Assessing eco-efficiency and the determinants of horticultural family-farming in southeast Spain

Journal of Environmental Management 204 (2017) 594e604

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Assessing eco-efficiency and the determinants of horticultural familyfarming in southeast Spain

n, Emilio Galdeano- Go
mez*, Juan C. Pe
rez-Mesa, Angeles Godoy-Dura ~ oz Laura Piedra-Mun Department of Economic and Business, University of Almería (Agrifood Campus of International Excellence), 04120 Almería, Spain

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 May 2017 Received in revised form 26 August 2017 Accepted 12 September 2017

Eco-efficiency is currently receiving ever increasing interest as an indicator of sustainability, as it links environmental and economic performances in productive activities. In agriculture these indicators and their determinants prove relevant due to the close ties in this activity between the use of often limited natural resources and the provision of basic goods for society. The present paper analyzes eco-efficiency at micro-level, focusing on small-scale family farms as the principal decision-making units (DMUs) of horticulture in southeast Spain, which represents over 30% of fresh vegetables produced in the country. To this end, Data Envelopment Analysis (DEA) framework is applied, computing several combinations of environmental pressures (water usage, phytosanitary contamination, waste management, etc.) and economic value added. In a second stage we analyze the influence of family farms' socio-economic and environmental features on eco-efficiency indicators, as endogenous variables, by using truncated regression and bootstrapping techniques. The results show major inefficiency in aspects such as waste management, among others, while there is relatively minor inefficiency in water usage and nitrogen balance. On the other hand, features such as product specialization, adoption of quality certifications, and belonging to a cooperative all have a positive influence on eco-efficiency. These results are deemed to be of interest to agri-food systems structured on small-scale producers, and they may prove useful to policymakers as regards managing public environmental programs in agriculture. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Eco-efficiency Family farm Horticulture DEA Socio-economic features

1. Introduction Humans have intervened in natural ecosystems to gather food in the most productive and efficient way possible for millennia. In recent decades, this activity has become extremely important due to an incessant population increase and the growing need to obtain basic goods amidst alarming scarcity of resources and greater pressure on the environment. In this context, eco-efficiency, meaning the relationship between economic and ecological efficiency, is a useful indicator in relation to the capacity of companies, sectors and economies to produce goods and services with less consumption of natural resources and less impact on the environment. This concept emerged in the 1990s as an indicator of sustainability (Schaltegger and Sturm, 1996; Bleischwitz, 2003). The OECD (Organization for

* Corresponding author.
mez). E-mail address: [email protected] (E. Galdeano- Go http://dx.doi.org/10.1016/j.jenvman.2017.09.037 0301-4797/© 2017 Elsevier Ltd. All rights reserved.

Economic Co-operation and Development) defined it as the efficiency with which ecological resources are used to meet human needs (OECD, 1998, 1999). While eco-efficiency has become a rather popular concept, it does have several limitations with regard to sustainability. On the one hand, it focuses on only two aspects of production processes and fails to consider social implications (Beltr
an Esteve, 2012). On the other hand, the term efficient by no means implies that something is sustainable; the former is merely a necessary condition (or intermediate step) for the latter. This is essentially due to the fact that eco-efficiency is simply a relative indicator, and it does not provide information regarding effectiveness (Callens and Tyteca, 1999). For example, although the relative level of a particular environmental pressure may be low, which would be indicative of high efficiency, it might exceed the absorption capacity of the ecosystem itself. Moreover, the appearance of the so-called ‘rebound effect’, e.g. changes in consumption patterns (Holm and Englund, 2009), can compensate for or neutralize the efficiency achieved at any given moment. In short, achieving environmental efficiency at the micro level does not guarantee the


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

achievement of environmental sustainability objectives in absolute terms at the macro level (Huppes, and Ishikawa, 2005). Despite these shortcomings, eco-efficiency analyses have shown to be of particular interest for two reasons. Firstly, progress in environmental performance may represent the most cost-effective way to achieve a reduction in environmental degradation. Secondly, policymakers may find it easier to adopt policies aimed at achieving improvements in performance than other more radical policies that directly restrict the level of economic activity
mez-Limo
n et al., 2012). (Kousmanen and Kortelainen, 2005; Go In the field of microeconomics, this indicator has grown in popularity, particularly with the World Business Council for Sustainable Development (WBCSD, 2000); it serves as a means of stimulating productive organization to seek out environmental improvements that yield parallel economic benefits (Keating et al., 2010). In agricultural production, measures of eco-efficiency not only allow pollutants to be tested but they can also incorporate the socalled balance principle (Coelli et al., 2007), which implies that nutrients not contained in good outputs return to the environment as potential pollutants (e.g. fertilizers, wastes, etc.). Hoang and Nguyen (2013) expand upon this analysis of agri-food activities by including nonmaterial inputs (e.g. labor, capital, farm services, etc.), offering a more general study (material balance and energy balance) of environmental problems (Oude Lansink and Wall, 2014). The most common measuring techniques in this context are production frontier models, DEA (Data Envelopment Analysis) and SFA (Stochastic Frontier Analysis), essentially because their operationalization relates environmental and economic outcomes rather than just conventional inputs and outputs (Callens and Tyteca, 1999; Lauwers, 2009). In keeping with this approach, Kousmanen and Kortelainen (2005) propose a definition of ecoefficiency based on what they refer to as ‘a pressure-generating technology set’, which represents all feasible combinations of economic value and environmental pressures (treated as conventional inputs). This particular definition has been applied in ecoefficiency analyses conducted in a variety of sectors, as well as in various agri-food production studies (see for a review e.g. Oude Lansink and Wall, 2014). Following this methodology, numerous analyses in the agri-food sector have focused on eco-efficiency at farm level, using individual farms as the basic production units (e.g. De Koeijer et al., 2002;
nPicazo-Tadeo et al., 2011; Hoang and Nguyen, 2013; Beltra Esteve et al., 2014). This has been done so fundamentally because it is believed that individual farmers are responsible for making the most important decisions concerning the use of resources and the implementation of greener technologies (Webster, 1999). In addition, and perhaps more importantly, when conducting specific analyses on farming activity, it must be taken into account that there are factors determined by issues of agricultural policy and others influenced by socioeconomic features, e.g. farmers' attitudes, organization structures, environmental concerns, etc. (Keating et al., 2010). In effect, although business people, policy makers and society in general are always interested in indicators, the study of determinants of eco-efficiency is no less important, yet it has been addressed to a lesser extent in the literature (Oude Lansink and Wall, 2014). This work presents an eco-efficiency analysis of the horticultural farming system in southeast Spain, which is characterized by a structure consisting of small-scale family farms. This farming sector represents about 30% of Spanish fresh vegetables produced and about 18% consumed in Europe (Cajamar, 2016). Over the course of decades, the growers in this region have progressively incorporated technologies adapted to the environmental conditions of the area,

595

such as limited basic resources (mainly water and land), as well as various techniques to reduce negative externalities resulting from the increasing intensification of farming methods (e.g. water scarcity, chemical contamination, waste, etc.), although not always with ~ oz et al., 2016). This agricultural homogeneous results (Piedra-Mun development has been greatly influenced by a series of socioeconomic components of farming itself (e.g. family, cooperative organization, education, etc.) as well as interrelationships with the agri-food cluster surrounding said activity, such as auxiliary and
nchez et al., 2011). service companies (Aznar-Sa The aims of this paper are: i) to obtain a series of indicators for eco-efficiency by computing several ratios for environmental pressures and economic value added in this farming activity; ii) and to analyze their socioeconomic and environmental influences on these indicators. The methodology followed is the Data Envelopment Analysis, DEA, framework on a sample of horticultural family farms as decision-making units (DMUs). In a second stage, by using a truncated regression and bootstrapping techniques, the determinants of eco-efficiency scores in this farming system are analyzed. Unlike other research papers along this line, the present study identifies both aggregate and specific individual environmental pressures as well as the determinants that influence them. Such information is important for intensive horticultural sectors, especially those in semiarid regions with specific environmental characteristics. In the present case, examples of these features include the concentration of crops in the study area and the traditional pressures exerted on its natural surroundings. Moreover, as we are dealing with family farms, this study is of particular interest at the micro-level given the determinants being analyzed, i.e. social, economic and environmental. The former are the most relevant in this regard due to the sector's family orientation (e.g. the role of farm succession to the next generation) and organizations (cooperatives). These aspects, related to the type of farm managers and their behavior, are of considerable importance as well. In addition, while one of the main objectives of other eco-efficiency studies is usually to focus on the assessment of environmental programs or the support to farming, in the present case, such initiatives are traditionally of little relevance. As a result, in this context, the influence of the specific characteristics of the sector, e.g. type of manager, become even more important. The results of the present study can provide a wide overall perspective on ecological-economic relationships and orientations. This can be of great help for the drafting of agri-environmental policies in the horticultural sector under study, as well as a contribution to methodological applications in other similar farming systems. The rest of the paper is structured as follows. In the second section a description of the sector under study is outlined. In the third section, methods and data are expounded. In the fourth section the estimations and results are presented. Finally, the conclusions and policy implications are detailed in the fifth section.

2. Description of horticulture in southeast Spain The current agricultural system had its beginnings some five decades ago, initially cultivating fresh vegetables out in the open and later in greenhouses. With some 30,000 ha of crops at present, located in the coastal area of the provinces of Almería and Granada, this system accounts for approximately one third of all vegetables grown in Spain and this produce is destined for both the domestic and foreign markets (Cajamar, 2016). The evolution of this horticultural system has been particularly characterized by a process of adaptation to environmental conditions and challenges of resource efficiency (Figs. 1 and 2). This agricultural area is semi-arid with scarce precipitation, but

596


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

Fig. 1. Location of horticultural area.

Fig. 2. Evolution of the farming system in surface and in production (Cajamar, 2016).

it has subterranean water resources and a mild climate due to the proximity of the sea and the nearby mountain ranges that afford it protection. These climatological characteristics and several technologies adapted to local requirements have allowed a ten-month long growing season (from September to June) and give rise to a highly productive sector. Farming activity has become specialized in certain crops: pepper, tomato, cucumber, green bean, zucchini,

eggplant, water melon and melon. Farmers traditionally tend to rotate the production of these different crops (e.g. two to three crops a year) although increased crop specialization has been observed in recent years (Valera et al., 2016). Among the local technologies the “enarenado” technique (protecting the natural soil with a layer of sand and natural manure) is of particular interest, as it allows reduced water consumption and the condensation of


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

atmospheric humidity. On the other hand, the sector has also developed through the use of greenhouses with plastic covering held in place between wire mesh and supported on metal posts (“parral-type” greenhouses). This simple system, combined with the area's natural sunlight (over 3000 h per year), is highly efficient as it requires almost no extra energy, improves water savings and reduces the effects of erosion caused by heavy occasional rainfall and strong winds which are characteristic of semi-arid regions such
mez et al., 2013). as this (Galdeano-Go

2.1. Persistence of environmental pressures Despite the adoption of local technologies and adaptations
n made to crops, various environmental concerns still persist (Tolo Becerra and Lastra Bravo, 2010): - Water scarcity. Traditionally, the increasing surface area dedicated to crop production has caused concern over the limited hydrological resources. Since the late 1980's this sector has progressively incorporated generalized systems of water economy (drip irrigation, hydroponic systems and reusage techniques) and is currently considered a highly technified irrigation
ndez et al., 2007). Nevertheless, the manarea in Spain (Ferna agement efficiency results for this resource could be much more heterogeneous, depending on the technologies utilized and the ~ oz et al., 2016). Consequently, the overfarm type (Piedra-Mun exploitation of aquifers is an ongoing concern, despite attempts in recent years to compensate by increasing the water supply using desalinization plants and residual water treatment methods. On the other hand, from a water balance approach, i.e. the amount of water required to vegetables produced in this case, there also exist differences among farms due to the varied levels of yield, which depend on the technologies used and the
n Becerra types of vegetables in which farms specialize (Tolo et al., 2013). - Waste. Crop concentration has also generated substantial amounts of material waste (plastics, containers, etc.) and plant waste. In recent decades, various different Rural Hygiene Plans have been implemented (by Municipalities in conjunction with farmer organizations) to improve the collection and treatment of many types of waste materials, which has helped to reduce
n-Ferre et al., 2011). Thus, a wide variety of this problem (Callejo waste is almost completely recyclable, mainly plastic, cardboard, and the metal from greenhouse structures. In contrast, with respect to plant waste (about 30 tons per hectare), it is necessary to intensify hygiene plans and treatment processes as a pro
n Becerra portion of this waste is still not properly recycled (Tolo and Lastra Bravo, 2010). - Chemical contamination. The use of chemical products has traditionally caused soil and water contamination, mainly as a consequence of nitrates and the use of phytosanitary products (Martínez-Vidal et al., 2004; Pulido-Bosh, 2005). Although over the last decades there has been a trend toward adopting biological technologies such as IPM, integrated pest management, and organic production (Van der Blom, 2010), in addition to
rez-Mesa and quality certifications imposed by retailers (Pe
mez, 2015), the implementation of these pracGaldeano-Go tices is still not homogeneous among farmers. For example, although IPM is the most widespread technology (over 80% of total production), its use depends on crop type; and, as for organic farming, it only accounts for a small percentage of production (about 10%), and in many cases these types of farms are in their initial phase (e.g. some allocate 1 ha to organic and another to more traditional practices). Additionally, there are

597

various types of farms, which range from the most traditional (‘enarenado’ technique) to soilless crops (hydroponics). Among these different systems contamination, particularly from fertilizers at present, tends to be rather varied as well (Valera et al., 2016).

2.2. Socieconomic features This horticultural sector represents the largest economic activity in the region; farming activity alone represents 24% of GDP (Gross
mez Domestic Product) and 27% of employment (Galdeano-Go et al., 2013). Furthermore, the services and the associated auxiliary industry account for approximately 32% of the GDP in this area
nchez et al., 2011). Indeed, this sector's activity has (Aznar-Sa generated, in parallel, local development of a number of commercial structures, currently represented by cooperative organizations, as well as an agro-industrial cluster dedicated to the supply of inputs (fertilizers, bees, seeds, etc.) and various services (financing, consulting, R&D, training, etc.). The local nature of this progress is essential from a socioeconomic point of view (endogenous nature), bearing in mind that public programs and aid traditionally have been rather insignificant, only accounting for approximately 1.5% of
mez the income generated by the farming activity (Galdeano-Go et al., 2013). The social component is also extremely relevant as the sector was founded on a system of family farms and has been a key source of employment (more than half of the total is family members), which has favoured its economic feasibility. The farmland is widely divided among 13,500 small-scale farmers, approximately 2 ha on average (Valera et al., 2016). Authors such as Tout (1990) and García-Latorre et al. (2001) describe the role of family farms in this productive activity, explaining how they have created their own organizations and built up a close relationship with other stakeholders in the sector, to whom the farmers transmit not only their economic concerns but also those regarding the management of
mez et al., 2016). For instance, Downward resources (Galdeano-Go and Taylor (2007) highlight the social nature of this agrarian system and its effects on efficient water usage. Furthermore, organization through cooperatives has proved essential for the development of commercialization, access to services and the channelling of subsidies from the Common Agricultural Policy (CAP); in addition, these entities have also played an important role in the training of growers and the promotion of innovations and ~ oz et al., 2017). environmentally respectful practises (Piedra-Mun 3. Methods 3.1. Analysis of eco-efficiency following DEA methodology Estimation of eco-efficiency can be done using ratios, i.e. outputs divided by inputs, which relate the goods produced to the environmental pressures and impacts caused by the activity.1 In recent years, varying combinations of environmental pressures have been regarded as inputs, mainly to determine the possible effects substitutions may have on business results, while value added has usually been considered as the output. Therefore, eco-efficiency

1 Among other traditional indicators of environmental impact of products is life cycle accounting (LCA), which deals with identifying and evaluating each one of the impacts associated with the different phases of production and commercialization. This approximation can be combined with life cost cycle analysis (LCC) to include economic aspects. Nevertheless, these techniques offer various indicators which are
n Esteve, 2012). difficult to add (Beltra


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

598

would improve if environmental pressures were reduced and valued added remained constant, or, if value added increased and environmental pressures remained stable, or, if value added increased as pressures were decreased (Kortelainen and Kousmanen, 2007). Measurement of this indicator can be conducted using various methodologies (Tyteca, 1996), yet all generally feature the existence of an eco-efficient frontier which represents the best possible practices for a given technology of reference. Essentially, these approaches are usually grouped into parametrics, characterized by the assumption of a specific functional form for the technology, and non parametrics, normally using DEA (e.g. Cooper et al., 2007; Cook and Seiford, 2009). The latter makes it possible to characterize the productive units, which are considered DMUs, as efficient or inefficient by means of measuring the distance of a technological frontier based on the best practices observed. One of the advantages of DEA is the flexibility of its functional form and its use in terms of calculating weights, specifically because these are not exogenous and/or subjective, e.g. the weights of the various environmental pressures in the construction of aggregate environmental pressure; instead they are generated endogeneously and are different for each produc
n Esteve, tion unit (Kousmanen and Kortelainen, 2005; Beltra 2012). Furthermore, these weights adapt to the units of the measure chosen, allowing a more suitable normalization of the measures themselves. These characteristics constitute an advantage for overcoming the uncertainty of measuring environmental pressures from the point of view of the benefit-of-thedoubt principle (Cherchye et al., 2007). In this line, outstanding works include Kousmanen and Kortelainen (2005), Kortelainen and Kousmanen (2007), Zhang et al. (2008), and other studies applied to the agri-food sector,
mez-Limo
n et al. (2012) and such as those by Serrao (2008), Go Beltr
an-Esteve et al. (2014), all of which utilized DEA in the context of the neoclassical theory of production. Following the methodology of the aforementioned works, the present study adopts an economic-environmental ratio approximation at micro level (Huppes and Ishikawa, 2005; Picazo-Tadeo et al., 2011), which is detailed below. Assuming that the economic indicator is value added, v, which generates a series of N environmental pressures, represented by vector p ¼ (p1, …, pn), for a k ¼ 1, …, K set of horticultural farms, we can define the set of pressure generating technologies, PGT (which represents all possible combinations of value added and environmental pressures) as follows: PGT ¼ [(v, p) 2 R1þN j added value v can be generated with pressures p] (1) Having defined technology, the calculation of eco-efficiency for a horticultural farm k is formulated as: Eco-Efficienyk ¼ vk / P(pk)

(2)

where P is the function that aggregate the n environmental pressures in one single value of pressure or environmental damage. For the calculation of P, the most common approximation in the literature is followed, which consists of using a weighted linear average of individual environmental pressures as the aggregation function. In this way, the following calculation is obtained:

Pðpk Þ ¼

N X

wn pnk

n¼1

where wn is the weight assigned to pressure n.

(3)

Following the DEA methodology, the score of eco-efficiency for a horticultural farm k can be computed using a mathematical program as follows:

Maximize wnk; Eco  efficiencyk ¼ vk=

N X

wnk pnk

(4)

n¼1

Subject to:

, vk

N X

wnk pnk  1;

n ¼ 1; …; N

n¼1

wnk  0; k ¼ 1; :::; K where wnk is the weight with which each pressure n enters into the computation of the composite environmental pressure score of farm k. On the other hand, the programming proposed in (4) has an equivalent dual formulation (Picazo-Tadeo et al., 2011), and the computational program is written as follows: Minimize qn, zk Eco-efficiencyk ¼ qk

(5)

Subject to:

vk 

K X

zk vk

k¼1

qk pnk 

K X

zk pnk ; n ¼ 1; :::; N

k¼1

zk  0; k ¼ 1; :::; K where zk is a variable that represents the intensity weighting with which each farm k observed enters in the composition of the ecoefficient frontier. In this formulation, it is also assumed that technology represents constant returns to scale as it is the most widelyused supposition in these cases.2 The solution to program (5), namely the parameter qk, describes the maximum radial or equiproportional contraction of the environmental pressures vector at farm level, for a given value added. This score of eco-efficiency is upper-bounded to 1, the score which represents best performance, and the lower the score computed the lower eco-efficiency. In other words, this computation determines the inefficiency margin that could be potentially cut (1qk) for a farm k, in terms of proportional reduction of the environmental pressures, while maintaining its valued added. Another point of interest in the present analysis is the estimation of the potential reduction of a specific i pressure (or a specific group) without increasing the remaining environmental pressures ei and for a given value added. To this end, the following programming is proposed (Beltr
an Esteve, 2012):

2 While it is important to consider variable returns to scale from an economic perspective, from ecological perspective production is characterized by constant returns to scale (Picazo-Tadeo et al., 2011). As the topic of interest in this case relates to the total pressures exerted on the environment and not their distribution among different farms, we opt for the constant scale economies (Kousmanen and Kortelainen, 2005). This assumption was also considered in other analyses of efficiency and productivity in the horticultural sector in southeast Spain (Galdeano
mez et al., 2006; Rodríguez-Rodríguez et al., 2012). Additionally, our analysis Go will calculate the output and most of the inputs as average data per unit of surface area.


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

Minimize qi, zk Eco-efficiencyk ¼ qik

(6)

Subject to:

vk 

K X

zk vk

k¼1

qik pik 

K X

zk pik ; i 2 n and i ;  i

k¼1

pik 

K X

zk pik ;  i2n

k¼1

zk  0; k ¼ 1; :::; K

3.2. Analysis of determinants of eco-efficiency It is considered that eco-efficiency or eco-inefficiency is affected by the characteristics of companies and other variables that may affect the production process. These influences can be analyzed in a second stage following a multivariate regression model for each computed score of eco-efficiency, qk, as follows:

q k ¼ b x k þ εk

(7)

where εk ~ N (0, s2), and b is the vector of parameters for the vector of independent variables xk. These independent variables are mainly characteristics of the farms and others related to the specific sector context of horticultural production in the region under study. They are described in further detail in the following subsections. In order to perform the estimation of (7), a traditional method has been censored OLS, ordinary least squares, or Tobit regression, but these might lead to inaccurate results due to the existence of correlation between the measures of eco-efficiency and the error terms (εk). On the other hand, Simar and Wilson (2007) proposed a procedure based on truncated regression3 to avoid said correlation, thereby allowing better statistical inference. This procedure involves the following steps (for more details, see ‘first algorithm’ of Simar and Wilson, 2007: 41e42): [1] The computation of all b q k solving (5) and (6). [2] The use the method of maximum likelihood to obtain an estimate b b of b, as well as an estimate sb ε of sε in the truncated regression of b q k on xk in (7) using the subset of ecoinefficient observations. [3] The computation of L bootstrap estimates for b and sε in the following way: b 2ε and left Draw εk from a normal distribution with variance s * b b truncation at 1e b xk and compute q k ¼ b xk þ εk. Then estimate the truncated regression of q*k on xk by maximum likelihood, b *, b yielding a bootstrap estimate ( b s *ε). [4] Finally, with a large number of bootstrap estimates (e.g. L ¼ 1000), it becomes to test hypotheses and to construct confidence intervals for b and sε.

3 In a truncated normal distribution, εk is not observed when it would fall below 1-b0 xk. In a censored model, εk is always observed, even if there is some information loss.

599

4. Sample and data The information has been gathered by conducting surveys on a set of individual farms. In this case an ad hoc questionnaire was created for the present study, included as supplementary material (available upon request). This survey features a variety of questions with the objective of obtaining data about: a) environmental issues (use of natural resources, waste management, environmental certifications, proactivity to agro-ecological innovations, etc.); b) social features (education, age, experience, degree of cooperative integration, etc.); c) economic details (sales value, production cost, yield, specialization, subsidies, etc.) including information related to connections with the service sector (mainly R&D centers) and other auxiliary activities linked to farming. The surveys were conducted during the 2014e2015 growing season, from September to June. The farms were selected by simple random sample, without replacement, carrying out a total of 360 surveys; however, several measurement errors were detected, which brought down the final sample to 327 farms.4 In general, relative homogeneity was detected among the farms surveyed, specifically in terms of family ownership (98.7%) and land area, which was between 1 and 5 ha, with an average of 2.19 ha. Additionally, the statistical data from our surveys have been contrasted with the data of Department of Agriculture of the Andalusian Regional Government to corroborate the statistical suitability.

4.1. Description of output and environmental pressures As explained in the methodology, the study of eco-efficiency requires the elaboration of ratios which relate economic output and a series of environmental pressure indicators. As economic component, a gross value added, GVA, of a horticultural farm, k, is considered and calculated using the following formula: GVAk ¼ (Sales valuekeDirect costsk)/Land areak

(8)

Sales and direct costs5 (fertilizers, contracted services and fixed costs edepreciation and maintenance of facilities-) are measured in euros, while the surface or land area is measured in hectares, which provides an estimation of the average economic output per hectare. As for environmental pressures, these were estimated using five indicators relevant to environmental management in horticultural production in this region, as described above, according to the
n Becerra and Lastra Bravo (2010) and Tolo
n Becerra works by Tolo et al. (2013): - The use of water resources was measured using two indicators, given the relevance of this environmental factor: a) On one hand, water usage by area is considered, which is designated as WU and estimated as WUk ¼ Water usagek/Land areak

(9)

measuring water usage in m3 and obtaining average usage per hectare.

4 This final sample, from a total of 13,500 farms, for 95% confidence level and 0.5standard deviation, implies a margin of error of ± 5.3%. 5 The subsidies are not included in the GVA because these economic aids are disconnected from the production. Also, labor costs are not included in direct cost since these costs represent more than 50% of the cost production of horticultural farm, 60% of which is family labor (Hortyfruta, 2010), making it an important source of farm income.

600


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

WBk ¼ Water usagek /Produce tonsk

(10)

- The use of fertilizers, measured by the nitrogen balance (as it is the main component used in fertilizers), which is designated as NB and estimated as NBk ¼ Nitrogen in inputk /Nitrogen in outputk

(11)

where both ratio components are measured in kilograms per hectare, and nitrogen in output is the amount measured according
mez-Limo
n and Arriaza Balmo
n, to the vegetable yield obtained (Go 2011). - The effect of phytosanitary products is considered an approximation of average risk. Previous analyses carried out for this sector indicate that with an average of 0.2 tons of phytosanitary products per hectare, there is a certain risk of presence in plant material and contamination ground water (Martínez-Vidal et al., 2004; Valera et al., 2016). Therefore, this pressure indicator, designated as PHT, is calculated in terms of variation with regard to the risk averages in the following manner: PHTk ¼ Kilograms per hectarek /200 kg

(12)

b) On the other hand, the volume of water required per ton of produce obtained on the farm is designated as WB (as an
n Becerra approximation to the so-called ‘water balance’ e Tolo and Lastra Bravo, 2010), estimated as In this way, values greater than the unit will indicate an increase in risk of contamination, while values less than the unit will indicate a decrease. - Waste generated on farms, considering mainly vegetable waste
n Becerra and Lastra Bravo, 2010), and the corresponding (Tolo proportion that is not recycled or is not transported to municipal landfills, herein designated as AMW, AMWk ¼ Wastek/Land areak

(13)

where the volume of plant waste is measured in tons, providing the amount per hectare of crop.6 4.2. Description of influencing variables on eco-efficiency Farms' characteristics considered as possible determinants of eco-efficiency which were obtained from the surveys are grouped into social, economic and environmental factors. Several social factors are considered according to the specific
mez characteristics of the family farming under study (Galdeano-Go ~ oz et al., 2017). For instance, in relation to et al., 2016; Piedra-Mun farmer features, there are working parents, children and in some cases grandparents (about three generations in the development of this horticulture) as family owners and workers, and it is desirable

6 The measurement output and environmental pressure as ratios might imply some cautions in DEA analysis (Hollingsworth and Smith, 2003) mainly when dealing with VRS analysis (and the consideration of a BCC model). Nevertheless, as indicated above, the authors opted to consider the assumption of CRS, due to characteristics of the sample itself and the objectives of the eco-efficiency analysis.

to gather key data from the decision-maker of the farm. Also, although the surveys distinguish between male and female, and women represent about 20% of the farmer's decision-makers, no differences have been detected in eco-efficiency estimates. In addition, inheritance factor is considered due to the possible implication of the next generation in environmental concerns, ac~ oz cording to other studies on this horticultural sector (Piedra-Mun et al., 2017) (see Table 1). Additionally, some economic variables related to the agri-food auxiliary sector surrounding this system are evaluated (AznarS
anchez et al., 2011, 2014; Valera et al., 2016). For instance, several industries, as well as cooperatives, have played an important role in providing advice and consultation services to vegetable
mez et al., 2016), as observed in the surveys farms (Galdeano-Go carried out. Also, some research centers around the sector (Universities of Almería and Granada, Cajamar Foundation eLas Palmerillas-, the Association of Producers and Exporters of Fruit and Vegetables eCoexphal-, etc.), have made important contributions
nchez et al., 2014). in innovations applied in the sector (Aznar-Sa Table 2 shows descriptions of these variables. The hypotheses to be tested propose that higher values of these variables have positive effects on eco-efficiency, i.e. reduction of environmental pressures, with the exception of age and specialization, which is considered to be the reduction of the number of crop rotations on the farm. Table 3 shows the descriptive statistics of these variables. 5. Estimation and results 5.1. Estimations of the eco-efficiency Table 4 displays the mean results of the computed scores for eco-efficiency, both those of the pressures set and those obtained for each one of the environmental pressures used for reference. The average values obtained indicate that farmers produce with an average eco-inefficiency margin of nearly 11%, i.e. the average score of eco-efficiency is 0.89; this means that the aggregate environmental pressures could potentially be reduced by about this proportion (11%) while maintaining the same level of value added. The minimum required reduction in environmental pressures to achieve eco-efficiency correspond to water usage and nitrogen balance, with inefficiency margins of 5% and 7.3% respectively; these data are indicative of better efficiency in the usage of water resources and fertilizers, a trend which has been observed in other ~ oz et al., 2016), and indicative of imstudies (e.g. Piedra-Mun provements to irrigation technologies in recent years, which have additionally made it easier to control dosages of specific fertilizers. The indicator of water balance shows an inefficiency margin of 8.7% that could be potentially reduced, which we believe is the consequence of the low yield on farms with more conventional green
n Becerra et al., 2013). The environmental house structures (Tolo pressure related to phytosanitary products reveals that the

Table 1 Estimate values of the economic output and environmental pressures. Indicator

Mean

Std. Dev.

Minimum

Maximum

GVAa WUb WBb NB PHT AMW

24.693 4.735 0.918 0.324 0.179 0.368

20.531 1.207 1.740 0.183 0.120 0.240

18.127 4.284 0.279 0.092 0.115 0.168

37.168 5.406 2.070 0.431 0.274 0.533

a b

In thousands of euros. In thousands of m3.


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

601

Table 2 Farms' characteristics variables. Dimension

Variable

Description

Social

Age (AGE) Experience (EXP) Education (ED)

Age of the decision-maker of the horticultural farm Number of years of experience in vegetable production Average education of family farm decision-maker. The education of each was measured on a scale of 1 (no education), 2 (primary education), 3 (middle school), 4 (high school or vocational training) or 5 (university or higher education) Percentage of family workers in total employment of the farm Dummy variable scoring 1 if the farmer thinks the next generation will inherit the family farm or 0 when he does not Variable weighted as 0 when no member of any cooperative organization, as 1 when a member, and 2 when a member of more than one (e.g. one marketing cooperative and one supply cooperative) Number of tons of vegetable produced per hectare Number of crops cultivated by the family farm. Thus, the lower this variable is, the higher the family farm's specialization Weighted mean of the farmer valuation of the efficiency of the auxiliary industries in the sector, scored from 1 to 5 Farmer proactive work with research centers and universities on new cultivation techniques and structural innovations in the farm to improve his economic performance, scored from 1 to 5 Total subsidies (in euros) over total income of the farm, obtaining the percentage which represents subsidies Integrated Pest Management and/or other certification of agroecological production (in kilograms) per hectare. A weighted mean of all the crops was calculated This variable measures the extent of the family farm awareness of efficient use of natural resources and openness to specific innovation to improve the relationship of its activity with the natural environment. This variable is scored from 1 to 5

Family labor (FL) Inherit (INH) Cooperativism (COOP) Economic

Yield (YLD) Specialization (SP) Auxiliary sector (AUX) R&D proactivity (RDP)

Environmental

Subsidies (SUB) Quality certification (QC) Environmental innovation (EIN)

utilization of chemical products can potentially be reduced by 9.2% while avoiding a reduction in economic output; thus, although there has been an increase in integrated pest control and organic
n et al., 2010), the results production systems in recent years (Beltra are heterogeneous depending on the type of crop and the implementation of these practices by the farmers themselves; in fact, these techniques are most likely not applied by farmers on the actual crops but rather, for example, on weeds and pest control on the land of farms with more traditional greenhouse structures (Valera et al., 2016). The most eco-inefficient situation is found to be that of plant waste, indicating that the environmental pressure caused by this waste could be potentially cut by 24.3% without cost in terms of reducing value added; and it is considered that, as already mentioned in the previous section, even though efforts

Table 3 Descriptive statistics of variables. Variable

Mean

Std. Dev.

Minimum

Maximum

Age (AGE) Experience (EXP) Education (EDU) Family labor (FL) Inherit (INH) Cooperativism (COOP) Yield (YLD) Specialization (SP) Auxiliary sector (AUX) R&D proactivity (RDP) Subsidies (SUB) Quality certification (QC)a Environmental Innovation (EIN)

42.537 28.185 3.519 1.072 0.879 1.230 23.108 1.741 3.961 3.602 0.034 17.618 3.746

22.018 22.740 1.026 0.638 0.306 0.865 11.617 0.790 0.932 1.026 0.028 11.207 2.219

21.000 7.000 1.000 0.609 0.000 0.000 15.860 2.333 2.000 1.000 0.000 12.180 2.000

66.000 48.000 5.000 1.480 1.000 2.000 29.215 4.000 5.000 5.000 0.065 25.300 5.000

a

Thousands of kilograms.

Table 4 Computed scores of eco-efficiency on horticultural farms. Eco-efficiency ratios

Mean

St. Dev.

Maximum

Minimum

Eco-efficiencyTOTAL Eco-efficiencyWU Eco-efficiencyWB Eco-efficiencyNB Eco-efficiencyPHT Eco-efficiencyAMW

0.891 0.950 0.913 0.927 0.908 0.757

0.969 0.908 0.964 0.985 1.102 0.886

1.000 1.000 1.000 1.000 1.000 1.000

0.620 0.718 0.645 0.730 0.623 0.485

have been made in recent years to improve recycling and waste
n-Ferre et al., 2011), a negative environmental management (Callejo impact still exists that must be substantially reduced. 5.2. Analysis of determinants of eco-efficiency Table 5 shows the estimations obtained by the procedure described in Section 3.2.7 The results obtained reveal coefficients with signs suitable to the hypotheses presented in the description of the variables considered to influence eco-efficiency, although some of them do not display a significant impact. Although the acceptable level of significance is normally 0.05 or higher, in this case the three typical levels of significance are shown in order to extract more conclusions from the estimated parameters in spite of the diversity of the variables considered and the various influences they can have on eco-efficiency. With regard to social factors, the young age of farmers (AGE), although it displays generally positive signs for parameters, does not imply a significant reduction of inefficiencies. In contrast, high level of experience (EXP) and level of education (EDU) of farmers do influence on production efficiency; these two variables prove relatively significant (p > 0.05 and p > 0.10, respectively) for reducing eco-inefficiency, both aggregated and individually, namely in terms of water usage and waste management. This coincides with results from other studies on ecoefficiency in Spanish farming which found that higher level of education and specialized training affect the improvement of effi
mez-Limo
n et al., 2012). However, the percentage of ciency (Go family labor (FL) does not prove to be a differential element in the reduction of environmental pressures, bearing in mind the generalized family-oriented nature of the farms sample. The next variable, inheritance (INH), positively impacts the reduction of inefficiency, particularly in terms of improving waste management. Cooperativism (COOP) has a positive influence and affects improvement of efficiency to an even greater extent (p > 0.05), taking into consideration the support these organizations provide in terms of training of new practices and new technology investments. The results of these last two variables are in accordance with other studies in the sector, which confirm that they affect the

7

For more details on the estimation in Stata package, see e.g. Tauchmann (2015).


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

602

Table 5 Results of truncated regressions on eco-efficiency influences. Variable

Eco-efficiencyTOTAL

Eco-efficiencyWU

AGE

0.009 (0.027, 0.008)

0.001 (0.014, 0.011) 0.004 (0.038, 0.029)

EXP EDU FL INH COOP YLD

0.033** (0.024, 0.041) 0.021* (0.016, 0.027) 0.009 (0.018, 0.037) 0.019** (0.012, 0.027) 0.008** (0.005, 0.011) 0.002 (0.009, 0.012)

0.018** (0.010, 0.027) 0.019* (0.012, 0.026) 0.011 (0.007, 0.029) 0.005* (0.002, 0.009) 0.023*** (0.014, 0.033) 0.011 (0.027, 0.006)

0.013* (0.008, 0.019) 0.007* (0.000, 0.015) 0.005 (0.026, 0.039) 0.000 (0.000, 0.000) 0.005** (0.003, 0.008) 0.007 (0.018, 0.030)

SP AUX RDP SUB

0.007** (0.010, 0.005) 0.016** (0.010, 0.022) 0.010* (0.007, 0.015) 0.002 (0.013, 0.019)

0.003* (0.004,0.002) 0.012** (0.008, 0.017) 0.031** (0.022, 0.041) 0.000 (0.008, 0.010)

0.012** (0.016,0.008) 0.002* (0.000, 0.003) 0.018* (0.015, 0.024) 0.004 (0.012, 0.021)

QC EIN Constant b ε) Sigma( s

0.017** (0.012, 0.023) 0.015** (0.010, 0.021) 3.108*** (2.054, 4.163) 0.631*** (0.456, 0.805)

0.005* (0.003, 0.008) 0.024** (0.016, 0.033) 1.964*** (1.215, 2.714) 0.723*** (0.537, 0.910)

N Log likelihood

327 294.073

327 264.902

Eco-efficiencyWB

Eco-efficiencyNB

Eco-efficiencyPHT

Eco-efficiencyAMW

0.000 (0.017, 0.020)

0.003 (0.017, 0.023)

0.007 (0.031, 0.046) 0.018* (0.009, 0.027) 0.004 (0.005, 0.013) 0.010* (0.005, 0.014) 0.004* (0.000, 0.009) 0.009 (0.021, 0.004)

0.020* (0.012, 0.027) 0.024** (0.011, 0.036) 0.002* (0.000, 0.005) 0.021** (0.011, 0.030) 0.010** (0.006, 0.013) 0.016 (0.039, 0.008)

0.011** (0.015, 0.007) 0.010* (0.006, 0.013) 0.020** (0.013, 0.028) 0.003 (0.016, 0.024)

0.005* (0.007, 0.003) 0.004* (0.002, 0.007) 0.011* (0.007, 0.015) 0.008 (0.005, 0.023)

0.012* (0.009, 0.017) 0.008* (0.005, 0.012) 3.512*** (2.618, 4.406) 0.785*** (0.619, 0.952)

0.010 (0.056, 0.032) 0.016* (0.009, 0.024) 0.001 (0.002, 0.005) 0.000 (0.014, 0.017) 0.008* (0.003, 0.012) 0.0012** (0.008, 0.015) 0.005 (0.013, 0.004) 0.000 (0.001, 0.000) 0.001 (0.000, 0.003) 0.006* (0.004, 0.009) 0.002 (0.010, 0.004) 0.021*** (0.014, 0.029) 0.016** (0.010, 0.023) 2.721*** (1.813, 3.628) 0.609*** (0.512, 0.708)

0.008** (0.006, 0.012) 0.019** (0.013, 0.027) 3.084*** (2.037, 4.132) 0.721*** (0.508, 0.933)

0.015** (0.009, 0.020) 0.002* (0.000, 0.005) 2.427*** (1.604, 3.249) 0.703*** (0.486, 0.920)

327 197.030

327 282.560

327 211.086

327 302.255

Confidence interval 0.95% in parentheses (lower bound, upper bound). Signification level: regression).

improvement of natural resource use, namely, water (Piedra~ oz et al., 2016). Mun Regarding economic variables, the results are also varied. Firstly, greater yield (YLD) does not significantly affect the reduction of inefficiency related to environmental pressures; even for certain ratios, such as water usage, chemical use, or waste, the sign of the coefficients is negative, likely stemming from the trend in this farming sector in recent years to focus more on product quality
rez-Mesa and Galdeano-Go
mez, 2010). On the than on quantity (Pe other hand, product specialization (SP) does tend to be a determining factor (in many cases with a p > 0.05), as does the support of a local auxiliary industry (AUX) and the proactivity for R&D (RDP) when it comes to achieving greater efficiency and reducing environmental pressures, which is also demonstrated in other recent ~ oz works on sustainability in the sector under study (Piedra-Mun et al., 2017). However, in contrast with other studies on Spanish
nchez Ferna
ndez, 2009), subsidies (SUB) do not appear farming (Sa to have any significant effect on the improvement of eco-efficiency, which might even seem obvious given the scarce influence that public aid programs have in this sector, as previously mentioned
mez et al., 2016). (Galdeano-Go As for the increases in the number of quality certifications (QC), as expected, these positively affect the improvement of ecoefficiency overall (p > 0.05) and in terms of individual pressures, particularly in the use of phytosanitary products and waste management. Similarly, farmers' assessment of innovation in environmental-friendly techniques (EIN) proves to be a determining factor with regard to reducing overall eco-inefficiency margin and most of its individual aspects. These results are in accordance with similar studies related to Spanish agriculture
mez-Limo
n and Arriaza Balmo
n, 2011; Torres et al., 2016) and (Go others specifically focusing on this horticultural sector (Piedra~ oz et al., 2017). Mun

6. Conclusions and policy implications This work carried out an assessment of eco-efficiency of horticultural farms in southeast Spain. This area represents the highest

***

p < 0.01;

**

p < 0.05; *p < 0.1. Number of replications: L ¼ 1000 (for each

production and supply concentration for this agri-food sector in the EU. The goal of finding a balance between economic output and the use of natural resources is an ongoing goal, one that is rooted in tradition yet also fundamental for the future sustainability of this vegetable sector. This is especially true in the semi-arid climate of this area, whose resources are limited by environmental conditions, yet, at the same time, is constantly under pressure to supply a highly intensive farming sector. Therefore, not only is it important to evaluate efficiency indicators by considering various environmental pressures, both aggregated and individually, it is also crucial to analyze factors that might influence the reduction of these pressures. Particularly, the study at micro-level proves to be of great interest as this sector is characterized by the presence of many small-scale farms with a heterogeneous make-up based on technological aspects, their family-oriented production structure, and positive attitude toward environmental respectful practices. These features could be of considerable use for establishing more suitable management actions. The findings of this analysis reveal that, on average, the farms in the sample could reduce their environmental pressures by 11%, while maintaining their economic performance. Nevertheless, inefficiencies vary depending on the type of resource used and its environmental impact. For example, in terms of water usage and fertilizer application, inefficiency margins stand at 5% and 7% respectively, which is the result of implementing new technologies in resource control in the sector over recent decades. Moreover, the necessary reduction of phytosanitary products to achieve a better environmental-economic balance stands at 9%. The highest inefficiencies are found for plant waste management (at approximately 24% inefficiency margin), thereby making it an aspect where a great deal of effort must be made toward reducing this environmental pressure. In the second part of the study, the results obtained reveal that certain characteristics of the farms analyzed, such as experience, farm transfer to subsequent generations, and integration of family farms into cooperatives, influence the reduction of the inefficiencies identified. From the economic point of view, aspects like product specialization, the influence of a local auxiliary industry, and relationships with R&D centers also display a positive


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

influence on improving the use of resources and reducing environmental pressures. Additionally, the adoption of quality certifications and the increased interest of family farms in environmental innovation prove to be important determinants of improving ecoefficiency. From said results emerge various political implications. On one hand, it is necessary to continue promoting programs that reduce the environmental pressures still facing this sector, specifically those which would provide larger investments in recycling (e.g. technologies in the framework of a ‘circular economy’) and aid for waste management training, as well as the promotion of major collaboration between farms and local services and administration in their management. Furthermore, maintaining these familyoriented structures by supporting new generations and their respective cooperative organizations also proves to be an important factor. In contrast, it is observed that current support programs (mainly the Operational Programs of CAP) are of minor significance in terms of improving efficiency, which is why it would be necessary to establish initiatives that more directly promote, for example, R&D proactivity or support for new environmental innovation on horticultural farms. For example, carrying out specific environmental programs and the practice of technology transfer through the local auxiliary industry and the farm organizations themselves could be key tools for reducing ecoinefficiencies. Although the methods applied and the results obtained herein may be of interest to other vegetable production systems or agrifood sectors, particularly those with family-farm structures, the present work has several limitations which could serve as reference for future research. For example, since the present work focused on only one growing season, a posterior eco-efficiency analysis could be conducted to compare subsequent ratios with those initially obtained to determine their evolution over time; additionally, a comparison with other national and international vegetable sectors displaying similar indicators could be of interest as well. Similarly, future work could focus on expanding the factors and characteristics of farms which have an influence on eco-efficiency, as this is an issue that could prove to be of great use for designing initiatives and policy programs aimed at the sustainable of farming sectors. Acknowledgments This research was partially funded by Spanish MCINN and FEDER aid (project ECO2014-52268-P) and by Andalusian Regional Government (project SEJ-2555, Consejería de Economía,
n y Ciencia). Innovacio Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jenvman.2017.09.037. References
nchez, J.A., Galdeano-Go
mez, E., Pe
rez-Mesa, J.C., 2011. Intensive hortiAznar-Sa culture in Almería (Spain): a counterpoint to current European rural policy strategies. J. Agrar. Change 11 (2), 241e261. http://dx.doi.org/10.1111/j.14710366.2011.00301.x.
nchez, J.A., Galdeano-Go
mez, E., Tapia Leo
n, J.J., 2014. Innovacio
n y centros Aznar-Sa
n en la agricultura intensiva de Almería. Cuad. Estudios Agrode investigacio aliment. 6, 205e227.
n, F.D., Parra, A., Rolda
n, A., Soler, A., Vila, E., 2010. Pasado, presente y futuro Beltra del control integrado de plagas en la provincia de Almería. Cuad. Estudios Agroaliment. 1, 27e43.
n Esteve, M.M., 2012. Essays on the Assessment of Eco-efficiency in AgriculBeltra ture (Doctoral Thesis). University of Alicante, Spain.
n-Esteve, M., Go
mez-Limo
n, J.A., Picazo-Tadeo, A.J., Reig-Martínez, E., 2014. Beltra A metafrontier directional distance function approach to assessing eco-

603

efficiency. J. Prod. Anal. 41, 69e83. http://dx.doi.org/10.1016/j.eiar.2017.01.001. Bleischwitz, R., 2003. Cognitive and institutional perspectives of eco-efficiency. Ecol. Econ. 46, 453e467. http://dx.doi.org/10.1016/S0921-8009(03)00186-1.
lisis de la Campan ~ a Hortofrutícola de Almería, Campan ~ a 2014/ Cajamar, 2016. Ana 2015. Cajamar, Almería, Spain.
n-Ferre, A.J., Vel

pez-Martínez, J.A., ManzanoCallejo azquez-Martí, B., Lo Agugliaro, F., 2011. Greenhouse crop residues: energy potential and models for the prediction of their high heating value. Renew. Sustain. Energy Rev. 15, 948e955. http://dx.doi.org/10.1016/j.rser.2010.11.012. Callens, I., Tyteca, D., 1999. Towards indicators of sustainable development for firms: a productive efficiency perspective. Ecol. Econ. 28, 41e53. http:// dx.doi.org/10.1016/S0921-8009(98)00035-4. Cherchye, L., Moesen, W., Rogge, N., van Poyenbroeck, T., 2007. An introduction to 'benefit of the doubt' composite indicators. Soc. Indicat. Res. 82 (1), 111e145. http://dx.doi.org/10.1007/s11205-006-9029-7. Coelli, T., Lauwers, L., Huylenbroeck, G.V., 2007. Environmental efficiency measurement and the materials balance condition. J. Prod. Anal. 28, 3e12. http:// dx.doi.org/10.1007/s11123-007-0052-8. Cook, W.D., Seiford, L.M., 2009. Data envelopment analysis (DEA)-Thirty years on. Eur. J. Oper. Res. 192, 1e17. http://dx.doi.org/10.1016/j.ejor.2008.01.032. Cooper, W.W., Seiford, L.M., Tone, K., 2007. Data Envelopment Analysis. A Comprehensive Text with Models, Applications, References and DEA-solver Software. Springer, New York. De Koeijer, T.J., Wossink, G.A.A., Struik, P.C., Renkema, J.A., 2002. Measuring agricultural sustainability in terms of efficiency: the case of Dutch sugar beet growers. J. Environ. Manag. 66 (1), 9e17. http://dx.doi.org/10.1006/ jema.2002.0578. Downward, S.R., Taylor, R., 2007. An assessment of Spain's Programa AGUA and its implications for sustainable water management in the province of Almería, southeast Spain. J. Environ. Manag. 82 (2), 277e289. http://dx.doi.org/10.1016/ j.jenvman.2005.12.015.
ndez, M.D., Gonza
lez, A.M., Carren ~ o, J., Pe
rez, C., Bonachela, S., 2007. Analysis Ferna of on-farm irrigation performance in Mediterranean greenhouses. Agric. Water Manag. 89, 251e260. http://dx.doi.org/10.1016/j.agwat.2007.02.001.
mez, E., Aznar-S

rez-Mesa, J.C., 2013. Sustainability diGaldeano-Go anchez, J.A., Pe mensions related to agricultural based-development: the experience of 50 years of intensive farming in Almería (Spain). Int. J. Agric. Sustain. 11 (2), 125e143. http://dx.doi.org/10.1080/14735903.2012.704306.
mez, E., Ce
spedes-Lorente, J., Rodríguez-Rodríguez, M., 2006. ProGaldeano-Go ductivity and environmental performance in marketing cooperatives: an analysis of the Spanish horticultural sector. J. Agric. Econ. 57 (3), 479e500. http:// dx.doi.org/10.1111/j.1477-9552.2006.00061.x.
mez, E., Pe
rez-Mesa, J.C., Godoy-Dura
n, A., 2016. The social dimension Galdeano-Go as a driver of sustainable development: the case of family farms in southeast Spain. Sustain. Sci. 11 (2), 349e362. http://dx.doi.org/10.1007/s11625-015-03184.
nchez-Pico
n, A., 2001. Dealing with aridity: García-Latorre, J., García-Latorre, J., Sa socio-economic structures and environmental changes in an arid Mediterranean region. Land Use Pol. 18, 53e64. http://dx.doi.org/10.1016/S0264-8377(00) 00045-4.
mez-Limo
n, J.A., Arriaza Balmo
n, M., 2011. Evaluacio
n de la Sostenibilidad de las Go
micos de Andalucía, Explotaciones de Olivar en Andalucía. Analistas Econo Unicaja, M
alaga, Spain.
mez-Limo
n, J.A., Picazo-Tadeo, A.J., Reig-Martínez, E., 2012. Eco-efficiency Go assessment of olive farms in Andalusia. Land Use Pol. 29, 395e406. http:// dx.doi.org/10.1016/j.landusepol.2011.08.004. Hoang, V.-N., Nguyen, T.T., 2013. Analysis of environmental efficiency variations: a nutrient balance approach. Ecol. Econ. 86, 37e46. http://dx.doi.org/10.1016/ j.ecolecon.2012.10.014. Holm, S.-O., Englund, G., 2009. Increased ecoefficiency and gross rebound effect: evidence for USA and six European countries 1960-2002. Ecol. Econ. 68, 879e887. http://dx.doi.org/10.1016/j.ecolecon.2008.07.006. Hollingsworth, B., Smith, P., 2003. Use of ratios in data envelopment analysis. Appl. Econ. Lett. 10, 733e735. http://dx.doi.org/10.1080/1350485032000133381. Hortyfruta, 2010. I Radiografía del Empleo en la Horticultura Andaluza de Invernadero. Hortyfruta, Frutas y Hortalizas de Andalucía, Almería, Spain. Huppes, G., Ishikawa, M., 2005. A framework for quantified eco-efficiency analysis. J. Ind. Ecol. 9, 25e41. http://dx.doi.org/10.1162/108819805775247882. Keating, B.A., Carberry, P.S., Bindraban, P.S., Asseng, S., Meinke, H., Dixon, J., 2010. Eco-efficient agriculture: concepts, challenges, and opportunities. Crop Sci. 50, 109e119. http://dx.doi.org/10.2135/cropsci2009.10.0594. Kortelainen, M., Kousmanen, T., 2007. Eco-efficiency analysis of consumer durables using absolute shadow prices. J. Prod. Anal. 28, 57e69. http://dx.doi.org/ 10.1007/s11123-007-0046-6. Kousmanen, T., Kortelainen, M., 2005. Measuring eco-efficiency of production with data envelopment analysis. J. Ind. Ecol. 9 (4), 59e72. http://dx.doi.org/10.1162/ 108819805775247846. Lauwers, L., 2009. Justifying the incorporation of the materials balance principle into frontier-based eco-efficiency models. Ecol. Econ. 68, 1605e1614. http:// dx.doi.org/10.1016/j.ecolecon.2008.08.022.
nzalez-Rodríguez, M.J., Belmonte-Vega, A., GarridoMartínez-Vidal, J.L., Go
n por pesticidas en aguas ambiFenich, A., 2004. Estudio de la contaminacio entales de la provincia de Almería. Revis. Cient. Ecosis. 13 (3), 30e38. OECD, Organization for Economic Co-operation and Development, 1998. Eco-efficiency. OECD, Paris.

604


Godoy-Dura
n et al. / Journal of Environmental Management 204 (2017) 594e604 A.

OECD, Organization for Economic Co-operation and Development, 1999. Organization for Environmental Indicators for Agriculture. Volume Economic Cooperation 1-Concepts and Framework. OECD, Paris. Oude Lansink, A., Wall, A., 2014. Frontier models for evaluating environmental efficiency: an overview. Econ. Bus. Lett. 3 (1), 43e50.
rez-Mesa, J.C., Galdeano-Go
mez, E., 2010. Agrifood cluster and transfer of techPe nology in the Spainish vegetables exporting sector: the role of multinational enterprises. Agric. Econ. 10, 478e488.
rez-Mesa, J.C., Galdeano-Go
mez, E., 2015. Collaborative firms managing perishPe able products in a complex supply network: an empirical analysis of performance. Supply Chain Manag. J. 20 (2), 128e138. http://dx.doi.org/10.1108/SCM06-2014-0185.
mez-Limo
n, J.A., 2011. Assessing farming Picazo-Tadeo, A.J., Reig-Martínez, E., Go eco-efficiency: a data envelopment analysis approach. J. Environ. Manag. 92, 1154e1164. http://dx.doi.org/10.1016/j.jenvman.2010.11.025. ~ oz, L., Galdeano-Go
mez, E., Pe
rez-Mesa, J.C., 2016. Is sustainability Piedra-Mun compatible with profitability? An empirical analysis of family farming activity. Sustainability 8 (893). http://dx.doi.org/10.3390/su8090893. ~ oz, L., Vega-Lo
pez, L.L., Galdeano-Go
mez, E., Zepeda-Zepeda, J.A., 2017. Piedra-Mun Drivers for efficient water use in agriculture: an empirical analysis of family farms in Almería (Spain). Exp. Agric. 1e14. http://dx.doi.org/10.1017/ S0014479716000661. Pulido-Bosh, A., 2005. Recarga en la Sierra de G
ador e Hidrogeoquímica de los
n Experimental Cajamar, Almería, Spain. Acuíferos del Campo de Dalías. Estacio
mez, E., Carmona-Moreno, E., GodoyRodríguez-Rodríguez, M., Galdeano-Go
n, A., 2012. Environmental impact, export intensity and productivity inDura teractions: an empirical index analysis of the agri-food industry in Spain. Can. J. Agric. Econ. 60 (1), 33e52. http://dx.doi.org/10.1111/j.1744-7976.2011.01230.x.
ndez, G., 2009. An
S
anchez Ferna alisis de la Sostenibilidad Agraria mediante indi
ticos: Aplicacio
n empírica a sistemas agrarios de Castilla y Leo
n cadores sinte
cnica de Madrid, Madrid. (Doctoral Thesis). Universidad Polite Schaltegger, S., Sturm, A., 1996. Managerial eco-control in manufacturing and process industries. Greener Manag. Int. 13 (1), 78e91. Serrao, A., 2008. Measuring Eco-efficiency of Agricultural Activity in European Countries: a Malmquist Index Analysis. American Agricultural Economics

Association, Orlando, Florida. Annual Meeting 2008, July 27e29. Simar, L., Wilson, P., 2007. Estimation and inference in two-stage, semi-parametric models of production processes. J. Econometrics 136, 31e64. http://dx.doi.org/ 10.1016/j.jeconom.2005.07.009. Tauchmann, H., 2015. Simarwilson: DEA Based Two-step Efficiency Analysis. German Stata Users Group Meeting, June 26, 2015 (Nuremberg, IAB).
n Becerra, A., Lastra Bravo, X., 2010. La agricultura intensiva del Poniente Tolo
stico e instrumentos de gestio
n ambiental. MþA Revis. Almeriense. Diagno
n. Medio Ambiente 8, 18e40. Electro
n Becerra, A., Lastra Bravo, X., Fern
Tolo andez Membrive, V.J., 2013. Huella hídrica y
n al Poniente Almersostenibilidad del uso de los recursos hídricos. Aplicacio
n.Medio iense. Estudios previos y medidas de eficiencia. MþA Revis. Electro Ambiente 14, 56e86. Torres, J., Valera, D.L., Belmonte, L.J., Herrero-S
anchez, C., 2016. Economic and social through organic agriculture: study of the restructuring of the citrus sector in the “Bajo Andarax” district (Spain). Sustainability 8 (918). http://dx.doi.org/ 10.3390/su8090918. Tout, D., 1990. The horticulture industry of Almería province, Spain. Geograph. J. 156 (3), 304e312. Tyteca, D., 1996. On the measurement of the environmental performance firm e a literature review and a productive efficiency perspective. J. Environ. Manag. 46, 281e308. http://dx.doi.org/10.1006/jema.1996.0022.
pez, A., 2016. Greenhouse Agriculture in Valera, D., Belmonte, L., Molina, F., Lo Almería. A Comprehensive Techno-economic Analysis. Cajamar, Almería, Spain. Van der Blom, J., 2010. Applied entomology in Spanish greenhouse horticulture. Pro. Netherlands Entomol. Soc. Meet. 21, 9e17. WBCSD, World Business Council for Sustainable Development, 2000. Measuring Eco-efficiency. A Guide to Reporting Company Performance. WBCSD, Geneva. Webster, P., 1999. The challenge of sustainability at the farm level: presidential address. J. Agric. Econ. 50 (3), 371e387. http://dx.doi.org/10.1111/j.14779552.1999.tb00888.x. Zhang, B., Bi, J., Fan, Z., Yuan, Z., Ge, J., 2008. Eco-efficiency analysis of industrial system in China: a data envelopment analysis approach. Ecol. Econ. 68 (1e2), 306e316. http://dx.doi.org/10.1016/j.ecolecon.2008.03.009.