From frontier governance to governance frontier: The political geography of Brazil’s Amazon transition

World Development 114 (2019) 59–72

Contents lists available at ScienceDirect

World Development journal homepage: www.elsevier.com/locate/worlddev

From frontier governance to governance frontier: The political geography of Brazil’s Amazon transition Gregory M. Thaler a,⇑, Cecilia Viana b, Fabiano Toni c a

Department of International Affairs, University of Georgia, 328 Candler Hall, Athens, GA 30602 USA Lancaster Environment Centre, Lancaster University, LEC Building, LA1 4YQ Lancaster, United Kingdom c Center for Sustainable Development, University of Brasília, Campus Darcy Ribeiro, Asa Norte, Brasília, DF 70910-900 Brazil b

a r t i c l e

i n f o

Article history: Accepted 24 September 2018

Keywords: Frontier Governance Deforestation Agricultural intensification Brazil Amazon

a b s t r a c t The ‘frontier’ is central to a new wave of development scholarship, but the broad deployment of the concept has blurred several key dimensions of frontier development. We focus on the Brazilian Amazon to synthesize classical frontier theory and emerging perspectives with special attention to the role of governance in frontier development. Since 2004, primary deforestation in the Brazilian Amazon has declined over 70 percent while agricultural production in the region has increased. Contrary to narratives that view this transition as the result of ‘frontier governance’ – i.e., the imposition of order on a preexisting frontier – we propose the concept of a ‘governance frontier,’ which recognizes the role of politics in constructing and transforming frontier spaces. This concept politicizes economic accounts of frontier development and spatializes abstract notions of governance. We employ a ‘follow the policy’ methodology to trace the evolution of a governance frontier in the eastern Brazilian Amazon, drawing on original fieldwork across four Amazonian municipalities and inside an environmental non-governmental organization. We show that a key feature of the Amazonian governance frontier has been a distinct geographical configuration of ‘model municipalities’ that function as nodes of policy experimentation, legitimation, and transfer. Our findings support an integration of frontier theory and governance theory in a placebased, political geography approach to regional political-economic transformation, which demands greater attention to the political dimensions of frontiers and to the spatial dimensions of governance. Ó 2018 Elsevier Ltd. All rights reserved.

The ‘frontier’ is a persistent spatial imaginary in academic and popular understandings of development. Although the frontier has been prominent historically in scholarship on New World settlercolonial societies (Billington, 1971; Hennessey, 1978; Turner, 1921; Webb, 1952), a new wave of frontier thinking is underway. A spate of recent publications focuses on ‘commodity frontiers’ (Moore, 2000, 2015) or ‘resource frontiers’ (World Development, 2018); a 2016 Social Science Research Council workshop was devoted to ‘‘Frontier Assemblages” in Asia; and a series of five sessions at the 2018 Annual Meeting of the American Association of Geographers examined ‘‘Frontiers in the Contemporary Moment.” These engagements with the frontier concept are producing valuable new theoretical and empirical insights. As frontier theory is deployed more broadly, however, several weaknesses are becoming apparent, especially with regard to the literature on ‘resource frontiers,’ which views frontiers as constituted through the extraction and commodification of specific resources such as ⇑ Corresponding author. E-mail address: [email protected] (G.M. Thaler). https://doi.org/10.1016/j.worlddev.2018.09.022 0305-750X/Ó 2018 Elsevier Ltd. All rights reserved.

gold, carbon storage, or ‘scenery’ (Barney, 2009; Rasmussen & Lund, 2018, 391; Tsing, 2003; World Development, 2018). In particular, the resource frontier approach struggles to explain several key dimensions of frontier development, namely, the articulations among different resource frontiers within a single landscape, the ways that distant frontiers may be linked with each other, and agricultural intensification as a frontier phenomenon. This article returns to the New World settler-colonial frontier of the Brazilian Amazon to develop a synthesis of classical frontier theory and emerging perspectives with special attention to the role of governance in frontier development. In Brazil, the Amazon region has been a lodestar of frontiermaking projects, and the inspiration for diverse strands of frontier theory, in both English and Portuguese (e.g., Becker, 1982; Foweraker, 1981; Martins, 1972; Schmink & Wood, 1992; Velho, 1972). The Brazilian Amazon since the 1970s has undergone massive land use change, resulting in deforestation of an area larger than France (INPE, 1989, 2018; Skole & Tucker, 1993). Forest conversion was driven initially by small farmer migration, extractive logging and mining, and extensive cattle ranching, followed by industrial

60

G.M. Thaler et al. / World Development 114 (2019) 59–72

ranching and soy cultivation. Since 2004, however, regional deforestation rates have declined over 70 percent,1 while soy production in the Legal Amazon has nearly doubled and the cattle herd has grown by 20 percent (IBGE, 2018a, 2018b).2 This transition towards a lowdeforestation, high-productivity agro-industrial economy has been strongly influenced by new policy measures that Nepstad and colleagues term ‘frontier governance’ (Nepstad et al., 2002, 2014). Taking as a starting point the crucial role of governance in transforming the Amazon frontier, this article explores the political geography of Brazil’s Amazonian transition, i.e., how governance develops through space and time. Our goal is not to explain the full range of policies and socio-economic factors that led to reductions in Amazonian deforestation. These factors have been analyzed in a number of important articles (Assunção, Gandour, & Rocha, 2015; Hecht, 2011; Nepstad et al., 2014). Deforestation reduction is part of a broader ‘‘pervasive transition” of the Amazonian land use system (Lapola et al., 2013), however, which includes transformations in productive practices and policy frameworks in the agricultural zone outside of conservation areas and indigenous territories.3 We focus on this zone of agricultural colonization, where politicaleconomic developments have depended not just on reduced deforestation and agricultural intensification, but also on political changes that help direct and stabilize the emerging ‘green’ agro-industrial order. Political change, like forest conservation and agricultural intensification, is a process that develops through time and space, and it does not occur everywhere equally and all at once. Governance on the ground takes shape through the interplay of local policy processes with governance initiatives at other levels and scales, which vary from place to place in their implementation. We approach this political geography of Amazonian governance through an engagement with the literature on Amazonian frontier development. Based on our reading of this literature, we conceptualize frontier change as the expansion and contraction of particular political-economic systems of rural production. This approach revises the ‘resource frontier’ perspective by incorporating resource frontiers within a systemic political economy. At the same time, we share with the resource frontiers literature an attention to multiple, interrelated political-economic and socio-ecological dimensions of frontier formation. Consequently, we critique the narrative of Amazonian ‘frontier governance’ for its narrow definition of the frontier and its abstract understanding of governance. Land cover change and governance are both spatiallydifferentiated processes and do not necessarily evolve in congruence. We employ a ‘‘follow the policy” approach (Peck & Theodore, 2012) to reveal how abstract constructs of ‘governance’ are constituted through specific political processes situated in space and time. Based on original fieldwork following policy within an environmental non-governmental organization (NGO) and across four municipalities in the eastern Brazilian Amazon, we demonstrate the importance of ‘model municipalities’ as nodes for the development of Amazonian environmental governance, and we argue for a reconceptualization of environmental governance and frontier theories through the frame of a ‘governance frontier.’ The ‘governance frontier’ concept works to politicize economic accounts of frontier development, to spatialize abstract notions of governance, and to problematize narratives of ‘frontier governance.’ Additionally, we advocate moving beyond conventional distinctions between ‘hard’ and ‘soft’ governance (based on

1 Deforestation figures in this article are from INPE’s PRODES program (INPE, 2018), which uses Landsat imagery to report annual clear-cutting of areas over 6.25 ha in Amazonian primary forest from 1988 to present. 2 The Legal Amazon is an area designated by the Brazilian Government for regional policy and includes the states of Acre, Amapá, Amazonas, Mato Grosso, Pará, Rondônia, Roraima, Tocantins, and western Maranhão. 3 Agriculture refers to both cultivating crops and raising livestock.

‘state’ versus ‘non-state’ actors) to recognize the essential hybridity of policy processes, and we conclude with a call for research to examine the entanglements of the governance frontier with other social, economic, and ecological frontier dimensions. 1. Background Large-scale deforestation of the Brazilian Amazon began in the 1970s with the expansion of small farmer migration, extractive logging and mining, and extensive cattle ranching (Hecht & Cockburn, 2010; Schmink & Wood, 1992). Since the 1980s, industrial ranching and soy cultivation have expanded dramatically (Rudel, DeFries, Asner, & Laurance, 2009), accelerating annual deforestation rates to the point where in 2004 nearly 28,000 km2 of forest was cleared. After 2004, however, deforestation rates declined 84 percent to reach their lowest point in 2012, while regional agricultural production continued to increase (cf. Macedo et al., 2012). Deforestation rates have crept upward since then, but they remain at historically low levels as the agricultural economy continues to intensify (Gibbs et al., 2015; Pacheco & Poccard-Chapuis, 2012; zu Ermgassen et al., 2018). Deforestation reductions have been driven significantly by new environmental governance measures (Assunção, Gandour, & Rocha, 2012; Nepstad et al., 2014), including the creation of new conservation areas and indigenous territories (Schwartzman, Alencar, Zarin, & Souza, 2010; Soares-Filho et al., 2010), soy and cattle industry moratoria on commercialization of commodities from newly-deforested areas (Gibbs et al., 2015, 2016), and enhanced enforcement of the Federal Forest Code.4 New governance measures have also supported agricultural intensification, for example through the creation of new credit incentives such as the federal government’s Low-Carbon Agriculture Program (ABC), and through the promotion of integrated pasture-cropping systems by the federal agricultural research corporation (Embrapa). Governance measures are by no means the only factors behind deforestation reductions and agricultural intensification. They interact with other forces such as social movement activism, macroeconomic factors, technological developments, and environmental changes. At the same time, an understanding of governance processes is necessary for a full understanding of the dynamics of regional transformation. Crucially, the multiple political-economic dimensions behind the Amazonian transition do not materialize everywhere equally, nor do they always cooccur. In particular, the governance apparatus that supports ‘green’ agro-industry in the Brazilian Amazon has emerged through specific policy processes with particular temporal and spatial manifestations. Municípios, or municipalities, have been a key level for the development of new policy models and the implementation of state and federal policies. We adopt a place-based approach in this article that situates municipalities in relation to each other and to regional and national-level governance processes. Before elaborating on our methodology, we frame our approach to the political geography of Amazonian development through an engagement with the literatures on frontiers and governance. 2. Frontier theory Nepstad et al. (2002) describe the Amazonian transition as a process of ‘frontier governance.’ In geography, history, and political 4 The Brazilian Forest Code prohibits clearing along bodies of water and on steep slopes and hilltops, which are called ‘permanent preservation areas,’ and it generally requires landowners in the Amazon to maintain 80 percent of their property area under natural vegetation as a ‘legal reserve,’ deforesting no more than 20 percent of their properties for other uses. The Forest Code was passed in 1965 and later modified by presidential decrees, but it went largely unenforced until 2004. A revised Forest Code was passed in 2012.

G.M. Thaler et al. / World Development 114 (2019) 59–72

ecology, the frontier is a deep-rooted concept, denoting the edge of a system and a zone of contact (Billington, 1971; Hennessey, 1978; Lattimore, 1962; Sauer, 1963; Turner, 1921). In the contemporary Brazilian Amazon, scholarship has focused on two major frontier formations. The first is the deforestation frontier, the infamous ‘Arc of Fire’ across southern and eastern Amazonia where forest is cleared for other land uses, predominantly extensive cattle pasture. The second is the intensification frontier, or what Brannstrom (2009) describes as the ‘neoliberal agricultural frontier’ of industrial soy cultivation and intensive cattle ranching that expands largely over previously cleared areas (VanWey, Spera, de Sa, Mahr, & Mustard, 2013). These two frontiers correspond respectively to the boundaries of the extensive agricultural economy, dominated by pastures with stocking rates often below 1 head/hectare (Pacheco & Poccard-Chapuis, 2012), and the intensive agricultural economy, characterized by high levels of capital investment and agricultural productivity (Arvor, Meirelles, Dubreuil, Bégué, & Shimabukuro, 2012; Macedo et al., 2012). In this article, we refer to the deforestation front as the extensive frontier and the intensification front as the intensive frontier, and we focus primarily on the latter. Frontier dynamics in the Brazilian Amazon are the subject of a voluminous literature comprising diverse theoretical perspectives. We group theoretical understandings of Amazonian frontier development into two large tents: economic geography and political economy.5 Approaches in economic geography include land rent theory, new institutional economics and new economic geography, and household life cycle theory. Land rent theory is based in the von Thünen tradition, and holds that land use (and hence the geography of frontier expansion) is determined by location rents that materialize as a function of distance to markets (Walker, 2004). While land rent theory recognizes that location rents emerge within a policy framework (Walker et al., 2009), it is not attuned to spatial differentiation in the construction and implementation of policy, and it defines the frontier entirely through its economic dimension. New institutional economics and new economic geography subsume neoclassical land rent approaches within a meso-level perspective that attends to particular configurations of economic organization in the production of the frontier. Jepson (2006), for example, focuses on the role of agricultural cooperatives and colonization firms in reducing transaction costs and mitigating risks to explain the geography of industrial row-crop expansion on the intensive frontier in Mato Grosso State. Other research highlights how differentiated environmental regulations and land tenure institutions determine the emergence of agglomeration economies that structure the geography of Amazonian soy production (Garrett, Lambin, & Naylor, 2013). Despite increased attention to institutional factors, the frontier imaginary in this perspective remains narrowly economic, a question of land use. A third approach within economic geography is household life cycle theory, rooted in the Chayanovian tradition, which argues that on smallholder properties, household land use decisions are partially determined by household labor supply and consumption dynamics (McCracken et al., 1999; Walker & Homma, 1996). Household life cycle models are useful for explaining micro-level dynamics on smallholder plots (Perz, 2001, 2002). At higher levels of spatial aggregation, the ability of these models to explain frontier development depends on the mix of actors present, since largescale deforestation and agro-industrial expansion dominate land change dynamics in many parts of the Amazon (Pacheco, 2012). Brondizio and Moran (2012) demonstrate that household life cycles, new institutional economics, and land rent theory can come together to identify distinct deforestation trajectories at different

5 Summers (2008) provides a useful overview of frontier theory in the Brazilian context.

61

levels of analysis. Even in combination, however, these economic geography perspectives understand the ‘frontier’ through a single dimension of land use change, be it deforestation or agricultural intensification. Politics enter on a technical level as ‘policy,’ which alters the economic calculus, but contention and the operation of power are largely omitted, despite the fact that conflict and violence are at the core of the frontier narrative and experience. Political economy approaches place greater emphasis on these political dimensions of frontier development. We categorize political-economic frontier theory into structuralist and poststructuralist accounts. Structuralist accounts conceptualize frontier development as the expansion or reproduction of capitalist relations of production. One strand of structuralist scholarship describes linear stages of settlement and agro-industrial expansion that emerge through dialectical, violent contestation between peasant and capitalist frontier actors (Foweraker, 1981; Martins, 1972). This process of frontier expansion may be punctuated by the opening of ‘‘speculative fronts” and retractions (Sawyer, 1984), but tends ultimately to the displacement of peasants and growth of modernizing agro-industry. In a second strand of structuralist scholarship, associated with world systems analysis, frontier development is traced through the international division of labor to link expanding Amazonian commodity production with world market demand and restructuring, mediated by the political domination of large landowners and political elites (Bunker, 1985; Frank, 1967). Perspectives differ among structuralist accounts on the degree to which they expect the Amazon to become a zone of modernized, industrial production; in the first strand, parts of the Amazon may undergo industrial development while pushing the deforestation frontier further into the interior, while in the second strand the Amazon is conceived as a perpetual periphery experiencing continual extractive cycles of underdevelopment. Post-structuralist political-economic theory sees a more variegated Amazonian landscape where frontiers emerge and are shaped through contested socio-ecological relations at multiple spatial and temporal scales.6 Diverse actors come together variously and often violently on ‘contested frontiers’ (Schmink & Wood, 1992) that are dynamic products of sedimented socio-ecological processes ‘disarticulated’ from each other and from linear narratives of national development and capitalist structure (Browder & Godfrey, 1997; Schmink, Hoelle, Gomes, & Thaler, 2017). The ‘frontier’ in post-structuralist theory has multiple dimensions, including land use, political institutions, and cultural narratives. Arima, Walker, Perz, and Souza (2015) detect these variegated frontier forms in patterns of forest fragmentation, of which they recognize five types produced by different economic and socio-political processes of road construction and property formation. In the cultural dimension, Thypin-Bermeo and Godfrey (2012) identify three different frontier imaginaries in their interviews with residents of the ‘Amazonian boomtown’ of Marabá. Outside the Brazilian context, contemporary political-economic frontier theory demonstrates similar structuralist and poststructuralist tendencies, with a structuralist, world systems perspective evident in discussion of ‘commodity frontiers’ in the literature on capitalism as world ecology (Moore, 2010, 2012), and a poststructuralist perspective prominent in the literature on land control and ‘resource frontiers’ (Blomley, 2003; Eilenberg, 2014; Peluso & Lund, 2011; Rasmussen & Lund, 2018; Tsing, 2003; Watts, 2012). 6 The ‘post-structuralist’ label has many connotations. In this case, we use it to denote political-economic theories that do not see the frontier as a direct and predictable manifestation of capitalist organization. As Summers writes, ‘‘Disarticulated and contested frontiers emphasize the existence of a plurality of frontier types that are not necessarily structured by a single overarching economic system. On the contrary, this frontier theory emphasizes . . . the importance of frontiers as contested spaces that are not necessarily linked to the regional economy in a predictable way” (2008, 28).

62

G.M. Thaler et al. / World Development 114 (2019) 59–72

The resource frontiers perspective takes the ‘resource’ as the unit of analysis.7 According to Rasmussen and Lund, ‘‘a frontier emerges when a new resource is identified, defined, and becomes subject to extraction and commodification” (2018, 391). They observe that the resource frontier dissolves existing social orders, while processes of spatial ordering or ‘‘territorialization” create new systems of resource control (Rasmussen & Lund, 2018, 388). Our consideration of Amazonian frontier theory highlights several weaknesses in this perspective. As the Amazonian experience demonstrates, the ‘resource’ is not always the best unit of analysis for understanding regional transformations that are commonly interpreted through the frame of the ‘frontier.’ Resource frontier theory has difficulty capturing how exploitation of different resources in the same landscape is articulated systemically; for example, how gold mining in Indonesia links with agricultural plantation development (Peluso, 2018, 403), or how protected area creation in the Amazon links with cattle ranching (Thaler, 2017). Further, resource frontiers are inadequate for thinking about agricultural intensification. Expansion of intensive agro-industry in the Amazon is widely understood as a frontier process, but this frontier develops not through extraction of new resources but rather through transformation of the political-economic organization of production. The settler-colonial economy of the Brazilian Amazon thus illuminates the frontier as the boundary of political-economic systems characterized by distinct configurations of rural production. This conceptualization echoes Lattimore’s understanding of the frontier as a zone of transition on an economic scale from extensivity to intensivity (1962, 473–74), and Bunker’s (1985) distinction between extractive and productive economies. Forest extractivism, conversion of land for extensive agriculture, and intensive agro-industry have emerged over the history of Amazonian development as three distinct political-economic formations.8 Each form of rural production is characterized by distinct political institutions and labor processes and generates distinct socio-ecological landscapes at a regional scale. Transition from one form of organization to another, for example at the intensive frontier where intensive agro-industry replaces extensive ranching and subsistence farming, involves processes of spatial reorganization or ‘reterritorialization’ and legitimation, as resource frontier theory explains. ‘Resource frontiers’ are thus subsumed within the frontiers of regional political-economic systems. 3. Governance theory Political-economic theory takes seriously the role of power and conflict in shaping frontiers. ‘Governance’ is the counterpart of the frontier in the lemma of ‘frontier governance’ by which much contemporary scholarship explains Brazil’s Amazon transition (Nepstad et al., 2002, 2014). Beyond its broadest definition as institutionalized socio-political authority, governance has taken on a specific meaning as the exercise of authority by heterogeneous actors (such as corporations, intergovernmental organizations, and NGOs), in contrast to ‘government,’ which is the exercise of authority by the state (Rhodes, 1996; Rosenau, 1995; Stoker, 1998). The extensive literature on ‘environmental governance’ includes on one hand institutionalist scholarship that seeks to 7 Examples of resources given by Rasmussen and Lund include ‘‘oil, gold, new crops like soy or oil palm, carbon storage, or ‘scenery’” (2018, 391). 8 These three categories simplify complex histories and geographies. The extensive agricultural economy in particular has involved diverse and contentious relations between land grabbers, ranchers, small farmers, and traditional and indigenous populations. With the expansion of the cattle economy since the 1980s, however, an increasingly unified extensive economy emerged (Hoelle, 2011; Smeraldi & May, 2008; Walker et al., 2009), and that extensive cattle sector is distinct from forest extractivism and from an agro-industrial economy of plantations, intensive ranching, and smallholder cash-cropping.

improve environmental management (Dietz, Ostrom, & Stern, 2003), and on the other hand more critical scholarship that analyzes environmental governance as ‘the nexus of nature and neoliberalism’ (Himley, 2008). The institutionalist tradition has identified institutional design requirements for the sustainable governance of common-pool resources (Ostrom, 1990; Ostrom, Joanna Burger, Field, & Policansky, 1999), and a large literature has emerged on the question of multi-level governance, which explores how actors and institutions interlink across levels and scales to determine socioenvironmental change (Brondizio, Ostrom, & Young, 2009; Mwangi & Wardell, 2012; Ravikumar, Larson, Duchelle, Myers, & Tovar, 2015). In the Brazilian Amazon, institutionalist approaches to environmental governance highlight how policies at multiple levels and across multiple scales have interacted to reduce deforestation, linking supply chain interventions such as the zerodeforestation Soy Moratorium and Cattle Agreement with federal, state, and municipal policies for protected area creation, land tenure regularization, and Forest Code enforcement (Duchelle et al., 2014; Nepstad et al., 2014). A common distinction is between ‘hard’ and ‘soft’ governance, where hard governance comprises government regulation and soft governance comprises ‘hybrid,’ ‘multi-stake holder’ approaches (Brannstrom, Rausch, Brown, Andrade, & Miccolis, 2012; Lemos & Agrawal, 2006). Brannstrom et al. analyze the Forest Code compliance program in the municipality of Lucas do Rio Verde and the market exclusion strategy of the Soy Moratorium as examples of soft governance, which they argue helps to ‘‘bridg[e] the divide between environmental and agricultural interests” (2012, 365). Critical political economies of environmental governance, on the other hand, have emphasized new forms of enclosure and commodification and the reconfiguration of regulatory frameworks as elements of the neoliberalization of environmental governance (Castree, 2010; Heynen, McCarthy, Prudham, & Robbins, 2007; Himley, 2008; McCarthy & Prudham, 2004). ‘Green neoliberal’ (Goldman, 2005) policies include the creation of private nature preserves, privatization of water supplies, eco-labeling schemes, and emissions trading systems. This literature at times invokes the concept of the frontier to describe not just the spatial expansion of capital accumulation, but also transformations in forms of governance. Thus, the creation of ecosystem services markets is a ‘‘new frontier in the production of nature” (Robertson, 2012; Smith, 2007, 33), and global land grabbing and industrial crop booms are ‘‘new frontiers of land control” (Peluso & Lund, 2011). The ‘frontier’ in these formulations generally refers to an intangible boundary of socio-technical change, and is seldom considered as a concrete spatial form. In the Amazon, critical environmental governance scholarship has argued that soy farmers and cattle ranchers have used compliance with some environmental regulations to enhance their environmental credentials and legitimize agribusiness expansion (Azevedo et al., 2017; Baletti, 2014; Brannstrom, 2009), and it has described a complex of state, agribusiness, and NGO actors that develops environmental policy as a means to promote agro-industrial development and state building (Thaler, 2017). While this scholarship identifies processes of spatial reorganization or ‘reterritorializations’ as a dimension of environmental governance (Baletti, 2012; Thaler, 2017), it has been largely silent on the specific spatial trajectories of these governance processes. Indeed, in both the critical and institutionalist literatures, governance is rarely viewed dynamically across both time and space. Governance theory has given inadequate attention to how governance develops and spreads across a landscape. For the goal of an explicitly spatialized account of the development of environmental governance, the question of frontier governance holds great potential.

G.M. Thaler et al. / World Development 114 (2019) 59–72

63

4. From ‘Frontier Governance’ to the ‘Governance Frontier’

5. Methodology

This article brings together frontier theory and governance theory to reinterpret frontier theories through a focus on the role of governance in socio-ecological change; and to reinterpret governance theory through a focus on space and place. Glossing Brazil’s Amazonian transition as a process of ‘frontier governance’ (Nepstad et al., 2002, 2014) implies a pre-existing, apolitical frontier that is ‘out there’ to be governed. We argue instead that the extensive and intensive Amazonian frontiers can be conceptualized as the boundaries of distinct political-economic formations, each with a characteristic set of governance arrangements. The Amazonian extensive zone of extractive and speculative ranching and smallholder agriculture has been an extremely violent and contentious space. While the state has often appeared absent on the ground, the extensive economy has nonetheless been structured by a set of political-economic relations where large landholders wield authority and patronage and extract resources abetted by state tax policies, fiscal incentives, and land distributions (Hecht & Cockburn, 2010; Schmink & Wood, 1992). This latifundista socio-political order has faced organized challenges from smallholder and indigenous social movements advancing a socio-environmental vision for rural production (Campos & Nepstad, 2006; Schwartzman et al., 2013). These social movements and their allies have prevented the conversion of large areas of forest, especially through the creation of conservation areas and indigenous territories. In the agricultural zone outside protected areas (and within some sustainable-use conservation areas), a ‘‘convergence on cattle” (Hoelle, 2011) since the 1980s turned the extensive economy into a ‘cattle kingdom’ (Smeraldi & May, 2008) stabilized by governance arrangements including ranching credit lines, control of foot-and-mouth disease, export-oriented monetary policy, lax enforcement of environmental regulations, and a tenure regime that encouraged deforestation as proof of productive occupation (Walker et al., 2009). Meanwhile, the expansion of the intensive frontier, i.e., the development of an intensive agro-industrial economy in the Amazon, has involved the development of a new set of multi-level governance arrangements favoring deforestation control and agricultural intensification, as described in the Background section above. The Amazon transition can therefore be conceptualized as a regional shift from a deforesting, extensive political economy towards a ‘green,’ intensive political economy. This consolidation of a regional space of ‘green’ agro-industry has both economic and political dimensions, which are co-constituted, but which do not evolve with perfect spatial congruence. In other words, the political and economic dimensions of the intensive frontier develop through distinct though interconnected geographies. For example, whereas the municipality of Sorriso is the epicenter of soy production in north-central Mato Grosso, the neighboring municipality of Lucas do Rio Verde became the epicenter of environmental governance. Thus, rather than speaking of ‘frontier governance,’ which ignores the political dimensions of frontier formations, we propose the concept of a ‘governance frontier’ to capture the particular spatial relations through which governance formations develop and expand. This governance frontier is one component of the multidimensional intensive political-economic frontier. Lastly, our findings lead us to question the distinction between ‘hard’ and ‘soft’ governance deployed by Brannstrom et al. (2012). Neoliberal governance practices elide the already artificial distinction between state and society (cf. Mitchell, 1991), such that governance becomes the irreducibly ‘hybrid’ product of interactions among farmers’ and ranchers’ organizations, social movements, NGOs, corporations, and government. We use our empirical findings to elaborate on these arguments in the Discussion.

This article explores the development of the Amazonian governance frontier through a focus on municipalities that have become models of environmental governance. Brazil is a federation with three levels: the nation, states, and municípios or municipalities. In 2008, the federal government created a ‘priority list’ of municipalities for combating deforestation in the Amazon. The list singled out 36 high-deforestation municipalities for enhanced environmental monitoring and enforcement, and stipulated criteria for exiting the list that included reductions in deforestation and a review of private land titles. Due especially to the priority list, the municipal level has been a key focus for environmental governance measures over the past decade. Two municipalities have become particularly prominent models of ‘green’ municipal governance and development: Lucas do Rio Verde (Lucas) in Mato Grosso and Paragominas in Pará. A project in Lucas was instrumental to the creation of CAR (Cadastro Ambiental Rural), the Rural Environmental Registry that facilitates landowner compliance with the Forest Code. A project in Paragominas helped it become the first municipality to exit the priority list. The governance models developed in these municipalities have been widely promoted and replicated. CAR became a national requirement under the revised Forest Code passed in 2012, and Paragominas inspired a state-wide Green Municipalities Program in Pará that now includes three-quarters of the municipalities in the state. These models of municipal governance did not emerge independently of each other. They are connected, not least by the involvement of The Nature Conservancy (TNC), a major international environmental NGO, in both Lucas and Paragominas. To illuminate these connections and the geography of policy evolution and transfer in Amazonian municipal environmental governance, we use a follow the policy approach inspired by Peck and Theodore (2012). This method links global ethnography (Burawoy, 2001; Hart, 2006) and critical policy studies under the idea that policy designs are complex social constructions developing through multi-sited webs of experimentation, emulation, and evolution. As Peck and Theodore affirm: ‘‘It may be commonplace to represent policy models as rational abstractions – essentializing some supposedly definitive cluster of design features or packaging some programmatic philosophy – but in fact these are coconstituted through the networks, and across the landscapes, over which they travel. . . .[T]hrough their very movement they (re)make connections between these sites, evolving in form and effect as they go” (2012, 23 original emphasis). They call for: ‘‘reconceiving the spatial and scalar relations between research sites as conjunctural nodes within three-dimensional webs of relations (scalar  spatial), across which transformative processes operate, evolve, break down, trigger countervailing forces, and so on. . . This involve[s] dispensing with preconceived assumptions regarding the (linear, radial, or top-down) sequencing of policy invention and emulation, in favor of an understanding of policy ‘reinvention’ as a continuous, multisite process. . .” (2012, 27 original emphasis). By ‘following the policy’ in the Amazon, we provide a perspective on the spatiality and social processes of the expanding green governance frontier. Our research followed the evolution of municipal governance through the tropical forest conservation programs of The Nature Conservancy, informed by an organizational ethnography conducted within these programs by the first author. Numerous other

64

G.M. Thaler et al. / World Development 114 (2019) 59–72

policy pathways exist and could be followed to illuminate other aspects of the development of municipal governance. Nonetheless, by following TNC we capture key places and events in regional policy evolution. We focus on a 10-year period from 2004 to 2014, which leads us from the municipality of Santarém, at the mouth of the Tapajós River in western Pará, to Lucas in Mato Grosso, to Paragominas in northeastern Pará, and lastly to São Félix do Xingu (São Félix) in southeastern Pará (Fig. 1). This article is based on original fieldwork in all four of these municipalities, comprising 120 semi-structured interviews, participant observation at events and NGO activities, and the collection of documentary materials including NGO reports, news stories, and policy documents. Interviewees included farmers and ranchers; representatives of community associations and rural syndicates; government officials at the municipal, state, and federal levels; and staff of domestic and international NGOs. Interviewees were asked about their experiences with and perspectives on local environmental and agricultural change. We made brief field visits to Santarém (January– February 2011, second author) and Lucas (April–May 2014, first author), and we rely more on secondary accounts for those municipalities. We conducted more extensive fieldwork in Paragominas (February–June 2011/June 2012, second author) and São Félix (January–March 2014/July 2014, first author). The organizational ethnography of TNC in Brazil was conducted from September 2013 to August 2014. The next section presents our empirical account of policy evolution across these four municipalities. 6. Following municipal environmental governance 6.1. From Santarém. . . The Nature Conservancy is a US-based environmental NGO founded in 1951. It began working in Brazil in the late 1980s, focusing initially on protected areas management. By the end of the 1990s, TNC staff felt that Brazilian protected areas had become more consolidated, and they wanted to shift to thinking about development pressures, in particular deforestation driven by soy cultivation and cattle ranching.9 In 2003, multinational commodities trader Cargill opened a soy port on the Amazon River in Santarém, which primarily exported soy grown in Mato Grosso and Rondônia, but also drove the expansion of industrial soy cultivation in the area around the city (Greenpeace, 2006a; Steward, 2007). Industrial soy expansion is typical of the Amazonian intensive frontier, which today generally succeeds the extensive deforestation frontier, occupying previously cleared areas and displacing smallholders and extensive ranchers to the forest margin. In the early 2000s, however, soy was expanding rapidly into primary forest areas (Morton et al., 2006). The soy boom around Santarém materialized through processes of deforestation and land grabbing similar to those in the ranching sector (Baletti, 2014; Greenpeace, 2006a), and this rapid frontier development prompted the local Rural Workers’ Union to launch a campaign against soy production and to discourage smallholders from selling land to soy farmers (Steward, 2007). On a vacation in Santarém at this time, a TNC manager noticed clearing for soy in the area around the Cargill port. He called the TNC Director in Minnesota, where Cargill is headquartered, who put him in touch with Cargill’s Sustainability Department. A Cargill representative flew to Brazil to visit the area with the TNC manager, who warned him that environmental activists would likely target Cargill for this deforestation.10 In 2004, TNC began working 9 This account of the Responsible Soy project in Santarém is based primarily on an interview with a TNC manager on July 13, 2015, corroborated by interviews with other TNC employees, TNC and Cargill documents (Cargill, n.d.; Dias, 2013; TNC, 2017), and secondary literature (Adams, 2015; Baletti, 2014; Steward, 2007). 10 Interview with TNC manager, July 13, 2015.

with Cargill to set up a monitoring system for Santarém and the neighboring municipality of Belterra to ensure that Cargill’s local soy purchases were not driving deforestation, an effort that became the ‘Responsible Soy’ project.11 There was internal debate at Cargill whether to pursue ‘zero deforestation’ or ‘zero illegal deforestation,’ where forest clearing could still be allowed as long as a property was in compliance with Forest Code protections. The company initially decided to pursue ‘zero illegal deforestation,’ i.e., compliance with the Forest Code. In 2006, however, environmental NGO Greenpeace launched a campaign targeting Cargill for driving deforestation through soy purchases (Greenpeace, 2006a), with actions including a blockade of Cargill’s port (Greenpeace, 2006b). In response, Cargill assembled the other major soy traders in cooperation with environmental NGOs to agree to a regional Soy Moratorium against the purchase or trade of soy produced in areas of the Amazon Biome deforested after July 2006.12 TNC’s existing monitoring efforts with Cargill in Santarém provided a template for implementing the Soy Moratorium, with the individual soy-producing property as the unit of governance. Thus, although Brannstrom et al. (2012) divide ‘soft’ governance into ‘compliance’ and ‘market exclusion’ categories, the Soy Moratorium, an example of market exclusion, was based on the collaboration between TNC and Cargill, which began as a compliance effort. While the Responsible Soy project and the Soy Moratorium exclude new deforestation from the soy commodity chain within the Amazon Biome, they leave other governance issues unresolved.13 Soy producers may have deficits of legal reserve or permanent preservation areas (áreas de preservação permanente – APPs) and may lack legal title to their properties, but so long as they do not clear new land for soy, they can continue to sell to Cargill and other multinationals. Baletti reports that the Santarém soy farmers, Cargill, and TNC attempted to use their Responsible Soy initiative to ease Forest Code regulations and tenure regularization for participating farmers. In February 2006, they submitted a proposal to state agencies: ‘‘Considering the ‘relevance and social and economic impact of [soy] agriculture production in the national and regional level’, they argued the agencies should facilitate soy development by regularizing the property rights of all soy farmers participating in the TNC program, reducing all of their existing fines to 10% of their economic value, accepting geographic and production data provided by TNC as ‘official’ data, and expediting authorizations for deforestation.” (Baletti, 2014, 16) The agencies refused to make these concessions. Nonetheless, the emergence of coordinated property-level environmental monitoring in Santarém by non-governmental actors and its regionalization under the Soy Moratorium established the Santarém experience as a model of environmental governance in the Ama11 In developing its Santarém model, TNC drew on its experience with property registration in Mato Grosso, where in 2000 the World Bank helped set up the System for Environmental Licensing of Rural Properties (SLAPR), a system for managing property-level environmental licensing, monitoring, and enforcement related to the Forest Code using remote sensing and GIS technology (interview with TNC employee, October 08, 2013). ‘Following the policy’ risks infinite regress. We begin our narrative in Santarém because it allows us to examine how the interaction of different regulations, technologies, and actors in a specific landscape drove the emergence of a municipal-level policy model. 12 After passage of the 2012 Forest Code, the reference date for the Soy Moratorium was shifted to July 2008. 13 One of the most important impacts of the Soy Moratorium may have been the development of a new governance model. The environmental impacts of the Soy Moratorium are complicated to evaluate. Prior to the Moratorium, soy was a minor but growing direct driver of Amazon deforestation (Morton et al., 2006). The Moratorium deterred further forest conversion for soy in the Amazon Biome, but may have displaced soy expansion to the Cerrado (Gibbs et al., 2015).

G.M. Thaler et al. / World Development 114 (2019) 59–72

65

Fig. 1. Model Municipalities and Deforestation in the Brazilian Amazon. Forest is shown as percentage of tree cover (vegetation >5 m in height) and forest loss is defined as stand-replacement disturbance (Hansen et al., 2013). Area shown corresponds to the Legal Amazon plus eastern Maranhão.

zon. Of the 29,672 ha of soy planted in Santarém and Belterra in 2016 (IBGE, 2018b), just 189 ha (0.6 percent) was planted in violation of the Soy Moratorium (GTS Soybean Working Group, 2016).14 6.2. . . . to Lucas do Rio Verde. . . At about the same time as the Soy Moratorium came into effect, the mayor of Lucas, who was also the owner of Fiagril, an important regional agricultural conglomerate, approached TNC. In the 1980s to 1990s, Lucas had grown into a major soy production center in northern Mato Grosso. Lucas lies predominantly in the Cerrado (savanna) biome as opposed to the Amazon (humid forest) biome, but it is located within Brazil’s Legal Amazon region, and the mayor was concerned to protect Lucas from the association of Amazonian soy with forest destruction. Aware of TNC’s compliance work with soy farmers, the mayor forged a partnership for a municipal-level property registration and compliance project called Lucas Legal or ‘Legal Lucas’ (Ferreira, 2010; Rausch, 2013). The Brazilian agri-food corporation Sadia, which operates a pork and poultry processing plant in Lucas, and agricultural inputs multinational Syngenta joined Fiagril in sponsoring the project. The municipal rural producers’ syndicate, Ministério Público Estadual (State Public Ministry),15 Rio Verde Foundation (a local agricultural development body), and State Environmental Secretariat (SEMA) joined as partners. The first step toward achieving environmental compliance was creation of a land-use map to register individual properties and map APPs and legal reserves at the farm level. Unlike in Santarém, where only soy farms were monitored, Legal Lucas included all 14 ‘Zero deforestation’ in the soy sector, however, does not mean zero deforestation in Santarém. Although long-settled and not an area of active frontier expansion, the municipality lost 356km2 of primary forest in 2007–2017, while the Soy Moratorium was in effect. 15 The Public Ministry in Brazil is an independent body of government prosecutors who bring cases in the public interest.

rural properties in the municipality, making the municipality a functional level of environmental governance. The project paid the costs of mapping, increasing the attractiveness of participation for producers. The second stage of the project would involve reforestation or compensation of deficit areas to achieve full Forest Code compliance. Mato Grosso had since 2000 been implementing a System for Environmental Licensing of Rural Properties (SLAPR) through which landowners could register property boundaries, pay fines for prior illegal deforestation, sign an agreement to reforest areas, and obtain authorization for new clearing if permitted (Azevedo, 2009). Legal Lucas involved SEMA as a partner with an understanding that the agency would not prosecute producers who had illegally deforested but were seeking compliance through the project. That understanding did not hold with the federal environmental agency (IBAMA), and in September 2008, IBAMA fined 15 property owners in Lucas. This action triggered intense lobbying by municipal leaders, TNC, and SEMA, culminating in a state law that officially created CAR, breaking environmental licensing into stages. CAR comprised a registry of geo-referenced property boundaries, including APP and legal reserve areas and existing land cover, allowing for identification of forest deficits. Producers who completed CAR were granted a grace period to bring their properties into compliance without being fined (Rausch, 2013, 263–64). By 2007, all rural properties in Lucas were registered with the project, though only around 20 percent were registered with the state through SLAPR. After the creation of CAR, many more producers completed state registration. The municipal government mandated that producers register, and Sadia resolved to purchase only environmentally-compliant production. Producers also started to isolate degraded APPs, and by 2014 over 60 percent of Lucas’ APP deficit was under restoration.16 CAR has become a key component of the policy framework for ‘green’ agro-industrial development in the Amazon, but it is diffi-

16

Interview with TNC manager, April 22, 2014.

66

G.M. Thaler et al. / World Development 114 (2019) 59–72

cult to attribute deforestation reductions to CAR itself, and most registered landholders in the region appear not to be meeting their forest restoration requirements (Azevedo et al., 2017). Moreover, even as Lucas became a model for Forest Code compliance, it was also a crucible for discussions about how to make compliance less demanding. To comply with the Forest Code, producers in Lucas had to surrender valuable agricultural land for restoration, giving impetus to proposals such as allowing APPs to be included in legal reserve calculations and allowing legal reserve deficits to be compensated through a broader range of off-property options.17 These ruralista18 proposals gained credibility when supported by environmentalist actors such as TNC within environmental governance projects such as those in Santarém and Lucas. These modifications were included in the revised 2012 Forest Code, which reduced the environmental debt of rural properties in Brazil by 29 million hectares, forgoing restoration of an area the size of Italy (Soares-Filho et al., 2014). The wall-to-wall municipal-level approach piloted in Lucas, the creation of CAR as an environmental registration system, and the Legal Lucas model of multilevel government-NGO-corporate-land owner partnership have all proven hugely influential for Amazonian environmental governance. As one municipal environment official opined, ‘‘TNC’s projects today are basically copies of Legal Lucas,”19 and a TNC regional manager concurred that ‘‘Lucas is our laboratory. Everything that is new we start out in Lucas.”20 Following creation of the Federal Government’s list of priority municipalities in 2008, TNC, other NGOs, and the Brazilian government have targeted numerous environmental governance projects to the municipal level, usually working through a multi-stakeholder partnership model. CAR has also been widely replicated and scaled up, thanks in particular to the transfer of the Lucas model to Paragominas. 6.3. . . . to Paragominas. . . Paragominas was one of the original municipalities placed on the federal priority list.21 The former logging boomtown was developing during the 2000s as a hub of intensified ranching and grain production. After the creation of the Soy Moratorium, elite landowners in Paragominas, who were aware of TNC’s work in Lucas and Santarém, had met with TNC representatives in 2007 to discuss replicating the Lucas model in their municipality.22 Concomitantly, members of this landowning elite ran for leadership of the municipal rural producers’ syndicate (SPRP) on a platform promoting environmental compliance, hoping to replicate Lucas’ success in attracting agricultural investments. In January 2008, Paragominas entered the priority list. Federal enforcement actions shuttered sawmills, confiscated illegal timber and charcoal, and imposed fines and credit restrictions on private landowners. These sanctions hamstrung the rural economy. In response to the economic crisis and civil unrest (including the burning of government vehicles) (Viana et al., 2016), the mayor convened local organizations to commit to a ‘pact for zero deforestation,’ garnering 51 signatories including government bodies, rural producers’ unions, and community associations. The mayor also negotiated with the federal government to loosen the property registration requirements of the priority list, using the example of Lucas to convince the Ministry of Environment (MMA) to accept

CAR as a form of registration instead of a full revision of land titles.23 The new regulation, published in March 2009, stipulated that priority municipalities must register 80 percent of their private property area through CAR to exit the list (MMA, 2009). Under the rubric of the ‘Green Municipality’ project, the municipal government and landowning elite partnered with TNC and Imazon (a Brazilian environmental NGO) to obtain technical support. TNC worked with SPRP to convince medium and large landowners to complete CAR, with TNC subsidizing registration costs. As in Mato Grosso, landowners feared that by registering they would be subject to fines for past illegal clearing. Following the Lucas model, the municipal government negotiated a grace period with the Ministério Público Federal (Federal Public Ministry – MPF) so landowners completing CAR would not automatically be fined. Beginning in 2009, the project received resources from the Vale Sustainable Development Fund (Fundo Vale), which later expanded its operations into a Green Municipalities Initiative, based on the Paragominas example and active in over a dozen Amazonian municipalities (Fundo Vale, 2012). In 2010, Paragominas became the first municipality to exit the priority list. The Lucas model of municipality-wide CAR registration now became the Paragominas model of CAR and municipal anti-deforestation pacts, and this model was heavily promoted by government, NGOs, and Paragominas elites (Guimarães, Veríssimo, Amaral, & Demachki, 2011). ‘‘Paragominas was an advance in the methodology,” a TNC manager asserts, viewing the addition of anti-deforestation pacts as an evolution of the organization’s municipal governance approach.24 In 2011, the Pará State Government launched a Green Municipalities Program to combat deforestation and promote municipal land use management based on the Paragominas model. Other state programs have been made conditional on adhering to this program, and 121 out of 144 Pará municipalities have joined. While Paragominas has been promoted as a model of participatory environmental governance, the pact and the Green Municipality project did not include all social groups equally. The pact resulted from an agreement among political and economic elites without substantial civil society consultation.25 CAR efforts targeted the largest properties to meet the criterion of registering 80 percent of private property space as quickly as possible, and the project began registration of small properties only in 2012. Charcoal production, formerly a major source of income for smallholders, was virtually eliminated under the new environmental regime (Carmenta, Coudel, & Steward, 2018), while initiatives supporting smallholder production have been sparse (Viana et al., 2016). As one smallholder recalls: ‘‘I went to the CAR launch meeting at the end of 2010. They told smallholders not to deforest anymore; they said that farmers should work differently, without burning, but I still wonder how we are going to do this. There is one tractor, but for 1000 families. If it takes one day on each property, even in three years it won’t finish.”26 Another smallholder elaborates: ‘‘It is more difficult for smallholders to stop deforesting, to stop making charcoal. Even I deforested five hectares last year. I deforested right away, because I was afraid it would be more difficult later on with enforcement. We have to be better trained not to deforest. The government has not prepared anyone.”27

17

Interview with municipal environment official, May 02, 2014. Ruralista, ‘ruralist,’ refers to the agribusiness and landowners’ caucus in the Brazilian Congress and to large landowner political interests generally. 19 Interview with municipal environment official, May 02, 2014. 20 Interview with TNC manager, April 22, 2014. 21 Lucas, located primarily in the Cerrado, and Santarém, a more substantial urban center, were never placed on the list. 22 Interviews with SPRP leader, March 28, 2011, and TNC manager, November 11, 2013. 18

23 24 25 26 27

Interview Interview Interview Interview Interview

with with with with with

former Ministry of Environment official, April 09, 2014. TNC manager, November 11, 2013. TNC manager, November 11, 2013. family farmer, April 06, 2011. family farmer, April 01, 2011.

G.M. Thaler et al. / World Development 114 (2019) 59–72

In contrast to the lack of investment in smallholder agriculture, NGOs, foundations, and the municipal government continue to invest in legalizing and consolidating intensive agro-industrial production on large properties. Following Paragominas’ exit from the priority list, for example, Fundo Vale and Dow AgroScience corporation supported a US$1.4 million Green Cattle Program to promote ranching intensification on six pilot properties. Piketty et al. (2015) indicate that mechanized agriculture and intensive ranching will likely be limited to favorable soils with good road access, in areas already under elite control, reinforcing unequal experiences of environmental governance and agricultural intensification in the municipality. Nonetheless, the Paragominas model, evolving out of the models pioneered in Lucas and Santarém, offered a powerful example of the reconciliation of environmental governance with profitable agro-industry on an active deforestation frontier. In the words of Mauro Lúcio Costa, former president of SPRP, on a TNCsponsored visit to São Félix in 2012, ‘‘No one should be different from Paragominas” (O Estado de São Paulo, 2012). 6.4. . . . to São Félix do Xingu In São Félix, TNC sought to transfer the CAR-and-pact municipal governance model to the extensive deforestation frontier. Whereas in Paragominas, Lucas, and Santarém, governance projects were linked with a significant soy sector, São Félix lay beyond the intensive frontier, a land of smallholder farmers and large-scale extensive ranchers. At 84,213 km2, São Félix covers a vast section of the Xingu River basin, and in the early 2000s, municipal deforestation rates exceeded 1000 km2/year. State and federal environmental governance efforts targeted São Félix in the mid-2000s, when the federal government created two new strictly-protected conservation areas overlapping the municipal territory, and the remaining unprotected area west of the Xingu River was declared a sustainable-use conservation area (APA) by the State of Pará. These measures largely halted westward expansion of the extensive frontier in the municipality and helped to reduce deforestation from 1408 km2 in 2005 to 878 km2 in 2007. Rapid deforestation for small farmer colonization and extensive ranching continued in the APA and outside of conservation areas and indigenous territories, however, and São Félix entered the priority list in 2008 as the municipality with the highest total annual deforestation. This status as ‘‘deforestation champion” made São Félix a particular focus for additional anti-deforestation efforts (Schmink et al., 2017; Thaler, 2017, 1438). It was at this point that TNC began to develop a program in São Félix. Whereas TNC’s projects in Paragominas and Lucas were undertaken partly on the invitation of local actors, São Félix was strategically targeted by the NGO as a demonstration site for Reducing Emissions from Deforestation and Forest Degradation (REDD). TNC first experimented with REDD in Bolivia in the 1990s, but the inclusion of REDD in the United Nations Framework Convention on Climate Change (UNFCCC) negotiations in 2007 touched off a new wave of REDD experimentation worldwide. A 2008 TNC feasibility study identified São Félix for a REDD pilot in the Amazon,28 and TNC arrived in the municipality in 2009, at a time when many ranchers were still bitterly resistant to environmental regulations. A member of the ranchers’ syndicate recalls the first meeting between the syndicate, TNC, and government agencies that April: ‘‘The reception [of the outside organizations] was terrible. People hung up banners [protesting environmental enforcement]. We had a very hostile meeting of the syndicate. Two months 28

Interview with former TNC manager, August 15, 2014.

67

later there was a full embargo of São Félix’s production.29 When producers realized they could not sell their production, they came to the syndicate to see what to do, and we explained that we would have to bring all the groups together for another meeting. We were number one on the priority list: radical consequences are going to come. In August 2009, we organized another meeting of all the institutions. People were afraid to attend after the poor reception at the first meeting, but the second meeting was good. From there we created partnerships and began to discuss what to do.”30 The REDD pilot that TNC launched in São Félix was far broader than the Responsible Soy, Legal Lucas, or Green Municipality projects, eventually including support for monitoring deforestation, management of indigenous territories, ranching intensification, and smallholder cacao production.31 Nonetheless, TNC’s work in São Félix intersected with projects launched by IEB (a Brazilian NGO) in 2009 and by MMA in 2011, with a common strategy centered on the Paragominas model of CAR registration and a municipal anti-deforestation pact. Thanks to Paragominas, completion of CAR became a requirement for exiting the priority list, and MPF began to require municipalities to agree to a municipal ‘pact for the control of deforestation’ in order to extend licensing deadlines for rural producers. These mandated pacts became a condition for inclusion in the statewide Green Municipalities Program, also based on the Paragominas model. In December 2010, the São Félix municipal government signed an agreement with MPF that required a municipal pact. The mayor’s office convened a working group including NGOs, producers’ groups, and the Municipal Environment Office. Here arose the first major modification to the Paragominas model. In many Pará municipalities, the pact is effectively rubber-stamped, without local engagement in negotiating its content. In the first working group meeting, however, participants called for a Commission to facilitate pact development and implementation (Cavalcante, Maciel, & Carvalho, 2013b).32 The resulting Commission expanded to include over 20 entities.33 The Commission organized meetings in rural areas of São Félix and in the municipal seat to raise awareness about environmental regulations and solicit input on commitments from the state, civil society organizations, and rural producers that would enable the decoupling of agricultural production and deforestation. These commitments were incorporated into the pact, signed in August 2011, and became the basis for a ‘post-pact agenda’ implemented by the Commission. The creation of a multi-stakeholder commission to accompany the pact and the consultation process that preceded the pact’s signing were significant innovations, as the pact became an organized endeavor involving civil society.34 Following promulgation of the pact, however, the Commission was largely driven by the NGOs and MMA, with little effective civil society participation and little municipal government leadership (Cavalcante et al., 2013a). 29 This ‘embargo’ refers to heavy enforcement actions under the federal government’s Operação Arco de Fogo (Operation Arc of Fire). 30 Interview with former ranchers’ syndicate official, February 04, 2014. 31 Though conceived as a REDD project that might attract carbon finance, the TNC project in São Félix never marketed carbon credits, and TNC eventually abandoned the REDD label, referring to its project as a ‘‘green development” or ‘‘sustainable landscapes” initiative (interviews with TNC manager, November 11, 2013 and July 17, 2014). 32 Interview with TNC employee, January 31, 2014. 33 The nine members who regularly participated in Commission activities were: the municipal environment and agriculture offices, federal Ministry of Environment, federal Indian affairs agency (FUNAI), Pará agricultural extension agency (EMATER), managing council of the Triunfo do Xingu environmental protection area (APA), ADAFAX (a local NGO), IEB, and TNC (Cavalcante, Maciel, & Carvalho, 2013a). 34 Interview with TNC employee, January 31, 2014.

68

G.M. Thaler et al. / World Development 114 (2019) 59–72

Even as the pact was signed, substantial progress had been made on CAR registration. In 2010, TNC sent one of its most experienced technicians, who had worked on Legal Lucas, to São Félix to advance CAR. TNC focused on registering large properties in collaboration with the Municipal Environment Office, while MMA began a project in 2011 to complete CAR for smallholders. By November 2011, São Félix had more than 80 percent of its non-protected areas registered in CAR. During this period of priority government enforcement, CAR registration, and the pact process, deforestation in São Félix declined rapidly, from 761 km2 in 2008 to 140 km2 in 2011. New governance measures, while effective in reducing deforestation, also caused economic activity to stagnate. Unlike on the intensive frontiers in Santarém and Paragominas, São Félix lacked intensification pressure from expansion of the soy supply chain, which elsewhere serves to raise land values and promote capital investments and pasture-to-cropland conversion (Richards, Walker, & Arima, 2014). In the extensive ranching and smallholder economy, large ranchers may have capital to intensify, and some in São Félix have begun to adopt improved pasture management. For many small and medium producers, however, environmental governance produced economic difficulties, and informants reported an accelerating tendency for smallholders to sell their properties due to economic stress. Real GDP per capita declined in São Félix following the 2008 governance interventions, suggesting that residents were economically worse-off in 2010 than they were in 2007 or even 2000, while in Paragominas, Lucas, and Santarém, 2010 GDP per capita remained substantially higher than a decade earlier (Fig. 2).35 In response to the stagnation of the extensive frontier economy in São Félix and continuing annual deforestation rates exceeding the 40 km2/year mark necessary to exit the priority list, TNC pursued two important modifications of the Paragominas model: First, TNC began to invest directly in intensifying agricultural production in São Félix. In 2013, the organization launched a sustainable cacao project to support smallholders to diversify production and reforest degraded areas with cacao agroforestry (TNCBrazil, 2013).36 In a second project, TNC worked with medium and large ranchers to raise stocking rates and rehabilitate degraded pastures (TNC-Brazil, 2016). In 2016, Walmart began marketing ‘‘deforestation-free, climate-smart” beef sourced from this project (Moore Foundation, 2016), and TNC supported the municipal government to launch a Municipal ABC Plan aimed at attracting training and credit from the federal Low-Carbon Agriculture (ABC) Program. Second, TNC collaborated with the municipal government and ranchers’ syndicate to petition the federal government to make the deforestation threshold for exiting the priority list proportional to the area of the municipality, so a large municipality like São Félix need not reduce deforestation to the same level as Paragominas, a municipality less than a quarter its size. MMA has refused this change, and actors in São Félix have had difficulty maintaining political pressure for the amendment.37 The impact of governance efforts on deforestation in São Félix is undeniable. After São Félix’s inclusion in the priority list, municipal deforestation fell over 80 percent in 2008–2011, and despite rebounding, 2016 deforestation was nearly 60 percent below its 2008 level. These reductions are most directly attributable to earlier protected area creation, as well as to continued federal envi35 Due to rapid demographic change on Amazonian frontiers, official population estimates are often inaccurate. We therefore report GDP per capita figures only for the years of the 2000 and 2010 censuses and 2007 population count. 36 Cacao production in the municipality has been supported by other organizations as well, notably the federal agency CEPLAC (Executive Commission of the Cacao Production Plan), national NGO Imaflora, local NGO ADAFAX, and local cooperative CAMPPAX (Mixed Alternative Cooperative of Small Producers of the Upper Xingu). 37 Interview with former ranchers’ syndicate official, February 04, 2014.

Fig. 2. Municipal GDP per capita in constant Brazilian reais. Municipal GDP per capita for year 2000 is in current reais; subsequent years are adjusted using the annual average consumer price index. Calculated from IBGE (2016, 2017, 2018).

ronmental enforcement, which was facilitated by CAR registration. The Santarém-Lucas-Paragominas policy model we have traced was seen as successful in those municipalities not so much due to direct reductions in deforestation, however, as to the construction of a governance framework for ‘green’ agroindustrial production. In São Félix, this governance model was imposed prior to the economic transformation of the intensive frontier. During the painful period of economic adjustment that followed, the effort to revise priority list criteria failed and TNC’s investments in intensification were able to reach only a small proportion of the population. Notwithstanding this mixed experience, TNC always conceived of its São Félix intervention as a pilot project to further policy evolution and serve as a model for replication. As TNC’s local manager in São Félix observes, ‘‘As environmental compliance evolves, TNC creates pilots to support the state in developing policies.”38 The elevation of São Félix to a model was thus virtually inevitable, and is justified by TNC in terms of its major reductions in deforestation, large area of CAR registration, inclusive pact process, and ranching and cacao demonstration projects. TNC’s 2015 Global Lands Report features a quotation from Wilson Batista of the São Félix ranchers’ syndicate, in a section with glossy photos of a Santarém soy farmer and Paragominas cattle herder: ‘‘We can be a model,” Batista says of São Félix, ‘‘not just for Brazil, but for the world” (TNC, 2015, 42).39 7. Discussion The policy pathway that we have followed (Fig. 3) from Santarém to Lucas to Paragominas to São Félix, focusing on the decade of 2004–2014, is just one dimension of the development of environmental governance and agro-industrial expansion in the Brazilian Amazon, but it captures key events and dynamics within more general processes. We draw several lessons from ‘following the policy’ of TNC’s municipal governance programs. 38

Interview with TNC local manager, February 17, 2014. On a subsequent field visit in June 2018, the first author found that TNC is expanding its cacao and ranching projects in eastern Pará, but the organization is moving away from a municipal-level focus in favor of a supply-chain approach. Medium and large ranchers in São Félix continue to invest in intensification, and some have begun mechanized grain cultivation for cattle feed. Small farmers receiving support from TNC and other entities have diversified their production and improved their livelihoods, but smallholders residing farther from the municipal seat are less likely to receive assistance and have continued to struggle or sell their properties. Especially west of the Xingu River, smallholder communities have been depopulated while large ranches grow larger. These recent developments will be discussed in detail in future work. 39

G.M. Thaler et al. / World Development 114 (2019) 59–72

69

Fig. 3. Policy pathway of Amazonian environmental governance. This multilevel chronology illustrates the web of policy development narrated in the article. Event time placement is approximate. Arrows indicate the most direct causal connections between events.

First, the distinction between ‘hard’ and ‘soft’ governance is empirically untenable and thus of limited analytical utility. Twenty-first century environmental governance in the Brazilian Amazon has been developed and implemented largely within a heterogeneous complex of state, agribusiness, social movement, and NGO actors (Campos & Nepstad, 2006; Schwartzman et al., 2010; Thaler, 2017). A ‘hard’ government regulation such as CAR, for example, emerged from the ‘soft,’ ‘multi-stakeholder’ initiative of Legal Lucas (which itself was premised on the ‘hard’ Federal Forest Code), and CAR implementation has likewise involved ‘soft’ cooperation among government, NGOs, corporations, and rural syndicates. A ‘soft’ governance initiative such as the Soy Moratorium, meanwhile, has relied on state monitoring capacity and biome boundaries and evolves dialectically with ‘hard’ government regulation such as the 2012 Forest Code revision. The distinction between ‘hard’ and ‘soft’ governance mirrors the distinction between ‘state’ and ‘society,’ which, as Mitchell observes, has always been ‘‘a line drawn internally, within the network of institutional mechanisms through which a certain social and political order is maintained” (1991, 90). Neoliberalization implies the blurring of even this internal line, such that both empirically and analytically, neoliberal authority is never purely hard or soft, but is always ‘hybrid’ to begin with.40 Second, ‘frontier governance’ is more fruitfully conceptualized as the expansion of a governance frontier. The governmental order that promotes and reproduces ‘green’ agro-industry emerges through processes of experimentation, emulation, and evolution by heterogeneous actors at particular interconnected spatial and temporal nodes. The model municipalities in our study constitute nodes of policy evolution that have been central to the spatial expansion of the governance frontier, a form of translocal connection in the development of frontier spaces that does not register in the ‘resource frontier’ perspective. This governance frontier is co-produced with other frontier socio-economic and ecological formations of an intensive political-economy, such as mechanized soy agriculture and intensive ranching. Governance stabilizes and regulates this organization of production, but the multiple political-economic dimensions of the intensive frontier are not always perfectly superimposed. These dimensions are dynamically

40 To paraphrase Swyngedouw (1996), governance has no ‘hard’ or ‘soft’ ontological or essential foundation in ‘state’ or ‘society,’ but rather the process of hybridization has ontological and epistemological priority.

inter-related with each other and with other political-economic projects, such as socio-environmental livelihoods and activism of indigenous peoples and smallholders, or large-scale deforestation and land-grabbing by extensive ranchers. There is thus a specific geography of the governance frontier that is related but not identical to the geography of the frontier of agricultural intensification. Third, each of the nodes of policy evolution that we examine becomes a municipal model through what Brazilians call the flexibilização (flexibilization)41 and the legitimation of the governance regime.42 Thus, in Santarém, TNC’s monitoring project allowed for the Soy Moratorium as a flexibilization or resolution of Greenpeace’s campaign against Cargill and blockade of the Cargill port. TNC, Cargill, and the soy farmers then unsuccessfully petitioned the government for a flexibilization of Forest Code enforcement based on their participation in the Responsible Soy project. In Lucas, CAR flexibilized SLAPR to allow landowners to register their properties without being fined. In Paragominas, CAR flexibilized the priority list to avoid a review of municipal land titles. In São Félix, there was an unsuccessful attempt to flexibilize the priority list further by changing the 40 km2/year deforestation limit to be proportional to municipal area. Each successful flexibilization facilitates compliance at a cost of virtual amnesty, forgoing fines and legitimizing land grabs, at the same time as it transforms illegal deforesters and land grabbers into ‘responsible producers’ in ‘model municipalities’ (cf. Azevedo, 2009). Model municipalities are thus also crucibles of legitimation, consolidating and validating agro-industrial expansion.43 In sum, the governance frontier in the Amazon develops through the negotiation and contestation of a policy framework that legitimizes, stabilizes, and reproduces an agro-industrial economy. This argument has been made before with different framings (Baletti, 2014; Viana et al., 2016). With our empirical data from ‘following the policy’ and our theoretical interpretation of the governance frontier, we contribute a better understanding of the specific political geography of this process of frontier development.

41 Roughly equivalent would be the ‘relaxation’ of the governance regime, but ‘flexibilization’ is more euphemistic. 42 Legitimation as a key dynamic of frontier expansion is discussed also by Rasmussen and Lund (2018). 43 At the same time, these models demonstrate the agility of a set of producers and corporations who not only participate in the contestation and construction of policies, but also alter their practices and make investments as protagonists of agro-industrial ‘green development.’

70

G.M. Thaler et al. / World Development 114 (2019) 59–72

8. Conclusion

References

Brazil’s Amazonian transition since 2004 of reduced deforestation and intensified agricultural production has been attributed to processes of ‘frontier governance,’ but this formulation depoliticizes the frontier and despatializes governance. We argue for a reinterpretation of frontiers and governance through a theorization of the governance frontier, and we use a ‘follow the policy’ approach to provide a view of the evolution of a governance frontier in the Brazilian Amazon. This governance frontier expands through a distinct geographical configuration of nodes of policy experimentation, legitimation, and transfer rooted in Amazonian municipalities. The governance frontier comprises one dimension of the Amazonian transition towards a regional political economy centered on intensive agro-industry, developing alongside related dimensions of socio-economic and ecological change. These findings help to identify several promising pathways for future research. First, we have described one policy pathway in this article, but other pathways could be mapped through other NGOs, producers’ organizations, and social movements working at the regional level. Following other policy pathways would further enhance our understanding of the spatial and temporal dimensions of the Amazonian policy process. Second, additional research is needed on the spatial entanglements of the governance frontier with other frontier dimensions. Can the Paragominas model of municipal environmental governance be effective in a place where it precedes the establishment of agro-industrial socio-economic relations? São Félix remains a critical case for answering this question as a node where the governance model was deployed over high levels of deforestation and extensive ranching absent conversion pressures from field crops. Just as soy expansion in northern Mato Grosso catalyzed governance expansion in those areas, is it possible that governance expansion in São Félix will catalyze agro-industrial intensification? TNC projects for cacao production and ranching intensification aim at this outcome, but have yet to entrain landscape-level transformation. At a more aggregate level, future research should examine patterns in the interconnected expansions of environmental governance and intensive agro-industry. Researchers might trace with GIS the spatial development of multiple policy strands to map the governance frontier and to analyze its evolution in relation to socio-economic and land use dimensions of the intensive frontier. There is broad consensus that environmental governance has driven substantial reductions in Amazonian deforestation, but a deeper understanding of the role of governance in Amazonian development requires the integration of frontier theory and governance theory. The utility of this integrated political geography approach is not limited to the Amazon, but rather is broadly applicable to studies concerned with regional political-economic transformation. Going forward, we hope this contribution encourages scholars of economic geography and political economy to adopt systemic perspectives that pay greater attention to the political dimensions of frontiers and to the spatial dimensions of governance.

Adams, Ryan Thomas (2015). Neoliberal environmentality among elites: becoming ‘responsible producers’ in Santarém, Brazil. Culture, Agriculture, Food and Environment, 37(2), 84–95. Arima, Eugenio Y., Walker, Robert T., Perz, Stephen, & Souza, Carlos (2015). Explaining the fragmentation in the Brazilian Amazonian forest. Journal of Land Use Science (April), 1–21. https://doi.org/10.1080/1747423X.2015.1027797. Arvor, Damien, Meirelles, Margareth, Dubreuil, Vincent, Bégué, Agnès, & Shimabukuro, Yosio E. (2012). Analyzing the agricultural transition in Mato Grosso, Brazil, using satellite-derived indices. Applied Geography, 32(2), 702–713. Assunção, Juliano, Gandour, Clarissa, & Rocha, Rudi (2012). Deforestation slowdown in the legal Amazon: Prices or policies? Rio de Janeiro: CPI Working Paper. Assunção, Juliano, Gandour, Clarissa, & Rocha, Rudi (2015). Deforestation slowdown in the Brazilian Amazon: Prices or policies? Environment and Development Economics, 20(6), 697–722. Azevedo, Andréa (2009). Legitimação Da Insustentabilidade? Análise Do Sistema de Licenciamento Ambiental de Propriedades Rurais - SLAPR (Mato Grosso). Universidade de Brasília. Azevedo, Andréa, Rajão, Raoni, Costa, Marcelo A., Stabile, Marcelo C. C., Macedo, Marcia N., dos Reis, Tiago N. P., et al. (2017). Limits of Brazil’s forest code as a means to end illegal deforestation. Proceedings of the National Academy of Sciences, 114(29), 7653–7658. https://doi.org/10.1073/pnas.1604768114. Baletti, Brenda (2012). Ordenamento territorial: Neo-developmentalism and the struggle for territory in the lower Brazilian Amazon. Journal of Peasant Studies, 39(2), 573–598. Baletti, Brenda (2014). Saving the Amazon? Sustainable Soy and the new extractivism. Environment and Planning A, 46(1), 5–25. Barney, Keith (2009). Laos and the making of a ‘relational’ resource frontier. The Geographical Journal, 175(2), 146–159. Becker, Bertha (1982). Geopolítica Da Amazônia: A Nova Fronteira de Recursos. Rio de Janeiro: Zahar. Billington, Ray (1971). The genesis of the frontier thesis: A study in historical creativity. San Marino, CA: Huntington Library. Blomley, Nicholas (2003). Law, property, and the geography of violence: The frontier, the survey, and the grid. Annals of the Association of American Geographers, 93(1), 121–141. Brannstrom, Christian (2009). South America’s neoliberal agricultural frontiers: Places of environmental sacrifice or conservation opportunity? AMBIO: A Journal of the Human Environment, 38(3), 141–149. Brannstrom, Christian, Rausch, Lisa, Brown, J. Christopher, de Andrade, Renata Marson Teixeira, & Miccolis, Andrew (2012). Compliance and market exclusion in Brazilian agriculture: Analysis and implications for ‘soft’ governance. Land Use Policy, 29(2), 357–366. Brondizio, Eduardo S., & Moran, Emilio F. (2012). Level-dependent deforestation trajectories in the Brazilian Amazon from 1970 to 2001. Population and Environment, 34(1), 69–85. Brondizio, Eduardo S., Ostrom, Elinor, & Young, Oran R. (2009). Connectivity and the governance of multilevel social-ecological systems: The role of social capital. Annual Review of Environment and Resources, 34(1), 253–278. Browder, John, & Godfrey, Brian J. (1997). Rainforest cities: Urbanization, development, and globalization of the Brazilian Amazon. New York: Columbia University Press. Bunker, Stephen (1985). Underdeveloping the Amazon: Extraction, unequal exchange, and the failure of the modern state. Urbana: University of Illinois Press. Burawoy, Michael (2001). Manufacturing the global. Ethnography, 2(2), 147–159. Campos, Marina, & Nepstad, Daniel (2006). Smallholders, the Amazon’s new conservationists. Conservation Biology, 20(5), 1553–1556. Cargill. (n.d.). Cargill and the nature conservancy. https://www.cargill.com/doc/ 1432101097009/cargill-tnc-partnership-fact-sheet-pdf.pdf. Carmenta, Rachel, Coudel, Emilie, & Steward, Angela M. (2018). Forbidden fire: Does criminalising fire hinder conservation efforts in swidden landscapes of the Brazilian Amazon? Geographical Journal (February), 1–15. https://doi.org/ 10.1111/geoj.12255. Castree, Noel (2010). Neoliberalism and the biophysical environment: A synthesis and evaluation of the research. Environment and Society: Advances in Research, 1 (1), 5–45. Cavalcante, Gilma da Costa, Maciel, Norton Peres Cavalcante, & Carvalho, Raimundo José Coroa de. (2013). Diagnóstico Do Status Atual E Proposta de Reestruturação Organizacional Da Comissão Do Pacto Municipal Para O Fim Do Desmatamento Ilegal de São Félix Do Xingu/PA. São Félix do Xingu. Cavalcante, Gilma da Costa, Maciel, Norton Peres Cavalcante, & Carvalho, Raimundo José Coroa de. (2013). Relatório Preliminar: Diagnóstico Do Status Atual E Proposta de Reestruturação Organizacional Da Comissão Do Pacto Municipal Para O Fim Do Desmatamento Ilegal de São Félix Do Xingu/PA. São Félix do Xingu. Dias, Peri. (2013). The green wave of Brazilian Soy. http://www.nature.org/ ourinitiatives/regions/southamerica/brazil/explore/the-green-wave-ofbrazilian-soy.xml. Dietz, Thomas, Ostrom, Elinor, & Stern, Paul C. (2003). The struggle to govern the commons. Science, 302(5652), 1907–1912. Duchelle, Amy, Cromberg, Marina, Gebara, Maria Fernanda, Guerra, Raissa, Melo, Tadeu, Larson, Anne, et al. (2014). Linking forest tenure reform, environmental compliance, and incentives: Lessons from REDD+ initiatives in the Brazilian Amazon. World Development, 55, 53–67.

9. Declarations of interest None. Acknowledgments We are deeply grateful to the inhabitants of Santarém, Lucas, Paragominas, and São Félix, and the staff of TNC-Brazil. We thank the anonymous reviewers for their helpful comments. Thaler was supported by a US National Science Foundation Graduate Research Fellowship [grant number DGE-1144153].

G.M. Thaler et al. / World Development 114 (2019) 59–72 Eilenberg, Michael (2014). Frontier constellations: Agrarian expansion and sovereignty on the Indonesian-Malaysian border. The Journal of Peasant Studies, 41(2), 157–182. Ferreira, Igor (2010). Parcerias Para a Gestão Ambiental Em Propriedades Rurais: O Caso de Lucas Do Rio Verde - MT. Universidade de Brasília. Foweraker, Joe (1981). The struggle for land: A political economy of the pioneer frontier in Brazil from 1930 to the present day. New York: Cambridge University Press. Frank, Andre Gunder (1967). Capitalism and underdevelopment in Latin America: Historical studies of Chile and Brazil. New York: Monthly Review Press. Fundo Vale (2012). Municípios Verdes. Rio de Janeiro, Brazil: Série: Integração Transformação - Desenvolvimento. Garrett, Rachael D., Lambin, Eric F., & Naylor, Rosamond L. (2013). The new economic geography of land use change: Supply chain configurations and land use in the Brazilian Amazon. Land Use Policy, 34, 265–275. Gibbs, Holly, Munger, Jacob, L’Roe, Jessica, Barreto, Paulo, Pereira, Ritaumaria, Christie, Matthew, et al. (2016). Did ranchers and slaughterhouses respond to zero-deforestation agreements in the Brazilian Amazon? Conservation Letters, 9 (1), 32–42. Gibbs, Holly, Rausch, Lisa, Munger, J., Schelly, I., Morton, D. C., Noojipady, P., et al. (2015). Brazil’s Soy Moratorium. Science, 347(6220), 377–378. Goldman, Michael (2005). Imperial nature: The World Bank and struggles for social justice in the age of globalization. New Haven, Connecticut: Yale University Press. Greenpeace. (2006). Eating up the Amazon. Greenpeace. (2006). Greenpeace activists in Brazil Block Cargill Soy facility, May 22. https://www.greenpeace.org/archive-international/en/news/features/greenpeaceactivists-in-brazil/. GTS Soybean Working Group. (2016). Soy Moratorium: Monitoring satellite images of soy plantings in the Amazon Biome. Guimarães, Jayne, Veríssimo, Adalberto, Amaral, Paulo, & Demachki, Adnan (2011). Municípios Verdes: Caminhos Para a Sustentabilidade. Belém: Imazon. Hansen, Matthew, Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., et al. (2013). High-resolution global maps of 21st-century forest cover change. Science, 342(6160), 850–853. Hart, Gillian (2006). Denaturalizing dispossession: Critical ethnography in the age of resurgent imperialism. Antipode, 38(5), 977–1004. Hecht, Susanna (2011). From eco-catastrophe to zero deforestation? Interdisciplinarities, politics, environmentalisms and reduced clearing in Amazonia. Environmental Conservation, 39(1), 4–19. Hecht, Susanna, & Cockburn, Alexander (2010). The fate of the forest: Developers, destroyers, and defenders of the Amazon (Updated Ed). Chicago: University of Chicago Press. Hennessey, Alistair (1978). The frontier in Latin American history. University of New Mexico Press. Heynen, Nik, McCarthy, J., Prudham, Scott, & Robbins, Paul (2007). Neoliberal environments: False promises and unnatural consequences. New York: Routledge. Himley, Matthew (2008). Geographies of environmental governance: The nexus of nature and neoliberalism. Geography Compass, 2(2), 433–451. Hoelle, Jeffrey (2011). Convergence on cattle: Political ecology, social group perceptions, and socioeconomic relationships in Acre, Brazil. Culture, Agriculture, Food and Environment, 33(2), 95–106. IBGE (2018a). Pesquisa Pecuária Municipal Accessed August 12 https://sidra. ibge.gov.br/pesquisa/ppm/quadros/brasil/2016. IBGE (2018b). Produção Agrícola Municipal Accessed August 12 https://sidra. ibge.gov.br/pesquisa/pam/tabelas. IBGE. (2016). Censo Demográfico. https://sidra.ibge.gov.br/pesquisa/censodemografico/demografico-2010/inicial IBGE. (2017). Sistema Nacional de Índices de Preços Ao Consumidor. http://www. ibge.gov.br/home/estatistica/indicadores/precos/inpc_ipca/defaultseriesHist.shtm IBGE. (2018). Produto Interno Bruto Dos Municípios - Referência 2002 - 1999-2012. https://sidra.ibge.gov.br/pesquisa/pib-munic/tabelas INPE. (1989). Avaliação Da Floresta Amazônica. São José dos Campos, Brazil. http:// www.obt.inpe.br/prodes/Prodes1989.pdf INPE. (2018). Projeto PRODES: Monitoramento Da Floresta Amazônica Brasileira Por Satélite. Jepson, Wendy (2006). Producing a modern agricultural frontier: Firms and cooperatives in Eastern Mato Grosso, Brazil. Economic Geography, 82(3), 289–316. Lapola, David, Martinelli, Luiz, Peres, Carlos, Ometto, Jean, Ferreira, Manuel, Nobre, Carlos, et al. (2013). Pervasive transition of the Brazilian land-use system. Nature Climate Change, 4(1), 27–35. Lattimore, Owen (1962). Studies in frontier history: Collected papers, 1928–1958. New York: Oxford University Press. Lemos, Maria Carmen, & Agrawal, Arun (2006). Environmental governance. Annual Review of Environment and Resources, 31(1), 297–325. Macedo, Marcia, DeFries, Ruth, Morton, Douglas, Stickler, Claudia, Galford, Gillian, & Shimabukuro, Yosio (2012). Decoupling of deforestation and Soy production in the Southern Amazon during the late 2000s. Proceedings of the National Academy of Sciences, 109(4), 1341–1346. Martins, José de Souza. (1972). Frente Pioneira: Contribuição Para Uma Caraterização Sociológica. Cadernos (Universidade de São Paulo, Centro de Estudos Rurais E Urbanos). McCarthy, James, & Prudham, Scott (2004). Neoliberal nature and the nature of neoliberalism. Geoforum, 35(3), 275–283. McCracken, Stephen, Brondizio, Eduardo, Nelson, Donald, Moran, Emilio, Siqueira, Andrea, & Rodriguez-Pedraza, Carlos (1999). Remote sensing and GIS at farm property level: Demography and deforestation in the Brazilian Amazon. Photogrammetric Engineering & Remote Sensing, 65(11), 1311–1320.

71

Mitchell, Timothy (1991). The limits of the state: Beyond statist approaches and their critics. American Political Science Review, 85(1), 77–96. MMA (2009). Portaria No. 103, de 24 de Março de 2009. Diário Oficial da União. Moore, Jason W. (2000). Sugar and the expansion of the early modern world economy: Commodity frontiers, ecological transformation, and industrialization. Review, 23(3), 409–433. Moore, Jason W. (2010). ‘Amsterdam is standing on Norway’ part I: The alchemy of capital, empire and nature in the diaspora of silver, 1545–1648. Journal of Agrarian Change, 10(1), 33–68. Moore, Jason W. (2012). Cheap food & bad money: Food, frontiers, and financialization in the rise and demise of neoliberalism. Review: A Journal of the Fernand Braudel Center, 33(2/3), 225–261. Moore, Jason W. (2015). Capitalism in the web of life: Ecology and the accumulation of capital. New York: Verso. Moore Foundation. (2016). Deforestation-free beef in Brazil. https://www.moore. org/article-detail?newsUrlName=deforestation-free-beef-in-brazil Morton, Douglas C., DeFries, Ruth S., Shimabukuro, Yosio E., Anderson, Liana O., Arai, Egidio, del Bon, Fernando, et al. (2006). Cropland expansion changes deforestation dynamics in the Southern Brazilian Amazon. Proceedings of the National Academy of Sciences, 103(39), 14637–14641. Mwangi, Esther, & Wardell, Andrew (2012). Multi-level governance of forest resources. International Journal of the Commons, 6(2), 79–103. Nepstad, Daniel, McGrath, D., Alencar, A., Barros, A. C., Carvalho, G., Santilli, M., et al. (2002). Frontier governance in Amazonia. Science, 295(5555), 629–631. Nepstad, Daniel, McGrath, D., Stickler, C., Alencar, A., Azevedo, A., Swette, B., et al. (2014). Slowing Amazon deforestation through public policy and interventions in beef and soy supply chains. Science, 344(6188), 1118–1123. O Estado de São Paulo. (2012). Paragominas É O Exemplo a Ser Seguido, March 18. http://www.estadao.com.br/noticias/impresso,paragominas-e-o-exemplo-aser-seguido,849971,0.htm Ostrom, Elinor (1990). Governing the commons: The evolution of institutions for collective action. New York: Cambridge University Press. Ostrom, Elinor, Burger, Joanna, Field, C. B., Norgaard, R. B., & Policansky, David (1999). Revisiting the commons: Local lessons, global challenges. Science, 284 (5412), 278–282. Pacheco, Pablo (2012). Actor and frontier types in the Brazilian Amazon: Assessing interactions and outcomes associated with frontier expansion. Geoforum, 43(4), 864–878. Pacheco, Pablo, & Poccard-Chapuis, Rene (2012). The complex evolution of cattle ranching development amid market integration and policy shifts in the Brazilian Amazon. Annals of the Association of American Geographers, 102(6), 1366–1390. Peck, Jamie, & Theodore, Nik (2012). Follow the policy: A distended case approach. Environment and Planning A, 44(1), 21–30. Peluso, Nancy (2018). Entangled territories in small-scale gold mining frontiers: Labor practices, property, and secrets in Indonesian gold country. World Development, 101, 400–416. Peluso, Nancy, & Lund, Christian (2011). New frontiers of land control: Introduction. Journal of Peasant Studies, 38(4), 667–681. Perz, Stephen (2001). Household demographic factors as life cycle determinants of land use in the Amazon. Population Research and Policy Review, 20(3), 159–186. Perz, Stephen (2002). Household demography and land use allocation among small farms in the Brazilian Amazon. Human Ecology Review, 9(2), 1–16. Piketty, Marie-Gabrielle, Poccard-Chapuis, René, Drigo, Isabel, Coudel, Emilie, Plassin, Sophie, Laurent, François, et al. (2015). Multi-level governance of land use changes in the Brazilian Amazon: Lessons from Paragominas, State of Pará. Forests, 6(5), 1516–1536. Rasmussen, Mattias Borg, & Lund, Christian (2018). Reconfiguring frontier spaces: Territorialization and resource control. World Development, 101, 388–399. Rausch, Lisa (2013). Environmental governance as a development strategy: The case of Lucas Do Rio Verde Legal. University of Kansas. Ravikumar, Ashwin, Larson, Anne M., Duchelle, Amy E., Myers, Rodd, & Tovar, Jazmin Gonzales (2015). Multilevel governance challenges in transitioning towards a national approach for REDD+: Evidence from 23 subnational REDD+ initiatives. International Journal of the Commons, 9(2), 909–931. Rhodes, R. A. W. (1996). The new governance: Governing without government. Political Studies, XLIV, 652–667. Richards, Peter, Walker, Robert T., & Arima, Eugenio Y. (2014). Spatially complex land change: The indirect effect of Brazil’s agricultural sector on land use in Amazonia. Global Environmental Change, 29, 1–9. Robertson, Morgan (2012). Measurement and alienation: Making a world of ecosystem services. Transactions of the Institute of British Geographers, 37(3), 386–401. Rosenau, James N. (1995). Governance in the twenty-first century. Global Governance, 1(1), 13–43. Rudel, Thomas, DeFries, Ruth, Asner, Gregory, & Laurance, William (2009). Changing drivers of deforestation and new opportunities for conservation. Conservation Biology, 23(6), 1396–1405. Sauer, Carl (1963). In John Leighly (Ed.), Land and life: A selection from the writings of Carl Ortwin Sauer. Berkeley: University of California Press. Sawyer, Donald (1984). Frontier expansion and retraction in Brazil. In Marianne Schmink & Charles H. Wood (Eds.), Frontier expansion in Amazonia. Gainesville, FL: University of Florida Press. Schmink, Marianne, Hoelle, Jeffrey, Gomes, Carlos Valério A., & Thaler, Gregory M. (2017). From contested to ‘green’ frontiers in the Amazon? A long-term analysis of São Félix Do Xingu, Brazil. The Journal of Peasant Studies. https://doi.org/ 10.1080/03066150.2017.1381841.

72

G.M. Thaler et al. / World Development 114 (2019) 59–72

Schmink, Marianne, & Wood, Charles (1992). Contested frontiers in Amazonia. New York: Columbia University Press. Schwartzman, Stephan, Alencar, Ane, Zarin, Hilary, & Souza, Ana Paula Santos (2010). Social movements and large-scale tropical forest protection on the Amazon frontier: Conservation from chaos. The Journal of Environment & Development, 19, 274–299. Schwartzman, Stephan, Boas, André Villas, Ono, Katia Yukari, Fonseca, Marisa Gesteira, Doblas, Juan, Zimmerman, Barbara, et al. (2013). The natural and social history of the indigenous lands and protected areas corridor of the Xingu River Basin. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 368, 1–12. rstb.2012.0164. Skole, David, & Tucker, Compton (1993). Tropical deforestation and habitat fragmentation in the Amazon: Satellite data from 1978 to 1988. Science, 260, 1905–1910. Smeraldi, Roberto, & May, Peter (2008). O Reino Do Gado: Uma Nova Fase Na Pecuarização Da Amazônia Brasileira. Amigos da Terra - Amazônia Brasileira. Smith, Neil (2007). Nature as accumulation strategy. Socialist Register, 43, 1–21. Soares-Filho, Britaldo, Moutinho, Paulo, Nepstad, Daniel, Anderson, Anthony, Rodrigues, Hermann, Garcia, Ricardo, et al. (2010). Role of Brazilian Amazon protected areas in climate change mitigation. Proceedings of the National Academy of Sciences, 107(24), 10821–10826. Soares-Filho, Britaldo, Rajão, Raoni, Macedo, Marcia, Carneiro, Arnaldo, Costa, William, Coe, Michael, et al. (2014). Cracking Brazil’s forest code. Science, 344, 363–364. Steward, Corrina (2007). From colonization to ‘environmental Soy’: A case study of environmental and socio-economic valuation in the Amazon Soy frontier. Agriculture and Human Values, 24(1), 107–122. Stoker, Gerry (1998). Governance as theory: Five propositions. International Social Science Journal, 50(155), 17–28. Summers, Percy M. (2008). The post-frontier: Land use and social change in the Brazilian Amazon (1992–2002). Virginia Polytechnic Institute and State University. Swyngedouw, Erik (1996). The city as a hybrid: On nature, society and cyborg urbanization. Capitalism Nature Socialism, 7(2), 65–80. Thaler, Gregory M. (2017). The land sparing complex: Environmental governance, agricultural intensification, and state building in the Brazilian Amazon. Annals of the American Association of Geographers, 107(6), 1424–1443. Thypin-Bermeo, Sam, & Godfrey, Brian J. (2012). Envisioning Amazonian frontiers: Place-making in a Brazilian boomtown. Journal of Cultural Geography, 29(2), 215–238. TNC. (2015). Lands can do more: An integrated approach to conservation and development. https://global.nature.org/content/global-lands-report

TNC. (2017). Responsible Soy in South America: The nature conservancy and Cargill. http://www.nature.org/about-us/working-with-companies/companies-we-workwith/cargill-and-conservancy-case-study.xml TNC-Brazil. (2016). Carne Sustentável: Do Campo À Mesa. Accessed January 28. http://www.tnc.org.br/por-que-a-tnc/trabalhamos-em-parceria/nossosapoiadores/pecuaria-legal-boas-escolhas-do-campo-a-mesa.xml TNC-Brazil. (2013). TNC, Cargill, CEPLAC E CAPPRU Lançam Projeto Cacau Mais Sustentável Em São Félix Do Xingu (PA). http://portugues.tnc.org/nossashistorias/noticias-recentes/lancamento-cacau-mais-sustentavel.xml Tsing, Anna (2003). Natural resources and capitalist frontiers. Economic and Political Weekly, 38(48), 5100–5106. Turner, Frederick Jackson (1921). The frontier in American history. New York: Henry Holt and Company. VanWey, Leah, Spera, Stephanie, de Sa, Rebecca, Mahr, Dan, & Mustard, John (2013). Socioeconomic development and agricultural intensification in Mato Grosso. Philosophical Transactions of the Royal Society: B, 368(1619), 20120168. Velho, Otávio (1972). Frentes de Expansão E Estrutura Agrária: Estudo Do Processo de Penetração Numa Área Da Transamazônica. Rio de Janeiro: Zahar. Viana, Cecilia, Coudel, Emilie, Barlow, Jos, Ferreira, Joice, Parry, Luke, & Gardner, Toby (2016). How does hybrid governance emerge? Role of the elite in building a green municipality in the Eastern Brazilian Amazon. Environmental Policy and Governance, 26(5), 337–350. Walker, Robert (2004). Theorizing land-cover and land-use change: The case of tropical deforestation. International Regional Science Review, 27(3), 247–270. Walker, Robert, Browder, John, Arima, Eugenio, Simmons, Cynthia, Pereira, Ritaumaria, Caldas, Marcellus, et al. (2009). Ranching and the new global range: Amazônia in the 21st century. Geoforum, 40(5), 732–745. Walker, Robert, & Homma, Alfredo (1996). Land use and land cover dynamics in the Brazilian Amazon: An overview. Ecological Economics, 18, 67–80. Watts, Michael (2012). A tale of two gulfs: Life, death, and dispossession along two oil frontiers. American Quarterly, 64(3), 437–467. Webb, Walter (1952). The great frontier. Houghton Mifflin. zu Ermgassen, Erasmus K. H. J., de Alcântara, Melquesedek Pereira, Balmford, Andrew, Barioni, Luis, Neto, Francisco Beduschi, Bettarello, Murilo M. F., et al. (2018). Results from on-the-ground efforts to promote sustainable cattle ranching in the Brazilian Amazon. Sustainability, 10, 1301. https://doi.org/ 10.3390/su10041301. World Development. (2018). Special section: Reconfiguring frontier spaces: The territorialization of resource control. 101, 388–488.