DF A Product Chain Organization study of Brazilian soy The role of financial actors Master’s thesis in Industrial Ecology FRANCESCA MAGNOLO Department of Technology Management and Economics Division of Environmental System Analysis CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2019 Report No. E2019:106 Master’s thesis 2019, E2019:106 A Product Chain Organization study of Brazilian soy The role of financial actors Francesca Magnolo Supervisor and Examiner: Henrikke Baumann, Professor at Technology Management and Economics, Chalmers Co-supervisor: Alice Dauriach, PhD candidate, Global Economic Dynamics and the Biosphere, Royal Swedish Academy of Sciences Department of Technology Management and Economics Division of Environmental System Analysis Chalmers University of Technology Gothenburg, Sweden 2019 A Product Chain Organization study of Brazilian soy The role of financial actors FRANCESCA MAGNOLO © FRANCESCA MAGNOLO, 2019. Master’s Thesis 2019 E2019:106 Department of Technology Management and Economics Divison of Environmental System Analysis Chalmers University of Technology SE-412 96 Gothenburg Telephone +46 (0) 31 772 1000 Cover: Land use change between 1986 and 2018, São Félix do Xingu, Brazil (NASA). Printed by Chalmers digitaltryk Gothenburg, Sweden 2019 iv A Product Chain Organization study of Brazilian soybean The role of financial actors FRANCESCA MAGNOLO Department of Technology Management and Economics Chalmers University of Technology Abstract Deforestation in Brazil is a result of a complex network of actors and their interests around farmland and agricultural products’ chains. A particular interest by foreign investors on Brazilian farmland has also increased after the global financial crisis, leading to a process of financialization. The complexity of interests and financial- ization make sustainable management and governance of global product chains very difficult. What is needed, is an understanding of how this system of interests works and how this leads to negative impacts, in order to find mitigation solutions. The aim of this thesis project has been to describe and model the actor-network around the soybean product chain in order to support its sustainable management. The method used consisted in a Product Chain Organization (PCO) study which has looked at the actors, especially focusing on financial actors, along different steps of the chain, identified using a life-cycle perspective. The focus on financial actors resulted in following both the product chain and the financial chain. For doing so, a document study on soybean companies, foreign investors, Brazilian communities and deforestation policies has been performed as well as two interviews to some important financial actors. The model created showed the importance of specific types of actors along the chain and the links between them. Some investors own very large shares of agricultural funds or companies for farmland investments. Hence, they can exert a strong influ- ence on how agricultural funds and their farmland investments are managed, with the power of having a highly positive or negative impact on a local level. So far, the main limitation of investors in their approach to sustainable investing is due to the subordination of environmental and social benefits to financial ones. Thus, there is still a gap between sustainability commitments and actual outcomes. The method used proved that LCA studies and PCO studies can complement each other. PCO studies, looking at the interactions of actors responsible for the actual global flow- ing of products, provide opportunities for minimizing the impacts identified by LCA studies. Moreover, this thesis project has shown that there is scope for all types of financial actors to reduce deforestation and build more sustainable global product chains. Remote actors, such as foreign investors, have a concrete influence on a local level and the power of changing the fate of vast areas of territories. Thus, the financial root is of interest and worthy of further study for sustainable management. Key words: Deforestation, soy, Brazil, Financialization, Product Chain Organiza- tion, Sustainable management, Life Cycle Assessment. v Acknowledgements I would like to thank my examiner and supervisor Henrikke Baumann, for teaching me so much in these five months. For her constant support and guidance, for all our chats and discussions and for always being present when I needed the most. I would also like to thank my co-supervisor Alice Dauriach for all the valuable inputs and comments, and for always encouraging me. I wish to thank all the Environmental System Analysis division, Professors and PhD students, for always warmly welcoming me in all their activities. That made writing my thesis a much more pleasant work. I would also like to thank Allison Spector, Sustainability Director for Nuveen and Christina Olivecrona, Sustainability Analyst for Andra AP-fonden, for their assis- tance and help, always open to answering all my questions. I’m eternally grateful to my family, for always being there for me despite the dis- tance and for giving me the chance to be here to follow my dreams and ambitions. A special thank to my sister, for all the calls, laughs and visits, also in the rainiest and coldest periods. Finally, my gratitude and love goes to Marco, who has always believed in me and supported me in every possible way. We made it together. vii Contents List of Figures xiii List of Tables xv 1 Introduction 1 1.1 The rise of soy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.1 Soy production in Brazil . . . . . . . . . . . . . . . . . . . . . 3 1.2 Soy deforestation in Brazil . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2.1 The Brazilian Amazon . . . . . . . . . . . . . . . . . . . . . . 4 1.2.2 The Cerrado . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.3 The political threat to the Amazon, the Cerrado and to their indigenous communities . . . . . . . . . . . . . . . . . . . . . 6 1.3 The need for a bigger picture of interests . . . . . . . . . . . . . . . . 9 1.4 Aim of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 The state of knowledge on the link between finance and farmland 13 2.1 The role of financial actors in Brazilian farmland and soy production 13 2.1.1 Land grabbing and financialization of land . . . . . . . . . . . 14 2.1.2 Foreign investors in Brazil and their link to deforestation . . . 15 3 Method 17 3.1 Framework: IE and PCO . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2 Design of the study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2.1 The product chain from an LCA’s perspective . . . . . . . . . 18 3.2.2 The influence of different types of actors: a PCO’s approach . 19 3.2.2.1 Global Wealth Chains . . . . . . . . . . . . . . . . . 19 3.3 Data collection, method, and analysis . . . . . . . . . . . . . . . . . . 20 4 The Soy Product Chain Organization 23 4.1 The product chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 4.1.1 Agricultural land creation . . . . . . . . . . . . . . . . . . . . 24 4.1.1.1 The process . . . . . . . . . . . . . . . . . . . . . . . 25 4.1.1.2 Environmental implications . . . . . . . . . . . . . . 25 4.1.1.3 Policies and commitments . . . . . . . . . . . . . . . 26 4.1.2 Agricultural production . . . . . . . . . . . . . . . . . . . . . 28 4.1.2.1 The process . . . . . . . . . . . . . . . . . . . . . . . 28 ix Contents 4.1.2.2 Enviornmental implications . . . . . . . . . . . . . . 29 4.1.2.3 Yields and production . . . . . . . . . . . . . . . . . 29 4.1.2.4 Some powerful actors . . . . . . . . . . . . . . . . . . 31 4.1.2.5 The relationship between production and finance . . 32 4.1.3 Processing and trading . . . . . . . . . . . . . . . . . . . . . . 34 4.1.3.1 The process . . . . . . . . . . . . . . . . . . . . . . . 34 4.1.3.2 The environmental implications . . . . . . . . . . . . 34 4.1.3.3 The power of soy traders . . . . . . . . . . . . . . . . 34 4.1.3.4 Archer Daniels Midland . . . . . . . . . . . . . . . . 36 4.1.3.5 Cargill . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4.1.3.6 Bunge . . . . . . . . . . . . . . . . . . . . . . . . . . 38 4.1.3.7 COFCO and the Asian influence . . . . . . . . . . . 39 4.1.4 Final processing . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.1.5 Retail and consumption . . . . . . . . . . . . . . . . . . . . . 41 5 Agricultural land creation 43 5.1 Actors in the product chain . . . . . . . . . . . . . . . . . . . . . . . 43 5.1.1 Capitalized farmers . . . . . . . . . . . . . . . . . . . . . . . . 43 5.1.2 "Sem terras" and small farmers . . . . . . . . . . . . . . . . . 44 5.1.3 Ranchers and land grabbers ("grileiros") . . . . . . . . . . . . 45 5.2 The wealth chain: The responsibility of foreign farlmand investors. . . 47 5.2.1 Foreign pension funds . . . . . . . . . . . . . . . . . . . . . . 48 5.2.1.1 The TIAA’s "empire" in Brazil . . . . . . . . . . . . 48 5.2.1.2 TCGA I . . . . . . . . . . . . . . . . . . . . . . . . . 51 5.2.1.3 The link between TGCA I and land grabbing . . . . 54 5.2.1.4 TCGA I and deforestation . . . . . . . . . . . . . . . 56 5.2.1.5 TIAA/Nuveen steps and strategies to tackle environ- mental and social issues . . . . . . . . . . . . . . . . 56 5.2.1.6 The role of the National Swedish Pension Fund AP2 58 5.2.1.7 Disentangling TIAA/Nuveen from land grabbing . . 59 5.2.2 University Endowments . . . . . . . . . . . . . . . . . . . . . 60 5.2.3 Harvard University Endowment in Brazil . . . . . . . . . . . . 60 5.2.3.1 Harvard’s link to land grabbing and deforestation . . 61 6 Analysis 63 6.1 Financial actors along the chain . . . . . . . . . . . . . . . . . . . . . 63 6.1.1 Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 6.1.1.1 Agricultural land creation: from pension funds to farmland . . . . . . . . . . . . . . . . . . . . . . . . 66 6.1.1.2 Production, trading and processing . . . . . . . . . . 69 6.1.2 Ways of approaching sustainability . . . . . . . . . . . . . . . 69 6.1.2.1 Main differences observed along the chain . . . . . . 70 6.1.2.2 Wealth and sustainability. For who? . . . . . . . . . 70 6.1.2.3 A new meaning of fiduciary duty is needed . . . . . . 72 6.2 Zooming on agricultural funds’ investments . . . . . . . . . . . . . . . 73 6.3 The methodological approaches . . . . . . . . . . . . . . . . . . . . . 75 x Contents 7 Discussion 77 8 Conclusion 79 8.0.1 Further research . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Bibliography 81 xi Contents xii List of Figures 1.1 Products derived from soy (WWF, 2014) . . . . . . . . . . . . . . . . 3 1.2 Frontiers of soy expansion in Brazil, 2010-2016, (West et al., 2018) . . 5 1.3 Proportion of remaining vegetation and agricultural land in the Cer- rado and Amazon (CDP, 2018) . . . . . . . . . . . . . . . . . . . . . 6 1.4 Total deforested area in the Amazon, comparison between 2017-2018, (Boadle, 2019) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.5 The Munduruku are an example of indigenous community that has lived in the Sawré Muybu in the heart of the Amazon, for generations. In addition to preserving their way of life, the demarcation of Sawré Muybu ensures the conservation of 178,000 hectares of Amazonian rainforest. Markus Mauthe / Greenpeace 2016 . . . . . . . . . . . . . 8 2.1 Percentage of soy exported under zero-deforestation commitments (ZDCs) in the Amazon and the Cerrado in 2015 (Trase data, 2017) . 16 3.1 Framework of the study . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2 Visual representation of the method used in the report . . . . . . . . 20 4.1 Product chain of soy. Each box represent the technical steps along the soybean product chain. The arrows between each step give ap- proximate proportions of the soybean flows. The two circles between agricultural land creation and agricultural production represent the area (ha) used respectively in the Cerrado and Amazon biome for soy- bean cultivation. The blue box surrounding the two arrows represent the total production coming from both biomes. . . . . . . . . . . . . 23 4.2 Soybean flows from the two biomes to countries of consumption (Trase data 2017) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4.3 End-point land transformation impacts ($) for one tonne of soybean produced in 2010 in the Amazon and Cerrado biomes of Mato Grosso, Brazil, considering direct conversion of natural vegetation to soybean (Amazon and Cerrado), and a pasture transition (Pasture/Amazon, Pasture/Cerrado) (Lathuilliere et al, 2017) . . . . . . . . . . . . . . . 26 4.4 Soy expansion, 2005-2016 (West et al., 2018) . . . . . . . . . . . . . . 27 4.5 Trends in soy production in Brazil, Argentina and Paraguay (West et al., 2018) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.6 GHG sources in soybean production in Mato Grosso (Raucci et al., 2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 xiii List of Figures 4.7 Average soybean yield between 1990 and 2011, and yield, area and total production in the 2010/11 growing season in the main Brazilian production states (Santhelas et al., 2015) . . . . . . . . . . . . . . . . 30 4.8 Average ESG policies of soy producers compared to their investors’ scores. (Levicharova et al., 2018) . . . . . . . . . . . . . . . . . . . . 32 4.9 Soybean productivity from 1976 to 2016 (Data from: Network for social justice and human rights, 2018) . . . . . . . . . . . . . . . . . . 33 4.10 Major soy traders and destinations (Kuepper et al., 2017. Data from Panjiva) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 4.11 China’s growing soybean consumption and import demands (Brown- Lima et al., 2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.1 Deforestation on the borders between Parà, Tocantis and Maranhão . 45 5.2 Deforestation in Rondônia . . . . . . . . . . . . . . . . . . . . . . . . 46 5.3 Deforestation along the BR-163 Highway, linking Mato Grosso and Parà . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 5.4 TIAA-CREF farmland investments in Brazil. In blue: foreign enti- ties; In red: Brazilian entities (Data from: GRAIN 2015; TIAA, 2017; Fian international 2018) . . . . . . . . . . . . . . . . . . . . . . . . . 50 5.5 TIAA’s Brazilian farmland holdings. (Source: TIAA, 2016) . . . . . . 51 5.6 Ludmila Farm in Pauì (Nuveen Farmland Map, 2018) . . . . . . . . . 52 5.7 Deforestation in Santa Filomena, Pauì, between 2001 and 2018, where the area of the Ludmila farm is located. In pink are highlighted the areas with a loss of forest cover higher than 30%. Spots in light green represent forest cover higher than 75% (globalforestwatch.org) . . . . 57 5.8 Harvard’s corporate structure in Brazil and its link to three local operators: Caracol, Insolo and BGE (GRAIN, 2018) . . . . . . . . . . 61 6.1 Illustrative framework of the actor system along the first three steps of the product chain. . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 6.2 Interactions from pension funds to farmland . . . . . . . . . . . . . . 66 6.3 Main differences along the chain . . . . . . . . . . . . . . . . . . . . . 70 6.4 Simplified scheme of farmland investment structures. Institutional investors are marked in red because of their substantial ownership shares in agricultural funds. It is important to note that accord- ing to the investment entity the influence of other types of investors vary, and other types of investors may become more relevant. Source: Ouma, 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.5 One step process, from biosphere to remote actors . . . . . . . . . . . 75 xiv List of Tables 5.1 TCGAI ownership shares by different pension funds (Source: Fian international, 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 xv List of Tables xvi 1 Introduction Brazil is home to two of the most important biomes on Earth, the Amazon and the Cerrado. Observing them from above, they show very different landscapes, pass- ing from the homogeneous, dark green-colored vegetation of the Amazon’s tropical rainforest in the North of the country, to a more arid and rugged-looking Cerrado’s savannah as we look more South and to the West. What they have in common is the importance of their vegetation for climate stability, biodiversity, and other ecosystem services, of which we all, globally, can benefit. However, as shown by the cover image of this thesis project (NASA), for the past 30 years or so (Kaimowitz and Smith, 2001), they have both been subjects of extensive deforestation. The landscape has rapidly transformed to a mosaic of swathes of cleared vegetation, now used for agriculture or cattle grazing to serve global commodity markets. More and more resources are necessary for a continuously increasing global pop- ulation that needs to be fed. Findings means and spaces for producing enough food necessary to sustain all of us, has become one of the greatest challenges of our time. So far, this has come at the expense of native vegetation in different parts of the world, including important biomes such as the Amazon and the Cerrado. In addition, the agricultural sector, such as the soybean sector, has branched out dif- ferent supply chains, from food to animal feed and cosmetics. This has undoubtedly contributed to the clearing of land for extensive monocultures. Thus, deforestation is affecting climate and environmental resilience through increased greenhouse gas emissions, loss of biodiversity and mineralization of soils. When looking at a product chain, such as that of the soybean, the system built around it is extremely complex. The product flows all around the globe in many different forms, passing through a range of technical processes, involving different stakeholders and other types of actors, with many social and environmental impacts. All these interactions and their development are both difficult to track and assess. When it comes to soybean or other agricultural commodities, the most visible conse- quence is exactly that mosaic of swathes of agricultural fields that once was a forest or a savannah. For more sustainable management and governance of commodity chains, minimizing their social and environmental impacts, there is a need for understanding how com- plex systems around them work. Many specialized studies can become too narrow or provide a too fragmented picture of how environmental, social, political, financial aspects interact to create the current situation. A more comprehensive picture of 1 1. Introduction product chains, build on all these different angles that have been specifically ex- plored, might give new and interesting insights and lead to more environmentally, socially and economically sustainable solutions. The soybean constitutes an interesting case since its growing importance on the global market, implying growing sustainability challenges. Greater sustainability challenges in countries such as Brazil, where soybean production compete with the protection of important biomes. Soy expansion in this country threatens not only the existence of vast vegetation areas but also the lives of all indigenous commu- nities that have always lived there. In addition, the current political situation is threatening areas that so far have been environmentally and socially protected, to favor the agribusiness. In order to paint a picture of the increasing importance for sustainable production and consumption of soybean, some background is given on market development, ecological situation, and logistics. 1.1 The rise of soy Soy is a high protein and energy agricultural product which can be used in dif- ferent ways: as feed, food and fuel. Around three-quarters of soybean production is destined to animal feed, due to the rise in global meat production and the de- cline of its relative cost. Between 1967 and 2007 pork production rose by 294 per cent, egg production by 353 per cent and poultry meat by 711 per cent (McLeod, 2011), with consumption of soy and industrial farming being crucial elements of these trends. China, the most populous country in the world supporting over 20 per cent of world’s population, has also become the world’s largest pork producer and consumer, therefor the largest soy importer. Giving some proportions, China imports of soy are three times the amount of EU and 12 times that of Japan, but this change has come only in the last twenty years (Brown-Lima et al., 2010) Only 6 per cent of soy produced is used directly as food. Whole beans can be eaten as vegetables or further processed and incorporated into other food products, such as soy flours and lecithin, a protein additive and emulsifier. Soy can also be used in baked products as margarine, in fried products and as cooking oil. Two per cent of total soy production is converted into biodiesel, which is predicted will con- tinue driving demand for soy, with a sharp rise in production by 2025 (WWF, 2014). As shown in Figure 1.1, soy plays an important role in many different supply chains and its demand has exponentially grown in the last decades. In the last 50 years, the production of soy has grown tenfold, from 27 to 269 million tons, with an extensive land use equivalent to the area of France, Germany, Belgium and the Netherlands (WWF,2014). With an increasing world population demanding for more protein sources, soy is especially important as agricultural crop producing more protein per hectare than any other major crop. However, this comes at a cost. The increased demand for soy requires the direct or indirect conversion of forest, grassland and savannah to agriculture, with enormous environmental impacts. In fact, the fastest growth in recent years occurred in South America, home to the biggest and most 2 1. Introduction Figure 1.1: Products derived from soy (WWF, 2014) important biomes on Earth, which have been threatened by conversion to agricul- tural land at a tremendous pace. In South America, soy production grew by 123 per cent between 1996 and 2004, and this growth is not predicted to stop: the Food and Agriculture Organization of the United Nations (FAO, 2007) suggests soy production will almost double by 2050 (WWF, 2014). 1.1.1 Soy production in Brazil The first big increase of soy production in South America started in 1973 after US imposed an embargo on soy exports (Faminow and Hillman, 1987), worried about the repercussions of a rising overseas demand over its domestic feed supply chain. This event lead South American governments to increase soy production. They supported the expansion of agricultural frontiers through subsidized credit, invest- ments in infrastructures and agricultural research, technical assistance to producers and price support. These politics made soy production increase twelve-fold between 1970 and 1980 in Brazil, Argentina and Paraguay. In the next decades, this trend continued to be sustained by the increasing global demand for animal feed. The expansion took place mainly in northeastern Argentina, southern Brazil and eastern Paraguay, where conditions were most favorable to soy varieties introduced from the US (West et al., 2018). What lead soy production to move toward tropical regions, which traditionally were not suitable for this type of crop, were substantial government investments which quickly pushed agronomic improvements, expanding soy production northwards, reaching the Amazon and the Cerrado (Kaimowitz and Smith, 2001). In Brazil, the great boom in soy expansion was driven by both the US embargo in 1973 and by the need to generate currency to pay for imports such as petroleum. At the time, Japan provided technical assistance to increase soybean production on marginal frontier land (Brown-Lima et al., 2010). Until the 1980s Brazilian soy- bean production was concentrated in the traditional farming regions in the south of 3 1. Introduction the country, including São Paulo, Santa Catarina, Paranà and Rio Grande do Sul (Kaimowitz and Smith, 2001). In the following years soybean production continued to grow and the Brazilian government started investing in developing new soybean varieties as well as different planting techniques. In 1997, the agricultural expansion reached the Cerrado and the Amazon. While in 2000 the top exporters of soy were still the traditional regions in the south of the country, this situation drastically changed in 2009. EU tripled its imports of soy and China’s imports increased so much to to just dwarf those of EU. China became Brazilian soy first importer and since its sourcing regions are mainly Mato Grosso, Mato Grosso do Sul, Bahia and Goias, these states became top producers and exporters of soy (Brown-Lima et al., 2010). These huge quantities of soy ex- ported have required an extensive work in agricultural land conversion from native forest, savannah, grassland or pasture. At the same time, US could not bring that much new land into agricultural production, due to the competition of soy with other crops such as corn. This implies that Brazil is now competing with US for the position as first global soy producer, both converting land, increasing yields and replacing other crops with soybeans. For this reason, the Brazilian soy production system has received pressures from conservation NGOs (Gibbs et al., 2015), the media and consumers, being concerned about high deforestation rates and eager to a more responsible sourcing of soy. However, most of the concern has come from European more than Chinese buyers (Brown-Lima et al., 2010). 1.2 Soy deforestation in Brazil Increasing soy production is associated with deforestation. In order to create agri- cultural land has occurred and still occurs in different Brazilian states. In the last ten years, most soy deforestation in Brazil has occurred in the transition area be- tween the Amazon and the Cerrado in Mato Grosso state, and in the Matopiba region of the Cerrado, defined as the world’s fastest growing soy frontier. In this period, Chinese imports have been associated with much of the soy deforestation. The biome that has been mostly affected by soy deforestation has been the Cerrado, while just for a minor extent the Amazon biome, mainly due to stricter regulations imposed by the Brazilian Forest Code. As showed in Figure 1.2, the Amazon biome only have some areas with low direct soy expansion, while Cerrado has been affected more or less everywhere. The whole Matopiba region has been subject to high soy expansion at a very high rate, as well as some areas in the Mato Grosso State. 1.2.1 The Brazilian Amazon The Amazon rainforest has been highlighted as one of the fundamental biomes able to influence global climate stability due to its importance in the global carbon cycle. In the Earth System, the carbon cycle acts as a buffer to maintain the planetary 4 1. Introduction Figure 1.2: Frontiers of soy expansion in Brazil, 2010-2016, (West et al., 2018) environment within well-defined limits (Steffen, 2006), which ensure stability to the System itself. It stores between 135 to 180 billion tons of carbon on its extension of 5.5 billion of square meters (Gaffney et al., 2018), and its contribution to evapo- transpiration has a high influence on rainfall. It is also the largest and most diverse tropical rainforest, home to one in every 10 species on Earth. One-fifth of the Ama- zon has already been cleared from native vegetation and during 2000-10 around 3.6 million hectares of forest were lost yearly (WWF, 2014). Using the Brazilian gov- ernment deforestation monitoring data from PRODES, an estimated 1.8 million ha of soy in 2016 were under native vegetation in the year 2000 (West et al., 2018). Together with other types of crops and cattle ranching, soy undoubtedly played a major role in deforestation of this area. 1.2.2 The Cerrado The Cerrado, which lies mostly in Brazil, is recognized as the world’s most biodi- verse savannah (CDP, 2018). It’s a vast land that includes dry grassland, woodland, forests and wetlands, holding 5% of the world’s biodiversity, and of its more than 11.000 plant species, nearly one half are found nowhere else on Earth. Its trees may seem smaller than the ones in the Amazon, but they have a deep root system which holds 70% of the biomass underground. In fact, this "upside-down forest" is able to store about 265 tonnes of carbon per hectare (Castro and Kauffman, 1998). This important ecosystem is now threatened by extensive land conversion. Estimates on how much native land remains intact are diverse, with the Brazilian government stating that a 53% still remains untouched (MMA, 2010), while other studies assessed that the percentage might be as low as 35% (Klink and Machado, 5 1. Introduction 2005; Durigan et al., 2007) In the Cerrado, soy accounts for 90% of its cultivated crops and, in 2013/2015, 52% of Brazilian soy was produced in this area. From 2000 to 2014 its agricultural area grew by 87% (Filho and Costa, 2016), with a conversion to soy mainly from pastures in Mato Grosso and Goias and mainly from native vegetation in the region of Matopiba, northeast of the Cerrado, where the states of Maranhao, Tocantis, Paui and Bahia are merged. Considering the Cerrado’s aptitude for modern agriculture and its capacity to feed the world, a balance between conservation and agricultural expansion has become one of the biggest challenges for this area. Within the Cerrado, the Matopiba region is considered one of the most threat- ened regions by soy expansion and conversion from native vegetation in the near future. In Matopiba only, from 2000 to 2017, soy expanded by 310% (West et al., 2018) at very high deforestation rates, and there are different signals indicating that soy-associated deforestation will increase in the coming years. Firstly, there is still a large availability of native vegetation which could be potentially be cleared for new agricultural land; secondly, a large amount of investments in soy infrastructure such as crushing facilities and soy storage are ongoing in the region. The combination of these two factors can be considered as preconditions for deforestation because of the lack of legal protection of the area, or at least much lower than the protection guaranteed to the Amazon region by the Forest Code (Vieira et al., 2018). 1.2.3 The political threat to the Amazon, the Cerrado and to their indigenous communities Figure 1.3: Proportion of remaining vegetation and agricultural land in the Cer- rado and Amazon (CDP, 2018) Agricultural land conversion has led to impacts on both the Amazon and Cerrado areas. As shown in Figure 1.3 today only about 55% of the Cerrado biome remains intact, while the Amazon, thanks to stricter Legal Reserve Requirements from the Forest Code, still holds 82% of its native vegetation (CDP, 2018). Furthermore, 6 1. Introduction since 2000, the deforestation rate in the Cerrado has been 2,4 times higher com- pared to that of the Amazon. Indigenous land has been linked to healthier forests and lower carbon dioxide emis- sions from deforestation and forest degradation (Stevens et al., 2014). It accounts about 13% of Brazil territory, corresponding to 117 billion ha of land. This land is home to about 850,000 people organized in 300 tribes (Boadle, 2019). At the moment there are 462 indigenous lands officially registered (Gonzales, 2018). Unfortunately, the data recorded in 2018 seem to increasingly threaten the Amazon and indigenous lands. The number of protected areas is probably going to change by the switch of responsibility from FUNAI to the Ministry of Agriculture on organizing these reserves. Such change of responsibility puts a lot of risks not only to the lives of thousands of tribal people, but also to the rich ecosystems they have always been living in balance with. In addition, according to data from the Instituto Nacional de Pesquisas Espaciais (INPE), during the presidential election season between Au- gust and October 2018, there were 8,000 cases of deforestation alerts in the Amazon rainforest, and more than 1,690 square kilometers of forest were cleared in just three months. Compared to the previous year, deforestation rate in 2018 was 48.8 percent higher than the same period. Rondônia has been the third-most deforested state in the Amazon, making up to 20 percent of the total deforestation during that period. Bolsonaro’s electoral speeches have indeed started a process of legitimization of land grabbing. His political project has given hope to all those land grabbers aiming to increase land deals with big companies, usually backed by foreign investors. Figure 1.4: Total deforested area in the Amazon, comparison between 2017-2018, (Boadle, 2019) Bolsonaro presidential election’s campaign has been based on a propaganda in favour of large infrastructural projects and commercial development by opening up more native territory. His words in 2017 in Mato Grosso: "If I become president, there 7 1. Introduction won’t be one square centimeter of land designated for indigenous reservations” or "[...]I’m not getting into this nonsense of defending land for Indians" (Watson, 2018) have led dozens of men to enter protected indigenous land claiming for their stakes. Armed with machetes, chainsaws and firearms, they threat indigenous people to set fire to their villages or use violence to force them out of their land. Bolsonaro’s election has further worsened the situation and such type of attacks have increased 150 percent since he was elected in late October, according to the Indigenous Mis- sionary Council (CIMI), a Brazilian advocacy group. "The number of invasions and attacks on indigenous reservations rose and deforestation skyrocketed nearly 50 per- cent during Bolsonaro’s presidential campaign. Bolsonaro has railed against what he sees as excessive federal protections for indigenous people" threatening also the ecosystems which these people have always sustained and protected. (Boadle, 2019). According to him, the excessive federal protections to these communities represent an impediment to agribusiness, and he aims to weaken these protections in their favour. To fulfill this aim, one of his first political actions as president has been to take FUNAI’s (Fundação Nacional do Índio - Brazil’s Indian Affairs Department) authority to set reservation boundaries to give it to the Ministry of Agriculture, which is dominated by rural interests. In charge of dealing with indigenous land is- sues is Nabhan Garcia, a right-wing farming organizer who has always been fighting against reservations (Boadle, 2019). Although these latest events are particularly Figure 1.5: The Munduruku are an example of indigenous community that has lived in the Sawré Muybu in the heart of the Amazon, for generations. In addition to preserving their way of life, the demarcation of Sawré Muybu ensures the conser- vation of 178,000 hectares of Amazonian rainforest. Markus Mauthe / Greenpeace 2016 dangerous for tribal people and the environment they live in and protect, it is also important to note that land grabbing has always been a problem in the country. The Constitution adopted by Brazil in 1988, establishing the end of the dictatorial period, recognizes the respect for the cultural identity of indigenous people and the rights they have on the lands they live in. But the spread of land grabbing and 8 1. Introduction illegal logging demonstrate that the state protects the rights of indigenous peoples only theoretically (Rylands and Brandon, 2005). Much more often, in the case of disputed territories, it is the oligarchies of the agricultural sector that prevail. In addition to indirectly allow this type of process - thanks to a "bland" legislation - the Brazilian administrations have for decades also financed an invasive infrastructure and mining sector (Network for social justice and human rights, 2018). Other risks threatening tribal people are also coming from the new Minister of In- frastructure Tarcísio Gomes de Freitas, considered one of President Jair Bolsonaro’s most capable ministers, which aims at creating "a second revolution in Brazilian agribusiness" (Branford, 2019). His strategy includes seeking foreign investors, to push forward with new roads and railways which could potentially open the Amazon and the Cerrado biomes to land grabbers, illegal loggers, illicit ranchers and indus- trial agribusiness (Branford, 2019). Two important infrastructural examples that would affect the biomes are the Ferrogrâo (Grainrail) and FIOL (the Railway for the Integration of the Center-West). Infrastructural projects such as railroads, high- ways and port have always been fundamental for the economic development of the region, but at the same time have always been related to an increase in deforestation. Studies have shown that more than 70 percent of deforestation occurs within 50 km of paved roads, while at most 7 percent occurs along unpaved roads (IPAM, 2000). Moreover, infrastructure investments usually create strong connections between big agricultural traders and specific regions of production, which gain more control over such regions, shaping development trajectories and the sustainability of agriculture, including soy (West et al., 2018). What happens locally in Brazil, however, is not only determined by local dynamics. The political plans mentioned above, can’t be implemented without capital, which is channeled here from different countries. Foreign capital invested in farmland or in agricultural companies with unsustainable practices, trigger local dynamics with potential impacts on the environment and local populations. The main purpose of land grabbers, is to sell the land to big companies or big investors. Land grabbers’ actions are induced by the concrete possibility of finding a market also for land acquired clearing native vegetation or forcing populations to leave their homeland. A "revolution of Brazilian agribusiness" is also a possible option because there are international economic interests in such plans. 1.3 The need for a bigger picture of interests The network and the interests surrounding agricultural production as that of soy are complex and diverse. This makes sustainable management of global product chains difficult, especially since a lot of the literature is specialized, providing a fragmented view of these complex systems. There is, therefore, a need for more comprehensive studies integrating material, social, environmental, political and financial aspects in order to find sustainable, coordinated solutions. In addition, global product chains are shaped by global interests. This interest 9 1. Introduction is expressed in channeling capital in areas of production, with possible environmen- tal and social repercussions. Financial actors, given their economic power, are the driving force of product chains. What is known about them, is that they have a strong link with local dynamics around product chains. However, the type of influ- ence that they have, their interactions with other different actors and their link with environmental and social issues have seldom been elaborated in literature. Hence, it is necessary to better understand their influence on global product chains in order to identify the positive and negative aspects of it. As Brazil is a resource-rich coun- try from which different countries benefit, we need to find ways to have a "remote control" with sustainable outcomes. 1.4 Aim of the project This project is an attempt to make a comprehensive description and model of the soybean product chain. Such a comprehensive model should look at the technical processes, at the local and remote actors involved in each process and at their in- teractions. In particular, the focus on the relationship between financial actors and farmland, aims at re-materialize concepts such as capital, resources and value into practical activities, giving them a context, the soybean market in Brazil. In fact, financial capital is often imagined as an abstract entity, which "circulates around the globe as a function of its profit-seeking imperative, impacting on households, communities, companies, regions, and ecosystems" (Ouma, 2016). However, as the impacts are far from abstract, so is financial capital and whoever directs it. The purpose of such model is to provide support to sustainable management and governance of product chains. Looking at the product chain of Brazilian soybean from agricultural land creation to retail and consumption, the report will try to answer to the following research questions: With a perspective on more steps of the product chain (agricultural land creation, production and trading): • What are the differences between financial actors at these different steps of the chain? Do they have different ways of interactions or different ways of approaching sustainability? Zooming on the product chain’s step linked directly to high environmental and social impacts (agricultural land creation): • How financial actors operate on land? • What is their relationship with local actors? Can this relationship be improved to reduce deforestation and land grabbing? Moreover, since this type of comprehensive and interdisciplinary approaches are still not very diffused, a reflection will be done on the methodology used in this report, answering these other research questions: • How the understanding of the relationship between actors in the wealth chain and the product chain can complement LCA studies? • What are the differences when looking at the whole product chain compared to focusing on only one step? 10 1. Introduction 1.5 Limitations This thesis project has looked at different steps along the product chain and have been generally described from agricultural land creation to retail and consumption. However, the analysis has been focused on agricultural land creation, production and trading, being these three steps placed in Brazil and their direct relation to de- forestation and local social problems. Moreover, thanks to the abundance of studies on foreign investors in the first step of agricultural land creation, a special focus has been given to those. Consequently, for the first step it was possible to better reconstruct the relationships between investors at different levels (from those closer to the product flow to those more remote and "indirect"), while for the production and trading steps it was only possible to study the interactions between local actors and investors directly linked to them. Hence, it is possible that the analysis of the influence of some financial actors in the first step compared to the others may be biased, given the greater amount of data available for this phase. In the production and trading phase, the relation between investors closer to local actors (more often asset managers) with more remote ones (different asset managers’ beneficiaries) and their influence on the product chain, is left for further research. 11 1. Introduction 12 2 The state of knowledge on the link between finance and farmland Financial actors and capital are often considered as abstract entities. This section aims at explaining their concrete influence in the Brazilian context. Therefor, finan- cial actors’ relation to farmland and the soybean product chain will be described. 2.1 The role of financial actors in Brazilian farm- land and soy production Investors’ interest in farmland has increased after the world financial crisis of 2007/2008. This growing interest has lead to a process called "financialization", the transforma- tion of land into a "materialized financial asset" (Fian International, 2018), as a consequence of the power and influence of global finance. Financial actors chan- nel capital into land purchases and land-based activities aiming to diversify their portfolios, increasing returns and lowering risks. They saw in farmland a good op- portunity of investment due to the increasing demand for food and the growing per-capita meat consumption requiring an increase in production of protein-rich feed such as soy (Steinweg et al. 2018). Financial actors operate in different ways with different implications on a local scale. Two principal ways by which capital is channeled into Brazilian land purchases are (Van Gelder et al., 2002): • Debt, by banks and credit agencies; • Equity, by insurance companies, pension funds, mutual funds, university en- dowments, investment firms, private investors. In addition to these more traditional ways of investing, investors have increasingly turned to a strategy of direct farmland purchases, building complex structures of local subsidiaries to directly control the land. Pension firms, endowments and pen- sion funds are among the leading actors involved in this process of financialization and creation of "global wealth chains" (Seabrooke and Wigan, 2014), but also the main actors indirectly causing episodes of land grabbing (Fian international, 2018). Global Wealth Chains (GWCs) are defined as connected forms of capital aiming at creating wealth through the construction of opaque corporate structures. The re- location of wealth operated by the GWCs usually results in braking loose from the location of value creation, increasing inequalities (Seabrooke and Wigan, 2014) and environmental problems, such as deforestation and its consequences in the specific case of Brazil (Steinweg et al., 2018). An example of GWCs are the net of interna- 13 2. The state of knowledge on the link between finance and farmland tional financial actors around farmland investments achieved by complex corporate structures and ownership chains. In Brazil, the interest in farmland has been focused particularly in the Cerrado, because of its geographical position, ecological characteristics, and the reduced re- quirements of Brazil Forest Code in this area, a regulatory framework for land use and environmental conservation on rural properties, with the aim of protecting na- tive vegetation. According to the Forest Code, all rural landowners must maintain a proportion of their land as Legal Reserve (LR), and while in the Amazon biome the proportion is 80%, in the Cerrado is only 20% (Rogerson and Døvre, 2018). An- other important and investment-attractive region within the Cerrado biome, now considered the new soybean frontier (GRAIN and Rede Social de Justiça e Direitos Humanos, 2018), is the Matopiba region. Here, institutional investors together with agribusiness have adopted business models based on the acquisition of land, clear- ing it from its native vegetation and transforming it into farmland (Steinweg et al., 2017). The Brazilian State has also played an important role in facilitating these processes, helping agribusiness expansion through significant subsidies. Just con- sidering soy monocultures, they have started to expand into Matopiba from 2000 and since then they have expanded continuously, due to the need of new area of investments by global finance. The drop of commodity prices on the world market after the financial crisis of 2007/2008 and the parallel rise of land prices in Brazil, led to extensive land spec- ulation and the creation of "land-companies", more interested in acquiring, selling, leasing and managing land, than in actual agricultural production (Fian interna- tional, 2018). 2.1.1 Land grabbing and financialization of land Land speculation and large-scale farming nourished by financial actors’ need of rev- enue streams have resulted in significant environmental (Steinweg et al., 2018) and social impacts (GRAIN and Rede Social de Justiça e Direitos Humanos, 2015). In fact, both phenomena are often linked to the process of "grilagem", a particular form of land grabbing. A lot of new farms are created from land formally owned by the state where local communities have lived for generations, using the land for hunting, harvesting fruit and collecting fire-wood (GRAIN and Rede Social de Justiça e Di- reitos Humanos, 2015). Land grabbers, more commonly called "grileiros", are local criminals which often use violence and intimidation to displace these communities from their land, putting fences to prevent them from accessing the land. Police and local governments are usually corrupted and involved in the land speculation process and rarely defend the people forced to leave their land. Grileiros mainly act with the hope of selling the land to potential interested actors such as big companies or investors. Thus, once the they have fenced and illegally acquired the land, they falsify land titles to legitimise the illegal land occupation (Fearnside, 2008). The next step for them is to sell the land to big companies which are most of the times connected to international investors (GRAIN and Rede Social de Justiça e Direitos 14 2. The state of knowledge on the link between finance and farmland Humanos, 2015). Channeling enormous flows of money in these areas of Brazil, investors have thus, indirectly and unconsciously, fuelled this process. 2.1.2 Foreign investors in Brazil and their link to deforesta- tion Using Land Matrix, an open access platform sharing information about land deals, it is possible to observe that there have been more than 1,6 million hectares of land transferred from Brazil to foreign investors since the year 2000. When looking specifically at soy, land deals involve 447,566 hectares, almost 30% of all land deals with foreign investors. From the Chain Reaction Research (Steinweg et al. 2018) on foreign farmland investors in Brazil, it is reported that 10 foreign investors hold 1.5 million hectares of Brazilian farmland, with the Cerrado and especially the Matopiba region, being "at the heart of farmland investment growth". Only eight companies backed by foreign investors control 868,488 ha of farmland. In these foreign-held farms there have been 423,242 ha of deforestation since 2000, even though defor- estation rates have decreased significantly in the last six years (Steinweg et al. 2018). In 2010, the Brazilian government tried to contain this massive land acquisitions by foreign investors delivering a law, Brazil’s Attorney General, which stipulated that foreign investors couldn’t own more than 25% of rural land in any munici- pality, and investors of the same nationalities could own a maximum of 10% of a municipality’s lands. However, some foreign investors have find ways to bypass the legislation, creating "opaque structure" to make their investments appear more Brazilian than they are (GRAIN and Rede Social de Justiça e Direitos Humanos, 2015). Even though financial institutions are not physically based or operating in Brazil, they are subject to potential financial downsides and business risks that might com- promise the profitability of their portfolio. Related to soy, but also to other types of crops and farmland investments, some of the major risks identified for companies and investors are (CDP, 2018): • Operational risks: if soy production involves deforestation or is not sustainably produced, this has direct and indirect impacts on ecosystem services, which can lead to lower productivity and higher production costs; • Market risks: investors or companies without responsible environmental poli- cies may suffer from a reduced market appetite for deforestation-related soy and other commodities, and not being able to access markets for deforestation- free products. • Reputational risks: increasing global awareness about climate change, defor- estation and environmental issues in general, has simultaneously increased public attention to such problems, putting pressure on companies linked to deforestation and other social and environmental risks in the Cerrado and the Amazon; • Regulatory risks: as international pressure to act on climate change and to halt deforestation increases, regulations are likely to change, posing risks to 15 2. The state of knowledge on the link between finance and farmland companies and investors which haven’t started a path toward more responsible and more sustainable sourcing yet; A recent study, (Galaz et.al 2018), which have started from market-based infor- mation about the soybean sector in Brazil, have identified big soybean companies according to their market share and then linked them to a small set of financial actors. The companies identified starting from their share on the market are mainly responsible for processing and trading, and export soybean to consumer countries. Among these companies, many of them are committed to zero-deforestation policies as part of their sustainability strategy and are part of the Soy Moratorium, a zero- deforestation agreement to protect the Brazilian Amazon, and also among Cerrado Manifesto Statement of Support’s signatories, to demonstrate their committment to the production of more sustainable soy from the Cerrado. However when it comes to financial actors, only one big foreign investor in Brazilian farmland, Nuveen, holding almost 300,000 hectares of land in Brazil (Nuveen, 2018) and 1/3 of it dedicated to grains, has drafted a zero-deforestation policy last year (Steinweg et al. 2018) valid both in the Amazon and in the Cerrado biomes (Nuveen, 2018). Hence, looking at Figure 2.1 showing the total amount of soy that is exported from Brazil under zero- deforestation commitments, it is possible to observe that it amounts to only 44% of soy produced in Cerrado, and 46% of the soy from the Matopiba region (CDP, 2018). Regarding the Amazon, thanks to the stricter regulations imposed by the Forest Code, the percentage of soy exported under zero-deforestation commitments is much higher, 93%. Given the current climate crisis and the importance of the Cerrado biome in preserving fundamental ecosystem services, improvements for in- cluding zero-deforestation commitments in this area are needed, both by companies and big investors. Figure 2.1: Percentage of soy exported under zero-deforestation commitments (ZDCs) in the Amazon and the Cerrado in 2015 (Trase data, 2017) 16 3 Method 3.1 Framework: IE and PCO The method used in this report is built on several ways to look at product chains. The study of the interconnections between different types of actors along product chains and their interaction with the biosphere is an Industrial Ecology’s peculiar feature. For this reason, different approaches commonly used within this field have been mixed together to provide novel insight to product chains and to the role that some specific actors, such as financial actors, not ordinarily considered in this type of analyses, can have on the chain. Figure 6.1 visually represents the framework used for this study, where the prod- uct flow represents the physical flow of a product or material along the chain, from resources extraction to disposal. It’s the backbone of the whole system, and it has interactions with both physical and human actors. A product flow relies on the resources provided by the biosphere inevitably creating impacts, whether in the form of extraction of material or in releasing pollutant emissions. These technical processes are usually covered by Life Cycle Analysis (LCA), which models the flow looking at environmental information along the chain. However, the product doesn’t flow by itself. Figure 3.1: Framework of the study A number of actors such as people working in factories, companies, organizations, 17 3. Method governments, are responsible for shaping the product flow according to specific needs. In this study, the actors considered are divided into two classes: local actors, di- rectly involved in handling the product, such as laborers and companies managing and organizing the flow, in the country where each step of the chain is located; re- mote actors, such as organizations, investors, companies from different sectors and countries, that have an interest in the product or a relationship with local actors, but don’t directly put their hands on the flow. Despite their physical distance, such remote actors can have an important role in shaping the product chain, with social and environmental effects. The inclusion of the net of human actors when think- ing about product chains is a typical element of Product chain organization studies (PCO), which can take into account social and economic factors. 3.2 Design of the study This study combines the vision of technical processes related to the Brazilian soy- bean product chain, its ability to flow and its environmental implications typically considered in LCAs, with the organizational side covered by PCOs, where the net- work of human actors, through the work of people, companies, and organizations that allows the flow to exist and move forward, is also taken into account. In the following sections, both the methodologies are defined, explaining in which ways they have been used in this study. 3.2.1 The product chain from an LCA’s perspective Environmental factors and impacts within a product chain are considered when performing Life Cycle Analysis (LCA). With LCA products are followed from raw material extraction from natural resources to disposal, and in order to perform this type of studies, a lot of environmental information are collected and analyzed. The goal is to assess the amount of resources utilized throughout the life cycle of the prod- uct, but also to measure the outflows in terms of pollution, expressed with different types of environmental impact categories. An LCA is used to understand specific critical points in a product life cycle, whether is excessive use of natural resources or release of pollutant emissions in the environment for one single product, but also to compare environmental performances of products with the same function. The definition of system boundaries depends on the goal and scope of the study, which can include the whole life cycle of the product from "cradle to grave", or be focused on only some of the processes excluding such as, for example, the user and disposal phase. Geography is also included in the system boundaries, but usually, the time dimension is excluded. (Baumann and Tillman, 2004) In this report, the life cycle thinking has been initially used to map all different steps characterizing soybean product chain, from agricultural land creation to re- tailing and consumption. Life cycle findings on soybean - sourced in Brazil or in different parts of South America - have been used to understand and describe the technical processes involved in soybean cultivation, production and processing and to have information about the differences in environmental impacts among the product 18 3. Method chain’s steps, from agricultural land creation to export. The consultation of LCAs including land use change (LUC) impact assessment have been prioritized, given the importance of this process for the biomes considered in the study. However, since social and economic aspects are not included in LCA, only used as weighting fac- tors to compare environmental impacts among each other, another method to look at product chains has been combined to have a more comprehensive picture of the actor-network around the product chain, its relation with the technical processes and their impacts. 3.2.2 The influence of different types of actors: a PCO’s approach Behind every product or material chain, moving in different parts of the world, there is a complexity of actors that allow the product to be manufactured, handled, shipped, transported, transformed. The PCO methodology aims at showing the tangled relationship between "people and matter" (Baumann, 2012), a link which is often missing in current academic approaches. It consists in mainly two steps, using first a life cycle approach in drawing a basic life cycle of the product, and then mapping all the actors involved with it, studying how these actors are organized, how they interact with each other and with the product flow, reconstructing the network they create around it. Depending on the problem and on the scope of the study, PCOs can be focused on different aspects (Baumann, 2012). In this project, PCO has been used when tracking the actors involved in each step of the soybean product chain, previously identified with a life cycle approach. Specifi- cally, this study has mapped some of the most important actors operating in Brazil and more distant ones, especially focusing on financial actors, looking at how they relate with Brazilian organizations, companies, workers, its environment, and the soybean product flow. The focus on financial actors derives from the recent atten- tion that has been given to the connection between the Earth systems and finance (Galaz et. al 2015; Galaz et.al 2018). In this regard, the need for scholars to fur- ther investigate different financial actors’ link to social-ecological effects has been highlighted. Hence, a particular focus on data collection will be given on the steps on the chain where the relation among financial actors, local actors, and social and environmental impacts, has not been thoroughly described before. 3.2.2.1 Global Wealth Chains Within the study of financial actors as remote actors affecting the soybean product chain, this report will also look at the complex corporate structures of some big investors, such as pension funds and endowments, linked to the product chain step identified as more directly linked to environmental degradation in Brazil. In this regard, it has been acknowledged the need for having a clearer picture on GWCs’ im- pacts on developing countries, given the currently visible contradiction of the latter hosting GWCs as a developmental strategy, but on the contrary, being tremendously impacted by them. Pension funds and endowments are responsible for this process 19 3. Method (Seabrooke and Wigan, 2014), hence their corporate structure will be described more in depth. 3.3 Data collection, method, and analysis Figure 3.2: Visual representation of the method used in the report A literature study has first been conducted to have a better understanding of the characteristics of the soybean product chain. In this phase, LCA and LCIA on land use change, land occupation, and transformation for soybean production in Brazil have been used. The product chain with all its steps have been constructed, collecting technical information about processes and environmental impacts for the different steps. Then, the study has focused on identifying the actors involved. Once identified they have been placed in the step or steps of the chain where they were recog- nized having the most influence on. In addition, they have been differently classified and vertically placed on the framework if local actors, more directly involved with handling soybean, or remote actors, such as financial actors. Once collected the in- formation about the different steps and its actors, an overall picture of the product chain has been depicted, describing with a general view each step of the chain. Thereafter, a more detailed description of agricultural land creation has been given, choosing to zoom in this step given its importance for land clearing and its environ- mental and social implications. Here, the "intricate interplay" between local actors, such as farmers, laborers, ranchers, land grabbers, big soybean production compa- nies, and financial ones, such as pension funds and endowments, have been more thoroughly described. Of these financial actors, involved in GWCs, it is reported how their corporate structures is built, how they engage with their subsidiaries, what are the social impacts related to the presence of their companies on land and 20 3. Method the deforestation related to their land occupation. Always within financial actors in the agricultural land creation’s step, two inter- views have been conducted in order to have more information about one specific GWC operating in Brazil that has been linked to social and deforestation problems, build on a system of farmland investments by pension funds from different coun- tries. The first interview has been carried out with Nuveen, the asset manager of an agricultural fund owning extensive areas in Brazil. The second one has been carried out interviewing the Second Swedish National Pension Fund (AP2), investing in the Brazilian agricultural fund managed by Nuveen. The questions asked focused on their zero-deforestation policy, on how the asset manager-beneficiaries relationship influenced the drafting of the policy, and on future improvements planned for more sustainable investing. The purpose has been to understand the type of interactions that occur between asset manager-beneficiaries, and how the interactions could be related to impacts on land. Thereafter, the role of financial actors in different steps of the chains is analyzed, comparing financial actors interaction and approaches to sustainability in agricul- tural land creation, production and trading. Moreover, a more detailed analysis of the first step is to be given. The analysis also includes a reflection on the interdis- ciplinary method used, especially reflecting on the additional understanding that it can provide compared to more traditional approaches. Finally, results are discussed comparing them with previous related findings. 21 3. Method 22 4 The Soy Product Chain Organization This chapter will provide the reader with a general overview over the soybean prod- uct chain. All the steps are also presented, explaining their impacts and the principal actors identified. Since the importance of the first step of the chain in the deforesta- tion problem, its actors are not be presented here but deeply described in Chapter 5. 4.1 The product chain Figure 4.1: Product chain of soy. Each box represent the technical steps along the soybean product chain. The arrows between each step give approximate propor- tions of the soybean flows. The two circles between agricultural land creation and agricultural production represent the area (ha) used respectively in the Cerrado and Amazon biome for soybean cultivation. The blue box surrounding the two arrows represent the total production coming from both biomes. Many studies have looked at the soybean product chain to understand its steps and related environmental impacts. Dalgaard et al. (2008) and Castanheira and Freire 23 4. The Soy Product Chain Organization (2013) have looked at the whole product chain until transport to Europe, with Cas- tanheira and Freire (2013) taking into account land use change scenarios in their LCA; Lathulliere et al. (2017)’s LCA have focused on land use change impacts of soybean; West et al. (2018) have looked at soybean trade flows among different steps of the chain. Based on these study, it has been possible to build a schematic figure (Figure 4.1) on soybean product chain. The flowchart describes soybean product chain from agricultural land creation to retail and consumption, including approximate proportions of soybean volumes be- tween each step. The agricultural land used for soybean in the Cerrado is around twelve-fold the land used in the Amazon biome. The two arrows coming out of the agricultural production step give some proportions of the soybean produced re- spectively in each biome, of which a 6% is used as whole soybeans mainly for food production and directly sent to final processing; the remaining 94% is sent to crush- ing facilities to be converted into soy meal (about 79%) and crude oil (about 18%), while 3% of the input is left as waste. Once processed in the crushing plant, soy meal and oil are sent to final processing industries, mostly in the consumer countries. Whole soybeans are also an input of the transport process, since they are also sent to final processing industries but without previous treatment. The final processing step includes processing from different types of industries, hence the outflows are different types of products. To roughly estimate the different sizes of the outflows, approximate information for global productions have been used (WWF, 2014), since exact proportions for Brazilian soybean have not been found. The majority of the soybean produced is converted into feed (about 75%); 2% is used as biodiesel; 6% is used as food without any further processing. The remaining 20% is used for food oil and by the chemical industry for cosmetics or pharmaceuticals. The consumption of soybean produced in Brazil occurs in Brazil itself, but in major part is exported abroad. In Figure 4.2 it is possible to observe the soy flows sourced in the Cerrado and Amazon, which through the work of the soy traders reaches different countries in the world. Excluding domestic consumption, China is the first importer and the Netherlands appear to be the second one, accounting for more than one-fifth of the soy that arrives to EU (WWF, 2014). 4.1.1 Agricultural land creation Soy product chain starts with agricultural land creation. According to the World Bank database, in 2016 the percentage of agricultural land in Brazil was 33.9% of the total land, and of this 33.9% of land, soy is about 12%. Such percentages are necessary to support the enormous internal and particularly external demand. When looking at the whole life cycle of soybean (Castanheira and Freire, 2013), land use change appears to be a very crucial element within soybean greenhouse gas (GHG) balance, including also other different impacts. 24 4. The Soy Product Chain Organization Figure 4.2: Soybean flows from the two biomes to countries of consumption (Trase data 2017) 4.1.1.1 The process The agricultural land used for soy cultivation can be either created converting native vegetation (direct conversion), or converting pastures (indirect conversion). To cre- ate agricultural land, trees are usually removed by charcoal producers and the rest of the vegetation is collected by tractors and bulldozers and then burned (Mattson et al., 2000). To better understand how this process is carried out, social aspects are very important. Chapter 5 will provide more information about the actors involved and how they relate to driving this process. 4.1.1.2 Environmental implications Land use change due to conversion to soybean crops has multiple impacts, and they largely change according to the previous state of the land converted (Castanheira and Freire, 2013) Soybean is usually grown in monocultures, with high repercussion on biodiversity. Clearing diversified vegetation to replace it with monocultural crops in such biodiverse region such as the Cerrado or the Amazon, always carries the risk of removing a great number of species per hectare. Today, especially in the Cerrado, only a small part of the land is protected, and those protected areas are not even widely distributed, but grouped in areas not to interfere with the airplaine spread of pesticides (Ratter et al., 1997) A more evenly distribution of protected areas would have helped, to some extent, biodiversity protection, creating the possibility to an- imals and plants species to use habitat corridors, making them less vulnerable to land conversion. (Mattsson et al., 2000). 25 4. The Soy Product Chain Organization Lathuilliere et al., (2017) performed a LCA of soybean focusing on land transforma- tion and occupation impacts to biodiversity and ecosystem services in Mato Grosso, a region comprising both the Amazon and Cerrado biomes. What emerged from the Figure 4.3: End-point land transformation impacts ($) for one tonne of soybean produced in 2010 in the Amazon and Cerrado biomes of Mato Grosso, Brazil, con- sidering direct conversion of natural vegetation to soybean (Amazon and Cerrado), and a pasture transition (Pasture/Amazon, Pasture/Cerrado) (Lathuilliere et al, 2017) study, were lower impacts when converting soybean cultivation from pasture, com- pared to native vegetation. Clearing native vegetation from the Amazon resulted the most impactful choice, as showed in Figure 4.3. Another study (Castanheira and Freire, 2013), taking into account different land use prior to soybean cultivation and focusing on GHG emissions, also found that degraded grassland should preferably be used for soybean cultivation. However, the main risk of the use of conversion from pasture or degraded grass- land that could be used for cattle grazing, is to push this type of activities in other areas of further deep in the Amazon, where there is still land available, increasing the risk of deforestation (West et al., 2018). 4.1.1.3 Policies and commitments Agricultural land creation is a crucial step within soy product chain because of all the environmental aspects that it involves, described above. Depending on the biome, whether is the Amazon or the Cerrado, agricultural land creation is regulated by the Brazilian Forest Code, which imposes different percentages of native vegetation that must be preserved on private lands, much higher in the Amazon compared to the Cerrado. Moreover, the "Soybean Moratorium" has been established in 2006 to control agricultural land creation related to soybean production. It is a multi- stakeholders initiative that limit soybean cultivation and its sponsors are committed not to buy or finance soybean crops established in the Amazon Biome after 2008. It has been agreed between the Brazilian government, environmental NGOs and 26 4. The Soy Product Chain Organization different companies and traders among the Brazilian Association of Vegetable Oil Industries (ABIOVE) and the National Association of Cereal Exporters (ANEC), which together account for the commercialization of more than 90% of Brazilian soybean (Cattelan and Dell’Agnol, 2018). The benefits in the Amazon biome have been impressive, since the soy planted in recently cleared land decreased from 30% in 2006 to only 1% by 2014. However, its main limitation is that it doesn’t include the Cerrdo biome in its scope, which is a highly biodiverse savannah, important for climate stability and biodiversity as well as the Amazon biome. Not covering this area, the Soy Moratorium might decrease agricultural expansion in the Amazon, but it also might push soybean-associated land conversion in the Cerrado, which already holds less strict regulations from the Forest Codes and where, in fact, took place the most recent soy expansion, as shown in Figure 4.4. Figure 4.4: Soy expansion, 2005-2016 (West et al., 2018) . In response to that, the Cerrado Manifesto has been released in 2017 which includes a call for action from civil society stakeholders and an implementation plan to re- duce deforestation in the Cerrado. Over 70 large corporations had already signed on to the agreement in 2018, including major retailers, fast-food chains and big brands such as McDonalds, Tesco, L’Oreal, Nestlè, Unilever, Carrefour, IKEA and ICA gruppen. However, large companies such as Cargill, ADM and Bunge – which all benefit from the agricultural land in the Cerrado – haven’t taken part of the ini- tiative yet. Nevertheless, their participation is considered essential to the Cerrado Manifesto’s success (Belmaker, 2018). Although many improvements in reduction of deforestation and direct conversion 27 4. The Soy Product Chain Organization from native vegetation have been made between the first and second half of the 2000s, with a reduction in deforestation related to soybean passing from 455 m2 per year per tonne of soybean in 2001 - 2005 to 97 m2 per year per tonne of soybean in 2006 - 2010, there has been a parallel increased use of land, water and fertilizer equal to 30% linked to likewise environmental impacts (WWF, 2014) 4.1.2 Agricultural production Since 2018, Brazil is the world’s biggest soybean producer in the world (West et al., 2018). Soybean is an annual crop grown in moderate, sub-tropical and tropical climates (WWF, 2014), and in Brazil average productivity rated have been able to triple since 1970. Total soy production in Brazil has significantly grown also in the last two decades, and according to Trase data (West et al., 2018), from 2000 to 2016, total production has increased from about 45 million tons per year to more than 90 million tons, and the area of production has risen accordingly. 4.5) The Figure 4.5: Trends in soy production in Brazil, Argentina and Paraguay (West et al., 2018) production from 2003 to 2017 in the Cerrado biome increased from 19 million tones to 46 million tons; in the Amazon from 2 million tones in 2003 to almost 15 million tons (West et al., 2018). 4.1.2.1 The process Once the clearing is complete, the soil is ploughed and prepared for planting. Soy- beans are usually planted in October or November when rains start, and then is harvested in April or May. They can be grown by themselves occasionally, but mainly in rotation with other crops, such as winter wheat or maize (Mattson et al., 1999). Although the low natural fertility of the soil, especially in the Center-West of the country, this region has the highest agricultural potential, needing the ap- plication of very little quantities of nitrogen input for very high yields. (Raucci et al.,2014) In fact, 70-85% of the nitrogen requirement is supplied by biological fixa- tion (Alves et al., 2003). However, due to problems with erosion, most of soybean 28 4. The Soy Product Chain Organization areas in Brazil are cultivated under the no-tillage system (Raucci et al., 2014) and this increase the need for herbicide treatment. Also, insecticides are used rather extensively (Mattson et al., 2000). 4.1.2.2 Enviornmental implications Focusing the Cerrado, where much of the production is located, related to soybean cultivation the soil loss amounts to 8 tonnes per ha and year, (Mattson et al., 2000) and loss soil organic matter is also a serious problem, mainly due to warm climate, dry winters, quick decomposition of crop residues (Castro and Logan, 1991). More- over, the use of heavy machinery reduces the porosity of the soil (soil compaction), which can contribute to a reduction of crop yields (Mattson et al., 2000) Figure 4.6: GHG sources in soybean production in Mato Grosso (Raucci et al., 2014) Agricultural cultivation is considered an "hotspot" for global warming potential (Dal- gaard et al,. 2008), and when considering GHG emissions from the operations of production, crop residues result to be the main factor of contribution, which repre- sents 33-40% of total emissions, followed by the use of fossil fuels for the agricultural operations, 20% of total emissions. The use of pesticides (herbicides, fungicides and insecticides) accounts for 6-10% of GHG emissions, but has other important envi- ronmental impacts Figure 4.6. 4.1.2.3 Yields and production What makes growing soybean very attractive in agriculture, is its ability to tie up nitrogen in the soil. Being nitrogen a primary element for growing crops, the rota- tion of soybeans with other types of crops means that the next ones will require less fertilizer inputs. This is also linked to its ability to respond more quickly to changes in the global market prices, compared to other multi-annual crops such as palm oil, because if prices are lower, agricultural land can be easily converted in other crops and viceversa (WWF, 2014). Recently, due to droughts in Argentina and a 29 4. The Soy Product Chain Organization possible trade war between US and China, there has been a marked rise in Brazil- ian soy prices. In this regard, and also because of the growth of Chinese demand and domestic consumption, the United States Department of Agriculture (USDA) had forecasted a 2% increase in soy harvest in Brazil in 2018. (Einstein-Curtis, 2018) Crop yields are an important factor regulating soybean production, as well as other steps of the chain. High yields, in fact, are critical to secure food and energy supply while preserving environmental quality and natural resources. Agricultural yields differ according to the growing region in the country, with substantial gaps between one region and another. According to 2011 data (Figure 4.7, Mato Grosso, in the northern part of Brazil, was the region with the highest yield (2778 kg/ha) while the lowest yield was in Rio Grande do Sul, in the very South of the country (1880 kg/ha). It is important to notice though that in this southern region agricultural expansion is no longer an option, whereas states such as Mato Grosso it is still occurring together with an increase of yields (Santhelas et al., 2015). In general, Brazil has increased Figure 4.7: Average soybean yield between 1990 and 2011, and yield, area and total production in the 2010/11 growing season in the main Brazilian production states (Santhelas et al., 2015) both production and agricultural area since 1961, with yields increasing more than the production area. This has been possible thanks to mechanization, improvements in soil fertility, investment in agricultural research and innovation. However, higher yields are linked to an increased use of fertilizer and pesticides, with impacts on soil and groundwater (Santhelas et al., 2015). Genetically modified (GM) crop technology has played a major role in increasing agricultural yields. Today, more than 90% of the soy production in Brazil is GM. Introduced in 1990, it has considerably stimulated soy production and expansion. New seed types allowed producers to both increase yields and start cultivating areas that were not agriculturally viable before (West et al., 2018). 30 4. The Soy Product Chain Organization 4.1.2.4 Some powerful actors In this section some of the most powerful actors - mainly big companies - in soy- bean production will be listed, linking them to their investors when the information retrieved allow that. • O Telhar Agropecuaria, backed by Altima Partners (UK hedge fund) and Capital Group (US private equity firm), is a major soy producer in Mato Grosso (Levicharova et al., 2018). This company began as an association of Argentine cattle farmers in the 1980s and started to be involved in grain production in the 1990s. According to GRAIN data (Grain, 2012), in 1999 O Telhar was one of the largest soy producers in Brazil, farming on rented land. Foreign investors entered the company after 2006, and in 2012 it was farming 800,000 ha of Brazilian land. According to Steinweg et al. (2018) today O Telhar holds 86,574 ha of land in the Matopiba region. • SLC Agricola, backed by Valiance Asset Management limited (a UK-private equity fund), is also one of the biggest soy producers (Fian international, 2018) operates 15 large farms, spread across six Brazilian states – Mato Grosso, Goiás, Bahia, Piauí, Maranhão and Mato Grosso do Sul, including farmland holdings in the Matopiba region (Kuepper et al., 2017). Through its real estate joint venture SLC LandCo it holds 87,000 ha, although its total holding amount to about half million he of land in Brazil, with some 300,000 ha planted with soy (Fian international 2018). Between 2011 and 2017, the deforestation associated with its farmland holdings amounted to 39,887 ha of land (Kuepper, 2017), and a deforestation of 66,234 ha only in the Matopiba region between 2000 and 2017 (Steinweg et al., 2018). SLC farms are located in areas, such as Santa Filomena in Pauì, where illegal land grabbing is common, and SLC Agrícola’s business partners have faced legal charges (Kuepper, 2017). SLC Agricola is also linked to another big investor, Mitsui and Co, from Japan, forming with them the joint venture SLC-MIT Empreendimentos Agrícolas S.A., another soy producer with about 87,000 ha of land holdings in Matopiba, linked to about 12,000 of land deforested. • Agricola Xingu SA is another big soy producer linked to the Japanese investor Mitsui and Co. With SLC-MIT and Xingu, Mitsui and Co formed the so- called Multigrain Group in 2011, which now owns 120,000 ha of farmland in Brazil. Xingu’s soybean production is non-genetically modified, with the aim for Mitsui to ensure a stable and trustworthy supply to Japan and other markets (including China) (Mitsui and co., 2019). • BrasilAgro is a Brazilian rural real estate firm, which produces soybean holding 11 properties, but mainly focuses on acquiring ’non-productive’ or ’underuti- lized’ land, generating revenues by clearing, developing land and selling it. It is backed by Cresud, an Argentine corporation. • Other important producers are Gruppo Amaggi, Agrex do Brazil, Terrasanta, Brookfield Brasil and Grupo Bom Jesus. Levicharova et al., (2017) looked at the ESG policies of these last soy producers together with the ones listed above, giving them a score between 0 and 100 according different criteria, evaluating 31 4. The Soy Product Chain Organization the overall presence and scope of ESG policies, the inclusion or exclusion of environmental standards, the adherence to key criteria on human and labor rights, and performance on transparency and good governance practices. The companies registered a score of 31 out of 100, showing a lack of environmen- tal awareness and commitment to zero-deforestation policies, even though the trend is tacking off by key buyers. In fact, what emerged is that the scores are driven by policies on governance and disclosure, human and labor rights and the overall scope of the commitments, more than being driven by envi- ronmental standards. Figure 4.8 devides the companies mentioned above in low-scoring and high-scoring, comparing them with the scores in ESG policies of their investors. Figure 4.8: Average ESG policies of soy producers compared to their investors’ scores. (Levicharova et al., 2018) • Commercial banks, international asset managers and self-funding are the prin- cipal forms of financing for agricultural activities. Domestic lending is associ- ated to high costs, since private sector banks are afraid to provide long-term credit, because of the unpredictable risks related to the agricultural sector (Levicharova et al., 2018). • The Brazilian Development Bank or Banco Nacional de Desenvolvimento Econômico e Social (BNDES) is also a very important actor, disbursing rural credit and mostly benefiting large commercial farmers. In 2016/17, 21% of subsidized credit lines were addressed to large producers and about 12% to family farms (Levicharova et al., 2018). 4.1.2.5 The relationship between production and finance Soybean is usually grown on big plantations by very large agricultural enterprises but, less frequently, it can also be cultivated on a smaller scale on family farms work- ing in cooperatives. In big farms, manual labor is usually limited and it involves some direct planting, manual weeding and stone removal to avoid damages to har- vesters. The massive increase of production since 1970 has been possible thanks to a "debt-fueled" modernization of agriculture (Network for social justice and human rights, 2018), using financial mechanisms such as subsidized credit, tax exemptions, above-cost pricing policies and the cancellation of already subsidized debts (Jùnior, 2002) 32 4. The Soy Product Chain Organization One of the main causes driving territorial expansion of soybean production, espe- cially in the Matopiba region, has been asset price inflation’s stimulation. Financial asset price inflation occurs when there is a speculative increase in the price of a certain asset which is attracting new investors, consequently dropping when the fi- nancial bubble bursts. Commodity prices started to drop when the global financial crisis erupted in 2008, hence speculative capital was transferred to low risk securities, such as agricultural land and food commodities, included soybean. In this situation pension and hedge funds in search for returns, invested in this type of commodity basing their trades on specific future prices and promises of expanded production. As a result, soybean producers as well as the processing industry and traders, started acquiring more and more resources in order to obtain more capital in exchange for pledges to expand production. This mechanism, called financial "simulation", is in turn able to push market prices of soybean and other commodities to increase. Con- sequently, high commodity prices led both the territorial expansion and the increase of soybean production, with a greater expansion of big corporations, able to exploit land as an asset, promising large-scale production and increased profits to investors (Network for social justice and human rights, 2018). Observing the graph 4.9, it is Figure 4.9: Soybean productivity from 1976 to 2016 (Data from: Network for social justice and human rights, 2018) possible to notice three main peaks in soybean productivity beyond the trend line. The first two correspond to the years 2000 and 2002, and the third one in 2010. Such high productivity values followed, respectively, two important events. In 1999, BNDES powered agricultural production offering subsidized credit and in 2008 the global financial crisis begun, pushing foreign investments in Brazilian farmland. In both cases, after a couple of years, soybean producers increased their production and productivity levels by expanding into new areas and adopting new techniques to in- crease their yields, such as the use of genetically modified (GM) crops. Right after the crisis in 2008, the fall of productivity as well as production was linked to the momentary restriction on credit, which affected the 2008-2009 harvest. However, the agricultural area kept expanding also in periods of decreased production and productivity. Only in Matopiba, this resulted in a 235% increase of area dedicated to soybean between 2000 and 2014, jumping from 1 million to 3.4 million hectares. 33 4. The Soy Product Chain Organization 4.1.3 Processing and trading As showed in Figure 1.1, soybean is used for many different purposes. The first processing step starts in crushing facilities, where raw soybeans are turned into the principal products of soybean: animal feed and oil. 4.1.3.1 The process Once harvested, soybeans are bought, collected and sent to crushing facilities or processing industries. Globally, of total soybean production, around 87% is crushed (Oil World, 2019). The rest is usually used for direct human consumption, sent to specific food industries or used as seeds. Some products derived from non-crushed seeds are for example tofu, soy-sauce or other meat or dairy substitutes. Crushing plants are used to mechanically crushing the soybeans to extract soy oil and separate it from soy meal. During this process, about 79% of the soybean is turned into meal, which is subsequently toasted, dried and grinded. The protein content of the end- product is between 44% and 48% - depending if the hulls are removed or not before the crushing - and that makes it a perfect ingredient for livestock feed, but also for protein-rich food or non-food purposes. In crushing plants, soybean can also be processed into crude soy oil (around 18%) which is then transferred to oleochemical plants or refineries. Crushing plants can either be located close to soybean growing area o near a harbour. (Van Gelder and Dros, 2002) In Brazil, most of the total soy-crushing capacity is in the south of the country, even though Matopiba has seen a 75% increase in crushing capacity between 2005 and 2015 (West et al., 2018). After being crushed, soy is exported to consumer countries or used for domestic consumption. 4.1.3.2 The environmental implications According to Norris et al., (2016) - considering soybean produced and crushed in US, but not shipped overseas - soybean milling in crushing facilities has most contri- bution in the impact categories of resource depletion and climate change, compared to the other impact categories included (human health, ecosystem quality, water withdrawal). Another LCA study (Da Silva et al., 2010) - which considered soy- bean production and shipping to Europe, but without including crushing processing prior to export - highlighted the importance of the transport phase compared to the production phase, which in terms of climate change, acidification and cumulative energy demand in Central West Brazil, is responsible for 40% of the impacts. Also, it is suggested that the mode of transport chosen and the distance to be traveled strongly influence environmental impacts. In this regard, (Dalgaard et al., 2008) suggests that shipping is much more environmentally friendly than transport by truck, although both choices have higer impacts on eutrophication and acidification compared to agricultural production. 4.1.3.3 The power of soy traders Crushing plants are an important element in the processing and trading step, since they are mainly owned and managed by soybean traders, which can be very small, 34 4. The Soy Product Chain Organization local companies, but more often large, international corporations. Traders buy the product from farms, process it into crushing facilities and then export it to consumer countries. Only a bunch of traders are responsible for most of soy exports in Brazil: in 2016, the six largest traders - Archer Daniels Midland (ADM), Cargill, Bunge, Louis Dreyfus, Amaggi and COFCO - accounted for 57% of all Brazilian soy exports (West et al., 2018). Figure 4.10: Major soy traders and destinations (Kuepper et al., 2017. Data from Panjiva) Each of them owns between 2 and 8 crushing facilities and has different sourcing regions across Brazil. The relationship between traders and municipalities is very strong and, according to latest West et al. (2018) it was estimated that for some two-thirds of Brazilian municipalities exporting soybean, more then half of the ex- ports were managed by a single trader, and in about one-fifth of municipalities only one trader was registered to handle all the exports. The soy infrastructures operated by big traders are not limited to crushing facilities, they also include silos, ware- houses, but also railroads and port terminals. Hence, they play a very important role in shaping the economic development of the areas they are controlling, and are also very important players in shaping a more sustainable soybean sourcing (West et al., 2018). In the following sections, some of the most important traders will be presented, describing their operations, their main sourcing areas and their steps to tackle defor- estation problems. Also, for ADM, Cargill and Bunge, information about investors and financial characteristics will also be provided. In Figure 4.10 an overview of all different traders and soy destination is given. 35 4. The Soy Product Chain Organization 4.1.3.4 Archer Daniels Midland Archer Daniels Midland (ADM) one of the world’s largest agricultural commodity traders (GRAIN, 2012), and among the largest processors in Brazil (Steinweg and de Wilde, 2018), it is headquartered in US and started operating in Brazil since 1997. It sources its soy mainly in Mato Grosso (West et al., 2018) and owns more than 30 silos in Brazil, as well 13 oilseeds processing plants. It also has operations in eight Brazilian ports (Steinweg and de Wilde, 2018). According to West et al., (2018), it also owns Brazil’s largest crushing plant. ADM is the fifth largest soy trader in Brazil’s Matopiba region, the last soy frontier, and it has been associ- ated with 13,873 hectares (ha) of deforestation in its main sourcing municipalities in 2017. However, ADM doesn’t publish a list of its soy suppliers - which does for its palm oil commerce - not willing to disclose the names of the farmers from which they source soy. Moreover, ADM doesn’t even support the Cerrado Manifesto, which many of its clients support instead. To show its environmental commitment, ADM has published a no-deforestation policy in 2015, mainly focusing on traceabil- ity, transparency of its soy supply chains and on engagement with their suppliers (Steinweg and de Wilde, 2018). However, unlike its competitors Cargill and Bunge, ADM hasn’t given a time-frame of commitment. The company has a Responsible Soybean Standard, a certification program which aims at bolster environmentally and socially soy production, but that tackles only illegal deforestation, leaving legal deforestation processes still possible (Steinweg and de Wilde, 2018). When it comes to its financing strategy, its assets are 50% financed by its equity and that BlackRock, Vanguard, State Street (Galaz et al., 2018) and Maquarie Group (Steinweg and de Wilde, 2018) are among their biggest shareholders, all investment management firms which, on average, have weak forest policies. There are, however, other shareholders that, even though having less stake, have demonstrated to be more concerned to deforestation related to soy production, divesting by some other companies (SLC Agricola) which had been highly related to deforestation (Govern- ment Pension Fund of Norway) and engaging with Bunge about that. Some 20% of ADM’s assets are either financed by gross debt. Also in this case, there are more en- vironmentally concerned investors and less concerned ones. Among the first group, engagement or divestment possibilities are very low. Among the most concerned ones are Aegon and Aviva (pension and insurance funds, and asset management firms respectively from the Netherlands and from UK), and Deutsche Bank, with deforestation policies in place and signatories of the Cerrado Manifesto (Steinweg and de Wilde, 2018). 4.1.3.5 Cargill Cargill is the second biggest soy exporter, preceded by Bunge and followed by ADM. It is a privately owned company registered in US, which started to operate in Brazil since 1965. Its assets include 6 crushing facilities, 140 silos and 5 port terminals, and it sources it soy mainly from the North and North-West regions in the Amazon (West et al., 2018). Cargill constructed a soy export harbour terminal in Santarém, Parà state, in 2003, one of the city where many land conflict have emerged. Af- 36 4. The Soy Product Chain Organization ter establishing the port in fact, this area has become attractive for soy cultivation, leading large landowners from the south of the country to buy or — more commonly — grab land, furthering conflicts (Van Solinge, 2010).However, the state from which Cargill holds much of the soy export market share (21%) is Maranhão, where is the largest exporter through its subsidiary Cargill Agricola and where, between 2010 and 2017, about 336,426 ha of land has been cleared for soy. More than half of these export are destined to the Chinese market. Also in Maranhão, the presence of Cargill, together with the one of other big traders, has had not only environmental, but also social impacts. Here, a number of agrarian land conflicts are present, but also cases of slave labor have been reported. Furthermore, the social and economic development of the territory is particularly difficult given the demand for skilled technicians, which lack in the area. Local research in Maranhão in 2018 also showed local communities to suffer from agrochemical pollution, lack of water and pesticide poisoning, all linked to soy production (Steinweg and Rijk, 2018). Cargill, in a ven- ture with a London-based tractor maker co