Reverse logistics of empty pesticide containers: solution or a problem?

ABSTRACT

Pesticide residues has become a serious environmental concern due to the intensification of agricultural production, its universal use and the risk of contamination. It is an issue that is inserted in the dichotomy between the search for increased food production, waste generation and environmental preservation. There are two main objectives for this work. First, the descriptive analysis of the reverse logistics (RL) of empty pesticide containers (EPCs) as it is practiced under the Campo Limpo system (Brazil) so that the factors enabling its success are identified. Second, the proposal of strategies to improve the collection system. Technical visits and a system dynamics approach were combined, using the Behaviour Over Time graph and a causal loop model to describe the problem and propose strategies. Results indicate three key areas of the Campo Limpo system that can be improved: producer awareness, alternative means of returning containers and data transparency. Small and medium producers are more penalised in the return process due to the lack of information, difficulty in accessing disposal centres and the cost of transportation.

KEYWORDS:

• Reverse logistics
• waste minimisation
• end-of-life management
• system dynamics
• Agricultural Plastic Waste

1. Introduction

Brazilian agricultural production is characterised by the massive use of pesticides. Between 1998 and 2008, pesticide sales increased by 945.5% (Sharma et al. 2019). In family farming, there was a notable increase in the use of pesticides between 2006 and 2017 (Valadares, Alves, and Galiza 2020). From 2005 to 2015, the average of new registrations was 140, in 2016 there was an increase to 277 new registrations and in 2020, 493 new registrations (Anvisa 2020). This is a record increase in Brazil, going on the opposite direction of other countries trying to curb the use of these chemicals. Moreover, it shows that the country’s economic dependence on agribusiness leads to permissive policies and regulations to ensure higher crop yields (Bassani et al. 2018).

Empty pesticide containers (EPCs) are the most common agricultural residue and are potentially hazardous for both human health and the environment due to pesticide residue in the containers (Marnasidis et al. 2018). Developing countries do not have enough data to report on management of agricultural residues (Jin, Bluemling, and Mol 2018), without a regulatory body, EPCs can be left or buried in the field as well as disposed of in water bodies, which has been reported happening in Greece (Damalas, Telidis, and Thanos 2008) and Ethiopia (Mengistie, Mol, and Oosterveer 2016). This was also the case for Brazil before the implementation of the National Solid Waste Policy (NSWP). This environment-oriented legislation introduces the concept of extended producer responsibility and institutes a compulsory reverse logistics (RL) for products such as, lamps, batteries, electronic products and lubricant oils, but also for empty pesticide containers. The importance of an adequate end of life management system for EPCs is evidenced by the creation of specific waste management programs throughout the world, such as Pamira (Germany), CleanFARMS (Canada) and Sigfito (Spain).

As for Greece, Marnasidis et al. (2018) proposed three scenarios to aid producers comply with the national management plan regarding EPCs. Shammi et al. (2020) studied perception and behaviour of farmers in Bangladesh towards environmental and health risks concluding that pesticide contamination in farmers happen not only in Bangladesh, but also other developing countries. In South Africa, Aqiel Dalvie et al. (2006) highlights farmer difficulty with EPCs management and the need for public policies addressing damage caused by pesticide usage.

In the Brazilian scenario, to ensure compliance with regulations, pesticide manufacturers made a sectoral agreement to enable EPC return for revaluation through RL and created the National Institute for Empty Container Processing, or InpEV (acronym in Portuguese). InpEV currently represents almost all pesticide manufacturers that market in Brazil, being responsible for coordinating the entire RL process for containers through the creation of the Campo Limpo system (CLS) which is responsible for the RL of all agricultural EPC. In Brazil, small farmers claim a lack of information on how to perform suitable rinsing for EPCs (Freitas, Hoppe, and Murini 2015; Mecabô 2018). Motivations and main success drivers of EPCs RL are evidenced in Oliveira and Camargo (2014), while producer motivation for EPCs return in different regions in the country were surveyed and examined by Carvalho, Ponciano, and de Souza (2016); Ladeira, Maehler, and Nascimento (2012); Mecabô (2018). Marsola, Olivera, and Neto (2020); Oliveira (2019) used Life Cycle Assessment to propose improvements in the RL through transportation adjustments and reaffirmed the importance of the RL system. These studies point to legal pressure and environmental concerns as the main return motivation factors.

Farmers are the end users of EPCs and are exposed to pesticides. The correct use and disposal of pesticides and their packaging are fundamental in reducing pollution, natural resources contamination and pesticide poisoning in agricultural regions. The aim of this research is to describe and evaluate whether the RL process of EPCs through the Campo Limpo System (CLS) is efficient. The main question to be answered is if the CLS is capable of coordinating the reverse supply chain of pesticides given the distinct producer profiles. The hypothesis of this study is that the RL of EPC in Brazil, despite having a shared responsibility structure between the agents involved, does not have a transparent and independent system to efficiently manage the RL. Since, with RL, it is possible to reduce the amount of packaging exposed to the environment and stored by producers, poisoning and the amount of raw material extraction for container production (Li and Huang 2018). After identifying the main obstacles, action strategies will be proposed to improve the reverse collection system.

This research differs from previous work in the following aspects. First, the RL of EPCs is analysed based on observation from technical visits and primary data. Second, the choice of São Paulo state, an area where pesticide consumption is vast and has units of all agents from the analysed system. Third, the EPCs RL was analysed from a holistic perspective, through System dynamics (SD). SD is an appropriate methodology to demonstrate causal relationships between interacting variables (Perlow et al., 2002), it is particularly useful for this study since it can be used to evaluate the interactions between different stakeholders (Davis, Eisenhardt, and Bingham 2007). Finally, improvement strategies where proposed for the CLS concerning: producer conscientization, alternative EPC return means and data transparency. Field specialist surveys were used to validate the results of this work.

For a complex system analysis, be it techno-social or economic-political, as is the RL of EPCs, System Dynamics (SD) is a method that aids in understanding policy resistance origins and problem complexity to conceive of more effective strategies (Sachan, Sahay, and Sharma 2005; Sterman 2000). SD allows flexibility using both qualitative and quantitative methods for a holistic analysis (Maani and Cavana 2000), enabling strategic decisions and policies to guide towards desired behaviours (Rebs, Brandenburg, and Seuring 2019) and improve complex systems (Elias 2017). The use of SD to assess different spheres of RL is found in the literature, including waste electrical and electronic equipment (Ghisolfi et al. 2017), electric car batteries (Alamerew and Brissaud 2020) and social responsibility (Sudarto et al. 2016). The SD method in this study was used for a comprehensive assessment of CLS operation that enabled strategy proposals to improve RL efficiency.

2. Literature review

Concerns about safe commercialisation and use of pesticides led the Food and Agriculture Organization (FAO) and the World Health Organization (WHO) with the support of the United Nations to establish the international code of ethics for the management of pesticides, known as ‘International Code of Conduct on Pesticide Management’ (FAO/WHO 2018; FAO, 2008). However, although most countries follow this Code, its application varies greatly, from a strict regulatory framework in the European Union (EU) to the much more flexible framework in Brazil, influenced by the power of the pesticide industry (Pelaez, Terra, and Silva 2010; Pelaez, Rodrigues, and Ri 2015).

Brazil is considered the pioneer in Latin American and Caribbean countries to implement this waste management legislation (Guarnieri, Cerqueira-Streit, and Batista 2020). In order to ensure compliance with the law, Brazil established, in the Brazilian Solid Waste Policy framework, the creation of sectorial agreements, which are guided by the principle of ‘Shared Responsibility’ (Guarnieri, Cerqueira-Streit, and Batista 2020). This principle covers all actors involved in the different supply chains, making them responsible for waste management and assigning different responsibilities depending on their role in the sector. However, as Singh et al. (2019) points out, the establishment of sustainable systems that lead to the circular economy must consider not only the positions and opinions of representatives of the sector, but also those of formulators and implementers of public policies. In the case of Brazil, it is also noted the absence of a management body that accounts for the complexity of the reverse system of empty pesticide containers (Oliveira and Camargo 2014).

Reverse logistics are the processes to which a given material is submitted after its useful life in order to manage the amount of toxic and non-toxic waste disposed in the environment (de Brito and Dekker 2003). If the residue must return as raw material in its own production chain or to produce another good, RL is necessary (Guarnieri, Cerqueira-Streit, and Batista 2020). A few obstacles in RL implementation are: technology and infrastructure, knowledge and support, operational issues (Moktadir et al. 2020). Late insertion of RL may limit its economic viability (Campos et al. 2017).

The high consumption of plastics in agriculture makes the destination of waste an important environmental issue (Blanco et al. 2018). To this end, logistics are required, specifically on transportation and packaging which, on their own, generates relevant economic and environmental impacts due to the short life span of containers (Barros et al. 2018). When end-of-life is assessed, recycling is an alternative, as it is possible to obtain recycled plastics with good properties (Briassoulis, Hiskakis, and Babou 2013; Picuno et al. 2020). However, a reduced amount of resources are destined for RL and new materials which are specific obstacles that developing countries encounter as opposed to developed ones (Kinobe et al. 2015).

In Brazil, the biggest hurdle to be overcome for a RL system is the high transportation costs due to long distances (Rebehy et al. 2019), a precarious logistics infrastructure (Arkader and Ferreira 2004) and the predominance of the road modal (Oliveira et al. 2021). There are regions in Brazil with low populational density and lack of infrastructure that lead to low mobility conditions, marked by the lack of road network and/or high level of road degradation (Cattini Junior 2015).

Specifically for EPC, there is an additional concern regarding the variety of agents in the reverse chain which was a challenge for EPC RL that had to be overcome by both private and public sectors (Guarnieri, Cerqueira-Streit, and Batista 2020; Oliveira and Camargo 2014). The effectiveness of the RL of EPC is directly affected by the final disposal of the retrieved products, which depend on the adequate rinsing of the containers (Agrawal and Singh 2019; Picuno et al. 2020). Pesticide residue that can remain in the containers after use is a major environmental concern due to massive use of these containers and the contamination risk they represent (Li and Huang 2018).

3. Material and methods

In EPCs logistics flow, industry produces, distributes and the producer begins the RL flow of containers after use. The farmer must first rinse the container properly, by either triple rinsing or pressure rinsing, then return the empty container to the location previously identified on the purchase receipt, which is called collection point, although, in practice, the producer returns it to any collection point, central collection or itinerant collection. If proper rinsing was performed, the container is considered non-hazardous waste and is sent to recycling along with the lids. However, if rinsing was not adequate EPCs are sent for incineration due to the possibility of toxic pesticide residue remaining in the container. The Campo Limpo System only considers two end of life options, recycling or incineration. The logistics and RL flow of EPCs is represented in Figure 1. Pesticide containers are made from different materials. The majority of containers in the CLS are manufactured using high-density polyethylene (HDPE), but they can also be made from co-extruded polyethylene (COEX) or polypropylene (PP) and steel or other metals. Metallic containers constitute less than 1% of the packages that return to the CLS. Empty pesticide plastic containers (EPPC) carry liquid products to be diluted in water and thus are washable, this type of container represents 99% of the packages used for pesticides (INPEV 2018).

Figure 1. Pesticides containers flow through logistics and RL chain in Brazil

The responsibility of waste management that was previously assigned to consumers and municipalities now includes the industry that manufactures these products (Demajorovic and Massote 2017), being an integrated RL chain that owns its success in large part to the mandatory shared responsibility (Oliveira and Camargo 2014; Veiga 2013a, 2013b). For empty pesticide containers, the end-of-life is a mandatory and shared responsibility between several public and private agents, including: the final consumer, the manufacturer/importer, retailers and the public regulatory agency Each chain agents’ responsibility is listed in Table 1.

Table 1. Responsibilities of logistics chain agents

Logistic Chain Links	Responsibilities
Farmers	Rinsing, disposal (puncturing the bottom of the container), temporary storage (for up to one year), return and keeping the return receipt
Cooperatives and Distributors	Identification of return location for the containers in the product purchase receipt, certification of container return and educate farmers about the importance of the container collection and recovery program
InpEV (Pesticide Industry)	All container transportation between receiving units and end of life facilities, disposal of them through recycling or incineration and promotion of educational programs
Public Sector	Licencing of receiving units, supervision of the attributions of all agents and carrying out educational programs

With the intention of improving system understanding and its main obstacles and enabling factors for EPCs collection, the first research stage was the assessment of processes that are carried out in the CLS as well as its primary data. The choice of the State of São Paulo as study area was due to specific reasons that contribute to the quality and quantity of information available, such as: it is the second largest pesticide consumer in Brazil, there are units from all system agents, there is a large volume of collected containers; there is a large rural area with an agriculture dependent economy. This was the first step for the identification of variables for the SD model which was structured as illustrated in Figure 2.

Figure 2. Methodology structure

Phase I. problem structuring

Phase I is constituted by the systematic structuring of the problem (Maani and Cavana 2000). To this end, a behaviour over time graph (Sterman 2000) is developed and stakeholders are identified.

i. Development of behaviour over time (BOT) graph: A BOT graph is a tool used to show trends and patterns of the main variables of a system under analysis. The patterns identified in the BOT can indicate variations and trends, for example, growth, decline, fluctuations or a combination of these, over an extended period of time, usually from several months to several years. The important elements captured by the BOT are drawn in an approximate sense, without exact numerical values (Maani and Cavana 2000).

ii. Reference Literature: the theoretical model for RL of general packaging in Brazil from Demajorovic and Massote (2017), was used as a basis for the causal loop model and the RL of pesticides.

Phase II. causal loop modelling

i. Variable identification: based on the guidelines provided by Sterman (2000), the variables will be extracted from the information collected in the previous steps and will be used to develop the causal loop model.

ii. Development of the causal loop model: the model will be developed from the connection of the main identified variables that are related to the variables in the BOT graph. A causal loop model captures the structure of a system by mapping the cause-and-effect relationships to offer a holistic view and serve the purpose of communication between parties (Sachan, Sahay, and Sharma 2005). Each variable in the casual loop model is followed by an arrow that carries positive or negative signs. Each arrow indicates the direction of change from one variable to another. The negative sign, means that an increase/decrease in variable X will lead to a decrease/increase in variable Y. The positive sign, indicates an increase/decrease in the model of variable X would result in an increase/decrease in the model of variable Y. Combinations of causal relationships, whether positive or negative, give rise to feedback loops, which can be positive (reinforcing) or negative (balancing) (Sterman 2000). The causal loop model was developed with the software Vensin version 9.0.

As a result of this step, we have separate loops that represent different sections of the central problem and improvement strategies for the CLS are proposed.

Phase III. Validation

Additionally, as a complement to data collected on technical visits and in available literature, these strategies are then validated by interviews with experts in the field. These experts were chosen according to their specialisation, picked among those who could contribute with improvement strategies for EPCs RL. The names of the experts were omitted, and interviews were conducted by email and telephone. The interview was composed of four topics: knowledge of producer profile, alternative container collection options, data transparency and improvement suggestions for EPCs RL functioning. This approach was used in de Oliveira and Alvim (2017) and some authors refer to it as rapid assessment or quick appraisal (Dunn 1994). In this methodology, data from secondary sources are used together with non-random samples and semi structured interviews with key players that can be applied in research that necessitates gathering of data and/or more information to understand the dynamics of a sector.

4. Results

To understand and analyse trends in the CLS, a BOT was developed (Figure 3). It shows the counter-intuitive nature of the problem across 10 years of data, from 2008 to 2018 through the normalisation of five variables: rate of return of EPCs, which is the annual rate of return for EPCs according to InpEV; number of collection units, which is the number of units that receive EPCs, composed of collection points and central collections that are in operation each year; number of industries members of the CLS, these are pesticide manufacturers associated with InpEV that collaborate with the CLS; crop area in São Paulo state, which is the total amount of land used for growing produce in the state and sales of pesticides in São Paulo

Figure 3. Behaviour over time chart for the campo limpo system

From the BOT graph we can see the following potential problems of the CLS:

(i) Crop area increases over the whole period while the number of collection points decreases.

(ii) The quantity of returned containers has increased, but the collection points have been decreasing.

(iii) The sale of pesticides grows at a higher rate than the number of collection points.

(iv) Number of industry members in the CLS increase without a corresponding increase in collection points

Sugarcane is the main agricultural product in the state of São Paulo, however, soybean production has been playing an increasing role in the last five years as an alternating culture in areas where sugarcane soil is been renewed and in agricultural expansion areas (IEA, 2019). Soybean production interferes in the competition for productive areas due to its high liquidity in the domestic and international grain markets, this dynamic stimulates the process of crop replacement. This is relevant since soybean production uses 17.7 liters/ha of pesticide while sugarcane only requires 4.8 liters/ha (Pignati et al. 2017).

An important factor for the correct understanding of Figure 3, is that in the last 3 years of the depicted time period there was an atypical movement in the volume of sales of pesticides due to the fall in agricultural production, attributed to a drought in the state of São Paulo in the expected harvest period. Also, in 2016 sales of pesticides fell in Brazil due to currency devaluation relative to USD and illegal pesticide usage went up as well as new plague control technologies (Cruz 2017).

From the preliminary theoretical model for RL of packaging in Brazil (Demajorovic and Massote 2017) and the potential problems presented in Figure 3, it was possible to build the causal loop model presented in Figure 4. The casual loop model was analysed by identifying seven feedback loops operating in the system, namely: R1 loop represents the CLS, R2 loop represents the pesticide industry, R3 loop represents large Scale producers, B1 loop represents the cost of Transportation for small scale producers, B2 loop describes producer awareness, B3 loop represents CLS infrastructure and B4 is the data loop.

Figure 4. Casual loop model for the Campo Limpo system

Positive feedback loops, for this case under analysis: R1, R2 and R3 accelerate the change within the systems, which can result in rapid growth. On the other hand, negative feedback loops are balancing and dampen changes (Kotir et al. 2016), they are B1, B2, B3, and B4.

The use of pesticides involves several stakeholders: government, industry, retailers and farmers. In EPCs logistics flow, the pesticide industry manufactures, distributes, collects and disposes of containers, while producers must rinse and return EPCs, with all stages of the process being done by InpEV (R1 loop). For the functioning of this system it is important to highlight population density, urbanisation index, consumer location, environmental awareness and collection costs, which are affected by logistics organisation and infrastructure (number of collection points, central collections, itinerant collections and amount of EPCs). Companies must adopt measures such as EPCs return procedure implementation and collection point provision (B3 loop) (Rebehy et al. 2019). Having a shared RL system is cheaper for the companies than creating their own RL processes due to economies of scale that are fundamental for the functioning of CLS (R2 loop). The proper rinsing of EPCs by the producers is crucial for InpEV since containers with any remaining pesticide residue are considered hazardous waste and require special transportation and disposal, increasing costs (Jin, Bluemling, and Mol 2018)(R3, B1 and B2 loops).

An EPCs returned in the CLS always follows the flow described in Figure 1, therefore, as soon as the producer returns it, the container will be transported by trucks at least twice (producer – collection central – end of life). Long distances are widely discussed in studies on this subject due to the vast territorial span of the country that affect any logistics systems’ economic efficiency. In (Lopes et al. 2017; Oliveira and Magrini 2017), the obstacle for small producers to return containers due to transport costs is pointed out (B1 loop). This obstacle highlights the distinct characteristic of the system stakeholders, small and large producers cannot be held to the same standards. In order to comply with the NSWP, large rural producers have the advantages of diluting transportation costs in their revenue and using their already available transport fleet and employees (B3 loop).

The number of EPCs return units is still limited in São Paulo state (B3 loop). In 2018, rural producers were able to return their EPCs to 37 collection points and 14 collection centrals as well as itinerant collections that were implemented in 2013 across the country. However, São Paulo only had 13 itinerant collection in 2018 according to InpEV as illustrated in Figure 5, which shows that itinerant collections only happened on a central strip spanning from west to southeast regions of the state. In addition to a spatial analysis of the collection units, the volume and distribution of agricultural production in the state of São Paulo must be evaluated. This analysis took into account the main agricultural products of the state in 2017, namely: cotton, peanuts, rice, oat, banana, potato, coffee, sugarcane, onion, beans, orange, manioc, corn, soy, sorghum, tomato, wheat and grapes (IBGE 2017). The number of collection units is directly associated with the distance that the rural producer must travel to return an EPCs, which, in turn, is directly associated to transportation costs (B1 and B3 loops).

Figure 5. Spatial representation of the Campo Limpo system unites and agricultural production in the state of São Paulo

Pesticide sales is causally linked to producer exposure to the product, which is also related to the number of containers to be returned (R1 and B4 loops). It is important to highlight that EPCs, being plastic residue, has an extended natural degradation period and, through time, it can form barriers that affect soil aeration, water body and soil contamination, becoming hazardous for the environment and to human health. Most organisms are exposed to toxic organic residues from pesticides that impact the environment (Jin, Bluemling, and Mol 2018; Li and Huang 2018). Therefore, the appropriate return of agricultural residues can mitigate its environmental impacts and is an important issue to be discussed and improved. The absence of regulations relative to the use of pesticides in developing countries, contributes to the high incidence of intoxications (Hvistendahl, 2013). The enactment of the NSWP was able to dimmish the EPCs disposal problem through the implementation of RL and has avoided a significant amount of pesticide residue in the environment, for previous to it producers burned containers in the field, contaminating air, soil, water.

Compliance with NSWP is based on shared responsibility, for which it is necessary that all parties are aware of their obligations. For small rural producers, it is usually harder to access and understand information. Medium and small producers, due to lack of awareness and resources, can build inappropriate EPCs storage facilities or do not do it at all (Wandscheer and de Carvalho 2016). It is the duty of the manufacturing industry, the government and InpEV to guide and educate rural producers (B2 loop). The operational cost for municipalities that implement RL, in terms of environmental education, is around U$2,000.00 per month for municipalities with up to 30,000 inhabitants, U$5,000.00 for municipalities with up to 100,000 inhabitants and for municipalities with over 250,000 inhabitants, U$10,000.00 (IBAM, 2012 apud; Rebehy et al. 2019).

Although the programs statistical data points to an efficient system, it was not evident in any of the studies referenced that small farmers collaborate effectively to sustain the high return rate of containers at 94% (B4 loop). Another important factor that should be emphasised is that there is no impartial database on pesticide use or sale in Brazil. The industry providing the data is the same as the industry that controls InpEV. The reliability of the data may compromise the validity of any study. In this case, we observed that the change in the container return rate calculation methodology has been biased over the years to increase the return rate.

In the early years, InpEV calculated the rate of return based on the return of primary packaging (which has direct contact with pesticide) and secondary packaging (which does not have direct contact but stores primary packaging). Then the rate of return was calculated only through the primary packaging returned and, in 2019, it appeared on the InpEV website that only plastic containers were considered for calculating the rate of return.

Moreover, the CLS disregards the estimated 20 to 25% of the total volume of pesticides used in Brazil, which originates from non-legal sources (INPEV 2018) a much higher figure than that of developed countries, which usually varies between 5 and 7% (Płonka, Walorczyk, and Miszczyk 2016).

5. Discussion

In order to improve the CLS, the following strategies are proposed by this group of authors, the points in which they are inserted in the causal loop model are illustrated in Figure 6.

Figure 6. Strategies to solve the problems of the campo limpo system

(i) Producer awareness

The search for more sustainable agricultural practices and the rational use of pesticides has a multidisciplinary character, requiring that all interested parties (for example, farmers, consumers, agrochemical and food industries) receive adequate training, taking advantage of the acquired experience, and work together (Loha et al. 2018; Lykogianni et al. 2021).

Since it is common among small Brazilian producers to have a low level of education (Recena et al. 2006), technical language and letter size on EPCs labels and instructions are an entry barrier for the producer (Marques et al. 2019; Recena and Caldas 2008). Therefore, members of this group may have difficulties in reading pesticide container labels and instructions, which provide specifications in technical language regarding dosage, dilution, storage, among others, increasing the incidence of poisoning (Oliveira-Silva et al. 2001; Valadares, Alves, and Galiza 2020). Added to this is the tendency of producers to fail to comply with what is required by law when there is no enforcement (Marques et al. 2019). Thus, the first alternative is to continue investing in educational programs for farmers, so that they can understand the need for the correct return of EPCs through the CLS.

Frische et al. (2018) determined that one of the principles for the implementation of sustainable pesticide use is to provide support to farmers, both regarding training and application as well as minimising the use of pesticides. The main source of information on pesticides for producers is the retailers (Recena and Caldas 2008)), and these transmit the information directly to the producers. So, oral communication was identified as the main means of communication for farmers regarding the use and danger of pesticides (Yang et al. 2014). Investment is required to train this group, which is not foreseen in the NSWP. Schools in rural communities must carry out educational work with children, which can be a vector of knowledge for parents. Awareness and education programs can start from municipalities, reaching isolated rural producers. Labels and package inserts must be written in a simple and objective way to facilitate the understanding of the products. Simple changes, such as increasing the label letters and simplifying language would contribute positively to producers. Environmental education and awareness practices can guarantee the maintenance of this broad reverse system (Oliveira and Camargo 2014).

Participants in food production chains must be constantly informed about the safe use of pesticides. This should include access to up-to-date information on pesticides, use of generic and counterfeit pesticides, and proper waste disposal (Zikankuba et al. 2019). In addition, alternative handling and safe practices for pesticide use can be taught to small producers through government and rural development programs. Since the small producer is the one who suffers most from the lack of information. Farmers’ knowledge and awareness of the risks associated with the life cycle of pesticides is essential to ensure the efficiency of the system and protect stakeholders (Yang et al. Yang, et al., 2014.). All specialists reinforce the importance of making producers aware for the good performance of the RL.

(ii) Alternative means of returning containers

In Brazil, the biggest hurdle to be overcome for a RL system is the high transportation costs due to long distances (Rebehy et al. 2019), a precarious logistics infrastructure (Arkader and Ferreira 2004; Alamerew and Brissaud 2020) and the predominance of the road modal (Coleti and Oliveira 2019). There are regions in Brazil with low populational density and lack of infrastructure that lead to low mobility conditions, marked by the lack of road network and/or high level of road degradation (Cattini Junior 2015). RL of EPCs in Brazil was designed for locations where there is possibility of economies of scale (Veiga 2013a).

Through the decentralisation of collection points the producer can be closer to disposal units. Even if this increases the number of individual trips, the freight to end of life facilities is done in an optimised manner since the cargo is already processed. In Fatoretto and Oliveira (2019) it was found that the reduction in freight distances, through the decentralisation of collection points, was a positive factor for the reduction in environmental impact.

Several alternatives for the return of EPCs have already been provided by the literature. As a suggestion of improvement in the system, it was proposed by Castro (2011) to shred the container in all units in the chain, in order to reduce the trips between these units and recycling companies. Veiga (2013b) proposes co-collection, in which a vehicle has several compartments for different materials or bag them separately and a decentralised system, as also suggested by specialists.

The pesticide resale channels (distributors) and cooperatives that appear in Figure 1, in the direct logistic flow should be able to store clean containers. However, if the container has not been properly rinsed, complications arise for its transportation, storage and end of life, since this type of packaging is described by NSWP as ‘hazardous waste’ and a series of regulations must be followed to carry out the processes described above. The farmer should receive a note that confirms the delivery of EPCs, it could easily be given if the above strategy of a single system for all interested parties was in practice. The itinerant collection system organised by the collection points and InpEV must be maintained and expanded, in these events it is also important to emphasise for the agricultural producer the importance of adherence to the CLS and correct rinsing of EPCs.

The CLS is subsidised by rural producers because the cost of the system is already included in the final price of the pesticide paid by the producer. In this way, the pesticide manufacturing industry has an obligation to effectively provide this value in favour of the rural producer, by expanding the number of collection points and central collections. Currently, this does not happen and the extra charge for rural producers is being turned into more profit for the pesticide industry. Public policies that limit the maximum displacement the producer can make to return EPCs and stricter sanctions for not promoting initiatives and alternatives for producers distant from collection units could help sustain the system.

(iii) Data transparency

Data transparency is a very controversial topic, some specialists believe that the data released is clear, others that there is a lack of precision. As pointed out by Rebehy et al. (2019), centralised management has a greater capacity to implement efficient systems, however there are risks associated with low transparency and the formation of a cartel or monopoly. Qian et al. (2018) points out that policy makers, especially in developing countries, should review, provide transparency and enforce pesticide regulations. (Frische et al. (2018) argues that an effective, independent and transparent monitoring system is needed. The legal obligation of farmers to document pesticide use (in application records) provides an adequate basis for traceability. This requires random or targeted inspections that must be conducted by a government agency. The conflict of interest in the Brazil case is evident, since the pesticide manufacturing industry is the same responsible for supporting InpEV, operator of the CLS.

Transparency of practices for compliance with NSWP is an industry duty. As an alternative to improve the transparency of information and data released by the Institution, it would be advisable to carry out audits through external companies. Another measure that could be adopted by the pesticide industry is to supply sales data indirectly, such as by volume in kilograms of the packages sold. It is worth mentioning that currently the calculation of the rate of return is based only on plastic containers not considering other materials (such as cardboard), which distorts and bloat the real value of the rate of return. Another important aspect that deserves mentioning is the illegal entry of pesticides and the commercialisation of pirated products in Brazil. InpEV itself estimates that ¼ of the pesticides used in Brazil are of illegal origin and these products are also not considered in the rate of return statistics.

Also, although the legislation determines the mandatory use of the term pesticides, one of the main industry communication strategies is the use of the term agricultural defensives in communication and dissemination materials, as an attempt to understate risks inherent to the product (Leal and Lopes 2019).

6. Conclusion

In this work, data collected from the Brazilian system of reverse logistics of empty pesticide containers was analysed and discussed to propose system improvement strategies. The RL of pesticide containers through the CLS is a solution for the correct end-of-life destination. The contributions of this study are the RL optimisation strategies conceived through a holistic system analysis. The results indicate that strategies to improve the performance of EPCs RL focus on three major areas: producer awareness, alternative means of collection and data transparency.

Reverse Logistics for pesticide packaging is essential to recover waste that is harmful to the environment, improve the use of resources and ultimately avoid long-term damage to the environment and society. This increase in efficiency can be achieved with the expansion in the number of collection units, itinerant collections and alternative means of returning EPCs without increasing costs for producers. This practice still applies an uneven burden on small producers due to storage and transportation costs while being the ones that are ultimately penalised for non-compliance. There is industry lethargy on investment for the expansion of collection units.

Awareness and environmental educational strategies for rural producers can be improved by more effective targeting. There should be educational programs specifically aimed at improving farmers’ awareness, especially small and medium-sized ones, about the life cycle of pesticides, that is, including the process of storage, rinsing, transportation and proper disposal. On the other hand, for large farmers, awareness must be directed towards reducing the use of pesticides and more sustainable production practices.

The lack of transparency in a mandatory process by the NSWP must be overcome, especially in the case of pesticide residues that are notably hazardous toxic waste and that can cause damage to public health and the environment, from handling to improper disposal.

Lastly, limitations of this study are the geographic cut, since not all states have the same collection infrastructure, and that the strategies were corroborated by experts and researchers, but are yet to be tested. For future work, expanding the analysis for these chain agents can subsidise broader and more assertive public strategies and policies, and studies discussing facility locations for the CLS.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001 and The Brazilian National Council for Scientific and Technological Development (CNPq), grant #144.566/2019-2.

Notes on contributors

Karina Braga Marsola

Karina Braga Marsola is a Ph.D. candidate at the School of Agricultural Engineering at the University of Campinas (UNICAMP), Brazil. She is a researcher at the Agroindustrial Logistics and Commercialization Laboratory (LOGICOM/UNICAMP). Her main research interests are in the areas of reverse logistics, life cycle assessment, food supply chain management and agricultural commercialization. She holds a Master of Agricultural Engineering from UNICAMP with an academic excellence scholarship stay at the University of Porto, Portugal.

Andréa Leda Ramos de Oliveira

Andréa Leda R. Oliveira is an Associated Professor at the School of Agricultural Engineering at the University of Campinas (UNICAMP), Brazil. She is the Coordinator of Agroindustrial Logistics and Commercialization Laboratory (LOGICOM). Her main research interests are in the areas of food supply chain management, sustainable supply chain management, food loss and waste, transport management and operational research. She holds a Split PhD in Economic Development from UNICAMP and Technical University of Dresden (TUD University), Germany and a Master of Electrical Engineering from UNICAMP.

Monique Filassi

Monique Filassi is a Ph.D. candidate at the School of Agricultural Engineering at the University of Campinas (UNICAMP). She is a researcher at the Agroindustrial Logistics and Commercialization Laboratory (LOGICOM/UNICAMP). Her main research interests are in the areas of food supply chains, agricultural commercialization, multidimensional analysis, sustainable development, sustainability indicators. She holds a Master of Agricultural Engineering from UNICAMP.

Arun A. Elias

Arun A. Elias is the Associate Dean (International and Accreditation) at the Wellington School of Business and Government, Victoria University of Wellington, New Zealand. His main research interests are in the areas of systems thinking and modelling, stakeholder management, and sustainable supply chain management. He holds a PhD in Management from Victoria University of Wellington, a Master of Industrial Engineering and Management from IIT Kharagpur and a Master of Agricultural Engineering from Allahabad University.

Fernando Andrade Rodrigues

Fernando A. Rodrigues is a Postdoctoral Research Associate at the Department of Energy of the Instituto Tecnológico de Aeronáutica (ITA - Aeronautics Institute of Technology), Brazil. His main research interests are in the areas of sustainable energy solutions, numerical modeling, and fluid mechanics. He holds a PhD in Mechanical and Aeronautical Engineering from ITA, and a Master of Mechanical Engineering from the University of Massachusetts, USA.