A path towards sustainability through the circular economy phenom: indicators from the Spanish agri-food sector

According to the Organization for Economic Cooperation and Development (OECD, 2012), the world population is going to surpass 9 billion by 2050. This forecast makes it necessary to increase food production by 70%. In this context, the agri-food sector has become a key factor in the path towards achieving this objective. However, the sector must face the challenge of increasing its production to supply the world market while also looking for new, more environmentally friendly production methods. Nevertheless, the agri-food industry is one of the key contributors to environmental impacts. On the one hand, the production of food requires the use of resources such as fuel, land, water and raw materials. On the other hand, the application of chemical inputs, such as fertilizers or fungicides, creates direct emissions of nitrous oxides contributing to climate change (Scherhaufer et al., 2018). Indeed, there is a great potential in the sector to reduce environmental pressures related to the use of limited natural resources by developing more sustainable business models. Moreover, the growing awareness of human health due to the undesirable effects of hazardous synthetic chemical inputs has encouraged the search for eco-friendly alternatives (De Corato, 2020).

In this international context, the circular economy (CE) model has gained widespread recognition in recent years owing to the goal of maintaining components, materials and products at their highest utility to eliminate waste. However, defining CE is not an easy task. A literature review found 221 different CE definitions (Kirchherr et al., 2023). These numerous definitions apply the 3R principles (Reduce, Reuse and Recycle). What is more, many definitions arise from 2017, highlighting not only the role of business models and consumers as CE agents, but also policymakers and scholars’ implications in the CE transition (Kirchherr et al., 2017). Thus, said review concluded that the definition provided by the Ellen MacArthur Foundation is the most prominent. According to the authors, CE can be defined as “an industrial system that is restorative or regenerative by intention and design. It replaces the “end-of-life” concept with restoration, shifts towards the use of renewable energy, eliminates the use of toxic chemicals, which impair reuse, and aims for the elimination of waste through the superior design of materials, products, systems, and, within this, business models” (Ellen MacArthur Foundation, 2012, p. 7). Therefore, CE is characterized by low consumption of materials, elimination of inputs from fossil or non-renewable sources, low pollution levels and high circulation rates (Jun and Xiang, 2011). As a result, it contributes to the three dimensions of sustainable development (society, economy and environment) as well as towards achieving the main sustainable development goals, such as hunger, health and wellness (Zabaniotou, 2018; Sehnem et al., 2019).

The application of CE in agri-food companies has been studied by several investigations and methodologies. Elia et al. (2017) evaluate a set of indicators and methodologies according to five CE characteristics provided by the European Environmental Agency (EEA, 2016):

1. reducing input and use of natural resources;

2. reducing emission levels;

3. minimizing losses of valuable materials;

4. increasing the share of renewable and recyclable resources; and

5. increasing the durability of products.

Iacovidou et al. (2017) review the methods for achieving resource recovery from waste to promote CE. Pauliuk (2018) proposed a framework of indicators to be used in CE analysis based on material flow, material flow cost accounting and life cycle assessment tools. Furthermore, the European Academies’ Science Advisory Council provides a list with more than 300 indicators that could potentially be used to measure progress in CE (EASAC, 2016). Other studies are focused on CE analysis in a specific context of study. For example, Patricio et al. (2018) and Grimm and Wösten (2018) focused on mushroom production. Although other authors, such as Fernandez-Mena et al. (2016), Kristensen et al. (2016), Pagotto and Halog (2015) and Caruso et al. (2019), concentrate on the entire agri-food sector, most of the research on CE in this sector analyses mono-dimensional indicators (Strazza et al., 2015; Tua et al., 2019; Kurańska et al., 2019; Loizia et al., 2019), which results in a methodology unable to provide recommendations for achieving CE objectives. In addition, few works include a combination of diverse indicators (Genovese et al., 2017; Aravossis et al., 2019; Marino and Pariso, 2020). In this regard, Poponi et al. (2022) have developed a comprehensive dashboard of 102 indicators to assess sustainability at macro, meso and micro levels. In addition, the UNI/TS 11820:2022 standard provides a framework for measuring circularity in organizations, including 71 indicators at the micro- and meso-levels (see e.g. Amicarelli and Bux, 2023 or Amicarelli et al., 2023). This standard is proposed as a certification to homogenize the assessment of circularity across various sectors, including the agri-food sector. In any case, as pointed out by Corona et al. (2019), Heras-Saizarbitoria et al. (2023) or Perramon et al. (2024) and others, more research is necessary to address the difficulties of measuring CE implementation, particularly regarding managerial decisions and strategies.

Accordingly, the first objective of the present study is to offer an overview of the emerging academic literature on CE indicators in the agri-food sector, focusing on the four circular strategies of narrowing, slowing, closing and regenerating. This aims to establish an accurate framework of reference for the application of CE objectives at micro level. Secondly, this CE indicator framework is comprehensively applied to the Spanish agri-food sector.

Within the European context, the agri-food industry has gained significant attention due to its substantial contribution to greenhouse gas emissions (EU, 2021). As Spain is a leading agri-food producer and exporter, this sector is highly internationalized, with exports representing approximately 10% of the EU-27’s total agricultural exports (Ministerio de Agricultura, Pesca y Alimentación, 2023). Moreover, the majority of these exports are destined for European and North American markets (about 70%), which often impose stringent sustainability requirements.

Furthermore, the agri-food sector is embedded within broader supply chains, including complementary sectors such as packaging, transport and distribution. These interrelations, particularly within internationalized supply chains, create an ideal context for analysing CE, as the adoption of circular practices in one sector can have multiplier effects on others (Malik et al., 2023). As the Food and Agriculture Organization of the United Nations (FAO, 2004) highlights, the agri-food sector acts as a positive catalyst for other economic sectors. This is due to the surpluses it generates (including labour, foreign exchange and internal savings), which stimulate economic development in other areas. Furthermore, a recent study by the Spanish Chamber of Commerce (Cámara de Comercio de España, 2024) demonstrates the significant multiplier effect of the sector on other industries and its positive impact on the labour market, solidifying its strategic importance for a country’s economic development. Specifically, the study’s findings reveal that for every euro of gross value added (GVA) generated by the agri-food sector in 2021, at least 1.4 euros of additional GVA were produced in the rest of the economy. Therefore, the Spanish agri-food industry offers a particularly valuable case study for exploring how CE strategies can be applied to complex and global supply chains, with potential transferability to other sectors that share these characteristics (Pérez-Mesa and Galdeano-Gómez, 2015).

Consequently, our study offers a novel contribution to the CE literature by introducing a multidimensional indicator framework comprehensively applied to the agri-food sector at a microeconomic level, integrating miscellaneous pillars of circularity strategies and considering key business determinants (financial, internalization, knowledge, awareness and digitalization). This not only allows for a more complete evaluation of the implementation of circularity in companies but also provides a framework that can be used to guide public policies and business decisions in sectors with a high environmental impact.

The remainder of this paper is organized as follows: Section 2 provides the theoretical framework on drivers and CE indicators in the agri-food sector. Section 3 explains the case study and the methodology procedures. Next, Section 4 presents the results of the empirical analysis. Finally, Section 5 concludes the study, summarizing the main findings and presenting suggestions for future research.

CE activity is a complex process influenced by different factors, one that includes a diverse range of practices that can be classified into distinct strategies (Bocken et al., 2016). Thus, developing a scale for their measurement, achieved by identifying their key indicators, is essential for accurately assessing CE.

Regarding agri-food economic activity, the CE phenom has recently become a focal point of interest in literature (e.g. Poponi et al., 2022), highlighting the importance that the agri-food sector has in the transition towards circular models. This acknowledgement is attributed to its strong environmental impact, along with its intensive use of resources, extensive transport requirements and the considerable amount of generated waste (Crovella et al., 2021; Castillo-Díaz et al., 2023; Correani et al., 2023).

Numerous studies about the implementation of CE strategies at firm level have been analysed (Gusmerotti et al., 2019; Ünal and Shao, 2019; Perramon et al., 2024); however, none is specifically applied to the agri-food sector. What is more, most studies analyse the CE without integrating the four strategies and without being established in a systematic way (Prieto-Sandoval et al., 2019; Heras-Saizarbitoria et al., 2023). Thus, taking into consideration the research carried out by Velasco-Muñoz et al. (2021) applied to the agricultural sector, and following the recommendation given by Bocken et al. (2016), Gallego-Schmid et al. (2020) and Konietzko et al. (2020), those circular indicators related to the CE strategies of narrowing, slowing, closing and regenerating have been considered. Narrowing strategy aims at using fewer resources and involves eco-efficient solutions that reduce resource intensity and the environmental impacts per unit of product or service provided (Mendoza et al., 2017). Slowing strategy involves prolonging and intensifying the use of products to retain their value over time (Gallego-Schmid et al., 2020). Closing strategy aspires to create new value through the reuse and recycling of materials, thereby closing the loop between post-use and production (Bocken et al., 2016). Regenerating strategy includes actions aiming to preserve and enhance natural capital (EMF, 2016).

Table 1 emphasizes key findings from the literature concerning the most common indicators of CE in the agri-food sector.

The inputs used to make a product determine its characteristics and, at the same time, its environmental impact. Thus, reducing the use of pollutant inputs or substituting them for cleaner materials contributes towards decreasing both waste and CO2 emissions. In this context, the materials used to make a product comprise one of the indicators that many studies highlight among the factors necessary for creating products that are more environmentally friendly. Moretti et al. (2020) highlight the substitution of mineral fertilizers for recycled organic alternatives to promote CE mitigation of N20 emissions. In the same line, Yilmaz Balaman et al. (2018) and Poponi et al. (2022) insist on the use of biofertilizers and biomass in the first phase of the agri-food production chain. Other authors, such as Genovese et al. (2017) and Marino and Pariso (2020), emphasize the importance of optimizing the use of raw materials to obtain products. In addition, the reduction in the utilization of non-renewable inputs constitutes a key aspect of environmental efficiency in a product (Pagotto and Halog, 2015). In this panorama, ecological production is positioned as a new agricultural production method committed to efficiency in the use of fertilizers and organic amendments, water and energy (Cajamar, 2020). According to the Codex Alimentarius Commission (2006), ecological/integrated production promotes and improves the health of agroecosystems by enhancing biodiversity, supporting natural biological cycles, and stimulating soil biological activity. It emphasizes the use of management practices, preferring these to the use of external inputs on farms, considering that regional conditions will require locally adapted systems. This is achieved by using, whenever possible, cultural, biological and mechanical methods, rather than relying on synthetic materials, to fulfil each specific function within the system. Thus, ecological/integrated production removes the use of pollutant inputs, such as synthetic fertilizers and pesticides, and reduces the use of non-renewable energy (FAO, 2019). In addition, some environmental policies have focused on eco-design packaging, such as Directive 94/62/EC in the European Union (EU). The reason for this is the large amount of waste that disposable packaging generates and its negative environmental impact (González-Torre et al., 2004). Thus, the use of returnable packaging, which can be reused, contributes by increasing product efficiency while reducing waste and resource consumption. In this line, Hart (1995), Shrivastava (1995) and Christmann (2000) highlight the importance of designing reusable packaging to improve firms’ circular performance. Other authors (Zailani et al., 2012; Wever and Vogtlander, 2014; Wilkström et al., 2016) focus on the importance of including “sustainable” packaging design to fulfil ecological requirements and to encourage customers to reduce food waste as well as recycle packaging. According to Pauer et al. (2019), the implementation of recycled food packaging can help to make production processes more circular. What is more, biodegradable packaging is positioned as a key tool in several sectors to satisfy the environmental requirements of the market, as it is made of non-pollutant materials (Ivankovic et al., 2017).

Moreover, waste management practices play an essential role, primarily because the volume of waste in a process is considered one of the main causes of pollution and, consequently, one of the greatest sustainability challenges for food systems (Poponi et al., 2022). Some authors emphasize the need to keep waste to a minimum to achieve sustainability (Shrivastava, 1995; Norberg-Bohm, 1999; Cheng and Shiu, 2012). In fact, the environmental impact of food waste covers all emissions derived from the different steps of the food supply chain. In this sense, FAO indicated that if food waste were a country, it would be the third biggest CO2 producer after China and the USA. Thus, food waste management is considered an extremely important socio-environmental issue. Marino and Pariso (2020) have introduced various indicators related to waste to analyse the progress towards achieving CE objectives in 28 EU member states. In their study, Fernandez-Mena et al. (2020) consider the indicator “waste management” with the aim of developing a theoretical tool to explore opportunities for reaching CE in agro-business. In the same line, Srivastava et al. (2020) and Gravagnuolo et al. (2019) emphasize the need to introduce waste management practices as a strategic plan for a successful change towards sustainable business models. Recently, concern for levels of food waste has expanded the search for new practices that can contribute towards fighting this problem. The common expression of food loss and waste includes a share of total food production that was originally intended for human consumption but not consumed (Gustavsson et al., 2011). In this sense, Ciccullo et al. (2021) highlight the role of technologies as a solution to tackle food waste. Moreover, authors such as Chaboud and Daviron (2017) and De Steur et al. (2016) defend the reduction of food loss and waste as an economic gain for all actors in the supply chain. Other studies, such as Gustavsson et al. (2011), Timmermans et al. (2014) and Chaboud and Daviron (2017), position the reduction of food waste as an environmental gain, as it also generates a waste of land, water, energy and inputs. Thus, decreasing food waste contributes towards reducing the pressure on natural resources. Furthermore, several investigations view waste management as an opportunity to create energy or power. In their study on the City of Napoli’s agro-industrial economy, Santagata et al. (2020) focus on the conversion of waste into biodiesel. In this line, Kurańska et al. (2019) advocate for the development of chemical components derived from waste materials. Moreover, Kalmykova et al. (2016) defend waste regeneration as a key practice to reduce negative environmental impact.

The total use of water or energy is a widely used method in economic literature for analysing the environmental impact of processes. The negative impact of climate change may also include more extreme weather, such as more periods with excessive rainfall and others with low rainfall resulting in droughts. In addition, in relatively dry climates, small variations in precipitation can cause significant changes in natural recharge of groundwater. This situation can be even more severe in the food production sector due to resulting fluctuations in production and shortages in food supply (FAO, 2010). The balance between food demand and available water for agriculture is a global issue, and water shortage has become the main constraint on food security worldwide (Hanjra and Qureshi, 2010). Thus, water management is the key to ensuring that more food can be produced for the growing population (Bertilsson, 2012). Agriculture is the sector responsible for most water use, consuming 70% of total water use in the world. Therefore, improving agricultural water productivity is an important measure for ensuring global food security, economic development and social stability, as well as diminishing adverse effects on human health (UNESCO, 2012). Consequently, agriculture is one of the sectors where the use of water has been most analysed, and many studies include the optimization of water usage to measure farmers’ environmental impact (Azad and Ancev, 2014; Galdeano-Gómez et al., 2017; Piedra-Muñoz et al., 2017, 2018). Kang et al. (2016) provide a practical application of highly efficient agricultural water use in China, introducing a novel irrigation method and integrative methods in the Shiyang River Basin of Northwest China. Muga and Mihelcic (2008) investigate the sustainability of different wastewater treatment technologies, whereas Bouwer (2000) explores more storage of protected water, including long-term storage to collect water reserves during times of water surplus for use in times of water shortage. According to Gravagnuolo et al. (2019), water management strategies are essential for fostering sustainable business models that help to preserve ecosystems. In this context, studies such as Aravossis et al. (2019), Ignacio et al. (2019) and Poponi et al. (2022) include indicators associated with water management in the search for tools to determine CE values.

Other works advocate for including energy management models in CE development (Hussain et al., 2020). In this line, studies such as Barros et al. (2020) and Sharma et al. (2020) introduce waste-to-energy techniques that help to reduce energy to create a nexus towards circular business models. The use of renewable energy and clean technologies are also essential aspects to developing circularity models. Frondel et al. (2008) highlight the environmental benefits of introducing end-of-pipe technologies in manufacturing processes, whereas Guziana (2011) concludes that clean technologies are more proactively innovative than the former. In this line, in a survey of environmental strategies carried out by companies located in the South of England, Garrod and Chadwick (1996) determined that investment in clean technology is one tool that can be implemented to fulfil ecological requirements. Finally, additional articles exist that address the importance of introducing renewable energies in company processes to improve the quality of life for current and future generations, as well as to meet public environmental objectives (Lacerda and Van den Bergh, 2014; Nesta et al., 2014; Nicolli and Vona, 2016).

Furthermore, material flow analysis is a well-established sustainability method for quantifying resource utilization, monitoring material efficiency and enhancing the adaptation of biological cycles when necessary (Reich et al., 2023; Rocchi et al., 2021). In this regard, material circularity indicator is applied to measure for improving the management of materials, such as optimizing resource exploitation, consumption and environmental protection (Binder, 2007). As a result, environmental sustainability is closely linked to environmental performance evaluation, as the latter is analogous to a control mechanism for pest and fertilizer levels (Frondel et al., 2008). Correspondingly, the most widely implemented system for assessing material control is through environmental auditing certifications (e.g. GlobalGap or GRASP). This implies that any organization seeking environmental certification must adopt pro-environmental activities (Pichlak and Szromek, 2022).

In addition, the current size increase in agricultural vehicles constitutes a key problem regarding energy requirements and CO2 emissions. Thus, developing vehicle management systems that help to optimize transport routes is essential to reduce said emissions as well as fuel use. According to Bochtis et al. (2012), implementing optimal fieldwork tracks contributes to reducing energy requirements and minimizing risk for soil compaction.

Embracing recycling practices is crucial for implementing the CE strategy of closing the material flow loop (Franklin-Johnson et al., 2016; Linder and Williander, 2017). Some investigations highlight the importance of recycling packaging to improve resource efficiency and reduce agri-food waste (Stock, 1992; Carter and Ellram, 1998; Silva et al., 2013). Furthermore, numerous studies uphold the implementation of recycled packaging as a key aspect in CE across the agri-food sector in recent years. For example, Šerešová and Kočí (2020) introduce an indicator related to packaging recycled to examine the environmental positive outcomes.

Few investigations have studied practices that contribute to extending product life in the agri-food sector. However, the most extensively analysed issue is temperature control. According to Beirne (2007), mismanagement of food storage temperature can result in rapid proliferation of pathogens, causing premature product spoilage and an increase in organic waste.

Regenerative practices in the agri-food sector cover a range of activities. Among these are regenerative cultivation methods, developing packaging design for decomposition, increasing carbon sequestration through plant waste management practices, material treatment processes, as well as other practices to preserve ecosystems and foster the transition to circularity (EMF, 2013, 2016; Velasco-Muñoz et al., 2021; Stathatou et al., 2023). However, the vast majority of investigations in this field are focused on the opportunities that regenerative agriculture offers (Rhodes, 2017; Zazo-Moratalla et al., 2019; Schreefel et al., 2020; Dudensing, 2023). Regenerative agriculture is described as having “the intention to improve the health of soil or to restore highly degraded soil, which symbiotically enhances the quality of water, vegetation and land-productivity”. However, no empirical indicators have been specifically developed for this strategy.

One line of CE research is focused on analysing those factors that determine the circularity of a sector. According to Stucki et al. (2023) and Velasco-Muñoz et al. (2021), the implementation of circular activities at the micro level is a question of innovation capability. Important explanatory variables in these models include export activities, innovation knowledge (absorptive capacity, which refers to a company’s ability to acquire, assimilate and use new knowledge), company size, competition intensity and financial resources. These variables are potentially relevant for CE adoption. Following these recommendations, we consider five dimensions in our analysis:

1. financial;

2. knowledge;

3. awareness;

4. internalization; and

5. digitalization.

Financial resources are the primary consideration before implementing any circular investment. In fact, the impact of a company’s financial dimension on innovative activities has been addressed in research (Canepa and Stoneman, 2008; Hall and Lerner, 2010; Hottenrott and Peters, 2012; Stucki et al., 2023). Particularly, cash flow is widely used to analyse the sensitivity of circular investments to changes in available financial resources with the aim of identifying potential financial constraints (Fazzari et al., 1987; Stucki et al., 2023).

Internationalization is an equally important aspect for highly circular companies. Exporting companies increase the number of stakeholders because of their presence in international markets. Thus, some investigations suggest that globalization increases institutional and customer pressure on companies to exceed environmental regulations (Christmann and Taylor, 2001; Shah and Rivera, 2007). In this regard, Ibrahiem and Hanafy (2021) argue that international trade enhances the shift towards renewable energy and energy-efficient technology. Managi et al. (2009) demonstrate that international trade reduces CO2 and sulphur oxides (SO) emissions in countries belonging to the OECD. Furthermore, Al-Mulali and Sheau-Ting (2014) and Al-Mulali et al. (2015) find a positive relationship between international trade and ecological footprint in a comparison study on developing and developed countries. This finding underscores the potential for international companies to contribute positively to the circularity path (Antweiler et al., 2001; Managi et al., 2009; Shapiro and Walker, 2018).

Awareness and environmental corporate culture are a subject that is attracting the attention of recent literature. Most studies have shown that organizational attitudes, governance and cultures may affect firm sustainability (Bleischwitz et al., 2012; Bossle et al., 2016; Dangelico, 2016; Ortiz-de-Mandojana et al., 2016; Tsai and Liao, 2017; García-Granero et al., 2020). The adoption of green awareness within an organization regarding a given environmental issue makes that firm more likely to implement sustainable practices (Liao, 2018). More specifically, corporate environmental performance is regarded as a key driver for improving CE strategies (Porter and Kramer, 2006; Glavas and Mish, 2015; Wijethilake et al., 2016). For example, Newton and Harte (1997) emphasized the significant impact that environmental corporate culture has on the transition towards circular practices. In this sense, the need for culture change rests on the premise that a comprehensive green business ecology is essential to advance towards this goal. Recently, various studies have regarded green business awareness as a good indicator of circularity in the agri-food sector. For instance, Istudor and Suciu (2020) introduce the indicator of corporate social responsibility to analyse the circularity of the food retail sector in the EU. Similarly, Matrapazi and Zabaniotou (2020) study the eco-social business model as an indicator in their circularity analysis of the food waste sector.

Other investigations suggest that knowledge is a decisive determinant for a company’s absorptive capacity, estimated through the level of human capital and the proportion of employees with a corresponding level of education (Cohen and Levinthal, 1989, 1990; Zobel, 2017). According to Stucki et al. (2023), the ability to absorb knowledge is critical for a company’s innovation performance, both in terms of resource inputs (R&D expenditures) and resource outputs (i.e. product or process innovations). As these authors mentioned: “it reflects not only a company’s innovation efforts per se but also its ability to understand, transfer and apply external knowledge internally. A company’s absorptive capacity was equated with its commitment to research and development in general and to product and/or process eco-innovation in particular”. Furthermore, some studies establish a positive relationship between absorptive capacity and firms’ circular performance, emphasizing the importance of knowledge acquisition, integration, translation and exploitation in the development and adoption of green circular practices (Delmas et al., 2011; Siddik et al., 2023).

Digitalization is another essential element in the implementation of CE practices. Companies are using advanced digital technologies such as big data, sensors, 5G, IoT, robotics, Blockchain and others to increase production efficiency, thereby contributing to the achievement of sustainable environmental and socioeconomic benefits and promoting overall CE (Bressanelli et al., 2019; Bag et al., 2020). Indeed, using digital technologies facilitates access to information, which is critical during decision-making processes to tackle sustainability circularity issues (Jun et al., 2009; Preut et al., 2021; Subramoniam et al., 2021). Within the context of CE, digital technologies can support waste reduction and improve resource efficiency (Schulze, 2016; PwC, 2019). For instance, the utilization of internet with sensors can enable smart water consumption in urban centres by collecting real-time information about the quality and quantity of drinking water and wastewater in these areas. This allows the development of more efficient treatment, distribution and collection services (Tsakalides et al., 2018). What is more, implementing IoT and sensors can aid in the development of new energy management practices and tools (Wang et al., 2016; Pawar and Vittal, 2019). Several studies have considered the capability of sensors to detect solid waste amounts, demonstrating their ability to improve collection practices (Reverter et al., 2003; Vicentini et al., 2009; Longhi et al., 2012). Furthermore, other authors have highlighted the contribution of Blockchain technology in enhancing recycling practices and extending product life. In this regard, Romero-Frias et al. (2021) examine its contribution to CE by promoting packaging recycling, improving process transparency, raising awareness and reducing costs in a study of Spanish companies. Magrini et al. (2021) affirm that the implementation of Blockchain helps to maintain control over products until the end-of-life stage. Thus, this technology can be used by manufacturers, service providers and repair centres to collect data on the lifetime of products/components, use phase, maintenance and repair cycles and geo-localize equipment, allowing for a potential life extension.

Based on the described strategies and determinants of agri-food firms, Figure 1 shows the model constructed for the empirical analysis.

The case study is focused on the agri-food sector located in southeast Spain (i.e. the provinces of Granada, Almeria and Murcia). The economic activity of this sector is rather significant, representing 12.7% of agricultural Gross Domestic Product (GDP) in the European context and 10.2% of European agriculture employment in 2020 (Salamanca and Maudos, 2021). The activity of the companies in this region consists of producing and marketing agri-food produce, integrating growers either as associates or as owners of commercial entities or organizations (e.g. co-operatives, limited liability companies, etc.). Greenhouses constitute the main production system in this area (Rodríguez-Rodríguez et al., 2012), and they require the intensive use of resources and generate considerable amounts of waste and residues (e.g. packaging materials, fertilizers, plastics, etc.). On the other hand, the agri-food sector also contributes to the development of services (e.g. financing, consulting, R&D, etc.) and an associated auxiliary industry with a strong environmental orientation (e.g. fertilizers, bees, seeds, etc.), which accounts for approximately 32% of GDP in the area (Aznar-Sánchez et al., 2011; Galdeano-Gómez et al., 2017). Furthermore, this sector clearly targets foreign markets and has a strong capacity for growth and adaptation to new demands – over 60% of the production of these firms is exported, which accounts for over 35% of total Spanish agricultural exports and about 18% of all vegetables consumed in Europe (Cajamar, 2022). Thus, these firms must operate in a highly complex environment and deal with international competitors, regulations, standards and requirements, making CE implementation a highly relevant topic for this group (Antonietti and Marzucchi, 2014; Hojnik et al., 2018). Consequently, said companies have been evolving towards environmental adaptation characterized by a more efficient use of resources and a reduction of environmental impact (Martos-Pedrero et al., 2019). This is particularly important in the agri-food context, where all supply chain members have a high environmental impact (Spielman and Birner, 2008; OECD, 2013). As a result, the agri-food model in southeast Spain has drawn international attention, as several studies show (e.g. Galdeano-Gómez et al., 2013, 2017; Piedra-Muñoz et al., 2016; Godoy-Durán et al., 2017; Castillo-Díaz et al., 2022; Duque-Acevedo et al., 2022), and it constitutes an adequate empirical frame of reference.

The data were collected in 2022 using a questionnaire targeted at individuals within the companies responsible for environmental management. The survey was designed specifically for this purpose based on field studies and the relevant literature on CE strategies at micro level obtained from the literature review. Next, the survey instrument was pre-tested on five firms’ environmental quality managers, and the questions were selected and modified according to their comments and suggestions. Following these steps, the final questionnaire was structured into three main sections:

1. company economic and financial information;

2. perception of drivers’ influences; and

3. a series of items on CE strategies’ dimensions.

According to the Iberian Balance Sheet Analysis System (Sistema de Análisis de Balances Ibéricos in Spanish, SABI), 302 firms commercialized fresh fruit and vegetables in southeast Spain during the period under study. A simple random sample was chosen without replacement. The final number of valid surveys was 93. This represents a satisfactory response rate of 30.8% (Menon et al., 1996).

Based on the previous review of indicators related to CE strategies and determinants, also according to the specific characteristics of the context of study, the selected sub-indicators to be applied to the empirical analysis are shown in Table 2.

The financial variable refers to financial resources and includes one item related to cash flow value (Fazzari et al., 1987; Stucki et al., 2023). Regarding the internalization variable, it represents the influence of market size on CE. The percentage of product exported is an effective measure of market size (Acemoglu and Linn, 2004; Stucki et al., 2023).

Adapted from previous studies (Stucki et al., 2023), the awareness construct includes three variables. It incorporates company age and company size on one hand, and on the other hand, five 5-point Likert scale items related to the introduction of environmental objectives and plans, adoption of circular implementation practices and compliance with environmental initiatives in the business model.

The knowledge variable represents business capacity for implementation of circular practices. In this sense, four variables were used following the recommendation proposed by Stucki et al. (2023). Firstly, R&D expenditure represents the ability to absorb knowledge in terms of resource inputs. Secondly, product and process innovations represent not only a company’s innovation efforts but also its ability to understand, transfer and apply external knowledge internally. Thirdly, employee education level and human capital are relevant elements in the absorptive capacity of a company. What is more, the digitalization variable was included, indicating the degree of digitalization within a company (De Marchi and Di Maria, 2020; Ranta et al., 2021; Stucki et al., 2023).

The CE construct relates to these green practices that companies implement to be more sustainable. Drawing upon previous research (Konietzko et al., 2020), this variable presents a second-order structure formed by narrowing, closing, slowing and regenerating CE strategy constructs. Narrowing strategy is determined by six items:

1. ecological and integrated production;

2. eco-design packaging input;

3. clean technology;

4. water management practices;

5. waste management practices; and

6. transport routes optimization.

In accordance with FAO (2012), the percentage of ecological production was added as a sub-indicator of environmentally friendly inputs. In fact, ecological production removes the use of pollutant inputs, such as synthetic fertilizers and pesticides, and also decreases the use of non-renewable energy. This CE indicator also includes the eco-design packaging sub-indicator, in accordance with Ivankovic et al. (2017). In addition, it encompasses sub-indicators related to water and waste management, incorporating practices aimed at reducing water use (Tang et al., 2013; Piedra-Muñoz et al., 2017) and optimizing transport routes (Bochtis et al., 2012). In addition, adapted from previous studies (Binder, 2007; Sönmez and Mamay, 2018; Pichlak and Szromek, 2022), the material circularity indicator variable reflects the voluntary adoption of proactive material management practices aimed at achieving product circularity.

Closing strategy is assessed by two items: packaging recycled and plastic recycled (Florida, 1996; Banco Interamericano de Desarrollo (BID), 2007; Johnstone et al., 2010; Dalhammar, 2015; Rodríguez and Wiengarten, 2017). In relation to the recycled packaging circularity indicator, the percentage of packaging sub-indicator was considered following the recommendations of Pauer et al. (2019) and Šerešová and Kočí (2020).

With regard to slowing strategy, it is measured by one item related to extending product life practices implementation. Finally, regenerating strategy includes one item related to those green practices that companies implement to preserve and restore the ecosystems.

Once an initial set of CE items was ready, a pilot-test was performed to ensure its reliability and validity. Performing a pilot-test is an important step in the scale development process because it can remove any invalid items (Anderson and Gerbing, 1991; Cheng et al., 2014). For this purpose, five environmental managers from five different marketing-producer companies were asked to review and comment on the items, their clarity, ambiguity, completeness, readability and structure. As a result, 21 multi-item scales were generated, including six constructs (financial, internalization, awareness, knowledge, digitalization and CE).

The reflective or formative relationships of the items with respect to their corresponding latent variables were proposed following the suggestions of Jarvis et al. (2003) and Mackenzie et al. (2005). According to these authors, all constructs have a formative character because they are determined by their items and present indicators that are established exogenously and are not correlated among one another (Chin, 1998). Finally, the relationships between financial, internationalization, awareness, knowledge and digitalization constructs with the CE construct, respectively, are formative; meanwhile, the relationship between the CE construct and its first-order structure (four CE dimensions) is reflective.

A partial least squares based-structural equation modelling method (PLS-SEM) is applied to test the research model and hypotheses proposed (Roldán and Sánchez-Franco, 2012). PLS-SEM method estimates complex cause-effect relationship models with latent variables or constructs. It is composed of two sub-models: the measurement model and the structural model. The former represents the relationships between the observed data and the latent variables. The latter considers the relationships between the latent variables. An iterative algorithm solves the structural equation model by estimating the latent variables using both sub-models in alternating steps. The measurement model estimates the latent variables as a weighted sum of its manifest variables. The structural model estimates the latent variables by using linear regression between the latent variables, as estimated by the measurement model. This algorithm repeats itself until convergence is achieved (Hair et al., 2018). PLS-SEM is considered the most appropriate technique when structural models are complex, with formative and reflective indicators, as in this study (Hair et al., 2014). This method was preferred over covariance approaches since it is designed to predict relationships among variables in relatively small samples (although representative) with less sensitivity to normality assumption (Henseler et al., 2016). It was also applied because it accounts for measurement error and corrects for attenuation, thereby overcoming many of the problems associated with regression models (Jaccard and Wan, 1996). Moreover, due to the shape of the proposed model, PLS was chosen because it allows evaluation of a composite measurement model (Henseler et al., 2014; Sarstedt et al., 2016). As it is a structural model that includes a second-order construct, a build-up approach used (Aldás-Manzano, 2012).

As previous researchers have suggested that unusual patterns of scores can disproportionately influence the results (Tabachnick and Fidell, 2006), an analysis of outliers was conducted with the aim of identifying and discarding them.

Common method variance (CMV) is addressed because the collected data were reported using a single informant from each of the companies, and they were collected from the same questionnaire during the same period of time. Therefore, an exploratory factor analysis was conducted that included all the measurement scales proposed in the model using SPSS. Similar methodological approaches have used CMV to assess the potential existence of common method variance (Cheng et al., 2014; Hojnik et al., 2018).

The results reveal that no single factor accounts for most of the variance and that the first factor captures only 24.7% of the variance, which demonstrates a low threat of common method variance.

Table 3 summarizes and describes the key features of the data set used in the empirical application.

The evaluation of the measurement model is intended to assess the relationships between the indicators and the constructs. Due to the fact that the study uses formative measurements, the measures of the variables were tested and validated in several ways. Two statistical tests were performed to evaluate the formative variables of the model in both steps of the build-up approach method: multicollinearity analysis and analysis of the weight-loading relationship of each indicator (Hair et al., 2014). The relative relevance of each formative indicator was supported by a comprehensive literature review, interviews with managers and previous questionnaires pre-tests (as reported in Section 3.2). Based on the feedback and insights from the interviews with managers, the wording of some items was slightly modified to an acceptable level of significance.

As for another aspect, the existence of collinearity in formative constructs can cause erroneous results. In this line, Hair et al. (2011) define variance inflation factor (VIF) values below 5.00 for each item to avoid multicollinearity problems. As shown in Tables 1 and 2 (see  Appendix), all VIF values are under this value in the proposed model. Therefore, the existence of multicollinearity problems can be rejected, which validates the formative constructs for the model composition.

Once the measurement model was assessed by testing the multicollinearity and the weight-loading relationship of the measurement scales for the formative indicators, partial least squares structural equation modelling (PLS-SEM) was used to test the hypothesized relationships between the latent variables. The steps suggested by Aldás-Manzano (2012) were followed as the proposed model is a second-order construct and it is necessary to apply the build-up approach method. With this approach, firstly, the structural model is estimated, ignoring the second-order construct to calculate the residual value of the first order dimensions. Secondly, these residual values are included as indicators of the second-order construct to estimate the model proposed. The evaluation of the structural model aims to determine the relationships between the constructs. Thus, three statistics were used:

1. structural model path coefficients;

2. coefficients of determination R²; and

3. the predictive relevance Q².

Standardized betas (β) for the path coefficients measure the strength and direction of the significance of the structural model (Wijethilake et al., 2016). According to Chin (1998) and Hair et al. (2014), path coefficients must be above 0.20 to be meaningful predictors. The model presented all path coefficients above 0.20, demonstrating that the relationships maintained are significant. However, following Chin (1998) and Hair et al. (2014), a bootstrapping technique (5000 re-samples) is used to generate standard errors and t-statistics that allow the evaluation of the statistical significance for the relationships hypothesized within the research model. Figure 2 shows the results. All correlations among latent variables are statistically significant.

The Coefficient of Determination (R²), which measures the predictive accuracy, is the central criterion for judging the quality of the partial least squares structural equation modelling (Chin, 1998; Wijethilake et al., 2016). The R² of the model is 0.34, which greatly exceeds the 0.1 minimum level proposed by Falk and Miller (1992), indicating that it is a good explanatory model. Concerning the cross-validated redundancy measure (Q²), it assesses the model’s predictive relevance, i.e. if the model has the ability to predict the reflective indicators of endogenous latent variables. Stone–Geisser’s Q² value was calculated by referring to a blindfolding sample reuse technique with a data omission distance (D) equal to 6 (Wold, 1982). Q² values larger than zero for a particular endogenous construct indicate the path model’s predictive relevance. The Q² value of the model is above zero (0.14), which indicates the satisfactory predictive relevance of the model.

The application of CE models in different economic activities has become a significant concern for achieving sustainability goals. This has implications not only for academia but also for public and private decision-makers. Particularly, the importance of CE analysis has emerged in the agri-food sector in recent years, attributed to the distinctive production features within this sector. It requires intensive use of resources and generates considerable amounts of waste and residues (e.g. packaging materials, fertilizers, plastics, etc.), having a strong impact on the environment and ecosystems. In this sense, the implementation of CE practices is positioned as a key aspect in the transition towards the achievement of more circular production processes.

This study presents novel empirical research in the circularity field, developing a multi-indicator framework on CE strategies, which is applied to southeast Spain’s agri-food sector. Testing the structural model by means of PLS-SEM, the study offers an analysis of the relationship between financial resources, internalization, environmental awareness, knowledge and digitalization with CE. This provides evidence regarding the importance of the drivers that motivate companies to adopt a more circular approach.

The investigation led to the main conclusion that among all the CE strategies, narrowing is the most significant (β = 0.983), followed by regenerating (β = 0.540). Slowing and closing strategies also display significance, though less than the other two (β = 0.083 and β = 0.105, respectively). These results call into question the effectiveness of public policies to support the implementation of CE regenerating and closing practices despite having a positive effect on CE path. In fact, the ecological and integrated production does not reach 24% of total production, regardless of its contribution to mitigating negative externalities. What is more, only three companies acknowledge using regenerating practices. These results are in line with other research that highlights the need to apply agroecology and regenerative principles to improve livelihoods and align the sector with critical global socio-economic boundaries (Stathatou et al., 2023).

Furthermore, the narrowing CE indicators related to clean technology, waste management and water management present great potential for improvement. Only those companies with major financial capacity implement more than one green technique. The reason lies in the fact that the adoption of these practices involves high costs. However, the implementation of these practices is closely linked to environmental and human well-being. In addition, in line with the conclusions obtained by Amicarelli and Bux (2023), and despite the agri-food sector seeming more oriented towards measuring circularity levels, a lack of energy data has been detected. However, the use of energy in agriculture is a key aspect in the transition to CE. In this sense, more efforts should be made with the aim of developing an energy management method that can help to analyse the introduction of renewable energies as well as the reduction of conventional energy consumption to create a stronger nexus towards more circular business models (Albino et al., 2009). On this matter, the application of the technical standard UNI/TS 11820:2022 to the context of study can be useful to investigate the users’ level of perception and awareness of selected indicators, contributing to identify those with difficulties in collecting data for measuring circularity (Amicarelli and Bux, 2023; Amicarelli et al., 2023). Also, while slowing strategies offer significant potential for reducing food waste and extending packaging lifespans, their implementation faces limitations in terms of measurement. Currently, few companies use dedicated indicators to track the effectiveness of these strategies. This lack of data makes it difficult to comprehensively assess their impact on achieving a circular system. Implementing proper storage techniques, extending shelf life through controlled environments and optimizing packaging to minimize spoilage all contribute to slowing down food waste. In addition, using reusable or refillable packaging for products reduces the need for single-use containers, thereby extending the lifespan of packaging materials. However, implementing certain slowing strategies, like robust packaging, might require upfront costs for producers. This phase is currently underdeveloped, despite its crucial importance. In fact, as Bocken et al. (2016) rightly point out, it is essential at this stage to focus on the design of durable products and the creation of service circuits that extend their useful life, such as repair, maintenance, remanufacturing and remanufacturing. It is worth remembering that the main objective of the transition to circular models is to avoid the final phase (end of life) by maximizing the useful life of the products we consume to save resources and improve efficiency.

In fact, the lack of knowledge is a main barrier to the development of CE models, as mentioned by D’Adamo et al. (2024). This leads to another interesting finding, the positive relationship that knowledge and green awareness have with the CE level of agri-food firms. In agreement with the findings of Amicarelli and Bux (2023), there is a positive relationship between the level of awareness and perception of the measurability of the CE increases and the level of education. Thus, it is necessary to provide training for the employees involved in the processes of measurement, especially in those indicators not susceptible to economic evaluation (i.e. recovered water in wastewater). Besides, in line with the results of other investigations (Parr, 2009; Bossle et al., 2016; Dangelico, 2016), greater environmental awareness of the company is reflected in a higher predisposition to introduce more circular practices. In this context, the role of senior staff is a key factor in promoting green values throughout the company (Andriopoulos, 2001; Halbesleben et al., 2003; Rajala et al., 2016). Managers can significantly influence the assessment of conditions to achieve the transition towards CE within their organizations. Also, the education level at organization level is a relevant aspect to have into consideration. For that purpose, it is necessary that regulators actively encourage companies to carry out these circular practices, offering tax benefits for those that incur high implementation costs or offering advisory services without cost to facilitate the path towards circularity. For example, the analysis of the circular capacity of each company by an expert can generate opportunities to improve economic-environmental impact at a lowest cost.

In addition, this investigation is in line with those studies that highlight the importance that market orientation has on the decision to implement circular practices (Sorroche-del-Rey et al., 2023; Stucki et al., 2023). Thus, organizations with a market orientation contribute to achieving CE with the aim of reducing costs, improving company reputation and increasing operational efficiency in terms of the output gained to run a business operation.

The practical implications of our study are significant for both public policy and agri-food companies. On the one hand, for governments, regulators and entrepreneurs, this study provides a conceptual framework that explains which CE indicators should be considered when developing policies that seek to promote and implement circularity. Thus, new policies on energy management, waste management, water management, ecological and integrated production and regenerating practices are the key to turning sector weaknesses into strengths. On the other hand, for companies, acquiring more environmental awareness and knowledge is essential for promoting more sustainable work habits that enhance CE. The adoption of CE strategies by companies is crucial not only to comply with European regulations but also to improve competitiveness in international markets. As highlighted in our research, Spanish agri-food companies must meet increasingly stringent sustainability requirements in their destination markets, especially in the EU and North America.

Our theoretical framework, which integrates multiple circularity indicators (e.g. waste management, clean technology, water management), is not only relevant for the analysis of the agri-food sector but can be applied to other industrial sectors seeking to improve their resource efficiency and sustainability. In this sense, the work contributes to the literature by providing a replicable methodology that can be adapted to different economic and geographic contexts, especially those linked to international supply chains and sectors with intensive use of resources.

The present study, while offering valuable insights, is subject to certain limitations that suggest avenues for future research. In particular, the size and geographic scope of the sample, focused in the agri-food sector, hinders the generalizability of results to other sectors or countries with distinct characteristics. Another aspect to improve is the expansion of the number of indicators analysed. While we included a set of key circularity indicators, the analysis of other critical factors, such as the use of renewable energies or energy management, could provide a more comprehensive view.

To address these limitations, future research could explore the proposed model with a more granular focus on micro-level sub-indicators to validate their applicability and effectiveness in the agri-food sector. This would enable a more nuanced assessment of implemented circular strategies. Furthermore, future research should include more variables related to energy efficiency, especially in sectors that rely heavily on conventional energy sources. In addition, extending this research to other sectors would facilitate the identification of sector-specific CE indicators, gaps in current measurement and the development of novel indicators to enhance circularity assessment. Likewise, analysing the evolution of CE adoption over time could provide valuable insights into how policies and business decisions influence circularity. Finally, a promising line of research would be to study the costs of implementing CE strategies in different sectors and their impact on the adoption of these practices, given that one of the main obstacles observed is the high initial cost of some clean or regenerative technologies.

Funding: This work was partially supported by the European Education and Culture Executive Agency (EACEA), project CIRCULAR (ref. 101082209).

CRediT authorship contribution statement: Conceptualization, investigation and methodology: E.M.G-G., L.P-M. and E.G-G.; Data collection and curation: E.M.G-G. and Y.S-R.; Formal analysis: E.M.G-G. and L.P-M.; Supervision: L.P-M. and E.G-G.; Writing – original draft: E.M.G-G. and Y.S-R.; and Writing-review & editing and project administration: L.P-M.

Declarations of competing interest: The authors declare no conflict of interest.