located above cost line C2 in Figure 7.1). Thus 92.7% of farms used less or were within the cost- efficient range (i.e., located below cost line C2 in Figure 7.1).
One missing element in the litearture is a trade-off analysis for technically inefficient farms. While the existing approach can benchmark technically inefficient farms agaist technically efficient farms, in this analysis we also benchmark inefficient farms against both nutrient and cost efficient farms. Diagaphical movements in Figure is illustrated through the movement from point A to point B then respectively to points C or N. This means that trade-off analysis is only relevant for those technically inefficient farms. Approaches to policy interventions, if any, are limited to making farms technically efficient first before a trade-off actually receives attention. As shown in the next section, we provide an analysis of benchmarking technically inefficient farms at point A. This has practical benefits as our proposed approach allows different farms to pursue different mamagement strategies to improve either cost and environmental performances.
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
7.4 Elaborative approach of MBP cost and environmental efficiency
It needs to be pointed out that a trade-off between cost and environmental efficiency may not necessarily exist for all farms. In Figure 7.2, points C and N are two unique points at which the iso-quant is tangent to iso-cost line C1 and iso-nutrient line
N1. At the cost-efficient point, an iso-nutrient line N2 can be constructed parallel to the iso-nutrient line N1. Similarly, iso-cost line C2 is parallel to the iso-cost C1 crossing the environmentally efficient point, N. These typical iso-cost lines, iso-nutrient lines and the iso-quant curve in Figure 7.2 divide the feasible production set, T, into four subsets, namely Tn, Tc, Tt, and Ti (hereby called group N, C, T and I respectively).
Figure 7.2: Cost-environment trade-off Some farms belong to the subset, Tn, with n being a typical farm in Figure 7.2,
n n n representing farms using and inputs, [ , ] ∈ T , C > C2 & N1 < N < N2.
c Some belong to the subset, T , with c being a typical farm, representing farms using
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
c c c t and inputs, [ , ] ∈ T , C1 < C < C2 & N > N2. Some belong to the subset, T ,
t with t being a typical farm, representing farms using and inputs, [ , ] ∈ T , C1
t t i < C < C2 & N1 < N < N2. The rest belong to the subset, T , with i being a typical farm,
i i i representing farms using and inputs, [ , ] ∈ T , C > C2 & N > N2. Note that some farms may stay on the iso-cost line or iso-nutrient line.
Figure 7.2 also illustrates differing options for farms and resultant changes in the total value of cost. Farms belonging to each subset should have different strategies for better cost and environmental performance, regardless of improving technical efficiency.
First, for farm n, NE is achieved by a mutually beneficial interaction between nutrient consumption and the resultant cost (i.e., moving to point N in Figure 7.2). This farm should, at least maintain the level of nutrient consumption and reduce its production cost to be more environmentally and cost-efficient. To achieve maximum cost-efficiency it must consume more nutrients. Second, for farm c, cost efficiency is attained by a mutually beneficial interaction between resultant cost and nutrient consumption (i.e., moving to point C in Figure 7.2). This farm should, at least, maintain the same level of production cost and opt for a reduction in nutrient consumption to be more cost and environmentally efficient. This farm must have a larger production cost to reach the environmentally efficient.
Third, for farm t, there is always a trade-off between cost and nutrient consumption if the farm opts for either cost or environmental efficiency. In other words,
CE or NE is achieved (i.e., moving to point C and N in Figure 7.2) by a trade-off between nutrient consumption and the resultant cost. Fourth, for farm i, either opting for
CE or NE could result in a reduction in both production costs and nutrient consumption. ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
In Figure 7.2, any movement of the farm (n, c, t and i) towards the iso-quant constrained between the two arrows could improve either cost performance or environmental performance or both.
In short, farms in different groups should pursue different strategies to improve economic and environmental performance. Farmers in Groups N and T must pay a higher production cost to be environmentally efficient and this cost can be interpreted as a shadow cost. In contrast, group C and T farmers must damage environment more to reach the cost-efficient operation and this provides a measure of harm to the environment. Unlike Nguyen et al., (2012), we can calculate this cost and harm of both technically efficient and inefficient farms, where applicable. However, group I farms have no trade-off between production cost and nutrients used to pursue either cost or environmental efficiency. Farms in group T always have a trade-off between cost and nutrients used when opting for either cost or environmentally efficient operations.
Therefore, identifying the group membership of farms is useful for policy interventions which can target individual groups of farms rather than treating them indifferently.
7.5 Empirical strategies
7.5.1 Study site and data
The research site of this study is the Central Highlands, by far the largest coffee- producing area in Vietnam. The Central Highlands region includes five provinces: Dak
Lak, Lam Dong, Dak Nong, Gia Lai, and Kon Tum. In these provinces, soil and climatic conditions are especially favourable for industrial crop cultivation (i.e., coffee, pepper, and avocado). In 2015, the Central Highlands accounted for approximately 88 percent of the coffee planting area and coffee production in Vietnam (Dries et al., 2015). ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
A mixed cluster and stratified random survey of 829 coffee farmers was conducted using a face-to-face interview technique in three provinces (327 farms in Dak Lak, 282 farms in Lam Dong and 220 farms in Gia Lai provinces). Data on three crop years
2012/13 to 2014/15 was collected for each farm. The selection of districts and communes was based on several criteria including the importance of coffee as the key livelihood for farmers, geographical and ecological representation as well as the popularity of certified coffee production. Consultation with the local offices of the
Department of Agriculture and Rural Development was carried out to reduce potential
variable label N Mean St. Dev. Min Max
biases due to variations of economic conditions, production scale as well as social dimensions across the three provinces. Each local administrative unit is considered as a cluster.
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
Dried coffee output (kg/ weighted Y tree) 1,994 3.593 1.058 0.753 7.053 x1 Labour (man-days/ weighted tree) 1,994 0.308 0.173 0.030 1.548 Table x2 Land area (m2/ weighted tree) 1,994 10.485 1.660 2.519 18.593 Chemical fertilisers (kg/ weighted 7.1: x3 tree) 1,994 2.562 1.288 0.333 11.642 Desc x4 Irrigation water (m3/ weighted tree) 1,994 1.201 0.534 0.100 5.643 Other production cost (USD/ riptiv x5 weighted tree) 1,994 0.451 0.389 0.027 3.896 w1 Labour rate (USD/day) 1,994 6.593 0.887 4.444 15.556 e w2 Land and tree depreciation (USD/m2) 1,994 0.119 0.038 0.005 0.278 Price index of chemical fertilisers statis w3 (USD/kg) 1,994 0.419 0.083 0.200 0.756 w4 Irrigation water price (USD/m3) 1,994 0.000 0.000 0.000 0.000 tics Price index of other production cost w5 (USD) 1,994 1.000 0.000 1.000 1.000 of Eutroying power of labour (kg/man- a1 day) 1,994 0.000 0.000 0.000 0.000 varia a2 Eutroying power of land area (kg/m2) 1,994 0.009 0.000 0.009 0.009 bles Eutroying power of fertilisers (kg/kg a3 of fertilisers) 1,994 0.609 0.156 0.118 1.120 a4 Eutroying power of water (kg/m3) 1,994 0.056 0.000 0.056 0.056 Eutroying power of other production a5 cost Kg/USD) 1,994 0.000 0.000 0.000 0.000 7.5.2 Model specifications
It is noted that different ages of perennial trees are biologically associated with different
doses of inputs, fertility, and yield (Hasnah et al., 2004). Therefore, to account for
different ages of coffee trees, we use a nonlinear regression model to construct weights
associated with different age ranges, i.e., under eight years, from nine to fifteen years,
from sixteen to twenty years and above twenty-one years24. To avoid estimation bias
due to different ages of coffee trees, all production factors (y and x) were calculated per
weighted tree.
24 The supplementary document to this paper provides details on how weights of coffee tree ages were constructed. ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
The production factors include one output and a vector of five inputs. The output
(y) was measured in kilograms of dried coffee beans per weighted tree. Labor (x1) included both family and hired employment measured in man-day per weighted tree.
Cultivating land area (x2) was measured in squared meters per weighted tree. Chemical fertilisers (x3) measured in kilogram per weighted tree, were aggregated from different types of NPK fertilisers - Urea, Kali and other chemical fertilisers - using a price- weighted Fisher quantity index. Irrigation water amount (x4) was measured in cubic meters per weighted tree. This variable was calculated using information on total irrigation time and capacity of pumps. Other production costs (x5) measured in US dollar per weighted tree, were aggregated from other costs such as costs for organic fertilisers, machinery, fuels, and transportation.
We made several treatments to derive the unit cost of inputs. First, for the price of labour (w1), we used labour rate for hired labour as the price of labour including family labour. Second, land rental and crop depreciation cost were used as the price of cultivating an area (w2). As coffee tree density in the research region is relatively similar and coffee trees are known as a fixed asset, the price of land rentals and coffee tree depreciation are derived from these variables. Potential coffee output can be determined from this data. We found that the common rate for both land rental and crop depreciation equated to about 20 percent of the value of coffee output. The price of the cultivating area including land rental and crop depreciation was therefore calculated as 20 percent of total income from coffee divided by the coffee output. Third, the price index of chemical fertilisers (w3) was calculated as the ratio of total costs of all chemical fertilisers to the Fisher quantity index of chemical fertilisers. As for the price of
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
irrigation water (w4), given farmers freely extracted water for irrigation, the unit price of this input was set to be zero. Last, other production costs were measured in their monetary value, and hence the price of this input (w5) was set at one.
Following previous literature (Nguyen et al. 2012 and Hoang et al. 2013), we focus our analysis on nitrogen (N) and phosphorous (P) which are identified as the primary cause of eutrophication in the water system (i.e., Nguyen et al., 2012). For labour (x1) and other production costs (x5) the nutrient contents are set zero (i.e., a1 and a5).
The rest are nutrient-containing inputs, namely land area (x2), chemical fertilisers
(x3) and irrigation water (x4). However, the nutrient contents of these inputs are not available. The unit nutrient contents of land area (a2), is, therefore, assumed to be equal to a constant across different land plots. We followed Nguyen et al., (2012) to assume the nutrient contents of each square meter of cultivating land is about 0.0092 kg of eutrofying power. Chemical fertilisers and irrigation water contain both N and P. As previous studies indicated that P has more eutrofying power than N (Gold and Sims,
2005), we used a fixed set of weights (1 for N and 10 for P) to aggregate total nutrient contents of chemical and irrigation water – a methodology used by Hoang and Nguyen
(2013). Percentages of N and P in chemical fertilisers were readily available (also see, for example, Nguyen et al., 2012). Regarding the N and P content of irrigation water, it was found that this data was unavailable. We therefore used the result of a previous study (Ebina et al., 1983) which found that one cubic meter contains about 0.0202 and
0.0036 kilograms of N and P respectively. Thus, a4 was calculated to be 0.0202 +10 x
0.0036 =0.056 kg/ cubic metre. While this treatment may not be perfect, it can be argued
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
that there is scope for farmers to be more efficient in using irrigation water, therefore reducing consumption of nutrients. In addition, for a given land plot, there may be different irrigation water sources with different levels of nutrient content. Thus, choosing a constant for the unit nutrient contents of irrigation water helps take account of this problem.
In the DEA approach, it is important to specify whether production technology exhibits variable return to scale (VRS) or constant return to scale (CRS). We follow the previous studies to assume that the technology exhibits CRS (Coelli et al., 2007; Nguyen et al., 2012; Welch and Barnum, 2009). The CRS-DEA models of CE and NE are specified as:
{:wx' cc−+ y Yλλλ ≥ 0,0,0} x − X ≥ ≥ 7.7 min ii i i x,λ
{:ax' ee−+ y Yλλλ ≥ 0, x − X ≥ 0,0} ≥ min ii i i 7.8 x,λ
7.5.3 Efficiency results
Table 7.2: Cost and environmental efficiency measures
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
As shown in Table 7.2, the average cost efficiency scores in the three crop years
N Mean St. dev. Min Max Crop year 2012/13 Technical efficiency (TE) 666 0.627 0.174 0.342 1.000 Cost efficiency (CE) 666 0.458 0.148 0.071 1.000 Cost allocative efficiency (CAE) 666 0.731 0.132 0.192 1.000 Environmental efficiency (NE) 666 0.243 0.112 0.061 1.000 Nutrient allocative efficiency (NAE) 666 0.388 0.136 0.107 1.000 Crop year 2013/14 Technical efficiency (TE) 679 0.625 0.172 0.319 1.000 Cost efficiency (CE) 679 0.456 0.147 0.130 1.000 Cost allocative efficiency (CAE) 679 0.732 0.131 0.239 1.000 Environmental efficiency (NE) 679 0.175 0.091 0.039 1.000 Nutrient allocative efficiency (NAE) 679 0.279 0.104 0.081 1.000 Crop year 2014/15 Technical efficiency (TE) 649 0.642 0.165 0.340 1.000 Cost efficiency (CE) 649 0.448 0.148 0.144 1.000 Cost allocative efficiency (CAE) 649 0.697 0.131 0.311 1.000 Environmental efficiency (NE) 649 0.197 0.102 0.038 1.000 Nutrient allocative efficiency (NAE) 649 0.307 0.122 0.074 1.000 were estimated to be 45.8 percent, 45.6 percent and 44.8 percent respectively. The average cost efficiency scores for the three crop years is 45 percent, suggesting that on average coffee farms could reduce total production cost by 55 percent without any reductions in output. On average, by doing so, farms could realise an average saving of
US $4,011 per crop year. Note that this saving is about 150 percent of average profits.
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
Hence, holding output level and prices constant, farms can convert these cost savings into significantly improved profit.
Two primary sources of cost inefficiency are technical inefficiency and cost allocative efficiency. The mean of technical efficiency scores in the three crop years ranges from 62.5 percent to 64.2 percent, indicating a potential for a proportional reduction in the consumption of all five defined inputs while maintaining a constant output. The average cost allocative efficiency scores were 73.1 percent, 73.2 percent, and 69.7 percent in the crop years respectively. These results suggest that adjusting the combination of inputs are as important as reducing input consumption in terms of cost savings.
The average NE scores of the three crop years were only 24.3 percent, 17.5 percent, and 19.7 percent respectively. This suggests that the coffee farms could use an input bundle that contains 75.3 percent, 82.5 percent and 80.3 percent less eutrofying
(i.e. polluting) power of nutrients. This indicates a remarkable potential to reduce the impact of the current nutrient balance on the water and soil environments.
7.5.4 Analysis of cost savings and reduction in nutrient consumption
Another main goal of this paper is to estimate cost and nutrient savings where farms improve the current levels relating to the three operational efficiency targets: technically efficiency, cost efficiency, and environmental efficiency. Table 7.3 summarises the relative changes in total production cost and total aggregate nutrient levels for three scenarios: (1) from the existing operation to the technically efficient operation, (2) from the current operation to the cost-efficient operation, (3) from the current operation to the environmentally efficient operation. ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
Table 7.3: Cost and environmental performance
Original cost Original Changes in cost and nutrient (%) per unit nutrient per Current to
output unit output TE Current to CE Current to NE Crop year/ (USD/kg) (kg/kg) Nutrient/ Cost Nutrient Cost Nutrient # of farms Cost change change change change change
2012/13 Mean 1.34 0.83 -37.27 -54.25 2.28 -1.81 -75.74 n1 = 666 S.D. 0.45 0.46 (55.4) (-94.9) (0.7) (-1.2) (-174.6)
2013/14 Mean 1.36 0.84 -37.47 -54.38 16.28 29.40 -82.48 n2 = 679 S.D. 0.41 0.48 (56.6) (-96.4) (4.7) (17.6) (-236.6)
2014/15 Mean 1.41 0.84 -35.80 -55.22 -39.18 21.34 -80.35 n3 = 649 S.D. 0.42 0.48 (55.2) (-95.3) (-22.0) (13.1) (-199.9)
All sample Mean 1.37 0.84 -36.86 -54.61 -6.44 16.35 -79.54 N = 1,994 S.D. 0.43 0.47 (96.4) (-165.5) (-3.7) (16.9) (-335.6) Asymptotic t-statistics testing the null hypothesis of changes in cost or nutrient equal zero are in parentheses
______Chapter 7: Trade-off analysis between cost and nutrient efficiency 137
First, this study shows movement to the technically efficient operation, on average, could reduce cost and nutrient levels by about 36.8 percent, without reducing coffee output. The average reduction in cost is equivalent to US $0.50 per kg of coffee output,
US $2,603 per farm or about US $1,764 per hectare25. The average planting area of farms in the sample was about 1.55 hectares. The reduction in nutrient is equivalent to
0.31 kilograms aggregate of nutrients per kilogram of coffee output or approximately
1,560 kg of eutroying power per farm or 1,031 kg per hectare on average26. Notably there is only minor variation in these figures across three crop years.
Reductions in both cost and nutrient consumption can be tracked to the potential improvement of technical efficiency; hence interventions on how to reduce waste in input consumption are crucial and can bring substantial improvement to both economic and environmental efficiency.
The movement of farms from the current level of operation to the cost-efficient position could reduce production costs by 54.6 percent, equivalent to an average of US
$3,856 per farm or US $2,547 per hectare. Being cost efficient also could reduce total aggregate nutrients on average by about 6.4 percent. While cost changes are similar across the three years, changes in total consumption of nutrients associated with moving to cost efficient positions in two latter crop years are higher than that in the first crop year (i.e. 2 percent increase in 2012/13 and 16 percent in 2013/2014).
25 In the sample, the average number of weighted trees per farm is 1,482; the average farm area is
1.514 hectares per farm and the average total cost per weighted tree is $4.765 US. Hence, US $2,603 = US
$4.765 per weighted tree x; 1,482 trees x 36.86%, and US $1,719 = 2,603/ 1.514.
26 The average aggregate nutrients per weighted tree is 2.857 kg per weighted tree. ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
This means that by choosing the cost-efficient bundle of inputs, farms were likely to cause more damage to the surrounding environment. However, in the last crop year, obtaining the efficient bundle of inputs could also reduce aggregate nutrients by 39.2 percent released into to the surrounding environment (equivalent to 1,658 kg of eutroying power per farm on average).
As indicated in the last two columns in Table 7.3, the movement from the current production to the environmentally efficient scenario could reduce the eutroying power by 79.5 percent while it requires a 16.3 percent increase in total production cost on average. This movement could save over 3,000 kilograms of aggregate nutrients per farm, while requiring an extra cost of about US $670 per farm. In the last two sampled crop years, the result indicates that being more environmentally efficient is costly. That is such a movement could enhance nutrient reduction by 75 to 80 percent, while requiring an increase in production cost of 21.3 percent in 2014/15, 29.4 percent in
2013/14 and a decrease of 1.8 percent in 2012/13.
Given existing relative prices of nutrient containing inputs, there is a trade-off between the cost-efficient operation and environmentally efficient operation, as these two points (in Figure 7.2) are located in two different locations on the iso-quant. As often discussed in MB-based efficiency studies, this type of trade-off between cost and nutrient outcomes requires interventions in the markets of nutrient-containing inputs (i.e. fertilisers).
7.5.5 Trade-off between economic and environmental performance analysis
Closer examination of the distribution of farms in the iso-quant, iso-cost and iso- nutrient framework is particularly useful in terms of its policy implications. Within the ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
production feasibility regions, farms fall in one of the four distinct groups as depicted in
Figure 7.2. The arrows associated with points, c, n, i and t in Figure 7.2 represent possible movements these particular farms can choose to achieve higher cost efficiency or environmental efficiency without increasing nutrients or production cost respectively.
The changes in nutrient usage and costs from moving from the current production to the cost associated with environmentally efficient operations indicate that the average farm locates in group C in Figure 7.2. This means being environmentally efficient is costlier while being cost efficient is also more environmentally efficient. The changes in costs and nutrient usage that farms in a group face when moving to the cost-efficient or environmentally efficient operation, are summarised in Table 7.4.
Table 7.4: Changes in cost and nutrient usage by different groups of farms
Number Percentage Current to CE Current to NE Group of obs (%) % in C % in N % in C % in N Group C 732 36.82 -54.81 -36.47 +32.45 -83.45 Group N 118 5.94 -56.07 +39.96 -18.11 -69.56 Group T 495 24.90 -49.86 +75.42 +50.47 -73.57 Group I 643 32.34 -58.11 -44.57 -21.91 -81.94 Total 1,988 100
First, farms in group C represent 36.8 percent of the sample. These farms only face a trade-off when moving towards the environmentally efficient point. Thus, farms in this group may choose to move to the cost-efficient point to increase both cost and environmental efficiency. This would reduce production by 54.8% and the cost of nutrients by 36.5%.
Second, farms in group N consist of only 5.9 percent of the sample. These farms face a trade-off only between cost and nutrient consumption when moving to the cost- ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
efficient point (C in Figure 7.2). However, if these farms move to the environmentally efficient point, they could also reduce their production costs. Thus, farms in this group may opt for environmental efficiency, thus decreasing their production costs. By doing so, they could save 18.1% of production costs and 69.5% of nutrient consumption.
Alternatively, they may maintain the same level of nutrients and move to a lower level of production cost to improve their cost efficiency.
Third, farms in group I represents the 32.3 percent of the sampled observations.
This group has the greatest potential to reduce both cost and nutrients, meaning there is no trade-off for farms in this group. This indicates that if these farms wish to increase cost efficiency, they can also gain in environmental efficiency. This could be the result of technical and/ or allocative efficiency improvement.
Fourth, farms in group T, which accounts for 24.9 percent of the sample, always have a trade-off between cost and nutrient consumption when either moving to the cost- efficient point or environmentally efficient point. That is, they could reallocate material inputs without increasing production costs to improve environmental efficiency or vice versa. Moving to any point between the two arrows associated with point c in Figure 7.2, represents these options.
As quantified in Section 3, estimating the shadow cost of being environmentally efficient and the harm to the environment of being cost-efficient is particularly useful in devising policy options regarding incentives to increase environmental accountability.
To attain NE, farms in groups C and T representing 61.0 percent of the sample, must be burdened with a higher production cost, involving 32.4 percent and 50.5 percent increases for groups C and T respectively. To achieve CE group N and T farms ______Chapter 7: Trade-off analysis between cost and nutrient efficiency
representing 30.0 percent of the sample, must consume more nutrients, involving 39.9 percent and 75.4 percent increases for groups N and T respectively. This information can help policy-makers to adjust the costs of material inputs via taxes and subsidies to cope with the problem of overuse of chemical inputs in the agricultural sector.
In addition to technical efficiency improvement27, it is shown that farms in different groups should operate differently to achieve both cost and environmental efficiency. Farms in group C must take on higher production costs to be more environmentally efficient, but in attaining CE, this group of farms can also increase environmental performance. Thus, group C farms should seek to improve TE and CAE.
Farms in group N should concentrate on improving TE and NAE. Farms in group I should focus on increasing TE, then NAE and CAE. However, group T farms should only increase TE and at least maintain the same level of production cost and nutrient consumption.
The four groups of farms also have distinct features associated with the allocation of material inputs. Figure 7.3 represents CAE and NAE of farms categorized by farm groups. The Kernel density presented in Figure 7.3 indicates that the CAE scores of farms in group C are significantly higher than those of group N. For group C, the majority of farms in the sample (37 percent as presented in Table 7.4), tend to be cost- saving oriented rather than environmentally oriented28. On the other hand, the NAE
27 As defined in Equations 5 and 6, TE is a component of both NE and CE. Thus, any improvement of TE results in an increase in both NE and CE.
28 This trend is also correct for all sampled farms. As presented in Annex 1, in all groups, CAE is significantly higher than NAE.
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
scores of group N are higher than those of group C. This group, the minority of farms in the sample (6 percent as presented in Table 7.4), is environmentally oriented rather than cost-saving oriented. The remaining groups, I and T are also likely to be more cost- saving oriented.
Figure 7.3: CAE and NAE of farms in different groups
7.5.6 The role of the sustainability certification program
Two of the primary objectives of all certification schemes is to promote environmentally friendly practices and to help farmers to become more economically viable. Thus, cost and environmental efficiency can be used as important indicators for the effect of certification schemes. As often used in the empirical efficiency literature,
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
the Wilcoxon test29 was performed to examine the difference in the distribution of efficiency scores between the two groups of farms: certified and non-certified farms.
Note that the economic benefit of certification includes price premiums for the certified coffee output, but this lies outside of the scope of this study.
Table 7.5: Efficiency between certified and non-certified groups
Certified farms Non-certified farms Wilcoxon Efficiency n = 1,067 n = 927 Test measures Mean St. dev. Min Max Mean St. dev. Min Max (p-value) TE 0.649 0.178 0.319 1.000 0.611 0.160 0.341 1.000 0.0000 CE 0.470 0.151 0.177 1.000 0.435 0.141 0.071 0.937 0.0000 CAE 0.727 0.125 0.313 1.000 0.713 0.140 0.192 0.970 0.0743 NE 0.238 0.119 0.052 1.000 0.218 0.095 0.042 1.000 0.0019 NAE 0.365 0.133 0.100 1.000 0.359 0.127 0.089 1.000 0.5464
Table 7.5 shows that certified farms outperformed non-certified farms in terms of technical efficiency and nutrient allocative efficiency, which explain why the former have better cost and nutrient performance. These results may imply that certification programs present both economic benefits to farmers and environmental benefits to the wider community. However, it should be noted that the self-selection issue in which more efficient farms opt to participate in the certification program is not controlled for in this study due to lack of data. Hence caution is needed in the interpretation of the results.
29 It is also known as the Wilcoxon rank sum test which is commonly used in the efficiency literature to test differences between distribution of efficiency scores of a group and that of another group
(see, for example, Henzel & Wollmershäuser 2008; Choi et al. 2015; Pereira & Marques 2017; Asmild &
Hougaard 2006a).
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
Table 7.6: Efficiency between certified and non-certified farms over time Wilcoxon test of Non-certified certified vs non-certified Certified production production (p-value) Crop year 2012 2013 2014 2012 2013 2014 2012 2013 2014 # of farms 333 377 357 333 302 292 TE 0.659 0.638 0.651 0.596 0.609 0.632 0.000 0.091 0.158 CE 0.482 0.469 0.461 0.433 0.440 0.432 0.000 0.041 0.019 CAE 0.734 0.738 0.708 0.729 0.724 0.683 0.826 0.412 0.023 NE 0.276 0.210 0.232 0.236 0.194 0.221 0.000 0.255 0.546 NAE 0.420 0.327 0.356 0.400 0.319 0.354 0.049 0.799 0.696
Table 7.6 describes temporal investigation of the mean efficiency variations between certified and non-certified farms over the three crop years. While the technical efficiency level appears to increase overtime, cost and nutrient efficiency do not exhibit a consistent trend. In the first crop year, certified farms were more cost and environmentally efficient than their non-certified counterparts due to higher technical and cost and nutrient allocative efficiency. This means that certified farmers use less inputs, their input bundles are cheaper and contain less aggregated nutrients. In the second crop year certified farms performed better than non-certified farms in term of cost efficiency but mainly due to the farmers using less inputs. In the last crop year, certified farms have higher cost efficiency which is mainly driven by cost allocative efficiency (i.e. cheaper combinations of inputs).
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
7.6 Conclusions and policy implications
This study used a DEA technique with the incorporation of material balance principles to estimate and decompose cost and environmental efficiency of 1,994 coffee farms over three consecutive crop years, 2012 -2015, in the largest coffee producing region in Vietnam. Several interesting findings are evident. First, on average, there is a substantial potential to improve technical efficiency by reducing the consumption of nutrient-containing inputs and by doing so reduce production costs and the potential to cause environmental pollution. Radical reduction in the consumption of inputs has a larger impact on cost efficiency than from choosing a cheaper combination of inputs (i.e. cost allocative efficiency). However, it has a less impact on environmental efficiency than by changing the combination of nutrient-containing inputs (i.e. nutrient allocative efficiency).
It is evident that the Vietnamese coffee industry could simultaneously substantially reduce production costs and improve environmental performance. Moreover, improving environmental efficiency does not necessarily increase the production costs and vice versa. Being cost-efficient could significantly reduce production costs and the aggregate nutrients required (i.e., group C and I farms), although some farms may consume more aggregate nutrients (i.e., group N and T farms). As well, it is important to note that the change in nutrient consumption and production costs varied from crop year to crop year.
Although there is, in general, a trade-off between economic and environmental performance, trade-offs have different end purposes (i.e., improving cost efficiency or environmental efficiency) and not all farms face a trade-off. A majority of farms which are cost-saving oriented (group C), are burdened with a larger production cost when
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
pursuing environmentally efficiency, although these farms have no trade-off when improving cost efficiency. Thus, these farms should target input use efficiency, and, in this way, they can maintain current production costs while enhancing both economic and environmental performance.
The smallest proportion of farms which are environmentally oriented (group N), are faced with a trade-off when pursuing cost-efficiency. These farms should reduce nutrient consumption to enhance both cost and environmental efficiency. Alternatively, they can maintain the same level of nutrients but reduce costs to gain cost efficiency.
24.9 percent of sampled farms (group T) always faced trade-offs between cost and nutrient levels when pursuing either cost or environmental efficiency. These farms should maintain the same level of production costs and reduce nutrients to gain environmental efficiency. If they wish to increase cost efficiency, they should maintain the same level of nutrient intake but reallocate inputs to reduce production costs. They cannot reach full cost or environmentally efficiency without pushing nutrients or costs above their current levels. This suggests that there are costs involved in being more environmentally friendly.
32.3 percent of sampled farms (group I) had no trade-off. These farms were significantly inefficient in terms of both cost and environmental performance. When these farms improve cost or environmental efficiency – through TE, CAE and or NAE improvements - they will gain in both cost and environmental performance.
The cost and environmental efficiency of farms varied across farms, crop-years, sustainability certification status and regions. Generally, sustainability certified farms
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
performed better than their non-certified counterparts in terms of both economic and environmental aspects. In the three sampled crop years, certified farms were likely to be more cost-efficient than non-certified farms. As for environmental efficiency, in the first sampled crop year certified farms were more environmentally efficient than non- certified farms. However, there was no statistical difference between the two groups in the last two sampled crop years. Regarding regional difference, certified farms from Gia
Lai province attained a higher level of both cost and environmental efficiency.
These findings confirm the considerable value in applying the DEA technique and the material balance principles as a joint methodology the results of which lead to consideration of several interesting policy options. First, technical training focusing on improving the technically efficient use of inputs could help reaching a higher level of both cost and environmental efficiency. Second, availability of market information on input prices and technical support on how to choose fewer nutrient contents in material inputs, could improve cost efficiency and environmentally efficiency respectively.
Third, there exists opportunities for the formulation of policies which incentivise the substitution of high nutrient content material inputs by material inputs with lower nutrient content or the imposition of restrictions on the consumption of high nutrient inputs. This is expected to produce a coincidence of the iso-cost line with the iso- nutrient line - hence achieving both cost efficiency and environmental efficiency simultaneously. Fourth, focussing farms on improving cost efficiency or environmental efficiency should be prioritised based on location of farms in the iso-cost, iso-quant and iso-nutrient framework (group C, N, T or I). Fifth, certification management should concentrate on improving the quality of certification schemes. Given the different
______Chapter 7: Trade-off analysis between cost and nutrient efficiency
scenarios involving efficiency measures and changes in cost and nutrients, a heterogeneous and integrated policy approach is clearly needed.
Irrigation water efficiency
This chapter aims to address to answer the following research questions: (1) what are the levels of eff of irrigation water efficiency with respect to meta-technology and group technologies? (2) Based on the meta-technology ratio, which irrigation technology is more advanced than the others? (3) How does irrigation water efficiency vary across certification status and crop years?
8.1 Introduction
Given the constant rise in global demand for coffee over recent decades
(International Coffee Organization - ICO, 2017), irrigation water management has become a major challenge in coffee production (Chemura et al., 2014; Amarasinghe et al., 2015). Coffee is a water-intensive crop (Luong and Tauer, 2006), and rainfall only supplies about 25% of the potential crop demand for water (Amarasinghe et al., 2015).
Improving efficiency of irrigation water in coffee production is therefore one of the foremost objectives in order to enhance sustainable coffee production (Adams and
Ghaly, 2007). Given the growing demand for agricultural products in general and depletion of water resources, agricultural water management requires considerable adjustments (Mueller et al., 2012). There have been several attempts aimed at improving resource use efficiency - i.e., irrigation water, and economic benefits for coffee farmers. ______Chapter 8: Irrigation water efficiency
These endeavours include changes in irrigation technologies and farming management.
First, there is a trend to switching from conventional irrigation technologies - i.e., micro- basin irrigation - to more advanced technologies, i.e., the overhead sprinkler irrigation system (see, for example, Sakai et al., 2015; D’haeze et al., 2003). In is found that more advanced technologies are usually associated with larger investment cost, but lower labour requirement (Cheesman and Bennett, 2008). As well, adoption of sustainability certification schemes is expected to help coffee farmers become more socially, economically and environmentally viable.
Despite the rapid increase in adoption of sustainability certification in agricultural production which is partly driven by consumer preferences, the actual impact of certification needs to be examined. There are widespread concerns about food safety, and the extent to which environmental and social values of agricultural production are preserved (Mergenthaler et al., 2009). It is clear that consumers are willing to pay more for sustainability certified coffee. They are able to differentiate between various certification schemes and pay higher premiums for particular certified brands (Loureiro and Lotade, 2005). The indicatdions are that coffee consumers not only appreciate sustainability labels but also value the specific information on sustainability attributes
(Van Loo et al., 2015). This includes the code of conduct of common certification schemes in relation to demonstrating water conservation and efficiency of water use30 in coffee cultivation. However, a literature search indications that little rearch has been carried out on the consequences of sustainability certification on irrigation water (Vos and Boelens, 2014) - especially in coffee cultivation.
30 The code of conduct of 4C and Utz-certified is available at 4C-certification, 2015; Utz-certified, 2015. ______Chapter 8: Irrigation water efficiency
Although there are a number of studies on coffee farming (i.e., Giuliani et al.,
2017; Chiputwa et al., 2015; Dzung et al., 2011; Luong and Tauer, 2006), none focus on irrigation water efficiency under different technologies or on the effect of sustainability certification on water efficiency. Some previous studies do provide evidence on water use efficiency (i.e., Liu et al., 2016). There is a survey of studies about the effect of certification on coffee farming by Blackman and Rivera (2011) together with further recent similar studies such as Blackman & Naranjo (2012), Hardt et al., (2015), Bose et al., (2016), Giuliani et al. (2017) and Hagga et al. (2017). However, there is no study focusing on the effect of certification on irrigation water efficiency.
The current analysis, therefore, primarily focuses on assessing the efficiency of using irrigation water in small coffee farmers. To do so, sub-vector data envelopment
(DEA) meta-frontier approaches is used to allow for different irrigation technologies being used on the farms in Vietnam. A systematic review of irrigation water efficiency studies indicates that sub-vector DEA, proposed by Färe et al., (1994), is the most commonly used technique for irrigation efficiency measurement (Pereira and Marques,
2017). This technique is used to calculate technical efficiency for a subset of inputs or a single input, i.e., irrigation water, given other inputs and outputs pre-determined.
Another technique employed is the non-parametric meta-frontier (O’Donnell et al.,
2008) when different distinct technologies exist. In the current empirical context of irrigation water efficiency measurement, micro-basin irrigation and sprinkler irrigation systems are considered as two group frontiers. Unlike previous studies, which used either sub-vector DEA or meta-frontier approaches separately, we combine the two
______Chapter 8: Irrigation water efficiency
techniques to measure irrigation water efficiency between the different irrigation technologies.
Another major objective of this study is to examine differences in the efficiency levels between sustainability certified coffee farms and non-certified farms. We compare irrigation water efficiency and the meta-technology ratio between certified and non- certified farms under each irrigation system, using standard OLS models. OLS is employed as it has its own merits (McDonald, 2009), although there are debates on the second stage DEA (Simar and Wilson, 2011). Irrigation water efficiency levels with respect to group frontiers and the meta-frontier were regressed against socio-economic characteristics of farms and farmers.
This study has several significant contributions to existing literature. First, different irrigation technologies are considered in a single framework to assess overall efficiency in water use. This allows obtaining irrigation water efficiency levels with respect to micro-basin irrigation technology, sprinkler irrigation technology and meta- technologies. The result also indicates that both two irrigation system farms are on average inefficient in water use relative to their own technology, but sprinkler irrigation farms are more productive than micro-basin irrigation farms. Second, there is a mismatch of water use efficiency between sustainability certification standards and the actual effect of certification. The impact of certification on irrigation water efficiency remains unclear for both irrigation system farms. Thus, note that for policy options, irrigation technology is more important than certification labels for improving water efficiency, although water efficiency and conservation are primary farming practice concerns of being certified. Switching from micro-basin irrigation method to sprinkler
______Chapter 8: Irrigation water efficiency
irrigation method possibly helps to enhance higher potential reduction in water use in coffee production. This is also used to develop irrigation water pricing policies in farming. However, further research on cost-benefit is necessary to conclude which irrigation methods are economically viable.
The remaining of this study is organised as follows. Section 8.2 reviews relevant literature on irrigation water efficiency in relation to different irrigation methods and the effect of sustainability certification on water use in crop cultivation. Empirical strategies including empirical methodologies, data description, and model specification are presented in Section 8.3. Section 8.4 provides results and discussions on water efficiency under different irrigation methods and the effect of certification on irrigation water performance. In Section 8.5, conclusions and policy options are drawn.
8.2 Irrigation water and certification in Vietnam coffee farming
Vietnam is a global leading coffee producing country, though there are growing concerns about irrigation water for coffee farming. Over-irrigation (D’haeze et al. 2003;
Dzung, Khanh, and Dzung 2011) and inefficient use of irrigation (Amarasinghe et al.,
2015) are identified. D’haeze et al. (2005b) argue that lack of regulation in the use of water is one of the reasons leading to over-irrigation in coffee production, which, in turn, may degrade groundwater resources. If there is no policy change concerning water use, some farms will over-irrigate, creating water shortages for others. In a broader context, continued overuse of water in coffee production will negatively affect entire agricultural production systems in Vietnam’s Central Highlands. Therefore, groundwater pumping should be reduced by improving water productivity in coffee production (Amarasinghe
______Chapter 8: Irrigation water efficiency
et al. 2015). As for smallholders in developing countries like Vietnam, farming practices, however, are usually based on ad hoc local knowledge and market mechanisms targeting at short-run purposes rather than long-term sustainable goals.
Therefore, research-based policies options for efficient use of irrigation water is crucial for coffee farming as well as agricultural cultivation in the region.
Given the important role of irrigation water for coffee farming, there is a dearth of previous research on assessing irrigation water use efficiency with respect to irrigation technologies or the effect of sustainability certification. Previous studies in Vietnam coffee farming sector only focus on, for example, irrigation water amount (D’haeze et al., 2005b) or the relationship between irrigation and productive efficiency in coffee production (Cheesman and Bennett, 2008; Rios and Shively, 2006). Rios and Shively
(2006) use a cross-sectional dataset of 209 farms in the Dak Lak Province, Vietnam31 to calculate technical and cost efficiency using standard data envelopment techniques.
Number of pumps and the length of irrigation pipelines are employed as the efficiency determinants in Tobit regression models. Number of pumps tends to be positively associated with efficiency level, while the length of pipelines is inversely correlated with efficiency. The authors also suggest examining irrigation systems for additional information about irrigation water efficiency. Although there are two predominant and distinctly different irrigation methods in Vietnam’s coffee farming, namely micro-basin
31 This province has the largest coffee planting area and production in Vietnam. The data used in the current study was also collected from this province and other two major coffee producing provinces in
Vietnam.
______Chapter 8: Irrigation water efficiency
and sprinkler systems (Cheesman and Bennett, 2008), the current study is the first one examining irrigation water efficiency under different technologies.
8.3 Empirical strategies
8.3.1 Sub-vector DEA meta-frontier models
The primary aim of this study is to examine irrigation water efficiency of coffee farming as coffee is a water intensive crop. This is relevant to the efficiency analysis of an individual input, i.e., irrigation water, known as of sub-vector efficiency or input- specific efficiency. The measure of sub-vector efficiency is the technical efficiency measure for a subset of inputs rather than for the all inputs. Some inputs are treated as discretionary or variable inputs, i.e., irrigation water and some other inputs are treated as non-discretionary or fixed inputs. It means that this measure indicates the possible reduction in an individual input or a subset of inputs, given all other inputs and outputs unchanged (Speelman et al., 2008; Lansink et al., 2002). One could also use stochastic frontier analysis (SFA) or DEA to measure sub-vector efficiency (also see, for example,
Reinhard et al., 2000). In efficiency literature, a number of studies pointed out the advantages and disadvantages of SFA and DEA approaches (i.e., Watto and Mugera,
2016). For sub-vector efficiency measure, however, DEA is considered as a more appropriate technique over SFA (Lansink et al., 2002). In the irrigation water use efficiency literature, sub-vector DEA techniques have been widely used (Frija et al.,
2009; Lilienfeld and Asmild, 2007; Speelman et al., 2008).
Regarding DEA efficiency measure, one could employ CRS or VRS specification
(as mentioned in Section 6.3.1). The CRS refers the linear relationship between inputs
______Chapter 8: Irrigation water efficiency
and outputs. It means that, for example, when doubling inputs results in doubling outputs. In the case of irrigation water efficiency, measured by sub-vector DEA techniques, it is inappropriate that there is a linear relationship between irrigation water and coffee output. In addition, VRS specification is considered as more appropriate over
CRS in agricultural production (Asmild and Hougaard, 2006b). Hence, in this study we assume that the relationship between irrigation water and coffee output exhibits VRS.
A conventional approach measuring efficiency consider that all DMUs possess the same production technology. Where appropriate, DMUs should be treated under different technologies and this is known as the meta-frontier concept. O’Donnell et al.,
(2008) applied this concept in estimating DEA efficiency. This involves dividing DMUs into sub-groups and estimating group frontiers of sub-group samples. Each group of
DMUs are assumed to exhibit the same production technology.
In this empirical context, we assume that there are two distinct irrigation technologies used by the coffee farmers. These are micro-basin or flooding irrigation system known as a conventional technology and sprinkle or drip, and spray irrigation system known as an advanced technology. The meta-frontier approach is illustrated in
Figure 8.1.
______Chapter 8: Irrigation water efficiency
Figure 8.1 Irrigation water meta-frontier efficiency A combination of DEA meta-frontier approaches and sub-vector DEA techniques is employed to measure irrigation water efficiency. Meta-technical efficiency, group technical efficiency, sub-vector meta-irrigation efficiency and group sub-vector irrigation efficiency models are specified as:
Meta-technical efficiency is calculated by solving following problem,
Min{|(θθθ w ,x, y )∈ T }. This is equivalent to the linear programming problem:
θ Minθλ, θλ−≥0 subject to wi w 0, θλ−≥0 xxi 0, 8.1 −+0λ ≥ yyi 0, λλ=≥1, 0 The meta-irrigation water efficiency is specified as Min{|(,θθww wx, y )∈ T }. The equivalent linear programming problem is
θ w Minθλ, θλw −≥0 subject to wi w 0, −≥0λ xxi 0, 8.2 −+0λ ≥ yyi 0, λλ=≥1, 0 where, x and y refers input and output vectors respectively, and w is irrigation water use.θ is a scalar, and λ is a vector of constants. In this study, as irrigation water was treated as a variable or discretionary input and other inputs, namely labour, chemical fertilisers, and other costs were treated as non-discretionary inputs. Hence, the
______Chapter 8: Irrigation water efficiency
input-oriented DEA technique is used to assess the possible reduction of irrigation water, while holding coffee output and the non-discretionary inputs constant.
For group frontiers, the above problems are solved for two sub-groups. These are the micro-basin irrigation system farms, (x, y, w) ∈Tm, and the spray or drip irrigation system farms, (x, y, w) ∈Ts, where T = {Tm ∪ Ts}. T refers meta-technology set, Tm is the micro-basin irrigation technology subset and Ts is the spray or drip irrigation technology subset.
8.3.2 Irrigation water-oriented meta-technology ratio (MTR)
Another important application of the meta-frontier approach is meta-technology ratio (MTR). MTR, also known as technology gap, shows the gap between a group frontier and the meta-frontier (also shown in Figure 8.1). Many studies used output- oriented DEA techniques to estimate efficiency then calculating MTR (O’Donnell et al.,
2008). Noted that input-oriented MTR scores are interpreted differently. Following up input-oriented DEA techniques, which is used in this study, the input-oriented irrigation water MTR is calculated as follow:
meta MTRIW = IWE 8.3 IWE group or IW Emeta=× M TR IW IW E group 8.4 where, IWEmeta and IWEgroup are irrigation water efficiency scores with respect to the meta-frontier and the group frontiers respectively.
The irrigation water MTR has a value between 0 and 1 as the meta-frontier envelop the group frontiers. MTR takes a value of 1, indicating the case where meta- efficiency equal to group-efficiency and the larger value of MTR suggests the smaller
______Chapter 8: Irrigation water efficiency
gap between the meta-frontier and the group frontier. Suppose that, for illustration purposes, the input-oriented irrigation MTR of farm n’ in group technology N is 0.70, which means that only 70% of the efficient amount of irrigation water corresponding to the technology of group N, will be needed if farm n’ is operated under the meta- technology in the efficient manner. In other word, farm n’ could further reduce an extra of 30% of their technically efficient irrigation water if they would have taken the meta- technology in the efficient way. Thus, in general, a technology with a larger average irrigation MTR score is more advanced than other existing technologies in the sense of irrigation water efficiency improvement.
8.3.3 Efficiency econometric models
As mentioned in Chapter 6, there have been substantial debates on the econometric approaches for the second stage analysis examining DEA inefficiency or efficiency models. There are a number of models which have been widely used in the literature
(also see Section 6.2 for the choice of efficiency models). One of the main purposes of this study is to answer the question of whether the irrigation technology or sustainability certification contributes to irrigation water efficiency in coffee production. To control other socio-economic characteristics of farms and farmers, we use a standard OLS regression to examine the relationship between efficiency variation and irrigation technologies, and certification status of inefficient farms.
The efficiency variation models are specified as:
=+ββ ++ β + ξ IWEit011 z t... 1010 z t i , 8.5
______Chapter 8: Irrigation water efficiency
where, IWE is irrigation water efficiency, zz110,..., are contextual variables ββ ξ representing socio-economic factors, 010,..., are coefficients to be estimated and i is regression errors. The contextual variables or environmental variables are hypothesized to be associated with irrigation water efficiency variation. These variables are defined and summarised in Table 8.1.
8.3.4 Data and variables
The data used in this study were extracted from a survey of 896 coffee farmers for three crop years between 2012 and 2014 as described in Section 3.4.2. The cleaned data for variables, including production factors and contextual variables, used in this study, are defined and summarised in Table 8.1.
The coffee production factors included coffee output (y), irrigation water (x1), labour (x2), chemical fertilisers (x3), other production direct cost (x4) and land area
(x5). These variables are calculated for a weighted coffee tree32 on average. The sample was divided into two groups based on irrigation systems. We also performed a pairwise t-test to compare the means of production factors and socioeconomic characteristics of two groups. In general, there were differences in most of production factors, while the result indicated no significant difference in most of socio-economic indicators. The micro-basin or conventional system farmers tended to use more inputs than the sprinkler system farmers did, except for other direct production cost, while the sprinkler system farms gain higher coffee yield. However, due to lack of data availability on fixed cost,
32 Also see Section 6.4.3 and results of a non-linear regression for constructing the weights associated with ages of coffee trees are presented in Appendix 1.
______Chapter 8: Irrigation water efficiency
i.e., investment costs for the sprinkle irrigation system, we cannot conclude that the sprinkler system is better than the conventional system in terms of economic benefits.
This requires a further investigation on cost-benefit analysis of these two irrigation systems which may provide useful information for the coffee farming industry.
However, we, in this paper, aimed at analysing irrigation water efficiency and seeking for drivers of irrigation water efficiency, i.e., whether irrigation technologies or sustainability certification matter.
In terms of socio-economic categories of the sampled coffee farming households, the farmers who adopted the advanced irrigation technology, the sprinkler system, tended to practice in a more environmentally farming approach. For example, 74.9% of the sprinkler system farms operated under a certification scheme, while this number of the micro-basin system farms was only 38.4%. Additionally, 73.9% of farmers from the sprinkler system group reported to have shade trees and 72.8% have wind-break trees standing around their coffee plantations. However, these numbers for the micro-basin system farms were 43.4% and 42.0% respectively. Thus, we controlled for farm elevation, family size, and coffee farming experience, extension service contact, shade trees, wind-break trees, farm size and coffee tree density to examine the effect of irrigation types and certification on irrigation water efficiency.
______Chapter 8: Irrigation water efficiency
Table 8.1: Descriptive statistics of production factors per weighted tree
Sprinkler system Micro-basin system Pooled Variable P-value N1 = 716 N2 = 1,102 N = 1,818 of t-test mean SD mean SD mean SD Production factors (calculated per weighted tree) Y Coffee output (kg) 3.459 0.927 3.233 0.970 0.000 3.322 0.959 3 x1 Irrigation water (m ) 1.177 0.481 1.246 0.483 0.003 1.219 0.483 x2 Labour (man-day) 0.283 0.138 0.315 0.176 0.000 0.303 0.162 x3 Chemical fertilisers (kg) 2.394 0.980 2.585 1.250 0.000 2.509 1.155 x4 Other direct cost (1,000 VND) 10.020 7.690 10.152 8.944 0.737 10.1 8.47 0.000 10.53 x5 Land area (m2) 10.314 1.358 10.682 1.605 7 1.523 Contextual variables: z1 Irrigation type (1 for sprinkle and 0 otherwise) 1.000 0.000 0.000 0.000 N/A 0.394 0.489 z2 Certification (1 for certified and 0 for otherwise) 0.749 0.434 0.385 0.487 N/A 0.528 0.499 z3 Elevation of the farm (metres from sea level) 676 253 752 207 0.000 722 228 z4 Family size (number of people) 4.454 1.346 4.632 1.579 0.010 4.562 1.494 0.196 19.04 z Coffee farming experience (years) 5 19.291 5.938 18.881 7.491 2 6.922 z6 Extension service (1 for yes and 0 for otherwise) 0.459 0.499 0.342 0.475 N/A 0.388 0.488 z7 Shade trees (1 for yes and 0 for otherwise) 0.739 0.440 0.434 0.496 N/A 0.554 0.497 z8 Wind-break trees (1 for yes and 0 for otherwise) 0.728 0.445 0.420 0.494 N/A 0.541 0.498 z9 Coffee farm size (hectares) 1.609 1.330 1.497 1.022 0.055 1.541 1.154 z10 Coffee tree density (number of trees per hectare) 1,082 85 1,083 114 0.854 1,082 104
______Chapter 8: Irrigation water efficiency 162
8.4 Result and discussion
8.4.1 Efficiency and irrigation meta-technology ratio.
Examining three key elements, technical efficiency (TE), irrigation water efficiency (IWE) and irrigation meta-technology ratio (MTR) is expected to provide some insights into the performance of the coffee farms and irrigation technologies. The results, presented in Table 8.2, were obtained from six efficiency models using DEA techniques. First, the meta-technical efficiency was estimated from a standard DEA model (i.e., equation 8.1), which incorporated all sampled data. Second, we used a sub- vector DEA technique (i.e., equation 8.2) to calculate meta-irrigation water efficiency.
This model employed all sampled data, in which irrigation water was treated as a discretionary variable and other inputs were treated as non-discretionary variables.
Likewise, the rest of four models we used similar techniques but only captured different subgroups of the dataset, sprinkler system and micro-basin system. These were to estimated group technical efficiency and group irrigation water efficiency. The MTR scores were obtained as indicated in Equation 8.3.
In terms of TE, generally coffee farms have opportunities to improve the efficiency of input use. The meta-TE scores of coffee farmers range from 12.7% to
100%, with an average meta-TE of 69.4%. Majority of farms obtain a TE range of between 70% and 80%. This means that coffee farms may reduce their current input use level of about 30% on average and still obtain the same current level of output. The results of group-frontiers indicate that the current level of inputs is 86.3% of their
______Chapter 8: Irrigation water efficiency
efficient input mix for the sprinkler system farms and this number for the micro-basin system farms is 79.3% on average.
The irrigation water sub-vector DEA meta-efficiency result indicates a possible significant reduction in water use, while holding other inputs and output unchanged.
Majority of farms are inefficient in terms of irrigation water, i.e., half of IWE scores are smaller than 40%. The result is in line with previous empirical evidence reporting the issue of overuse of irrigation water in Vietnam coffee farming sector (D’haeze et al.,
2005b). Although coffee farmers in Vietnam extract water without paying any user fee, irrigation involves in serious environmental problems, i.e., damaging underground water resources (Dang and Shively, 2008), and economic burdens for farmers, i.e., labour and energy costs (Amarasinghe et al., 2015). Therefore, improving irrigation water efficiency could bring both environmental and economic benefits to the farmers and it is also beneficial to the coffee farming sector in long-run.
______Chapter 8: Irrigation water efficiency
Table 8.2: Distribution of efficiency scores and irrigation MTR Sprinkler system Micro-basis system Meta-technology TE IWE MTR TE IWE MTR TE IWE Efficiency and MTR distribution < 20 0 111 0 0 85 0 0 397 20-30 0 101 0 0 202 0 0 392 30-40 0 93 0 0 247 6 0 271 40-50 0 68 3 3 160 53 9 181 50-60 4 75 1 72 112 120 112 145 60-70 36 46 25 209 88 233 357 104 70-80 181 51 66 300 48 190 511 72 80-90 363 39 501 382 24 389 410 50 90-99 21 21 20 22 22 21 234 43 100 111 111 100 114 114 90 185 163 Descriptive statistics of efficiency and irrigation MTR scores Mean 0.863 0.527 0.926 0.793 0.483 0.768 0.694 0.427 St. Dev. 0.101 0.295 0.010 0.129 0.250 0.161 0.127 0.272 Min 0.549 0.080 0.474 0.471 0.116 0.334 0.467 0.068 Max 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Obs. 716 716 716 1,102 1,102 1,102 1,818 1,818 With respect to group frontiers, in general, the coffee farms using the sprinkler irrigation system operates closer to their efficient frontier in comparison to the coffee farmers using the micro-basin irrigation system. The average IWE score of the sprinkler system farms is 0.527, meaning that they could save up to 47.3% of their current amount of irrigation water, while maintaining coffee output and other inputs. This is equivalent to about 543 cubic metre per ha on average33. Similar calculations, with the mean IWE of the micro-basin system farms is 0.483, a potential reduction of irrigation water of about 606 cubic metre per ha may be achieved, as their current level is 1,173 cubic
33 For the sprinkler irrigation system group, the mean irrigation water per weighted tree is 1.177 m3
(x1 in Table 1), number of weighted trees is 1,568 per farm and average area per farm is 1.601 hectares (z7 in Table 1). Total amount of irrigation water per hectare per crop year is 1.177 x 1,568/1.601 = 1,147 m3 and the potential irrigation water saving per hectare per crop year is (1-0.473) x 1,147 = 543 m3.
______Chapter 8: Irrigation water efficiency
metre per hectare on average. In addition, IWE scores of the sprinkler system farms are more equally distributed from 20% to 99%, while the IWE scores of the other irrigation system farms are mostly distributed to a lower range of efficiency.
______Chapter 8: Irrigation water efficiency
The irrigation MTR also indicates that the sprinkler irrigation technology is more productive than the micro-basin irrigation technology. The average irrigation MTR of
the sprinkler system farms is 0.926 which is statistically higher34 than that of the conventional irrigation system with an average MTR of 0.768. This means that the gap between the sprinkler irrigation technology frontier and the meta-frontier (the average gap is only 7.4%) is very small. However, there is a significant gap between the micro- basin irrigation technology frontier and the meta-frontier. This, for example, explains that if the micro-basin irrigation system farmers choose to use a sprinkler irrigation system the efficient irrigation water needed equals 76.8% of the efficient amount of irrigation water under the micro-basin system. Figure 8.2 also shows the difference in irrigation MTR between the two technologies. In short, it clearly indicates that irrigation technologies are a key driver of irrigation water efficiency improvement.
34 We performed a common test, Wilcoxon rank sum test, and the p-value is 0.000. ______Chapter 8: Irrigation water efficiency
Figure 8.2: Irrigation MTR distribution: sprinkle vs micro-basin
8.4.2 Variation in irrigation water efficiency and sustainability certification.
In addition to irrigation technologies, another key aim of this study is to examine the effect of sustainability certification and other determinants of irrigation water efficiency. Improving water efficiency is considered to be a core objective of certification programs, at least from the viewpoint of the certification organisations. One could argue that whether being certified could help improving irrigation water efficiency or other water efficiency drivers, i.e., irrigation technologies, matters. Table 8.3 presents three efficiency variation models using a standard OLS regression technique35. In this study we also aim at factors contributing to irrigation water efficiency improvement, hence the efficient observations (i.e., IWE equals one) were excluded from the efficiency variation models.
It may be argued the existence of systematic differences between certified and non-certified groups of coffee farmers, although we attempted to minimise this problem
35 As also explained in Section 8.3.3, the choice of second stage DEA model depends on the purpose of empirical studies. The use of standard OLS is to examine the relationship between efficiency variation and its determinants (McDonald, 2009). We performed Hauseman tests and the results suggest using random-effect models for the sprinkler system efficiency ( χ2 =−=4.67,p value 0.458 ) and micro- basin system efficiency ( χ2 =−=9.07,p value 0.170) models. For the meta-technology irrigation water efficiency model, the Huasman test suggests using a fixed effect model ( χ2 =−=19.45,p value 0.013 ).
However, our interest is to examine the effect of socio-economic factors on irrigation water efficiency variation of inefficient farms and the panel data only consists of three crop years. Hence, we employed a random effect model.
______Chapter 8: Irrigation water efficiency
during sampling process as mentioned in Section 3.4.2. To address this concern, after estimating efficiency scores we employed propensity score matching36 techniques and examine the difference in key elements of this study, i.e., efficiency scores and MTR scores before and after matching. The results presented in Appendix 5 indicate no significant difference before and after matching. Therefore, we assume no issue associated with systematic differences between certified and non-certified groups.
It is consistent with the result of irrigation MTR that irrigation technologies have a significant impact on irrigation water efficiency variation, as the estimated coefficient associated with irrigation systems is statistically significant in the meta-IWE model.
This is an indication that it is appropriate to treat the two irrigation systems into two different IWE models.
36 For the use of propensity score matching to examine the effect of certification, Blackman and
Naranjo (2012) and Ibanez and Blackman (2016) provide examples of relevant covariates and outcomes.
Blackman and Naranjo (2012) demonstrate on the use of pesticides, chemical fertilisers and herbicides as outcome variables. Besides, these interested variables, Ibanez and Blackman (2016) focus on sewage disposal as an negative practice and other environmental (i.e., the use of organic fertiliser) and economic
(i.e., total cost, yield and income) positive practices. Differently from theses, we particularly evaluate the effect of certification on irrigation water efficiency under differently distinct irrigation technologies.
______Chapter 8: Irrigation water efficiency
Table 8.3: Irrigation water efficiency random effect OLS models
______Chapter 8: Irrigation water efficiency
Irrigation water efficiency Sprinkle Micro-basin Meta-frontier Irrigation type (1 for sprinkle and 0 N/A N/A 0.094*** otherwise) (0.011) Certification (1 for certified and 0 for -0.016 -0.002 0.0002 otherwise) (0.021) (0.013) (0.011) Elevation of the farm (metres from sea 0.0003*** 0.0003*** 0.0003*** level) (0.00004) (0.00004) (0.00003) Family size (number of people) 0.024*** 0.004 0.011*** (0.006) (0.004) (0.003) Coffee farming experience (years) -0.003** 0.002** -0.0003 (0.002) (0.001) (0.001) Public extension contact (1 for yes and -0.059*** -0.046*** -0.046*** 0 for otherwise) (0.018) (0.013) (0.011) Shade trees (1 for yes and 0 for -0.068*** -0.005 -0.011 otherwise) (0.022) (0.014) (0.012) Wind-break trees (1 for yes and 0 for 0.086*** 0.003 0.041*** otherwise) (0.023) (0.014) (0.013) Coffee farm size (hectares) 0.030*** 0.002 0.024*** (0.007) (0.007) (0.005) Coffee tree density (number of trees per 0.0002* 0.0002*** 0.0002*** hectare) (0.0001) (0.0001) (0.00005) Year 2013 -0.019 -0.00003 -0.004 (0.022) (0.014) (0.012) Year 2014 -0.025 -0.012 -0.018 (0.021) (0.014) (0.012) Constant -0.043 -0.037 0.172** (0.128) (0.064) (0.060) Observations 605 989 1,6604 R2 0.210 0.129 0.153 F-statistic 14.31*** 13.18*** 24.86*** Note: *p<0.1; **p<0.05; ***p<0.01 Although practicing sustainability certification is expected to improve environmental performance, i.e., IWE, the result indicates week evidence about the effect of certification on IWE improvement. Regarding irrigation technology frontiers, the sprinkler system efficiency model shows insignificant impact of certification on IWE
______Chapter 8: Irrigation water efficiency
variation, while the micro-basin model indicates a positive impact of certification on irrigation water efficiency improvement. Figure 8.3 also indicates that the irrigation water technologies are more important than being sustainability-certified. We also controlled for characteristics of farm elevation, socio-economic status and environmentally friendly farming practices, i.e., shade trees and wind-break trees to isolate the effect of certification. However, sustainability certification has an unclear impact on irrigation water efficiency, except for the micro-basin system farms.
F igur
e
8.3:
Met
a- tec
hnology ratio to the meta-frontier
Further examining other socio-characteristics of the coffee farms and farmers is expected to unveil useful information on how farms could attain irrigation water efficiency improvement. Coffee farming experience is likely to positively contribute to
IWE of the micro-basin system farmers, while it has a negative impact on IWE of the sprinkler system. It means that farming experience is a key factor for conventional irrigation practices, whereases it is not necessarily important for advanced irrigation practices.
______Chapter 8: Irrigation water efficiency
The result also shows that farm management has an important role rather than public support in water efficiency enhancement. The public extension service even has an unclear impact on helping farmers to improve IWE. This public support should focus more about environmental benefits rather than technical knowledge and skills targeting at only economic gains. In terms of farm management, coffee farmers may seek for a lower tree density to improve IWE. Regardless of irrigation technologies, both systems are likely to have a negative farm-scale effect, meaning that coffee production exhibits water inefficiency of scale. Having wind-break trees may also be beneficial to those seeking to IWE improvement, although the impact is statistically insignificant. However, another environmentally friendly practice, availability of shade trees, appears to have a significant and negative impact on IWE of the sprinkler irrigation system. Shade trees may also consume inputs, i.e., fertilisers and water, and share land area with coffee trees.
As such, this may reduce coffee output, then affecting the efficiency of coffee production. Coffee tree density is likely too high for irrigation water efficiency improvement.
8.5 Conclusions and policy implications
As identified in the literature, excessive extraction of water for irrigation in coffee production appeared to be serious threats for both environmental and economic aspects of the Vietnam coffee farming sector. Improving irrigation water efficiency of coffee production serves both economic and environmental purposes for sustainability development of the coffee industry. This study achieved several important findings and answered the question that whether irrigation water technologies or sustainability certification matters and what are drivers for improving irrigation water efficiency. ______Chapter 8: Irrigation water efficiency
First, it is consistent with previous studies that irrigation water of coffee production in Vietnam is inefficient. This is not only about the issues of overuse of irrigation water in coffee production, but also some farms may not have sufficient water for their coffee farming. As such, there is a substantial possibility to improve irrigation water efficiency in the Vietnam coffee farming sector. Second, with respect to the two distinct irrigation systems, the advanced technology is more productive than the conventional irrigation technology. The efficient farm using meta-technologies requires
92.6% of the efficient amount of water under the sprinkler irrigation technology, while it needs only 76.8% of the efficient amount of water under the micro-basin irrigation technology. Third, the sprinkler system farm can opt for a reduction of 47.3% of current irrigation water level about 543 m3 per hectare per crop year, while maintaining the same level of output and other non-discretionary inputs on average. This number for the micro-basin system farm is 52.7% of current irrigation water level or about 606 m3.
Fourth, the finding also clearly indicates that irrigation water technologies really matter, but sustainability certification appears to have week impact on improving IWE.
Additionally, some socio-economic characteristics may have an impact on water efficiency variation. These include family sizes, the availability of shade trees, wind- break trees, and coffee farm sizes. However, public extension service appears to be not effective in improving IWE.
These findings may provide several useful policy implications. First, government policies should support advanced irrigation technologies, i.e., sprinkler, drip and spray irrigation. This may include credit policies, financial support or public extension service which should focus more on irrigation water use. Second, certification standards should
______Chapter 8: Irrigation water efficiency
be restructured to incorporate more advanced or water-saving irrigation technologies, as water resources are key elements of sustainable coffee development. Third, some farm management area should also be supported, i.e., reducing coffee tree density and promoting wind-break trees. However, details on types of wind-break trees or shade trees should be investigated and they may provide different level of benefits to the coffee growers.
This study also has some limitations and further research may provide more comprehensive information on net benefits of different irrigation systems, i.e., advanced and conventional irrigation technologies. Although the findings suggest that the sprinkler irrigation system has more benefits over the micro-basin irrigation system, cost of investment and depreciation cost on these two types of irrigation are lack. As such, further research focusing on cost-benefit analysis of advanced and conventional irrigation technologies is desired. In addition, some other natural condition dimensions, such as rainfall, soil quality and the distance to a surface water source, may also complement to the irrigation water efficiency studies.
______Chapter 8: Irrigation water efficiency
Syntheses, conclusion and policy options
This chapter offers answers for the following questions: (1) What are the main findings of the studies presented in this thesis? (2) Based on the findings, what are possible focal areas that policy makers could take into account to support the sustainable future development of Vietnam’s coffee farming sector? (3) What are the limitations of this thesis and future research needs?
9.1 Introduction
Vietnam – which is the second largest coffee producer globally – has been unsustainably increasing its coffee production over the last few decades. Thus while this industry has an important role in the country’s economy and an influence on the global coffee market, it has faced major economic and environmental issues. These include inefficient use of input factors, particularly over-use of chemical fertilisers, and unsustainable extraction of water for irrigation; hence economic and environmental performance exhibit significant potential to improve.
Participating in sustainability certification programs is a common market oriented strategy that is expected to improve sustainability in coffee production, although it is accepted that, in the case of Vietnam, this is insufficient and yet more environmentally friendly farming techniques need to be adopted. That is, firstly, the choice of farming systems by coffee farmers are likely to have an effect on economic and environmental viability. Second, unstainable coffee farming may have occurred through inefficient use of environmentally detrimental inputs or environmental pressures. Third, the choice of
______Chapter 9: Syntheses, conclusion and policy options
material input combinations may also have reduced environmental sustainability. Forth, environmentally unsustainability in coffee farming may as well be a product of inefficient use of irrigation water. Identifying the relative influence of these factors on sustainable farming is therefore important in developing an appropriate policy response to achieve sustainable development in coffee farming.
Four empirical studies presented in this thesis examines a variety of aspects of productive, cost and environmental efficiency of coffee farmers in Vietnam and the impact of the above-mentioned drivers on coffee production and develops policy advice for coffee farming stakeholders on the means to increase both sustainability and efficiency. This chapter also indicates limitations of this study and provides directions for future research.
9.2 Key findings
To address the research questions, we conduct four studies corresponding with research aims 1 to 4. Each of these studies offers useful information about Vietnam’s coffee farming sector that can help its management and policy making. The main findings are summarised as follows.
9.2.1 Efficient farming systems
The first aim of this thesis is to examine different farming systems in regard to productive efficiency and identify the drivers of existing farming systems of Vietnamese coffee farmers. We measure the efficiency of coffee production and examine the variation in inefficiency levels. We derive efficiency diversification and complementary effects from the multiple outputs and inputs of the distance function. In the inefficiency
______Chapter 9: Syntheses, conclusion and policy options
model, categorical variables representing three common farming systems are employed.
We also provide an analyses which seeks to explain the presence of inefficient farming systems (see Chapter 5).
The results indicated that the technical efficiency level is relatively low, although some farmers are highly efficient. Besides the important role of several socio-economic factors, such as education, family size and ethnicity, farming diversification strategies are shown to be an important driver of inefficiency variation. Both the inefficiency model and complementary effect measures indicate that synchronised farming of coffee and the other industrial crops is the most efficient system. In contrast, farms with coffee and rice - known as the segregated system - appear to have a negative complementary effect (or diseconomies of scope) on the two crops.
We also provide detailed explanations of the findings. As indicated in previous studies, the synchronized farming system may produce bio-diversification or agronomic benefits and which translate into efficiency improvements. Another reason for the economies of scope may relate to efficient use of input factors. On the other hand, at least in the short run, the existence of inefficient segregated farming systems may partly be due to the issue of food security.
9.2.2 Eco-efficiency of Vietnam’s coffee farming sector
The second aim of this thesis is to examine eco-efficiency, its variation and the differences in eco-efficiency between sustainability certified farms and non-certified farms. We measure eco-efficiency as being where, for a given value added of coffee production, the environmental pressures are minimised. Common non-parametric tests
______Chapter 9: Syntheses, conclusion and policy options
are performed to examine the difference in eco-efficiency between certified and non- certified groups of coffee farmers. In the second stage, eco-efficiency is regressed against relevant characteristics (see Chapter 6).
We find that the eco-efficiency level is very low and exhibits wide variation across coffee farms. Although certified coffee farms appear to perform better than their non- certified counterparts, this difference converges over the sampled time period. This poses an important issue. That is, reductions in the eco-efficiency of certified farms may occur over time, during which non-certified farms are showing evidence of catching-up.
This suggests further investigation is needed on the role of positive externalities on sustainability programs. However, it does appear that the price premium derived from certification is not sufficiently significant to offset the extra effort required. Thus, it can be assumed that the quality of sustainability certification programs are more important than drastic expansion of these schemes.
The eco-efficiency variation analysis indicates several important drivers that could improve eco-efficiency. First, modern irrigation technologies, i.e., overhead sprinkler systems or spraying systems, are viable options for improving eco-efficiency. Second, availability of wind-break trees can also be an important factor in improving eco- efficiency. However, public extension services which provide trainings for farmers seem not to be effective.
9.2.3 The nexus between cost and environmental efficiency
The third aim of this thesis is to examine the cost and environmental efficiency of
Vietnam’s coffee farming sector and to analyse trade-offs between cost and
______Chapter 9: Syntheses, conclusion and policy options
environmental efficiency. The cost and Material Principle Balance -based nutrient efficiency (also known as environmental efficiency) is estimated. Cost efficiency is decomposed into technical efficiency and cost allocative efficiency. Environmental efficiency is decomposed into technical efficiency and nutrient allocative efficiency.
Different movement scenarios were assessed to analyse trade-offs between cost and environmental benefits in terms of different crop years, geographic regions and certification status (see Chapter 7).
It is found that not all farms necessarily face a trade-off between cost and environmental efficiency. There are possibilities to improve technical efficiency and hence enhance both cost and environmental efficiency. However, it is observed from the current study that if farms choose to be more cost-efficient, in doing so some will reduce nutrient consumption. Such a choice indicates farmers are moving towards a more environmentally efficient operation. As for the technically efficient farms, there exist trade-offs between cost and environmental efficiency. The level of trade-off varies across differing crop years.
Overall, sustainability certified farms show relatively higher cost and environmental efficiency levels than non-certified farms. This is primarily due to higher technical efficiency. Over the sampled period, in the first crop year, certified farms had both higher costs and environmental efficiency. However, in the next two crop years, this difference disappeared. In terms of regional differences, the farmers from Dak Lak province shows no statistical difference in cost and environmental efficiency between certified and non-certified farms. In Gia Lai province, it is evident that certified farms achieved higher cost efficiency as well as higher environmental efficiency than non-
______Chapter 9: Syntheses, conclusion and policy options
certified farms. However, in Lam Dong province, non-certified farms obtained higher environmental efficiency than certified farms.
9.2.4 Irrigation efficiency
Aim 4 of this thesis is to examine water efficiency under different irrigation technologies of certified and non-certified farms. We estimate irrigation water efficiency with respect to the meta-frontier and group frontiers corresponding to two distinct irrigation technologies: the micro-basin system and the sprinkler system. The meta- technology ratio is calculated and common non-parametric tests are performed to investigate difference in these indicators, irrigation efficiency and meta-technology ratios between certified farms and non-certified farms (see Chapter 8).
It is found that there is significant potential for Vietnam’s coffee farming sector to reduce irrigation water use. The result also confirm that sprinkler irrigation technology is more advanced than the micro-basin irrigation system. Regarding sustainable certification, there is no statistical difference in irrigation water efficiency between certified and non-certified farms: the difference in irrigation water efficiency being primarily due to irrigation technology.
9.3 Management and policy implications
The economic and environmental performance of Vietnam’s coffee farming sector is not only important for the Central Highlands communities, but also for the global coffee market. However, economic and environmental aspects that concern Vietnam’s coffee farming sector have not been sufficiently considered in past development strategies. Based on the results of this thesis, the following recommendations provide
______Chapter 9: Syntheses, conclusion and policy options
supportive measures for the sustainable future development of the Vietnam’s coffee farming sector. This addresses aim 5 of this thesis.
9.3.1 The choice of farming systems by coffee farmers
An important objective of the Vietnamese government’s sustainable coffee development plan (described in Section 2.4) is to reduce the coffee planting area from
650 thousand hectares to be about 600 thousand hectares. Of note, then, is that the coffee monoculture or the farming system with a low proportion of other industrial crops is a dominant farming system in the Central Highlands of Vietnam. Indeed in any coffee plantation there are always less productive coffee trees which may be due to aging, diseases or/and other problems.
One of the important findings of this study is that inter-cropping between coffee and other industrial crops is an efficient system. This may be due to bio-diversification or agronomic benefits (see, for example, Singh, 2000; Ogundari, 2013). In addition, diversifying to other industrial crops is an important means to alleviate market risk, as high volatility of coffee market price have been a frequent feature in the past.
Based on these observations, it is suggested that coffee farmers should be supported in diversifying their farming enterprises by increasing the proportion of other industrial crops. For example, in the short-run, unproductive coffee trees should be replaced by appropriate local industrial crops, i.e., pepper, avocado and fruit trees. In long-run, coffee farmers should be encouraged to move from rice towards other industrial crops. However, such interventions should also take into account capital investment, rice food security, risk-averse attitudes, soil features, and land terrain
______Chapter 9: Syntheses, conclusion and policy options
characteristics. In this way, diversifying to other industrial crops is also in line with the main objective of the national sustainable coffee development plan which is to reduce the coffee planting area.
9.3.2 Improving quality rather than rapid expansion of certification
schemes
The studies presented in this dissertation indicate that certified farms generally deliver a higher economic and environmental performance level than their non-certified peers in the region. During the first two sampled crop years, coffee farms obtained higher eco-efficiency, although there was no statistical difference in the last sampled crop year (Chapter 6). This is similar to the result of the studies of cost and environmental efficiency. In the first sampled crop year, both cost and environmental efficiency of certified farms were significantly higher than those of non-certified farms, whereas no difference was found in the last two sampled crop years (Chapter 7). As indicated in Chapter 7, the economic benefit of being certified, and receiving a price premium, were not sufficient to offset the extra effort required by certification standards.
It is also important to note that the difference in economic and environmental sustainability varied across geographical regions. Nevertheless, as indicated in Section
2.3, there has been rapid expansion of certified production in Vietnam, with approximately 80% of the country’s coffee production expected to be certified in 2016.
Based on these observations, policies need to be designed to focus on quality of certification rather than extensive expansion of certification schemes. In doing so, several focal areas should be taken into account in an integrated manner for the development of management and policy designs for Vietnam’s coffee farming sector. ______Chapter 9: Syntheses, conclusion and policy options
Certification standards should include financial incentives for coffee growers and include environmentally friendly farming practices. As found in Chapter 6 and 8, the use of sprinkler irrigation technology was associated with higher eco-efficiency and irrigation water efficiency. The availability of wind-break trees was also beneficial for both the economic and environmental performance of coffee production. Promoting these environmentally friendly practices should therefore be an important part of
Vietnam’s coffee farming sector management strategy based on a cost- benefit analysis which draws on the date provided in this thesis.
This thesis also provides evidence of significant differences in economic and environmental performance between certified and non-certified farms across the three provinces. As found in Chapter 6, certified farmers from Gia Lai province had both the highest cost and environmental efficiency. In terms of management and policy development, this information needs to be taken into account in identifying reasons for these differences and identifying what should be learnt from the best practices to enhance economic and environmental viability.
9.3.3 Promoting efficient use of material inputs
The findings from the study set out in Chapter 6 indicate that there is a potential to improve both cost and environmental efficiency by improving technical efficiency.
However, technical support from the government and public extension services appear to be ineffective in promoting eco-efficiency performance. This implies that a more effective extension service needs to focus on both technical and environmental perspectives. For example, training on how to efficiently use material inputs such as chemical fertilisers could help increase both cost and environmental efficiency. ______Chapter 9: Syntheses, conclusion and policy options
The findings of cost and environmental efficiency analysis in Chapter 7 also show there are trade-offs if farms endeavour to be more cost-efficient or environmentally efficient. Some technically efficient farms also face trade-offs between cost and environmental efficiency. However, there are some farms which are not faced with a trade-off between cost and environmental efficiency. This helps to confirm the view that there is a significant gap between the iso-cost line and the iso-nutrient line. Therefore, interventions from the government should promote the use of material inputs with low nutrient content and place restrictions on consumption of material inputs with high nutrient content. To do so, taxes and subsidies should be designed so that prices of material inputs have a significantly positive correlation with their nutrient content level.
This would help make the iso-cost line coincide with the iso-nutrient line and thereby eliminating the need for a trade-off. In such an environment, farms would pursue technically, cost and environmentally efficient operations which would all contribute to improving both economic and environmental sustainability of the industry.
9.3.4 Promoting the use of advanced irrigation technologies
The findings of eco-efficiency analysis presented in Chapter 6 show that advanced irrigation technologies, i.e., drip, sprinklers or spray system could help enhance eco- efficiency. In addition, the findings of irrigation water efficiency and its determinants presented in Chapter 8 also indicate that advanced irrigation technologies are more water- productive than the conventional irrigation water technologies, i.e., micro-basin or flood irrigation system. However, the effect of certification is unclear. As such, government policies, i.e., technical support and credit programs, should focus on promoting advanced irrigation technologies. For certification standards, a more specific ______Chapter 9: Syntheses, conclusion and policy options
requirement on advanced irrigation technologies could also help improve the quality of certification schemes.
9.4 Limitations of the thesis
In previous chapters, these limitations in the individual studies have been discussed. However, we restate the shortcomings here as they may be related to each other and have influenced the key results of this dissertation.
In terms of background information, data about global coffee farming and
Vietnam’s coffee sector relating to environmental aspects are limited, although multiple sources were used. For example, participating in a certification scheme is voluntary and certifiers provide farmers or groups of farmers, certification services through exporting companies or roasters. In fact, there are many competing companies and roasters operating in the various regions. However, the provincial departments of Agricultural and Rural Development do not have up-to-date and precise statistics of, for example, the number of certified farms, certified areas and the times at which they were certified.
The study presented in Chapter 4 has two major shortcomings. First, it is possible that there is a sampling bias given the sample size used in the study was only 167 coffee farms. As small scale farming is one of the typical characteristics of the Vietnam’s coffee farming sector, the sample size was relatively small. Second, there was a lack of information on specific types of the industrial crops employed. This leaves open the question as to what specific industrial crops should be chosen to diversify farming enterprises. Therefore, the results and implications in this chapter should be interpreted with caution.
______Chapter 9: Syntheses, conclusion and policy options
The studies presented in Chapters, 4, 5, 6 and 7 use surveyed datasets. It is required by certification schemes that the certified farmer keep a farming logbook recording the use of major production inputs and farming activities. Keeping a farming logbook is common in the research region as coffee production is commonly considered as running a business. However, there were cases, especially for noncertified farms, where there were difficulties in recording data where the information was unclear.
A further important shortcoming of this thesis concerns the unavailability of ecological, and climatic information. That is, ecological and climatic factors, i.e., availability of species, soil types, rainfall, and temperature, are important for coffee production. This affects the use of inputs as well as coffee output. Therefore, the results of sustainability measures should be interpreted with caution.
9.5 Further research
As for the study presented in Chapter 4, further research could add robustness to some of this study’s conclusions by examining larger datasets and panel data to better address the issue of heterogeneity in coffee farming. Further analysis of specific types of crops that deliver both efficiency gains, economies of scope and higher profitability to coffee farmers in Vietnam could also complement this study. In addition, it is desirable to investigate further the nature of the market failure in order to achieve a socially optimal level of production of ecological benefits from the diversification of synchronized crops.
There are several other directions for further research which could be undertaken based on the studies presented in Chapters 5, 6 and 7. First, other ecological and
______Chapter 9: Syntheses, conclusion and policy options
environmental dimensions such as the number of biological species and climatic conditions could be included in the analysis to provide a more comprehensive perspective. For example, the current study presented in Chapter 5 found a negative effect of shade trees on eco-efficiency improvement. However, the cultivation of shaded coffee may be highly correlated with the conservation of biodiversity which could provide considerable ecological benefits (Borkhataria et al., 2012). Second, other socio- economic aspects relevant to the welfare of coffee farming households should be incorporated in order to examine the social sustainability of coffee production. Third, a more direct assessment of the effect of certification programs on efficiency and farming efficiency is needed. Last, stochastic analysis of the relationship between economic benefits and ecological pressure could provide a useful alternative modelling approach where a stochastic production environment is more in evidence.
9.6 Conclusion
Vietnam’s coffee farming sector has experienced a significant growth in production and farming area in past decades. Yet, the literature has indicated that economically and environmentally unsustainable production is a seriously challenge facing the industry.
As a market-oriented strategy, sustainability certified production has been adopted and rapidly expanded. However the effect of these sustainability schemes remains unclear globally. Thus, there is a knowledge gap about the economic and environmental viability of Vietnam’s coffee farming sector. The purpose of this thesis is therefore to enhance the understanding of the economic and environmental performance of the Vietnam’s coffee farming sector.
______Chapter 9: Syntheses, conclusion and policy options
The four studies presented in this thesis validate the hypothesis that there is potential to improve the economic and environmental viability of Vietnam’s coffee farming sector. We concluded that the coffee farming development strategies should promote, to a far greater extent than currently, sustainable farming practices. Based on the results of the four studies several specific policy and management recommendations are provided. The focal areas for Vietnam’s coffee farming management and its policy development should be on diversification of farming systems, the quality of certification schemes and material input use efficiency. Given the biological complexities of farming activities, the proposed recommendations should be taken into account in an integrated manner for sustainable future development of the industry. However, some areas also require further research as evidenced by the limitations outlined in this thesis.
______Chapter 9: Syntheses, conclusion and policy options
______Chapter 9: Syntheses, conclusion and policy options
Bibliography
4C-certification, 2015. The 4C code of conduct [WWW Document]. 4C. URL http://cas- veri.com/files/CAS_Doc_01_4C-Code-of-Conduct_v2.1_en.pdf (accessed 7.17.17). Adams, M., Ghaly, a. E., 2007. Maximizing sustainability of the Costa Rican coffee industry. J. Clean. Prod. 15, 1716–1729. https://doi.org/10.1016/j.jclepro.2006.08.013 Ahmed, E.M., 2012. Green TFP Intensity Impact on Sustainable East Asian Productivity Growth. Econ. Anal. Policy 42, 67–78. https://doi.org/10.1016/S0313- 5926(12)50005-6 Ahmed, N., Zander, K.K., Garnett, S.T., 2011. Socioeconomic aspects of rice-fish farming in Bangladesh: Opportunities, challenges and production efficiency. Aust. J. Agric. Resour. Econ. 55, 199–219. https://doi.org/10.1111/j.1467- 8489.2011.00535.x Aigner, D.D.J.D., Lovell, C.A.K.A.K., Schmidt, P., 1977. Formulation And estimation Of Stochastic Frontier Production Function Models. J. Econom. 6, 21–37. https://doi.org/10.1016/0304-4076(77)90052-5 Aldanondo-Ochoa, A.M., Casasnovas-Oliva, V.L., Almansa-Sáez, M.C., 2017. Cross- constrained Measuring the Cost-environment Efficiency in Material Balance Based Frontier Models. Ecol. Econ. 142, 46–55. https://doi.org/10.1016/j.ecolecon.2017.06.006 Alvarez, A., Arias, C., 2004. Technical efficiency and farm size: a conditional analysis. Agric. Econ. https://doi.org/10.1111/j.1574-0862.2004.tb00192.x Amarasinghe, U.A., Hoanh, C.-T.T., D’haeze, D., Hung, T.-Q.Q., 2015. Toward sustainable coffee production in Vietnam : More coffee with less water. Agric. Syst. 136, 96–105. https://doi.org/10.1016/j.agsy.2015.02.008 Amirteimoori, a, Kordrostami, S., Rezaitabar, A., 2006. An improvement to the cost efficiency interval: A DEA-based approach. Appl. Math. Comput. https://doi.org/10.1016/j.amc.2006.02.005 Ara, S., 2002. Environmental Evaluation of Organic Rice : A Case Study in the Philippines. Analysis. Arabi, B., Doraisamy, S.M., Emrouznejad, A., Khoshroo, A., 2017. Eco-efficiency
measurement and material balance principle: an application in power plants Malmquist Luenberger Index. Ann. Oper. Res. 255, 221–239. https://doi.org/10.1007/s10479-015-1970-x Asmild, M., Hougaard, J.L., 2006a. Economic versus environmental improvement potentials of Danish pig farms. Agric. Econ. 35, 171–181. https://doi.org/10.1111/j.1574-0862.2006.00150.x Asmild, M., Hougaard, J.L., 2006b. Economic and Environmental Efficiency of Danish pig farms. Agric. Econ. 35, 171–181. Athukorala, W., Hoang, V., Wilson, C., 2015. Impact of agricultural chemicals on the environment and human health in Asia, in: Managi, S. (Ed.), The Routledge Handbook of Environmental Economics in Asia. Taylor and Francis. Ayres, R.U., 1995. Thermodynamics and Process Analysis for Future Economic Scenarios. Environ. Resour. Econ. 207–230. Bacon, C., 2005. Confronting the Coffee Crisis: Can Fair Trade, Organic, and Specialty Coffees Reduce Small-Scale Farmer Vulnerability in Northern Nicaragua? World Dev. 33, 497–511. https://doi.org/10.1016/j.worlddev.2004.10.002 Banker, A.R.D., Charnes, A., Cooper, W.W., 1984. Some Models for Estimating Technical and Scale Inefficiencies in Data Envelopment Analysis. Manage. Sci. 30, 1078–1092. https://doi.org/10.1287/mnsc.30.9.1078 Banker, R.D., Natarajan, R., 2008. Evaluating Contextual Variables Affecting Productivity Using Data Envelopment Analysis. Oper. Res. 56, 48–58. https://doi.org/10.1287/opre.1070.0460 Barham, B.L., Callenes, M., Gitter, S., Lewis, J., Weber, J., 2011. Fair Trade/Organic Coffee, Rural Livelihoods, and the “Agrarian Question”: Southern Mexican Coffee Families in Transition. World Dev. 39, 134–145. https://doi.org/10.1016/j.worlddev.2010.08.005 Barham, B.L., Weber, J.G., 2012. The Economic Sustainability of Certified Coffee: Recent Evidence from Mexico and Peru. World Dev. 40, 1269–1279. https://doi.org/10.1016/j.worlddev.2011.11.005 Basu, A.K., Hicks, R.L., 2008. Label performance and the willingness to pay for Fair Trade coffee: a cross-national perspective. Int. J. Consum. Stud. 32, 470–478.
https://doi.org/10.1111/j.1470-6431.2008.00715.x Battese, G.E., Coelli, T.J., 1995. A model for technical inefficiency effects in a stochastic frontier production function for panel data. Empir. Econ. 20, 325–332. https://doi.org/10.1007/BF01205442 Battese, G.E., Greg, C.S., 1977. Estimation of a production function model: with application to the pastoral zone of Eastern Australia. Aust. J. Agric. Econ. 21, 169– 179. Baumol, W.J., Panzar, J.C., Willig, R.D., 1988. Contestable markets and the theory of industry structure, Harcourt Brace Jovanovich. Harcourt Brace Jovanovich, Inc., New York. https://doi.org/10.2307/134928 Bell, L.W., Moore, A.D., 2012. Integrated crop-livestock systems in Australian agriculture: Trends, drivers and implications. Agric. Syst. 111, 1–12. https://doi.org/10.1016/j.agsy.2012.04.003 Bell, S., Morse, S., 2008. Sustainability Indicators Measuring the Immeasurable?, Second Edi. ed. Earthscan, London. Berg, H., 2002. Rice monoculture and integrated rice-fish farming in the Mekong Delta, Vietnam - Economic and ecological considerations. Ecol. Econ. 41, 95–107. https://doi.org/10.1016/S0921-8009(02)00027-7 Besstremyannaya, G., 2013. The impact of Japanese hospital financing reform on hospital efficiency: A difference-in-difference approach. Japanese Econ. Rev. 64, 337–362. https://doi.org/10.1111/j.1468-5876.2012.00585.x Binam, J.N., Sylla, K., Diarra, I., Nyambi, G., 2003. Factors affecting technical efficiency among coffee farmers in Cote d’Ivoire: Evidence from the centre west region. African Dev. Rev. 15, 66–76. Binam, J.N., Tonyè, J., Wandji, N., Nyambi, G., Akoa, M., 2004. Factors affecting the technical efficiency among smallholder farmers in the slash and burn agriculture zone of Cameroon. Food Policy 29, 531–545. https://doi.org/10.1016/j.foodpol.2004.07.013 Blackman, A., Naranjo, M. a., 2012. Does eco-certification have environmental benefits? Organic coffee in Costa Rica. Ecol. Econ. 83, 58–66. https://doi.org/10.1016/j.ecolecon.2012.08.001
Blackman, A., Rivera, J., 2011. Producer-level benefits of sustainability certification. Conserv. Biol. 25, 1176–85. https://doi.org/10.1111/j.1523-1739.2011.01774.x Bolwig, S., Gibbon, P., Jones, S., 2009. The Economics of Smallholder Organic Contract Farming in Tropical Africa. World Dev. 37, 1094–1104. https://doi.org/10.1016/j.worlddev.2008.09.012 Borkhataria, R., Collazo, J. a., Groom, M.J., Jordan-Garcia, A., 2012. Shade-grown coffee in Puerto Rico: Opportunities to preserve biodiversity while reinvigorating a struggling agricultural commodity. Agric. Ecosyst. Environ. 149, 164–170. https://doi.org/10.1016/j.agee.2010.12.023 Bose, A., Vira, B., Garcia, C., 2016. Does environmental certification in coffee promote “business as usual”? A case study from the Western Ghats, India. Ambio 45, 946– 955. https://doi.org/10.1007/s13280-016-0796-3 Boselie, D., 2016. The True Price of Coffee from Vietnam. Bozoğlu, M., Ceyhan, V., 2007. Measuring the technical efficiency and exploring the inefficiency determinants of vegetable farms in Samsun province, Turkey. Agric. Syst. https://doi.org/10.1016/j.agsy.2007.01.007 Bray, J.G., Neilson, J., 2017. Reviewing the impacts of coffee certification programmes on smallholder livelihoods. Int. J. Biodivers. Sci. Ecosyst. Serv. Manag. 13, 216– 232. https://doi.org/10.1080/21513732.2017.1316520 Bruno, I.P., Unkovich, M.J., Bortolotto, R.P., Bacchi, O.O.S., Dourado-Neto, D., Reichardt, K., 2011. Fertilizer nitrogen in fertigated coffee crop: Absorption changes in plant compartments over time. F. Crop. Res. 124, 369–377. https://doi.org/10.1016/j.fcr.2011.07.004 Callens, I., Tyteca, D., 1999. Towards indicators of sustainable development for firms. Ecol. Econ. 28, 41–53. https://doi.org/10.1016/S0921-8009(98)00035-4 Castro-Tanzi, S., Dietsch, T., Urena, N., Vindas, L., Chandler, M., 2012. Analysis of management and site factors to improve the sustainability of smallholder coffee production in Tarrazú, Costa Rica. Agric. Ecosyst. Environ. 155, 172–181. https://doi.org/10.1016/j.agee.2012.04.013 Chambers, R.G., Fare, R., Jaenicke, E., Lichtenberg, E., 1998. Using dominance in forming bounds on DEA models : The case of experimental agricultural data. J.
Econom. 85, 189–203. Charnes, A., Cooper, W.W., Rhodes, E., 1978. Measuring the efficiency of decision making units. Eur. J. Oper. Res. 2, 429–444. https://doi.org/10.1016/0377- 2217(78)90138-8 Chavas, J.-P., 2011. On the microeconomics of diversification under learning. J. Econ. Zeitschrift fur Natl. 104, 25–47. https://doi.org/10.1007/s00712-011-0202-3 Chavas, J.-P., Di Falco, S., 2012. On the Role of Risk Versus Economies of Scope in Farm Diversification With an Application to Ethiopian Farms. J. Agric. Econ. 63, 25–55. https://doi.org/10.1111/j.1477-9552.2011.00319.x Chavas, J.-P., Kim, K., 2010. Economies of diversification: A generalization and decomposition of economies of scope. Int. J. Prod. Econ. 126, 229–235. https://doi.org/10.1016/j.ijpe.2010.03.010 Chavas, J.P., 2008. On the economics of agricultural production. Aust. J. Agric. Resour. Econ. 52, 365–380. https://doi.org/10.1111/j.1467-8489.2008.00442.x Cheesman, J., Bennett, J., 2008. Smallholding Size, Irrigation Infrastructure, and the Efficiency of Coffee Production in Viet Nam Revisited. For. Trees Livelihoods 18, 321–335. https://doi.org/10.1080/14728028.2008.9752641 Chemura, A., Mahoya, C., Kutywayo, D., Chidoko, P., Services, S., Services, S., 2014. The growth response of coffee plants to organic manure , inorganic fertilizers and integrated soil fertility management under different irrigation. Int. J. Recycl. Org. Waste Agric. 3, 59; 1-9. https://doi.org/10.1007/s40093-014-0059-x Chen, B., Chen, G.Q., 2007. Resource analysis of the Chinese society 1980–2002 based on exergy—Part 3: Agricultural products. Energy Policy 35, 2065–2078. https://doi.org/10.1016/j.enpol.2006.07.007 Chiputwa, B., Spielman, D.J., Qaim, M., 2015. Food Standards, Certification, and Poverty among Coffee Farmers in Uganda. World Dev. 66, 400–412. https://doi.org/10.1016/j.worlddev.2014.09.006 Choi, Y., Oh, D. hyun, Zhang, N., 2015. Environmentally sensitive productivity growth and its decompositions in China: a metafrontier Malmquist-Luenberger productivity index approach. Empir. Econ. 49, 1017–1043. https://doi.org/10.1007/s00181-014-0896-5
Chung, Y.H., Färe, R., Grosskopf, S., 1997. Productivity and Undesirable Outputs: A Directional Distance Function Approach. J. Environ. Manage. 51, 229–240. https://doi.org/10.1006/jema.1997.0146 Cinemre, H. a., Ceyhan, V., Bozoğlu, M., Demiryürek, K., Kılıç, O., 2006. The cost efficiency of trout farms in the Black Sea Region, Turkey. Aquaculture 251, 324– 332. https://doi.org/10.1016/j.aquaculture.2005.06.016 Coelli, T., 1998. A multi-stage methodology for the solution of orientated DEA models. Oper. Res. Lett. 23, 143–149. https://doi.org/10.1016/S0167-6377(98)00036-4 Coelli, T., Fleming, E., 2004. Diversification economies and specialisation efficiencies in a mixed food and coffee smallholder farming system in Papua New Guinea. Agric. Econ. 31, 229–239. https://doi.org/10.1016/j.agecon.2004.09.010 Coelli, T., Lauwers, L., Huylenbroeck, G., 2007. Environmental efficiency measurement and the materials balance condition. J. Product. Anal. 28, 3–12. https://doi.org/10.1007/s11123-007-0052-8 Coelli, T., Perelman, S., 1999. A comparison of parametric and non-parametric distance functions: With application to European railways. Eur. J. Oper. Res. 117, 326–339. https://doi.org/10.1016/S0377-2217(98)00271-9 Cook, W.D., Seiford, L.M., 2009. Data envelopment analysis (DEA) - Thirty years on. Eur. J. Oper. Res. 192, 1–17. https://doi.org/10.1016/j.ejor.2008.01.032 D’haeze, D., Deckers, J., Raes, D., Phong, T.A., Loi, H. V., 2005a. Environmental and socio-economic impacts of institutional reforms on the agricultural sector of Vietnam: Land suitability assessment for Robusta coffee in the Dak Gan region. Agric. Ecosyst. Environ. 105, 59–76. https://doi.org/10.1016/j.agee.2004.05.009 D’haeze, D., Deckers, J., Raes, D., Tran, A.P., Nguyen, D.M.C., 2003. Over-irrigation of Coffea canephora in the Central Highlands of Vietnam revisited Simulation of soil moisture dynamics in Rhodic Ferralsols. Agric. Water Manag. 63, 185–202. https://doi.org/10.1016/S0378-3774(03)00181-1 D’haeze, D., Raes, D., Deckers, J., Phong, T.A., Loi, H.V., 2005b. Groundwater extraction for irrigation of Coffea canephora in Ea Tul watershed, Vietnam—a risk evaluation. Agric. Water Manag. 73, 1–19. https://doi.org/10.1016/j.agwat.2004.10.003
Dang, H., Shively, G., 2008. Coffee Boom, Coffee Bust and Smallholder Response in Vietnam’s Central Highlands. Rev. Dev. Econ. 12, 312–326. https://doi.org/10.1111/j.1467-9361.2007.00391.x Dang, T., 2003. Coffee production, social stratification and poverty in a Vietnamese Central Highland community1. Underst. Poverty Vietnam Philipp. 71. Dang, T.H., Shively, G., 2005. Coffee vs. cacao: A case study from the Vietnamese central highlands. J. Nat. Resour. Life Sci. Educ. 34, 107–111. Doutriaux, S., Geisler, C., Shively, G., 2008. Competing for Coffee Space: Development-Induced Displacement in the Central Highlands of Vietnam. Rural Sociol. 73, 528–554. https://doi.org/10.1526/003601108786471422 Dries, M., Huong, N., Ward, M., 2015. Vietnam Coffee Annual May 2015. USDA Foreign Agricultural Service. Dzung, N.A., Khanh, V.T.P., Dzung, T.T., 2011. Research on impact of chitosan oligomers on biophysical characteristics, growth, development and drought resistance of coffee. Carbohydr. Polym. 84, 751–755. https://doi.org/10.1016/j.carbpol.2010.07.066 Eakin, H., Winkels, A., Sendzimir, J., 2009. Nested vulnerability: exploring cross-scale linkages and vulnerability teleconnections in Mexican and Vietnamese coffee systems. Environ. Sci. Policy 12, 398–412. https://doi.org/10.1016/j.envsci.2008.09.003 Ebina, J., Tsutsui, T., Shirai, T., 1983. Simultaneous determination of total nitrogen and total phosphorus in water using peroxodisulfate oxidation. Water Res. 17, 1721– 1726. https://doi.org/10.1016/0043-1354(83)90192-6 Elder, S.D., Zerriffi, H., Le Billon, P., 2013. Is Fairtrade certification greening agricultural practices? An analysis of Fairtrade environmental standards in Rwanda. J. Rural Stud. 32, 264–274. https://doi.org/10.1016/j.jrurstud.2013.07.009 Emrouznejad, A., De Witte, K., 2010. COOPER-framework: A unified process for non- parametric projects. Eur. J. Oper. Res. 207, 1573–1586. https://doi.org/10.1016/j.ejor.2010.07.025 Emrouznejad, A., Yang, G., 2016. A framework for measuring global Malmquist– Luenberger productivity index with CO 2 emissions on Chinese manufacturing
industries. Energy 115, 840–856. https://doi.org/10.1016/j.energy.2016.09.032 Emrouznejad, A., Yang, G. liang, 2017. A survey and analysis of the first 40 years of scholarly literature in DEA: 1978-2016. Socioecon. Plann. Sci. 1–5. https://doi.org/10.1016/j.seps.2017.01.008 FAO, 2014. Food Agricultural Organization [WWW Document]. Statistics (Ber). Färe, R., Grosskopf, S., Lowell, K., Lovell, C.A.K., 1994. Production Frontiers. Cambridge University Press. https://doi.org/http://dx.doi.org/10.1017/CBO9780511551710 Färe, R., Grosskopf, S., Pasurka, C., Lovell, C.A.K., Pasurka, C., 1989. Multilateral Productivity Comparisons When Some Outputs are Undesirable : A Nonparametric Approach. Rev. Econ. Stat. 71, 90–98. Farrell, M.J., 1957. The Measurement of Productive Efficiency. J. R. Stat. Soc. Ser. A 120, 253–290. https://doi.org/10.2307/2343100 Fousekis, P., 2003. Technical efficiency determinants for fisheries: a study of trammel netters in Greece. Fish. Res. https://doi.org/10.1016/S0165-7836(03)00019-5 Frija, A., Chebil, A., Speelman, S., Buysse, J., Van Huylenbroeck, G., 2009. Water use and technical efficiencies in horticultural greenhouses in Tunisia. Agric. Water Manag. 96, 1509–1516. https://doi.org/10.1016/j.agwat.2009.05.006 Garcia, A.F., Shively, G.E., 2011. How Might Shadow Price Restrictions Reduce Technical Efficiency? Evidence from a Restricted DEA Analysis of Coffee Farms in Vietnam. J. Agric. Econ. 62, 47–58. https://doi.org/10.1111/j.1477- 9552.2010.00269.x Gemech, F., Mohan, S., Reeves, A., Struthers, J., 2011. Market-Based Price-Risk Management: Welfare Gains for Coffee Producers from Efficient Allocation of Resources. Oxford Dev. … 39, 49–68. https://doi.org/10.1080/13600818.2010.550399 Giuliani, E., Ciravegna, L., Vezzulli, A., Kilian, B., 2017. Decoupling Standards from Practice: The Impact of In-House Certifications on Coffee Farms’ Environmental and Social Conduct. World Dev. 96, 294–314. https://doi.org/10.1016/j.worlddev.2017.03.013 Gold, A.J., Sims, J.T., 2005. Eutrophication. Encyclopedia of Soils in the Environment,
Elsevier, Oxford. Goldfarb, D.S., Fischer, M.E., Keich, Y., Goldberg, J., 2005. A twin study of genetic and dietary influences on nephrolithiasis: A report from the Vietnam Era Twin (VET) Registry. Kidney Int. 67, 1053–1061. https://doi.org/10.1111/j.1523- 1755.2005.00170.x Haggar, J., Soto, G., Casanoves, F., Virginio, E. de M., 2017. Environmental-economic benefits and trade-offs on sustainably certified coffee farms. Ecol. Indic. 79, 330– 337. https://doi.org/10.1016/j.ecolind.2017.04.023 Hajargasht, G., Coelli, T., Rao, D.S.P., 2006. A Dual Measure of Economies of Scope, Economics Letters. https://doi.org/10.1016/j.econlet.2008.01.004 Haji, J., 2006. Production Efficiency of Smallholders’ Vegetable-dominated Mixed Farming System in Eastern Ethiopia: A Non-Parametric Approach. J. Afr. Econ. 16, 1–27. https://doi.org/10.1093/jae/ejl044 Hardt, E., Borgomeo, E., Rozely, F., Fernando, L., Pinto, G., Paul, J., dos Santos, R.F., Pinto, L.F.G., Metzger, J.P., Sparovek, G., 2015. Does certification improve biodiversity conservation in Brazilian coffee farms? For. Ecol. Manage. 357, 181– 194. https://doi.org/10.1016/j.foreco.2015.08.021 Hasnah, Fleming, E., Coelli, T., 2004. Assessing the performance of a nucleus estate and smallholder scheme for oil palm production in West Sumatra: a stochastic frontier analysis. Agric. Syst. 79, 17–30. https://doi.org/10.1016/S0308-521X(03)00043-X Henzel, S., Wollmershäuser, T., 2008. The New Keynesian Phillips curve and the role of expectations: Evidence from the CESifo World Economic Survey. Econ. Model. 25, 811–832. https://doi.org/10.1016/j.econmod.2007.11.010 Ho, T.Q., Hoang, V.-N., Wilson, C., Nguyen, T.-T., 2017. Which farming systems are efficient for Vietnamese coffee farmers? Econ. Anal. Policy 56, 114–125. https://doi.org/10.1016/j.eap.2017.09.002 Ho, T.Q., Yanagida, J.F., Illukpitiya, P., 2014. Factors Affecting Technical Efficiency of Small- holder Coffee Farming in the Krong Ana Watershed , Vietnam. Asian J. Agric. Extension, Econ. Sociol. Sociol. 3, 37–49. Hoang, V.-N., 2011. MEASURING AND DECOMPOSING ECONOMIC, ENVIRONMENTAL AND ECOLOGICAL EFFICIENCY AND
PRODUCTIVITY CHANGE IN AGRICULTURAL PRODUCTION. The University of Queensland. Hoang, V.-N., Alauddin, M., 2010. Assessing the eco-environmental performance of agricultural production in OECD countries: the use of nitrogen flows and balance. Nutr. Cycl. Agroecosystems 87, 353–368. https://doi.org/10.1007/s10705-010-9343- y Hoang, V.-N., Coelli, T., 2011. Measurement of agricultural total factor productivity growth incorporating environmental factors: A nutrients balance approach. J. Environ. Econ. Manage. 62, 462–474. https://doi.org/10.1016/j.jeem.2011.05.009 Hoang, V.-N., Nguyen, T.T., 2013. Analysis of environmental efficiency variations: A nutrient balance approach. Ecol. Econ. 86, 37–46. https://doi.org/10.1016/j.ecolecon.2012.10.014 Hoang, V.-N., Rao, D.S.P., 2010. Measuring and decomposing sustainable efficiency in agricultural production: A cumulative exergy balance approach. Ecol. Econ. 69, 1765–1776. https://doi.org/10.1016/j.ecolecon.2010.04.014 Hou, Y., Gao, Z., Heimann, L., Roelcke, M., Ma, W., Nieder, R., 2012. Nitrogen balances of smallholder farms in major cropping systems in a peri-urban area of Beijing, China. Nutr. Cycl. Agroecosystems 92, 347–361. https://doi.org/10.1007/s10705-012-9494-0 Huang, C.J., Liu, J.-T., 1994. Estimation of a non-neutral stochastic frontier production function. J. Product. Anal. 5, 171–180. Ibanez, M., Blackman, A., 2016. Is Eco-Certification a Win-Win for Developing Country Agriculture? Organic Coffee Certification in Colombia. World Dev. 82, 14–27. https://doi.org/10.1016/j.worlddev.2016.01.004 ICO, 2017a. Historical data on the global coffee trade [WWW Document]. Int. Coffee Orgnization. URL http://www.ico.org/new_historical.asp?section=Statistics (accessed 8.29.17). ICO, 2017b. Exports of all forms of coffee by all exporting countries [WWW Document]. Int. Coffee Organ. URL http://www.ico.org/historical/1990 onwards/PDF/2a-exports.pdf (accessed 5.29.17). ICO, 2017c. Total production by all exporting countries [WWW Document]. Int. Coffee
Orgnization. URL World coffee consumption (accessed 5.29.17). IDH, 2014. Sustainable Coffee Program in Vietnam. Illukpitiya, P., Yanagida, J.F., 2010. Farming vs forests: Trade-off between agriculture and the extraction of non-timber forest products. Ecol. Econ. 69, 1952–1963. https://doi.org/10.1016/j.ecolecon.2010.05.007 Jena, P.R., Chichaibelu, B.B., Stellmacher, T., Grote, U., 2012. The impact of coffee certification on small-scale producers’ livelihoods: a case study from the Jimma Zone, Ethiopia. Agric. Econ. 43, 429–440. https://doi.org/10.1111/j.1574- 0862.2012.00594.x Kilian, B., Jones, C., Pratt, L., 2006. Is sustainable agriculture a viable strategy to improve farm income in Central America ? A case study on coffee. J. Bus. Res. 59, 322–330. https://doi.org/10.1016/j.jbusres.2005.09.015 Kim, K., Chavas, J.-P., Barham, B., Foltz, J., 2012. Specialization, diversification, and productivity: a panel data analysis of rice farms in Korea. Agric. Econ. 43, 687– 700. https://doi.org/10.1111/j.1574-0862.2012.00612.x Kline, A., 2009. Sustainable Coffee Certifications A Comparison Matrix, SCAA Sustainability Committee. Kostov, P., Lingard, J., 2004. Subsistence Agriculture in Transition Economies: Its Roles and Determinants. J. Agric. Econ. 55, 565–579. https://doi.org/10.1111/j.1477-9552.2004.tb00115.x Kuit, M., Rijn, F. van, Vu, T.M.T., Pham, V.A., 2013. The Sustainable Coffee Conundrum: A study into the effects, cost and benefits of implementation modalities of sustainable coffee production in Vietnam. Kumbhakar, S.C., Ghosh, S., McGuckin, J.T., 1991. A generalized production frontier approach for estimating determinants of inefficiency in US dairy farms. J. Bus. Econ. Stat. 9, 279–286. Kuosmanen, T., Kortelainen, M., 2005. Measuring eco-efficiency of production with data envelopment analysis. J. Ind. Ecol. 9, 59–72. https://doi.org/10.1162/108819805775247846 Lansink, A.O., Pietola, K., Backman, S., 2002. Efficiency and Productivity of Conventional and Organic Farms in Finland 1994-1997. Eur. Rev. Agric. Econ. 29,
51–65. https://doi.org/10.1093/erae/29.1.51 Latynskiy, E., Berger, T., 2017. Assessing the Income Effects of Group Certification for Smallholder Coffee Farmers: Agent-based Simulation in Uganda. J. Agric. Econ. https://doi.org/10.1111/1477-9552.12212 Lauwers, L., 2009. Justifying the incorporation of the materials balance principle into frontier-based eco-efficiency models. Ecol. Econ. 68, 1605–1614. https://doi.org/10.1016/j.ecolecon.2008.08.022 Letourneau, D.K., Armbrecht, I., Rivera, B.S., Lerma, J., Carmona, E.J., Daza, M.C., Escobar, S., Galindo, V., GutiéRrez, C., LóPez, S.D., Mejía, J.L., Rangel, A.M.A., Rangel, J.H., Rivera, L., Saavedra, C.A., Torres, A.M., Trujillo, A.R., 2011. Does plant diversity benefit agroecosystems? A synthetic review. Ecol. Appl. https://doi.org/10.1890/09-2026.1 Li, Q., 1996. Nonparametric testing of closeness between two unknown distribution functions. Econom. Rev. 15, 261–274. https://doi.org/https://doi.org/10.1080/07474939608800355 Lilienfeld, A., Asmild, M., 2007. Estimation of excess water use in irrigated agriculture: A Data Envelopment Analysis approach. Agric. Water Manag. 94, 73–82. https://doi.org/10.1016/j.agwat.2007.08.005 Liu, X., Li, F., Zhang, Y., Yang, Q., 2016. Effects of deficit irrigation on yield and nutritional quality of Arabica coffee (Coffea arabica) under different N rates in dry and hot region of southwest China. Agric. Water Manag. 172, 1–8. https://doi.org/10.1016/j.agwat.2016.04.007 Loureiro, M.L., Lotade, J., 2005. Do fair trade and eco-labels in coffee wake up the consumer conscience? Ecol. Econ. 53, 129–138. https://doi.org/10.1016/j.ecolecon.2004.11.002 Luong, Q. V., Tauer, L.W., 2006. A real options analysis of coffee planting in Vietnam. Agric. Econ. 35, 49–57. https://doi.org/10.1111/j.1574-0862.2006.00138.x Lyngbaek, A.E., Muschler, R.G., Sinclair, F.L., 2001. Productivity and profitability of multistrata organic versus conventional coffee farms in costa rica. Agrofor. Syst. 53, 205–213. https://doi.org/10.1023/A:1013332722014 Macpherson, A.J., Principe, P.P., Smith, E.R., 2010. A directional distance function
approach to regional environmental–economic assessments. Ecol. Econ. 69, 1918– 1925. https://doi.org/10.1016/j.ecolecon.2010.04.012 Mandal, R., 2014. Flood, cropping pattern choice and returns in agriculture: A study of Assam plains, India. Econ. Anal. Policy 44, 333–344. https://doi.org/10.1016/j.eap.2014.08.001 Marchand, S., 2012. The relationship between technical efficiency in agriculture and deforestation in the Brazilian Amazon. Ecol. Econ. 77, 166–175. https://doi.org/10.1016/j.ecolecon.2012.02.025 Marsh, A., 2007. Diversification by smallholder farmers: Viet Nam Robusta Coffee, in: Food and Agriculture Organization of the United Nations. Masuda, K., 2016. Measuring eco-efficiency of wheat production in Japan: a combined application of life cycle assessment and data envelopment analysis. J. Clean. Prod. 126, 1–9. https://doi.org/10.1016/j.jclepro.2016.03.090 McDonald, J., 2009. Using least squares and tobit in second stage DEA efficiency analyses. Eur. J. Oper. Res. 197, 792–798. https://doi.org/10.1016/j.ejor.2008.07.039 Meeusen, W., van Den Broeck, J., 1977. Efficiency estimation from Cobb-Douglas production functions with composed error. Int. Econ. Rev. (Philadelphia). 18, 435– 444. Méndez, V.E., Bacon, C.M., Olson, M., Petchers, S., Herrador, D., Carranza, C., Trujillo, L., Guadarrama-Zugasti, C., Cordón, A., Mendoza, A., 2010. Effects of Fair Trade and organic certifications on small-scale coffee farmer households in Central America and Mexico. Renew. Agric. Food Syst. 25, 236–251. https://doi.org/10.1017/S1742170510000268 Mergenthaler, M., Weinberger, K., Qaim, M., 2009. The food system transformation in developing countries: A disaggregate demand analysis for fruits and vegetables in Vietnam. Food Policy 34, 426–436. https://doi.org/10.1016/j.foodpol.2009.03.009 Meyfroidt, P., Vu, T.P., Hoang, V.A., 2013. Trajectories of deforestation, coffee expansion and displacement of shifting cultivation in the Central Highlands of Vietnam. Glob. Environ. Chang. 23, 1187–1198. https://doi.org/10.1016/j.gloenvcha.2013.04.005
Mohan, S., 2007. Market-based price-risk management for coffee producers. Dev. Policy Rev. 25, 333–354. https://doi.org/10.1111/j.1467-7679.2007.00373.x Morrison-Paul, C.J., Johnston, W.E., Frengley, G.A.G., 2000. Efficiency in New Zealand Sheep and Beef Farming: The Impacts of Regulatory Reform. Rev. Econ. Stat. 82, 325–337. https://doi.org/10.1162/003465300558713 Mueller, N.D., Gerber, J.S., Johnston, M., Ray, D.K., Ramankutty, N., Foley, J. a., 2012. Closing yield gaps through nutrient and water management. Nature 490, 254–257. https://doi.org/10.1038/nature11420 Munksgaard, J., Christoffersen, L.B., Keiding, H., Pedersen, O.G., Jensen, T.S., 2007. An environmental performance index for products reflecting damage costs. Ecol. Econ. 64, 119–130. https://doi.org/10.1016/j.ecolecon.2007.02.006 Murthy, P.S., Madhava Naidu, M., 2012. Sustainable management of coffee industry by- products and value addition—A review. Resour. Conserv. Recycl. 66, 45–58. https://doi.org/10.1016/j.resconrec.2012.06.005 Mutoko, M.C.M., Hein, L., Shisanya, C. a. C., 2014. Farm diversity, resource use efficiency and sustainable land management in the western highlands of Kenya. J. Rural Stud. 36, 108–120. https://doi.org/10.1016/j.jrurstud.2014.07.006 Nguyen, T.-T., Hoang, V.-N., Seo, B., 2012. Cost and environmental efficiency of rice farms in South Korea. Agric. Econ. (United Kingdom) 43, 369–378. https://doi.org/10.1111/j.1574-0862.2012.00589.x O’Donnell, C.J., Coelli, T., 2005. A Bayesian approach to imposing curvature on distance functions. J. Econom. 126, 493–523. https://doi.org/10.1016/j.jeconom.2004.05.011 O’Donnell, C.J., Rao, D.S.P., Battese, G.E., 2008. Metafrontier frameworks for the study of firm-level efficiencies and technology ratios. Empir. Econ. 34, 231–255. https://doi.org/10.1007/s00181-007-0119-4 Ofori-Bah, A., Asafu-Adjaye, J., 2011. Scope economies and technical efficiency of cocoa agroforesty systems in Ghana. Ecol. Econ. 70, 1508–1518. https://doi.org/10.1016/j.ecolecon.2011.03.013 Ogundari, K., 2013. Crop diversification and technical efficiency in food crop production: A study of peasant farmers in Nigeria. Int. J. Soc. Econ. 40, 267–287.
https://doi.org/10.1108/03068291311291536 Padrón, B.R., Burger, K., 2015. Diversification and Labor Market Effects of the Mexican Coffee Crisis. World Dev. 68, 19–29. https://doi.org/10.1016/j.worlddev.2014.11.005 Papke, L.E., Wooldridge, J.M., 1996. Econometric Methods for Fractional Response Variables With an Application to 401 ( K ) Plan Participation Rates. J. Appl. Econom. 11, 619–632. Pereira, H., Marques, R.C., 2017. An analytical review of irrigation efficiency measured using deterministic and stochastic models. Agric. Water Manag. 184, 28–35. https://doi.org/10.1016/j.agwat.2016.12.019 Philpott, S.M., Bichier, P., Rice, R., Greenberg, R., 2007. Field-testing ecological and economic benefits of coffee certification programs. Conserv. Biol. 21, 975–985. https://doi.org/10.1111/j.1523-1739.2007.00728.x Picazo-Tadeo, A.J., Beltrán-Esteve, M., Gómez-Limón, J.A., 2012. Assessing eco- efficiency with directional distance functions. Eur. J. Oper. Res. 220, 798–809. https://doi.org/10.1016/j.ejor.2012.02.025 Picazo-Tadeo, A.J., Gómez-Limón, J. a, Reig-Martínez, E., 2011. Assessing farming eco-efficiency: a Data Envelopment Analysis approach. J. Environ. Manage. 92, 1154–64. https://doi.org/10.1016/j.jenvman.2010.11.025 Piorr, H.-P., 2003. Environmental policy, agri-environmental indicators and landscape indicators. Agric. Ecosyst. Environ. 98, 17–33. https://doi.org/10.1016/S0167- 8809(03)00069-0 Pitt, M.M., Lee, L.F., 1981. The measurement and sources of technical inefficiency in the Indonesian weaving industry* 1. J. Dev. Econ. 9, 43–64. Pittman, R., 1983. Multilateral Productivity Comparisons with Undesirable Outputs. Econ. J. 93, 883–891. Rahman, S., 2009. Whether crop diversification is a desired strategy for agricultural growth in Bangladesh? Food Policy 34, 340–349. https://doi.org/10.1016/j.foodpol.2009.02.004 Ramalho, E.A., Ramalho, J.J.S., Coelho, L.M.S., 2015. Exponential regression of panel data fractional response models with an application to firm capital structure.
Ranjan Jena, P., Grote, U., 2016. Fairtrade Certification and Livelihood Impacts on Small-scale Coffee Producers in a Tribal Community of India. Appl. Econ. Perspect. Policy 0, ppw006. https://doi.org/10.1093/aepp/ppw006 Reifschneider, D., Stevenson, R., 1991. Systematic departures from the frontier: A framework for the analysis of firm inefficiency. Int. Econ. Rev. (Philadelphia). 32, 715–723. Reinhard, S., Lovell, C.A.K.A.K., Thijssen, G., 1999. Econometric estimation of technical and environmental efficiency: an application to Dutch dairy farms. Am. J. Agric. Econ. 81, 44–60. https://doi.org/10.2307/1244449 Reinhard, S., Lovell, C.A.K., Thijssen, G.J., 2000. Environmental efficiency with multiple environmentally detrimental variables ; estimated with SFA and DEA. Eur. J. Oper. Res. 121, 287–303. Reinhard, S., Thijssen, G., 2000. Nitrogen efficiency of Dutch dairy farms : a shadow cost system approach. Eur. Rev. Agric. Econ. 27, 167–186. Rios, A.R., Shively, G.E., 2006. Farm size, irrigation infrastructure, and the efficiency of coffee production in Vietnam. For. Trees Livelihoods 16, 397–412. https://doi.org/10.1080/14728028.2006.9752575 Rødseth, K.L., 2016. Environmental efficiency measurement and the materials balance condition reconsidered. Eur. J. Oper. Res. 250, 342–346. https://doi.org/10.1016/j.ejor.2015.10.061 Ruben, R., Fort, R., 2012. The Impact of Fair Trade Certification for Coffee Farmers in Peru. World Dev. 40, 570–582. https://doi.org/10.1016/j.worlddev.2011.07.030 Rubio, N., Barros, S., 2017. Brazil Coffee Annual Report. USDA Foreign Agricultural Service Sakai, E., Barbosa, E.A.A., de Carvalho Silveira, J.M., de Matos Pires, R.C., 2015. Coffee productivity and root systems in cultivation schemes with different population arrangements and with and without drip irrigation. Agric. Water Manag. 148, 16–23. https://doi.org/10.1016/j.agwat.2014.08.020 Simar, L., Wilson, P.W., 2011. Two-stage DEA: Caveat emptor. J. Product. Anal. 36, 205–218. https://doi.org/10.1007/s11123-011-0230-6 Simar, L., Wilson, P.W., 2007. Estimation and inference in two-stage, semi-parametric
models of production processes. J. Econom. 136, 31–64. https://doi.org/10.1016/j.jeconom.2005.07.009 Simar, L., Zelenyuk, V., 2006. On Testing Equality of Distributions of Technical Efficiency Scores. Econom. Rev. 25, 497–522. https://doi.org/10.1080/07474930600972582 Singh, S., 2000. Contract farming for agricultural diversification in the India Punjab: A study of performance and problems. Indian J. Agric. Econ. 56, 139–165. Smith, V.H., Joye, S.B., Howarth, R.W., 2006. Eutrophication of freshwater and marine ecosystems. Limnol. Oceanogr. 51, 351–355. https://doi.org/10.4319/lo.2006.51.1_part_2.0351 Solís, D., Bravo-Ureta, B.E., Quiroga, R.E., 2009. Technical efficiency among peasant farmers participating in natural resource management programmes in Central America. J. Agric. Econ. 60, 202–219. https://doi.org/10.1111/j.1477- 9552.2008.00173.x Speelman, S., D’Haese, M., Buysse, J., D’Haese, L., D’Haese, M., Buysse, J., D’Haese, L., 2008. A measure for the efficiency of water use and its determinants, a case study of small-scale irrigation schemes in North-West Province, South Africa. Agric. Syst. 98, 31–39. https://doi.org/10.1016/j.agsy.2008.03.006 TECHNOSERVE, 2013. Vietnam: A Business Case For Sustainable Coffee Production. Technoserve: Business Solutions for Poverty. Tellarini, V., Caporali, F., 2000. An input/output methodology to evaluate farms as sustainable agroecosystems: an application of indicators to farms in central Italy. Agric. Ecosyst. Environ. 77, 111–123. https://doi.org/10.1016/S0167- 8809(99)00097-3 The Sustainable Coffee Program, 2013. A business case for sustainable coffee production, The sustainable coffee program. Thiam, A., Bravo-Ureta, B.E., Rivas, T.E., 2001. Technical efficiency in developing country agriculture: a meta-analysis. Agric. Econ. 25, 235–243. https://doi.org/10.1111/j.1574-0862.2001.tb00204.x Tran, Q., Smith, G., 2017. Vietnam Coffee Annual Report. USDA Foreign Agricultural Service
Ullah, A., Perret, S.R., Gheewala, S.H., Soni, P., 2016. Eco-efficiency of cotton- cropping systems in Pakistan: an integrated approach of life cycle assessment and data envelopment analysis. J. Clean. Prod. 134, 623–632. https://doi.org/10.1016/j.jclepro.2015.10.112 USDA, 2017. Coffee: World Markets and Trade. United Stages Department of Agriculture (USDA) - Foreign Agricultural Service. http://apps.fas.usda.gov/psdonline/circulars/coffee.pdf Utz-certified, 2015. The Utz-certified code of conduct [WWW Document]. Utz. URL https://utz.org/?attachment_id=3309 (accessed 7.17.17). Utz-certified, 2014. Utz certified Impact Report. https://utzcertified.org/attachments/article/26582894/utz-impact-report-web-1.pdf Valkila, J., 2014. Do Fair Trade Pricing Policies Reduce Inequalities in Coffee Production and Trade? Dev. Policy Rev. 32, 475–493. Valkila, J., 2009. Fair Trade organic coffee production in Nicaragua — Sustainable development or a poverty trap ? Ecol. Econ. 68, 3018–3025. https://doi.org/10.1016/j.ecolecon.2009.07.002 van Asten, P.J.A., Wairegi, L.W.I., Mukasa, D., Uringi, N.O., 2011. Agronomic and economic benefits of coffee-banana intercropping in Uganda’s smallholder farming systems. Agric. Syst. 104, 326–334. https://doi.org/10.1016/j.agsy.2010.12.004 van de Walle, D., Gunewardena, D., 2001. Sources of ethnic inequality in Viet Nam. J. Dev. Econ. 65, 177–207. https://doi.org/10.1016/S0304-3878(01)00133-X Van Loo, E.J., Caputo, V., Nayga, R.M., Seo, H.S., Zhang, B., Verbeke, W., 2015. Sustainability labels on coffee: Consumer preferences, willingness-to-pay and visual attention to attributes. Ecol. Econ. 118, 215–225. https://doi.org/10.1016/j.ecolecon.2015.07.011 Varghese, S.K., Veettil, P.C., Speelman, S., Buysse, J., Van Huylenbroeck, G., 2013. Estimating the causal effect of water scarcity on the groundwater use efficiency of rice farming in South India. Ecol. Econ. 86, 55–64. https://doi.org/10.1016/j.ecolecon.2012.10.005 Vedenov, D., Houston, J., Cardena, G., Cardenas, G., 2007. Production Efficiency and Diversification in Mexican Coffee-Producing Districts. J. Agric. Appl. Econ. 39,
749–763. Vellema, W., Buritica Casanova, A., Gonzalez, C., D’Haese, M., 2015. The effect of specialty coffee certification on household livelihood strategies and specialisation. Food Policy 57, 13–25. https://doi.org/10.1016/j.foodpol.2015.07.003 Villano, R., Fleming, E., Fleming, P., 2010. Evidence of farm-level synergies in mixed- farming systems in the Australian Wheat-Sheep Zone. Agric. Syst. 103, 146–152. https://doi.org/10.1016/j.agsy.2009.11.005 Vlontzos, G., Niavis, S., Manos, B., 2014. A DEA approach for estimating the agricultural energy and environmental efficiency of EU countries. Renew. Sustain. Energy Rev. 40, 91–96. https://doi.org/10.1016/j.rser.2014.07.153 Vos, J., Boelens, R., 2014. Sustainability standards and the water question. Dev. Change 45, 205–230. https://doi.org/10.1111/dech.12083 Wall, G., 1977. EXERGY- A USEFUL CONCEPT WITHIN RESOURCE ACCOUNTING. Waller, J.M., Bigger, M., Hillocks, R.J., 2007. Coffee Pests, Diseases and Their Management. CABI Publishing. Wang, X., 2010. Irrigation Water Use Efficiency of Farmers and Its Determinants: Evidence from a Survey in Northwestern China. Agric. Sci. China 9, 1326–1337. https://doi.org/10.1016/S1671-2927(09)60223-6 Watto, M., Mugera, A., 2016. Wheat farming system performance and irrigation efficiency in Pakistan: A bootstrapped metafrontier approach. Int. Trans. Oper. Res. 0, 1–21. https://doi.org/10.1111/itor.12314 Welch, E., Barnum, D., 2009. Joint environmental and cost efficiency analysis of electricity generation. Ecol. Econ. 68, 2336–2343. https://doi.org/10.1016/j.ecolecon.2009.03.004 Willig, R.D., 1979. Multiproduct Technology and Market Structure. Am. Econ. Rev. 69, 346–351. Wollni, M., Brümmer, B., 2012. Productive efficiency of specialty and conventional coffee farmers in Costa Rica: Accounting for technological heterogeneity and self- selection. Food Policy 37, 67–76. https://doi.org/10.1016/j.foodpol.2011.11.004 Wollni, M., Zeller, M., 2007. Do farmers benefit from participating in specialty markets
and cooperatives ? The case of coffee marketing in Costa Rica 1. Agric. Econ. 37, 243–248. Wossink, A., Denaux, Z.S., 2006. Environmental and cost efficiency of pesticide use in transgenic and conventional cotton production. Agric. Syst. 90, 312–328. https://doi.org/10.1016/j.agsy.2006.01.004 Wu, J., Wu, T., 2012. Sustainability Indicators and Indices: An Overview, in: Handbook of Sustainable Management. Imperial College Press, London, pp. 65–86. Yu-Ying Lin, E., Chen, P.-Y., Chen, C.-C., 2013. Measuring the environmental efficiency of countries: a directional distance function metafrontier approach. J. Environ. Manage. 119, 134–42. https://doi.org/10.1016/j.jenvman.2013.01.015 Zhou, H., Yang, Y., Chen, Y., Zhu, J., 2018. Data envelopment analysis application in sustainability: The origins, development and future directions. Eur. J. Oper. Res. 264, 1–16. https://doi.org/10.1016/j.ejor.2017.06.023
Appendix
Appendix 1: Derivation of output complementary effect
Deriving cross derivative of the distance function with respect to Yi and Yj:
pqp 111αβτ pqpijlnXYX j q ij ln j q ij ln j αβ 222 = ii jjj DAXYX(,xy ) ∏∏∏ii i ∏ Y i ∏ Y i iii i i
qp p 11βτ 1 α qqijlnYX j q ij ln j p p ij ln X j βα22 2 = i jji j DYYYAXX(,xy ) ∏∏ii ∏ i ∏ i ∏ i ii i i i
11qp 1 p qpββ++lnYX τ ln αα + ln X iijjijj22 iijj 2 = jj j DY(,)xy ∏∏ii AX ii
qp ββ++11 τ >= call fii(y )=and ijjijj ln Y ln X , j i i 1.... q 22jj
qpp =+++αα1 lnDfii (y )ln YA ln ( i ijji ln XX )ln iij2
∂∂∂∂lnDDDD ln 1 ∂ff()y ===+iilnY ∂∂∂∂∂i YDYDYYYiiiii
∂∂∂∂lnDDDD ln 1 ∂ff()y ===+iilnY ∂∂∂∂∂i YDYDYYYiiiii
∂∂DD∂∂ff()yy f f () ⇔=(lniiYD + ),sincej>ithen0 i = = i D ∂∂i ∂ ∂ YYii Y i Y i YY ii
2 ∂∂DD1 ∂f ()y ∂D f ()y =+f ()y i D , likewise, = j D ∂∂YY Y ∂ Yi ∂ Y ∂ ij i j j YYjj
∂2DD1 fff()yy∂∂ff()yy () ()y () =+=+Dfjji()y ii D ∂∂i ∂ ∂ YYij Y i Y j Y j Y i Y j Y j
qp ββ++11 τ >= Since, fii(y )=and ijjijj lnYXjiiq ln , 1.... then 22jj
∂f ()y 11 i = β ∂ ij YYjJ2
2 ∂ DDffji()y ()y 11 D 1 Therefore, =+=+ββff()y ()y ∂∂ ij j i ij YYij Y i Y j Y j22 YY ij
=β =β Since the variables are normalized at their means, then fii()y and fjj()y .
Then the second derivative of the input distance function with respect to Yi and Yj is negative if:
1 ()0ββ+< β ij2 ij
Appendix 2: Weighted number of coffee trees
Std. Label Coef. Err. t P>t 1 wg grafted trees β 2.5299 0.12 21.16 0.0000 2 w8- trees with ages under 8 years β 2.0060 0.08 25.27 0.0000 3 w9-15 trees with ages from 9 to 15 years β 2.6676 0.06 45.07 0.0000 4 w16-20 trees with ages from 16 to 20 years β 3.1348 0.05 58.05 0.0000 5 w21+ trees with ages over 21 years β 2.8356 0.06 47.06 0.0000
Appendix 3: Mean efficiency of farms dropped out and joined in certification programs
Dropped out Joined in Mean eco-efficiency Mean eco-efficiency No. of levels No. of levels Period farms Before Afterfarms Before After
2012 to 2013 26 0.4244 0.4116 60 0.3364 0.3303 2013 to 2014 5 0.3955 0.4496 3 0.5561 0.6083 2012 to 2014 31 0.4533 0.4099 59 0.3526 0.3956
Appendix 4: Ethics approval and survey
PARTICIPANT INFORMATION FOR QUT RESEARCH PROJECT – Survey –
Measuring socio-economic and environmental sustainability of coffee production using econometric and frontier-based models: Vietnam as a case study
QUT Ethics Approval Number 1500000663
RESEARCH TEAM Principal Researcher: Thong Quoc Ho PhD student Associate Researchers: Dr Vincent Hoang Principal Supervisor Professor Clevo Wilson Associate Supervisor School of Economics and Finance, QUT Business School, Queensland University of Technology (QUT)
DESCRIPTION This project is being undertaken as part of a PhD study for Thong Quoc Ho.
The purpose of this project is to take an in-depth examination at factors affecting efficiency in coffee production and other related issues, such as the use of chemical fertilizers, irrigation water, sustainability-certified production and living conditions of coffee farmers. This is to formulating appropriate policies to better manage current coffee production in Vietnam, profitably, environmentally and socially.
You are invited to participate in this project because you have involved in coffee production and coffee output from your province has a significant role for Vietnam coffee industry and global coffee market as well.
PARTICIPATION Participation will involve completing a set of survey that will take approximately 2 hours of your time. Questions will include information about coffee production related issues, characteristics of farms and farmers, and other socio-economic factors.
Example questions: Coffee production Crop year 12/13 Crop year 13/14 Crop year 14/15
What was/is the size of your ______ha ______ha ______ha farm in ha?
How many coffee trees did/ do ______trees ______trees ______trees you have in total?
What is the main type of O (1) flooding O (1) flooding O (1) flooding irrigation that you used? O (2) dripping O (2) dripping O (2) dripping O (3) sprinkling O (3) sprinkling O (3) sprinkling O (-99) N/A O (-99) N/A O (-99) N/A
Your participation in this project is entirely voluntary. If you agree to participate you do not have to complete any question(s) you are uncomfortable answering. Your decision to participate or not participate will in no way impact upon your current or future relationship with QUT or any associated external organisations. If you do agree to participate you can withdraw from the project at any time until 2 weeks after the survey has been submitted. Any identifiable information already obtained from you will be destroyed.
EXPECTED BENEFITS It is expected that this project will not directly benefit you. However, it may benefit researchers and policy makers to providing better policy options for the coffee industry.
To compensate you for your contribution should you choose to participate the research team will provide you with out-of-pocket expenses 80,000 VND.
RISKS There are no risks beyond normal day-to-day living associated with your participation in this project.
PRIVACY AND CONFIDENTIALITY Any information obtained in connection with this research project that can identify you will remain confidential unless required by law.
Any data collected as part of this project will be stored securely as per QUT’s Management of research data policy. And non-identifiable data collected in this project may be used as comparative data in future projects or stored on an open access database for secondary analysis.
The project is funded by Department of Foreign Affairs and Trade of Australia and Economy and Environment Program for Southeast Asia and they will not have access to individually identifiable or raw data obtained during the project.
CONSENT TO PARTICIPATE We would like to ask you to sign a written consent form (enclosed) to confirm your agreement to participate.
QUESTIONS / FURTHER INFORMATION ABOUT THE PROJECT If you have any questions or require further information please contact one of the researchers listed below.
Thong Quoc Ho [email protected] +61 7313 87409 Vincent Hoang [email protected] +61 7 3138 4325 Clevo Wilson [email protected]
CONCERNS / COMPLAINTS REGARDING THE CONDUCT OF THE PROJECT QUT is committed to research integrity and the ethical conduct of research projects. However, if you do have any concerns or complaints about the ethical conduct of the project you may contact the QUT Research Ethics Unit on +61 7 3138 5123 or email [email protected]. The QUT Research Ethics Unit is not connected with the research project and can facilitate a resolution to your concern in an impartial manner.
Thank you for helping with this research project. Please keep this sheet for your information.
-Household survey- General notes: -88: don’t know -99: N/A General information
What is the gender of respondent? O (1) Male O (2) Female
What is the marital status of the respondent? O (1) Married O (2) Single O (3) Widow(er)
Is the respondent household’s head? O (1) Yes O (0) No
If not, what is relationship of respondent to O (1) Wife household head? O (2) Husband O (3) Daughter O (4) Son O (5) Other relative O (6) Other, specify ( )
What is the name of the village you live in? ( )
What is the name of the commune you live in? O (1) Ea Quang O (15) Ia Yok O (2) Hòa Tiến O (16) Ea Phê O (3) Ea Yông O (17) Ea Tóh O (4) Ea Kênh O (18) Hòa Thắng O (5) Quảng Tiến O (19) Ia Tiêm O (6) Ea Tân O (20) Chư Pơng O (7) Buôn Trấp O (21) Nghĩa Hưng O (8) Ea Kao O (22) Trà Đa O (9) Cư Êbua O O (10) Al Ba (23)______O (11) Ia H'Lop O O (12) Dun (24)______O (13) Thị trấn Chư Sê O O (14) Ia Sao (25)______O (26)______O (27)______O (28)______
What are the coordinates of the farmer’s house? Zone ( ) X m ( ) Y m ( ) O (-88) Don’t know
What is the altitude above sea level? Meters ( ) O (-88) Don’t know
Respondent Husband / Wife
What is your and your partner’s age? ( ) ( )
What is your and your partner’s ethnicity? O (1) Kinh O (1) Kinh O (2) Other, please O (2) Other, please specify specify ( ) ( ) O (-99) N/a
Education level of the respondent O (0) none O (1) primary school not finished O (2) primary school finished O (3) secondary school not finished O (4) secondary school finished O (5) secondary school not finished O (6) secondary school finished O (7) more advanced
Education level of husband/ wife O (0) none O (1) primary school not finished O (2) primary school finished O (3) secondary school not finished O (4) secondary school finished O (5) high school not finished O (6) high school finished O (7) more advanced O (-99) N/a
How many people in your family? ( )
How many children (<16 years) do you have? ( )
To what extent are you/family dependent on coffee ( Percentage ) farming?
Respondent Husband / Wife
Were you born in the same region that you O (0) No O (0) No currently live in? O (1) Yes O (1) Yes O (-88) Don’t know O (-88) Don’t know
Were your parents coffee farmers? O (0) No O (0) No O (1) Yes O (1) Yes O (-88) Don’t know O (-88) Don’t know
How many years have you been active in coffee farming? ( ) ( )
Are you member of any coffee related organization, Crop year Crop year Crop year program or project, if so, which? 12/13 13/14 14/15 O (0) No O (1) Yes
Cooperative O (0) O (1) O (0) O (1) O (0) O (1)
Farmers union O (0) O (1) O (0) O (1) O (0) O (1)
Research project O (0) O (1) O (0) O (1) O (0) O (1)
Training project O (0) O (1) O (0) O (1) O (0) O (1)
Certification project O (0) O (1) O (0) O (1) O (0) O (1)
Other, specify: O (0) O (1) O (0) O (1) O (0) O (1) ( ______)
Crop year 12/13 Crop year 13/14 Crop year 14/15
Did/do you need credit to O (0) No O (0) No O (0) No cover the costs of coffee O (1) Yes O (1) Yes O (1) Yes production?
If you needed credit, were/are O (0) No O (0) No O (0) No you able to obtain it? O (1) Yes O (1) Yes O (1) Yes O (-99) N/A O (-99) N/A O (-99) N/A
In previous crop years, what O (1) Red book O (1) Red book O (1) Red book type of ownership did you O (2) Rent or lease O (2) Rent or lease O (2) Rent or lease have of the farm? O (3) Traditional land O (3) Traditional land O (3) Traditional land use right use right use right O (4) Squatting O (4) Squatting O (4) Squatting O (5) Other, specify O (5) Other, specify O (5) Other, specify
( ) ( ) ( )
Coffee production practices Production
---. Planting area Crop year 12/13 Crop year 13/14 Crop year 14/15
What was/is the size of your ______ha ______ha ______ha farm in ha?
What was/ is the size of your ______ha (Robusta) ______ha (Robusta) _____ha (Robusta) farm in ha for different types of ______ha (Arabica) ______ha (Arabica) ______ha (Arabica) coffee? ______ha (Others) ______ha (Others) ______ha (Others)
How much was/is planted with ______ha ______ha ______ha coffee in ha)
How much dried bean coffee in Mt did you produce? ______tons ______tons ______tons
How many coffee trees did/ do ______trees ______trees ______trees you have in total?
How many trees with ages ______trees ______trees ______trees ranged between 1–8 years
How many trees with ages ______trees ______trees ______trees ranged between 9–15 years
How many trees with ages ______trees ______trees ______trees ranged between 16 – 20 years
How many trees with ages ______trees ______trees ______trees ranged between 21 years and above
Did/do you use shade trees in O (0) No O (0) No O (0) No your coffee field(s)? O (1) In some of my O (1) In some of my O (1) In some of my fields fields fields O (2) In all of my fields O (2) In all of my fields O (2) In all of my fields
Did/do you use windbreaks in O (0) No O (0) No O (0) No
your coffee field(s)? O (1) In some of my O (1) In some of my O (1) In some of my fields fields fields O (2) In all of my fields O (2) In all of my fields O (2) In all of my fields Irrigation
--- Crop year Crop year Crop year 12/13 13/14 14/15
What is the main type of O (1) flooding O (1) flooding O (1) flooding irrigation that you used? O (2) dripping O (2) dripping O (2) dripping O (3) sprinkling O (3) sprinkling O (3) sprinkling O (-99) N/A O (-99) N/A O (-99) N/A
On average, how many irrigation rounds do you use ______round(s) ______round(s ______round( per season? ) s)
For round 1, what was the ______hp ______hp ______hp capacity of pump?
How long did it take for round 1 ______hours ______hours ______hours
For round 2, what was the ______hp ______hp ______hp capacity of pump?
How long did it take for round 2 ______hours ______hours ______hours
For round 3, what was the ______hp ______hp ______hp capacity of pump?
How long did it take for round 3 ______hours ______hours ______hours
For round 4, what was the ______hp ______hp ______hp capacity of pump?
How long did it take for round 4 ______hours ______hours ______hours
For round 5, what was the ______hp ______hp ______hp capacity of pump?
How long did it take for round 5 ______hours ______hours ______hours
Did/do you have access to O (0) No O (0) No O (0) No electricity at the farm house? O (1) Yes O (1) Yes O (1) Yes O (-99) N/A O (-99) N/A O (-99) N/A
--. What are/were sources of irrigation water? Crop year Crop year Crop year 12/13 13/14 14/15
Underground water, if no go to q61 O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
How many wells do/did you have on your farm? ______well(s) ______well(s) ______well (s)
What is/was average depth of wells? ______metres ______metres ______metres
Ground water (Lakes/ reservoirs/ stream) O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
What was/is the average distance from your farm to the ______metres ______metres ______ground water sources? metres
How many pumps do you own for coffee production? _____pump(s) _____pump(s) _____pum p(s)
What were/is capacity of the pump most commonly ______hp ______h ______used? p _hp
---. Crop year Crop year Crop year 12/13 13/14 14/15
Did you hire irrigation-related O (0) No O (0) No O (0) No services? O (1) Yes O (1) Yes O (1) Yes If yes, how much did you pay for following items:
Fuel or electricity for pumping ( ) ( ) ( ) VND VND VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A
Equipment rental (pumps, pipes) ( ) ( ) ( ) VND VND VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A
Full package rent ( ) ( ) ( ) VND VND VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A
*Excluding labour cost
Fertiliser
Chemical fertiliser
--. How much chemical fertiliser Crop year Crop year Crop year of which type did you apply on 12/13 13/14 14/15 your coffee planted area during the last season?
Urea ______kg ______kg ______kg
What was the average price per ______kilogram of Urea? vnd/kg vnd/kg vnd/kg
Sulphate Ammonium (SA) ______kg ______kg ______kg
What was the average price per ______kilogram of SA? vnd/kg vnd/kg vnd/kg
Phosphorus 16.5%37 ______kg ______kg ______kg
What was the average price per ______kilogram of Phosphorus? vnd/kg vnd/kg vnd/kg
Potassium ______kg ______kg ______kg
What was the average price per ______kilogram of Potassium? vnd/kg vnd/kg vnd/kg
NPK ______kg ______kg ______kg
(____/_____/____) (____/_____/____) (____/_____/____) (indicate type) (indicate type) (indicate type)
What was the average price per ______kilogram of NPK? vnd/kg vnd/kg vnd/kg
Other, specify: ( ) ______kg ______kg ______kg
What was the average price per ______kilogram of the above fertiliser? vnd/kg vnd/kg vnd/kg
Other, specify: ( ) ______kg ______kg ______kg
What was the average price per ______kilogram of the above fertiliser? vnd/kg vnd/kg vnd/kg
37 Also called Lan, Dam Lan
Organic fertiliser
Did you apply compost, mulch or O (0) no O (0) no O (0) no organic manure during the last O (1) Yes O (1) Yes O (1) Yes crop year?
How much did you apply? ______tons ______tons ______tons (Excluding the compost made from coffee pulp)
How much did you spend for ( ) ( ) ( ) these organic fertilisers? VND VND VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A
Cover pulp waste O (0) Used to make O (0) Used to make O (0) Used to make compost compost compost O (0) Sold O (0) Sold O (0) Sold O (1) Burned O (1) Burned O (1) Burned O (2) Thrown away O (2) Thrown away O (2) Thrown away
If coffee pulps were used to make compost, how much compost did you produce? ______tons ______tons ______tons
How much did it cost for making ( ) ( ) ( ) compost from coffee pulp? VND VND VND (excluding labour cost) O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A Pesticides
Insecticides (target insects)
---. Crop year 12/13 Crop year 13/14 Crop year 14/15
Did you use insecticides? O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
What is the amount of insecticides applied? ______litter/kg ______litter/kg ______litter/kg
How much did you spend ( ) ( ) ( ) on insecticides during the VND VND VND last season? O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know O (-99) N/A O (-99) N/A O (-99) N/A
What are common brands ______of the insecticides (indicate ______3 brands)? ______
Herbicides (target grasses)
---. Crop year 12/13 Crop year 13/14 Crop year 14/15
Did you use herbicide? O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
What is the amount of herbicides applied? ______litter/kg ______litter/kg ______litter/kg
How much did you spend on ( ) ( ) ( ) herbicides during the last VND VND VND season? O (-88) Don’t O (-88) Don’t O (-88) Don’t know know know O (-99) N/A O (-99) N/A O (-99) N/A
What are common brands of ______the herbicides (indicate 3 ______brands)? ______Fungicides (target fungal pets)
---. Crop year 12/13 Crop year 13/14 Crop year 14/15
Did you use fungicides? O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
What is the amount of fungicides applied? ______litter/kg ______litter/kg ______litter/kg
How much did you spend ( ) ( ) ( ) on fungicides during the last VND VND VND season? O (-88) Don’t O (-88) Don’t O (-88) Don’t know know know O (-99) N/A O (-99) N/A O (-99) N/A
What are common brands ______of the fungicides (indicate 3 ______brands)? ______Credit
---. Crop year 12/13 Crop year 13/14 Crop year 14/15
During last season, did you use O (0) No O (0) No O (0) No credit? O (1) Yes O (1) Yes O (1) Yes
How much money did you ( ) ( ) ( ) borrow last season? VND VND VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t
O (-99) N/A know know O (-99) N/A O (-99) N/A
How much interest did you pay ( ) ( ( on that loan? VND ) VND ) VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
If so, what was/is the main O (1) Bank O (1) Bank O (1) Bank source of credit for you? O (2) People’s O (2) People’s O (2) People’s Credit Fund Credit Fund Credit Fund O (3) Collector O (3) Collector O (3) Collector O (4) Exporter O (4) Exporter O (4) Exporter O (5) Projects from O (5) Projects O (5) Projects social organisations from social from social O (6) Family organisations organisations O (7) Friends O (6) Family O (6) Family O (8) Other, please O (7) Friends O (7) Friends specify O (8) Other, O (8) Other, ______please specify please specify O (-99) N/A ______O (-99) N/A O (-99) N/A Labour
---. Crop year Crop year Crop year 12/13 13/14 14/15
During last season, did you use O (0) No O (0) No O (0) No hired labour? O (1) Yes O (1) Yes O (1) Yes
How many man-days did you and your family members ______days ______days ______da spend on your coffee ys production?
How many man-days did you exchange with other farmers or ______days ______days ______da you relatives? ys
How many man-days did you ______days ______days ______da hire? ys
How much money per man-day ( ) ( ) ( did you pay for hired labour? VND VND ) VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Other production costs
---. How much did you spent last Crop year Crop year Crop year season on the following items? 12/13 13/14 14/15
Warehouse facilities ( ) ( ) ( ) VND VND VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Renting machines (excluding ( ) ( ( for irrigation) VND ) VND ) VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Transportation ( ) ( ( VND ) VND ) VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Drying costs (energy) ( ) ( ( VND ) VND ) VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Other production cost, please ( ) ( ( specify: VND ) VND ) VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t ( O (-99) N/A know know ) O (-99) N/A O (-99) N/A
Utz Certification
---. Utz certification Crop year Crop year Crop year 12/13 13/14 14/15
Are you certified Utz (if not, go to O (0) No O (0) No O (0) No q105)? O (1) Yes O (1) Yes O (1) Yes
When did you get certified? _____/_____ (month/ year)
Can you indicate if certification has ( ) ( ( led to additional costs in cash, and VND ) VND ) VND how much per season? O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
How much labour time do you spend to ensure compliance with ( ) ( ( the Utz code that you would days ) days ) days otherwise not have spent? O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
And what is this time spent on?
Administration ( ) ( ) ( )
days days days O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Additional work time in field ( ) ( ) ( ) (putting signage, etc) days days days O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Attending agronomy trainings ( ) ( ) ( ) days days days O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A
Attending meetings ( ) ( ) ( ) days days days O (-88) Don’t know O (-88) Don’t O (-88) Don’t O (-99) N/A know know O (-99) N/A O (-99) N/A Other certification
---. Crop year Crop year Crop year 12/13 13/14 14/15
Are you certified by other O (0) No O (0) No O (0) No certification, please specify: O (1) Yes, specify: O (1) Yes, specify: O (1) Yes, (if No, move to q114) specify: ______
When did you get certified? _____/_____ (month/ year)
Could you indicate if the ( ) ( ) ( certification has led to additional VND VND ) VND costs in cash, and how much per O (-88) Don’t know O (-88) Don’t know O (-88) Don’t season? O (-99) N/A O (-99) N/A know O (-99) N/A
How much labour time do you ( ) ( ) ( spend to ensure compliance with days days ) days the certification code that you O (-88) Don’t know O (-88) Don’t know O (-88) Don’t would otherwise not have spent? O (-99) N/A O (-99) N/A know O (-99) N/A
And what is this time spent on?
Administration ( ) ( ) ( days days ) days O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Additional work time in field ( ) ( ) ( (putting signage, etc) days days ) days O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Attending agronomy trainings ( ) ( ) ( days days ) days O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Attending meetings ( ) ( ) ( days days ) days O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A Turn over
---. How much did you earn from Crop year Crop year Crop year selling your coffee? 12/13 13/14 14/15
Production volume (dried coffee bean) (Kilograms)
Volume sold in that crop year (kilograms)
Average price per unit (VND/kg)
Total revenue (VND)
---. Did you earn additional money from O (0) No O (0) No O (0) No premiums from certification? O (1) Yes O (1) Yes O (1) Yes
Volume (dried coffee) on which ( ) kg ( ) kg ( ) premium was registered as certified O (-88) Don’t know O (-88) Don’t know kg coffee? O (-99) N/A O (-99) N/A O (-88) Don’t know O (-99) N/A
Volume (dried coffee) on which ( ) kg ( ) kg ( ) premium was received O (-88) Don’t know O (-88) Don’t know kg O (-99) N/A O (-99) N/A O (-88) Don’t know O (-99) N/A
Premium per unit ( ) ( ) ( VND/kg VND/kg ) VND/kg O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Total premium received ( ) ( ) ( VND VND ) VND O (-88) Don’t know O (-88) Don’t know O (-88) Don’t O (-99) N/A O (-99) N/A know O (-99) N/A
Other crop production for crop year 2014/15 (diversification models)
--. Pepper Cocoa Fruit trees
What are physical outputs of the corresponding crops? ______kg ______kg kg
What is revenue from the corresponding crops? ______vnd ______vnd vnd
How much Urea did you apply for corresponding crops ______kg ______kg during the last season? kg
What was the average price per ______kilogram of Urea? vnd/kg _ vnd/kg vnd/kg
Sulphate Ammonium (SA) ______kg ______kg kg
What was the average price per ______kilogram of SA? vnd/kg _ vnd/kg vnd/kg
Phosphorus 16.5%38 ______kg ______kg kg
What was the average price per ______kilogram of Phosphorus? vnd/kg _ vnd/kg vnd/kg
Potassium ______kg ______kg kg
What was the average price per ______kilogram of Potassium? vnd/kg _ vnd/kg vnd/kg
NPK ______kg ______kg _kg (____/_____/____) (____/_____/____) (indicate type) (____/_____/___ (indicate type) _) (indicate type)
What was the average price per ______kilogram of NPK? vnd/kg _ vnd/kg vnd/kg
Other, specify: ( ) ______kg ______kg kg
38 Also called Lan- Dam –Lan in Vietnamese
What was the average price per ______kilogram of the above fertiliser? vnd/kg _ vnd/kg vnd/kg
Other, specify: ( ) ______kg ______kg kg
What was the average price per ______kilogram of the above fertiliser? vnd/kg _ vnd/kg vnd/kg
How much did you apply? ______tons ______t ______tons (Excluding the compost made ons from coffee pulp)
How much did you spend for ( ) ( ( ) these organic fertilisers? VND ) VND VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t know O (-99) N/A know O (-99) N/A O (-99) N/A
Labour (man-days) ( ) ( ( ) O (-88) Don’t know ) O (-88) Don’t know O (-99) N/A O (-88) Don’t O (-99) N/A know O (-99) N/A
Other costs ( ) ( ( ) VND ) VND VND O (-88) Don’t know O (-88) Don’t O (-88) Don’t know O (-99) N/A know O (-99) N/A O (-99) N/A
--. Economic of Total production cost Total Revenue Labour other activities (Million VND) (Million VND) (man-day)
Other crops
Keeping livestock
Income from services ______vnd ______vnd ______vnd
Other non-farm income ______vnd ______vnd ______vnd Institutional environment
---. Di you attend agronomy Crop year Crop year Crop year trainings provided by 12/13 13/14 14/15 following organisations?
Extension services O (1) Yes O (1) Yes O (1) Yes O (0) No O (0) No O (0) No O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know
Other organisations such as O (1) Yes O (1) Yes O (1) Yes Farmer Unions, Women O (0) No O (0) No O (0) No Unions, Companies, … O (-88) Don’t know O (-88) Don’t know O (-88) Don’t know
Multidimensional deprivation Housing conditions
---. Which of the characteristics below do Crop year Crop year Crop year you have in your 12/13 13/14 14/15 house/living area?
Area of owned houses ______m2 ______m2 ______m2
Type of roof material of (0) Reinforcement (0) Reinforcement (0) Reinforcement the house (mostly living concrete concrete concrete in) (1) Tiles (cement. (1) Tiles (cement. (1) Tiles (cement. terracotta) terracotta) terracotta) (3) Roof slabs (3) Roof slabs (cement, (3) Roof slabs (cement, metal) metal) (cement, metal) (4) Leave/ straw/ (4) Leave/ straw/ rolled (4) Leave/ straw/ rolled roofing roofing rolled roofing (5) (-99) (5) (-99) (5) (-99)
Type of toilet (0) Flush toilet (0) Flush toilet (0) Flush toilet (1) Pour flush toilet (1) Pour flush toilet (1) Pour flush toilet (2) Double vault (2) Double vault (2) Double vault compost latrine compost latrine compost latrine (3) Toilet directly over (3) Toilet directly over (3) Toilet directly over the water the water the water
Type of waste disposal (0) Collected by (0) Collected by (0) Collected by someone someone someone (1) Dumping in ponds (1) Dumping in ponds (1) Dumping in ponds and lakes and lakes and lakes (2) Dumping in (2) Dumping in nearby (2) Dumping in nearby site site nearby site (3) Landfill (3) Landfill (3) Landfill burial/burning burial/burning burial/burning
Electricity O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
(Mobile) phone O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Computer O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Internet O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Air conditioner O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Motorbike O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Car O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Gas/electricity stove O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Health and environment
---. Crop year Crop year Crop year 12/13 13/14 14/15
Self-assessed food (rice, basic (1) Insufficient (1) Insufficient (1) Insufficient food grains and staples) adequacy (2) Sufficient (2) Sufficient (2) Sufficient (3) Sufficient (3) Sufficient (3) Sufficient
Self-assessed foodstuff (meat, (1) Insufficient (1) Insufficient (1) Insufficient vegetables and condiments) (2) Sufficient (2) Sufficient (2) Sufficient adequacy (3) Sufficient (3) Sufficient (3) Sufficient
Separate area for livestock O (0) No O (0) No O (0) No O (1) Yes O (1) Yes O (1) Yes
Source of drinking water (0) Tap in house (0) Tap in house (0) Tap in house (1) Public tap (1) Public tap (1) Public tap (2) Well (2) Well (2) Well (3) Stream water (3) Stream water (3) Stream water (4) Bought water (4) Bought water (4) Bought water (5) Rain water (5) Rain water (5) Rain water
When applying pesticides or (0) No (0) No (0) No herbicides do you get protected by (1) Yes (1) Yes (1) Yes gloves, masks, special clothes and (2) -99 (2) -99 (2) -99 boots? Access to education
---. Crop year Crop year Crop year 12/13 13/14 14/15
How many children who are of ( ) ( ) ( ) school-going age* do you have? O (-99) N/A O (-99) N/A O (-99) N/A
Is there any child who had to stop O (0) No O (0) No O (0) No schooling? O (1)Yes O (1)Yes O (1)Yes O (-99) N/A O (-99) N/A O (-99) N/A
And how many of them went/ have ( ) ( ) ( ) been to school? O (-99) N/A O (-99) N/A O (-99) N/A
Does your child(ren) have to O (0) No O (0) No O (0) No undertake farming activities O (1)Yes O (1)Yes O (1)Yes O (-99) N/A O (-99) N/A O (-99) N/A
If yes, how many hours per year O ______hours O______hours O ______hours per child? O (-99) O (-99) O (-99)
---. If yes, why?
Kid(s) have to work to support the O (0) No O (0) No O (0) No family O (1)Yes O (1)Yes O (1)Yes O (-99) N/A O (-99) N/A O (-99) N/A
Kid(s) prefer to work (voluntarily) O (0) No O (0) No O (0) No O (1)Yes O (1)Yes O (1)Yes O (-99) N/A O (-99) N/A O (-99) N/A
Too little money to pay school O (0) No O (0) No O (0) No fees, uniforms, etc O (1)Yes O (1)Yes O (1)Yes O (-99) N/A O (-99) N/A O (-99) N/A
Other, please specify: O (0) No O (0) No O (0) No O (1)Yes O (1)Yes O (1)Yes ( ______) O (-99) N/A O (-99) N/A O (-99) N/A
* School-going age is until 16