<%BANNER%>

Roads, Governance and Land Use in the Brazilian State of Acre


PAGE 1

1 ROADS, GOVERNANCE AND LAND USE IN THE BRAZILIAN STATE OF ACRE By JEFFREY BENSON LUZAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

2 Copyright 2006 by Jeffrey Benson Luzar

PAGE 3

3 To the people of Alto Acre

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my chair, Dr. Marianne Schmink, for her help and guidance from my first days at UF through the completion of the di ssertation. I am especi ally grateful for her helpful comments and insights throughout the writi ng process. I also would like to thank my committee members Dr. Karen Kainer, Dr. Stephen Perz, Dr. Janaki Alavalapati and Dr. Rick Stepp, each of whom has provided valuable a dvice and insights throughout my doctorate program. In addition to the members of my com mittee, I would also like to thank several other individuals at UF who have helped me transform research questions into a completed dissertation. In anthropology, Dr Russ Bernand offered very helpful guidance in the process of proposal writing and the creation of testable hypotheses. Dr. Gr enville Barnes of the SFRC has been helpful in encouraging me to contemplat e the various dimensions of land tenure as I developed the dissertation. Dr. Robert Buschbacher, of WEC and the WFT program, has helped me think through my ideas and has served as a me ntor from my first weeks at UF. Finally, Dr. Chuck Wood in Latin American Studies provided truly invaluable guida nce during the tedious process of data analysis I am also grateful for financial support from the Hewlitt Foundation, the Working Forests in the Tropics Program, th e National Science Founda tion and the FulbrightBrazil Program for making fieldwork possible I also thank my family, including my pare nts Katherine Ann Luzar and Louis Frantz Luzar, for their encouragement as I pursued my dr eams and my aunt Dr. E. Jane Luzar, who first introduced me to the opportunities available at UF and helped me in the transition to life in Gainesville. I would like to offer an additional wo rd of thanks to my friends and colleagues at UF and the WFT program, especially Hilary de l Campo, who has offered encouragement, hugs and laughter through both my greatest and most diffi cult moments while at UF. I would also like to offer a word of thanks to the many people at UFAC, PZ and SETEM in Rio Branco, including,

PAGE 5

5 though not limited to, Elsa Mendoza and Dr. Irving Foster Brown who have helped me develop my ideas, and have offered logistical support and friendship from the earliest days of my research process. Finally, an es pecially heartfelt word of thanks goes to the people of Alto Acre who have invited me into their homes, shared w ith me their lives, a nd many of whom I have come to consider among my closest friends. Specia l thanks goes to Gilcilene Vale da Silva who has walked with and encouraged me in recent journeys and, se Deus quiser, will continue to do so for many journeys yet to come.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ........10 LIST OF FIGURES................................................................................................................ .......12 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 Roads, Governance and Land Use in the Brazilian Amazon..................................................15 Research Questions............................................................................................................. ....20 Organization of Dissertation................................................................................................... 21 Research Site.................................................................................................................. ........22 PAD Quixad...................................................................................................................2 4 PAE Sta. Quitria............................................................................................................27 RESEX Chico Mendes(Seringal Etelvir)....................................................................28 Fazenda Sta. Rita(Seringal Etelvir).............................................................................29 Terra Solta.................................................................................................................... ...29 Porto Carlos................................................................................................................... ..30 Methods........................................................................................................................ ..........31 Site Selection................................................................................................................. ..31 Sampling....................................................................................................................... ...32 Data Collection................................................................................................................ 33 Data Analysis.................................................................................................................. .36 Limitations.................................................................................................................... ...36 2 THE POLITICAL ECOLOGY OF ROAD S, GOVERNANCE AND LAND USE IN THE AMAZON..................................................................................................................... .42 Land Use and Land Cover Change in the Brazilian Amazon.................................................43 Deforestation.................................................................................................................. .44 Cattle Ranching and Pasture Formation..........................................................................45 Fire........................................................................................................................... ........48 Roads.......................................................................................................................... .....50 Public Policy.................................................................................................................. ..52 Political Ecology.............................................................................................................. .......56 Governance and Institutional Design..............................................................................60 Political Ecology in the Dissertation...............................................................................62 Variables Used in Analysis..................................................................................................... 65 Dependent Variables.......................................................................................................65 Deforestation............................................................................................................65

PAGE 7

7 Cattle herd size.........................................................................................................66 Brazil nuts................................................................................................................66 Annuals.....................................................................................................................66 Last burn...................................................................................................................66 Independent Variables.....................................................................................................67 Place of origin..........................................................................................................67 Distance from road...................................................................................................67 Bi-localism (male and female).................................................................................67 Adult labor (male and female)..................................................................................68 Years of residence....................................................................................................68 Property size.............................................................................................................68 Governance (administrative unit).............................................................................68 Participation in governance......................................................................................68 Hypotheses..................................................................................................................... .........69 Roads and Land Use........................................................................................................69 Governance and Land Use...............................................................................................70 3 ROADS AND LAND USE CHANGE IN ALTO ACRE......................................................72 Roads and the Occupation of the Amazon..........................................................................73 Road Paving and Colonization in Alto Acre..........................................................................75 Testing the Relationship between Place of Origin and Land Use...................................80 Modeling Place of Origin and Land Use.........................................................................82 Discussion..................................................................................................................... ...82 Intra-Regional Migration, Gender and Land Use...................................................................83 Testing the Relationship between Bi-lo calism, Out-Migration and Land Use...............86 Female bi-localism and land use..............................................................................87 Male bi-localism and land use..................................................................................87 Adult females and land use......................................................................................88 Adult males and land use.........................................................................................88 Modeling Bi-localism, Household Composition and Land Use......................................89 Discussion..................................................................................................................... ...90 Access to Roadway and Land Use..........................................................................................92 Testing the Relationship between Road Access and Land Use.......................................96 Discussion..................................................................................................................... ...96 Cattle and Land Use............................................................................................................ ....98 Testing the Relationship between Herd Size and Other Land Use Practices..................98 Modeling Herd Size and Brazil Nut Producti on, Area in Annuals, and Use of Fire in Pastures..................................................................................................................10 0 Discussion..................................................................................................................... .101 Conclusions.................................................................................................................... .......102 4 GOVERNANCE AND LAND USE CHA NGE IN THE INTER-OCEANIC HIGHWAY CORRIDOR.....................................................................................................109 Research Questions............................................................................................................. ..112 Land Rights and Environmental Conser vationthe Birth of a Coalition............................113

PAGE 8

8 Legal and Institutional Structures of E nvironmental Governance in Alto Acre..................116 Governance at the Federal Level...................................................................................119 Governance at the State Leve l: the Forest Government............................................121 Municipalities................................................................................................................1 24 CONDIAC.....................................................................................................................125 Administrative Unit.......................................................................................................125 Non Governmental Organizations (NGOs)...................................................................126 STR............................................................................................................................ ....127 Associations................................................................................................................... 129 Land Tenure and Forest Conservation..................................................................................131 Associations and Scale......................................................................................................... 134 AMOREB......................................................................................................................135 AMPAESQ....................................................................................................................136 AMPAESQ and Local Associations..............................................................................137 Representation and Effectiveness in Association Models.............................................140 Associations and ScaleConclusions..........................................................................141 Governance and Land Use....................................................................................................142 Testing the Relationship between Governan ce (Administrative Unit) and Land Use..144 Modeling the Effect of G overnance at the Administrative Unit Level Upon Land Use............................................................................................................................ .145 Governance and deforestation................................................................................145 Governance and cattle herd size.............................................................................146 Governance and years since burning pasture.........................................................146 Testing the Relationship Between Particip ation in Governance and Land Use............146 Discussion..................................................................................................................... ........148 5 RECONCILING POLITICAL AND ECONOM IC REALITIES IN ALTO ACRES BR-317 CORRIDOR............................................................................................................155 The Political Ecology of Land Use in Alto Acre..................................................................155 The Alto Acre ParadoxWeighing the Costs and Benefits of Environmental Compliance..................................................................................................................... ..158 The Alto Acre Paradox and Institutional Design..................................................................166 Reconciling Centralized and Part icipatory Forms of Governance.......................................168 Implications of Findings for Forest and Agrarian Policy in Alto Acre................................169 Urban and Rural Connections.......................................................................................169 Environmental Infraction Penalties...............................................................................170 Cattle......................................................................................................................... .....171 ProAmbiente and Environm ental Service Payments.....................................................172 Fire........................................................................................................................... ......172 Directions for Future Research.............................................................................................173 Gender and Migration....................................................................................................173 Land Use Governance...................................................................................................174 APPENDIX A HOUSEHOLD SURVEY.....................................................................................................176

PAGE 9

9 LIST OF REFERENCES............................................................................................................. 190 BIOGRAPHICAL SKETCH.......................................................................................................199

PAGE 10

10 LIST OF TABLES Table page 1-1 Sampling by Administrative Unit......................................................................................39 3-1 Descriptive Statistics for Inde pendent Variables in Chapter 3........................................105 3-2 Descriptive Statistics for De pendent Variables in Chapter 3...........................................105 3-3 Brazil Nut Production Regressed on Place of Origin (OLS Regression Coefficients)....105 3-4 Correlations among Adult Males, Adult Females, Area in Annuals, Brazil nut Production and Cattle Herd Size......................................................................................106 3-5 Herd Size and Brazil Nut Production Regressed on Female Bi-Localism (OLS Regression Coefficients)..................................................................................................106 3-6 Brazil Nut Production and Cattle Herd Size Regressed on Adult Male Labor (OLS Regression Coefficients)..................................................................................................106 3-7 Cattle Herd Size Regressed onAdult Female Labor (OLS Regression Coefficients)......107 3-8 Correlations among Distance to Road, Deforestation and Cattle Herd Size...................107 3-9 Correlations among Cattle and Brazil Nuts, Area in Annuals and Years Since Burning Pasture................................................................................................................ 107 3-10 Area in Annuals and Frequency of Past ure Burning Regressed on Herd Size (OLS Regression Coefficients)..................................................................................................108 4-1 Administrative Units in Field Site : Range of Environmental Governance......................150 4-2 Descriptive Statistics for Bivariate an d Continuous Measures of Participation in Governance..................................................................................................................... .151 4-3 Frequencies for Categorical Measures of Governance....................................................152 4-4 Means Test for Effects of Governan ce (Administrative Unit) for Household Deforestation.................................................................................................................. ..152 4-5 Means Test for Effects of Governan ce (Administrative Unit) for Herd Size..................153 4-6 Means Test for Effects of Governance (A dministrative Unit) for Pasture Burning........153 4-7 Governance (Administrativ e Unit) and Land Use (OLS Regression Coefficients).........153 4-8 Factor WeightingsPartici pation in Governance Index.................................................154

PAGE 11

11 4-9 Correlations among Participation in G overnance, Household Deforestation, Cattle Herd Size and Years since Burning Pasture....................................................................154

PAGE 12

12 LIST OF FIGURES Figure page 1-1 Field Site in Context of Acre and Brazil............................................................................40 1-2 Research Site with Administrative Units...........................................................................41

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ROADS, GOVERNANCE AND LAND USE IN THE BRAZILIAN STATE OF ACRE By Jeffrey Benson Luzar December, 2006 Chair: Marianne Schmink Major Department: Interdisciplinary Ecology The dissertation examines the relationships among roads, governance and land use in the Western Brazilian Amazon. Given the often negativ e social and environmental impacts of past road construction projects in the Amazon, in re cent decades, the Brazilian government and civil society have come together to develop various forms of environmental governance, such as extractive reserves and regulations on land cl earing and burning, designed to minimize these negative impacts. The dissertation draws on a ca se study in the Inter-Oce anic Highway corridor in the state of Acre to examine both the impacts th at roads have had for land use and the role that governance has had in influe ncing these impacts. The dissertation focuses on two forms of migr ation that have accompanied the opening and subsequent paving of the roadin-migration and rural-urban migra tion. Ethnographic and multivariate regression analyses revealed that many of the differences in terms of land uses that once characterized agriculturally-oriented migran ts and extractivism-oriented Acreanos have disappeared after more than two decades of co -existence. An exception is the tendency of Acreanos to more heavily exploi t Brazil nuts than do migrants. More recently, the paving of the road has led to the emergence of bi-localisma tendency of households to invest their time and resources simultaneously in both rural and urban settings. This trend disproportionately

PAGE 14

14 affected women. Furthermore, households in which female household heads spent greater amounts of time in town had lower Brazil nut production and households with fewer adult women present had significantly smaller cattle herds. Qualitative analysis s uggested that this last tendency is related to gendered cattle produc tion systemsin which herds own by women tend to grow more quickly than those owned by men. Findings provide an optimistic evaluation of the potential for governance to moderate some of the most critical land use and land cover ch anges that have occurred in the road corridor. Multivariate analysis showed that environmen tal governance, as represented by governmentmandated administrative units, was significantly rela ted with lower levels of deforestation and smaller cattle herds.

PAGE 15

15 CHAPTER 1 INTRODUCTION Roads, Governance and Land Use in the Brazilian Amazon A central issue in discussions about sustai nable development in Brazil involves the implications of roads in the Amazon Basin. While road construction and paving carry the promise to improve access to goods and services among previously isolated rural and urban communities, they have also been shown to br ing about negative environmental and social consequences, including deforestation and the disp lacement of rural smallholders, under pressure from in-migration and land-speculation. In academic and policy circles, a considerable debate exists regarding the capacity of governance to prevent these potential repercussions from highway paving (e.g., Nepstad et al., 2001, 2002; Laurance et al., 2001, Soares-Filho et al., 2004). Relevant examples of governance incl ude the implementation of a deforestation monitoring program in Mato Grosso (Fearnsid e, 2003), ecological-economic zoning in Rondnia and Acre (Mahar, 2002, Governo do Acre, 2000) and, at smaller scal es, the activities of rural workers syndicates to promote sustainable land use among members. While much of the existing literature uses landscape-level models to predict social and environmental changes arising from road construction and governance, to date, relatively little detailed research has focused upon these processes at the household and community level. The dissertation addresses the relationship between roads, governance and land use at household and community levels in the municipalit ies of Brasilia and As sis Brasil in the Alto Acre region of Acre, Brazil. It draws upon a political-ecological framework to contextualize household-level land use decisions within a multi-scaled political, economic and ecological context. Through such a perspective, it beco mes easier to interpret the complex and often

PAGE 16

16 contradictory economic, ecological and political forces shaping land use decisions in Acres highway corridors. Historically, the opening of roads in the Brazilian Amazon has been accompanied by largescale deforestation, often followed by even more massive deforestation events when these roads are eventually paved (Hecht and Cockburn, 199 0; Schmink and Wood, 1992; Fearnside, 2000b; Laurance et al., 2002). In the Amazon region, while deforestation within 50 km buffers of unpaved roads generally ranges from 0-9%, within similar buffers of paved roadways 29-58% of the land area has been deforested (Nepstad, 2002) For instance, in the state of Rondnia, the paving of the BR-364 highway in the early 1980s le d to a massive influx of settlers from outside the region who brought with them livelihood systems that relied not upon the forest, as had the existing indigenous and rubber tapper populations but upon the removal of the forest and its replacement with agricultural fields and pasture (Browder and Godfrey, 1997). While deforestation in Acres early road corridors in the 1970s and 1980s generally occurred on a smaller scale than in neighboring Rondnia, the pr esence of an organized resistance among the rubber tapper community led to direct and wide ly-publicized conflicts between incoming groups seeking to acquire and clear land and the rubber tappers who sought to both maintain land rights and preserve the forest that provided a significant por tion of their livelihoods. The impact of roads upon land use change has b ecome an especially salient issue given the Avana Brasil (Advance Brazil) program of the fe deral government. As of the mid-2000s, the program had begun providing all-weather roads in regions previously served by seasonal unpaved roads. Avana Brasil consists of 338 projects in Amazni a with US $43 billion allocated for the first eight years (Fearnside, 200 2b). In all, some 6245 km of existing seasonal roads have either been paved or are scheduled to be paved (Nep stad et al., 2001). As the amount

PAGE 17

17 of land in the Brazilian Amazon within 50 km of paved highways will nearly double under the Avana Brasil programfrom 16-28%extrapolation from past trends shows that the project could greatly accelerate deforestation in the Amaz on Basin. In a socio-environmental evaluation of the project, Nepstad and co-authors ( 2001) predict that, due to the project, 120,000-270,000 square kilometers of primary forest coul d be lost in the next 25-35 years. In Acre, as in other parts of the Amaz on Basin, both the greatest challenges and opportunities for combing economic development with forest conservation will likely occur in highway corridors (Kainer et al ., 2003). On one hand, given in creased accessibility, following road paving, adjacent areas become subject to greater population movements and, due to the access to new markets, land uses such as largescale cattle ranching become more economically feasible, leading to further deforestation. On the other hand, the increased access to urban centers and markets may also prove to be be neficial in promoting forest-based economic development. For example, in roadside areas su ch as PAE Cachoera and Porto Dias and in the newly established Antimari state forest, all in the state of Acre, sustainable timber projects take advantage of the improved access to urban market s. While improved transportation in Acres road corridors has facilitated out-migration to cities, it has also, by pr oviding both better access to urban services and easier tr ansportation for urban-based teach ers, extensionists and medical care providers, improved living conditions in many rural communities (personal observation). Ironically, despite the increased threats to the forest, the capacity of governance regimes may actually be strongest in highway corridors. The director of the Brasilia office of the National Institute for Colonization and Agrarian Reform (INCRA) attested that the paving of the BR-317 highway had eased his offices task in wo rking with local associ ations and enforcing environmental regulation; whereas a visit to some of the more distant ar eas administered by the

PAGE 18

18 agency (up to 110 kilometers each way) once coul d take several days, with the paving of the road, it became possible for an INCRA team to ma ke field visits and retu rn within a matter of hours (Regional Director, Brasilia Office of INCRA, personal communication). In Acre, the arrival of roads ha s brought many benefits to local residents. It has eased the isolation once felt by rural familiesfacilitating their access to urban-based services such as schools, banks and hospitals. It also has brought increased acce ss to markets and to outside capital. With the notable exception of Braz il nut extraction, which has benefited from improvements in road access, these changes have led to a general shift away from forest-based land uses to other more lucrative and less labor intensive practicesespecially cattle ranching giving rural families an opportunity to purchase goods and services that might otherwise remain out of reach (Faminow, 1998; pers onal observation). For example, in the field site, numerous families attributed their ability to build new homes, acquire solar panels and generators and purchase motorcycles or other vehicles and pay off farm debt and emergency health bills directly to profits from cattle ranching. However, in addition to increased access to urban-based services and unprecedented prosperity brought by the road, the people of the region have found themselves facing unprecedented economic and environmental risks. Increased connectivity to regional and global markets has also brought new economic risks, such as the sudden shifts in prices for the forest and agricultural commodities upon which many families depend for monetary incomeas witnessed by recent price fluctuations for beef and Brazil nuts. Due to the often lucrative nature of cattle production in the region, many fa rmers have chosen to clear forest and plant pasture grasses. Due to the diffi culty of growing crops in land that has previously been used as

PAGE 19

19 pasture, increased allocation of land to pasture represents a decr eased opportunity to allocate land to other uses should beef prices fall. Whereas deforestation under trad itional shifting cultivation ge nerally allowed the eventual regeneration of forest cover, deforestation fo llowed by pasture creation leads to the long-term displacement and fragmentation of ecologically -important tropical forest. Compounding the direct impacts of deforestation a nd pasture establishment, the asso ciated use of fire in clearing forest and maintaining pasture has brought furt her negative impacts for the region. During the burning season, thick smoke has led to increases in traffic accidents and respiratory illnesses (Mendona, 2004), and during increasingly common dry year s (such as 2005), uncontrolled fires threaten not only the states forests, but also devastate pastures, corrals fences and livestock, ironically threatening the cattle industry itself in more extr eme instances. Furthermore, deforestation and fire may combine to bring devastating climactic a nd hydrological impacts at both regional and global levels. Within th e Amazon Basin, rivers run dry during the increasingly-intense dry seas ons, game and timber become scarce and adverse meso-climate change (including regional warming and drying) has shifted from a hypothetical proposition to an acknowledged fact of life for many of the areas residents, as noted in numerous interviews. And, at the global scale, cha nges in the Amazons hydrology resu lting from deforestation and fire may in turn change rainfall patterns elsewh ere in Brazil, North America and Eurasia (Maslin et al., 2006). Environmental governance, by shaping the tr ajectory of land use change in highway corridors, has direct implications for both fo rest conservation and economic development both within and beyond the highway corridors. In recent decades, governance in the Brazilian Amazon has been oriented toward the often contradictory tasks of promoting economic

PAGE 20

20 prosperity and avoiding environmental risk through mandated stewar dship of the regions forests. It is the complex relationship among these political, economic, and ecological contexts in affecting household and community -level land use change that my research project seeks both to describe and explain. Research Questions My research project addresses three broad que stions related to the relationship between roads, governance and land use in the Alto Acre region. The first question addresses the relationship between the Inter-O ceanic Highway and land use. As will be shown in the following chapter, researchers have pointed to the negative impacts of increased migration and market inte gration in past road construction projects, predicting that they are likely to continue in more recent road projects, such as the paving of the BR-317. Regarding migration, as will be shown in Chapter 3, in the Alto Acre region, the BR317 has facilitated the migration of families from outside of the state into the region. It has also led to intra-regional migration between rural and urban areas. The road has also provided improved access to and the growth of urban market s, and in so doing has dramatically changed the land use options available to the regions rural residents. Addressing these linkages between the road and land use, I ask how has the In ter-Oceanic Highway, by affecting migration and market integration, in turn affected land use in the road corridor? This question of the roads impacts upon land use leads to a second question about the capacity of governance to mediate these impacts. As will be shown in chapter 2, a considerable debate has emerged regarding the capacity of go vernance to mediate the impacts that new roads will have for land use in the Brazilian Amazon. While some researchers anticipate that past patterns of predatory land use will continue in current road projects, others argue that the development of legal mechanisms designed to pr otect Amazonias natural resources and the rise

PAGE 21

21 of an increasingly empowered civil society following Brazils return to democracy in the mid1980s can avert much of the negative social and environmental consequences of past infrastructure projects in th e region, even bringing new opport unities for integrating economic development and forest conservation. In order to address this debate in the context Acres BR317 highway corridor, I ask What role does governance play in in fluencing land use change in the BR-317 corridor?The final question addresse s the differential impacts of two forms of governancecentralized and participatoryupon la nd use in this region. In this region, centralized governance, while based in a democr atic system, relies heavily upon decisions by governmental bureaucracies. The establishment and enforcement of legal limits on private property deforestation would be one example. In the research site, this form of governance is most visibly represented by the administrative un it (e.g., extractive reserv e, colonization area, etc.), each of which are differentiated by legal de forestation limits, by presence of governmental agency and by the availability of environmental courses and pr ojects. The other form of governance is more participatory in nature and is most conspicuous at a household level. Examples would include participation in asso ciations, rural workers syndicates and in NGOsponsored environmental courses. In Alto Acre this participatory form of governance stems largely from the grassroots social movement that emerged in the 1970s and early 1980s to protect land rights and livelihoods of forest-dependent peoples su ch as rubber tappers. To address the differential impacts of these two fo rms of governance upon land use, I ask what is the relationship among political-e conomic context, various forms of governance, environmental subjectivity and land use in Alto Acre? Organization of Dissertation In this chapter, I offer an introduction to the research questions, the research site and methodology used in the collection and analysis of data. The following chapter outlines the

PAGE 22

22 theoretical background, highligh ting especially the possibil ities for applying a politicalecological perspective to land use change in the Amazon. Drawing upon this foundation, I operationalize the various independent and depende nt variables that are used in subsequent chapters to test relationships among the road, governance and land us e. Using these variables, I outline the hypotheses that will be tested in chapters 3 and 4. In chapter 3, I study the impacts of the Tran s-Oceanic Highway for land use. In this chapter, I first address the relationship among roads, migration and land usefocusing upon the role that place of origin and ru ral-urban migration have for land use. Chapter 3 then addresses the question of market integration, focusing spec ifically upon the impacts of road access and the expanding cattle economy, for land use. Chapter 4 addresses the impacts of governance for land use in the region. The first portion of the chapte r discusses the structure of governance in the region, including the intersections among land tenu re, institutional design and land use. The latter portion of the chapter addresses the im plications of centrali zed and participatory governance for land use. Chapter 5 applies an explicitly political-ecological framework to consider the findings in chapters 3 and 4. Special attention is given to the ways in which different forms of governance affect (or do not affect) land use and the im portance of the regiona l political-economy in mediating this relationship. Drawing upon this analysis, the chapter c oncludes with several policy recommendations for reconciling environmen tal governance with local economic realities. Research Site The research was conducted in the Alto Acre region of the state of Acre in the Western Brazilian Amazon (Figure 1-1). The Alto Acre region is so named due to its location in the upper regions of the Acre River watershed and is comprised of four muni cipalitiesAssis Brasil, Brasilia, Epitaciolandia and Xapuri. It is home to some 55,000 people (IBGE, 2004), who,

PAGE 23

23 while rarely differentiating themselves along ethnic lines, tend to come from a combination of European, African and Amer-Indian ethnic b ackgrounds, with non-Indians tending to have arrived in the region either by ri ver from Northeastern Brazil in the late 1800s and early 1900s to tap rubber or followed roads from South and Central Brazil in the mid to late 20th century. The population is roughly half rural and half urban. The entire ur ban population is concentrated in the towns of Assis Brasil, Brasilia, Epitaci olandia and Xapuri, capita ls of their respective municipalities and the only urban areas in the region. While the service sector, light industry, a nd, increasingly, the public sector, employ many people in rural areas, agricultu ral and forest production form the backbone of the private economy, especially in rural areas, but also among urbanites who often have land and/or cattle in surrounding rural areas. With the exception of some specialized professions, most workers earn the national minimum wagewhich at the time of research was $300 reais (US$140) per month. Predominant land cover includes open semi-deciduous forests, which consist of a broken canopy interspersed with palms and bamboo, clos ed semi-deciduous forest, which has a denser canopy and more open understory, anthropogenic gras slands (cattle pasture) and agricultural plots, and abandoned agricultural lands in vari ous states of succession (Governo do Acre, 2000). In the study region, soil types are dominated by Red and Yellow Agri ssols; while these soils tend to have high aluminum content, it rarely reach es the point of aluminum toxicity (Governo do Estado do Acre, 2000). In my research area, in formants commonly offered a perception of the areas soils and climate as being apt for agriculture. The study lies on the border of Brasilia and Assis Brasil, measuring some 50 kilometers east to west and 30 kilometers north to south an d is bisected east to west by the Inter-Oceanic Highway (Figure 1-2). The majority of the resear ch site lies in the municipality of Brasilia

PAGE 24

24 with the westernmost portion lying in Assis Bras il. As of 2004, the 4,852 individuals in Assis Brasil and 16,940 living in Brasilia were roughl y divided between urban and rural residence (IBGE, 2004). The site is bounded by the Rio Acre and Bolivia to the south and the Rio Xapuri and the RESEX Chico Mendes to the north. The study area formerly was comprised of four adjacent seringais (plural of seringal or rubber estate)So Pedro, Sacado, Etelvir and Porto Carlos, each of which was bisected by the roadway. Each seringal came to a different fate after the coming of the road in the late 1970s, the subsequent violent land disputes between ranc h workers and rubber tappe rs in the late 1970s and early 1980s, and the subsequent implementati on of agrarian reform and extractive reserves in the area. As of the time of research in the mid-2000s, the four original seri ngais were represented by six contiguous administrative units: 1) Quixad, a Colonization Project ( Projeto de Assentamento Dirigido or PAD) 2) Santa Quitria, an Extractive Settlement Project ( Projeto de Assentamento Extractivista or PAE) 3) RESEX Chico Mendes (the portion lying within Seringal Etelvir), 4) the privately-owned fazenda (ranch) Santa Rita, 5) th e largely un-administrated Terra Solta and 6) Seringal Porto Carlos. The admini strative units differed by land tenure system, government agency responsible (i f any), legal deforestation limit and the presence of government and NGO-sponsored courses and projects related to forest conservation. These differences and their implications for land use will be analyzed in depth in chapter 4. PAD Quixad Following the axis of the Inter-Oceanic Highw ay lays the colonization area officially known as Projeto de Assentamento Dirigida Quix ad, Gleba 6, or what is locally known more simply as Quixad. PAD Quixad is one of the numerous colonizat ion areas the federal government has implemented in the Brazilian Amaz on. In Acre alone, colonization areas occupy

PAGE 25

25 some 1.4 million hectares (Governo do Acre, 2000) Historically, these areas have been administered by INCRA and divided into rectangular family farms or lotes. In Acre, lotes ranging from 60 to 100 hectares have been given to current occupants and to landless families from other parts of the countryall of whom are commonly referred to as colonos (Governo do Acre, 2000). These areas have generally been designed in distant government offices, with little consideration for on-the-ground terrain, agricult ural suitability and water availability. Furthermore, due to the small size and the fact that few lotes coincided with pre-existing rubber trails, traditional forms of forest extractivism pr oved unviable, even for occupants with traditions of forest extractivism (Governo do Acre, 2000). In more recent years, due to a decrease in available land and increased emphasis upon fore st conservation, in the state of Acre, the formation of new PADs has slowed considerab lyand the size of newly-demarcated lotes has decreased considerably as well1. The history of PAD Quixad lies in the vi olent confrontations between the areas traditional rubber tapping populations and newly-arrived ranch workers2. In an effort to stem the growing violent land conflict in th e region that had led to the assa ssination of the president of Brasilias Rural Workers Syndicate, Wilson Pi nheiro, and a retaliatory lynching of the ranch manager widely believed to have been respons ible, INCRA expropriated several ranches located in this area to form the Projeto de Assentamen to Dirigido (PAD) Quixad. The initial project was implemented in 1982, with the westernmost s ectionGleba 6 (the portion of the PAD lying 1 In QuixadGleba 6, during 2005, some 11 families were settled onto a 300 hectare area that had not been developed by INCRA when the project was formed in the early 1980s. Landholdings were considerably smaller than past loteswith most families receiving slightly less than 30 hectares of land. 2 By ranch workers, I refer to both the ranch managers ( gerente ) and ranch hands ( pees ) that directly operated the ranches in the region. The word rancher ( fazendeiro ) is more properly applied to the ranch ownersindividuals who, due to the fact that they have often maintained residency in other parts of the state or country, rarely were directly involved with the day-to-day operations and with personal confrontations with rubber tappers during this era.

PAGE 26

26 in the research site) implemented shortly ther eafter in 1983. In PAD Quixad, Gleba 6, lots ranging from 70-100 hectares in size were dema rcated in areas of Seringais So Pedro and Sacado formerly occupied (though not necessarily cleared) by the ranch that had been managed by Nilo. The area stretched along the edge of the highway and, in some areas, along access roads leading further to the north and south. Fi rst priority was given to the areas current (Acreano) inhabitantsmost of whom received the lot in which their homes happened to be located. Unoccupied lots were given to landless families who were brought by INCRA from Paran and other states in the south in 1983. Many of the areas current residents had r eceived their lots fr om INCRA in 1983. Many other current inhabitants represent later migrants or children of the first recipients who purchased lots from prior inhabitants who moved, most co mmonly to urban centers in the region. As a colonization area, among the administrative units included in the study, Quixad contained the highest concentration of residents born outside of Acre, with 72% of sampled households in the PAD having at least one head born outside of Acre or neighboring areas of Peru and Bolivia. INCRA carried official respons ibility for overseeing the colo nization project. However, some families had received full title to their lots, and the agency intended to fully privatize the project in the near future. Livelihoods were dominated by su bsistence annual and perennial agriculture as well as some ma rket-oriented agriculturemost notably cattle a nd, to a lesser degree, Brazil nuts. Timbermostly from land de stined to be converted to agricultureserved as an occasional source of supplemental income for many families. Deforestation was very advanced in this area and nearly all lots had far exceeded the 20% deforestation level permitted by law. According to estimates from the direct or of INCRAs regional o ffice in Brasilia, some 60%-70% of the project area had been deforest ed as of 2003. No over-a rching local governance

PAGE 27

27 structure existed in the region. Ra ther, most residents belong to one of eight local associations, each of which is autonomous from the other. PAE Sta. Quitria Covering approximately 44,205 hectares, and ho me to at least 277 families, PAE Santa Quitria was created in 1988, the second PAE to be created in Acre (Governo do Acre, 2000). Under INCRAs original plans, the entire area was to be divided in to lots and incorporated into the PAD. However, the implementation of lots was halted as various stakeholders, including local residents and the National Council of Rubber Tappers (CNS), recognizing that traditional extractivist livelihoods were impossible on 70-1 00 ha lots, lobbied INCRA to implement a different system of agrarian reform that woul d recognize local families traditional rights to multiple rubber tapping trails (ranging from three to eight trails for most families, with each trail being considered roughly equivale nt to 100 hectares) and would gi ve land to its original rubber tapping occupants and not to landless agriculturalist families from outside of the region. The PAEs were established in an attempt to implement agrarian reform while respecting traditional colocao land tenure systems and forest-based livelihoods. As such, the PAEs served as an early experiment in the integration of agrarian refo rm and forest conservation. In these precursors of extractive reserves, such as Santa Quitria, INCRA maintained official ownership of the area (extrac tive reserves, such as the gl obally-renowned Chico Mendes Extractive Reserve, would be administered by IBAMA, the federal agency charged with environmental protection, without a role for INCRA, an agency tradit ionally charged with agrarian reform issues). PAEs were developed as an alternative to the PAD modality, an alternative in which traditional rubber tapper tenure systems would be ma intained. While the di stribution of lots in the PAD recognized rubber tapper rights to th e land they occupied, they recognized neither

PAGE 28

28 traditional livelihoods nor land tenure structures. As in other extractive reserves, land in the PAE is owned by the Brazilian state, and families have use rights to large areas (generally 200300 ha) that are defined not by firm boundaries, but by the rubber-t apping trails within them (generally three to four per family). Local governance systems tend to focus upon local associations and the Association of the Re sidents of PAE Sta. Quitria (AMPAESQ)a centrally-located association intended to serve as an umbrella organization for local associations and to represent the projects resident s in external poli tical concerns. RESEX Chico Mendes(Seringal Etelvir) Seringal Etelvir was sub-divided in to three tracts and sold to ranching interests in the early 1980s, but the owner of the northwestern tract, as of the late 1980s, ha d still not expropriated local families nor cleared the forest for pasture, and unlike the more accessible southern segment of the former seringal, it was never expropriated for colonizati on by INCRA. As a result, this approximately 8000 hectare area was still largely fo rested and lightly inhabited when the Chico Mendes Reserve was created in 1990 Following its appropriation by IBAMA and incorporation into the reserve, this seringal fragment stands at the southern fringe of the RESEX, bordering PAD Quixad and lying a mere four kilometers fr om the highway at its southernmost point. The area is arguably more directly impacted by road construction and market access than most other seringais within the RESEX. Indeed, evidenced in Figure 1-2, Seringal Etelvir is not only a bulwark of the RESEX facing southward toward th e highway, it is also a major exit point and first or final resting point for th e residents of at least five seri ngais located further inside the reserve. As in PAE Sta. Quitria, land tenure in th e RESEX is based upon the relatively large colocao property unit, and propert ies are subject to a 10% ceili ng on legal deforestation (as in the PAE but in contrast to the 20% in all other parts of the research site). Like Sta. Quitria, a

PAGE 29

29 two-tiered association structure exists, with loca l associations (two exist in this area) falling under the umbrella of the Association of Reserve Residents of Brasilia (AMOREB) situated in Brasilia. Fazenda Sta. Rita(Seringal Etelvir) The northeastern segment of Seringal Etelvi r was purchased to create three adjacent ranches, the largest of which is Santa Rita. The ranchs former owner, locally known as O Alemo (The German), until his deat h in a traffic accident in th e late 1980s enjoyed relatively harmonious relations with his rubber tapper neig hbors compared to neighboring ranchers and ranch workers who often killed and at times were killed by rubber tappers in land disputes. The Alemo, on the other hand, differe ntiated his ranch by purchasing land from not only the estate owner, but from the resident rubber tappers who lacked formal title to the land. This practice lowered animosity among displaced rubber tappers and is generally believed by locals to explain the continued existence of Santa Rita while so many other ranches were expropriated by INCRA as it attempted to stamp out land conflict in the region. While representing nearly half of the deforested area in the research site, the privat ely-owned Sta. Rita and its two ranch neighbors have played a relatively minimal role in more recent deforestation. Unlike other areas where deforestation rates have steadily increased, the defo restation of Sta. Rita largely occurred in two massive phases, first in the early 1980s and ag ain in 1988, just as Amazonian deforestation was catching the attention of the worl d and the Brazilian government. Terra Solta The area alternatively known as the Cinturo Verde (Green Belt) or the Terra Solta (Loose Land) received its names from the facts that it forms a two kilometer largely-forested belt extending some six kilometers between the fazenda and RESEX Chico Mendes and, as land that was left out of the RESEX due to a su rveying error, neither of the two major federal-

PAGE 30

30 level agencies involved in land use policy INCRA and IBAMAhas responsibility for the area. Most of the residents were either Acr eanos or the grown children of Quixadas aging original settlers. Although there was a gene ral property boundary understanding between most (though not all) residents, no formal land titli ng system existed, documentation of residence through Rural Workers Syndicate being the only doc umentation possible. Nor did residents have a common association. Rather, most residents bel onged to one of three local associations located outside the Terra Solta while severa l others remained unassociated. Porto Carlos The approximately 100 families of Porto Carlos lived in what was, as of the time of fieldwork, one of the last exis ting seringais in Alto Acre, a nd the only one lying along the BR317 highway corridor. Traditi onally, seringais were owned by a rubber baron, with the rubbertapper residents having some use ri ghts, but no legally recognized rights to tenure of the land on which they lived. While Porto Carlos had long c eased to be a functiona l seringal, the owners, while still managing a few prime areas for Brazil nut and cattle ranching, had allowed most of the seringal to fall under the de facto tenure of its ex-rubber tapper resi dents. Residents were still subject to federal environmental law (for instan ce, all landowners were subject to a 30 hectare deforestation limit, calculated as 20% of the 150 hectares officially held in each property3). And while kinship ties were strong in nearly all the communities, Porto Carlos demonstrated a remarkable level of kin-based social capital, w ith nearly all residents belonging to one of two families (which traditionally tended to inter-marry ). Compared to the other more heterogeneous and divided communities, Porto Carlos tended more often than not to speak with a (near) 3 While all properties were recently documented as 150 hect ares, properties could be up to 600 or more hectares. This has encouraged some of the larg er owners to settle family members on their property so as to maintain de facto ownership of their original colocaes

PAGE 31

31 common voice in matters of local and regional politi cs. While not all residents participated, a single association existed repres enting the peoples of Porto Ca rlos (Association Porto Carlos, located near the geographical center of the seringal). From both a theoretical and policy perspectiv e, as no formal governmental decision had been issued regarding the future of the area, Porto Carlos was an especially interesting area in the research site. Faced with the mixed successes an d failures of other land tenure systems, both the state government and INCRAwhich was gradually assuming responsibility for the areawere challenged with the task of implementing a system that maintains the areas wealth in forest resources while appeasing both the regions land less and the residents of this politically important seringal. Hence, according to local IN CRA officials in Brasilia, it appeared likely that parts of the area would be sub-divided fo r agrarian reform while other areas would be destined for a Sustainable Development Project ( Projeto de Desenvol vimento Sustentvel or PDS ) managed by INCRA. However, as of the tim e of research, neither the proportion of the area to be opened to colonization nor the exact nature of the PDS to be created were yet known. Methods Site Selection An original research site c onsisting of PAE Sta. Quitria and PAD Quixad, Gleba 6, was selected through analysis of maps and in discussion with key informants knowledgeable about the region during an initial visit in 2002. A key factor in selecting the site was the proximate location of these two land tenure units to the fi nal stretch of BR-317 which was to be completed later that year. During subsequent visits, the other sites were a dded, as I felt that they offered interesting comparisons with th e original two sites and also contributed to a more thorough understanding of the variation in the drivers and proce sses of land use change and governance in this area. Seringal Porto Carlos represented a la nd tenure system dating from the rubber era, yet

PAGE 32

32 was in a state of transition as INCRA was cons idering the implementation of an extractive reserve and/or colonization area within it. Als o, due to a growing awareness of the ecological importance of the area as one of the few areas in Alto Acre in which a heavily-forested area bordered the road, NGOs and the government had recently begun offering courses and projects in the area. The Terra Solta represented a co ntrol scenario, in which no formal land tenure system had been implemented and no environmen tal courses or projects had been offered. Finally, due to both the l ack of permanent inhabitants and it s relatively static land use system, Fazenda Rita was not included in quantitative an alysis; however it was included in descriptive portions of the dissertation, due to its signi ficance in the history of the area. Sampling Strictly random sample selecti on proved impractical due to local social realities and census data limitations in the region4. A total of 92 households were selected (aside from Santa Rita which has no permanent inhabitants and where only the ranch manager and several part time workers were interviewed) 41 (44.5%) from PAE Santa Quitria 26 (28%) from PAD Quixad, 15 (16%) from Seringal Porto Carlos, 6 (6.5%) from RESEX Chico Mendes Seringal Etelvir, and 4 (4.5%) from the Terra Solta. Using a st ratified sampling strategy, for each administrative unit, I attempted to include 20% of the total population parameter of households in the sample5 (see Table 1-1). Households were selected usi ng a stratified snow ball strategy that relied both on 4 Due to their remoteness, it sometimes proved impractical to make return visits to households where no one was home during the initial visit. Additionally, due to inform al land sales and exchanges, in most areas, it was not possible to obtain a fully accurate list of residents. Furthermore, early attemp ts to follow a strictly random design drew suspicion from some potential informantswho learned that they had been chosen as respondents due to their location on a list provided by a government agency. While potentially less statistically significant, as a foreign researcher, I found that a stratified sn owball approach allowed me to balance the goals of obtaining a representative sample and operating within local social norms (that is, as king to conduct interviews after being introduced to an individual by a prior acquaintance / informant in a non-interview setting). 5 In most cases, due to time and logistical constraints, samples more closely approximated 15%. Exceptions were the Terra Solta and RESEX Chico Mendes where, in an effort to compensate for small populations, I sampled larger proportions of the total population.

PAGE 33

33 satellite imagery and past informants. I used sate llite imagery of the area to ensure even spatial distribution, and that households with large, medium and sma ll-scale land clearing would be represented. In most cases, I relied upon past informants to introduce me to potential new informantsthus reducing the anxiety and distrust that might otherwise be associated with an interview by an unknown researcher. I also reli ed upon information from early informants to maximize the representativeness of sampled house holds in each community, such as including both migrants and non-migrants (if present) and members of various different extended families in a community. Data Collection Data collection occurred during May-June, 2002, May-June of 2003 and again from June, 2004 until July, 2005. A subsequent visit during May-August 2006 provided an opportunity to obtain information about changes during the preceding year (e.g., wildfires in various communities in the research site and a state-wide ban on pasture burning). During 2002 and again in 2003, I met with potential research collaborators and also with local community leaders, in bot h Brasilia and the actual field site, to discuss the proposed research project, both explaini ng my research and seeking feedback regarding its proposed content, form and potential relevance to the community. In the 2004-2005 field visit, while many informants were already somewhat aware of my research project, prior to interviews, the projects content, direct relevance and limitations were explained. Potential interview subjects were also informed about the optional nature of participa ting in the study. Semi-structured interviews with heads of sampled households represented a central research method. Interviews were conducted with household heads, most commonly male though sometimes female. Interviews ranged from 30 minutes to three hours. I applied a semistructured questionnaire (Appe ndix A). The questionnaire addr essed questions of household

PAGE 34

34 demography, participation in governance, land uses and environmental attitudes. The information from the questionnaire was writte n in a notebook and recorded using a handheld audio recording device. Semi -structured and unstructured in terviews with other household members (spouses, elderly parents, etc) gave additional insights. Accompanying on-farm interviews, when possible (approxi mately half of all interviews), I conducted farm-transects accompanied by the landowner. These walks provi ded an opportunity to assess the accuracy of interview data and gain additional input from farm ers that did not always arise within the more formal context of a sit-down interview. In addition to the household sample, key inform ants in the research area provided further data relevant to the research question. I conducted key informant interviews with governmental leaders at various levels, includ ing the presidents of 11 local associations, th e president of AMPAESQ, the vice-president of AMOREB, the president of th e STR, employees of SEATER ( Secretary of Technical Assi stance and Rural Extension ) various NGO representatives, members of the directory of the Center for Small Rura l Farmer Associations of Epitaciolndia and Brasilia ( Central de Associaes de Pequenos Produtor es Rurais de Epitaciolndia e Brasilia or CAPEB), Brasilias Secretary of the Environment and representatives of INCRA, IMAC and IBAMA in Assis Brasil, Brasilia and Rio Bran co. These interviews were complemented by informal conversations on other o ccasions with many of the same individuals as well as members of the local media and the prefeitos of Brasilia and Assis Brasil. Finally, in order to develop a deeper et hnographic understanding of governance and its articulation in this region, I relied heavily upon participant observation. Fo r instance, in both the field site and Brasilia, I freque ntly participated in associatio n meetings and assemblies as well as meetings and courses organized by IBAMA, STR and other NGOs. In addition to offering an

PAGE 35

35 opportunity to witness the on-th e-ground articulation of governan ce, they provided a forum for me to explain and answer questions about my re search project to members of the community. Participant observation in day-to-day farm ac tivities such as hunting, fishing, collection of cupuau fruits and cattle v accination gave me a better unders tanding of the material and symbolic significance of productive activities based on both forest and non-forest land covers. In order to create a spatially-linked databa se, GPS points were taken for all landholdings included in the study. Points were collected for im portant features such as homes, associations, road intersections and farm and administrative boundaries. Sate llite-based maps proved useful for several purposes. They assisted me in the se lection of spatially distributed farms with low, medium and high levels of defore station. Prior to entering the fi eld, a land cover map of the area was created using recent LandSat 7 ETM imagery in Erdas Imagine. Administrative and (in the case of PAD Quixad) property boundaries were obtained through the IN CRA office in Rio Branco and overlaid onto the image in order to create a field map. Using a handheld GPS unit, I obtained UTM coordinates for houses, fields a nd property boundaries wh ich could then be correlated with their location in the imagery. I conducted participatory map interpretation with those informants who were both able and interested in doing so6. This served several purposes it proved a useful tool in aiding informants memories about past land use and in offering comments about land use on neighboring farms. This often led to rich disc ussions by local resident s including, for instance, associating various family histor ies to specific areas, and thus allowing me to link people to 6 Due to the difficulty that some (mostly older) informan ts faced in understanding and interpreting the map, this technique was neither applied systematically nor used in quantitative analysis.

PAGE 36

36 pixels 7 (Liverman, et. al., 1998). It also allowed fo r an analysis of social networkswith many informants correctly associating most or all of the satellite-det ected land clearing plots in the community with their owners. Data Analysis In the process of data anal ysis, I first conducted descrip tive statistics to gain an understanding of the means and variances inhere nt in the various independent and dependent variables. Subsequently, I conduc ted bivariate analysis of indi vidual independent and dependent variables8. Bivariate analysis allowed me to te st for both linearity and significance of relationships between independent and dependent variables. Give n the fairly small size of the sample of households, significance was established as p<.1. Rela tionships that were both linear (or curvilinear) and significant were then subjecte d to multiple regression analysis in order to control for possible effects of c ovariates. An exception to this rule was made when bivariate analysis revealed non-significan t linear relationships and a suppression effect from covariates was suspected. In these cases, multiple regres sion analyses were conducted to control for suppression effect from covariates, even when re sults from initial bivariate analyses were not significant. Limitations Several data limitations should be noted as caveat s in the data analysis that appears later in the dissertation. As previously noted, due to on-the-ground limitations, a strictly random sample design was not followed. Hence, caution is du e when extrapolating these results beyond the 7 Examples included, for example, a woman who told me that an area of her property was light green because she and her husband had lost their cattle herd at the time th e imagery was taken, causing the pasture to grow up. Another family explained that a small fore st fragment on their property visible on the map was an area of forest they had left, upon a parents suggestion, in order to preserve an exceptionally strong spring that lay within it. 8 The specific bivariate test employed depending on the nature of the independent and dependent variables (e.g., continuous versus categorical). Bivariate analysis included Pearson correlations, t-tests and F tests.

PAGE 37

37 sample. However, I have attempted, to the greates t degree possible, to mini mize potential bias in the sample by including households from the various communities in the field site. Within each community, I also attempted to minimize bias in the sample by including households from various extended families and with small, medium and large amounts of forest clearing. While diverse, the sample does not necessarily accura tely represent the propor tion of different subgroups in the population. A second caveat applies to analyses of the impacts of governance, at the level of the administrative unit, for land use. Given both th e small population parameters and small samples from two of the administrati ve unitsTerra Solta and RESEX Chico Mendessamples cannot be assumed to be representative of the entire administrative unit. In the most extreme case of Terra Solta (4 households sampled from a population parameter of approximately 14), I ultimately elected to remove this administra tive unit when testing the relationship between governance at the level of the admi nistrative unit and land use. A third limitation comes from the possibil ity of inaccuracies, both deliberate and accidental, in reported data from informants. This was especially of concern when obtaining data about household deforestation. Due to th e fact that many households had exceeded their legal deforestation limits and were continuing to clear land, this variable was especially susceptible to under-reporting. In order to minimize the implications of this possibility in data collection, prior to arriving at a given househol d, I developed an approximate estimate of deforested area on the property using both visual analysis of satellit e imagery and personal inspection of the pastures that usually surr ounded rural homes. When reported estimates of deforestation differed greatly from my own, I would attempt to triangulate deforestation estimates with those given by neighbors to determine the source of error. As a last resort, in the

PAGE 38

38 few cases where I was unable to reconcile reporte d figures with my own estimates, deforestation values were recorded as missing.9 9 An example included an elderly divorcee who estimated th at she had some 10 hectar es of pasture, when both satellite imagery and my own visual calculation were closer to 40 hectares. As the sons who cared for her cattle herd were not present during my initial and subsequent visits, I coded the area of pasture and of total deforestation as missing in this case.

PAGE 39

39 Table 1-1. Sampling by Administrative Unit Households Total Households % of Households Sampled (approximate) Sampled Terra Solta 4 14 28% Seringal Porto Carlos 15 100 15% PAD Quixad 26 150 17% PAE Santa Quitria 41 277 15% RESEX Chico Mendes 6 17 35%

PAGE 40

40 Figure 1-1 Field Site in Context of Acre and Brazil

PAGE 41

41 1 1 2 3 4Administrative Units: 1=PAE Santa Quiteria2=PAD Quixada3=RESEX Chico Mendes 4=SeringalPorto Carlos 5=Terra Solta6=FazendaSanta Rita Green=Forest (Primary and Secondary); Red=Deforested (Pasture and Crops) Bolivia2 10 kms N5 6 Figure 1-2 Research Site with Administrative Units

PAGE 42

42 CHAPTER 2 THE POLITICAL ECOLOGY OF ROADS, GOVERNANCE AND LAND USE IN THE AMAZON A large body of literature addresses the issue of land use and land cover change, with a significant portion focusing specifically upon the Amazon Basin. While some broad treatments and meta-analyses of the topic of LULCC exist (e.g. Geist and Lambin, 2001), due to the breadth of the topic, most research has focused on specifi c manifestations (e.g. defo restation) and drivers (e.g. agricultural expansion, public policy) of LULCC. However, due to the diverse disciplinary backgrounds of LULCC researchers, while research in this subject area has given great insights into the various factors driving LULCC, much less work has been done to advance theoretical understandings of LULCC. In this study, I draw upon political-ecol ogy to both focus upon certain phenomena and processes occurring in th e Inter-Oceanic corridor and to understand the relations among them. In this chapter, I begin by offering a discussi on of the LULCC literature as it relates to the topic of roads and governance in the Brazilian Amazon. I first discuss the major land cover change affecting the study site —deforestation. I then discuss two land use practices that are intimately tied with deforest ation in many portions of the Brazilian Amazon—cattle ranching and the use of fire. I then address two driver s of LULCC that are of special concern to the research topic—roads and public policy. Subsequently, I discu ss the theoretical background of political ecology, especially as it relates to the issues of govern ance and scale. Then, drawing upon the LULCC and political ecolo gy literature, I outline and operationalize the independent and dependent variables used in the dissertation as well as th e hypothesized links existing among them.

PAGE 43

43 Land Use and Land Cover Change in the Brazilian Amazon Considered broadly, the over-arching dependent variable considered in the study is land use and land cover change (LULCC). In recent years, the subject of land use and land cover change has gained recognition as an emerging research agenda of profound theoretical and policy implications. While theoretically diverse, much of the LULCC literature is united in striving to transcend simplified narratives of causality such as population growth and shifting agriculture, looking instead to develop a more nuanced understanding of the causes of LULCC (Angelson and Kaimonwitz, 1999; Geist and La mbin, 2001; 2002; Schmink and Wood, 1992). Countering meta-theories which attempt to offer br oad explanations of deforestation, researchers have generally argued that tropi cal deforestation is determined by different combinations of proximate causes and underlying driving forces, varying between geogra phical and historical contexts (Geist and Lambin, 2002). Land use and land cover change is becoming an increasingly important interdisciplinary field of study. Growing human populations and technological change are associated with rapid changes in global land use and land cover— with anthropogenic land covers becoming increasingly common. In the early 21st century, crop and pasture la nd has come to represent a dominant land cover globally—rivaling forests— extending to approximately 40% of the Earth’s terrestrial surface (Foley et al., 2005). Land use change often has direct impacts upon land cover. For example, the shift among inhabitants of the rural Brazilian Amazon from extractivism toward cattle ranching is associated with a shift fr om forest to grassland (pasture) land cover. Furthermore, land use and land use change can have more indirect impacts upon ecosystems, including for example, changes in water cycles, evaporation and run-off (Foley et al., 2005).

PAGE 44

44 Deforestation Deforestation, at the household level, is a key dependent variable cons idered in the study. A large amount of literature has been written about deforestation in the Brazilian Amazon, including both its po litical-economic and ecological im plications (e.g., Schmink and Wood, 1992, Wood and Porro, 2002). Paleoecological and paleoclima tic records suggest that the Amazonian forest ecosystem has been in continuous existence for more than 55 million years. However, despite the Amazon’s resilience to the various climatic and geological changes occurring during its existence, th e continued resilience of the Amaz onian forest is uncertain in light of recent anthropog enic changes, including, most cons picuously, deforestation (Maslin et al., 2006). During the twentieth century, approx imately 16% of the Amazon forest area was cleared, primarily for cattle pasture and, more recently, for the cultivation of soybeans (Fearnside, 2001a; Maslin et al., 2006). Estimates of recent rates of deforestation range between approximately .38% and .5% of existing forest per year (Maslin et al., 2006; Houghton et al., 2000; Davidson and Artaxo, 2004). Under the current “business as usual” scenario (without increases in governance), resear chers have argued that, by 2050, nearly half of the Amazon’s original forest cover may be lo st (Soares-Fil ho et al., 2006). Much has been written about the actual and likely future impacts of deforestation at various scales. Climate change is an implication of tropical deforestation that is likely to be felt at local, regional and global levels. As the Am azonian forest plays an important role in both moderating temperatures and in recycling water into the atmo sphere during the dry season, on local and regional levels, large-sc ale deforestation may result in warmer and drier conditions in the region (Foley et al., 2005). This, in turn, would increase the region’s susceptibility to wildfire—further endangering the forest.

PAGE 45

45 Research has also addressed the risks that carbon released through deforestation and associated activities (such as fire) may have fo r regional and global climate. Research suggests that the Amazon forest, in an undisturbed state, functions as a net sink of atmospheric carbon (Carvalho et al., 2004). However, under curren t land use practices—inc luding logging, pasture burning and deforestation, Amazonia has become a globally significant source of atmospheric carbon. Amazonian deforestation and associated burning releases some 2-4% of annual global carbon emissions (200-300 million tons), representing more than 2/3 of Brazil’s total (Fearnside, 1997; Houghton et al., 2000; Carvalho et al., 2004). Cattle Ranching and Pasture Formation Cattle production represents an important depe ndent variable considered in the study. In the Brazilian Amazon, as in many parts of Latin America, cattle ranching has been the most common form of agricultural expansion into prev iously forested areas (Geist and Lambin, 2002). Brazil’s cattle herd is the worl d’s largest (Kirby et al., 2006). While the majority of Brazil’s cattle production occurs further south, production in Amazonia has been increasing in recent decades. Historically, the spread of cattle ranc hing in the Brazilian Amazon has been accelerated through governmental policies (Fearnside, 2000; 2002a). Seeing large-s cale ranching as a productive use of the Amazon Basin, the Brazilian government until recently offered heavy subsidies for this industry; be tween the years of 1971 and 1987, th e cattle industry received some $5 billion from the Brazilian government (Hall, 2000) After Brazil’s return to democracy, most subsidies for cattle ranching in the Amazon were removed. Due to both the scale and the relatively perm anent nature of most cattle pastures (as opposed to swidden plots which, if not planted in past ure, are usually allowed to return to forest), cattle ranching represents the single most importa nt proximate driver behind deforestation in the Brazilian Amazon (Hecht, 1985; Fearnside, 1990; Cardoso, 2002). Various authors have argued

PAGE 46

46 that cattle ranching, in the abse nce of governmental supports, is generally not profitable in the Amazon (Hecht and Cockburn; 1990; Schwartzman, 1992; Fujisaka et al., 1996). Due to mineral deficiencies in most Amazonian soils, it has been argued that pasture maintenance requires chemical treatments to maintain productiv ity—including phosphates—a substance with few known deposits in the Amazon region (Fearnside, 2002a ). Furthermore, due to the difficulty of preventing natural succession and weed development, in situations of ab undant inexpensive land (a scenario that no longer applie d in much of the Amazon and most of Alto Acre as of the mid00’s), it has sometimes proven less expensive to si mply clear new forest tracts for pasture than to rehabilitate older degraded ones (Fearnside, 2001). Many au thors further argue that land speculation has been a major driv er of pasture creation, with i ndividuals anticipating rising land prices as new roads gradually br ing forested lands into greater connectivity with market centers (Hecht and Cockburn, 1990). However, other authors, while considering pa sture expansion as an important driver of deforestation, present a different and somewhat more complex picture of its social and ecological implications. Faminow (1998) ar gues that, rather than being an outgrowth of governmental subsidies, the expansion in cattle production in the 1990’s stemmed primarily from the growth in regional demand, associated with an expanding and increasingly affluent urban Amazonian population (see also Browder and Godfrey, 1997). In rural Amazonian communities, in addition to serving as a capital investment, cattle re present an easily transportable and liquidable commodity, thus serving as an insu rance substitute. That is, cattle represent a fairly secure form of stored wealth that can be quickly transformed into cash in cases of medical or other emergencies (Perz, 2001a). Furthermore, unlik e horticulture and rubbe r extraction, the labor requirements of cattle ranching are relatively lo w and product transportation is generally simple

PAGE 47

47 (Arima and Uhl, 1997). And, given appropriate agri cultural practices, such as rotational stocking (Rueda et al., 2003) and grass-legume mixes (C arneiro and Valentim, 1997, Burns et al., 2004) productive pastures can be maintained for decad es or indefinitely while chemical inputs are reduced or eliminated. In an extensive discussion of the relations hip between cattle produc tion and deforestation in the Amazon, Faminow (1998) points to the fact that many of the dire predictions offered by scholars in the 1980’s regarding the impacts of ca ttle expansion assumed an exponential growth in deforestation rates, which in most cases has not been realized. Furthermore, he finds fault with arguments coupling land speculation and pa sture degradation as drivers behind pasture expansion. Were this train of thought correct, he argues, it woul d assume an infinite supply of irrational land speculators willing to pay increm entally higher prices fo r increasingly degraded land. Faminow begins his argument by countering a ssertions that Amazoni an soils are unfit for long-term agriculture, pointing to the similari ties between Amazonian soils and those of the Southeastern U.S., many of which have been in continuous cultivation for more than 200 years. Furthermore, in light of recent evidence from the Amazonian cattle industry, Faminow and other researchers have questioned the assumed lack of viability of cattle ranching as an economic endeavor—pointing out that, desp ite the reduction of government subsidies, cattle ranching has generally been shown to be economically productive (Faminow, 1998, Mattos and Uhl, 1994, Arima and Uhl, 1997; Machagata and Brown, 2003). Under such arguments, th e spread of cattle ranching arises less as a result of speculati on and land degradation th an as a function of economic decision-making by property owners oper ating in the existing political and economic system (Machagata and Brown, 2003).

PAGE 48

48 Various researchers have further argued th at it is less cattle pr oduction per se than inefficient grazing practices that threaten Br azil’s tropical forest s (Arima and Uhl, 1997, Faminow, 1998). Researchers ha ve found that heavier stocki ng rates (stocking rate > 1 head/hectare) and better managed pasture rota tion, by minimizing the accumulation of senescent forage, actually improve long-term pasture quality (Machagata a nd Brown, 2003; Faminow, 1998). Arguing that cattle-related defo restation currently arises from herd expansion rather than land speculation, Machagata and Brow n assert that inte nsification of cattl e production, in many cases, will slow deforestation. Coupled with improved enforcement of existing environmental laws limiting deforestation, such intensificati on should be able to improve local livelihoods without raising deforestation rates (Arima and Uhl, 1997; M achagata and Brown, 2003). Fire Another land use practice given cl ose attention in my research pr oject is the use of fire. In the Brazilian Amazon, fire is a commonly-used lowcost tool to establish and maintain pastures, especially in extensive cattle production systems. As most cattle production in the Amazon— especially at smaller scales—is based on con tinuous rather than ro tational stocking, cattle producers often rely upon fire as a strategy to re move the low-quality senescent plant material that tends to develop in such systems. Fire can have positive impacts for pasture development and maintenance, such as the elimination of invading plants, reduced shading, and improved carbon:nitrogen ratios through the release of nitrogen and the volat ilization of carbon in senescent plant material (H eringer and Jacques, 2002). However, it also results in negative effects as well—including the volatilizati on of nitrogen and potassium and increased susceptibility to erosion (Almeida n.d.). And while grass regrowth yields a short-term increase in palatable dry matter production, animal weight ga ins are generally lower than those in similar unburned pastures (Almeida, n.d.). Further negative impacts of fi re include the loss of crops,

PAGE 49

49 fodder, livestock, timber, fences and other infrastr ucture in the cases of wildfire (as occurred in several communities in the field site in 2005). From the farmer’s perspective, fire can have further negative impacts when forest areas are accidentally burned. For example, researchers ha ve noted a general decline in tree fruiting and in the populations of frugivorous vertebrates in bu rned forest in the Amazon (Barlow and Peres, 2006). This in turn can have negative imp acts for local populations who depend on non-timber forest products (NTFPs) and hunting for part of their livelihoods. Th e reduced production and eventual destruction of Brazil nut trees that have been affected by fire is a case in point. While these impacts are of great concern at the level of the household and local community, the impacts of wildfire upon the forest and for the larg er socio-ecological system are of even greater importance at regi onal and societal levels. Of part icular concern are the potential impacts of fire in damaging standing forests, aff ecting regional rainfall an d affecting the health and general well-being of local populations. Burning of pastures tends to kill existing tr ees and prevent regenera tion of new recruits— leading to the long-term replacem ent of native forest vegetation with exotic grasses and fireadapted weeds. Furthermore, wildfire enteri ng surrounding forests can le ad to positive localscale feedbacks setting the conditi ons for future fire (Cochrane et al., 2003; Hoffman et al., 2003; Alencar et al., 2004). While an initial fire will usually remain in the unde rstory, the increase in solar radiation from canopy gaps and in fuel lo ad from dead or damaged trees provides the conditions for successively more intense future fi res (Cochrane et al.,, 2003). Reinforcing this positive feedback is the probable im pact of fire upon regional climat e. Several researchers have drawn attention to the possibili ty that smoke, by augmenting the number of condensation points,

PAGE 50

50 may inhibit the creation of large atmospheric wa ter droplets and thereby reduce regional rainfall (Maslin et al., 2006; Cochrane et al., 2003). In addition to the impacts for rural livelihoods and the natural environment, fire and smoke have immediate negative implications reachi ng both rural and urban populations, including a notable rise in respiratory illnesses (Mendona, 2004). Mendona (2004) estimates that as much as 8% of all hospital-treated respiratory cases in the Brazilian Amazon are di rectly attributable to the burning of biomass; in 1998, an El Nio year in which large areas of the Amazon succumbed to wildfire, the number of cases reached 13,000. Roads Roads are generally recognized as the strongest single predictor of tr opical deforestation due to their numerous direct and indirect impacts upon land use (Kaimonwitz and Angelson, 1998; Geist and Lambin, 2001; 2002; Pfaff, 1997). Road construction in places such as the Brazilian Amazon generally increases the profitabi lity of timber and ag ricultural production and, by improving access, facilitates la nd speculation and colonization (Soares-Filho et al., 2004; Chomitz and Thomas, 2001). In an extensive cr oss-national review of prominent deforestation literature, Geist and Lambin (2001) found that road construction or paving served as a driving factor in two-thirds of the cases examined. Various analytical approaches have been used in the study of roads and their impacts on tropical deforestation. In a ground-breaking paper, Chomitz and Gray (1996) used a spatial econometric analysis in their study of deforestatio n in Belize. Using th is technique, they found that roads were correlated with agricultural land use, with soil fertility acting as a mediating variable. Roadside areas with good soils had an especially high (50%) probability of conversion to agriculture. Common in th e literature on roads and land use has been the extrapolation of historical patterns to future projections within ro ad buffers. This approach has been applied, for

PAGE 51

51 instance, in recent studies of the probable imp acts of current highway paving projects in the Brazilian Amazon (e.g., Laurance et al., 2001). Howeve r, the extrapolation of past to current and future trends is an inherently error-prone proced ure; it requires that othe r factors contributing to land use change (such as government policy) be held constant—a problem that has surfaced in widely differing predictions for future defore station due to the Avana Brasil highway paving program. Historically, Amazonian roads have contribute d to agricultural expans ion and deforestation by changing migration patterns and opening the re gion to populations from other portions of the country in search of land for subsisten ce or for investment (Moran et al., 1980 ; Schmink and Wood, 1992). In the case of Ac re, the opening of roads in the 1970’s and 1980’s led to an increase in in-migration, in turn leading to numerous land conflicts (Bakx, 1986). While government-sponsored colonization of the region has largely ceased, the impact of the nowpaved road for migration—especially between ru ral and urban areas—has continued into the mid 2000’s. Various predictions have been made regarding the likely implications of the Avana Brasil project. Predicted deforestation resulting from the project varies widely—due largely to differing assumptions underlying the predictive models used. In particular, some models extrapolate past patterns of governmental unwillingne ss or inability to create and enforce conservation policy to the future (e.g., Laurance et al., 2001, Fearnsid e 2003), while others (e.g., Nepstad et al., 2001, Silveira, 2001, Soares-Filho et al., 2004) allow for the possibility that recent political changes in Brazil indicate a greater likeli hood that viable systems of fo rest governance can emerge to control land use in roadside forests. This topic w ill be addressed in greater detail subsequently.

PAGE 52

52 Public Policy Public policy is a commonly cited factor affecting deforestati on (Skole et al., 1994), including both overtly pro-de forestation policy and policy th at has deforestation as an unintentional consequence. It also includes policies oriented towa rd forest conservation, as Acre has witnessed in the late 1990’s into the 2000’s (K ainer et al., 2004). As with roadways, policy rarely functions alone in its impacts on land use bu t rather most often occu rs in conjunction with other drivers. Geist and Lambin (2001; 2002) found policy and governance to be a factor in 78% of the cases of deforestation they considered—mos t often operating in tande m with other drivers. This tendency holds true in the case of recent lite rature addressing the combined effects of roads and governance in the Brazilian Amazon (e.g., Neps tad et al., 2001, Laurance, 2001, Nepstad et al., 2002; Fearnside, 2002, Soares-Filho et al., 2004). Among the recent government initiatives in tended to control land use change and deforestation in the Amazon has been agro-eco logical zoning. Dennis Mahar of the Economic Development Institution and formerly a me mber of the World Bank’s POLONOROESTE project, describes the mixed successes encount ered by an ambitious agro-ecological zoning initiative in the Amazonian state of Rondnia (Mahar, 2000). The prescriptive zoning program was intended to impose order in a state that had witnessed the squandering of forest resources and displacement of traditional peoples dur ing a 1980’s land rush accompanying the World Bank-funded paving of the BR-364 highway through th e state. Under the zoning plan, land use was to be regulated based upon both current prac tices and ecological suit ability. Despite the program’s relative successes in controlling defo restation, Mahar points to several limitations. Among the criticisms include the fact that few incentives exist for loca l-level governments to enforce zoning rules and that perver se incentives exist for land owners to deforest illegally in the hope of changing their land’s zoning status.

PAGE 53

53 A major platform of the Acre State Governo da Floresta has been Ecological-Economic Zoning (ZEE) (Kainer et al., 2004). The concep t of ZEE stems from language in the Federal Constitution promoting the decentralization of governmental decision making—especially in the area of environmental protection—including Ar ticles 21, 23, and 30. Presidential Decree number 99.540 of 21 September, 1990 created a Coordinated Commissi on for Ecological and Economic Zoning of the National Territory (Governo do Acre, 2001)—outlining a conceptual and legal framework for ecological and economic zoning to be implemented at the state level— with the Legal Amazon defined as a priority ar ea. ZEE represents an instrument to guide governmental investments in socio-economic deve lopment in a way that recognizes and builds upon the natural comparative advantages of each region. Through ZEE, planners in new agricultural and settlement fron tiers (such as many parts of Acre) have hoped to avoid the anarchy, environmental destruction and social in justice of past occupation processes in the Amazon. Implementation has been underwritten in part by financing from the PPG7 program of “integrated environmental management project s” (PGAI) (Governo do Acre, 2001), with further assistance from the GTZ. In the past, agrarian and social policy in Br azil, by subsidizing colonization and pasture expansion in the Amazon, tended to directly stimulate deforestation (Schmink and Wood, 1987; 1992). Since democratization in the mid-1980’s and the coinciding domes tic and international concern about the fate of the Amazon’s forests, many of the policies promoting deforestation have been changed and new environmental regula tions (e.g., establishment of national parks and 80% forest reserve regulations on private land) have come onto the books. However, the ability of the government to effectively enforce these re gulations is still in question and much of the land conversion occurring in the Brazilian Amazon t oday continues to be ill egal (Carvalho et al.,

PAGE 54

54 2002). Compounding this problem, under-staffed Brazilian authorities often have great difficulty detecting and combating violat ions (Carpentier et al., 2000). Perhaps the largest debate regarding road paving and deforestation in the Brazilian Amazon centers upon the ability of frontier governance systems to mediate this relationship. Several researchers have offered qualified optim ism in their projections for Avana Brasil’s impacts upon deforestation. Soares-Filho et al ., (2004) estimate that deforestation in the Brazilian Amazon may be reduced by 1.3 million hectares by 2050 given a governance scenario in which current environmental regulations are enforced. Such predictions are often based upon the shift at the federal and, in some cases, at the state level away fr om policies that have historically promoted deforestat ion in the region’s highway corrido rs. An early example of this shift was the creation and implementation of the RESEX model due to polit ical pressure from social movements in the 1980’s and early 1990’s (Schmink and Wood, 1992). More recently, at the local level, researchers have noted th e growing capacity of local governments in environmental and development planning and the ove rall increasing effectiveness of governance as a means of minimizing deforestation (N epstad et al., 2002; Silveira 2002). Nepstad points to the Environmental Crimes bill of 1998, authorizing IBAMA to levy fines and jail sentences for unauthorized fire and land clearing and the spec ific case of PROARCO—a recent fire control program—as evidence of th e potential effectiveness of governance in positively affecting land use change. Subseque nt to the implementation of PROARCO in 2000, Nepstad notes a two to four-fold decrease in the number of fires in satellite images of heavilysettled areas of the southern and eastern Amazon (Nepstad et al., 2002). Other researchers have found corroborating evidence for the effectiven ess of governance as a means of minimizing deforestation in other parts of Brazil that have not necessarily been recently affected by road

PAGE 55

55 paving. For example, Fearnside (2003) notes that in the state of Mato Grosso, a state-level deforestation licensing and en forcement program has demons trated positive results. Furthermore, areas of the state with the str ongest enforcement have witnessed the greatest declines in deforestation rates. A counter-argument asserts that Brazil’s return to democr acy and the introduction of environmental legislation will not be sufficient to significantly slow deforestation along highway corridors. While environmental regulation ha s generally increased since the mid-1980’s, Laurance et al., (2001) note th at, despite a dip during the early 1990’s, deforestation rates actually increased during the late 1990’s, largely due to improve d economic stability resulting from the federal government’s Plano Real monetary reform program. Like the Plano Real program, even various governmental incentives desi gned to be pro-forest or forest-neutral have had perverse effects. For example, Almeida a nd Uhl (1995) illustrate how the Brazilian Rural Land Tax—a tool intended to encourage effici ent and productive land use—actually penalized farmers who wished to invest in forest-conserva tion and intensification of cattle production. In the case of Avana Brasil, Laurance estimates that between 269,000 to 506,000 hectares of forest land will be lost annually due to the program (Laurance et al., 2001). Non-enforcement of existing laws has been a commonly cited concern; despite improvements in remote monitoring, it nonetheless remain s difficult to monitor land us e and enforce regulations, especially in remote areas. According to this perspective, the problem is exacerbated by corruption at various levels. Rese archers have also pointed to pr oblems intrinsic in the political process behind Avana Brasil and similar projects that may contribute to deforestation. Fearnside (2003) notes several sp ecific problems in the envi ronmental legislation process common in Brazil. In many projects, such as Avana Brasil, the project is proposed on its

PAGE 56

56 own—without a suite of alternative projects. Fu rthermore, efforts are often made to secure funding prior to impact assessments—consequen tly building political support that will help ensure the project’s ultimate vitality regardless of the environmental impacts. Finally, little impact assessment comes from within the project programs, but as in the case of Avana Brasil, studies tend to come largely from independent research organizations outside of the project. Political Ecology My research draws heavily upon a political-eco logical approach to land use and land cover change in this region. A polit ical-ecological perspective expl ores the relationships between human society and natural resources, giving special attention to the role of politics as individuals and groups contest rights to th ese resources (Bryant and Baile y, 1998). The term “political ecology” can be traced to the early 1970’s in Er ic Wolf’s call for an approach integrating land use with a local-global political ecology (Wolf, 1972; Peet and Wa tts, 1996). In response to both Neo-Malthusianism and closed-system ecologica l anthropology, early po litical ecology drew upon a Neo-Marxist perspective to bring politics and class struggl e into what previously had been a largely apolitical view of human-environment interac tions (Schmink and Wood, 1987; Peet and Watts, 1997; Bryant and Bailey, 1998). Informing the environmental movement of the 1960’s and 1970’s, Darwinism and NeoMalthusianism saw unregulated population growth as a key factor in global environmental degradation (Peet and Watts, 1997). For example, the acceleration of tropical deforestation at this time was, in both academic circles and in the popular media, often blamed on faceless rural peoples whose unabated population grow th drove them ever further in to the forest in search of cultivable land (Jarosz, 1996). S eeing these explanations as inad equate, early political ecologists looked to political and social factors to better explain phe nomena such as environmental degradation and famine (Wa tts, 1983; Moore, 1996).

PAGE 57

57 Systems-based ecological anthropology repres ented another key paradigm spurring the early development of political ecology (Watts, 1983, Peet and Watts, 1997). In particular, political ecologists drew excepti on to the homeostatic equilibriu m-based models then dominant in ecological anthropology (e.g. Rappaport, 1967, Orlove, 1980). Such work tended to focus upon energy and resource flows in closed-syste m societies (Bryant a nd Bailey, 1998). For instance, in the book Pigs for the Ancestors Roy Rappaport (1967) used a systems ecology approach to explain the liveli hood system of the Maring—an isolat ed tribal group living in New Guinea. As part of his argument, Rappaport claime d that ritual slaughterin g of pigs served as a means whereby the Maring maintain ed equilibrium within their environment and society. In addition to his use of an equi librium-based model, Rappaport was also criticized for his inattention to the wider politic al-economic context in which th e Maring lived. In contrast, political ecologists argued that human-environmen t interactions could only be understood within the context of local-g lobal articulations and linkages am ong the local community, the nationstate and international institutions (Biersack, 1999). Political ecology emerged as a distinct theoretical framew ork in the mid-1980’s through analyses of the political-economy of environmenta l degradation. In a discussion of soil erosion in the developing world, Piers Blaikie (1985) ar gued that environmental decision-making at a given scale is best explained when contextualized into the political-econo mic context of the next larger scale. Consequently, he called for a botto m-up research perspective which begins with the smallest decision-making level and progressively scales upwar d to the level of state and international political-economic st ructures in order to understand the forces shaping local level land use.

PAGE 58

58 In one of the first works to apply an explicitly political ecological approach to the Amazon, Schmink and Wood (1987) discussed the ways in which the wider socioeconomic and political context of the Brazilian Amazon have affected human use of natura l resources while also looking at the implications of this relationship for environmental po licy in the region. Schmink and Wood proposed a political-ecological model that links the Amazonian political-economy— including, among other factors, th e state, markets, and class st ructure—to patterns of resource exploitation in the region. Li ke Blaikie, Schmink and Wood focused upon local scale land use decisions, but saw them as being largely de termined by larger-scal e political-economic phenomena like capitalist expansion. In their study of frontier expansion in the Brazilian Amazon, Schmink and Wood (1992) considered frontier expansion and land conflicts surrounding the town of So Felix do Xingu in the state of Par. This discus sion is couched within analysis of the regional, national and international settings. The authors offered a temporal contextualizat ion of the problem—placing the processes occurring at these multiple levels within a broad historical context encompassing nearly 500 years. In this study, the authors e xpanded beyond this largely top-down framework in their earlier work. The authors addressed not only the ways in which political and economic structures have affected land use in the So Felix do Xingu region. They also addressed the ways in which individual actors, through their land use and through their contestation of rights to land, shaped the wider political and economic e nvironments. Hence, they brought a sense of individual agency to questions of political-economy and land use th at had largely been absent in earlier political ecological work. Wood and Porro (2002) offer an implicitly political-ecological framework for conceptualizing the relations hip between the socio-economic and biophysical drivers of

PAGE 59

59 deforestation in the Amazon Basin. In this mode l, the ultimate agent of land use and land cover change—the household or firm—is contextua lized within increasingly larger-scale socioeconomic and biophysical drivers. The h ousehold or firm’s land use and land cover outcomes, in turn, have feedback effects for all scales of socio-economic and biophysical drivers. Hence, Wood and Porro, like Schmink and Wood (1992), introduced a multi-directional chain of causality into their model of LULCC. However, the feedback effect occurs not through the direct agency of the household (or firm) respondi ng to and contesting the larger scale politicaleconomic context, but through the land use and land cover outcomes themselves. Through the 1980’s, political ecology tended to draw upon a neo-Marx ist perspective to look at local conflict largely in terms of capitalist penetrati on and class conf lict (Bryant and Bailey, 1997). Subsequently, however, political ecologists ha ve begun to draw upon a wider range of theory to explain human-environmen t relationships—including, among others, gender analysis. Using a gender analysis approach, research ers have found that the traditional focus upon social class in political ecology has overlooked other ke y variables affecting humanenvironment relations (Sontheimer, 1991; R oucheleau, et al., 1996; Schmink, 2000). Environmental policies may have very different impacts for men and women when they do not account for gendered patterns of resource access and use (Schroed er and Suryantata, 1996). For example, in her study of a government-sponsored agroforestry project in the Gambia, Carney (1996) found that, by overlooking th e role of gender in shapi ng local resource access and the traditional male control of tree resources, the proj ect effectively shifted the balance of resource control toward men and encouraged the economic marginalization of women in the affected communities.

PAGE 60

60 While offering a gender analysis of governance and land use is not the primary purpose of the dissertation, attention to gender provides an understanding of the va riable ways in which roads and governance affect households and individuals within households in the BR-317 highway corridor. Specifically, I address the variable ways in wh ich the road has affected men and women, addressing gender-sp ecific migration patterns between rural and urban areas. By tying this analysis with gendered land use pr actices (i.e. approaches to cattle production), I demonstrate the way in which gender can serve as a mediating factor between road paving and land use. Governance and Institutional Design Although it doesn’tWhile not falling directly wi thin the rubric of political ecology, other like-minded authors have addressed issues of govern ance and their implications for land use. Complementing political ecology, a considerable body of literature addresses the questions of institutions and institutional de sign and the implications for the management of collective resources (Ostrom, 1990, Agrawal and Gi bson, 1999, Ostrom, 2005). Ostrom defines institutions as “the prescriptions that humans use to organize all forms of repeated and structured interactions including those within families, neighborhoods, markets, firms, sports leagues, churches, private associations, and governments at all scales” (Ostrom, 2005: 3; see also: Agrawal and Gibson, 1999; Kiser and Ostrom, 1982; and Adger et al., 2003). Institutions may represent mechanisms of self-organization by communities or may be created and imposed by external entities such as state governments (Poteete and Ostr om, 2002). Institutions include both the overall institutional framework and the specific institutional arrangements of particular sets of rules—and often form nested structures in whic h larger-scale institutions limit the alternatives available at lower levels (Adger et al. 2003).

PAGE 61

61 Like political ecology, institutional theory developed in opposition to Neo-Malthusian theories forecasting the inevitable demise of co mmon pool resources not s ubject to privatization or state control. For example, in his much -cited study, Garret Hardin (1968) argued that in common pool resource systems—such as commonly held grazing pastures—the rational choice of each participant is to overuse the resource as s/he appropriates full benefit from each marginal unit of use while the impacts of resource degradation are distributed among all members. Such authors recommended alternatively for government control or privatiz ation (Demsetz, 1967; Posner, 1977). Implicit in both arguments is an assumption that local resource users are incapable of designing effectiv e means of self-governance (Ost rom, 1997). While institutional theorists generally accept that Hardin’s “trage dy of the commons” may in fact apply to truly unregulated open-access resources, in many cases communities have been able to avoid this situation through institutions of self-governance—allowing joint use that is not permitted to exceed some threshold of extraction (Ostrom, 1990; Ostrom et al., 1999; Agrawal, 2001) While institutions shape nearly every aspect of human life, a significant body of literature focuses specifically upon the institutions of land use. According to Agrawal and Gibson, institutions of land use face severa l shared tasks: they establish rules about the use, management and conservation of resources, the implementation of the rules that are created and the resolution of conflicts that consequently arise (Agraw al and Gibson, 1999). Ostrom (2005) points to several elements of institutiona l design that are generally requis ite for effective management of shared resources. Among the elements of su ccessful design are: clearly defined boundaries, proportional equivalence between benefits and co sts, graduated sancti ons, conflict-resolution mechanisms, collaboration between local and governm ent authorities in the enforcement of rules

PAGE 62

62 and trust between members—whethe r through shared cultural norms or repeated interactions, or both (Ostrom, 2005). Political Ecology in the Dissertation In the dissertation, I have utilized a polit ical-ecological appro ach to explore the relationship between government ag rarian, forest and infrastruc ture policies and household and association-level land use decisions. In this proc ess, while I have focused upon the household as the unit of analysis, I have framed the study of a household’s land management decisions within the association and the administrative unit, through the state, federal, and international contexts of environmental governance. While the disse rtation does not focus exclusively upon political drivers of land use change, it is grounded in an understand ing that the multi-directional relationship between a household an d wider political arenas plays an important role in shaping how land use decisions are made. By considering local-level gove rnance systems and the role of individuals in interpreting, responding to and reshaping governance system s through their land use practices, I have attempted to avoid deterministic interpretations based on top-down chains of causality between large-scale government policies and an individual ’s land use decisions. In the dissertation, I use a political-ecological approach to explore the relationship between centralized governance systems and household-level land use decisions. Similarly, I explore the implications that a household’s elective decision to participate in or opt out of th e various institutions of governance available to them affects their land use. I al so address the ways in which individuals and households, through their participa tion in wider structures of governance, reshape the landscape of environmental governance. In a 1983 essay entitled “Again st Political Ecology”, (Va yda and Walters, 1999), Andrew Vayda and Bradley Walters offer several cri tiques of political ecology, including the scant

PAGE 63

63 attention it often gives to ecology. However, most important, I would argu e, is their argument that political ecology makes an a priori assumption that politics is al ways central to the use of natural resource issues. In this article, they draw on a case study of mangrove deforestation in the Philippines to show how other non-political factors, such as tidal action and local preferences for certain timber species have been equally, if not more important than politics in shaping the extent and composition of the mangroves. An exclusive focus on political contestations for rights to resources would have overlooked othe r important explanations for deforestation occurring in this region. While not seeing it as a ground for the out-ri ght rejection of polit ical ecology, Vayda and Walter’s concern about the tend ency in political ecology to prioritize politics as the a priori explanation for any and all phenomena involving na tural resources has guid ed my thinking as I apply a political ecological pers pective to the topic of roads, governance and land use. For instance, in quantitative analyses I recognize the importance of and control for largely apolitical explanations of land use such as property size an d years of residence. Similarly, in the final chapter, when exploring the contradictions betw een the concern that pe ople express about the environment and their actual land use practices, I recognize that multiple factors, including, but not limited to politics, are likely involved. Another critique that has been leveled agains t much political ecological literature, even that which addresses questions of agency a nd feedback loops between political economic structure and local land use decision s, is the tendency to make an a priori judgment that political decisions made at the local level are preferable to those made at larger scales. For example, Brown and Purcell (2005) argue th at the question of scale should be treated as an object of inquiry in political-ecological st udies. They note that much polit ical-ecological work has fallen

PAGE 64

64 into a “local trap,” in which local scale actors and organizations are assumed to be more effective in promoting positive outcomes such as environmen tal sustainability and social justice. They argue that political-ecological research should carefully addre ss the question of scale without making a priori assumptions that one scale should be preferred over another. The question of scale is treate d implicitly throughout my research project—for example, consideration is given to the land use implications of regional and national scale migration on land use; similar consideration is given to the land use implications of the actions of governmental agencies from municipal to national levels. Furthermore, scale is treated as an explicit question of analysis in two discussions about the re lationship between governance and land use. First, addressing Brown and Purcell’ s concern about the “local trap” in political ecology, I consider the effectiveness of associations operating at two distinct spatial scales: at the level of the administrative unit and of the comm unity. That is, in e ach hypothesis concerning governance and land use, I use two measures of governance—one operati ng at the household level (e.g., participation in courses, STR, etc) an d one at the level of th e administrative unit (e.g., deforestation limits, IBAMA presence, etc) —to assess at which scale (household or administrative unit) the impacts of governan ce upon land use can be best measured. In drawing upon political ecology, I have al so recognized its limitations. While not necessarily a weakness, due to it s interdisciplinary nature and nu merous theoretical influences, political ecology has not develope d into a coherent unified theo ry but represents a broadlydefined approach and area of interest that can be integrated into a divers e array of perspectives on human-environment interactions. Furtherm ore, as political-ecology tends to focus upon narrative explanations of humanenvironment interaction it is rarely, in and of itself, well adapted to rigorous hypothesis tes ting (Peet and Watts, 1996). In th e case of my research, I have

PAGE 65

65 sought to overcome this limitation by drawing both on political ecology and the broader LULCC literature. While theoretically diffuse, the LU LCC literature does provide fertile material for developing testable hypotheses. A case in point is the argumen t between Laurance and Nepstad about whether or not governance re duces predatory land use in road corridors. Hence, I have drawn largely upon the LULCC literature to id entify key variables and hypothesize causal relationships between them. Political ecology a nd institutional theory perform an important complementary role by focusing the analysis on key factors not commonly discussed in the LULCC literature—including individual agency, contestation for the rights to resources and to define the “rules of the game”, and the implicatio ns of scale in affecting the role that governance has for land use. Variables Used in Analysis Drawing upon the LULCC literature, especially that which focuses upon issues of roads and governance in the Brazilian Amazon, I identifi ed five land use and land cover variables which are treated as dependent variables and 10 variables that can be considered drivers of LULCC and are treated as independent variables in chapters 3 and 4. Dependent Variables Deforestation Deforestation is a continuous vari able which refers to the net ar ea, in hectares, of mature or secondary forest that had been removed by the fa mily from the time of arrival on the property. As interviews occurred over a one year period, I used 2005 as a ba se year for calculating this variable. To obtain this inform ation, I subtracted the amount repor ted as deforested at the time of the household’s arrival on the property from the amount deforested at the time of the interview. Household-level defore station represents the area in hectares of forest (primary and secondary) that had been converted to non-forest during the family’s tenure on the property. As

PAGE 66

66 mentioned in chapter 1, this method can be subj ected to under-reporting. However, given the techniques available to minimize this bias (als o addressed in chapter 1) and the difficulty of determining the exact property boun daries in most portions of th e field site, I elected to use reported values rather than remote sensing. Cattle herd size Herd size reflects the total number of cattle owned by household members, including cattle held both on and off-farm. Data were also co llected for on-property cattl e herd size (regardless of owner). As my unit of analysis in the diss ertation was the household a nd not the land parcel, I elected to use the former meas ure rather than the latter. Brazil nuts Brazil nut harvest is a continuous variable measuring the amount in latas (1 lata = 18 kilograms) of Brazil nuts that were harv ested by the household in 2004—including those destined for consumption and for sale. Annuals Annuals is a continuous variable representing the area, in hectares, that the household had dedicated to the production of annua l crops (e.g., beans, rice, manioc maize) at the time of the interview. Last burn Years since last burn is a c ontinuous variable referring to th e number of years that had passed since the last time the household had inte ntionally burned an area of established (as opposed to new) pasture on their property1. Since interviews were conducted over a one year period, I used the year 2005 as a baseline for calc ulating this variable. Years since burning 1 While great variability exists between families in terms of fire use as a pasture maintenance strategy, given the general non-existence of technological alternatives to fire when forming pasture in the region, fire is used nearly universally when new pasture is established.

PAGE 67

67 established pasture reflects the number of years between 2005 and the last time fire was intentionally used in establishe d pasture on the property (excluding fire used to establish new pasture). Independent Variables Place of origin Place of origin was coded as a binomial vari able. Households with both heads born in Acre or neighboring regions of Bolivia or Peru2 were labeled “Acreano” and coded as 0. Households that had at least one head that was born outside of Acre or neighboring areas of Bolivia or Peru were labeled “Migrant” and were coded 1. Distance from road Distance from road is a continuous variable that measures the dist ance (in hours) that a household lies from the Inter-Oceanic Highway. As an individual may use various forms of transportation (walking, bicycle, horse, motorc ycle, truck, etc) on di fferent occasions, I standardized this measure by us ing travel time by foot. Bi-localism (male and female) Male and female bi-localism are dichotomous variables and represen t the frequency with which male and female household heads, respectiv ely, traveled to an urban area each month and were coded as binomial va riables (0=low and 1=high)3. 2 All individuals in the sample who were born in Peru and Bolivia were ethnically Br azilian, most of whom traced their ancestry to people from Brazil’s Northeast who had come to Acre and neighboring regions of Peru and Bolivia to tap rubber several generations earlier. 3 Male and female bi-localism were initially coded as fo ur-category categorical variables (0=never/almost never, 1=less than monthly, 2=monthly, 3=more than monthly, bu t less than weekly and 4=weekly or more). Given the lack of variance in the 0 category (n=1 for both male and female bi-localism) and the lack of large differences between some categories, I recoded the five categorical va riables as binomial variables. For both male and female bi-localism, the lower level (codes 0,1 and 2) was coded as 0 and the higher level (3 and 4) was coded as 1.

PAGE 68

68 Adult labor (male and female) Adult labor (male and female) are continuous variables and repres ent the number of “working age” (15+ years of age) males and fema les in each household. I used household labor as a proxy measure for out-migration. While the number of working age household members does, of course, reflect other factors beyond out-m igration (e.g. births, mortality), it is also intuitive that the effects of fe wer working-aged adults upon land us e would parallel the effects of adult out-migration. Years of residence Years of residence is a conti nuous variable measuring the numb er of years that a household has resided on the property. As fieldwork was conducted over the co urse of a year, this measure was standardized using 2005 as a ba se year (i.e. if a family arri ved in 1985, the value would be 20). This variable serves as a control in multiple regression analysis. Property size Property size is a continuous variable measur ing the area (in hectares) of the household’s land holding. This variable serves as a control in multiple regression analysis. Governance (administrative unit) Unit governance is an ordinal variable based on a four point index of the governance characteristics of administrative un its in the region. This index is explained in detail in chapter 4. Participation in governance Participation in governance is a continuous variab le. It is a factor-weighted index based on principle component factor analys is of 12 individual measures of participation in governance. This index is explained in detail in chapter 4.

PAGE 69

69 Hypotheses Roads and Land Use Within the first general research question a bout the impacts of the road for land use, I address four specific issues and how they relate to land use. First, rega rding the impacts of the road by opening the region to colo nization in the 1970’s and 1980’s, I posited the following four hypotheses about the lasting impacts of place of origin and land use: H 3.1 4 Migrant households tend to exhibit higher rates of deforestati on than those with heads born in Acre. H 3.2 Migrant households tend to own more cattle than those with heads born in Acre H 3.3 Migrant households tend to produce more Brazil nuts than those with heads born in Acre H 3.4 Migrant households tend to burn their pa stures more frequently than those with heads born in Acre Secondly, to address the implica tions of rural-urban migration and bi-localism that have been facilitated by the construc tion and especially the paving of the highway, I posited the following eight hypotheses: H 3.5 Higher levels of female bi-localism are associated with less production of annuals. H 3.6 Higher levels female bi-localism areassoci ated with less producti on of Brazil nuts. H 3.7 Higher levels of female bi-localism ar e associated with larger cattle herds H 3.8 Higher levels of male bi-localism are a ssociated with less production of annuals. H 3.9 Higher levels of male bi-localism are a ssociated with less production of Brazil nuts. H 3.10 Higher levels of male bi-localism ar e associated with larger cattle herds. H 3.11 Lower numbers of working age females are associated with less production of annuals. 4 Hypotheses starting with 3 ar e addressed in chapter 3. Hypotheses startin g with 4 are addressed in chapter 4. For example, H 3.1 is the first hypothesis addressed in chapter 3.

PAGE 70

70 H 3.12 Lower numbers of working age females are associated with less Brazil nut production. H 3.13 Lower numbers of working age females are associated will larger cattle herds. H 3.14 Lower numbers of working age males are associated with less production of annuals. H 3.15 Lower numbers of working age male s are associated with less Brazil nut production. H 3.16 Lower numbers of working age males are associated with larger cattle herds. Thirdly, to address the implications of acce ss to the highway for land use, I posit the following two hypotheses:: H 3.17 The greater the distance from the roadway, the lower the deforestation. H 3.18 The greater the distance from the ro adway, the smaller the cattle herds. To address the implications of the expandi ng cattle economy that ha s developed in the wake of the road’s construction and pavi ng, I posited the follow ing four hypotheses: H 3.19 The larger the cattle hers size, the less area dedicated to the production of annuals. H 3.20 The larger the cattle herd size, the lo wer the production of Brazil nuts. H 3.21 The larger the cattle herd size, the more fre quently pasture is burned. Governance and Land Use To address the role of governance, at the leve l of the administrative unit, in affecting the land use in the highway’s area of influen ce, I posit the following three hypotheses: H 4.1 Higher levels of governance at the admi nistrative unit level correspond with lower deforestation. H 4.2 Higher levels of governance at the admi nistrative unit level co rrespond with smaller cattle herds. H 4.3 Higher levels of governance at the admi nistrative unit level correspond with less frequent pasture burning.

PAGE 71

71 Finally, in order to test the im plications of participation in governance for land use, I posit the following three hypotheses: H 4.4 Higher levels of participation in govern ance correspond with lower deforestation. H 4.5 Higher levels of participation in govern ance correspond with smaller cattle herds. H 4.6 Higher levels of participation in govern ance correspond with less frequent pasture burning.

PAGE 72

72 CHAPTER 3 ROADS AND LAND USE CH ANGE IN ALTO ACRE A central issue in discussions about sustainabl e development in Brazil concerns the effects of highway paving on social-eco logical systems in the Amazon Basin. While road paving promises to improve access to goods and services among previously isolated rural and urban communities in Amazonia, it can also have negative environmental and social consequences, including deforestation, widespr ead use of fire, and the displa cement of rural smallholders. Roads in tropical ecosystems have brought nega tive ecological impacts including fragmentation of forest cover, the release of carbon into the atmosphere, increa sed fire frequency and species extinctions (Cumming et al., 2005). In the case of Amazonia, this has been accompanied by negative social impacts such as the increased violence and dispossessi on of traditional land holders (Schmink and Wood, 1992). In academic a nd policy circles, a he ated debate exists regarding the land use changes br ought by roadways and the capac ity of governance to prevent them (e.g., Nepstad et al., 2001, 2002; Laurance, 2001, Soares-Filho et al., 2004). Given that households are often the ultimate agents of land use change in the Brazilian Amazon’s highway corridors, a household-level research appro ach provides an important perspective in understanding the decision-making processes drivi ng land use change in areas such as Alto Acre’s Inter-Oceanic highway corridor. The first part of this chapter addresses th e question of migration—specifically how land use differs among migrants and natives, and am ong households of different age and gender combinations. It also addressed the impacts of bi-localism (the division of residence between rural and urban areas) for land use. The second portion of the chapter ad dresses the impacts that roadway access and expanding herd sizes have for land use.

PAGE 73

73 Descriptive statistics for the i ndependent and dependent variables1 used in the analysis are shown in Table 3-12. Of the sampled households, 38% had at least one head born outside of Acre. On average, households had more male adu lts present (1.95) than female adults (1.38). However, women scored slightly lower on the meas ure of bi-localism than did men. On average, members of households included in the study had to walk 1.5 miles to reach the paved highway and had been on their properties for nearly 20 ye ars. Property size varied greatly—from 40 to 1200 hectares, with the average property having 254 hectares. The ma jority (59%) reported using credit as of the time of the interview. Brazil nut production was highl y variable between propert ies, though most households produced modest amounts. The mean Brazil nut harvest for 2004 was 26.5 latas The average household reported having cleared 24.24 hectares of land during their tenure on the property, though this amount varied from 0 to 298 hectares. Of the land that had been cleared, households reported an average of 3.35 hectar es dedicated to the production of annuals as of the time of the interview. Like property size and Brazil nut pr oduction, cattle herd size was also variable— ranging from 0 to 800 head. The average househol d owned 58 head of cattle. On average, households had spent 3.19 years since the last time they had last in tentionally burned their cattle pasture. Roads and the “Occupation of the Amazon” In the early 1960’s Amazonia’s l ong-standing isolation from th e rest of Brazil came to an end. While previous administrations—most notably that of President Gertulio Vargas several decades earlier—spoke of the importance of inte grating Amaznia into the Brazilian mainstream, 1 The variable “cattle” is used as a dependent variable when testing the impacts of migration for land use. It is used as an independent variable when testing the relationship between expanding herd size and other land uses. Potential issues of endogeneity in the use of cattle as an independent variable are discussed in that section. 2 Descriptive statistics for variables occurring in both chapters 3 and 4 are displayed in Tables 3.1 and 3.2.

PAGE 74

74 by and large this did not result in concrete action until the 1960’s. In the early part of the decade, when the region’s first major highway was cons tructed linking Belm with the new national capital of Brasilia, Amaznia’s hi storical isolation began to ch ange rapidly. During the 1960’s, and even more so in the 1970’s and 1980’s, the federal government undertook an enormous effort to encourage the col onization of the Amazon through pr opaganda, road construction, and massive fiscal incentives (Mora n, 1981; Schmink and Wood, 1992). A number of explanations have been offered for this relati vely sudden interest in the colonization of the Amazon Basin on the part of the federal government. Subsequent to the military coup in 1964, the government developed a xenophobic fear of the “internationalization of the Amazon.” Informed by scholars such as A.F. Reis (Reis, 1960), government officials feared that a sparsely populated and poorly integrated Amaznia was vulnerable to threats from neighboring and northern nations in terested in the region’s vast resources. This fear was exacerbated by reports such as a 1971 United Nations Food and Agriculture Organization document suggesting that, if intensively fa rmed, the Amazon Basin could feed a global population of 36 billion (Pawley, 1971, Smith, 1982: 13). Furthermore, the government’s view of the Amazon as a “safety valve” for overpopulated and impoverished regions of the nation was a nother important incentiv e behind programs to construct roads and coloni ze the area (Bunker, 1985; Hall, 2000). In the Northeast, a series of droughts led to waves of out-migration to other parts of the country, including Amazonia. Exacerbating this problem, in both the Northeast and the South of Brazil, concentration of land ownership and the growth of agro-industry was forcing smallholders from their land (Smith, 1982). While hindsight shows that implementation of agrarian reform would likely have been a much more socially just and co st effective strategy to deal with these problems (Ozrio de

PAGE 75

75 Almeida, 1992), doing so would have run contrary to the right-wing dictatorship’s philosophical stance and, more importantly, by angering we althy land owners, would have undercut a fundamental component of the di ctatorship’s political base. Combined with the desire to avoid the internationalization of the Amazon and to relieve political pressure for agrarian reform was a de sire on the behalf of the government to exploit what was perceived as a vast untapped reserve of natural resources incl uding timber, minerals, and agricultural land (Bunker, 1985; Ozrio de Al meyda, 1992). The development of this region and its resources was intended to absorb inve stment capital and attract surplus labor from overpopulated and impoverished areas of Brazil such as th e Northeast (Schmink and Wood, 1992). Highway construction, especially in sparsely-pop ulated areas, often le d to negative social and environmental consequences including land speculation, rising land prices, population turnover and deforestation (Schmink a nd Wood, 1992; Hecht and Cockburn, 1990 ; Perz, 2001, Nepstad, 2001, 2002; Laurance, 2001). Road cons truction and colonization generally went handin-hand as large-scale colonizat ion was only feasible given over land connectivity to regional and national markets and population centers. I ndeed, the creation of the Amazon’s major colonization poles (e.g., Trans-Am aznia and Rondnia) was inextri cably linked to the roadways built through them (Browder and Godfrey, 1997). Road Paving and Colonization in Alto Acre In 1969, Acre’s isolation from the rest of Braz il began to break, with the construction of the unpaved BR-364 highway linking the capital c ity of Rio Branco with Rondnia and the remainder of the country (Bakx, 1986; Sawyer, 1 984). The following year, Governor Wanderlei Dantas, in a bid to stimulate outsi de investment in the state, offe red a now infamous invitation to an audience of So Paulo ranche rs and investors to “Produce in Acre, Invest in Acre, Export via

PAGE 76

76 the Pacific.” Many accepted the invitation and, by 1971, a full-scale land rush was well underway (Bakx, 1986). Initially, the roadway led to greater autonom y for rubber tappers, who consequently had freer access to markets for selling rubber and purchasi ng other goods (Cardoso, 2002). However, with the construction of the road a nd the arrival of Southerners came intense land conflict. Rubber barons, many of whom had larg ely abandoned their esta tes in prior decades, returned to reap the profits from soaring la nd prices, which had rise n up to 2000% in highway corridors (Bakx, 1986). State policies were enacted and land exchanged hands with little, if any, recognition of the rubber tappers who continued to occupy the land. Indeed, many ill-informed ranchers and investors were surprised and dism ayed to find their land already occupied (Bakx, 1986). As rubber tappers rarely possessed legal t itle to their land, they were commonly evicted, at times with threats of violence. Many cro ssed the border to Bolivia or relocated to the burgeoning periphery of Rio Branco ; others remained to mount a resistance movement to protect their rights to the land (Bakx, 1986; Cardoso 2002). By 1973, ranchers had begun resorting to violence as a means of removing rubber tappers (Keck, 1995; Schmink 1992) and, as a consequence, with the assist ance of political allies, rubber tappers began to organize into a social a nd political movement. A principal technique throughout the struggle was the use of empates— or non-violent stand-offs—when rubber tapper land was threatened with cleari ng by ranchers (Keck, 1995; Barbos a, 2000; Kainer et al., 2003). As in other parts of the Brazilian Amazon, grass-roots resistance movements emerged to challenge the expropriation of land and destruct ion of rubber tapper livelihoods. However, the movement in Acre was undoubtedly among the most dynamic, widely publicized, and ultimately successful of the grassroots resistance movement s of the Brazilian Amazon in this era. As it

PAGE 77

77 gained in political power and publicity, the rubber tapper movement expanded into increasingly broader political spheres (Keck, 1995)—from house hold-level conflicts to regional, state and ultimately the international environmental movement (Schmink and Wood, 1992; Keck, 1995; Kainer et al., 2003). In the early 1970’s the Alto Acre region became one of the first areas in the state to be affected by the arrival of ranchers and specula tors from Brazil’s South (Bakx, 1988). It also witnessed some of the most intense conflict betw een the newcomers and resident rubber tappers defending their rights to the land they had o ccupied for generations. During the 1970s and 1980s, ranchers obtained and cleared large trac ts of land in Alto Acre near the BR-317 highway—a spur from the BR-364 lead ing from Rio Branco to the Peruvian border. As a result, the Alto Acre region, and my research site in part icular, was a focus of conf lict, especially in the early years of the empate movement. Subseque nt to the assassinati on of STR leader Wilson Pinheiro in 1980 and the retaliator y killing of the ranch manager widely believed responsible, the government saw the need to intervene before social unrest further destabilized the region. These conflicts helped lead to the creation of the colonization area and, several years later in 1988, PAE Santa Quitria—one of the first experiment s in extractive reserves to be implemented in the Brazilian Amazon. Ironically perhaps, some of the mo st out-spoken proponents of the PAE and the preservation of rubber ta pper livelihoods (including some who are widely believed to have been actively involved in the violent sta nd-off with the rancher) have themselves become some of the area’s largest ranchers and vocal critics of environmental regulations in the PAE. While a handful of the older ranches, includi ng Fazenda Sta. Rita, ha ve persisted to the present, the majority were short-lived. In an effort to restore social calm in the region, many ranches in the area were ultimately appropriate d by INCRA and redistributed as 60-100 hectare

PAGE 78

78 lots in Directed Settlement Project Quixad to both displaced rubber tappers and to landless migrants arriving from outside of the state. During the early and mid1980’s, INCRA registered and settled families onto 301 lots in PAD Quixad, Gleba 6 (the portion of the Quixad Project included in my research project), offering land and occasionally building material, and transportation to the regi on. While migrants came from many areas of the country, the majority came from Brazil’s south, especially the largely defo rested former agricultural frontier of Paran. Despite free land, the journey and subsequent fi ght for survival that migrants faced were difficult, with many migrants abandoning their lots and returning to their homelands in Brazil’s South or relocating to the peri pheries of neighboring towns afte r finding the extreme isolation and the difficulties posed by the lack of basi c social services and technical isolation insupportable. This turnover is reflected in the fact that, tho ugh INCRA had discouraged the sale of lots, only a third of interviewed families from PAD Quixad were either the original owners or children of the original owners of the lot. While most of the migrants in the region ini tially settled in the co lonization area at the highway’s edge, as land became scarce in the co lonization area, many subs equently moved into PAE Santa Quitria and, to some degree, to th e RESEX Chico Mendes, bot h of which lie behind the colonization area. While this has at times occurred through land in vasion, the steady stream of out-migrants from the RESEX and PAE willing to sell their colocaes for low prices has provided a means of doing so with out outright social conflict. One family living immediately behind PAD Quixad in neighboring PAE Sta. Quitria told me that they had initially mi grated in the 1980’s from the st ate of Goais, along with the husband’s father, to a farm on the road’s edge in the colonization area. As the then-young couple began to start a family of their own, ra ther than purchasing a lot along the road, they

PAGE 79

79 decided upon a relatively inexpe nsive and completely forested 200 hectare land holding nearby in the PAE. While they would technically need community approval through AMPAESQ—an association representing the PAE’ s residents—they were told by a contact in INCRA that they could proceed with the purchase, but with no gu arantee that they would not subsequently be expelled by the community. Ultimately, the gamble paid off and they successfully received a carto de assentamento (“settlement card”) from INCRA, ent itling them to live in the PAE. While the majority of the residents of the PA E and RESEX were native-born Acreanos, such stories from migrants and their chil dren were nonetheless common. At the time of research, cultural difference s between native-born Acreanos and southernborn migrants were reflected in stereotypes commonly held by each about the other group. On one hand, migrants—especially older Southerner s—were often held to be rigid and moneyoriented, with little concern for th e forest, and at times insular a nd disinterested in participating in community organizations, such as associa tions. On the other hand, while many migrants boasted of their hard work in the hope of building a better future for themselves and their children, they would at times contrast their ow n ambitiousness with the complacency of many Acreanos who were perceived to be content to live on the thres hold of poverty as long as the immediate needs of the present were met. As evidence for such “ preguia ” (“laziness”), migrants would occasionally point out the modest homes, small clearings and lack of wellmaintained pasture of native-born Acreanos. Despite initial conflicts and lingering stereo types between these two groups, relations generally improved over subsequent years and decades. Inter-marriage between native-born Acreanos and migrants became commonplace, and many younger members of migrant families had little or no recollecti on of the South, referring to themselv es not by the place of their parents’

PAGE 80

80 birth but simply as “Acreanos.” This is re flected in a statement made by a daughter of Paranaense migrants in her late 20’s who told me that she took personal offense when people spoke derogatorily about Acreanos—she scarcely remembered Paran and despite her blond hair and blue eyes (commonly associated with southern migrants), she considered herself Acreana. Intermarriage between Acreanos and migrants wa s common; in PAD Quixad, for example, 25% of married couples had one spouse from Acre and the other from elsewhere (in all cases, the husband being a migrant and the wife being Acrean a). Through this process of inter-marriage and the coming of age of Acre-born children of migrants, the cultural distinctiveness of indigenous, northeastern and southe rn ethnic identities had largely melted into a shared Acreano identity, especially among younger individuals. Children of migrants tended to refer to themselves as Acreanos and not by their parents’ state of origin; inter-marriage between migrants and non-migrants was common (if not the norm); and, as shall be demonstrated, stereotypes aside, after controlling for other factors, few di stinctions persisted between these migrants and native-born Acreanos in terms of their land use practices. Testing the Relationship between Place of Origin and Land Use The migrants coming to Acre in the 1980’s di ffered from their Acreano neighbors in terms of livelihood systems and history. Whereas Ac reanos had a long tradition of forest-based livelihoods (e.g., rubber tapping), most migran t livelihoods depended upon mixed subsistence and market-oriented agriculture, including cattle production. Furtherm ore, it was almost entirely native Acreanos who participated in the rubber tapper movement that tied land rights with forest conservation, often in the form of extractive re serves. While many migrants had in fact been involved in struggles for land right s, rarely were they tied to fo rest conservation. Due to these historical differences, I anticipated a continui ng differentiation between the two groups in terms of land use. In order to test hypotheses 3.1, 3.2, 3.3, and 3.4 (listed in Chapter 2) about the

PAGE 81

81 impacts of place of origin and land use, I first conducted bivariate analysis of the relationship between place of origin and deforestation, catt le herd size, Brazil nu t production, and pasture burning. As each test involved a bivariate inde pendent variable (Acreano or migrant) and a continuous dependent variable I relied upon t-tests. Acreanos appeared to have a virtually e quivalent mean level of deforestation (25.3 hectares) to migrants (22.3 hect ares). An independent sample t-test of place of origin and household deforestation (equal variance assumed) failed to reveal a significant difference between native-born Acreanos and migrants (t =.360, df=84, p=.720). In contrast, migrants appeared to have a slightly hi gher mean cattle herd size (64.5) th an migrants (54). As with deforestation, an independent sample t-test of place and origin and herd size (equal variance assumed) failed to reveal a signi ficant difference between nativeborn Acreanos and migrants in terms of their land use (t=-.423, df=89, p=.673). Regarding Brazil nut production, Acreanos appeared to have a much larger mean 2004 harves t (52.3 latas) than migrants (8.9 latas). Unlike deforestation and herd size, an independent sample t-test (equal variance assumed) showed that Brazil nut production was significantly higher for Acreanos than for migrants (t=3.592, df=83, p=.001). Finally, regarding pasture burning, migrants appear to burn their pastures with slightly less frequency (3.47 years) than Acreanos (3.02 years). An inde pendent sample t-test (equal variance assumed) failed to reveal that the di fference was significant (t=-.895, df=78, p=.374). Based upon bivariate analysis, I rejected hypothe ses 3.1, 3.2, and 3.4. Place of origin did not have a significant impact on defore station, cattle herd size or use of fire in pastures. However, supporting hypothesis 3.3, place of origin did appear to be significantly related with Brazil nut production. Households originati ng in Acre produced signifi cantly more Brazil nuts than migrant households.

PAGE 82

82 Modeling Place of Origin and Land Use In order to further test the relationship be tween place of origin a nd Brazil nut production, I subjected it to a multiple regression analysis, cont rolling for years of residence and property size. The model indicated that Brazil nut production was not significantly affected by years of residence It was, however, significantly related to property size. Each additional hectare of property size was associated w ith an additional .08 latas of Brazil nut production. Of key concern to hypothesis 3.3, migrants were predicted to harvest 34.5 fewer latas of Brazil nuts than native-born Acreanos (Table 3-2). I therefor e accepted hypothesis 3.3, that households with place of origin in Acre produce more Brazil nuts than migrant households. Discussion The stronger correlation between property size and Brazil nut production is intuitive, given that larger properties can be expected to have more Brazil nut trees. The f act that, controlling for covariates, Brazil nut harvests remain significan tly related to place of origin lends a further insight into the relationship between place of or igin and land use change. On one hand, native Acreanos seem to have readily adopted non-forest -based land use practices, such as cattle, as they became viable in the region. On the other hand, even though Brazil nut production has recently become a very lucrative activity, non-Acr eanos have been less keen than their Acre-born neighbors to begin exploiting th is forest-based resource. Thus, while the BR-317 has, and still does, aff ect land use in the region (as will be further discussed in the next section), the cultural orientation of the migrants who followed the road to this area appears to play a limited role in this relationship several d ecades later. Cultural differences that once were very important in affecting land use between migrants and native Acreanos seemed to be, with some exceptions, diminishing as a common Acreano identity

PAGE 83

83 formed and all groups responded to shared market forces and governmental policies that have followed in the highway’s wake. Intra-Regional Migratio n, Gender and Land Use While two-thirds (66.3%) of house holds reported that they ha d modified their land use in some way because of the paving of the roadwa y, all but one (who complained of increased crime) indicated that their overall quality of life had improved because of the paving of the road—many referring to it as an abeno (blessing). Travel between their homes and the regional urban centers of Brasili a and Assis Brasil became much more feasible, putting markets, schools, churches and medical services more eas ily within reach. A journey to town that previously would take up to severa l days could be completed in as little as an hour, allowing the majority of residents in the research site to trav el to and from town in one day, even in the rainy season. This travel to and from town was generall y motivated by one or two primary goals: the economic transactions of selling farm produc ts and purchasing food and other products unavailable in the colonia and accessing a myriad of urban services—including banks, hospitals, government offices, secondary schools, brothels and forr dancehalls. Given the trade-offs between rural and urban residence, the paved hi ghway has facilitated temporary and permanent micro-level migration between the two as well as an emerging phenomenon of bi-localism as families split their residence between rural and ur ban areas. Most families made multiple-day trips to town on a monthly or greater basis and wh ile most stayed with family members while in town, 17% of interviewed households main tained a second resi dence in town. It is important to recognize that bi-localism ge nerally operates at the intra-household level. For various reasons, including the cost of transportation, on-farm responsibilities and fears of break-ins, it was usually individuals rather than entire families who traveled between the farm

PAGE 84

84 and city. For example, if a family needed to ma ke a bank transaction or tr eat a sick child at the hospital, one household head woul d usually go, leaving the other to care for the home and farm. Similarly, in families with adolescent children an d sufficient economic means, most would invest in an urban secondary education for one or two children while the others remained to provide onfarm labor. A telling indicator of gender-based differences in rural-urban movement is reflected in the numbers of men and women who had assumed l onger term in-town residency. On average, sampled households had sent .23 men and .37 women to the city in the prior four years. Given the fact that men contributed the largest part of farm labor, the continuation of most on-farm activities required either the presence of males or sufficient financial resources—such as salaries or cattle—to contract male labor. Conversel y, reproductive activities —including the care of children and the sick—were dominated by women and were often more easily carried out in an urban setting. Girls and young women were more likely to pu rsue advanced studies in town than boys and young men. Girls were often perceived as being both more adept at academics and less prone to becoming side-tracked by the myriad temptations of city life than boys, making investment in their education le ss risky than investment in boys’ education. Furthermore, in a growing service-based economy, women often found more opportuni ties for unskilled and semiskilled urban work than did men. And increasingl y, in the field of skille d urban labor (generally synonymous with the governmental sector in Al to Acre), a growing population of educated women were finding in-roads into positions once dominated by men as evidenced, for example, in the 2004 election of Brasilia’s first female prefeita And while unintended pregnancy was a great concern for many parents as they decided whether to send their daughters to study in town

PAGE 85

85 where they would be subject to less direct adult supervisi on, this risk was often accepted, especially given the preponderance of the same issu e in the rural area as well, even with closer parental supervision. Despite the advantages offere d by urban residence, the pos session of a rural land holding remained very desirable for most residents, th e outright sale of rura l land being most common among the elderly who desired to “cash-in” thei r accumulated farm investments and by the poor who did so out of necessity. Many informants boa sted of the peacefulness, quiet, healthfulness, and relative safety from crime and pollution that th ey enjoyed in the countryside. Furthermore, while economic opportunities were generally lim ited to the agricultura l sector, these were considered to be economically lower-risk ac tivities—especially cat tle raising—than urban economic ventures. For wealthier families, pos session of a rural land holding provided at least three services: as a weekend “vacation” home, as a capital investment—especially when it sustained a cattle herd—and as a hedge against economic failure in the city (e.g., lost job or failed business venture). In such an event, possession of a rural land holding allowed the owner to retreat to a relatively secure, if modest, on-farm subsistence. Natural population growth, coupled with tem porary and permanent rural-urban migration, has contributed to rapid growth of the region’s towns. As in many parts of the developing world (Lambin et al., 2001) and in the Amazon in par ticular (Browder and G odfrey, 1997), the direct impact of urbanization upon LUL CC by replacing fields and forest s with urban sprawl has been negligible in Alto Acre. Ur banization, however, by changing th e region’s political, social and economic landscapes, is a powerful indirect cause (and effect) of land use and land cover change in rural areas. For example, wage labor, refr igeration and the proxim ity of meat processing facilities contribute to a much larger per capita be ef consumption in the urban versus rural areas;

PAGE 86

86 hence as towns grow, so too does regional demand and price for beef. Also, the closely related phenomena of bi-localism and rura l exodus that are directly tied with the urbanization of the region also have impacts upon land use in the country side. In particular, one would expect that the part-time and permanent relocati on of rural residents to the city would result in shifts from labor intensive land uses to activ ities that require less labor. Testing the Relationship between Bi-l ocalism, Out-Migrati on and Land Use Most agricultural and extractivis t activities conducted in the fi eld site tended to be very labor intensive. This was especially true for the production of annual crops and for the extraction of Brazil nuts. In c ontrast, cattle ranching, while requi ring a larger cap ital investment (e.g., livestock, fencing, corrals vaccinations), was a much less labor intensive activity, especially once initial start-up labor investments such as the construction of fences and corrals had been completed. Hence, I posited that as rural household la bor supply diminishes—through both out-right migration to cities and through bi-localism, th at households would reallocate the remaining household labor from labor-intensive activities such as the production of annua ls and Brazil nuts to less labor intensive activities like cattle ranching In order to test hypotheses 3.5 through 3.16 (listed in Chapter 2) about the impacts of house hold labor and bi-localis m for land use, I first conducted bivariate analysis of the relationshi p between male bi-localism and deforestation, cattle herd size, Brazil nu t production, and pasture burning. I re peated the tests using female bilocalism as the independent variable. I then c onducted bivariate analysis of the impacts of household labor force for these same dependent vari ables. In the case of male and female bilocalism and land use, as each test involved a bivariate independent variable (high bi-localism and low bi-localism) and a continuous dependent variable, I relied upon t-te sts. In the case of

PAGE 87

87 household labor force and land use, as both th e independent and depe ndent variables were continuous, I used Pearson co rrelation analysis. Female bi-localism and land use Households with high levels of female bi-l ocalism had virtually the same mean area in hectares in annuals (3.2) as did households with low levels of bi-l ocalism (3.3). An independent sample t-test (equal variance assumed) did not reveal a significant difference between the two groups in terms of their cattle herd size (t=.093, df=76, p=.926). Households with high levels of female bi-localism had lower mean Brazil nut harvests (18 latas) than those with low levels of female bi-localism (49 latas). An independent sa mple t-test (equal variance assumed) revealed a significant difference between the two groups in terms of their Brazil nut production (t=2.274, df=75, p=.026). Households with high levels of female bi-localism had smaller mean cattle herds (37.69) than households with low levels of female bi-localism ( 75.26). An independent sample t-test (equal variance assumed) revealed a significant difference between the two groups in terms of their cattle herd size (t=1.345, df=77, p=.045). Hence, bivariate analysis failed to suppor t hypothesis 3.5 that households with higher female bi-localism would be associated with le ss area in annuals. It also failed to support hypothesis 3.7 that higher female bi-localism woul d be associated with larger cattle herds. However, it did support hypothesis 3.6 that higher levels of fe male bi-localism would be associated with smaller Brazil nut harvests. Male bi-localism and land use Households with high levels of male bi-localism had the same mean area in hectares in annuals (3.4) as did households with low levels of male bi-localism. An independent sample ttest (equal variance assumed) di d not reveal a significant difference between the two groups in terms of their cattle herd size (t=.015, df=77, p=.988) Households with high levels of male bi-

PAGE 88

88 localism appeared to have lower mean Brazil nut ha rvests (24.8 latas) than those with low levels of male bi-localism (46.5 latas). However, an independent sample t-test (equal variance assumed) did not reveal a signifi cant difference between the two groups in terms of their Brazil nut production (t=1.613, df=75, p=.111). Households with high levels of male bi-localism had smaller mean cattle herds (54.5) th an households with low levels of male bi-localism (65.9). An independent sample t-test (equal variance assu med) did not reveal a significant difference between the two groups in terms of thei r cattle herd size (t=.419, df=78, p=.676) Hence, bivariate analysis fa iled to support hypothesis 3.8 that high male bi-localism would be associated with smaller area in annuals. It also failed to support hypoth esis 3.9 that high male bi-localism would be associated with smaller Brazil nut harvests. Neither did bivariate analysis support hypothesis 3.10 that high male bi-localism would be associated with larger cattle herds. Adult females and land use Pearson correlations failed to show a signifi cant relation between adult female adults and area in annuals and Brazil nut produc tion (Table 3-4). Hence, biva riate analysis failed to support hypotheses 3.11 and 3.12 that lower female labor woul d be associated with less area dedicated to annuals and lower Brazil nut production. A signi ficant relationship did emerge between lower levels of adult female labor and smaller cattle he rds. However, the direction of the relationship countered hypothesis 3.13 that lower levels of adult female labor w ould be associated with larger cattle herds. Adult males and land use Pearson correlations showed a positive relatio nship between number of adult males and Brazil nut production and cattle herd size. No relationship emerged in the relationship between male labor and area dedicated to annuals. Henc e, bivariate analysis failed to support hypothesis 3.14 that lower levels of male la bor would correspond with a smalle r area dedicated to annuals.

PAGE 89

89 It did however support hypothesis 3.15 that lower levels of male labor would correspond with less Brazil nut production. Finally, while adult male labor was si gnificantly associated with cattle herd size, the direction of the relationshi p countered hypothesis 3.16 that lower levels of male labor would correspond with larger herd sizes. Modeling Bi-localism, Household Composition and Land Use I tested the first relations hip between female bi-local ism and Brazil nut production by using multiple regression to further test these relationships while controlling for other variables likely to affect land use (Table 3-5, column 1). Property size was significa ntly associated with Brazil nut harvest. For every additional hectare of property si ze, Brazil nut production increased by .084 latas. Of special concern to hypothesis 3.6, after controll ing for years of residence and property size, Brazil nut production was significantly associated with female bi-localism. Female bi-localism explained 3.3% of the varia tion in Brazil nut productio n. Households with a high level of female bi-localism produced 22.57 fewer latas of Brazil nuts in 2004 than did households with a low level of female bi-localism. I then replicated the test to determine the re lationship between female bi-localism and herd size (Table 3-5, column 2). In this model, the only significant pred ictor of herd size was years of residence. For every additional year of reside nce, herd size increased by 3.392 head. After controlling for years of residenc e and property size, th e relationship between female bi-localism and cattle herd size was no longer significant. To test the relationship between adult male s and Brazil nut producti on and between male adult males and cattle herd size, I ran multiple regr ession analyses to control for the influence of years of residence and property size in both rela tionships that proved si gnificant in Pearson correlation analysis. In the first model (Table 3-6, column 1), property size had a significant impact on Brazil nut production, with each additi onal hectare corresponding with an increase of

PAGE 90

90 0.104 in Brazil nut production. After controlling fo r years of residence and property size, the relationship between adult male labor and Brazil nut production was no longer significant. In the second model (Table 3-6, column 2), years of re sidence was the only significant predictor of cattle herd size. For every additional year of re sidence, herd size is expected to increase by 2.821 head. In the third model (Table 3-7), as in the second model, years of residence had a significant impact on herd size. For every additional year of residence, herd size is predicted to increase by 2.847 head. Of special concern to the question of household composition and land use, even after controlling for years of residen ce and property size, the nu mber of adult females remained significantly associated with larger ca ttle herds. For every additional adult female present in the household, cat tle herd size was predicted to increase by 34.327 head. Discussion Given the labor intensive nature of Brazil nut production and usual involvement of both sexes in its production, the relatio nship between female bi-local ism with lower production of Brazil nuts may be due to labor shortfalls. Lower availability of female labor may cause families to shift away from labor intens ive activities such as Brazil nut production. It is also possible that the relationship is spurious ; for instance, greater possession of capital and assets in town (something not directly measured in the study ) might necessitate that women travel more frequently to town while reducing the necessity of the family to focus on labor intensive farm activities for subsistence. The relationship between lower levels of adult female labor and smaller cattle herds directly contradicts hypothesis 3.13 th at lower levels of female, woul d be related with larger herd sizes. The finding initially seems counter-intuitive, especially given that the care (though not the ownership) of cattle, in this region, is an activit y almost entirely dominated by men. In this case, it seems unlikely that labor shortfalls are the pr imary explanation for the relationship. However,

PAGE 91

91 it is possible that variance in cattle production systems plays a ro le. Producers generally practice some combination of two general herd manage ment schemes—selling off calves or keeping them—especially salient in the case of female cal ves. One strategy is to sell off most or all offspring—practiced both due to lack of additio nal pasture and desire for short-term income— representing a subsistence-oriented herd manageme nt strategy in which both the size of the herd and annual revenues from calf sa les remain roughly constant. The other system, of keeping heifers with reproductive potenti al allows, under certain conditi ons (e.g., unconstrained forage availability), for exponential growth in herd size and, by extension, accumulated wealth in exchange for lower short-term income. Furthermore, under such a system, the sale of the majority of male calves not intended for reprodu ction can serve as a “dividend” payment while having a minimal impact on the long-ter m reproductive poten tial of the herd. While I did not measure the phenomenon directl y, in the course of interviews, subtle conflicts emerged between men and women in te rms of their approach to cattle production. Specifically, women seemed most likely to show caution in the management of their herds, being careful to build herd size as a resource for fam ily emergencies; men were often more prone to “cash-in” portions of their herds to purchase vehi cles and other consumer goods. For instance, one elderly widower told me that his late wife had always preferred, when possible, to keep calves rather than selling them. Upon her death, he told me, the herd had diminished as he gradually sold it off. In another instance, wh en I asked a female informant how the family (which seemed to be of modest means) had a fforded to purchase a car, she replied that her husband “ Perdeu a cabea ” (“lost his head”) and sold off a la rge part of the family’s herd to buy the vehicle. This conclusion was corroborated by feedback from numerous community members (though more often by women than by men) when I returned research results in 2006. Such

PAGE 92

92 anecdotal evidence, combined with statistical anal ysis, suggests that house holds with more adult females (and female decision-makers) present will tend to have the most rapidly expanding herds. Hence, the disproportional tendency of women to migrate, permanently or temporarily, may be having the effect of slowing the expansion of cattle herds in the region’s rural areas. Access to Roadway and Land Use Two important dates stand out regarding th e construction of the BR-317 roadway between Brasilia and Assis Brasil. First was the cons truction of a dirt road in the mid-1970’s which launched an initial wave of in-migration, land us e change and deforesta tion. Another milestone occurred in 2002, when the long-awaited paving of this highway became a reality, and it became possible to travel from Brazil’ s heartland through Alto Acre, to the Peruvian border without leaving asphalt. While this ha d been a perennial campaign promis e from numerous politicians, it was ultimately realized by Governor Jorge Viana, thus ingratiating him to the majority of Alto Acre residents, including ma ny who were otherwise opposed to his Worker’s Party. With the scheduled paving of the Peruvian ro adway stretching from Brazil to the Pacific coast, much speculation has been offered about the likely economic and environmental implications of this future asphalt link betw een Brazil’s heartland, the Brazilian and Peruvian Amazon, the populous Peruvian Andes, the Paci fic and, ultimately, Chin a and other rapidly growing Asian economies. This prospect has rais ed fears of devastati ng social and ecological changes, most notably an increased marketabil ity of soy from the Amazon’s southern fringes (and resulting deforestat ion) in that region. While the implications for Acre and the western Brazilian Amazon will undoubtedly be considerable, the extent of larger-ranging ramifica tions in Brazil is doubtful. In the case of soy, due to the distance involved and the formidable barrier posed by the Peruvian Andes—even on a paved road—it is unlikely that exporters will pref er this route over shorte r existing routes from

PAGE 93

93 Central Brasil through Bolivia and Chile or ch eaper water routes th rough the Amazon and Atlantic Ocean. During a 2005 discussion with Un iversity of Florida students and faculty, Governor Viana himself downplayed the wider-scal e national and internati onal consequences of the highway—stressing instead that the major implic ations of paving to the Peruvian border and beyond have occurred and will continue to occur at the regional level of Acre, the Bolivian state of Pando and the Peruvian state of Madre de Dios. As of 2005, Bolivia—especially the state of Pando and its ca pital Cobija—had become an important part of the social and economic lands cape for many residents of Alto Acre, including my research site. For example, most of the Brazi l nuts in the field site were sold to a Bolivian firm, and clothes, electronics and gasoline us ed by interviewed households commonly came from Cobija; many informants also spoke of family members who had gone to Bolivia searching for jobs or land. With the completion of the paved highway between the field site, Brasilia, and the neighboring city of Cobija, th e city’s free trade area had b ecome an important source for relatively inexpensive Chinese-made consumer g oods not only for Acre’s economic elite, but for all but the poorest of the field site’s rural reside nts. Additionally, for those with cattle or other liquid assets, Cobija’s private ho spital provided health services that, while expensive, were universally considered superior to Brasilia’s free public hospital. Furthermore, the availability of inex pensive nationally-subs idized gasoline— approximately one half of Brazilian prices—offered a considerable reduction in operating costs to tractor and chainsaw operators and especially to those with automobiles, including the owners of trucks running transportation routes between th e region’s rural and urba n areas. In effect, as in other Brazilian regions borde ring Bolivia, the economic development of Alto Acre has been able to free-ride upon Bolivia’s politi cally-popular gasoline subsidies.

PAGE 94

94 Additionally, the combination of improved transportation link s between Brazil and Bolivia and the growth of governance capacity in the former relative to the latter led Bolivia to be seen as a new frontier for many Brazilians. Brazilians who had been displaced by the closing frontier and rising land prices on the Brazilian side co uld, in exchange for compromised economic and physical security, find opportuni ties unavailable in Brazil. Brazilian ranchers and loggers frustrated with increased regulations in their home country had begun to view Bolivia as a lowcost and lightly regulated ar ena in which to expand opera tions. Likewise, many landless Brazilians had found that Bolivia offered an oppor tunity to find wage labor and land. However, this came at a risk as no legal recourse exis ted for Brazilians in unsafe labor conditions and, without legal tenure, any investments in land were subject to confiscation. It is possible, though still unclear, that the recent el ection of Evo Morales of the M ovement to Socialism Party as Bolivia’s president will change the generally lawless atmosphere that had prevailed in this region of Bolivia previously. As of the time of research, Peru had been less important than Bolivia in terms of its social and economic impacts upon the people of Alto Acre. However, with the completion of a permanent bridge between the two countries in late 2005, this could quickly change. Despite considerable population in the neighboring Madre de Dios province of Peru, little of the state’s area had been converted to pa sture, making the state a likel y purchaser of Acre beef upon completion of the bridge and import clearance. Similarly, should the paving of an asphalt link between Acre and Peru’s Pacific coast be achieve d, the implications for Acre will likely be considerable. While it is doubtful that the Braz ilian soy industry, concentrated in the nation’s center, will find the Inter-Ocean ic Highway to be a profitable export route, the road may, however, provide an incentive fo r soy production in Acre itself Likewise, rising demand might

PAGE 95

95 also occur for products from Acre’s forestry a nd cattle sectors with increased connectivity to East Asia’s expanding markets. As of the mid-2000’s, highway paving had le d to increased market integration and increased marketability for relatively high valu e and/or easily transportable farm products including cattle, coffee, bananas and some NTFPs such as Brazil nuts. However, due to their low value and the difficulty of tr ansportation, most annual crops su ch as rice, cassava and beans had been little affected by highway paving, as most of production was dedicated to household subsistence. While most families reported se lling surplus crop production, many claimed that this tended to consist of relatively small quan tities—as the majority of crops produced was dedicated to subsistence—even with improved road access. Various farmers told me that the paved highway had, in fact, hurt th e profitability of some crops such as rice and beans as cheaper produce from mechanized farms in Brazil’s So uth began to flood the region’s markets. Despite the recent paving of the road, the tide of migrants to the state of Acre has slowed, as less unoccupied land is now available for new co lonization projects. Most of the region’s land is currently occupied by extractive reserves, colo nization areas, indigeno us reserves or titled private property. In the occasional event of IN CRA redistribution of unoccupied or expropriated land, given the large number of landless families re lative to available land, new lots tend to be much smaller than in the past—tending to range from 5-30 hectares. However, the newly-paved road does provide improved access to markets for families already living in the area. Roads often serve as drivers of land use change by fac ilitating access to markets and promoting marketoriented agricultural activities that might not be viable in more distant locations (Chomitz and Gray, 1996). However, the benefits of road pa ving are not necessarily equal for all of the residents living near it. Due to the poor state of many of the region’s secondary roads, some

PAGE 96

96 households remain fairly isolated from regiona l population centers and markets, especially during the rainy season. Testing the Relationship betw een Road Access and Land Use One means of measuring the impacts of th e highway upon land use is by looking at the relationship between a househol d’s access to the roadway and its land use decisions, while controlling for other variables. Various technique s can be used to measure access to a roadway. Possibilities include measuri ng the linear distance between the home and the roadway and measuring the distance in terms of time required to travel from the home to the roadway. Due to the differences between paths and roads in terms of their quality and due to the fact that on-theground travel distances are often l onger than linear distances, I c hose the latter method. As area residents tend to rely on various forms of tr ansportation at different times—including foot, bicycle, horses/mules/oxen, motorcycle and truc k, I standardized the time-distance measurement by obtaining walking time-distance from each household. Bivariate analysis (Table 3-8) revealed that distance from the road showed no relationship with deforestation (r=-.073; p=.506) a nd with herd size (r=-.104; p= .325). Even after controlling for other variables—namely years of residence, years of cattle ownership, legal deforestation limit and propert y size—no relationships emerge d that could be clearly differentiated from chance. Hence, bivariate analysis failed to suppor t hypotheses 3.17 and 3.18 that increased distances from the road would co rrespond with lower deforestation and lower herd sizes. Discussion In order to avoid overstating th e lack of relationship between distance from the road and deforestation and cattle, it is impor tant to note that all sampled households lie within five hours walking distance of a paved road. It is he nce impossible, based upon the data, to draw a

PAGE 97

97 conclusion regarding the relations hip between distance to a road way and deforestation for the numerous families living even further from the roadway and outside the scope of this study. A possible explanation for the lack of a relationship between distance to the road and deforestation and with cattle herd size may be due to the fact that all househ olds could feasibly travel to and from urban market centers in one day if possi ble. This conclusion is supported by a study conducted in Columbia (Feaster, 1970) that found that distance from markets only significantly affected the land use among households where it wa s not possible to and from the market in a single day. Due to the difficulties imposed by sp ending a night away from home, such families enjoyed much less access to markets than those who could visit and return in the same day. As seen in Figure 1-2, in the Alto Acre region as a whol e, deforestation is most heavily concentrated in areas nearest the BR-317 while fore st cover tends to be more intact in more remote areas. This conforms to various studies such as Moran (1982) who found that deforestation had tended to follow road corridor s, especially in areas that have undergone colonization. While households living neares t the roadway certainly enjoyed easier market access for farm goods, including cattle, the distances at the scale of this study did not appear to be prohibitive in the production of cattle. Give n the relative abundance of trails and secondary (unpaved) roads in my research site and the fact that cattle, unlike most farm products, are able to transport themselves even during the rainy se ason, most if not all families could reasonably expect to transport cattle from farm to the edge of the roadway (which provided all-season access for cattle trucks) within a day or less. Though not directly evaluated in th e study, it is likely that the opportunity costs of cattle ranching for familie s living several days or more into the forest would manifest in lower herd sizes and less area de dicated to pasture.

PAGE 98

98 Cattle and Land Use In the Brazilian Amazon, cattle ranching has gained notoriety for replacing ecologically important forest with low-biodivers ity grassland. Many researcher s point to cattle ranching as a primary cause behind deforestation and social conflict in the Brazilian Amazon (Hecht, 1985; Cardoso, 2002). Others, such as Faminow (1998) see a more complex relationship between cattle and environmental degrad ation, with cattle production, in and of itself, being less important than the specific types of production used. It was clear, through both intervie ws and direct observation over the course of a year, that many households’ herds were rapidly increasing. While I did not systematically measure the change, on return visits to households both during fieldwork and when returning research results one year later, most households mentioned that their herds had increased since the time of the initial interview. Given the te ndency toward herd expansion that was occurring in the region, especially given the facilitated market access that the paved road provided, I tested the relationship between herd size and other land us e practices, particular ly the production of annuals, Brazil nut harvesting, and the use of fi re as a pasture maintenance mechanism. Testing the Relationship between Herd Size and Other Land Use Practices The strong correlation between herd size and defo restation seen in Ta ble 3-8 is intuitive— larger herds require larger areas of pasture to survive. Howeve r, the relationship between cattle herd size and other land us e practices is less obvious. In order to bette r understand the relationship between the size of a household’s cattle herd and other land uses, I tested the relationship between herd size and the production of annuals and Brazil nuts3. I also tested the 3 The relationship between cattle and Brazil nuts and annu als should be interpreted with caution. Based on qualitative observations of land use and of herd dynamics, it is arguable that cattle herd size functions as a driver of changes in other land uses such as annuals and Brazil nuts. For example, I found numerous cases in which Brazil nut production had been decimated as herds, and pasture expanded. I found no cases in which herd sizes had expanded as households reallocated pasture land to Brazil nuts. With very rare exception, the same can be said for

PAGE 99

99 relationship between herd size and use of fire as a pasture maintenance tool—a practice that can be detrimental to neighboring forest, crops and to the pasture itself (C ochrane, 1998, Nepstad et al.,, 2001, Almeida, n.d.). As households have a finite supply of la nd and labor, I expected that cattle production would result in an inevitable trade-off with ot her land uses and that la rger herd sizes would correspond with less production of annuals and Brazil nuts. I also expected that, due to their lower relative dependence on crops and Brazil nu ts, owners of large cattle herds would see wildfire as a minimal risk to their livelihoods consequently burning their pastures more frequently than families with smaller herds. In order to test hypotheses 3.19, 3.20 and 3.21, about the impacts of herd size for annual production, Brazil nut production and use of fire, I first conducte d bivariate analysis of the relationship between herd size and the three dependent variab les. As each test involved continuous independent and dependent variables, in each case, I used Pear son correlations (Table 3-9). Bivariate analysis revealed a significant relations hip between cattle herd size and area dedicated to annuals (r=.286, p=.007). However, th e direction of the rela tionship contradicted that which was posited in hypothesis 3.19 that larger herds would correspond with less area in annuals. This suggested that suggesting that, rather than presenting competing land uses, cattle ranching and annual production could be compleme ntary. Bivariate analysis did not reveal a significant relationship between ca ttle herd size and Brazil nut pr oduction. Hence, the analysis failed to support hypothesis 3.20 that larger he rds would be related with less Brazil nut production. Finally, bivariate analysis reveal ed a significant positive relationship between size annuals. However, as both the independent and dependent variables are land uses, the possibility for endogeneity exists.

PAGE 100

100 of cattle herd and the number of years since pa sture was last burned (r=.322, p=.004). As the relationship was positive rather than negative, it ra n contrary to expectations stated in hypothesis 3.21. I had expected that households with larg er herds would be less concerned about the impacts of wildfire upon other land uses and therefore use fire more freely than those with smaller herds. Instead, I found that households w ith larger herds had passed more years since the last time they had used fi re in their pastures. Modeling Herd Size and Brazil Nut Production Area in Annuals, and Use of Fire in Pastures In multiple regression models, I further tested the relationships between cattle herd size and area in annuals and between ca ttle herd size and years since usi ng fire in pasture. In both cases, I controlled for year s of residence and property size (Table 3-10). After controlling for the effect of years of residence and property size, a small, though significant, relationship did emer ge between herd size and area in annuals. According to the model, herd size explains 5% of the variation in area dedicate d to annuals. For every positive increment in the number of cattle owned, area in annuals increases by .007 hectares. As previously mentioned, due to the possibi lity of endogeneity between the two land use variables, it is difficult to ascertain the direc tion of causality between herd size and area in annuals. However, the model does, in the l east, suggest complementarity between the two variables. Countering hypothesis 3.19 that larger herds would be associated with smaller area dedicated to annuals, from this model we can conclude that cattle pr oduction does not, generally, represent a competing land use with annual pr oduction, and may possibly be a complementary activity. This conclusion conforms with inte rview data suggesting that production of annuals subsidizes cattle production, and vice versa. On one hand, due to the low market value of most annual crops, it is difficult for a household to re coup the large monetary and labor investments

PAGE 101

101 necessary to clear a given area of mature forest in the two to three year period that the field remains productive. On the other hand, the genera lly lucrative nature of cattle compensates the costs of deforestation—b ut only in the long run4. Hence, according to both the model and qualitative interview data, annuals and cattle may subsidize each other by compensating the costs of deforestation in the near and long-term respectively. Furt hermore, households with more cattle have both the financial means as well as the necessity to clear mo re forestland—land that under the predominant land use change trajectory, typically passes from forest to annuals to cattle pasture. I further tested the relationship between herd size and years since la st pasture burn using multiple regression, controlling for years of resi dence, legal limit, perennial production and use of credit (Table 3-10, column 2). Even after co ntrolling for these vari ables, the relationship between herd size and years since last burn remain s clear. In the model, herd size explains 10.1% of the variation in years since last past ure burn. For every add itional head of cattle owned, the length of time since the last burn increases by .006 years. In other words, for every head of cattle owned, a household is predicted to have spent an additi onal 2.2 days since burning pasture. While the incremental increase is modest given the wide variati on in herd sizes and the capacity of herds to expand rapidly under ideal ma nagement conditions, herd size arguably plays a significant role in ar ea households’ decisions to use fire both at present and in the future. Discussion Qualitative data and personal observation corrobo rated this relationship between herd size and fire use. Lower use of fire may be less a sign of concern that wildfire might damage crops or 4 The economic costs of deforestation include the household or hired labor as well as chainsaw fuel required to clear the area. In cases of illegal deforesta tion, further costs include any fines levied if the event is discovered by an environmental protection agency. Indirect costs (less likely to directly figure in a household’s decision to clear) may also include the opportunity costs of lost Brazil nut producti on and habitat for wild game species as well as reduced water availability if a stream passes through the area.

PAGE 102

102 forest than an expression of concern for the pa sture itself. Many individuals with large cattle herds told me that they had initially burned th eir pastures more freque ntly. However, through the course of experimentation, they learned th at forage quality and quantity and, by extension, animal performance, improved when pasture was bur ned less frequently. At least two additional explanations may exist for the lower use of fire among larger -scale producers. Larger-scale producers, by having less dramatic fluctuations in their herd sizes run less risk of weed invasion (and hence less need for burning) th an smaller-scale cattle producers5. Additionally, as largerscale cattle producers by definition have more res ources for capital investments such as fencing, they are more likely to use rotational stocking systems, a technique shown to improve forage quality and reduce the need for fire as a past ure management strategy (Rueda et al., 2003). Conclusions A number of tentative conc lusions about the impacts of highways on land use—and policies to control these impact s—can be drawn from the study. While the ini tial wave of migrants who followed the highway to Acre ma y have brought new land use practices to the region, among the people interviewed in this study, few differences remain today that allow one to distinguish native Acreanos and migrants in terms of their land use. On one hand, native-born Acreanos have operated according to an economic logic, adopting the practice of cattle ranching as economic conditions have favored it. The same, however, cannot be said for migrants in their adoption (or lack thereof) of Brazil nut collectio n as an economic strategy, even given favorable market conditions. Statistical analysis coupled with qualitative data fr om interviews suggests 5 For example, in the event of a medical emergency, a family may have to sell ten or more cattle, depending on the costs of treatment. For a smallholder, this would severely deplete, if not eliminate, the entire herd, leaving the empty pasture susceptible to weed invasion until a herd is re-e stablished. For a family with larger herds, such a sale would have minimal impact on overall stocking density and risk of weed invasion.

PAGE 103

103 that many of these migrants have tended to shy away from forest extractivist activities even given clear economic incentives to do otherwise. In the case of urban-rural migration and bi-l ocalism, it appears that the exodus of females from the countryside may have especially prof ound implications for rural land use. Due to limitations in the data, it is important not to overstate the explanations offered for the relationships between female bi-localism a nd land use. However, whatever the exact mechanisms by which it occurs, the models do show a very clear relationship between temporary and permanent out-migration of women and decreas ed on-farm production. Hence, it is possible that, at a regional level, the female rural exodus may be causing a shift from a rural and primary goods-based regional economy to an urban service economy. The expansion of cattle production that ha s followed the opening and paving of the highway appears among those interviewed to be wo rking in synergy with another key proximate driver of deforestation—the clear ing of forest for annuals. Hen ce, strategies aimed at reducing deforestation may need to address both land uses and, if possible, counter the synergies that exist between them. For instance, once a swidden field drops in productivity afte r two to three years, a farmer can make one of three decisions. S/he can allow the land to return to forest, thus forfeiting future returns on the initial investment in forest clearing (an option that seemed to be diminishing in popularity in the region). Altern atively, the farmer can continue to receive financial returns on the initial investment by seed ing the abandoned field with grass for pasture. A third option is to plant legumes such as tropical kudzu ( Pueraria phaseoloides) which, by fixing nitrogen, can extend the pr oductive life of the field for ma ny years, if not indefinitely (Valentim and Andrade, n.d.; Pereira, 2001; Shelton et al., 20056). In this case, agricultural 6 In regions such as Alto Acre where cattle are nearly ubiq uitous, the potential for invasiveness is minimal due to the fact that cattle prefer the legume for forage—c ompletely eliminating it when allowed to do so.

PAGE 104

104 extension and economic incentives may be able to promote the use of legumes such as tropical kudzu and therefore extend the life of swidden pl ots and, ultimately, slow the land use trajectory from forest to annuals to pasture. A further conclusion pertains to the use of fire in pastures. As cattle herds expand, evidence from this study suggests that it is like ly that farmers who had burned frequently will gradually diminish their use of fire as experience and word of mouth persuade them that better forage quality and animal performance can be at tained by burning less. Should this occur, while overall deforestation may or may not be affect ed, incomes from both cattle and Brazil nuts should increase and, due to lower mortality amon g remnant trees and the occasional recruitment of fire-intolerant tree seedlings,7 cattle pastures should retain a la rger presence of trees than the nearly tree-less pastures that dominate much of the region today. Reduction in pasture burning would also lower the number of ignition sources for the wildfires that have caused extensive damage to the region’s forests, especially in dry years such as 2005. A combination of increased extension activity and strengthened governance re garding the use and control of fire might accelerate this process, resulting in increas ed farm incomes, more productive and more biodiverse pastures and fewer w ildfires in surrounding forest. 7 Research has indicated the positive eff ects of partial shading upon both forage quality and animal performance in Amazonian pastures (Andrade et al, 200 2, 2004). Many informants in the field agreed and had stopped actively removing desirable shade and timber tr ee seedlings from their pasture. Howeve r, the long-term su rvival of these seedlings is only likely when fire is reduced or eliminated as a pasture maintenance technique.

PAGE 105

105 Table 3-1. Descriptive Statistics fo r Independent Variables in Chapter 3 Test Variables Min MaxMean Standard DeviationSkewness Valid n (n=92) Place of Origin 0 10.380.490.5 92 Adult Males 0 51.951.1290.87 84 Adult Females 0 51.380.861.13 84 Female Bi-Localism 0 10.360.490.56 79 Male Bi-Localism 0 10.410.50.36 80 Distance to Road 0 51.511.030.66 92 Control Variables Years of Residence 1 6619.913.551.16 91 Property Size 50 1200254.13234.641.89 91 Credit 0 10.590.49-0.39 86 Table 3-2. Descriptive Statistics fo r Dependent Variables in Chapter 3 Test Variables Min MaxMean Standard DeviationSkewness Valid n (n=92) Brazil Nut Production (latas) 0 35036.4857.312.86 85 Cattle (head) 0 80057.98113.054.49 91 Years Since Last Burn 1 73.192.160.65 80 Annuals 0 153.353.041.70 88 Deforestation 0 29824.2436.965.27 86 Table 3-3. Brazil Nut Production Regressed on Pl ace of Origin (OLS Regression Coefficients) Dependent Variable Brazil Nuts Independent Variable 1 Constant 22.568 Yrs of residence -0.097 Property Size .08** Place of Origin Acre (ref) Other state -34.522** R square 0.229 Notes: +p<.1, *p<.05, **p<.01

PAGE 106

106 Table 3-4. Correlations among Adult Males, Ad ult Females, Area in Annuals, Brazil nut Production and Cattle Herd Size Variable Adult Females Adult Males Annuals (ha) Brazil Nuts (latas) Cattle (head) Adult Females 1 Adult Males .513** 1 Annuals (ha) -0.098 0.0411 Brazil Nuts (latas) 0.122 .228*.199+1 Cattle (head) .316** .246*.286**-0.0641 Notes: +p<.1, *p<.05, **p<.01 Table 3-5. Herd Size and Brazil Nut Producti on Regressed on Female Bi-Localism (OLS Regression Coefficients) Dependent Variable Brazil Nuts (latas)Cattle (head) Independent Variable 12 Constant 22.769-1.3 Years of residence 0.0713.392** Hectares 0.084**0.007 Female Bi-Localism Low (ref) High -22.857+-21.302 R square 0.1940.173 Notes: +p<.1, *p<.05, **p<.01 Table 3-6. Brazil Nut Production and Cattle Herd Size Regressed on Adult Male Labor (OLS Regression Coefficients) Dependent Variable Brazil Nuts (latas)Cattle (head) Independent Variable 12 Constant 3.063-31.119 Yrs of residence 0.0242.821** Hectares 0.104**0.025 Adult Males 4.11915.019 R square 0.1820.174 Notes: +<.1, *p<.05, **p<.01

PAGE 107

107 Table 3-7. Cattle Herd Size Regressed on Adult Female Labor (OLS Regression Coefficients) Dependent Variable Cattle (head) Independent Variable 1 Constant -48.35 Yrs of residence 2.847** Hectares 0.016 Adult Females 34.327* R square 0.216 Notes: +<.1, *p<.05, **p<.01 Table 3-8. Correlations among Distance to Ro ad, Deforestation and Cattle Herd Size Variable Distance to Road (hours) DeforestationCattle Distance to Road 1 Deforestation (ha) -0.0731 Cattle (head) -0.1040.903**1 Notes: +p<.1, *p<.05, **p<.01 Table 3-9. Correlations among Cattle and Braz il Nuts, Area in Annuals and Years Since Burning Pasture Variable Cattle (head) Brazil Nuts (latas) Annuals (ha) Years Since Burning Pasture Cattle (head) 1 Brazil Nuts (latas) -0.0641 Annuals (ha) 0.286**.199*1 Years Since Burning Pasture .322**0.136.233*1 Notes: +<.1, *p<.05, **p<.01

PAGE 108

108 Table 3-10. Area in Annuals and Frequency of Pasture Burning Regressed on Herd Size (OLS Regression Coefficients) Dependent Variable Area in Annuals (ha) Years Since Burning Pasture (Years) Independent Variable 12 Constant 2.6742.921 Years of Residence 0.032-0.015 Hectares (ha) -0.0010.001 Cattle (head) 0.007*.007** R square 0.0990.114 Notes: +<.1, *p<.05, **p<.01

PAGE 109

109 CHAPTER 4 GOVERNANCE AND LAND USE CHANGE IN THE INTER-OCEANIC HIGHWAY CORRIDOR In the decades from the initial constructi on of the BR-317 in the 1970’s until the present, public perceptions and governmental policies to ward the Amazon have changed dramatically. The Brazilian government once saw the Amazon as an under-inhabited and vulnerable frontier in need of agricultural development and colonization; the forest was perceived as an obstacle to development rather than as a resource to be carefully managed. Furthermore, global policies toward the Amazon were virtually non-existent. However, from the 1980’s until the present, the region has become the focus of intense conserva tion efforts from civil society, the Brazilian government, environmental NGOs and in ternational development agencies. Recent conservation-oriented policies in the Am azon are largely rooted in two historical phenomena in the 1970’s and 1980’s. One was th e rise of grass-roots social movements responding to social conflict as a result of government moderni zation policies in the Amazon. The other was the growing awareness, both in Brazil and internati onally, of the Amazon’s ecological importance (Schwart zman, 1989; Schmink and Wood, 1992) As a response to external threats to th eir land and livelihoods, during this time period, indigenous and rubber tapper populations of the Brazilian Amazon began to form coalitions to fight against invasions of their land and the disruption of their livelihoods. While these efforts initia lly involved political organizing and coalition-building at the grassroots level, cases such as the Kayap tribe of Par (Schmink and Wood, 1992; Barbosa, 2000) and the rubber tappers of Acre eventually gained international publicity an d political alliances (Keck, 1995). Th e demands of both the social and environmental movements for the recognition of traditional land rights and for forest-based development have led to the creat ion and institutionalization of va rious systems of environmental

PAGE 110

110 governance—ranging from the creation of grass-root s associations and rural workers’ syndicates to the demarcation of extractive reserves and indigenous reserves. A major question regarding the impacts of su ch experiments in environmental governance in the Brazilian Amazon is their ability to control the negative LULCC dynamics commonly associated with road construction and paving. That is, can governance prevent or at least minimize the negative environmental and social impacts that have traditionally accompanied road construction in the Amazon? The term governan ce, as used in my research project, closely follows the definition proposed by Soares-Filho et al., (2004: 746): “ac tions by the State and civil society that protect public interests in natural resources and their utilization, including regulation, law enforcement, fiscal incentives, and the organization of informal networks of rural producers.” In the dissertation, governance is underst ood to occur at multiple scales. In the case of my research site, governance is represente d by the state and other governmental actors operating at various scales from international collaboration between Braz il and other countries like Germany in programs targeting the entire Amazon to actions by municipal governments designed to protect key local watersheds. In the specific case of Alto Acre, “administ rative units” (such as extractive reserves and settlement projects ) represent a bundle of such governmental actions and policies. For example, the RESEX and PAD are subject to differing en vironmental regulations and are overseen by different governmental agencies. Additionally, a second form of governance considered, which I term “household participation in governance” is less directly a ffected by the administrative unit and reflects, rather, decisions by households to participate in existing forms of governance implemented by government and civil society a nd to organize with other households into informal local governance organizations, such as a ssociations. This chap ter carefully addresses

PAGE 111

111 both forms of environmental governance—both co mpulsory and elective—in their impacts upon land use change in the region. In the Brazilian Amazon, environmental gove rnance—in the form of numerous public policies and environmental monito ring—has been implemented to control both deforestation and wildfire. Policy measures have included, for example, the creation of the Brazilian Institute of the Environment (IBAMA) in the 1980’s, which is a federal agency dedicated to environmental protection throughout the country, th e creation of specific administra tive units, such as extractive reserves, designed to preserve fo rest cover, the implementation of deforestation limits, and a licensing system for land clearing and burning. More recently, subs equent to the wildfires of 2005 (and the major portion of data collection), a state-wide moratorium on burning was put in place, with a permanent ban on fire in pasture areas appearing likely (personal communication, Francisco Rodrigues Chaves, Regional Director IMAC, Brasilia). At the grassroots level, other more particip atory forms of governance have emerged to control deforestation and wildfire in Alto Acre. The Rural Workers Syndicate (STR) of Brasilia and Epitaciolndia and various en vironmentally-oriented NGOs have played a crucial role in facilitating courses and projects at the community level in the region. Most families in the research site have gained some level of knowledge of the importance of forest preservation and fire control, through contact with such organizations. Furthermor e, many have gained practical knowledge of alternative land use practices that minimize the need for deforestation and use of fire. For example, a large minority of households (25 of 92) had elected to participate in the ProAmbiente program, a program first proposed by the STR, then accepted and funded by the federal government and later implemented by PESACRE, a Rio Branco-based NGO. Through this program, participating house holds receive technical assistance, credit and direct payments in

PAGE 112

112 exchange for environmental services rendered by reforestation and avoided deforestation and burning. Research Questions This chapter poses eight specific questions regarding the relationship between governance and land use in the BR-317 corridor of Alto Acre, f our of which are qualitativ e in nature and four of which present specific hypothe ses tested using both quantitativ e and qualitative data. Taking a longitudinal perspective from th e advent of forest conservation policy in Brazil to the time of research, in order to illuminate the complexiti es of the cross-scale system of environmental governance in the region, the chap ter begins with a qualitative discussion addressing the interrelated questions “what is the st ructure of the governance system affecting land use in the region in the mid-2000’s?” and “what are the relations hips between institutions of environmental governance operating at various scales?” The di scussion of these two qu estions provides the general background to contextua lize the latter questions. In the subsequent section, I discuss the historical relati onship between land tenure and forest conservation; using qualitative analyses I ask “what is the re lationship between land tenure and forest conservation att itudes and practices in the resear ch site?” This discussion gives special attention to the issue of continuities and discontinuities with past processes, using the example of land tenure and forest conservation as an example of the way in which politicaleconomic factors embedded in administrative units have led to apparent contradictions and paradoxes between environmental at titudes and environmental practices For example, given current political and economic realities, familie s with large cattle herds may find themselves supporting the creation of an extractive reserv e as a means of promoting their own economic self-interest. The chapter then returns to th e question of scale in lo cal governance systems,

PAGE 113

113 asking specifically “what are the on-the-ground implications in three distinctly-structured association types present in the research site?” The discussion then turns to a more quantitat ive analysis of the re lationship between two scales of governance and land us e. I pose four specific questi ons regarding the relationship between governance and key land use practices common in the region and with important implications for forest conservation. Speci fically, “what are the relationships between governance and deforestation, cattle production, pasture formation and pasture burning?” Land Rights and Environmental Conservation—the Birth of a Coalition Among the earliest collaborators in the rubbe r tapper movement were the Pastoral Land Commission and CONTAG (National Confederation of Agricultural Workers). Informed by Liberation Theology, the Pastoral Land Commission and local cler gy served as early advocates of rubber tapper land claims and helped to pub licize the movement beyond Acre’s borders. Early in the conflict between rubber tappers a nd ranchers in the mid-1970’s, CONTAG sent a delegate to the Acre rubber tapper movement CONTAG effectively moved the land struggle into the national court system—demanding (often effectively) compensation for expropriated land. CONTAG was also active in organizing rura l workers’ syndicates (commonly referred to by the acronym STRs), the first being in the town of Br asilia in 1975 followed by Xapuri in 1977 (Keck, 1995). The presidents of the Brasilia and Xapur i Rural Workers’ Syndicates—Wilson Pinheiro and Chico Mendes respectiv ely—quickly became outspoken leader s in the Acre rubber tapper movement, helping to lead their communities in organized peaceful protest against the appropriation of their land. The conflicts also at tracted the attention of national and international media and the leader of Brazil’s incipient Worker s’ Party—Luiz Ignacio “Lul a” da Silva (later to

PAGE 114

114 become Brazil’s president in 2002). Ultimatel y, Pinheiro and Mendes would be assassinated by ranchers and become the most renowned ma rtyrs of the rubber tapper movement. Meanwhile, the global environmental movement was growing rapidly, as was international concern over the destruction of Amazonian fore sts (Barbosa, 2000). During the middle and late 1980’s, several individuals with knowledge of the Brazilian Amazon and ties to the global environmental movement, including the anthr opologists Mary Alle gretti and Stephen Schwartzman, saw the possibility of mutual bene fit for the two movement s—the rubber tappers gaining financial and political support as well as publicity, an d the environmental movement being able to put a human face on the fight against Amazonian deforestation (Keck, 1995). Working with the grassroots Conselho Naci onal de Seringueros (Na tional Council of Rubber Tappers or CNS), they lobbied the federal govern ment to give recognition to rubber tappers as “traditional peoples,” thus st rengthening rubber tapper claims to land under Brazilian law. Subsequently, the CNS also collaborated with Am azonian indigenous groups to create a strategic alliance between the “peoples of the forest.” The Advent of Extractive Reserves Environmental activists and l eaders of the rubber tapper move ment devised a plan that would help preserve forests a nd protect seringueiro land rights th rough extractive reserves, a land tenure model designed to combine the two objective s. In 1985, the CNS held its first meeting in Brasilia. Shortly thereafter, in 1987, Schwartzman and Chico Mendes traveled to Washington D.C. to present the model to World Bank and ID B officials, ultimately securing publicity and international support and funding for the project. In the year following this visit, like Wilson Pinheiro, Chico Mendes was assassin ated in Acre. While the assa ssinations of some 982 rural workers, including Pinheiro, in Amazonian la nd conflicts between 1964 and 1988 had largely

PAGE 115

115 gone unnoticed outside of Brazil (Schwartzman, 1992: 54), the life and death of the high-profile Chico Mendes caught the imagin ation and attention of the international environmental movement. During his lifetime and subsequent to his death, a collaborative relationship between Amazonian social movements and the international environmental movement developed and has continued into the 21st century. While the lives of Pinheiro and Mendes were cut short by assassination, their movement lived on and gained force after their deaths. In subsequent decades, the Rural Workers’ Unions they headed in Brasilia and Xapuri respectivel y, as well as the CNS, have continued to be important political forces, lobbying government agencies and NGOs on behalf of rural communities. Their efforts also contributed in the late 1980’s and early 1990’s to the creation of agro-extractive settlement projects (PAEs), extractive reserves (RESEX) and other proconservation policies in Acre and other Amazonian states. With the rise of the rubber tapper movement, the assassination of Chico Mendes and the increased awareness—nationally and internationally—of the importance of the Amazon forest for biodiversity and global climate, development policies in the Amazon began to place greater emphasis upon the conservation of the region’s forests and less upon its conversion to non-forest ag ricultural production. An early sign of this shift appeared through the formulatio n and implementation of the extractive reserve model of integrated conservation and economic development in Acre and other parts of the Amazon in the late 1980’s and early 1990’s. Extractive reserves originated with a 1985 proposal by the National Council of Seringueiros under the leadersh ip of Chico Mendes (Fearns ide, 1989). In 1987, INCRA answered these demands through an internal de cree creating Extractivis t Settlement Projects (Projetos de Assentamento Extr ativistas or PAEs)—representi ng the first federally-recognized

PAGE 116

116 non-indigenous common property system in Brazil (Barbosa, 2000; Cardoso, 2002). Extractive reserves were government-administered land use units that recognized traditional communal land tenure systems and the exploitation of forest resources as “productive usage” while protecting the inhabitants of these areas from disp lacement by outsiders (Schwartzman, 1989, 1992; Allegretti, 1990; Brown and Rosendo, 2000; Mahar, 2000). Extractive reserves were to be “public lands designated for the sp ecific purpose of sustainable us e of forest products such as rubber, Brazil nut and palm heart by the resident population” (Allegretti, 1990: 353). Under this system, no individual deeds were issued, preven ting legal sale outside of the community; and with the exception of small-scale agriculture, only forest-based land uses were permitted (Fearnside, 1989). The following year, a federal-level envir onmental protection agency (IBAMA) was established under President Sarney in order to coordinate, execute and enforce national policy for the environment (Cardoso, 2002: 75). In 1990, newl y created extractive re serves came under the administration of IBAMA, though several pre-exis ting extractive settlem ent projects (including, PAE Santa Quitria) remained under INCRA juri sdiction. By the late 1990’s, three million hectares had been brought under extractive re serve management (Fearnside, 1998). While specific land tenure systems and environmenta l regulations have varied, numerous new extractive reserves have subsequently been created. By 2004, some 67 state and federal extractive reserves existed in the Brazilian Amazon, with ot hers still being created. Legal and Institutional Structures of Environmental Governance in Alto Acre The Alto Acre region serves as a case st udy of the relationship between governance and land use change in the Brazilian Amazon that, fr om both theoretical a nd policy perspectives, merits close attention. The c onfluence of several factors, in cluding a recently paved road,

PAGE 117

117 growing markets, especially for beef, a decades-o ld social movement that grew out of rubber tapper concerns for their livelihoods and the gradual conf luence of this social movement with the political mainstream have led to an envir onment in which rural households must balance complex trade-offs as they make land use decisions. This complex political-economic environment has led many households to view fore st conservation positivel y while continuing to make land use decisions that threaten the very forest that they desire to see protected. It is difficult to imagine a site in which the contrasting forces of infrastructure-related pressure toward deforestation and governance sy stems designed to promote forest conservation come into more stark contrast. On one hand, as described in Chapter Three, from the initial opening of a dirt path through the forest between Rio Branco and Assis Brasil to its present incarnation as a paved international highway, the BR-317 has exacerbated land conflict and deforestation. However, arising from the region’s unique history and relationship with the forest, especially as the birthpl ace of the rubber tapper movement and an early stronghold of Brazil’s Workers’ Party, it has developed an exceptio nally deeply-rooted popul ar and institutional dedication to forest conservation. In many respects, aside from the continuing pr essures toward deforestation, Alto Acre in the mid-2000’s represents a near-b est-case scenario of environmental governan ce. At the policy level, the region is home to various experiment s in environmental conservation, including PAEs and a RESEX and numerous c onservation-oriented projects the most notable being ProAmbiente, a credit and technological assistan ce program creating incentives for sustainable land use practices such as the reforestation of waterways and the reduction of the use of fire (Carvalho et al., 2004). ProAmb iente has been implemented in various pilot areas throughout the Brazilian Amazon, one of whic h is the Alto Acre region in which some 500 families have

PAGE 118

118 been enlisted. As a result of both firstha nd experience and the num erous environmentallyoriented courses and projects offered in rural ar eas of the region, rural re sidents—independent of their actual land use practices—te nd to be both aware of the impor tance of forest conservation and see it as a worthwhile ideal. And, give n the dominance of a si ngle political party (the Workers’ Party or PT) at all levels of gove rnance, the implementation of pro-conservation policies faces far fewer obstacles than in ot her regions of the Brazi lian Amazon with less political hegemony. And while local resistance to governmental policies certainly exists, the dominance of the PT at all governmental leve ls has also facilitated the creation of an environmental subjectivity, whereby many (though not all) rural households have come to accept and espouse governmental envi ronmental policies. Two governance entities deserving special note are the Syndicate of Rural Workers (STR) and the administrative unit. The STR is argua bly the most dynamic entity in Alto Acre’s governance structure. As a grassroots organiza tion, it is directly res ponsible to the rural households that fund it and elect its leaders. As the official representati ve of the region’s rural workers, it actively represents local concerns at the level of the munici pality, state and federal governments, while helping direct governmental and NGO resources toward rural communities. It is also a frequent liaison between rural communities and the government. For example, an individual may request the assist ance of the STR in a petition to be sent to a governmental agency, and a governmental agency or NGO may re quest the STR’s assistance in implementing a course or project in a rural community. As a bundle of environmental a nd agrarian policies, the admi nistrative unit is somewhat more abstract than other more concrete entities, such as the STR, IBAMA and the municipality. However, it plays an important role in shapi ng which governmental agencies are active, which

PAGE 119

119 external resources arrive, and which environmental regulations apply in a given area. For example, RESEX Chico Mendes is administered by IBAMA, receives a large amount of NGO and government-sponsored environmental course s and projects and is subject to a 10% deforestation limit. In contrast, PAD Quixad is administered by INCRA, tends to receive fewer environmentally-oriented courses and projects and is subject to a 20% deforestation limit. Governance at the Federal Level In my research area, the system of land use governance is composed of a network of agencies and organizations operating at macro, me so and micro-scales (see Figure 4.1). At the federal level, several government agencies, including the Ministry of the Environment ( Ministrio do Meio Ambiente or MMA) and the Ministry of Agrarian Development ( Ministerio do Desenvolvimento Agraria or MDA) execute federal environmental and agrarian law, with the sub-agencies of IBAMA (part of the MMA) and INCRA (part of the MDA) being the most visible in the region. In cases considered to be environm ental crimes (e.g., unauthorized deforestation or use of fire ), the Federal Police and the newly-appointed environmental prosecutor of Alto Acre (working within the state-level Ministry of Justice) assist in the enforcement of governmental environmental policies. Among the various federal and federally-linked agencies and organizations operating in the field site, three are most importa nt in terms of land use governance: the Brazilian Institute for the Environment and Renewable Resources (IBAMA), the National Institute for Colonization and Agrarian Reform (INCRA) and the Bank of Amaz onia (BASA). The first two are governmental agencies working primarily with legal administ ration and regularization of land use while the third is a regional banking operation that has be en instrumental in implementing governmental fiscal policies as it regulates financ ial flows into rural enterprises.

PAGE 120

120 IBAMA was created in 1989 by law number 7.735. It was formed through the fusion of four pre-existing environmental agencies: th e Secretary of the E nvironment (SEMA), the Superintendency of Rubber (SUDHEVEA), the Supe rintendency of Fisheries, and the Brazilian Institute for Forest Development (IBDF). Am ong its objectives, IBAMABrasilia is charged with issuing licenses for forest clearing and burning, in reserve areas in the four municipalities of Alto Acre, including PAE Santa Quitria a nd RESEX Chico Mendes. In addition to environmental law enforcement, IBAMA has also increasingly assumed the task of environmental education, for example collabora ting with NGOs in community-based workshops on land management programs. INCRA was established by D ecree 1.110 in 1970—through the combination of two earlier federal agencies—the Brazilian Institute of Ag rarian Reform (IBRA) and the National Institute of Agrarian Development (INDA). INCRA was char ged not only with agrari an reform, but with the colonization of Amazonia (INCRA, 1984). In 1996, INCRA was placed within the newly created Extraordinary Ministry of Land Polic y, later renamed the Ministry of Agrarian Development. In my research site, INCRA has historically played a very large role in demarcating and distributing properties in the co lonization area and in providing le gal recognition to traditional land holdings in the PAE. While the PAE and the PAD continue to be officially administered by INCRA, the agency maintains a fa irly low profile in most govern ance activities in the area now that the major tide of in-migration into the regi on has passed. INCRA co ntinues to administer cartes de assentamento or “settlement documents ”—which certify that the bearer is a recipient of INCRA and entitled to full benefits as such, a nd also function in lieu of title, as the bearers of such cards technically maintain us ufruct rights but not outright owne rship of their land.

PAGE 121

121 Governance at the State Leve l: the “Forest Government” Agencies such as the Secretary of Forests (SEF) and the Institute of the Environment (IMAC) implement and enforce land use policies at the state level. Th e state-level political environment in Acre changed dramatically in 199 8 with the election of forester-turned-politician Jorge Viana and the Worker’s Party / Governo da Floresta (“Forest Government”) in the gubernatorial elections. Campai gning on a platform against government corruption and on the vision of a forest-based economy and society, Vi ana’s Governo da Floresta has helped shape Acre into a pioneer among Brazil’ s Amazonian states in terms of innovative pro-forest policies. For example, in March, 2006, Acre became the s econd state in Brazil to create a special prosecutor for environmental crimes in the st ate legal system (personal communication, Dra. Nelma, Promotora do Alto Acre). The concept of Florestania or “Forest Citizenship” underlies the Governo da Floresta’s philosophy and approa ch to forest-based government. Florestania calls for responsible stewardship of the fo rest, respect for human rights and economic development built upon the sustainable manageme nt of the state’s forest resources. The Institute of the Environmen t (IMAC) functions as a statelevel counterpart to IBAMA, working with land use regulation. As a means of decentralizing land use law enforcement, in Acre, land use regulation on smaller scale propertie s (100 hectares or fewe r) has been delegated to state-level responsibility with IMAC as the executing organ. The agency has been gradually extending its reach into various parts of the state, with offices recently opened in key municipalities, including Brasilia, where it sh ared space with IBAMA before obtaining its own office in late 2005. In addition to its role as a state-level environm ental regulatory agency, IMAC has also played a central role in envir onmental education and in the implementation of

PAGE 122

122 state-wide Ecological-Economic Zoning (ZEE) and Regional Zoning (OTL) in the municipality of Brasilia. The State Secretary of Technical Assist ance and Rural Extension (SEATER) is an important state-level agency involved in bot h technical assistance and the approval of agricultural loans in the region. SEATER focuses especially upon technical assistance and loan approval for households applying to the federal government’s PRONAF credit program. The PRONAF program is divided into four categories—each targeted toward families in specific income ranges and with varying credit limits and repayment conditions. Among its benefits, PRONAF provides subsidized credit for smallscale farmers. Assuming that the credit implementation guidelines—as deve loped between the individual or association and SEATER— are adhered to and the loan is re paid within one year, the farmer receives a 22% discount. So, in the case of a R$1000 (US$450US) loan, R$780 (US $350) is repaid. PRONAF is available both for individuals and for associati ons, which can in turn use the f unding for collective or individual purposes. As collective loans tend to carry more favorable conditions for the borrowers, associations serve an important role in making cr edit available to their members, a point that will be further discussed in a subsequent section. An innovative program functioning at the st ate level but uniting various scales of governance from the international to the local association is th e State Program of EcologicalEconomic Zoning (ZEE). ZEE was created by Stat e Decree 503 of 6 April, 1999. The program was directly linked to the Gove rnor’s office, under the coordina tion of the State Secretary for Planning and Coordination (SEPLAN-AC) (Gove rno do Acre, 2001). In Acre, executing agencies include the Environment Institute of Acre (IMAC) and the Technological Foundation of Acre (itself created in the 1980’s with the assistance of Jorge Viana). The ZEE program

PAGE 123

123 functions through additional assistance fr om other governmental and non-governmental organizations, including the German Internationa l Aid Agency GTZ, which has been involved in similar projects in Par (Sombroek et al., 2000) As of 2006, the implementation process of ZEE in Acre had occurred at a broad state level, designating the comp arative advantages and risks in human settlement and development in various parts of the state, with the intention of creating a basis for more concrete and finer sc ale implementation subsequently. In Acre, the initial la unching of ZEE began in 1999 with th e systematization of mostly preexistent data and the creation of maps at th e 1:1,000,000 level covering the entire state, with finer scale zoning to be conducte d subsequently. The second phase of ZEE, underway at the time of research, focused upon Regional Zoning ( Ordenamento Territorial or OTL), with the municipality of Brasilia as a pilot area. As of 2006, the State Secretary of the Environment (SEMA) and IMAC, in collaboration with the Municipal Secretary of the Environment and funding from the GTZ, had begun participatory meetings among various types of landholders throughout the municipality to determine the goals, obstacles and opportunities they faced in key areas such as agricultural pr oduction, land conflicts, and educ ation. The purpose of these meetings was to gain an appreciation of th e variety of such pers pectives among various communities. For instance, in an OTL mee ting in AMPAESQ in June, 2006, the community mentioned as key priorities the increased comm ercialization of agricultural production, the avoidance of illegal land sales and the preservation of the forest. In a separate meeting held with large-scale ranchers in the municipality, they sh owed less concern with en vironmental issues, but still shared many of the same concerns a bout increasing agricultu ral production (through

PAGE 124

124 mechanization of pastures and fields) and divers ification into other ag ricultural products beyond cattle, likely reflecting concerns about the ca ttle price declines in the prior two years1. Municipalities The municipal government, or prefeitura, represents the smallest-scale institution implementing formal governmental policy, tending to work in tandem with the state and federal policies, especially in PT cont rolled municipalities such as Br asilia and Assis Brasil where municipal policy generally works in close synergy with policies at higher (PT controlled) levels of governance. Under the federal system outlined in the 1988 constitution, the federal government creates legal norms which are comple mented by laws and actions at the state and municipal levels. While the majority of Amazonian muni cipalities leave most environmental issues to the state and federal governments, this is much less true in Alto Acre where several municipal governments have taken strong measur es to create and implement municipal and regional-level environmental governa nce (Toni and Kaimonwitz, 2003). In Brasilia, the municipal government has recently created the U nder-Secretary of the Environment, within the Secretar y of Agriculture. While limit ed in terms of financial and human resources (it has only one full-time em ployee), the office has been active both in promoting environmental education in the municipa lity and in acting as a liaison between state and federal environmental agencies and the munici pal government. Among the activities of the Under-Secretary of the Environment have been the municipal-level implementation of Regional Zoning (OTL), the creation of a committee for th e prevention of wildfire; and environmental education (e.g., a local weekly ra dio broadcast addressing pertinen t environmental issues). 1 In 2006, prior interview respondents informed me that prices had fallen from approximately 250 to 190 reais for calves and from 500 to 370 reais for cows in this time period.

PAGE 125

125 CONDIAC In May, 2004, the Consortium of Inter-Municipal Development of Alto Acre and Capixaba (Consrcio de Desenvolvimento Intermunicipal do Alto Rio Acre e Capixaba or CONDIAC ) was created as an initiative between the four Alto Acre municipalities and ne ighboring Capixaba as a means of augmenting the bargaining power of the region at higher (especially federal) governmental levels, especially in terms of obt aining resources benefiti ng the entire region. Prior to the formation of COND IAC, due to their small populati ons, the individual municipalities faced difficulties obtaining more than modest federal funds, sin ce federal policies favored both municipalities with larger populations and thos e operating in consorti a (Jos Meneses Cruz, Director of CONDIAC; pe rsonal communication). CONDIAC has been involved with OTL and al so with the implementation of coordinated planning between the five municipalities. As such, CONDIAC has helped direct funding from the GTZ and the state intended for zoning in Brasilia. It has also been responsible for obtaining federal resources for infrastructure intended to benefit the entire regi on including, for instance, vehicles for municipal governments, aquacult ure projects, and a factory for making forestderived sweets. Such resources serve the entir e region and would have been difficult, if not impossible, for the municipalities to ha ve funded if operating individually. Administrative Unit Additionally, at an intermedia te level, specific land tenur e systems—both official and unofficial—form an important leve l of governance shaping both la nd use practices and the form of governance in a given area. These land te nure systems correspond with what I term administrative units; in the field site, these admi nistrative units are represented by RESEX Chico Mendes, PAE Santa Quitria, PAD Quix ad, Seringal Porto Carlos and the Terra Solta

PAGE 126

126 As will be discussed in further detail below, these administrative units affect both the form and strength of the environmen tal governance system operating within them. Administrative units can best be understood as “bundles” of e nvironmental governance. Administrative units affect both the form of govern ance implemented (e.g., deforestati on limits) and the presence and activities of governmental agencies and NGOs. For instance, NGOs working with a copaiba extraction project will be much more likely to ta rget an extractive reserve than a colonization area, and environmental infractions in a PAE or RESEX will be handled by the federal-level IBAMA while similar infractions in a colonization area will be handled by the state-level IMAC. Non Governmental Organizations (NGOs) Non governmental organizations (NGOs) play an important role both in facilitating government sanctioned projects and programs a nd also in representing rural communities at higher political levels. NGOs, esp ecially those interested in fore st conservation, maintain a very high profile in Rio Branco, Br asilia and in many rural communities. These NGOs include grassroots organizations, such as the Nati onal Council of Rubber Tappers (CNS) and the Federation of Rural Workers (FETACRE) that ar ose during the social movements of the 1970’s and 1980’s. They also include local and intern ational conservation and development-oriented NGOs such as PESACRE, CTA, WW F and Friends of the Earth ( Amigos da Terra ), which, while often working at the grassr oots level, are based in Rio Bran co, Brasilia or other countries. In various parts of my research site, NGOs ha ve been instrumental in facilitating courses and implementing projects related to forest c onservation. In the research site, WWF and CTA have helped implement a copaiba-tapping pr oject in AMPAESQ, and PESACRE has assumed responsibility for implementing the ProAmbiente pr ogram with four associations in the region. In many cases, NGOs are directly contracted by th e government. For example, in the case of the

PAGE 127

127 ProAmbiente program, local NGOs such as PESACRE are contracted to assume responsibility for local level implementation. As another example of the importance of NGOs in the area’s governance system, in 2005, INCRA requested a leadership capacitation course in an association in PAE Sta. Quitria. INCRA contracted the state-leve l Secretary of Forestry (SEF) which in turn contracted an NGO—the National Council of Rubber Tappers (CNS) —which ultimately facilitated the course. This is one of various examples in which an NGO has functioned as a link between large federal bureaucracies and local associations STR Like the environmental NGOs, the Syndicate of Rural Workers (STR) of BrasilaEpitaciolandia plays an important role as an intermediary between local communities and government at all levels and is perhaps the only organization operating at vi rtually all levels of the governance framework outlined in Figure 4. 1. That is, STR Brasilia-Epitaciolndia representatives (most commonly th e president) while frequently making appearances at local association meetings, are also re gularly present in governmental o ffices in Brasilia, Rio Branco and even Brasilia. However, unlike most NGOs, the accountability of the STR lies directly in the rural communities, as rural families support th e functions of the STR with their membership dues, and elect the union’s direct ory. Operating at an intermediate level between macro and local governance structures, the Rural Workers Un ion (STR) performs crucial roles. In Alto Acre, the STR is arguably the most important organization in this intermediary role. In addition to serving as a representativ e of local communities a nd associations, the STR also assists outside organizati ons and agencies interact with local communities. NGOs, the federal, state and municipal government common ly count on the assistance of the STR when arranging meetings or implementing projects in rural communities. The STR is a nearly

PAGE 128

128 omnipresent entity in my research site—its responsibilities includi ng the mediation of land disputes between neighbors, assisting families and associations to obtain credit, facilitating courses and projects in remote rural areas, a nd defending the interests of Alto Acre’s rural residents at all levels—from Bras ilia and Assis Brasil to Rio Bran co and the national capital of Brasilia. As in other parts of Acre, the STR of Brasil ia-Epitaciolandia arose with the help of the Pastoral Land Commission (CPT ) of the Catholic Church in the late 1970’s. Guided by Liberation Theology, clergy of the CPT formed Christian Base Communities (Comunidades Eclesiais de Base) among the rural working cla ss as a means to resist land dispossession and promote a philosophy that “land belongs to thos e who work it” (da Silva, 2004:46). The STR was initially created as an organization to repr esent rubber tapper rights in the face of forced evictions. It continued as such through the early 1990’s—as of 1990, of the BrasiliaEpitaciolandia STR’s 4,000 members, only 50 were co lonists from outside of Acre (da Silva, 2004:45). However, as colonists have begun to face many of the same concerns for their livelihoods as have rubber ta ppers—including, for example, transport and markets for agricultural products and the availability of s econdary education in ru ral schools—they have increasingly entered the ranks of the STR since the 1990’s, comi ng to represent approximately 1/6th of the STR Brasiliia-Epitaciolandia membership by 1998 (da Silva, 2004). The STR of Brasilia and Epitaciolandia, under the leadership of former president Chico in the mid and late 1990’s, began to open itself to colonists and rein vented itself as representative of all rural smallholders—regardless of their geographic orig in. Although several mi grant informants felt disfavored by the STR, most migrants, like na tive-born Acreanos, tended to hold positive views of the STR. One migrant informant told me that the current president of STR Brasilia-

PAGE 129

129 Epitaciolandia understood the differences and si milarities between rubber tappers and migrants and consequently was highly capable of repres enting and responding to the needs of both. Throughout its history, the STR has been closel y allied with the environmental movement. While this alliance has helped the STR in mee ting many of its goals, in cluding the creation of extractive reserves and the implementation of the ProAmbiente program, it has also brought some criticism from members who felt that the STR had maintained loyalty to the goals of forest conservation at the expense of the welfare of the rural workers it represented. Despite the perhaps inevitable complaints from some member s, the STR of Brasilia was largely recognized as being more inclusive, better organized and more effective than many other STRs in the region. This is reflected in the fact that, among info rmants within and outside of the STR, 75% expressed partial or complete a pproval of the organization. Associations Finally, community associations, which can ra nge from formal to informal institutions2 represent the smallest scale uni t of land use governance above the household. Associations serve to secure community needs—most often accessi ng agricultural credit, and caring for common property (e.g., rice husking facil ities) and local sec ondary roads—both through community labor and through lobbying municipal officials. Associati ons are also often the recipients of land use oriented meetings, courses and projects orga nized by or in collabo ration with government agencies and NGOs. Associations tend to work most closely with the STR and are often the target scale for environmental NGOs (as opposed to working with entire administrative units or with individual families). The administrative unit in which an association is located affects the types of resources and projects that arri ve from NGOs and the government. 2 The level of formality and institutional permanence of an association depends upon its structure and the scale at which it operates. This will be further discussed in a subsequent section.

PAGE 130

130 In many rural areas of the Brazilian Amazon, including Alto Acre, rural communities have organized into associations to represent community needs in governmental spheres, obtain credit and to create a forum for discussing local issues. In extractivist communities, associations have also served to create local-level governance syst ems to regulate extractiv ist activities, forest clearing and timber sales. Local associations re present channels through which to obtain credit, commercialize agricultural and extractive produc ts, and pressure local government into providing both road infrastructure as well as hea lth and education services. They also serve as arenas for debate and decision-making on local issues of general community concern (Brown and Rosendo, 2000; personal observation). In the field site, associations served as the governance institution operating clos est in scale to the household. In most cases, the association overlaps with tightly-knit family networks, with most associations representing several (often inter-marrying) extended families. Associations play a critical role in assisti ng households to process a nd market agricultural production. Most associations have communal rice husking machines—where members of the association can husk rice at a reduc ed cost prior to either saving for later consumption or selling in town. Many associations serve to secure purchasing contracts fo r farm production for all association members. For instance, AMPAES Q negotiates annual cont racts with Brazil nut purchasers who offer an agreed upon price in ex change for being the sole purchaser of Brazil nuts from association members. Another key purpose of associati ons is the collective pursuit of agricultural credit. Member s of an association can request credit collectively through the association, assuming collective responsibility for repayment while obtaining more favorable interest and repayment conditions.

PAGE 131

131 Land Tenure and Forest Conservation Another question addressed is the relationshi p between land tenure and forest conservation in the research site. That is, “what is th e relationship between land tenure and forest conservation attitudes and practi ces in the research site?” By shaping the land tenure system— particularly property size a nd land tenure security—specific administrative systems shape household livelihood strategies, leading to sometim es paradoxical land uses and attitudes toward forest conservation among the families residing within them. Numerous studies have documented the complex relationship between land tenure and land us e in the Amazon (e.g., Schmink and Wood, 1992; Alston and Liebcap, 1999). On one ha nd, especially in areas of competing land claims, occupants have been report ed to clear forest to demonstrate “productive use” of the land—and hence solidify their te nure claims (e.g., Hecht and Cockburn, 1990; Schmink and Wood, 1992; Kirby et al., 2006). On the other hand, using an opposing (but often effective) logic, from the time of Chico Mendes to the present, extractivists (and increasingly agriculturally-oriented p opulations) have used the need for forest conservation as a means of strengthening their claims to large forest land holdings. In limited instances in the field site, the prio r scenario of forest clearing as a means of establishing land tenure has occurr ed. For example, in a particularly densely populated region of PAE Santa Quitria in which residents are officia lly entitled to lot-sized (80 hectare) properties but (unlike PAD residents) do not have forma lly-demarcated boundaries, a land conflict erupted in the mid-00’s between a resident and his neighbor s. In order to solidify his land claim, the individual cleared and planted is olated swidden plots, in forest that his neighbors argued was theirs. While the clearings were only several hectares each, they effectively cut off the possibility of exploiting the more distant portions of th eir narrow properties. As of the time of

PAGE 132

132 research, the matter had been brought before IN CRA and it was as yet unclear if INCRA would be forced to change policy by formally demarcating lots within the PAE. Aside from this and a handful of similar instances, forest clearing as a means of establishing land tenure app eared uncommon in the region3. In the colonization area, land conflicts were virtually non-exista nt due to the clearly demarcat ed property boundaries. In other areas with colocao property structures, approximate property boundaries (based on rubber trails) were generally kn own and respected. Furthermore, due to the large size of properties, even families with large land clearings tended to clear on their own property only4. Finally, as will be discussed below, due to the political importance of forest conservation at various levels, preserved forest often proves to be an asset, ra ther than a liability, fo r households concerned about tenure security. Given the relative stability of land tenure and la nd claims in most of Alto Acre at the time of research (as opposed to other regions of th e Amazon or the same region in the 1970’s and 1980’s), a different scenario, of fo rest conservation (in discourse and/or actions) as a means of establishing land tenure, appeared most common in the research site. As an example, in the fairly sparsely populated Seringal Porto Carlos, most residents, especially larger landholders, were in favor of the creation of a PDS rather than a colonization settlement. Furthermore, in this seringal, the ProAmbiente program seemed to be well-accepted. Ironically, this was despite the fact that mean per household deforestation rates ha d been fairly high (3 ha/year) and some of the largest ranches in the field site were located in the area. Of the four communities of the research 3 While still rare, land conflicts more commonly involved the ownership and exploitation of Brazil nuts rather than deforestation. 4 However, in light of both growing population density and deforestation, it is unclear how long this situation will persist, especially in administrative units such as th e PAE and RESEX that lack clearly-demarcated property boundaries.

PAGE 133

133 site that were enlisted in ProA mbiente, Porto Carlos stood out in the fact that its ProAmbiente members had on average deforested larger area s of their properties than the ProAmbiente members in other communities [52.7 hectares for Porto Carlos (n=7) versus 27.7 hectares for all sampled ProAmbiente members (n=25)]. This seeming paradox is fairly easily explai ned. Regarding the possibility of the area becoming a PDS, through conversations with both proponents and opponents, it became clear that larger landholders saw this as a means of gaining governmental recognition of the generous colocao property size, and preventing the appropriation of land for new INCRA settlements. Large cattle producers, while reco gnizing that the implementation of a PDS might severely limit future pasture expansion, believed it would also prevent INCRA from divi ding their properties to settle landless strangers in the seringal. Sma ller landholders, on the othe r hand, saw little reason to accept stricter environmental laws implied by a PDS when they faced little risk of losing land and could even benefit th rough its redistribution. In the case of PAE Sta. Quitria, the infl uence of land tenure security on land use was similarly strong. In reserve areas especially the PAD, deforest ation was widely recognized as not only an ecological threat, but as a threat to land tenure securi ty as well. In INCRA areas dedicated to traditional land reform and settleme nt, property holdings usually did not exceed 100 hectares. Varying from this norm, due to the conservation component of PAEs, combined with the historical importance of the extractivist ec onomy, their land holdings were much larger, generally several hundred hectares However, as Santa Quitri a was facing on-going extensive deforestation, the envi ronmental services of the PAE we re becoming less evident and, as repeatedly noted in conversations with community leaders, should the area become perceived as a failed conservation unit, the justification for such large land holdings would become

PAGE 134

134 increasingly untenable politicall y, especially given governmental desires to appease members of the vocal landless movement through land distribu tion. An INCRA official confirmed that the prospect of converting the PAE to a coloni zation area (implying the appropriation of colocaes and distribution of smaller lots to residents and landless newcomer s), while unlikely to occur in the near-future, in light of rampant deforestation in the PA E, had occasionally been proposed in discussions within the agency and remained a po ssibility at some future point should current trends continue. These scenarios show both the strengths and li mitations of land tenure policy, as embodied in administrative units, as a means of influenci ng land use. On one hand, the acceptance of large properties ( colocaes ) in exchange for forest conservation, coupled with the threat of expropriation of deforested land, provided an in centive for forest conservation in both reserve areas and in areas in which rese rves were being considered. However, on the other hand, this incentive alone has not proven sufficient to co unter the lure of cattle ranching and pasture expansion among many reserve reside nts. In response to this c onflict between the individual and collective good, many indivi duals elected to deforest illegall y, finding the allure of wealth (or simply subsistence) from illegal land cleari ng to outweigh any collective danger imposed upon all members of the PAE by their illegal actions. Al so, given the land conflicts that have occurred in more densely-populated areas of the PAE, it is unclear how resilient the colocao property structure commonly associated with reserves will remain under conditions of population growth. Associations and Scale A fourth question regarding governance and land use concerns the relative effectiveness of the three association structures common in th e region: community-level associations and umbrella associations located locally and in regi onal urban centers. Sp ecifically, “what are the

PAGE 135

135 on-the-ground implications in terms of cross-scale trade-offs of these three distinctly-structured association types present in the field site?” While associations themselves rarely assume roles in creating and enforcing desired land use policy, they play a fundamental role in conn ecting larger-scale struct ures of governance with the household. In addition to agricultural credit, this is also the scale at which most projects and courses are offered and implemented. Of the 16 associations included in the study (1 in Seringal Porto Carlos, 5 in the PAE, 3 in the RESEX, and 7 in PAD Quixad), three distinct modalities of association structure were represen ted. Most associations operated directly at the level of the community. However, AMPAESQ and AMOREB served as umbrella associations for the smaller associations in the project/reserve, with an important distincti on being that while the former was geographically located in the comm unity, the latter was located in Brasilia. AMOREB Local governance in the Chico Mendes Reserv e was based upon an umbrella association structure, with local associations forming part of larger associations operating on a municipal level. Hence, for example, the portion of the reserve lying within the municipality of Brasilia was represented by the Associati on of the Residents of the Chico Mendes Reserve, Brasilia (AMOREB). As in other municipal-level associ ations of the RESEX, the main office of AMOREB was located in town. AMOREB is charged with, among other things, both implementing the management plan of the rese rve and representing community interests in Brasilia. On one hand, having an urban base allowe d AMOREB close access to government and NGO resources as well as high visi bility in Brasilia. Furthermor e, the association headquarters serve as a sort of central gather ing place for many of the associat ion’s residents while in town. However, the distance from the actual reserve ha d some negative impacts upon participation in

PAGE 136

136 the organization. In order to assume a leadership role, one had to have or find an urban residence and spend most days away from one’s colocao. The spatial distance between the association’s headquarters and its members may also have cont ributed to a fairly wi despread concern among members interviewed about lack of accountability on the part of the association’s directors. Some members of the associati on told me that he was concer ned about the loyalty of the association’s directory to its base, especially given its clos e collaboration with several governmental agencies. For example, one member of the association to ld me “They (AMOREB) don’t want to let us work (our land). Thos e people in AMOREB, they work directly with IBAMA (in enforcing envir onmental regulations).” AMPAESQ The Association of the Residents of PA E Santa Quitria (AMPAESQ) was founded in 1997 with the assistance of INCRA and CNS as an association to repres ent all of the PAEs residents. Like AMOREB, it served as an um brella association representing the five local associations within the PAE. Yet it differed in that it was physically located within the PAE. However, it nonetheless suffered similar problems with participation and representation. While the association was intended, on paper, to function as an umbrella association representing the local associati ons of the PAE, this was not always the case. Of the sampled households in the PAE, only 42.5% (17 of 40) had household members associated with AMPAESQ, the vast majority of whom lived in cl ose proximity to the association. Furthermore, of the households associated with AMPAESQ, less than half ( 8 of 17) were members of local associations. Hence, due to various factors—including c onflicts with local associations and the obstacle that sheer distance posed to many PA E residents who might otherwise participate— AMPAESQ, in practice, served less as an umbrella association than as a local association with a disproportionate amount of political power relative to its membership. Since most higher levels

PAGE 137

137 of governance viewed and treated AMPAESQ as representative of the entire Project, many courses and projects reached only the fraction of the PAE’s resident s living in close proximity to the association, while completely missing those li ving in more outlying areas who may have had little if any idea of AMPAESQ’s purpose or even existence. Similarly, many decisions representing the en tire Project—such as the creation of the project’s management plan and the allocation of external funds—have been controlled by a small number of the PAE’s residents, primarily those closely affiliated with the association. This led to one individual, an associated of both AMPAES Q and another local association, to complain “I think that the financial resources that they r eceive over there (in AMPAESQ) are being applied in the wrong way, that’s not really looking out for the pe ople of the region”. AMPAESQ and Local Associations AMPAESQ’s relationship with three local associ ations illustrates the limitations of the association’s effectiveness as a governing force for the entire reserve. In the case of Association AgroNorte—located in PAD Quixad—a sizeable minority of the association’s 50+ members resided within PAE Santa Quitria, with a se nse of community that spanned the boundary between PAD and PAE. While there seemed to be little antagonism between the two associations, due to the physical distance betw een them—approximately 25 kilometers from one association to the other—of the five AgroNort e families residing in the PAE included in the study, none were current members of AMPAESQ5. Not surprisingly, in th is distant portion of the PAE, properties were commonly divided and sold without the required consent from AMPAESQ. 5 While some AgroNorte residents of the PAE attended occasional AMPAESQ meetings that they felt to be of special importance, local informants told me that no members of the association were actual members of AMPAESQ.

PAGE 138

138 A very different scenario was presented by association Wilson Pinheiro, which, while lying a mere 15 minute walking distance from AMPA ESQ had an antagonistic relationship with it. The Wilson Pinheiro association was establ ished in the mid-1990’s by an extended migrant family that had purchased land within the PA E. Upon the creation of AMPAESQ a few years later, the members of Wilson Pinheiro refuse d to cede precedence to the newly formed association. The situation was further complicated as many of the (locally-born) AMPAESQ members were suspicious of the migrants living in their midst. As of 2006, while tensions had, at least supe rficially, cooled among individuals in both associations, neither recognized the legitimacy of the other. Members of Wilson Pinheiro complained about the undue power that AMPAES Q and the STR enjoyed, while the leadership of AMPAESQ expressed concern ab out the lack of collaboration fr om the association and that the lax attitude that some Wilson Pinheiro members had toward timber sales represented a violation of the PAE’s environmental regulations and might jeopardize th e collective well-being of the PAE’s residents. When I stated to one of the directors of AMPA ESQ that Wilson Pinheiro seemed very independent from AMPAESQ and ST R, he corrected me by saying that they not only operated independently, but that they were the active enemies of AMPAESQ and STR. The conflict between Wilson Pinheiro and AM PAESQ led to disillusionment among some Wilson Pinheiro members about the effectiveness and fairness of local level governance in the PAE. One informant, finding common cause with Brown and Purcell’s (2005) caution against the “local trap”, complained th at INCRA and other government ag encies were trying to “throw responsibility for the resolution of all these conflicts on us.” Wh ile well-intentioned, he felt that devolution of responsibility for local conflicts to AMPAESQ had opened the door for local and inter-family conflicts to come into play. In c ontrast, he argued, external government agencies,

PAGE 139

139 with little if any personal conn ections in the area, could make more objective decisions that could be better respected by all powers. Yet another different scenario of AMPAESQ’s relationship with local associations was the one it had with Vicente de Melo Association, loca ted in the PAE approximately 9 kilometers east of AMPAESQ. The two associatio ns tended to have very different attitudes toward conservation and the purpose of the PAE. AMPAESQ, in ge neral, was guided by a very pro-conservation philosophy, its members tended to have relatively m odest levels of household-level deforestation (mean 27.2 hectares), it tended to see IBAMA as an ally and it activel y courted collaboration with environmental NGOs including PESACRE, Am igos da Terra, WWF and CNS. Vicente de Melo, on the other hand, was named for its founde r Francisco “Macaxeira” Vicente de Melo, a former rubber tapper and leader of the empate movement who had become the largest rancher and likely the wealthiest individual in the PA E. Many other members of the association had followed Macaxeira’s lead, looking to cattle as a means of escaping pove rty while remaining on the land. Reflecting this fact, mean deforest ation levels stood at 71.2 hectares among sampled households belonging to the association Interestingly, despite dissimilar philosophies and approaches toward conservation and development, significant overlap and dialogue existed between the two associations. While Macaxeira was renowned in the PAE and Alto Acre in general for being a perennial voice of dissent about the perceived emphasis upon fore st conservation over economic development by AMPAESQ and its leaders, like many Vicente de Melo members, he did recognize the legitimacy of AMPAESQ and as of 2006, was consid ering a campaign to become its president. His complaint, he explained, was not against AM PAESQ per se, but against what he considered

PAGE 140

140 irresponsible management of association financ es and acceptance of environmental policies (e.g., 10% deforestation limits) that he found detrimental to the well-being of the PAEs residents. Representation and Effectiveness in Association Models For their part, local associat ions, on the whole seemed to have more satisfied members than the reserve-level AMOREB and AMPAESQ. This is likely due in part to the fact that most local associations have been constructed on top of existing kin networks and to the elective nature of participation in th e associations (unlike AMPAESQ and AMOREB, the membership in which, on paper, was required of all residents within their jurisdiction). However, such associations, due to their smaller size, often te nded to have less bargaining power and could easily be overlooked by government agencies and NGOs, especially when AMPAESQ or AMOREB were assumed to represent them. Lo cal associations also tended to be less institutionalized and more personality-driven than the umbrella associations. That is, the most successful associations tended to have char ismatic leaders who had proven themselves competent in working with both association memb ers and governmental agencies in Brasilia or Assis Brasil. Hence, it was not uncommon to fi nd a once-active association disappear when a president had left office. Several local associatio ns were defunct at the time of research; in one case, an association quickly fell apart when its popular and effective long -term president died. Fortunately, the reverse was also true, as in the example of an association in PAD Quixad that had previously nearly disbanded under weak lead ership later became one of the largest and most well-known associations in the regi on. Largely due to the influenc e of a particularly effective new association president, the association ha d become a recipient of various courses and projects, including ProAmbiente. In some ways, the reserve-level associat ions like AMOREB and AMPAESQ, which spanned multiple local communities and associati ons, stood at an advantage to local community

PAGE 141

141 associations. Such associations corresponded with governmentally-recognized administrative units (such as the PAE or RESEX) and were thus perceived as the legitimate representatives of the administrative unit in governmental decisions Due to their institu tionalization within the wider structure of governance, their power and pe rsistence has arguably been greater than local associations. However, their larger scale has al so brought some disadvantages to their members. The sheer question of distance often prevented the active participation of all members. Also, due to the fact that participation was often more compulsory than elective in nature and that membership was defined more by geographical loca tion within an administrative unit rather than by kin networks, more opportunities for dissent and for inter-family conflicts tended to arise than in the local community associations. Associations and Scale—Conclusions This discussion of a locally-based umbre lla association, an urban-based umbrella association and several local associ ations illustrates some of the important trade-offs that emerge due to the distinct scales at which each operates. Due to their high visibi lity and regular contact with governmental authorities, umbrella associ ations operating at admi nistrative unit level, especially those located in urban areas, tend to be more effective in attracting external resources to their members and can also more effectivel y lobby government represen tatives on behalf of their constituents. However, due to the large variance in proximity, both social and spatial, between the directors of umbrella associations and their constituents equal distribution of benefits in such associations tends to be more difficult th an in local associations. Differences in the legitimacy of the various a ssociations can be traced to similar questions of scale. Due to their higher visibility, greater institutionalization in the wider political structure, and tendency to have better educated and more po litically adept leaders, umbrella associations tend to be granted greater legitimacy than local associations by government agencies and NGOs.

PAGE 142

142 However, local associations tend to enjoy grea ter legitimacy among their constituents due to dayto-day interaction and due to kin ties between me mbers and between leadership and constituents. Governance and Land Use Given this understanding of both the structur e and the specific po licies of governance in the Alto Acre region, it is possible to operationali ze the concept into a measurable variable for hypothesis testing. I measured tw o distinct forms of governance in this study—one centralized and the other participatory. By measuring these two forms of gove rnance, it was possible to gain an understanding of the impacts of both tradit ional top-down environmental governance of the state and other governmental agencies and other more participatory forms of governance of civil society upon household-leve l land use decisions. The first measure of governance reflects the actions of the state referred to by Soares-Filho et al. (2004) and refers to the governance system associated with the administrative unit in which a family resides. The administrative unit can be thought of as a “bundle” of legally-represented rights and obligations that have been directly or indirectly assembled through governmental action. Each administrative unit is associated with a distinct land tenure system, norms and regulations regarding land use, and access to NGO and government-sponsored environmental courses and projects (Table 41). In the research site, ad ministrative units included three governmentally-sanctioned units: an extractive reserve, an ex tractive settlement project (PAE), and a colonization area (PAD), as well as two unofficial units: the Terra Solta and former Seringal Porto Carlos. While the PAE and RESEX were initially created with some degree of local input, the administrative unit, at the time of fieldwork, was larg ely a top-down structure created in government offices outlining the land tenure system and land use possibilities available to the households residing within them.

PAGE 143

143 The second form of governance, which I have termed “participation in governance,” refers to the actions by civil society referred to by Soares-Filho et al. (2004) and is, by and large, participatory and elective in nature. This form of governance, in the research region, represents the multiple forms through which households may el ect to participate in, challenge, and shape the wider structures of environmental governan ce. In my research project, I measured participation in governance through twelve household-level variable s (Tables 4-2 and 4-3) which were then used to create a f actor-weighted index of particip ation in governance (described below). The average household had one male adult who was a member of the STR; the number was slightly smaller for women. The majority of households (60%) expressed satisfaction with the STR. 84% of households had at least one member who was a member of an association, though less than half of all households had a memb er who had assumed a leadership role in an association. In most households (60%), at least one member atte nded association meetings on at least a monthly basis. Satisfication with the lo cal association was somewhat lower (47%) than for the STR. Nearly half of all households had at least one member who had taken an environmentally-oriented course, though the average number of men and women who had participated in an environmentally-oriented proj ect was much smaller. A considerable minority of households (27%) were involved in the ProAmb iente program. Nearly two-thirds were using credit. In order to operationalize governance at th e administrative unit leve l into a measurable variable, each administrative unit was assigned a governance score based on characteristics that I determined, through the course of fieldwork, to characterize the environmental governance

PAGE 144

144 system in a given administrative unit. A four point index of governance was created by asking the following questions: Has the administrative unit been targeted by any environmentally-oriented courses or projects—sponsored by either the gover nment and/or an NGO? (Yes=1, No=0) Is a specific governmental agency charged with the administrati on of the area? (Yes, IBAMA=2; Yes, INCRA=1; No=0)6 Is the area subject to a 10 % or 20% legal deforestat ion limit? (10%=1, 20%=0) Using this system, the environmental governance scores for the five administrative units were as follows: RESEX Chico Mendes=4 (targeted by courses and projects, admini stered by IBAMA and subject to a 10% deforestation limit). PAE Santa Quitria=3 (targeted by courses a nd projects, administered by INCRA and subject to a 10% deforestation limit). PAD Quixad=2 (targeted by courses and projec ts, administered by INCRA and subject to a 20% deforestation limit) Seringal Porto Carlos=1 (tar geted by courses and projects no official governmental administration, 20% deforestation limit) Terra Solta=0 (not targeted by courses and projects, no official governmental administration and 20% deforestation limit). Testing the Relationship between Govern ance (Administrative Unit) and Land Use In order to test hypotheses 4.1, 4.2, and 4.3 about the impacts of governance at the level of the administrative unit on deforestation, cattle herd size, and frequency of pasture burning, I first conducted means analyses and F-test s to determine the direction, li nearity and significance of the relationships. Means tests of the impacts of governance and land use showed linear relationships in the case of deforestation (Table 4-4) cattl e (Table 4-5) and fire (Table 4-6)7. Bivariate analyses of 6 While INCRA is charged with administration of issues like the inheritance of land and the illegal division of properties, unlike IBAMA, it is not charged with addressing explicitly environmental issues.

PAGE 145

145 the relationship between governance and deforestation, while not si gnificant, was suggestive that higher levels of governance might be associated with lower levels of deforestation, as was posited in hypothesis 4.1 (F=.960, df=3, p=.416)8. Similarly, bivariat e analyses of the relationship between governance a nd cattle herd size, while not si gnificant, was suggestive that higher levels of governance might be associated with smaller cattle herds, as posited in hypothesis 4.2 (F=1.248, df=3, p=.298). Finally, bivariate analyses of the relationship between governance and years since pasture burning, wh ile not significant, was suggestive of a relationship directly contradic ting hypothesis 4.3 that higher leve ls of governance are associated with more years since burning pasture (f=.156, df=3, p=.926). Modeling the Effect of Governance at the Administrative Unit Level Upon Land Use Due to the possibility of suppression effects from covariates, and gi ven the linearity in each relationship, I conducted multiple regressi on analysis of the effects of governance and deforestation, cattle herd size and years since bu rning pasture, despite the lack of significant relationships in bivariate analys is. In these models, I controll ed for years of residence and property size. Governance and deforestation I tested the first relations hip between governance and hous ehold deforestation using multiple regression to control for the effect of years of residence and property size (Table 4-6, column 1). In the model, years of residence was significantly related with deforestation. For 7 In each analysis, the Terra Solta (gove rnance value=0) violated the otherwis e linear tendency between governance and the three dependent variables. As the sample for the Terra Solta was small (n=4) and may not in fact be representative of the entire population I elected to remove it from the analysis. 8 When substituting “area in hectares in pasture” for “area in hectares deforested”, the relationship with the independent variable “administrative unit gove rnance” becomes somewhat stronger (F=1.264, df=3, p=.293). This suggests that governance is more effective in cont rolling deforestation for cattle ranching as opposed to deforestation in general—which includes other ends including annuals and perennials.

PAGE 146

146 every additional year of residen ce, the household is predicted to have cleared an additional 1.14 hectares of forest. Of special concern to hypot hesis 4.1, after controlling for years of residence and property size, area deforested was significantly related wi th deforestation. Governance accounted for 3.4% of the variation in household deforestation. For every additional increment on the index of governance, deforestation is predicted to decrease by 8.76 hectares. Governance and cattle herd size I tested the first relationship between gove rnance and cattle herd size using multiple regression to control for the effect of years of residence and propert y size (Table 4-6, column 2). In the model, years of residence was significan tly related with cattle herd size. For every additional year of residence, the household is expected to own an additional 2.659 head of cattle. Of special concern to hypothesis 4.2, after cont rolling for years of residence and property size, cattle herd size was significantly related with governance. Go vernance accounted for 4.4% of the variance in cattle herd size. For every additional increment on th e index of governance, cattle herd size is predicted to decrease by 31.155 head. Governance and years since burning pasture In the third model, I tested the relationship between governance and years since burning pasture, controlling for years of residence and property size (Table 4-6, column 3). The model did not show a significant relationship between governance and years since burning pasture. Hence, the model failed to prove hypothesis 4.3 that higher levels of administrative unit governance correspond to less frequent burning. Testing the Relationship Between Part icipation in Governance and Land Use In addition to governance as a product of actions by governmental agencies, I also considered governance as a household’s elective decision to participate in environmental governance, whether instigated by the government or the local community. In order to test

PAGE 147

147 hypotheses 4.4, 4.5 and 4.6 regarding the relationsh ip between participation in governance and land use, I created an index based on the 12 meas ures of participation in governance that were presented in Tables 4-2 and 4-3. The first objective in testing the relationship between participation in governance and land use was to determine to what degree an operationalizable concept of “participation in governance” existed at the household level. Participation in governance is a factor-weighted i ndex that accounts for the various commonly-available forms through which households in the study region may participate in, contes t or shape the wider system of environmental governance9. Principle components analys is generated factor loadings that I used to weight the relative importance of z-scores of these 12 indicators. A second principle component analysis gene rated factor loadings of the se ven variables that had yielded the highest scores (>.450) in the initial principl e component analysis (Tab le 4-8). The factor loadings from this second analysis were th en summed to create a new continuous variable representing “participation in governance” I then used bivariate analysis to test the relationships between the new participation in governance independent variable and deforestat ion, cattle herd size and years since burning pasture that were posited in hypotheses 4.4, 4.5 and 4.6. As participation in governance and the three dependent variables are all continuous, I used Pearson correlation analysis for each relationship (Table 4-9). Biva riate analysis failed to find a significant relationship between participation in governance and the dependent variab les. Hence, bivariate analysis failed to support hypotheses 4.4, 4.5 and 4.610 9 The eigenvalue threshold for including a component was set at 2 (which yielded one index of participation in governance). The factor loading threshold for including a variable in the index was set at .450. 10 As with governance at the level of the administrative unit, I further tested these relationships with multiple regression analysis to determine if years of residence or property size had a suppressing effect on the relationship between participation in governance and the three land use practices. As this did not yield any significant relationships, results are not reported.

PAGE 148

148 Discussion In this chapter, I have explored the comple x nature of environmen tal governance in Alto Acre. While environmental governance in this region is often represented by regulations imposed by governmental authorities, it is also a product of the actions of individuals, communities and other components of civil societ y (e.g., the STR). The history of Alto Acre shows the ways in which a social movement a nd a governmental apparatus, which were once in opposition, have co-evolved to the point that clear boundaries between the two are not always obvious. The close connection between the STR and Brazil’s Workers Party is a case in point. Similarly, many environmental regulations, such as deforestation limits in extractive reserves, which are largely enforced by the government, have grown from the actions of civil society as they contested earlier po licies that they saw as threatening their live lihoods (e.g., governmental incentives to clear land for pasture). A political ecological perspective shows the ways in which households not only are affected by environmental policy, but how they in turn shape the nature of governance as they pursue th eir individual or colle ctive goals. For example, the crucial support that the extractive rese rve model and the ProAmbiente program have enjoyed among Seringal Porto Carlos’ powerful ranching families derives, at least in part, from their perception that the benefits of such po licies—by buttressing th eir land tenure claims—outweighs the costs of stricter environm ental regulation. Finally, concerning the effectiv eness of governance to control land use, the analysis of governance at the level of the administrative unit level and land use is encouraging. The relationship between stronger “ bundles of governance” and reduced deforestation and smaller cattle herds suggests that, as ar gued by Nepstad et al. (2001), governance can be an effective means of moderating the environm ental degradation that has so often occurred in the Amazon’s roadways. At the same time, the lack of a cl ear relationship between pa rticipation in governance

PAGE 149

149 and land use shows certain limitations in the use of governance as a tool to shape land use, especially when the desire exists to transf orm “command and control” forms of governance to more participatory forms that positively shape land use without relying upon coercion. As will be seen in the following chapter, the root of this disjuncture between participation in governance and land use can be found through careful atten tion to the relative co sts and benefits of compliance with environmental regulations.

PAGE 150

150 Table 4-1. Administrative Units in Field Site: Range of Environmental Governance Unit Tenure Type Government Agency Responsible Legal Deforestation Limit Targeted by Courses or Projects? Governance Index Score RESEX Chico Mendes Long-term use rights; collective management plan IBAMA 10% Yes 4 PAE Santa Quitria Long-term use rights; collective management plan INCRA 10% Yes 3 PAD Quixad, Gleba 6 Long-term use rights; transitioning to private with individual titles. INCRA 20% Yes 2 Seringal Porto Carlos Private with squatters rights for residents; transitioning to long-term use rights or private property with titles. INCRA11 20% Yes 1 Terra Solta De facto private, no titles. None 20% No 0 Fazenda Sta. Rita12 Private, with title None 20% No N/A 11 As of the time of research, INCRA had completed a diagnostic study of Seringal Porto Carlos in anticipation of assuming responsibility for the administration of the region as a colonization area and/or Sustainable Development Project (PDS). 12 While Fazenda Santa Rita, like all private property in Amaznia, is subject to a 20% deforestation limit, most portions of the project were already cleared by the mid-1980s, ahead of the implementation of contemporary deforestation limits. Due to its fairly static LULCC traj ectory and lack of permanent households, it was not included in the analysis of gov ernance and land use.

PAGE 151

151 Table 4-2. Descriptive Statisti cs for Bivariate and Continuous Measures of Participation in Governance Test Variables MinMaxMean Standard DeviationSkewness Valid n (n=92) Male STR members 031.030.640.79 86 Female STR members 020.740.49-0.49 85 Association Member (0=no, 1=yes) 010.840.37-1.86 92 Association Cargo (0=no, 1=yes) 010.470.50.11 91 Environmental Course (0=no, 1=yes) 010.480.50.07 87 Male Participation in Project (0=no, 1=yes) 010.270.451.05 85 Female Participation in Projects (0=no, 1=yes) 010.070.253.51 89 Participation in ProAmbiente Program 010.270.451.04 92 Use of Credit (0=no, 1=yes) 010.590.49-0.39 86

PAGE 152

152 Table 4-3. Frequencies for Categor ical Measures of Governance Satisfaction with STR (0=no, 1=somewhat, 2=yes; 999=missing) Frequency Percent Valid 0 19 20.65 1 14 15.22 2 55 59.78 999 4 4.35 Total 92 100 Satisfaction with Local Association (0=no, 1=somewhat, 2=yes; 999=missing) Frequency Percent Valid 1 13 14.13 2 24 26.09 3 43 46.74 Total 80 86.96 Missing 999 12 13.04 Total 92 100 Frequency of Association Participation (0=never, 1=yearly, 2= alternating Months, 3=Monthly or More; 999=missing) Frequency Percent Valid 0 27 29.35 1 2 2.17 2 6 6.52 3 55 59.78 Total 90 97.83 Missing 999 2 2.17 Total 92 100 Table 4-4. Means Test for Ef fects of Governance (Adminis trative Unit) for Household Deforestation Governance Mean N Standard Deviation 1 (Porto Carlos) 37.8333 15 35.49 2 (PAD Quixad) 25.6957 23 19.26 3 (PAE Santa Quitria) 20.3947 38 47.98 4 (RESEX Chico Mendes) 13.1667 6 8.73 Total 24.5427 82 37.76

PAGE 153

153 Table 4-5. Means Test for Eff ects of Governance (Administrative Unit) for Herd Size Governance MeanN Standard Deviation 1 (Porto Carlos) 102.40 15 155.89 2 (PAD Quixad) 69.36 25 76.01 3 (PAE Santa Quitria) 43.85 41 123.9 4 (RESEX Chico Mendes) 20.17 6 14.68 Total 59.64 87 115.3 Table 4-6. Means Test for Eff ects of Governance (Administrative Unit) for Pasture Burning Governance MeanN Standard Deviation 1 (Porto Carlos) 3.42 12 2.47 2 (PAD Quixad) 3.21 24 2.21 3 (PAE Santa Quitria) 3.00 35 2.07 4 (RESEX Chico Mendes) 2.83 6 2.23 Total 3.12 77 2.15 Table 4-7. Governance (Adminis trative Unit) and Land Use (OLS Regression Coefficients) Dependent Variable Deforestation (ha) Cattle (head) Pasture Burning (years) Independent variable 123 Constant 16.36262.8313.541 Yrs of residence 1.140**2.659**0.005 Property Size (ha) 0.0210.0690.002 Governance (Ad. Unit) -8.760+-31.155*-0.398 R square 0.2650.1910.043 +p<.1,*p<.05, **<.01

PAGE 154

154 Table 4-8. Factor Weightings—P articipation in Governance Index Component Matrix(a) Component Male STR Members -Female STR Members -Satisfaction with STR -Association Membership 0.626 Association Satisfaction 0.661 Association Frequency 0.813 Association Cargo 0.468 Environmental Course 0.595 Environmental Project (Male) -Environmental Project (Female) -ProAmbiente 0.5 Credit 0.589 Extraction Method: Principal Component Analysis Initial Eigenvalue=3.005 Table 4-9. Correlations among Participation in Governance, Household Deforestation, Cattle Herd Size and Years since Burning Pasture Variable Participation Household Cattle Herd Years Since In Governance Deforestation (ha) Size (head) Burning Pasture Participation in Governance 1 Household Deforestation 0.0821 Cattle Herd Size 0.021.903**1 Years Since Burning Pasture 0.055.278**.322**1

PAGE 155

155 CHAPTER 5 RECONCILING POLITICAL AN D ECONOMIC REALITIES IN ALTO ACRE’S BR-317 CORRIDOR The Political Ecology of Land Use in Alto Acre A political-ecological perspective is useful in understanding the ways in which households make land use decisions in Alto Acre’s BR-317 hi ghway corridor. Such a perspective situates the household within a conceptual framework that addresses multi-scaled political, economic and ecological factors affecting and affected by th e household. Political ecology focuses on localscale land users, contextualizing them within their immediate social, ecological and economic environment, and then approachi ng these relationships within hi gher levels of decision-making power (Cumming et al., 2005, Schmink and Wo od, 1987, Blaikie and Brookfield, 1987). Households make land use decisions in a dialectical relationship with the political, economic and ecological environment. In this relationship, the household both shapes and is shaped by the wider political-economic context as it makes land use decisions. For instance, many administrative units such as extractive reserv es which now shape the land use possibilities available to residents are themselves largely a pr oduct of the grass-roots social movements of the 1980’s, of which many current residents were part. Analysis of field data reveal ed few lingering impacts of the cultural differences between migrants and native-born Acreanos interviewed in terms of land use, as they responded to similar ecological, social and economic forces. An exce ption, however, was the use of forest products such as Brazil nuts. The migrants who sett led in the region in the 1970’s and 1980’s had a profound impact upon the region by bringing land us e practices which often implied a new linear land use trajectory from forest to crops to past ure, replacing the shifting agricultural system which had previously left the vast majority of th e region’s forest cover in tact, as was evident in Figure 1-2. This in turn resu lted in ecological and economic changes; a region that was once

PAGE 156

156 dominated by forest cover and had a forest-bas ed economy had been transformed into a heavily deforested region in which cattle provide the mainst ay of the rural economy. Despite their initial differences, in most ways, both Acre-born a nd migrant families included in thestudy have responded similarly to these transformed ecologi cal and economic condi tions, with Brazil nut production being the only land use in which a si gnificant difference between the two populations appeared. As of the mid 2000’s, road paving had had grea ter implications for intra-regional migration than for inter-regional migration. That is, the vast majority of the movement of peoples in the region was occurring at a regional scale betw een rural and urban areas whereas migration between the region and other areas of the country while still existent, was both multi-directional and much less common. Intra-regional migration tended to especially affect women in interviewed households, as evid enced by the disproportionate num ber of men present in rural areas and the larger mean number per household of women who had left for town than men in the previous four years. This pattern may have surprising implic ations. A clear correlation was found between greater presence of ad ult females and larger cattle herds. During interviews and informal discussions with informants, a likely ex planation emerged that women tended to have a different approach to cattle produc tion than men. Whereas men tende d to see cattle as a readily available “bank account” for meeting short term e xpenses, women more often treated their herds as long-term capital investments to be used in cases of emergency or to meet long-term goals. Hence, the disproportionate migration of women to the city, by lowering mean household herd sizes, may be slowing the pressure toward defo restation in the countryside. However, it is unclear, from the perspective of a household’s economic well-being, whether the benefits from

PAGE 157

157 women’s urban economic pursuits out-weigh the loss of their la bor and influence in on-farm decision making. As of the time of my research project, it was clear that cattle production was increasing among most families. A promising finding was the tendency of larger-sca le cattle producers to use fire less frequently as a pasture maintenance tool. Hence, while the trend toward cattle herd expansion clearly had negative implications in terms of out-right defore station, the possibility existed that herd expansion was accompanied by reduc tions in wildfire in nearby standing forest. I considered two distinct forms of govern ance in terms of their impacts upon land use: governance in terms of the “bundle” of policies embodied by the admi nistrative unit and in terms of a household’s participation in governance, including, for instance, associations and environmentally-oriented courses and projects. Participation in governance, while not affecting land use, did have a very str ong relationship with attitudes to ward environmental rules among those interviewed. While househol d-level participation in governance did not show significant implications for land use, governance at the le vel of the administrative unit did. This was especially true in the case of deforestation and herd size. Taken generally, despite numerous obstacles su ch as the limited budgets of local offices of IBAMA and INCRA, governance seemed to be having an increasingly positive impact among those interviewed in the Alto Acre region. While many landowners continued to deforest 2-3 hectares per year, large-scale deforestation events like the ones that created Fazenda Sta. Rita in the 1980’s had become rare in the area, especially compared to recent decades. The paving of BR-317 improved the capacity for environmental governance, facilitating the oversight of the area by government agencies and also the entry of new projects like ProAmbiente, which would be difficult to carry-out in conditi ons of more diff icult access.

PAGE 158

158 Largely due to widespread environmental edu cation targeting rural residents, such as classes, projects, radio announcements, and STR meetings, the dangers of deforestation and climate change were widely known among the regi on’s rural residents. A virtual consensus existed that environmental regu lation of some form was necessary and a majority (61%) of informants agreed that existing environmental re gulations should be maintained or strengthened. This was often true even among households that had cleared large areas for pasture. For example, one informant, when asked if he approv ed of deforestation lim its, replied that, despite his desire to clear more forest to expand his agricultural operati ons, regulations on forest clearing were necessary. With ironic humor, he stated th at regulations on forest clearing were important; if there were no legal limits, he would probably clear every last hectare of forest that he owned. However, due to contradictions inherent in the contemporary political -economy of Alto Acre, such attitudes, while facilitating the tasks of politicians and governmental regulators in their efforts to preserve the forest, have not necessar ily translated into forest-conserving land use practices among households. The “Alto Acre Paradox”—Weighing the Costs a nd Benefits of Environmental Compliance The recent history of Alto Acre demonstrat es the multiple scales at which multi-scale political and economic forces a ffect household land use and also the ways in which households and individuals, through their act ions, in turn change the wide r political-economic system in which they live. Attention to the life stories of two influential individuals living in the research site shows the ways in which they have been ab le to successfully adapt to the rapidly changing political-economic context in which they live. It also shows the strugg les that more typical families face as they must negotiate the ofte n contradictory political and economic changes through which Alto Acre has passed in recent decades.

PAGE 159

159 As they make land use decisions, rural households in Alto Acre’s Inter-Oceanic highway corridor are influenced by the often contradi ctory forces of markets and a system of environmental governance that many of these sa me households helped create during the era of rubber tapper resistance to land invasions in the 1 970’s and 1980’s. This has led to what I term the “Alto Acre Paradox.” In Alto Acre, despite the large successes of th e social movement to create policies designed to promote forest conser vation, in the wake of the road’s construction and paving, many of the region’s residents, by fo llowing changing market forces, have adopted the cattle based livelihoods that many of them had once risked their lives to fight. On an ideological level, forest conser vation has made great advances, w ith many of the region’s rural and urban residents seeing it as an important priority. However, de spite these pro-forest policies, cattle production has become the backbone of the region’s private sector Hence, paradoxically, even the largest ranchers in the region expre ssed at least some degr ee of concern about the ecological impacts of deforestati on, yet even the most out-spoken advocates of conservation in the region owned at least some cattle. Two individuals, Z da Silva and Maria de Souza,1 both residents of PAE Santa Quitria, represent two examples of ways in which the area’s residents have responded to the changing political, economic and social contexts of Alto Acre. Both are household heads who have chosen very different means to negotiate a path to success for themselves and their families despite, or perhaps because of, the Alto Acre paradox. They are also, arguably, the two most powerful and controversial people living in the PAE, as they repr esent two competing models for the future of the PAE. The experiences of thes e two individuals give in sights into the strengths and limitations of various instituti ons of governance represented in this administra tive unit. 1 While the information about these two individuals is generally known in the region, pseudonyms are used.

PAGE 160

160 Maria, a woman in her late 30’s who, as of 2006 was preparing to begin study at a local university, followed the politics of forest conserva tion which led her from rural housewife with a fourth grade education to AMPAESQ vice-presiden t, local CNS representative, STR officer and, ultimately, to become Brasilia’s Municipal Secr etary of the Environment. Her successes came with some sacrifices in term s of land use on her family’s colocao as well as in terms of pushing local social norms in which women rarely assume such a conspicuously dominant role in the household. Prior to entering politics, she and he r husband sold off a large part of their cattle herd in order to become compliant with the 30 head limit mandated in the PAE’s management plan. They also drained their pond, which wa s built prior to the implementation of clear governmental guidelines, and built another in compliance with IMAC guidelines. In addition to adding to her moral authority as an environm ental advocate and government employee, these compromises may have long-term economic benefits as the surplus cattle were sold during a time of high beef prices and the well-designed pond provided a dependable source of protein in the form of fish, a product that might be commercialized in the near future.2 She and her husband also entered the ProAmbiente program, which w ould serve as another means of both increasing the family’s environmental credentials wh ile possibly increasing household income. In my discussions with her, Maria frequently mentioned the threats of deforestation in the PAE and in Alto Acre. These th reats included loss of water, lo ss of livelihoods and, in the case of Santa Quitria, a threat to the PAE itself and the colocao land tenure system it guaranteed. Furthermore, it was also clear that, due to her decision to build he r career within the governmental hierarchy of Alto Acre, sustai ned deforestation could also jeopardize her 2 As of 2006, given falling cattle prices, increased restricti ons on land clearing and fire use and subsidies from the local government to build ponds (usually in the form of free bulldozer use and man hours, with the landowner only responsible for fuel costs), many people had begun to express interest in aquaculture as a means of generating both cash income and readily-available protein for household consumption.

PAGE 161

161 environmentally-oriented career in the municipal government a nd, should the PAE be dissolved, she would also lose her influen tial behind-the-scenes role in AMPAESQ. However, should the municipality of Brasilia, a pilot project for ZEE, become a show-case of environmental conservation, she might likely attain even higher levels of political influence, especially given her collaboration with representatives of the st ate government and GTZ in the implementation of OTL in Brasilia. Responding very differently to the changing po litical, economic and eco logical realities of Alto Acre, Z was the PAE’s largest cattle ranche r and was among its most influential residents. In his mid-60’s at the time of research, Z had live d a life of seeming contradictions. In the late 1970’s and early 1980’s, he was part of the early STR, actively engaging in the peaceful and violent resistance to the expropr iation of colocaes by in-coming ranchers. He was also among the early proponents of the PAE land tenure mode l, having argued that, in stead of dividing the area into lots as INCRA had pl anned, the area’s residents were entitled to recognition of the colocao land tenure system in which residents could continue thei r traditional livelihood system of rubber tapping. However, beginni ng in the early 1990’s, as a response to falling prices for rubber and the growing market for b eef, he and his family began to shift their resources toward cattle rearing. As he only reluct antly sold female calves, his herd rapidly grew, leading him to become the largest rancher in the PAE and among the larg est ranchers in Alto Acre. Z had clearly been able to benefit from the various changes occurring in the region over recent decades, having taken advantage of social and economic conditions to both secure a large land holding and to become wea lthy through cattle ranching, the fo rmer made possible due to the successful rubber tapper social movement and th e collapse of earlier development models and

PAGE 162

162 the latter arising from the construction and paving of the road and the development of local and regional beef markets, both of which have bolster ed the profitability of cattle ranching in Alto Acre. Z was a perennial source of dissent at AMPAESQ meetings especially when INCRA or IBAMA agents were present. In his disagreeme nt with the environmental policies of the PAE, he drew upon a discourse of economic progress, soci al justice and attachment to place. In sum, his arguments emphasized that cattle ranching was th e only means in the PAE at that time for the generally poor residents to reach a point not necessarily of affluence but of being “ mais ou menos ” (more or less) in terms of their living stan dards (a viewpoint that Maria considered an out-dated discourse from “another era”). Like many other PAE re sidents, he had been born and raised on the colocao on which he continued to live. C onsequently, he resented the legal possibility that he could be evicted from his colocao and the PAE for violation of the management plan. He frequently made mention, however, of the value of forest preservation and his desire to see it preserve d. But doing so, he argued, require d greater assistance from the government, either through the subsidization of forest-based activities or, preferably, through assistance for intensified pasture management. He pointed to his passiv e reforestation of the streams crossing his pastures as proof of this conviction. He further argued that the PAE’s residents, by occupying large land areas, were in effect preserving the forest by preventing its occupation and clearing by outside rs, hence meriting compensation from the government. Like Maria, Z was also c oncerned about the possible di ssolution of the PAE and the colocao land tenure system given current levels of deforestation. However, rather than finding it necessary for the PAE’s residents to take urgent efforts to remain within legal limits, he saw the need to redefine the laws of the PAE to reflect the new economic reality in which its

PAGE 163

163 residents lived. Also, like other large cattle produ cers, he also faced the threat of falling beef prices, especially given his de cision to focus nearly exclus ively upon cattle ranching as an economic activity. However, assuming the recent dip in beef prices is te mporary, the large herd size should provide a buffer for hi m and other large-scale producers to ride out the slowdown. While the specific aspects of their feelings a bout the value of forest conservation differed—in some cases widely—it was clear that both saw this is a worthwhile goal. Both had played a role in the rubber tapper movement and had fought for forest conservation (th ough in very different time periods). As of the time of research, bot h were concerned about the destruction of the PAE’s forest cover, had taken some actions to mitigate their impact upon the environment (including minimized use of fire) and were advocates for governme ntal action to help preserve the forest while supporting local livelihoods. Ho wever, while Maria’s colocao remained well within legal deforestation and cat tle herd size limits, Z continue d to clear land for his growing herd, despite the fines he continually received. In the case of Maria, we see that the envir onmental attitudes arising from her experiences in the region’s social movement have transl ated into clear action. In Z’s case, his environmental attitudes and past experience in the rubber tapper move ment have had little impact upon his land use decisions. His decision not to use fire in his pastures and to reforest riparian areas, while signaling cooperation with the PAE’s management plan, also reflected his economic self-interest given th e scale of his cattle operation3. A fundamental difference, I argue, lies in the different sectors in which the two have sought to build their liv elihoods, social status and power within their communities. 3 Given the large size of his herd, he had capital necessary to invest in fire alte rnatives such as fences (for rotational stocking) and the practical experience to learn of the negative impacts of fire misuse upon forage quality and animal performance. Also, unlike smaller producers, he could maintain a more constant stocking intensity, further reducing weedy growth and the need for fire.

PAGE 164

164 Rising through the ranks of the grassroots soci al movement, Maria ultimately obtained a position in the municipal government. Given the close ties between the Workers’ Party and the rural social movement, the boundaries between the two have blurre d, with STR officers commonly running as PT candidates or assuming a ppointed positions in the PT government. Of further significance, both the rural social move ment (including the STR) and the PT have long histories of pro-forest stances that have continued into the mi d-00’s. While entailing some political and economic costs, such a stance has pr oven effective in attracting external funds and programs to the region. Examples include the Pr oAmbiente program, the selection of Brasilia as a pilot area for regional zoning, and funding for the development of forest-based economic activities like sustainable timber extraction and rubber processing. Z’s life-history stands in contrast to Maria’s. While he gained much of his local prestige as an advocate for the PAE, then for rubber subs idies, he ultimately secured a livelihood outside of the public sector. Rather, he followed market forces developi ng in the road’s presence that have made cattle production in the region lucrat ive. Hence, the livelihood which he built for himself and his family was founded in an economic system that rewards forest conversion and hence is directly contradictory to the philo sophy of the region’s e nvironmental governance system. His economic success through cattle ha s given him both the motivation and the means to openly criticize and violate e nvironmental regulations. This has given him the opportunity to make land use decisions as a form of political resistance to institutions of land use (e.g. 10% deforestation limits) that he felt had been co -opted by governmental authorities and no longer served the needs of the PAE’s residents. In pr ivate discussions and in speeches in association meetings, he emphasized that he was trying to pr omote a vision for development in the PAE in

PAGE 165

165 which local needs took precedence over the concerns of Rio Branco, Brasilia, the United States and Europe. Unfortunately, most people, in the PAE and in other parts of the research site, have been caught uncomfortably between these two extremes a nd have been less able to harness one of the current trends to their advantage. Like Z a nd Maria, most people were reluctant to clear the forest and agreed that forest conservation was a worthy goal. Yet most were also in a trajectory toward increased deforestation, and when they reach ed the limit, would be poorly able to either pay the fine to deforest illegall y or to adopt other land use practic es that did not require further deforestation. Unlike Z and Maria, land use deci sions of most individuals were driven less by calculated political strategy than by sheer necessity. While the growing public sector and the field of forest conservation provi ded an opportunity for some people, like Maria, to improve their livelihoods, such positions were nonetheless lim ited in number and only accessible to the wellconnected and those who managed to obtain more than the rudimentary education generally available in rural areas of Alto Acre. Si milarly, while cattle production had improved the economic lot of many households, the prospect of reaching the level of Z appeared unlikely for many. Z was an early adopter of cattle ranching and was conse quently able to build a large herd ahead of the implementation of tighten ed environmental governance. The typical household in the region neither had sufficient alte rnative income to shif t away from cattle as Maria had done, nor did they have enough cattle to pay the short-term cost of environmental fines for long-term gain from increased pa sture and herd size as Z had done. Most people in the study region were also mo re vulnerable to environmental and social risks than were Z and Maria. In dry years, th e prospect of losing their crops or Brazil nut trees through wildfire would have much more devasta ting implications than for those like Maria and

PAGE 166

166 Z for whom subsistence crops and Brazil nuts formed a much smaller part of their livelihood system. Likewise, while cattle investments can al ways be liquidated and reinvested in the event of lost land (e.g., if the PAE were to be dissolved to create a colonizati on area), the same cannot be done with the annual subsistence crops and Brazil nut trees upon which poorer families usually depend. Additionally, as poorer families were more likely to obtain their drinking water from streams (as opposed to wells and springs upon which families of greater means tended to rely), poorer families would be at greater risk if climate change, deforestation or pollution from animal waste reduced the availability of clean water in the area’s streams. An example of this contradiction househol ds often face when trying to balance the imperatives of forest conservation and economic de velopment can be found in the attempts of an association in the field site to begin producing and marketing copa iba, a tree resin that has been actively promoted by environmentally-oriented NGOs in the region as a product that can be extracted from the region’s forests. Some years after the project was implemented, the association was storing large quan tities of the resin in a shed adjacent to the association meeting place in the hope that market prices would increa se sufficiently to compensate the effort of extracting it. An informant, commenting on the project and the unsold product, told me he had become discouraged with copaiba, especially when neighbors who had remained outside the program good-humoredly reminded him that if he had dedicated the same amount of effort to raising cattle, he wouldn’t be wai ting for a buyer for his calves. The Alto Acre Paradox and Institutional Design Attention to issues of institu tional design lends insights into the Alto Acre paradox, as it has played out in the lives of Z and Maria. Se veral characteristics of robust institutions, as outlined by Ostrom (2005), are of particular releva nce to the experiences of Z and Maria and to

PAGE 167

167 others within the PAE and in the larger Alto Acre region. I will dr aw attention to two characteristics in particular: costs and benef its of compliance, and graduated sanctions. Regarding proportionality between costs and be nefits of compliance with environmental regulations, Maria is an exception to the rule in Alto Acre. She has parlayed environmental compliance (combined with a large amount of pers onal effort) into a risi ng career first in the social movement, then the municipal government. For most people, the benefits of compliance with environmental regulations are less clear and certainly less immediate. Many people are aware of and concerned about the implications of deforestation fo r regional climate change and water availability. In the case of PAE Santa Quit ria, most people are also aware of the jeopardy that deforestation and the shift from extractivism to cattle ranching is placing on their ability to justify their right to large colocaes to INCRA, a government agency that has faced increasing difficulty in finding sufficient land to settle the co ntinuing tide of landless families in search of a colonia. However, while these threats are generally perceived to be very real, they present themselves in a much longer time horizon than the much more immediat e benefits of illegal deforestation—including increased food secu rity (through crops) and increased household savings (through cattle). If fully funded, th e ProAmbiente program, with its promise of environmental service payments would address this problem. The existence of substantial monetary penalt ies, while helping to balance the costs and benefits of compliance with envi ronmental regulations, is far from perfect. An important factor limiting the effectiveness of environmental penalties is the lack of a clearly defined system of graduated sanctions, with repeated offenses r eceiving more severe penalties than first-time offenses. In fact, the opposite often occurs. Se veral informants told me that after repeated interactions with IBAMA or IMAC officials, farmers become more sophisticated in their ability

PAGE 168

168 to negotiate a fine than a first time offender. Th is is compounded by two further factors. First, at the time of research, IBAMA and IMAC generally offered reduced fines to those who could pay in a single payment, as opposed to a full fine for paying in parcels. This, combined with the fact that most repeat offenders, like Z de Souza, have significantly larger herds (that can be sold off to pay emergency expenses like fines) than firs t-time offenders, creates a system of sanctions that is actually regressive in nature—penalizi ng the first-time offender more severely than the generally wealthier repeat offender. Hence, the only hope a household that is unwilling or unable to emulate the conservati on-oriented model set by Maria, is to grow their herd while avoiding or at least weathering fines until they, like Z, can afford to pay the fines as a part of their operating costs. Reconciling Centralized and Participatory Forms of Governance The failure of household participation in gove rnance to significantly affect land use among those interviewed does not imply a rejection of the decentralization of governance or an argument that centralized governance is a prefer able means of forest conservation. By being active participants in the governance process, many households come to identify with environmental conservation as an important goal for the well-being of themselves and the wider community. Participation in governance may also l ead them to more readily consider alternative technologies and land uses that do not imply the de struction of the forest ecosystem. However the increased awareness of environmental issues and sense of empowerment to affect them that participation in governance brings may only transl ate into concrete acti on if households see any trade-offs implied by these actions as bei ng out-weighed by the benefits. While some households may in fact differ from this tendency and act in purely altruistic form, this is an unrealistic expectation for most of the region’s re sidents, especially those living at or below the poverty line.

PAGE 169

169 Until the economic costs of environmental conservation by a household are out-weighed by the benefits, it may be difficult to translat e environmental attitudes into environmental actions. Despite these obstacles, decentralized participatory governance is still a desirable goal in the region. Rather than placing the onus of enforcement u pon governmental agencies, the proconservation attitudes common in the region have the potential to contri bute to a governance system in which local institutions and individuals assume a complementary role in promoting sustainable land use. Both the economic and non -economic costs versus benefits of land use decisions appear to play a fundamental role in the link between particip ation in governance and land use changes on the ground. For example, in associations such as AM PAESQ in PAE Santa Qu itria, which brought together families that had not yet inter-married and with distin ct backgrounds (including nativeborn residents and indigenous and migrant fa milies who had relocated to the region, from elsewhere in Acre and southern Brazil respective ly), several informants expressed concern about the undue influence they felt certain fa milies had gained in the association.4 Hence, it is important to recognize that, wh ile removing economic obstacles is an important step in translating local governance into environmental action, other non-economic factors such as ethnicity, family identity and imagined histories will also play important roles, whether positive or negative, in this process. Implications of Findings for Forest and Agrarian Policy in Alto Acre Urban and Rural Connections The shift that the highway has brought from low to intermediate levels of rural-urban social and economic connectivity has inevita ble implications for rural land use and 4 Such sentiments were much less common in more homoge neous associations such as Association Porto Carlos—in which nearly all members were nativ e born Acreanos and were inter-related through blood or marriage.

PAGE 170

170 environmental policy. In regions such as Alto Acre, environmental degradation and deforestation cannot be considered solely as ru ral issues with rural implications and rural solutions. On one hand, rural land use change has increasingly significant urban implications as the region’s growing urban favelas fill with ex-ru bber tappers and ex-colonists, as urban water supplies are threatened by regional climate change and as smoke from forest fires threatens the health of the growing populations of young and elde rly urbanites, closes airports and causes deadly accidents on the region’s increasingly trafficked roads. On the other hand, urban public policy has increasingly la rge implications for rural land use; policies affecting the urban economy will have implications in the rural economy as well. In light of increased urban-rural connectivit y, state and municipal governments can utilize urban policy as an indirect means of rural fo rest governance. For in stance, by incorporating larger proportions of forest-based products in foods provided to school children and the needy, the local government can provide a larger in centive for rural produc ers to focus upon these products. And, by subsidizing the creation of forest -based industries in the region’s towns, not only can the burgeoning urban workforce find in creased job opportuni ties, but the rural landholder can find an added market for fore st-based goods. Recent government-sponsored initiatives in Alto Acre such as urban-base d operations processing fr uit pulp, timber and nontimber forest products are importa nt steps in this direction. Environmental Infraction Penalties Another way in which environmental policy c ould better integrate wi th local economic and social realities of rural livelihoods would be thro ugh a careful review of the ways in which fines for illegal land use are levied. As mentioned in th e discussion of the Alto Acre paradox, fines for illegal land use (1000 reais or US$4 55 per hectare of illegal clearing or burning of forest) tend to strike smallholders hardest, given that larger cattle owners generally have enough reserve capital

PAGE 171

171 to pay the fine as a short-term production cost. Such a fine can have a devastating impact on a smallholder without sufficient sa vings (usually implying cattle) to cover the cost. Possible solutions would be to eliminate the reduction of fines for early payment and to find alternative non-monetary means of paying fines. An opportunity exists to transform fines from strictly punitive measures into mechanisms to economically compensate environmental services That is, environmenta l fines could be made payable, in part or in whole, through environmental services. An example would be the option of reducing a 1000 real fine by 5 reais for each tree seedling successfully planted on a property. Not only would this give smallholders the opport unity to pay fines through labor rather than cash, it could also have the effect of ameliora ting the environmental damage incurred through the infraction while improving the l ong-term economic security of the household as seedlings develop into marketable trees. Cattle Another means of reconciling policy with economic realities would be through careful attention to policies directed toward cattle producti on in order to ensure the sustainability of this lucrative industry while compromising neither th e ecological integrity of the region nor other economic activities such as Brazil nuts. An ex ample would be the effective enforcement and promotion of best pasture management practic es, such as the protection of waterways, enforcement of overall land conversion limits, and increased funding for programs such as ProAmbiente that consider the entire household livelihood strategy, including cattle. Carpenter et al., (2001) assert that “the greater the number and equitability of potential uses of ecosystem goods and services, the higher the resilience of the SES” (Carpent er et al., 2001: 776). This argument lends support to an integrated m odel which sees cattle by providing a capital investment to numerous smallholders, as enhanc ing rural livelihoods, as long as it is contained

PAGE 172

172 and operates at a level that does not push othe r socio-ecological systems (e.g., forest cover and water quality) over critical thresholds. ProAmbiente and Environmental Service Payments The ProAmbiente program has emerged as an innovative means for average households to successfully navigate through the often contradi ctory economic, social, and ecological realities of contemporary Alto Acre. ProAmbiente offers technological assistance and credit for various perennial crops that have the potential to mimi c the income generating and capital investment qualities of cattle while capital izing upon the region’s transporta tion infrastructure and avoiding the predatory trajectory that tends to be associated with the produc tion of annuals and, especially, cattle. Importantly, the program also attempts to give greater economic value to forest conservation, providing cash payments in exchan ge for the environmental services rendered through maintained forest cover and avoided deforestation. Doing so creates a market mechanism to transform clean water, biodivers ity conservation and car bon sequestration into farm products capable of competing with beef as part of the house hold’s livelihood strategy. Furthermore, cash payments serve to boost the household’s resilien ce to shocks that could occur as they experiment with more environmen tally benign, but economically risky, land use practices. Unfortunately, funding issues have pl aced the program in jeopardy. However, should the program be fully implemented in the long-ter m, it will serve as a novel experiment in helping households navigate beyond the political-economic contradictions of the Alto Acre Paradox. Fire Acre’s state government took an enormous st ep in late 2005 toward promoting economic, ecological and economic resilience at all scales from the household farm to the entire state when it strengthened environmen tal policies oriented toward dimini shing the use of fire. As of 2006, fines for illegal pasture burning ha d been increased from 1000 to 5000 reais per hectare (that is,

PAGE 173

173 approximately US$450 and US $2275, respectiv ely), rightly reflecting the negative consequences of this activity. Fire, while necess ary for some activities such as clearing forest for annuals and perennials, has frequently been misu sed, with disastrous im plications not only for the region’s forests, but for crops, air qua lity and even for cattle production. Unlike deforestation, which could be argued to be a “ne cessary evil” given curren t livelihood strategies, fire, especially in pastures, has little if any positive aspects. This is evidenced by the fact that the region’s most successful ranchers have by and large eliminated fire as a strategy for pasture maintenance. By singling out this “lose-lo se” land use, Acre’s government has taken an important step in simultaneously protecting th e state’s forests and improving rural livelihoods. Other Amazonian states should consider following this example. Directions for Future Research Gender and Migration One impact of the Inter-Oceanic Highway that deserves further attention is the phenomenon of bi-localism and its implications fo r land use. In the case of Alto Acre, by disproportionately affecting women, rural-urban migration has distin ct implications for land use, especially of Brazil nuts and cattle. Further studies are need ed, both in the Inter-Oceanic highway corridor and in other areas of the Amaz on that have been recently impacted by road construction and paving. Such research will help determine to what degree these findings can be generalized beyond the context of these respondents in Alto Acre. Of special importance for future research is the relationship found in the dissertation between the presence of adult females in the household and higher cattle production. Base d upon qualitative data, I have explained this relationship as likely bei ng caused by gendered differences in approaches to cattle ranching, with women being more hesitant to sell calves than men. More research, both quantitative and qualitative, is needed to further evaluate this pr emise—both in Alto Acre and in other regions of

PAGE 174

174 the Amazon in which the expansion of cattle ranchi ng is threatening conservation efforts. Also, more research is needed to unders tand the full range of factors, and their relative importance, that influence a family’s decision to expand, maintain or decrease its cattle herd. Land Use Governance The findings in this dissertation indicate severa l important areas for future research. Based on this case study, the dissertation provides ev idence supporting the capacity of governance to stem the deforestation and spread of cattle ranchi ng that have traditionally followed Brazil’s road network as it expands into prev iously isolated regions of th e Amazon Basin. However, more research is needed in other regions of the Amazo n, especially in areas of the Amazon with other LULCC dynamics (e.g. large-scale logging, soybean fa rming) and in regions with other types of administrative units (e.g. indigenous reserves Sustainable Development Projects). In Alto Acre itself, at least two governance in terventions, which were nascent at the time of research, will likely prove fertile ground fo r evaluating the implications of governance for land use in the region. At the ti me of fieldwork, the ProAmbiente program was in the process of implementation and families had yet to receive th eir first payment for environmental services. As the program matures and becomes an establ ished part of Alto Acre’s environmental governance landscape, more research will be need ed to determine the e ffectiveness of this program in changing the cost and benefit equati on for small farmers as they decide among a variety of land use options—some of which being more environmentally sustainable than others. The use of fire is another important area of re search in regions of th e Amazon such as Acre that have recently placed strict restrictions on it s use. This dissertation has made the argument that effective environmental governance must bri dge the gap between the costs and benefits of compliance. It has also demonstrated that many households, through experimentation, have found that reducing or eliminating the use of fire, especially in pastures, has yielded beneficial

PAGE 175

175 results. Should these premises be true, recent restrictions on the use of fire should have much greater compliance than regulations, such as defo restation limits, that have a much less clear link to the household’s economic well-being.

PAGE 176

176 APPENDIX A HOUSEHOLD SURVEY Formulrio Entrevista Familiar 1. Nome do entrevistador(s): 2. Data___/___/____ 3. Nome do(s) entrevistado(s): 4. Localizao da propriedade UTM da casa: Leste:____________ Norte_______________ Latitude _______________ Longitude ______________ 5. Localizao de fronteiras da propriedade UTM: Leste__________ Norte__________ UTM: Leste__________ Norte__________ UTM: Leste__________ Norte__________ UTM: Leste__________ Norte__________ UTM: Leste__________ Norte__________ 6.Apelido (s): _________________________________________________________ 7. Identificao da Colocao / Lote: Municpio: Unit: Seringal: Colocao: Ramal: Numero de hectares na propriedade: Pode fazer um mapa de sua colocao / lote? 8. Propriedade da Terra

PAGE 177

177 Vocs tm o documento de sua terra? Sim____No Qual (caso que sim)? ______________________________________________________ Porque (caso que no)? ____________________________________________________ Em nome de quem est a terra de vo cs (casa que sim)? Esposo________ Esposa__________ Outro (quem) : 12. Vocs esto procurando outro tipo de documento agora? Sim______ No_______ Caso que sim, qual?__________________ Historia Migratria Migrao 13. Onde a senhora nasceu? 14. E o senhor? 15. Quanto tempo vocs moram no Acre (anos)?___________ (se o local de nascimento for outro) 16. Ano em que chegaram neste lote / colocao________________________ 17. Como vocs chegaram at essa colocao (estria migratria)? ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ ________________________________________________ 18. E vocs se conheceram como? 19. Vocs sabem quando este lote / coloca o foi criada? E quem era o dono antes de vocs?__________________________________________________________________ ______________________________________________________________________________ __________________________________________________________________ 20. Nos ltimos 4 anos, algum de sua famlia saiu da reserva? Quem?

PAGE 178

178 Caso que sim, para onde?___________________________________________________ Porque?_______________________________________________________________________ ______________________________________________________________________________ ____________________________________________________________ 21. Vocs tiveram que dividir sua colocao com algum? Sim _____No______ Com quem?____________________________ Porque?_______________________________________________________________________ __________________________________________________________________ 22. E fizeram algum documento para isso? Sim _____No______ Qual (caso que sim)? ______________________________________________________ Porque (caso que no) _____________________________________________________ 23. Vocs esto satisfeitos de morar no PAE/PC? (H) Sim______No______ (M) Sim______Nao______ Por que? (Homem) _______________________________________________________________ (Mulher)______________________________________________________________________ __________________________________________________________________ 24. Vocs pretendem ficar morando no PAE/PC? Sim _____No______ 25. Quanto tempo, por ms, vocs ficam na rua? Homem_________________________________________________________________ Mulher__________________________________________________________________ 26. Quando vo para rua, onde ficam? Prpria casa_____________ Com famlia_____________ Outro (que)__________________ 27. Vamos fazer um pensamento: O que vocs pensam para o futuro? Como vocs querem estar daqui a 4 anos? ________________________________________________________________________ ______________________________________________________________________________ __________________________________________________________________

PAGE 179

179 Demografia, Educao, Religio, Sade e Comunicao 28. Pessoas na famlia Nome Parentesco Idade Nvel de Estudo Estudando Agora? Onde mora? 29. Vocs tm alguma religio? Caso que sim, qual? Onde fica? 30. No ultimo ms, quantos vezes a senhora foi? E o senhor?_____________________________________________________________ 31. Tem um posto de sade aqui dentro? Sim___________ No_______________ Caso que sim, quem fez? ______________________________________ _______________________________________________________________________ 32. De que forma vocs decidem sobre produo, sa de, educao, coisas relacionadas para os filhos? ______________________________________________________________________________ ______________________________________________________________________________ ____________________________________________________________ 33. Vocs alguma vez j sarem daqui para morar na cidade? Sim______ No__________

PAGE 180

180 Porque?_______________________________________________________________________ ______________________________________________________________________________ ____________________________________________________________ Recursos Financeiros Quais desses itens vocs possuem? Itens Tipo/ Conservao Unidade Rdio Luz Fogo Curral para gado Cavalo Moto Carro/caminho Casa (tamanho, vedao, tijolo ou madeira conservao da madeira e tipo de madeira – torta ou no, fabricao prpria ou no ..) Vocs esto usando cr edito? Sim_________ No____________ Qual tipo (caso que sim)?_________________________________________________ Para fazer o que (caso que sim)?_________________________________________________________________________ __________________________________________________________________ No media, quanto vocs ganham por ms?________________________ De onde (e.g produo, dirio, bolsa familiar, etc)? Tipo Fonte Quanto (por ms)

PAGE 181

181 Tipo Fonte Quanto (por ms) Aceso a Estrada e LULCC Vocs lembram quando este ramal saiu?_____________________________ E entra carro durante quais meses por ano?_____________________________________ Como vocs fazem para sair daqui para o asfalto? ___________________________________________ Tipo De quem ? Carro / caminho Moto Bicicleta Cavalo / burro A pe Barco Outro Quantas horas vocs demoram para chegar ate a estrada no vero, sa indo da porta da sua casa?_________________ _______________________________________________________________________ O que mudaria em relao produo de vocs se o ramal fosse melhor?___________________________________________________ ______________________________________________________________________________ _____________________________________________________________ Quem vocs acham que responsvel pelo melhoramento dos ramais?______________ _______________________________________________________________________ _______________________________________________________________________ A vida de vocs mudou por causa do asfaltamento da estrada? Sim_________ No___________

PAGE 182

182 Caso que sim, como (sade, educao, lazer)?________________________________ _______________________________________________________________________ E o tipo de produo, depois do asfaltamento da estrada? Sim_____No______ Caso que sim, Por que? ____________________________________________________ ______________________________________________________________________ ______________________________________________________________________ ______________________________________________________________________ Governncia e Organizao Social: Vocs fazem mutiro por aqui? Sim__________ No_______________(com quem?) Vizinhos_____________ Famlia______________ associados da mesma associao_______________ outro (quem?) ___________________________ Quais so as pessoas mais podero sas aqui nesta comunidade?________________________

PAGE 183

183 Organizao Associa do? (quem?) Desde Quand o Para que serve esta organiza o? E sua famlia, com que freqncia assistem as reunies?** Algum desta coloca o teve cargo? (S/N) Qual funo? Quando? Voc est satisfe ito com essa organi zao? ** AMPAESQ Associao local Cooperativa (qual?) __________ Sindicato (qual?) __________ AMOREB Outro (qual)

PAGE 184

184 Para que vocs acham que serve uma associao de mulheres? Homem: Mulher: (1) 01 por ms, (2) 01 cada 03 meses (3) 01 cada 06 meses (4) 01 cada ano **(1) "timo (2) Bom (3) Ruim Quais so as regras ambientais que todo mundo segue? ________________________ ______________________________________________________________________ Por que? ________________________________________________________________________ ________________________________________________________________________ Quais so as regras ambienta is que devem ser mudadas? ______________________________________________________________________________ __________________________________________________________________ Instituio Para que serve este rgo? Vocs estao satisfeitos com essa organizao?** INCRA IMAC IBAMA Prefeitura

PAGE 185

185 Por que? _______________________________________________________________ ________________________________________________________________________ No ultimo ano, quantos vezes o senhor viu alguma pessoa do governo vindo fiscalizar na reserva? ________________Se viu, de que rgo era ele (a)? O senhor j recebeu um multa do IBAMA/IMAC? Sim ______No_______ Qual foi a ultima vez? _____________________________________________________ Caso que sim, por que?_____________________________________________________ Qual foi a ultima vez (conduzir para ltimos 4 anos) que o senhor soube de algum aqui da reserva que recebeu uma multa do IBAMA/IMAC? _________________________________ ________________________________________________________________________ ________________________________________________________________________ E lembra por que (caso que sim)? _____________________________________ ________________________________________________________________________ ________________________________________________________________________ Lembra se na sua associao quando a associa o mesmo fiscalizou o uso da terra dum scio? Sim_______ No___________ Caso que sim, quando?_____________ Por qu?____________________ Cursos Alguma vez participou em algum curs o (e.g. fogo, manejo, artesanato, etc)? Homem: Sim__ No___ Mulher: Sim No___ Caso que sim, quais? Nome do projeto________________ Quando?______________ Organizado por quem?___________________ Valeu a pena?___________________________ Por que?__________________________ Projetos Vocs j participarem em algum pr ojeto do governo ou outra organizao? Homem: Sim__ No___

PAGE 186

186 Mulher: Sim___ No___ Caso que sim, quais? _____________________________________________________ Nome do projeto 1________________ E ainda est participando? Sim____________ No______________ Mudou alguma coisa para sua famlia depoi s que o senhor participou deste programa? _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ Nome do projeto 2______________________ Est participando ainda? Sim____________ No______________ Mudou alguma coisa para sua famlia depois que vocs participaram deste programa? _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ ProAmbiente Conhece o programa ProAmbiente? Caso que no, omite o reste das perg untas sobre ProAmbiente Sim_______ No_______ Caso que sim, como conheceu este programa? _________________________________ ________________________________________________________________________ O que este programa significa para vocs? _______________________________________________________________________ _______________________________________________________________________ Vocs esto participando no programa? (Caso que no, omite o resto das perguntas sobre ProAmbiente) Sim___________ No_____________ Por que?________________________________________________________________ Quais so as mudanas que vocs acham que podem acontecer na sua produo e na sua renda com sua participao neste programa? ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ Vocs j fizeram o plano de uso do programa? Sim____________ No___________ Ele sarem como vocs queriam (caso que sim)? Sim____________ No___________ Por que ?________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________

PAGE 187

187 Vocs j conversarem com outros mora dores sobre o programa? Sim______ No_____ Com quem (caso que sim)?__________________________________________________ E o que vocs conversaram?_________________________________________________ ________________________________________________________________________ ________________________________________________________________________ Uso do da Terra e Desmatamento O que o senhor faz quando uma rea de terra no produz mais? ________________ ___________________________________________________________________ ___________________________________________________________________ ___________________________________________________________________ E como faz a deciso de pl antar capim ou deixar vira capoeira?______________________________________________________________________ _________________________________________________________________ Para que tipo de produo o senhor acha que sua terra d? ___________________________________________________________________ Quais os tipos de terra que sua rea possui hoje e quando chegaram nessa colnia/colocacao? (em hectares) Tipo atualmente quando chegaram Capoeira Pasto Roado Mata bruta Vocs botaram roado esse ano? Sim No___________ Quanto (caso que sim)? ________________________________ E vocs conseguiram tirar a licena no IBAMA? Deu tempo? Sim_____________ No_____________________ E ano passado? ___________________________________________________________ No ultimo ano, vocs plantaram arvores em seu terreno? Quantas (caso que sim)?________________ De qual tipo?______________ Por que?________________________________________________________________ Onde? __________________________________________________________________ J plantou algum tipo de leguminos a aqui? Qual (caso que sim)?____________________ E deu certo? _____________________________________________________________ Como chegaram deciso plantar isso?

PAGE 188

188 ______________________________________________________________________________ __________________________________________________________________ O que o senhor faz para colocar fogo em seu lote e ele no entrar no lote do seu vizinho? Existe alguns problemas na sua rea com fogos acidentais? Sim________ No Em sua propriedade, j tem algum tipo de produo agroflorestal? Sim_______ No Caso que sim, o que produz?_________________________________ Caso que sim, onde colocou? (m ata bruta, capoeira, etc)___________________________ E deu certo? Comercializao de produtos extrativistas no-madeireiros Qual a importncia destes produtos para sua renda Produto 2000* 2004** 2008* PRINCIPAIS Castanha Borracha (qual forma) Madeira FRUTOS Aa Pato (fruto) Bacaba Buriti Ouricuri Murmur (fruto) Tucum (fruto)

PAGE 189

189 Jatob (fruto) Juta (fruto) Bacuri PALMITOS LEOS/DERIVADOS Pato (leo) Copaba (leo) Jatob (leo) *1=mais, 2=mesmo, 3=menos **% de renda familiar Vocs tm gado? Sim No___________ Quanto?_________________ Desde quando?________________ Por que vocs criam gado (caso que sim)? Por que vocs no criam gado (caso que no)?__________________________________ E quanto pretende ter daqui a 4 anos? Sua vida mudou depois que vocs comearam a criar gado? Sim No__________ Por que? Que tipo de pasto o senhor tem? Como limpa seu pasto? Com que freqncia usa fogo para limpar o seu pasto (depois de ser formado)? Quando foi a ultima vez que usou fogo para limpar o seu pasto (depois de ser formado)? Comentrios sobre tcnica e fr eqncia de limpar pasto Tem algumas questes, duvidas ou sugestes sobre esta pesquisa?

PAGE 190

190 LIST OF REFERENCES Adger, W.N., and T.A. Benjaminsen 2001. Advancing a political ecology of global environmental discourses. Development and Change 32 : 681-715. Adger, W.N., K. Brown, J. Fairbrass, A. Jordan, J. Pavola, S. Rosendo, G. Seyfang 2003. Governance for sustainability: to wards a ‘thick’ analysis of environmental decision making. Environment and Planning 35 : 1095-1110. Agrawal, A., and C.C. Gibson 1999. Enchantment and disenchantment: the role of community in natural resource management. World Development 27 (4): 629-649. Alencar, A. C, and. L.A. Solrzano, D.C. Nepstad 2004. Modeling forest under story fires in an eastern Amazonian landscape. Ecological Applications 14 (4) (Supplement): S139– S149. Allegretti, M.H 1990. Extractive reserv es: an alternative for reconciling development and environmental conservati on in Amaznia. Pages 252-264 in A.B. Anderson, editor. Alternatives to deforestation: steps towa rd sustainable use of the Amazon rainforest. Columbia University Press, New York. Almeida, R. G (n.d.). Degradao, recuperao e sust entabilidade de pastagens cultivadas. [online] URL: http://www.forragicultura.com.br/arquivos/degrad acaorecuperacaosustentabilidadepastagens.pdf (last accessed on October, 23, 2006). Alston, L., G. Libecap, and B. Mueller 1999 Titles, conflict, and land use: the development of property rights and land refo rm on the Brazilian Amazon frontier The University of Michigan Press, Ann Arbor. Andrade, C.M.S., J.F. Valentim, and J. C. Carneiro 2002. Arvores de baginha ( Stryphnodendron guianense (Aubl.) Benth.) em ecossistemas de pastagens cultivadas na Amazonia Ocidental. Revista Brasileira Zootcnica 31 (2): 574-582. Andrade, C.M.S, J.F. Valentim, J.C Carneiro, and F.A. Vaz 2004. Crescimento de gramneas e leguminosas forrageiras tropicais sob sombreamento. Pesquisa Agropecuria Brasileira 39 (3): 263-270. Angelsen, A., and D. Kaimowitz 1999. Rethinking the causes of deforestation: lessons from economic models. The World Bank Research Observer 14 (1):73-98. Arima, E.Y., and C. Uhl 1997. Ranching in the Brazilian Amazon in a national context: economics, policy, and practice. Society and Natural Resources 10 :433-451. Bakx, K.S. 1988. From proletarian to peasant: rura l transformation in the State of Acre, 1870-1986. Journal of Development Studies 24 (2):141-160.

PAGE 191

191 Barbosa, L.C 2000. The Brazilian Amazon Rainforest: gl obal ecopolitics, development, and democracy. University Press of America, Lanham. Barlow, J., and C.A. Peres 2006. Effects of single and recurrent wildfires on fruit production and large vertebrate abunda nce in a central Amazonian forest. Biodiversity and Conservation 15 :985–1012. Biersack, A. 1999. From the “new ecology” to the new ecologies. American Anthropologist 101 (1):5-18. Blaikie, P 1985. The political economy of soil erosion in developing countries Longman Press, London. Blaikie, P., and H. Brookfield 1987. Land degradation and society. Methuen.: London. Browder, J.O., and B.J. Godfrey 1997. Rainforest cities: urbanization, development, and globalization of the Brazilian Amazon. Columbia University Press, New York. Brown, J.C., and M. Purcell 2005. There’s nothing inherent about scale: political ecology, the local trap, and the politi cs of development in the Brazilian Amazon Geoforum 36 :607–624. Brown, K. and S. Resendo 2000a. Environmentalists, rubber tappers and empowerment: the politics and economics of extractive reserves. Development and Change 31 (1):201-227. Brown, K. and S. Resendo 2000b. The institutional architectu re of extractive reserves in Rondnia, Brazil The Geographical Journal 166 (1):35-48. Bryant, R.L. and S. Bailey 1997. Third world political ecology Routledge Press, London and New York. Bunker, S.G. 1985. Modes of extraction, unequal exchange, and the progressive underdevelopment of an extreme pe riphery: the Brazilian Amazon 1600-1980. American Journal of Sociology 89 : 1017-1064. Cardoso, C 2002. Extractive reserves in Brazil ian Amaznia: local resource management and the global political economy Ashgate, Burlington, VT. Carneiro, J. C., and J.F. Valentim 1997. Recomendaes tecni cas para a intensificao da pecuaria de corte no Acre. Instrues Tcnicas EMBRAPA 7 :1-3. Carney, J. 1996. Converting the wetlands, engenderi ng the environment: the intersection of gender with agrarian change in Gambia. Pages 165-187. in R.Peet and M. Watts, editors. Liberation Ecologies: environment, development, social movements. Routledge, London and New York.

PAGE 192

192 Carpenter, S. B. J.M. Walker, and Anderies, N. Abel 2001. From metaphor to measurements: resilience of what to what? Ecosystems 4 (8): 765-781. Carpentier, C., S. Vosti, and J. Witcover 2000. Intensified production systems on western Brazilian Amazon settlement farms: could they save the forest? Agriculture, Ecosystem and Environment 82 :73-88. Carvalho, G., P. Moutinho, D. Nepstad, L. Mattos, and M. Santilli 2004. An Amazon perspective on the forest-climate connec tion: opportunity for climate mitigation, conservation and development? Environment, Development and Sustainability 6 :163–174. Chomitz, K. M., and D. A. Gray 1996. Roads, lands, markets and deforestation: a spatial model of land use in Belize. The World Bank Economic Review 10 (3):487-512. Chomitz K.M., and T. Thomas 2001. Geographic patterns of land use and land intensity in the Brazilian Amazon. Policy Research Working Paper 2687. The World Bank Development Research Group Infrastructure and Environment. Cochrane, A, A. Calendar, M.D. Schulze, C.M. Souza, D. Nepstad, P. Lefebvre, and E. Davidson. 2003. Positive feedbacks in the fire dynamic of closed canopy tropical forests. Science 284 :1832-1835. Da Silva, S.S. 2004. O espao agrrio acreano nas ultimas dcadas do sculo XX. Revista NERA 7 (4):42-49. Davidson E. and P. Artaxo 2004. Globally significant changes in biological processes of the Amazon Basin: results of the Larg e-scale Biosphere–Atmosphere Experiment. Global Change Biology 10 :519–529. Demsetz, H 1967. Toward a theory of property rights. American Economic Review 57 : 347359. Faminow, M.D 1997. Spatial economics of local demand for cattle products in Amazon development. Agriculture, Ecosystems and Environment 62 :1-11. Faminow, M. D 1998. Cattle, deforestation, and development in the Amazon: an economic, agronomic, and environmental perspective CAB International Publishers, Wallingford, CT. Fearnside, P.M 1989. Extractive reserves in the Brazilian Amazon BioScience 39 : 387-393. Fearnside, P.M 2000: Deforestation impacts, environmental services and the international community. Pages 11-24 in A. Hall, editor. Amaznia at the crossroads: the challenge of sustainable development. Institute of Latin American Studies, London.

PAGE 193

193 Fearnside, P.M 2001a. Soybean cultivation as a thre at to the environment in Brazil. Environmental Conservation 28 (1): 23-38. Fearnside, P.M 2001b. Land tenure issues as factor s in environmental destruction in Brazilian Amaznia: the case of southern Para. World Development 29 (8):1361-1372. Fearnside, P.M. 2002a. Can pasture intensificati on discourage deforestation in the Amazon and Pantanal regions of Brazil? Pages 299-314 in C. Wood and R. Porro, editors. Deforestation and land use in the Amazon. University of Florida Press, Gainesville. Fearnside, P.M 2002b. Avana Brasil: Environmental and social consequences of Brazil’s planned infrastructure in Amaznia. Environmental Management 30 (6): 35-747. Fearnside, PM 2003. Deforestation control in Ma to Grosso: A new model for slowing the loss of Brazil's Amazon Forest AMBIO 32 (5):343-345. Feaster, J.G 1970. An analysis of the relationship betw een infrastructure and agricultural development in Caqueta, Colombia. Ph.D. Dissertation. University of Kentucky, Lexington. Foley, J.A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin, M. T. Coe, G. C. Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C. J. Kucharik, C. Monfreda, J. A. Patz, I. C. Prentice, N. Ramankutty, and P. K. Snyder 2005. Global consequences of land use Science. 309: 570-574. Fujisaka, S. 1996. Slash-and-burn agricultu re, conversion to pasture, and deforestation in two Brazilian Amazon colonies. Agriculture, Ecosystems and Environment 59 :115-130. Geist H. and E. Lambin 2001. What drives tropical defore station? A meta-analysis of proximate and underlying causes of defore station based on subnational case study evidence. LUCC Report Series No. 4. LUCC Intern ational Project Office, Louvain, Belgium. Geist, H. and E. Lambin 2002. Proximate causes and underlying forces of tropical deforestation. Bioscience 52 (2):143-150. Gerwing J.J 2002. Degrading forests through logging and burning in the Eastern Amazon. Forest Ecology and Management 157 :131-141. Governo do Estado do Acre 2000. Zoneamento Ecologico-Economico do Estado do Acre Rio Branco, Acre. Hall, A ., editor. 2000. Amaznia at the Crossroads: The Challenge of Sustainable Development. Institute of Latin American Studi es, University of London, London. Hardin, G 1968. The Tragedy of the Commons. Science 162 :1243-1248.

PAGE 194

194 Hecht, S.B 1985. Environment, development and politics: capital ac cumulation and the livestock sector in eastern Amaznia. World Development 13 (6):663-684. Hecht, S.B. and A. Cockburn. 1990. The fate of the forest: developers, destroyers, and defenders of the Amazon. Verso, London. Herringer, I. A., and V. A. Jacques 2002. Acumulao de forragem e material morto em pastagem nativa sob distintas alternativas de manejo em relao as queimadas. Revista Brasileira Zootcnica 31 (2): 599-604. Hoffmann. W.A. W. Schroeder, and R.B. Jackson 2003. Regional feedbacks among fire, climate, and tropical deforestation. Journal of Geophysicial Research 108 Jarosz, L 1996. Defining deforestation in Madagascar. Pages 148-164 in R. Peet and M. Watts, editors. Liberation Ecologies: environment, development, social movements. Routledge: London and New York. IBGE. 2004. [online] URL: h ttp://www.ibge.gov.br/cidadesats (Last accessed on August, 10, 2006). Houghton R.A., D.L. Sk ole, and C.A. Nobre 2000. Annual fluxes of carbon from deforestation and regrowth in the Brazilian Amazon. Nature 403 :301–304. Kainer, K.A., M. Schmink, A.C.P. Leite, and M.J. da Silva Fadell 2003. Experiments in forest-based development in Western Amaznia. Society and Natural Resources 16 :869886. Keck, M.E 1995. Social equity and environmen tal politics in Brazil: lessons for the rubber tappers of Acre. Comparative Politics 27: 407-425. Kirby, K.K, W.F. Laurance, and A.K. Albernaz 2006. The future of deforestation in the Brazilian Amazon. Futures 38 :432–453. Kiser L. and E. Ostrom. 1982. The three worlds of action: a meta-theoretical synthesis of intuitional approaches. Pages 179-222 in E. Ostrom, editor. Strategies of Political Inquiry. Beverly Hills, CA, Sage Press. Laurance, W., M. Cochrane, S. Bergen, P. Fearnside, P. Delamonica, C. Barbgeer, C, S. D’Angelo, and T. Fernandes 2001. The Future of the Brazilian Amazon. Science 291 : 438-439. Laurance WF, A.K.M. Albernaz, G. Schrot h, PM. Fearnside, S. Bergen, E.M. Venticinque, and C. Da Costa 2002. Predictors of deforest ation in the Brazilian Amazon. Journal of Biogeography 29 (5-6):737-748.

PAGE 195

195 Liverman, D.M. E.F. Moran, R.R. Rindfuss, and P.C. Stern 1998. People and pixels: linking remote sensing and social science National Academy Press, Washington, D.C. Machagata, M. and K. Brown 2003. Cows, colonists a nd trees: rethin king cattle and environmental degradation in Brazilian Amaznia. Agricultural Systems 76 : 797-816. Mahar, D. J 2002. Agro-ecological zoning in R ondnia, Brazil: what are the lessons? Pages 115-128 in A. Hall, editor. Amaznia at the crossroads: th e challenge of sustainable development University of London, London. Blackwell Publishing, Ltd. Maslin, M., M. Yadvinder, O. Phillips, and S. Cowling 2005. New views on an old forest: assessing the longevity, resilience and future of the Amazon rainforest. Transactions Institute of British Geographers 30 477–499. Mattos M. and C. Uhl 1991. Social, economic, and ecological consequences of selective logging in an Amazonian frontier – the case of Tailandia. Forest Ecology and Management 46 (3-4): 243-273. Mendona M 2004. Estimao do Dano Sade Humana devido ao Uso do Fogo na Amaznia. Mrio Jorge Car doso de Mendona (IPEA/RJ). Moore, D 1996. Marxism, culture, and political ecology: Environmental struggles in Zimbabwe’s Eastern Highlands. Pages 125-147 in R. Peet and M. Watts, editors. Liberation ecologies: environment, development, social movements Routledge, London and New York. Moran, E.F. 1982. Colonization in the Tran samazon and Rondnia. Pages 285-303 In Schmink M. and C. Wood, editors. Gaines ville, University of Florida Press. Nepstad D., A. Verssimo, A. Alencar, C. Nob re, E. Lima, P. Lefebvre, P. Schlesinger, C. Potter, P. Moutinho, E. Mendo za, M. Cochrane, and V. Brooks 1999. Large-scale impoverishment of Amazonian forests by logging and fire. Nature 398: 505–508. Nepstad, D., Capobianco, J. P., Barros, A., Ca rvalho, G., Moutinho, P., Lopes, U., and P. Lefebvre. 2001. Avana Brasil: os custos ambientais para a Amaznia Instituto de Pesquisa Ambiental da Amaznia and Instituto Socioambiental. Nepstad, D, A. McGrath, A. Alencar, B. Barros, M. Carvalho, M. Santilli, and V. Diaz 2002. Frontier governance in Amaznia. Science 295 : 629-631 Orlove, B.S 1980. Ecological anthropology. Annual Review of Anthropology 9 : 235-273. Ostrom, E 1990. Governing the commons: the evolution of institutions for collective action. Cambridge University Press, Cambridge. Ostrom, E 1997. Self-governance and forest resources. CIFOR Occasional Paper 20 : 1-15.

PAGE 196

196 Ostrom, E., J. Burger, C. Field, R. Norgaard, D. Policansky 1999. Revisiting the commons: local lessons, global challenges. Science 284: 278-283. Ostrom, E 2005. Understanding Institutional Diversity. Princeton University Press, Princeton and Oxford. Ozrio de Almeida, A 1992. The colonization of the Amazon University of Texas Press, Austin. Pawley H 1971. In the year 2070. Ceres 4 (4): 22-27. Peet, R., and M. Watts, editors. 1996. Liberation Ecologies: environment, development and social movements. London, Routledge. Pereira, J.M 2001. Produo e persistncia de leguminosas em pastagens tropicais. Simpsio de Forragicultura e Pastagens NEFOR – UFLA. Peres, C.A 2001. Paving the way to the future of Amaznia. Trends in Ecology and Evolution 16 (5): 217-219. Perz, S.G 2001a. From sustainable development to “productive cons ervation: “forest conservation options and agriculture inco me and assets in the Brazilian Amazon. Rural Sociology 66 (1):93-112. Perz, S. G 2001b Household demographic factors as life cycle determinants of land use in the Amazon. Population Research and Policy Review 20 :159: 186. Pfaff, Alexander S. P 1997. What drives deforest ation in the Brazilian Amazon? Evidence from satellite and socioeconomic data. Policy Research Working Paper Series 1772, The World Bank. Posner, R 1977. Economic analysis of law Little, Brown and Company, Boston. Rappaport, R 1967. Pigs for the ancestors. Yale University Press, New Haven, CT. Reis, A.F 1960. A Amaznia e a cobia internacional Editoria Nacional, So Paulo. Rouchelou, D., B. Thomas-Slayter, and E. Wangari 1996. Feminist political ecology Routledge, London. Rueda, B.L., R. W. Blake, C. F. Nicholson, D. G. Fox, L. O. Tedeschi, A. N. Pell, E. C. M. Fernandes, J. F. Valentim, and J. C. Carneiro 2003. Production and economic potentials of cattle in pasture-based systems of the western Amazon region of Brazil. Journal of Animal Science 81 :2923–2937.

PAGE 197

197 Schmink, M 1992. Building institutions for sustaina ble development in Acre, Brazil. Pages 276-297 in K. Redford and C. Padoch, editors. Conservation of neotropical forests Columbia University Press, New York. Schmink, M. 2000. Gnero y conseracin: nuevos re tos en capacitacin y investigacin. Pages 36-41 In Manejo de recursos naturales de sde una perspective de gnero. Seminrio Taller. Seminrio Permanente de Investigacin Agrari a. Pontificia Catlica del Per, Lima. Schmink, M. and C.H. Wood 1987. The “political ecology” of Amaznia. Pages 38-57 in P.D. Little and M.M. Horowitz, editors. Lands at risk in the third wo rld: local level perspectives. Westview Press, Boulder, CO. Schmink, M., and C.H. Wood 1992. Contested Frontiers in Amaznia Columbia University Press, New York. Schroeder, R., and K. Suryanata 1996. Gender and class power in agroforestry systems: case studies from Indonesia and We st Africa. Pages 188-204 in R. Peet and M. Watts, editors. Liberation ecologies: environment, development, social movements London and New York, Routledge. Schwartzman, S 1989. Extractive reserves: the rubber tappers’ strategy for sustainable use of the Amazon rain forest. Pages 150-165 in J.O. Browder, editor. Fragile lands in Latin America: strategies for sustainable development. Westview Press, Boulder, CO. Schwartzman, S. 1992. Land distribution and the social costs of frontier development in Brazil: social and historical contex t of extractive reserves. Pages 51-66 in D. Nepstad and S. Schwartzman, editors. Non-timber products from tropical fore sts: evaluation of a conservation and development strategy. The New York Botanical Garden, New York. Shelton, H.M., S. Franzel, and M. Peters 2005. Adoption of tropical legume technology around the world: Analysis of success. Pages 149-166 in D.A. McGilloway, editor. Grassland: a global resource Wageningen Academic Publishers, The Netherlands. Silveira, A. 2002 Development of the Brazilian Amazon. Science 292 :1651. Skole, D., W.H. Chomentowski, W.A. Salas, and A.D. Nobre 1994. Physical and human dimensions of deforestation in Amazonia. BioScience 44 (5): 314-313. Smith, N. 1982. Rainforest corridors: The Tran samazon colonization scheme. University of California Press, Berkeley. Soares-Filho, B., A. Alencar, D. Nepstad, G. Cerqueira, M del C. Vera Diaz, S. Rivero, L. Solrzano, and E. Voll 2004. Simulating the response of land-cover changes to road paving and governance along a major Amazon hi ghway: the Santarm-Cuiab corridor. Global Change Biology 10 (5): 745-764.

PAGE 198

198 Sombroek, W. and A. de S. Souza 2000. Macroand micro ecological-economic zoning in the Amazon region: history, firs t results, lessons learned and research needs. German-Brazilian Workshop on Neotropi cal Ecosystems – Achievements and Prospects of Cooperative Research Hamburg, September 3-8. Sontheimer, S 1991. Women and the Environment: A Reader. Earthscan, London. Toni, F. and D. Kaimowitz 2003. Municpios e gesto florestal na Amaznia. A.S. Editores, Natal, Rio Grande do Norte. Valentim, J.F. and C.M.S. Andrade. (n.d.). Tropical kudzu ( Pueraria phaseoloides ): successful adoption in sustainable cattle production systems in the western Brazilian Amazon. XX International Grassland Conference Vayda, A.P 1983. Progressive contextualizatio n: methods for research in human ecology. Human Ecology 11 :265-281. Vayda, A.P.and B.B. Walters 1999. Against political ecology. Human Ecology 27 (1): 167179. Watts, M. 1983. Silent violence: food, fa mine and peasantry in Northern Nigeria. University of California Press, Berkeley. Wolf, E. 1972. Ownership and political ecology. Anthropological Quarterly 45 : 201-205. Wood, C.H. and R. Porro editors 2002. Land use and deforestation in the Amazon. University Press of Florida, Gainesville, FL.

PAGE 199

199 BIOGRAPHICAL SKETCH Growing up in a partially-foreste d rural area of the state of In diana, I have long had a love for the outdoors and for forests and have been intrigued by the interface of human activity with the natural environment. During subsequent trav el, especially in the we stern United States and South Asia, I became fascinated by the role that land management policy has had both for forest ecosystems and for the local communities that depend upon them. These combined experiences have been instrumental in directing my academ ic and career goals toward environmentally and socially sustainable forest management. Maste r’s fieldwork in the Peruvian Andes, including a brief visit to the Peruvian Amazon, has further refined my interests toward natural resource policy in Latin America. In 1998, I received a B.A. in Anthropology from Indiana University. Subsequently, I began a Master’s program in the Anthropology program at the University of Kentucky. As a part of my Master’s program, I conducted ethnographic fieldwork in the Department of Cusco in the southern Peruvian Andes on the social and ec onomic implications of community eucalyptus forestry. Subsequent to rece iving my M.A., in 2002, I began a Ph.D. program in the School of Natural Resources and the Environment (SNRE) at the University of Florid a. As part of my academic program, I concentrated in Tropical C onservation and Development (TCD). I spent a year and a half as a research assistant with UF’s Gender, Environment, Agriculture and Participation program (GEAP). S ubsequently, I spent two and a ha lf years as a fellow in UF’s NSF-funded Working Forests in th e Tropics Program, a program th at has been instrumental in my graduate training. While at UF, I have served as co-president of the SNRE Graduate Council and coordinator of the interd isciplinary Roadies Group. I traveled to Acre in the summer of 2002 and again in the summer of 2003 to conduct preliminary fieldwork. I returned to Acre in June of 2004, beginning fieldwork research in the

PAGE 200

200 Alto Acre region that would continue through July of 2005. In the summer of 2006, I again returned to Acre to return research results. Th is experience included pres entations in association meetings, rural schools and in homes in the field site.


Permanent Link: http://ufdc.ufl.edu/UFE0017264/00001

Material Information

Title: Roads, Governance and Land Use in the Brazilian State of Acre
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0017264:00001

Permanent Link: http://ufdc.ufl.edu/UFE0017264/00001

Material Information

Title: Roads, Governance and Land Use in the Brazilian State of Acre
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0017264:00001


This item has the following downloads:


Full Text





ROADS, GOVERNANCE AND LAND USE IN THE BRAZILIAN STATE OF ACRE


By

JEFFREY BENSON LUZAR












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2006

































Copyright 2006

by

Jeffrey Benson Luzar

































To the people of Alto Acre









ACKNOWLEDGMENTS

I would like to thank my chair, Dr. Marianne Schmink, for her help and guidance from my

first days at UF through the completion of the dissertation. I am especially grateful for her

helpful comments and insights throughout the writing process. I also would like to thank my

committee members Dr. Karen Kainer, Dr. Stephen Perz, Dr. Janaki Alavalapati and Dr. Rick

Stepp, each of whom has provided valuable advice and insights throughout my doctorate

program. In addition to the members of my committee, I would also like to thank several other

individuals at UF who have helped me transform research questions into a completed

dissertation. In anthropology, Dr. Russ Bernand offered very helpful guidance in the process of

proposal writing and the creation of testable hypotheses. Dr. Grenville Barnes of the SFRC has

been helpful in encouraging me to contemplate the various dimensions of land tenure as I

developed the dissertation. Dr. Robert Buschbacher, of WEC and the WFT program, has helped

me think through my ideas and has served as a mentor from my first weeks at UF. Finally, Dr.

Chuck Wood in Latin American Studies provided truly invaluable guidance during the tedious

process of data analysis. I am also grateful for financial support from the Hewlitt Foundation,

the Working Forests in the Tropics Program, the National Science Foundation and the Fulbright-

Brazil Program for making fieldwork possible

I also thank my family, including my parents Katherine Ann Luzar and Louis Frantz

Luzar, for their encouragement as I pursued my dreams and my aunt Dr. E. Jane Luzar, who first

introduced me to the opportunities available at UF and helped me in the transition to life in

Gainesville. I would like to offer an additional word of thanks to my friends and colleagues at

UF and the WFT program, especially Hilary del Campo, who has offered encouragement, hugs

and laughter through both my greatest and most difficult moments while at UF. I would also like

to offer a word of thanks to the many people at UFAC, PZ and SETEM in Rio Branco, including,









though not limited to, Elsa Mendoza and Dr. Irving Foster Brown who have helped me develop

my ideas, and have offered logistical support and friendship from the earliest days of my

research process. Finally, an especially heartfelt word of thanks goes to the people of Alto Acre

who have invited me into their homes, shared with me their lives, and many of whom I have

come to consider among my closest friends. Special thanks goes to Gilcilene Vale da Silva who

has walked with and encouraged me in recent journeys and, se Deus quiser, will continue to do

so for many journeys yet to come.









TABLE OF CONTENTS



A C K N O W L E D G M E N T S ..............................................................................................................4

L IS T O F T A B L E S ............................................................................................... ..................... 10

LIST OF FIGURES ............................................. .. ......... ............ ............... 12

A B S T R A C T .......................................................................................................... ..................... 13

CHAPTER

1 INTRODUCTION .................................... ........... ............................. 15

Roads, Governance and Land Use in the Brazilian Amazon.............................................15
Research Questions ........... .................... .. ........... .............................. 20
O organization of D issertation ... ........................................................................ ................ 2 1
R e se a rc h S ite ..........................................................................................................................2 2
P A D Q u ix a d ...................................................................................................................2 4
PAE Sta. Quiteria .......................................... ................ 27
RESEX Chico M endes- (Seringal Etelvir) .................................................................28
Fazenda Sta. Rita- (Seringal Etelvir) ................... ........................... ............... 29
T e rra S o lta .......................................................................................................................2 9
P o rto C a rlo s ..................................................................................................................... 3 0
M eth o d s ..................................................... ......................................... ......... ..........3 1
S ite S ele ctio n ........................................................................................ ........ ........... 3 1
Sampling .................................................... .................. 32
D ata C o lle ctio n ................................................................................................................3 3
D ata A analysis ................................................................................... ....................... 36
Limitations ............................................. .............................. 36

2 THE POLITICAL ECOLOGY OF ROADS, GOVERNANCE AND LAND USE IN
THE AMAZON .......................................................................... 42

Land Use and Land Cover Change in the Brazilian Amazon ............................... ...............43
D deforestation .............................................................. ......... ................... 44
Cattle Ranching and Pasture Formation.......................................................................45
Fire ........................................................ .................. 48
R o a d s ...................................................................................................... ......... . ....... 5 0
Public Policy ............................................ ................................. 52
P political E ecology ............................................................ ........................ ....................56
Governance and Institutional D esign ...........................................................................60
Political E cology in the D issertation ............................................................. ...............62
V ariab les U sed in A n aly sis.....................................................................................................6 5
D dependent V ariables ................................................................................................. 65
D deforestation ............................................................................................................65


6









C battle herd size ................................................................................................... 66
B razil nuts .............. ....................................................................... . .....66
Annuals ............................................ .................................. 66
L a st b u rn ...................................................................................................................6 6
Independent V ariables ............................................................................................... 67
P la c e o f o rig in ..........................................................................................................6 7
D instance from road ................................................................................................67
Bi-localism (male and female) ..................... .........................................67
A dult labor (m ale and fem ale)...............................................................................68
Y ears of residence ............................................................................................... 68
P ro p e rty siz e .............................................................................................................6 8
G overnance (adm inistrative unit)........................ ..................................................68
Participation in governance ................................................................................68
H ypotheses ................................................. ................ ..................69
R oads and L and U se ............................................................................ ...................... 69
G overnance and Land U se .........................................................................................70

3 ROADS AND LAND USE CHANGE IN ALTO ACRE.................................... ...............72

R oads and the "O occupation of the A m azon". .................................................... ...............73
Road Paving and Colonization in Alto Acre ......................... ....... .... ............... 75
Testing the Relationship between Place of Origin and Land Use..............................80
M odeling Place of Origin and Land Use......................................................................82
D isc u s sio n ............................................................. ...........................................................8 2
Intra-Regional M igration, Gender and Land Use ................................................................83
Testing the Relationship between Bi-localism, Out-Migration and Land Use ...............86
Female bi-localism and land use ..... ......................................................... 87
M ale bi-localism and land use ............................................................ ................ 87
A dult fem ales and land use ............................................... .............. ................ 88
A dult m ales and land use ............................................... ............. .. .. .. .......... .. ..88
Modeling Bi-localism, Household Composition and Land Use...................................89
D discussion ................................................................ .......... ....................90
A access to Roadw ay and Land U se......................................... .............. ............ ................ 92
Testing the Relationship between Road Access and Land Use ..................................... 96
D iscu ssion ....................................................... . .................. ................... 96
C battle an d L an d U se ...................................................................................... ................. ... 9 8
Testing the Relationship between Herd Size and Other Land Use Practices .............98
Modeling Herd Size and Brazil Nut Production, Area in Annuals, and Use of Fire
in P astu res ................................................................... 10 0
D iscu ssio n ............................................................................. ..... ........ ...........10 1
Conclusions ................................................. ................................102

4 GOVERNANCE AND LAND USE CHANGE IN THE INTER-OCEANIC
H IG H W A Y C O R R ID O R ......................................................................................... 109

R research Q questions ...................... ............. ...... ................... ................ 112
Land Rights and Environmental Conservation-the Birth of a Coalition.............................113


7









Legal and Institutional Structures of Environmental Governance in Alto Acre ................116
G overnance at the Federal L evel .............................................................. ............... 119
Governance at the State Level: the "Forest Government" .......................................... 121
M unicipalities.............. ....................................................................... . .... 124
C O N D IA C ............................................................................................ ..................... 12 5
A dm inistrative U nit ...................................................................... ........... ..... .. 125
Non Governmental Organizations (NGOs)............... .........................126
STR ................................................. .............................. 127
Associations ................................................ .......................... 129
Land Tenure and Forest Conservation....................................................... ............... 131
A sso ciation s an d S cale ............... ............................... .............................................. 134
A M O R E B ......................................................................................................................1 3 5
A M P A E S Q ....................................... .................................................. ..................... 13 6
AM PAE SQ and Local A ssociations.............................................................. 137
Representation and Effectiveness in Association Models.................. .................. 140
Associations and Scale- Conclusions ....... ......... ........ ...................... 141
G overnance and L and U se .............................................................. ......... ........ ............... 142
Testing the Relationship between Governance (Administrative Unit) and Land Use ..144
Modeling the Effect of Governance at the Administrative Unit Level Upon Land
U se ................................................ ................................................... ........ . ....... 1 4 5
Governance and deforestation...... ............. ............ ..................... 145
Governance and cattle herd size...... ......... ....... ..................... 146
Governance and years since burning pasture .................................... ................ 146
Testing the Relationship Between Participation in Governance and Land Use..........146
D discussion .................................................................................................... 148

5 RECONCILING POLITICAL AND ECONOMIC REALITIES IN ALTO ACRE'S
BR-317 CORRIDOR .......................................... .............................155

The Political Ecology of Land U se in Alto Acre............................................. ............... 155
The "Alto Acre Paradox"-Weighing the Costs and Benefits of Environmental
C o m p lian ce ............................................................................................... ......... .......... 15 8
The Alto Acre Paradox and Institutional Design.........................................166
Reconciling Centralized and Participatory Forms of Governance ..................................168
Implications of Findings for Forest and Agrarian Policy in Alto Acre ............................. 169
U rban and Rural Connections ................. ........................................................... 169
Environm ental Infraction Penalties ...... .............. ............ ..................... 170
C attle...................................................... .. ... ...... .........................17 1
ProAmbiente and Environmental Service Payments........................ .................. 172
Fire ......................................................... ..................... 172
D directions for Future R research .................................................................. ............... 173
Gender and Migration................................... .......... ........................ 173
L and U se G overnance ................................................. ............................................ 174

APPENDIX

A H O U SE H O L D SU R V E Y ................................................... ............................................ 176


8









L IST O F R E F E R E N C E S ....................................................... ................................................ 190

B IO G R A PH IC A L SK E T C H .................................................... ............................................. 199




















































9









LIST OF TABLES


Table page

1-1 Sam pling by A dm inistrative U nit ....................................... ....................... ................ 39

3-1 Descriptive Statistics for Independent Variables in Chapter 3 ..................................105

3-2 Descriptive Statistics for Dependent Variables in Chapter 3................ ...................105

3-3 Brazil Nut Production Regressed on Place of Origin (OLS Regression Coefficients).... 105

3-4 Correlations among Adult Males, Adult Females, Area in Annuals, Brazil nut
Production and Cattle H erd Size.................................... ....................... ............... 106

3-5 Herd Size and Brazil Nut Production Regressed on Female Bi-Localism (OLS
R egression C oefficients) .. ..................................................................... ............... 106

3-6 Brazil Nut Production and Cattle Herd Size Regressed on Adult Male Labor (OLS
R egression C oefficients) .. ..................................................................... ............... 106

3-7 Cattle Herd Size Regressed onAdult Female Labor (OLS Regression Coefficients)...... 107

3-8 Correlations among Distance to Road, Deforestation and Cattle Herd Size .................107

3-9 Correlations among Cattle and Brazil Nuts, Area in Annuals and Years Since
B u rn in g P a stu re ............................................................................................................... 10 7

3-10 Area in Annuals and Frequency of Pasture Burning Regressed on Herd Size (OLS
R egression C oefficients) .. ..................................................................... ............... 108

4-1 Administrative Units in Field Site: Range of Environmental Governance....................150

4-2 Descriptive Statistics for Bivariate and Continuous Measures of Participation in
G o v e rn a n c e ................................................................................................................... ... 1 5 1

4-3 Frequencies for Categorical Measures of Governance........................ .................. 152

4-4 Means Test for Effects of Governance (Administrative Unit) for Household
D e fo re station n ................................................................................................................. ... 1 5 2

4-5 Means Test for Effects of Governance (Administrative Unit) for Herd Size ................153

4-6 Means Test for Effects of Governance (Administrative Unit) for Pasture Burning ........ 153

4-7 Governance (Administrative Unit) and Land Use (OLS Regression Coefficients) .........153

4-8 Factor Weightings-Participation in Governance Index...................... .................. 154









4-9 Correlations among Participation in Governance, Household Deforestation, Cattle
Herd Size and Years since Burning Pasture .............. .......................... 154









LIST OF FIGURES

Figure page

1-1 Field Site in Context of A cre and B razil....................................................... ................ 40

1-2 R research Site w ith A dm inistrative U nits...................................................... ................ 41









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

ROADS, GOVERNANCE AND LAND USE IN THE BRAZILIAN STATE OF ACRE

By

Jeffrey Benson Luzar

December, 2006

Chair: Marianne Schmink
Major Department: Interdisciplinary Ecology

The dissertation examines the relationships among roads, governance and land use in the

Western Brazilian Amazon. Given the often negative social and environmental impacts of past

road construction projects in the Amazon, in recent decades, the Brazilian government and civil

society have come together to develop various forms of environmental governance, such as

extractive reserves and regulations on land clearing and burning, designed to minimize these

negative impacts. The dissertation draws on a case study in the Inter-Oceanic Highway corridor

in the state of Acre to examine both the impacts that roads have had for land use and the role that

governance has had in influencing these impacts.

The dissertation focuses on two forms of migration that have accompanied the opening and

subsequent paving of the road-in-migration and rural-urban migration. Ethnographic and

multivariate regression analyses revealed that many of the differences in terms of land uses that

once characterized agriculturally-oriented migrants and extractivism-oriented Acreanos have

disappeared after more than two decades of co-existence. An exception is the tendency of

Acreanos to more heavily exploit Brazil nuts than do migrants. More recently, the paving of the

road has led to the emergence of"bi-localism"-a tendency of households to invest their time

and resources simultaneously in both rural and urban settings. This trend disproportionately









affected women. Furthermore, households in which female household heads spent greater

amounts of time in town had lower Brazil nut production and households with fewer adult

women present had significantly smaller cattle herds. Qualitative analysis suggested that this last

tendency is related to gendered cattle production systems-in which herds own by women tend

to grow more quickly than those owned by men.

Findings provide an optimistic evaluation of the potential for governance to moderate

some of the most critical land use and land cover changes that have occurred in the road corridor.

Multivariate analysis showed that environmental governance, as represented by government-

mandated administrative units, was significantly related with lower levels of deforestation and

smaller cattle herds.










CHAPTER 1
INTRODUCTION

Roads, Governance and Land Use in the Brazilian Amazon

A central issue in discussions about sustainable development in Brazil involves the

implications of roads in the Amazon Basin. While road construction and paving carry the

promise to improve access to goods and services among previously isolated rural and urban

communities, they have also been shown to bring about negative environmental and social

consequences, including deforestation and the displacement of rural smallholders, under pressure

from in-migration and land-speculation. In academic and policy circles, a considerable debate

exists regarding the capacity of governance to prevent these potential repercussions from

highway paving (e.g., Nepstad et al., 2001, 2002; Laurance et al., 2001, Soares-Filho et al.,

2004). Relevant examples of governance include the implementation of a deforestation

monitoring program in Mato Grosso (Fearnside, 2003), ecological-economic zoning in Rond6nia

and Acre (Mahar, 2002, Governo do Acre, 2000) and, at smaller scales, the activities of rural

workers syndicates to promote sustainable land use among members. While much of the

existing literature uses landscape-level models to predict social and environmental changes

arising from road construction and governance, to date, relatively little detailed research has

focused upon these processes at the household and community level.

The dissertation addresses the relationship between roads, governance and land use at

household and community levels in the municipalities of Brasileia and Assis Brasil in the Alto

Acre region of Acre, Brazil. It draws upon a political-ecological framework to contextualize

household-level land use decisions within a multi-scaled political, economic and ecological

context. Through such a perspective, it becomes easier to interpret the complex and often









contradictory economic, ecological and political forces shaping land use decisions in Acre's

highway corridors.

Historically, the opening of roads in the Brazilian Amazon has been accompanied by large-

scale deforestation, often followed by even more massive deforestation events when these roads

are eventually paved (Hecht and Cockburn, 1990; Schmink and Wood, 1992; Fearnside, 2000b;

Laurance et al., 2002). In the Amazon region, while deforestation within 50 km buffers of

unpaved roads generally ranges from 0-9%, within similar buffers of paved roadways 29-58% of

the land area has been deforested (Nepstad, 2002). For instance, in the state of Rond6nia, the

paving of the BR-364 highway in the early 1980's led to a massive influx of settlers from outside

the region who brought with them livelihood systems that relied not upon the forest, as had the

existing indigenous and rubber tapper populations, but upon the removal of the forest and its

replacement with agricultural fields and pasture (Browder and Godfrey, 1997). While

deforestation in Acre's early road corridors in the 1970's and 1980's generally occurred on a

smaller scale than in neighboring Rond6nia, the presence of an organized resistance among the

rubber tapper community led to direct and widely-publicized conflicts between incoming groups

seeking to acquire and clear land and the rubber tappers who sought to both maintain land rights

and preserve the forest that provided a significant portion of their livelihoods.

The impact of roads upon land use change has become an especially salient issue given the

Avanga Brasil (Advance Brazil) program of the federal government. As of the mid-2000's, the

program had begun providing all-weather roads in regions previously served by seasonal

unpaved roads. Avanga Brasil consists of 338 projects in Amaz6nia with US $43 billion

allocated for the first eight years (Fearnside, 2002b). In all, some 6245 km of existing seasonal

roads have either been paved or are scheduled to be paved (Nepstad et al., 2001). As the amount









of land in the Brazilian Amazon within 50 km of paved highways will nearly double under the

Avanga Brasil program-from 16-28%-extrapolation from past trends shows that the project

could greatly accelerate deforestation in the Amazon Basin. In a socio-environmental evaluation

of the project, Nepstad and co-authors (2001) predict that, due to the project, 120,000-270,000

square kilometers of primary forest could be lost in the next 25-35 years.

In Acre, as in other parts of the Amazon Basin, both the greatest challenges and

opportunities for combing economic development with forest conservation will likely occur in

highway corridors (Kainer et al., 2003). On one hand, given increased accessibility, following

road paving, adjacent areas become subject to greater population movements and, due to the

access to new markets, land uses such as large-scale cattle ranching become more economically

feasible, leading to further deforestation. On the other hand, the increased access to urban

centers and markets may also prove to be beneficial in promoting forest-based economic

development. For example, in roadside areas such as PAE Cachoera and Porto Dias and in the

newly established Antimari state forest, all in the state of Acre, sustainable timber projects take

advantage of the improved access to urban markets. While improved transportation in Acre's

road corridors has facilitated out-migration to cities, it has also, by providing both better access

to urban services and easier transportation for urban-based teachers, extensionists and medical

care providers, improved living conditions in many rural communities (personal observation).

Ironically, despite the increased threats to the forest, the capacity of governance regimes

may actually be strongest in highway corridors. The director of the Brasileia office of the

National Institute for Colonization and Agrarian Reform (INCRA) attested that the paving of the

BR-317 highway had eased his office's task in working with local associations and enforcing

environmental regulation; whereas a visit to some of the more distant areas administered by the









agency (up to 110 kilometers each way) once could take several days, with the paving of the

road, it became possible for an INCRA team to make field visits and return within a matter of

hours (Regional Director, Brasileia Office of INCRA, personal communication).

In Acre, the arrival of roads has brought many benefits to local residents. It has eased the

isolation once felt by rural families-facilitating their access to urban-based services such as

schools, banks and hospitals. It also has brought increased access to markets and to outside

capital. With the notable exception of Brazil nut extraction, which has benefited from

improvements in road access, these changes have led to a general shift away from forest-based

land uses to other more lucrative and less labor intensive practices-especially cattle ranching-

giving rural families an opportunity to purchase goods and services that might otherwise remain

out of reach (Faminow, 1998; personal observation). For example, in the field site, numerous

families attributed their ability to build new homes, acquire solar panels and generators and

purchase motorcycles or other vehicles and pay off farm debt and emergency health bills directly

to profits from cattle ranching.

However, in addition to increased access to urban-based services and unprecedented

prosperity brought by the road, the people of the region have found themselves facing

unprecedented economic and environmental risks. Increased connectivity to regional and global

markets has also brought new economic risks, such as the sudden shifts in prices for the forest

and agricultural commodities upon which many families depend for monetary income-as

witnessed by recent price fluctuations for beef and Brazil nuts. Due to the often lucrative

nature of cattle production in the region, many farmers have chosen to clear forest and plant

pasture grasses. Due to the difficulty of growing crops in land that has previously been used as









pasture, increased allocation of land to pasture represents a decreased opportunity to allocate

land to other uses should beef prices fall.

Whereas deforestation under traditional shifting cultivation generally allowed the eventual

regeneration of forest cover, deforestation followed by pasture creation leads to the long-term

displacement and fragmentation of ecologically-important tropical forest. Compounding the

direct impacts of deforestation and pasture establishment, the associated use of fire in clearing

forest and maintaining pasture has brought further negative impacts for the region. During the

burning season, thick smoke has led to increases in traffic accidents and respiratory illnesses

(Mendonga, 2004), and during increasingly common dry years (such as 2005), uncontrolled fires

threaten not only the state's forests, but also devastate pastures, corrals, fences and livestock,

ironically threatening the cattle industry itself in more extreme instances. Furthermore,

deforestation and fire may combine to bring devastating climactic and hydrological impacts at

both regional and global levels. Within the Amazon Basin, rivers run dry during the

increasingly-intense dry seasons, game and timber become scarce and adverse meso-climate

change (including regional warming and drying) has shifted from a hypothetical proposition to

an acknowledged fact of life for many of the area's residents, as noted in numerous interviews.

And, at the global scale, changes in the Amazon's hydrology resulting from deforestation and

fire may in turn change rainfall patterns elsewhere in Brazil, North America and Eurasia (Maslin

et al., 2006).

Environmental governance, by shaping the trajectory of land use change in highway

corridors, has direct implications for both forest conservation and economic development both

within and beyond the highway corridors. In recent decades, governance in the Brazilian

Amazon has been oriented toward the often contradictory tasks of promoting economic









prosperity and avoiding environmental risk through mandated stewardship of the region's

forests. It is the complex relationship among these political, economic, and ecological contexts

in affecting household and community-level land use change that my research project seeks both

to describe and explain.

Research Questions

My research project addresses three broad questions related to the relationship between

roads, governance and land use in the Alto Acre region. The first question addresses the

relationship between the Inter-Oceanic Highway and land use.

As will be shown in the following chapter, researchers have pointed to the negative

impacts of increased migration and market integration in past road construction projects,

predicting that they are likely to continue in more recent road projects, such as the paving of the

BR-317. Regarding migration, as will be shown in Chapter 3, in the Alto Acre region, the BR-

317 has facilitated the migration of families from outside of the state into the region. It has also

led to intra-regional migration between rural and urban areas. The road has also provided

improved access to and the growth of urban markets, and in so doing has dramatically changed

the land use options available to the region's rural residents. Addressing these linkages between

the road and land use, I ask "how has the Inter-Oceanic Highway, by affecting migration and

market integration, in turn affected land use in the road corridor?"

This question of the road's impacts upon land use leads to a second question about the

capacity of governance to mediate these impacts. As will be shown in chapter 2, a considerable

debate has emerged regarding the capacity of governance to mediate the impacts that new roads

will have for land use in the Brazilian Amazon. While some researchers anticipate that past

patterns of predatory land use will continue in current road projects, others argue that the

development of legal mechanisms designed to protect Amazonia's natural resources and the rise









of an increasingly empowered civil society following Brazil's return to democracy in the mid-

1980's can avert much of the negative social and environmental consequences of past

infrastructure projects in the region, even bringing new opportunities for integrating economic

development and forest conservation. In order to address this debate in the context Acre's BR-

317 highway corridor, I ask "What role does governance play in influencing land use change in

the BR-317 corridor?"The final question addresses the differential impacts of two forms of

governance-centralized and participatory-upon land use in this region. In this region,

centralized governance, while based in a democratic system, relies heavily upon decisions by

governmental bureaucracies. The establishment and enforcement of legal limits on private

property deforestation would be one example. In the research site, this form of governance is

most visibly represented by the "administrative unit" (e.g., extractive reserve, colonization area,

etc.), each of which are differentiated by legal deforestation limits, by presence of governmental

agency and by the availability of environmental courses and projects. The other form of

governance is more participatory in nature and is most conspicuous at a household level.

Examples would include participation in associations, rural workers syndicates and in NGO-

sponsored environmental courses. In Alto Acre, this participatory form of governance stems

largely from the grassroots social movement that emerged in the 1970's and early 1980's to

protect land rights and livelihoods of forest-dependent peoples such as rubber tappers. To

address the differential impacts of these two forms of governance upon land use, I ask "what is

the relationship among political-economic context, various forms of governance, environmental

subjectivity and land use in Alto Acre?

Organization of Dissertation

In this chapter, I offer an introduction to the research questions, the research site and

methodology used in the collection and analysis of data. The following chapter outlines the









theoretical background, highlighting especially the possibilities for applying a political-

ecological perspective to land use change in the Amazon. Drawing upon this foundation, I

operationalize the various independent and dependent variables that are used in subsequent

chapters to test relationships among the road, governance and land use. Using these variables, I

outline the hypotheses that will be tested in chapters 3 and 4.

In chapter 3, I study the impacts of the Trans-Oceanic Highway for land use. In this

chapter, I first address the relationship among roads, migration and land use-focusing upon the

role that place of origin and rural-urban migration have for land use. Chapter 3 then addresses

the question of market integration, focusing specifically upon the impacts of road access and the

expanding cattle economy, for land use. Chapter 4 addresses the impacts of governance for land

use in the region. The first portion of the chapter discusses the structure of governance in the

region, including the intersections among land tenure, institutional design and land use. The

latter portion of the chapter addresses the implications of centralized and participatory

governance for land use.

Chapter 5 applies an explicitly political-ecological framework to consider the findings in

chapters 3 and 4. Special attention is given to the ways in which different forms of governance

affect (or do not affect) land use and the importance of the regional political-economy in

mediating this relationship. Drawing upon this analysis, the chapter concludes with several

policy recommendations for reconciling environmental governance with local economic realities.

Research Site

The research was conducted in the Alto Acre region of the state of Acre in the Western

Brazilian Amazon (Figure 1-1). The Alto Acre region is so named due to its location in the

upper regions of the Acre River watershed and is comprised of four municipalities-Assis Brasil,

Brasileia, Epitaciolandia and Xapuri. It is home to some 55,000 people (IBGE, 2004), who,









while rarely differentiating themselves along ethnic lines, tend to come from a combination of

European, African and Amer-Indian ethnic backgrounds, with non-Indians tending to have

arrived in the region either by river from Northeastern Brazil in the late 1800's and early 1900's

to tap rubber or followed roads from South and Central Brazil in the mid to late 20th century.

The population is roughly half rural and half urban. The entire urban population is concentrated

in the towns of Assis Brasil, Brasileia, Epitaciolandia and Xapuri, capitals of their respective

municipalities and the only urban areas in the region.

While the service sector, light industry, and, increasingly, the public sector, employ many

people in rural areas, agricultural and forest production form the backbone of the private

economy, especially in rural areas, but also among urbanites who often have land and/or cattle in

surrounding rural areas. With the exception of some specialized professions, most workers earn

the national minimum wage-which at the time of research was $300 reais (US$140) per month.

Predominant land cover includes open semi-deciduous forests, which consist of a broken

canopy interspersed with palms and bamboo, closed semi-deciduous forest, which has a denser

canopy and more open understory, anthropogenic grasslands (cattle pasture) and agricultural

plots, and abandoned agricultural lands in various states of succession (Governo do Acre, 2000).

In the study region, soil types are dominated by Red and Yellow Agrissols; while these soils tend

to have high aluminum content, it rarely reaches the point of aluminum toxicity (Governo do

Estado do Acre, 2000). In my research area, informants commonly offered a perception of the

area's soils and climate as being apt for agriculture.

The study lies on the border of Brasileia and Assis Brasil, measuring some 50 kilometers

east to west and 30 kilometers north to south and is bisected east to west by the Inter-Oceanic

Highway (Figure 1-2). The majority of the research site lies in the municipality of Brasileia-









with the westernmost portion lying in Assis Brasil. As of 2004, the 4,852 individuals in Assis

Brasil and 16,940 living in Brasileia were roughly divided between urban and rural residence

(IBGE, 2004). The site is bounded by the Rio Acre and Bolivia to the south and the Rio Xapuri

and the RESEX Chico Mendes to the north.

The study area formerly was comprised of four adjacent seringais (plural of seringal or

rubber estate)-Sdo Pedro, Sacado, Etelvir and Porto Carlos, each of which was bisected by the

roadway. Each seringal came to a different fate after the coming of the road in the late 1970's,

the subsequent violent land disputes between ranch workers and rubber tappers in the late 1970's

and early 1980's, and the subsequent implementation of agrarian reform and extractive reserves

in the area.

As of the time of research in the mid-2000's, the four original seringais were represented

by six contiguous administrative units: 1) Quixada, a Colonization Project (Projeto de

Assentamento Dirigido or PAD), 2) Santa Quiteria, an Extractive Settlement Project (Projeto de

Assentamento Extractivista or PAE), 3) RESEX Chico Mendes (the portion lying within Seringal

Etelvir), 4) the privately-ownedfazenda (ranch) Santa Rita, 5) the largely un-administrated Terra

Solta and 6) Seringal Porto Carlos. The administrative units differed by land tenure system,

government agency responsible (if any), legal deforestation limit and the presence of government

and NGO-sponsored courses and projects related to forest conservation. These differences and

their implications for land use will be analyzed in depth in chapter 4.

PAD Quixada

Following the axis of the Inter-Oceanic Highway lays the colonization area officially

known as Projeto de Assentamento Dirigida Quixada, Gleba 6, or what is locally known more

simply as "Quixada." PAD Quixada is one of the numerous colonization areas the federal

government has implemented in the Brazilian Amazon. In Acre alone, colonization areas occupy









some 1.4 million hectares (Governo do Acre, 2000). Historically, these areas have been

administered by INCRA and divided into rectangular family farms or "lotes. In Acre, lotes

ranging from 60 to 100 hectares have been given to current occupants and to landless families

from other parts of the country-all of whom are commonly referred to as colonos (Governo do

Acre, 2000). These areas have generally been designed in distant government offices, with little

consideration for on-the-ground terrain, agricultural suitability and water availability.

Furthermore, due to the small size and the fact that few lotes coincided with pre-existing rubber

trails, traditional forms of forest extractivism proved unviable, even for occupants with traditions

of forest extractivism (Governo do Acre, 2000). In more recent years, due to a decrease in

available land and increased emphasis upon forest conservation, in the state of Acre, the

formation of new PADs has slowed considerably-and the size of newly-demarcated lotes has

decreased considerably as well1.

The history of PAD Quixada lies in the violent confrontations between the area's

traditional rubber tapping populations and newly-arrived ranch workers2. In an effort to stem the

growing violent land conflict in the region that had led to the assassination of the president of

Brasileia's Rural Worker's Syndicate, Wilson Pinheiro, and a retaliatory lynching of the ranch

manager widely believed to have been responsible, INCRA expropriated several ranches located

in this area to form the Projeto de Assentamento Dirigido (PAD) Quixada. The initial project

was implemented in 1982, with the westernmost section-Gleba 6 (the portion of the PAD lying


1 In Quixadd-Gleba 6, during 2005, some 11 families were settled onto a 300 hectare area that had not been
developed by INCRA when the project was formed in the early 1980's. Landholdings were considerably smaller
than past lotes-with most families receiving slightly less than 30 hectares of land.
2 By "ranch workers," I refer to both the ranch managers (gerente) and ranch hands (peoes) that directly operated the
ranches in the region. The word "rancher" (fazendeiro) is more properly applied to the ranch owners-individuals
who, due to the fact that they have often maintained residency in other parts of the state or country, rarely were
directly involved with the day-to-day operations and with personal confrontations with rubber tappers during this
era.









in the research site) -implemented shortly thereafter in 1983. In PAD Quixada, Gleba 6, lots

ranging from 70-100 hectares in size were demarcated in areas of Seringais Sao Pedro and

Sacado formerly occupied (though not necessarily cleared) by the ranch that had been managed

by Nilo. The area stretched along the edge of the highway and, in some areas, along access

roads leading further to the north and south. First priority was given to the area's current

(Acreano) inhabitants-most of whom received the lot in which their homes happened to be

located. Unoccupied lots were given to landless families who were brought by INCRA from

Parana and other states in the south in 1983.

Many of the area's current residents had received their lots from INCRA in 1983. Many

other current inhabitants represent later migrants or children of the first recipients who purchased

lots from prior inhabitants who moved, most commonly to urban centers in the region. As a

colonization area, among the administrative units included in the study, Quixada contained the

highest concentration of residents born outside of Acre, with 72% of sampled households in the

PAD having at least one head born outside of Acre or neighboring areas of Peru and Bolivia.

INCRA carried official responsibility for overseeing the colonization project. However,

some families had received full title to their lots, and the agency intended to fully privatize the

project in the near future. Livelihoods were dominated by subsistence annual and perennial

agriculture as well as some market-oriented agriculture-most notably cattle and, to a lesser

degree, Brazil nuts. Timber-mostly from land destined to be converted to agriculture-served

as an occasional source of supplemental income for many families. Deforestation was very

advanced in this area and nearly all lots had far exceeded the 20% deforestation level permitted

by law. According to estimates from the director of INCRA's regional office in Brasileia, some

60%-70% of the project area had been deforested as of 2003. No over-arching local governance









structure existed in the region. Rather, most residents belong to one of eight local associations,

each of which is autonomous from the other.

PAE Sta. Quiteria

Covering approximately 44,205 hectares, and home to at least 277 families, PAE Santa

Quiteria was created in 1988, the second PAE to be created in Acre (Governo do Acre, 2000).

Under INCRA's original plans, the entire area was to be divided into lots and incorporated into

the PAD. However, the implementation of lots was halted as various stakeholders, including

local residents and the National Council of Rubber Tappers (CNS), recognizing that traditional

extractivist livelihoods were impossible on 70-100 ha lots, lobbied INCRA to implement a

different system of agrarian reform that would recognize local families' traditional rights to

multiple rubber tapping trails (ranging from three to eight trails for most families, with each trail

being considered roughly equivalent to 100 hectares) and would give land to its original rubber

tapping occupants and not to landless agriculturalist families from outside of the region.

The PAEs were established in an attempt to implement agrarian reform while respecting

traditional colocagdo land tenure systems and forest-based livelihoods. As such, the PAEs

served as an early experiment in the integration of agrarian reform and forest conservation. In

these precursors of extractive reserves, such as Santa Quiteria, INCRA maintained official

ownership of the area extractivee reserves, such as the globally-renowned Chico Mendes

Extractive Reserve, would be administered by IBAMA, the federal agency charged with

environmental protection, without a role for INCRA, an agency traditionally charged with

agrarian reform issues).

PAEs were developed as an alternative to the PAD modality, an alternative in which

traditional rubber tapper tenure systems would be maintained. While the distribution of lots in

the PAD recognized rubber tapper rights to the land they occupied, they recognized neither









traditional livelihoods nor land tenure structures. As in other extractive reserves, land in the

PAE is owned by the Brazilian state, and families have use rights to large areas (generally 200-

300 ha) that are defined not by firm boundaries, but by the rubber-tapping trails within them

(generally three to four per family). Local governance systems tend to focus upon local

associations and the Association of the Residents of PAE Sta. Quiteria (AMPAESQ)-a

centrally-located association intended to serve as an umbrella organization for local associations

and to represent the project's residents in external political concerns.

RESEX Chico Mendes-(Seringal Etelvir)

Seringal Etelvir was sub-divided into three tracts and sold to ranching interests in the early

1980's, but the owner of the northwestern tract, as of the late 1980's, had still not expropriated

local families nor cleared the forest for pasture, and unlike the more accessible southern segment

of the former seringal, it was never expropriated for colonization by INCRA. As a result, this

approximately 8000 hectare area was still largely forested and lightly inhabited when the Chico

Mendes Reserve was created in 1990. Following its appropriation by IBAMA and incorporation

into the reserve, this seringal fragment stands at the southern fringe of the RESEX, bordering

PAD Quixada and lying a mere four kilometers from the highway at its southernmost point. The

area is arguably more directly impacted by road construction and market access than most other

seringais within the RESEX. Indeed, evidenced in Figure 1-2, Seringal Etelvir is not only a

bulwark of the RESEX facing southward toward the highway, it is also a major exit point and

first or final resting point for the residents of at least five seringais located further inside the

reserve.

As in PAE Sta. Quiteria, land tenure in the RESEX is based upon the relatively large

colocado property unit, and properties are subject to a 10% ceiling on legal deforestation (as in

the PAE but in contrast to the 20% in all other parts of the research site). Like Sta. Quiteria, a









two-tiered association structure exists, with local associations (two exist in this area) falling

under the umbrella of the Association of Reserve Residents of Brasileia (AMOREB) situated in

Brasileia.

Fazenda Sta. Rita-(Seringal Etelvir)

The northeastern segment of Seringal Etelvir was purchased to create three adjacent

ranches, the largest of which is Santa Rita. The ranch's former owner, locally known as "0

Alemdo" ("The German"), until his death in a traffic accident in the late 1980s enjoyed relatively

harmonious relations with his rubber tapper neighbors compared to neighboring ranchers and

ranch workers who often killed and at times were killed by rubber tappers in land disputes. The

Alemdo, on the other hand, differentiated his ranch by purchasing land from not only the estate

owner, but from the resident rubber tappers who lacked formal title to the land. This practice

lowered animosity among displaced rubber tappers and is generally believed by locals to explain

the continued existence of Santa Rita while so many other ranches were expropriated by INCRA

as it attempted to stamp out land conflict in the region. While representing nearly half of the

deforested area in the research site, the privately-owned Sta. Rita and its two ranch neighbors

have played a relatively minimal role in more recent deforestation. Unlike other areas where

deforestation rates have steadily increased, the deforestation of Sta. Rita largely occurred in two

massive phases, first in the early 1980s and again in 1988, just as Amazonian deforestation was

catching the attention of the world and the Brazilian government.

Terra Solta

The area alternatively known as the Cinturdo Verde ("Green Belt") or the Terra Solta

("Loose Land") received its names from the facts that it forms a two kilometer largely-forested

"belt" extending some six kilometers between the fazenda and RESEX Chico Mendes and, as

land that was left out of the RESEX due to a surveying error, neither of the two major federal-









level agencies involved in land use policy-INCRA and IBAMA-has responsibility for the

area. Most of the residents were either Acreanos or the grown children of Quixada's aging

original settlers. Although there was a general property boundary understanding between most

(though not all) residents, no formal land titling system existed, documentation of residence

through Rural Workers Syndicate being the only documentation possible. Nor did residents have

a common association. Rather, most residents belonged to one of three local associations located

outside the Terra Solta while several others remained unassociated.

Porto Carlos

The approximately 100 families of Porto Carlos lived in what was, as of the time of

fieldwork, one of the last existing seringais in Alto Acre, and the only one lying along the BR-

317 highway corridor. Traditionally, seringais were owned by a rubber baron, with the rubber-

tapper residents having some use rights, but no legally recognized rights to tenure of the land on

which they lived. While Porto Carlos had long ceased to be a functional seringal, the owners,

while still managing a few prime areas for Brazil nut and cattle ranching, had allowed most of

the seringal to fall under the de facto tenure of its ex-rubber tapper residents. Residents were still

subject to federal environmental law (for instance, all landowners were subject to a 30 hectare

deforestation limit, calculated as 20% of the 150 hectares officially held in each property3). And

while kinship ties were strong in nearly all the communities, Porto Carlos demonstrated a

remarkable level of kin-based social capital, with nearly all residents belonging to one of two

families (which traditionally tended to inter-marry). Compared to the other more heterogeneous

and divided communities, Porto Carlos tended more often than not to speak with a (near)



' While all properties were recently documented as 150 hectares, properties could be up to 600 or more hectares.
This has encouraged some of the larger owners to settle family members on their property so as to maintain de facto
ownership of their original colocag5es.









common voice in matters of local and regional politics. While not all residents participated, a

single association existed representing the peoples of Porto Carlos (Association Porto Carlos,

located near the geographical center of the seringal).

From both a theoretical and policy perspective, as no formal governmental decision had

been issued regarding the future of the area, Porto Carlos was an especially interesting area in the

research site. Faced with the mixed successes and failures of other land tenure systems, both the

state government and INCRA-which was gradually assuming responsibility for the area-were

challenged with the task of implementing a system that maintains the area's wealth in forest

resources while appeasing both the region's landless and the residents of this politically

important seringal. Hence, according to local INCRA officials in Brasileia, it appeared likely

that parts of the area would be sub-divided for agrarian reform while other areas would be

destined for a Sustainable Development Project (Projeto de Desenvolvimento Sustentdvel or

PDS) managed by INCRA. However, as of the time of research, neither the proportion of the

area to be opened to colonization nor the exact nature of the PDS to be created were yet known.

Methods

Site Selection

An original research site consisting of PAE Sta. Quiteria and PAD Quixada, Gleba 6, was

selected through analysis of maps and in discussion with key informants knowledgeable about

the region during an initial visit in 2002. A key factor in selecting the site was the proximate

location of these two land tenure units to the final stretch of BR-317 which was to be completed

later that year. During subsequent visits, the other sites were added, as I felt that they offered

interesting comparisons with the original two sites and also contributed to a more thorough

understanding of the variation in the drivers and processes of land use change and governance in

this area. Seringal Porto Carlos represented a land tenure system dating from the rubber era, yet









was in a state of transition as INCRA was considering the implementation of an extractive

reserve and/or colonization area within it. Also, due to a growing awareness of the ecological

importance of the area as one of the few areas in Alto Acre in which a heavily-forested area

bordered the road, NGOs and the government had recently begun offering courses and projects

in the area. The Terra Solta represented a control scenario, in which no formal land tenure

system had been implemented and no environmental courses or projects had been offered.

Finally, due to both the lack of permanent inhabitants and its relatively static land use system,

Fazenda Rita was not included in quantitative analysis; however it was included in descriptive

portions of the dissertation, due to its significance in the history of the area.

Sampling

Strictly random sample selection proved impractical due to local social realities and census

data limitations in the region4. A total of 92 households were selected (aside from Santa Rita

which has no permanent inhabitants and where only the ranch manager and several part time

workers were interviewed), 41 (44.5%) from PAE Santa Quiteria, 26 (28%) from PAD Quixada,

15 (16%) from Seringal Porto Carlos, 6 (6.5%) from RESEX Chico Mendes, Seringal Etelvir,

and 4 (4.5%) from the Terra Solta. Using a stratified sampling strategy, for each administrative

unit, I attempted to include 20% of the total population parameter of households in the sample5

(see Table 1-1). Households were selected using a stratified snowball strategy that relied both on

4 Due to their remoteness, it sometimes proved impractical to make return visits to households where no one was
home during the initial visit. Additionally, due to informal land sales and exchanges, in most areas, it was not
possible to obtain a fully accurate list of residents. Furthermore, early attempts to follow a strictly random design
drew suspicion from some potential informants-who learned that they had been chosen as respondents due to their
location on a list provided by a government agency. While potentially less statistically significant, as a foreign
researcher, I found that a stratified snowball approach allowed me to balance the goals of obtaining a representative
sample and operating within local social norms (that is, asking to conduct interviews after being introduced to an
individual by a prior acquaintance / informant in a non-interview setting).
5 In most cases, due to time and logistical constraints, samples more closely approximated 15%. Exceptions were
the Terra Solta and RESEX Chico Mendes where, in an effort to compensate for small populations, I sampled larger
proportions of the total population.









satellite imagery and past informants. I used satellite imagery of the area to ensure even spatial

distribution, and that households with large, medium and small-scale land clearing would be

represented. In most cases, I relied upon past informants to introduce me to potential new

informants-thus reducing the anxiety and distrust that might otherwise be associated with an

interview by an unknown researcher. I also relied upon information from early informants to

maximize the representativeness of sampled households in each community, such as including

both migrants and non-migrants (if present) and members of various different extended families

in a community.

Data Collection

Data collection occurred during May-June, 2002, May-June of 2003 and again from June,

2004 until July, 2005. A subsequent visit during May-August 2006 provided an opportunity to

obtain information about changes during the preceding year (e.g., wildfires in various

communities in the research site and a state-wide ban on pasture burning).

During 2002 and again in 2003, I met with potential research collaborators and also with

local community leaders, in both Brasileia and the actual field site, to discuss the proposed

research project, both explaining my research and seeking feedback regarding its proposed

content, form and potential relevance to the community. In the 2004-2005 field visit, while

many informants were already somewhat aware of my research project, prior to interviews, the

project's content, direct relevance and limitations were explained. Potential interview subjects

were also informed about the optional nature of participating in the study.

Semi-structured interviews with heads of sampled households represented a central

research method. Interviews were conducted with household heads, most commonly male

though sometimes female. Interviews ranged from 30 minutes to three hours. I applied a semi-

structured questionnaire (Appendix A). The questionnaire addressed questions of household









demography, participation in governance, land uses and environmental attitudes. The

information from the questionnaire was written in a notebook and recorded using a handheld

audio recording device. Semi-structured and unstructured interviews with other household

members (spouses, elderly parents, etc) gave additional insights. Accompanying on-farm

interviews, when possible (approximately half of all interviews), I conducted farm-transects

accompanied by the landowner. These walks provided an opportunity to assess the accuracy of

interview data and gain additional input from farmers that did not always arise within the more

formal context of a sit-down interview.

In addition to the household sample, key informants in the research area provided further

data relevant to the research question. I conducted key informant interviews with governmental

leaders at various levels, including the presidents of 11 local associations, the president of

AMPAESQ, the vice-president of AMOREB, the president of the STR, employees of SEATER

(Secretary of Technical Assistance and Rural Extension), various NGO representatives, members

of the directory of the Center for Small Rural Farmer Associations of Epitacioldndia and

Brasileia (Central de Associa(5es de Pequenos Produtores Rurais de Epitacioldndia e Brasilkia

or CAPEB), Brasileia's Secretary of the Environment and representatives of INCRA, IMAC and

IBAMA in Assis Brasil, Brasileia and Rio Branco. These interviews were complemented by

informal conversations on other occasions with many of the same individuals as well as members

of the local media and the prefeitos of Brasileia and Assis Brasil.

Finally, in order to develop a deeper ethnographic understanding of governance and its

articulation in this region, I relied heavily upon participant observation. For instance, in both the

field site and Brasileia, I frequently participated in association meetings and assemblies as well

as meetings and courses organized by IBAMA, STR and other NGOs. In addition to offering an









opportunity to witness the on-the-ground articulation of governance, they provided a forum for

me to explain and answer questions about my research project to members of the community.

Participant observation in day-to-day farm activities such as hunting, fishing, collection of

cupuagu fruits and cattle vaccination gave me a better understanding of the material and

symbolic significance of productive activities based on both forest and non-forest land covers.

In order to create a spatially-linked database, GPS points were taken for all landholdings

included in the study. Points were collected for important features such as homes, associations,

road intersections and farm and administrative boundaries. Satellite-based maps proved useful

for several purposes. They assisted me in the selection of spatially distributed farms with low,

medium and high levels of deforestation. Prior to entering the field, a land cover map of the area

was created using recent LandSat 7 ETM imagery in Erdas Imagine. Administrative and (in the

case of PAD Quixada) property boundaries were obtained through the INCRA office in Rio

Branco and overlaid onto the image in order to create a field map. Using a handheld GPS unit, I

obtained UTM coordinates for houses, fields and property boundaries which could then be

correlated with their location in the imagery.

I conducted participatory map interpretation with those informants who were both able and

interested in doing so6. This served several purposes-it proved a useful tool in aiding

informant's memories about past land use and in offering comments about land use on

neighboring farms. This often led to rich discussions by local residents including, for instance,

associating various family histories to specific areas, and thus allowing me to "link people to





6 Due to the difficulty that some (mostly older) informants faced in understanding and interpreting the map, this
technique was neither applied systematically nor used in quantitative analysis.









pixels" 7 (Liverman, et. al., 1998). It also allowed for an analysis of social networks-with many

informants correctly associating most or all of the satellite-detected land clearing plots in the

community with their owners.

Data Analysis

In the process of data analysis, I first conducted descriptive statistics to gain an

understanding of the means and variances inherent in the various independent and dependent

variables. Subsequently, I conducted bivariate analysis of individual independent and dependent

variables8. Bivariate analysis allowed me to test for both linearity and significance of

relationships between independent and dependent variables. Given the fairly small size of the

sample of households, significance was established as p<.1. Relationships that were both linear

(or curvilinear) and significant were then subjected to multiple regression analysis in order to

control for possible effects of covariates. An exception to this rule was made when bivariate

analysis revealed non-significant linear relationships and a suppression effect from covariates

was suspected. In these cases, multiple regression analyses were conducted to control for

suppression effect from covariates, even when results from initial bivariate analyses were not

significant.

Limitations

Several data limitations should be noted as caveats in the data analysis that appears later in

the dissertation. As previously noted, due to on-the-ground limitations, a strictly random sample

design was not followed. Hence, caution is due when extrapolating these results beyond the


7 Examples included, for example, a woman who told me that an area of her property was light green because she
and her husband had lost their cattle herd at the time the imagery was taken, causing the pasture to grow up.
Another family explained that a small forest fragment on their property visible on the map was an area of forest they
had left, upon a parent's suggestion, in order to preserve an exceptionally strong spring that lay within it.

8 The specific bivariate test employed depending on the nature of the independent and dependent variables (e.g.,
continuous versus categorical). Bivariate analysis included Pearson correlations, t-tests and F tests.









sample. However, I have attempted, to the greatest degree possible, to minimize potential bias in

the sample by including households from the various communities in the field site. Within each

community, I also attempted to minimize bias in the sample by including households from

various extended families and with small, medium and large amounts of forest clearing. While

diverse, the sample does not necessarily accurately represent the proportion of different sub-

groups in the population.

A second caveat applies to analyses of the impacts of governance, at the level of the

administrative unit, for land use. Given both the small population parameters and small samples

from two of the administrative units-Terra Solta and RESEX Chico Mendes-samples cannot

be assumed to be representative of the entire administrative unit. In the most extreme case of

Terra Solta (4 households sampled from a population parameter of approximately 14), I

ultimately elected to remove this administrative unit when testing the relationship between

governance at the level of the administrative unit and land use.

A third limitation comes from the possibility of inaccuracies, both deliberate and

accidental, in reported data from informants. This was especially of concern when obtaining

data about household deforestation. Due to the fact that many households had exceeded their

legal deforestation limits and were continuing to clear land, this variable was especially

susceptible to under-reporting. In order to minimize the implications of this possibility in data

collection, prior to arriving at a given household, I developed an approximate estimate of

deforested area on the property using both visual analysis of satellite imagery and personal

inspection of the pastures that usually surrounded rural homes. When reported estimates of

deforestation differed greatly from my own, I would attempt to triangulate deforestation

estimates with those given by neighbors to determine the source of error. As a last resort, in the










few cases where I was unable to reconcile reported figures with my own estimates, deforestation

values were recorded as missing.9
















































9 An example included an elderly divorcee who estimated that she had some 10 hectares of pasture, when both
satellite imagery and my own visual calculation were closer to 40 hectares. As the sons who cared for her cattle
herd were not present during my initial and subsequent visits, I coded the area of pasture and of total deforestation as
"missing" in this case.









Table 1-1. Sampling by Administrative Unit
Total
Households Households
Sampled (approximate)


% of
Households
Sampled


Terra Solta 4 14 28%
Seringal Porto
Carlos 15 100 15%
PAD Quixada 26 150 17%
PAE Santa
Quiteria 41 277 15%
RESEX Chico
Mendes 6 17 35%









heasonaW Grad
IN.

spirit San.o
Fge d. JFS C t c B
~Catarina






FC


Figure 1-1 Field Site in Context of Acre and Brazil






























Administrative Units: 1=PAE Santa Quiteria 2=PAD Quixada 3=RESEX Chico Mendes
4=Seringal Porto Carlos 5=Terra Solta 6=Fazenda Santa Rita
Green=Forest (Primary and Secondary); Red= Deforested (Pasture and Crops)
Figure 1-2 Research Site with Administrative Units









CHAPTER 2
THE POLITICAL ECOLOGY OF ROADS, GOVERNANCE AND LAND USE IN THE
AMAZON

A large body of literature addresses the issue of land use and land cover change, with a

significant portion focusing specifically upon the Amazon Basin. While some broad treatments

and meta-analyses of the topic of LULCC exist (e.g. Geist and Lambin, 2001), due to the breadth

of the topic, most research has focused on specific manifestations (e.g. deforestation) and drivers

(e.g. agricultural expansion, public policy) of LULCC. However, due to the diverse disciplinary

backgrounds of LULCC researchers, while research in this subject area has given great insights

into the various factors driving LULCC, much less work has been done to advance theoretical

understandings of LULCC. In this study, I draw upon political-ecology to both focus upon

certain phenomena and processes occurring in the Inter-Oceanic corridor and to understand the

relations among them.

In this chapter, I begin by offering a discussion of the LULCC literature as it relates to the

topic of roads and governance in the Brazilian Amazon. I first discuss the major land cover

change affecting the study site-deforestation. I then discuss two land use practices that are

intimately tied with deforestation in many portions of the Brazilian Amazon-cattle ranching

and the use of fire. I then address two drivers of LULCC that are of special concern to the

research topic-roads and public policy. Subsequently, I discuss the theoretical background of

political ecology, especially as it relates to the issues of governance and scale. Then, drawing

upon the LULCC and political ecology literature, I outline and operationalize the independent

and dependent variables used in the dissertation as well as the hypothesized links existing among

them.









Land Use and Land Cover Change in the Brazilian Amazon

Considered broadly, the over-arching dependent variable considered in the study is land

use and land cover change (LULCC). In recent years, the subject of land use and land cover

change has gained recognition as an emerging research agenda of profound theoretical and

policy implications. While theoretically diverse, much of the LULCC literature is united in

striving to transcend simplified narratives of causality such as population growth and shifting

agriculture, looking instead to develop a more nuanced understanding of the causes of LULCC

(Angelson and Kaimonwitz, 1999; Geist and Lambin, 2001; 2002; Schmink and Wood, 1992).

Countering meta-theories which attempt to offer broad explanations of deforestation, researchers

have generally argued that tropical deforestation is determined by different combinations of

proximate causes and underlying driving forces, varying between geographical and historical

contexts (Geist and Lambin, 2002).

Land use and land cover change is becoming an increasingly important interdisciplinary

field of study. Growing human populations and technological change are associated with rapid

changes in global land use and land cover-with anthropogenic land covers becoming

increasingly common. In the early 21st century, crop and pasture land has come to represent a

dominant land cover globally-rivaling forests-extending to approximately 40% of the Earth's

terrestrial surface (Foley et al., 2005). Land use change often has direct impacts upon land

cover. For example, the shift among inhabitants of the rural Brazilian Amazon from extractivism

toward cattle ranching is associated with a shift from forest to grassland (pasture) land cover.

Furthermore, land use and land use change can have more indirect impacts upon ecosystems,

including for example, changes in water cycles, evaporation and run-off (Foley et al., 2005).









Deforestation

Deforestation, at the household level, is a key dependent variable considered in the study.

A large amount of literature has been written about deforestation in the Brazilian Amazon,

including both its political-economic and ecological implications (e.g., Schmink and Wood,

1992, Wood and Porro, 2002). Paleoecological and paleoclimatic records suggest that the

Amazonian forest ecosystem has been in continuous existence for more than 55 million years.

However, despite the Amazon's resilience to the various climatic and geological changes

occurring during its existence, the continued resilience of the Amazonian forest is uncertain in

light of recent anthropogenic changes, including, most conspicuously, deforestation (Maslin et

al., 2006). During the twentieth century, approximately 16% of the Amazon forest area was

cleared, primarily for cattle pasture and, more recently, for the cultivation of soybeans

(Fearnside, 2001a; Maslin et al., 2006). Estimates of recent rates of deforestation range between

approximately .38% and .5% of existing forest per year (Maslin et al., 2006; Houghton et al.,

2000; Davidson and Artaxo, 2004). Under the current "business as usual" scenario (without

increases in governance), researchers have argued that, by 2050, nearly half of the Amazon's

original forest cover may be lost (Soares-Filho et al., 2006).

Much has been written about the actual and likely future impacts of deforestation at

various scales. Climate change is an implication of tropical deforestation that is likely to be felt

at local, regional and global levels. As the Amazonian forest plays an important role in both

moderating temperatures and in recycling water into the atmosphere during the dry season, on

local and regional levels, large-scale deforestation may result in warmer and drier conditions in

the region (Foley et al., 2005). This, in turn, would increase the region's susceptibility to

wildfire-further endangering the forest.









Research has also addressed the risks that carbon released through deforestation and

associated activities (such as fire) may have for regional and global climate. Research suggests

that the Amazon forest, in an undisturbed state, functions as a net sink of atmospheric carbon

(Carvalho et al., 2004). However, under current land use practices-including logging, pasture

burning and deforestation, Amazonia has become a globally significant source of atmospheric

carbon. Amazonian deforestation and associated burning releases some 2-4% of annual global

carbon emissions (200-300 million tons), representing more than 2/3 of Brazil's total (Fearnside,

1997; Houghton et al., 2000; Carvalho et al., 2004).

Cattle Ranching and Pasture Formation

Cattle production represents an important dependent variable considered in the study. In

the Brazilian Amazon, as in many parts of Latin America, cattle ranching has been the most

common form of agricultural expansion into previously forested areas (Geist and Lambin, 2002).

Brazil's cattle herd is the world's largest (Kirby et al., 2006). While the majority of Brazil's

cattle production occurs further south, production in Amazonia has been increasing in recent

decades. Historically, the spread of cattle ranching in the Brazilian Amazon has been accelerated

through governmental policies (Fearnside, 2000; 2002a). Seeing large-scale ranching as a

productive use of the Amazon Basin, the Brazilian government until recently offered heavy

subsidies for this industry; between the years of 1971 and 1987, the cattle industry received some

$5 billion from the Brazilian government (Hall, 2000). After Brazil's return to democracy, most

subsidies for cattle ranching in the Amazon were removed.

Due to both the scale and the relatively permanent nature of most cattle pastures (as

opposed to swidden plots which, if not planted in pasture, are usually allowed to return to forest),

cattle ranching represents the single most important proximate driver behind deforestation in the

Brazilian Amazon (Hecht, 1985; Fearnside, 1990; Cardoso, 2002). Various authors have argued









that cattle ranching, in the absence of governmental supports, is generally not profitable in the

Amazon (Hecht and Cockburn; 1990; Schwartzman, 1992; Fujisaka et al., 1996). Due to mineral

deficiencies in most Amazonian soils, it has been argued that pasture maintenance requires

chemical treatments to maintain productivity-including phosphates-a substance with few

known deposits in the Amazon region (Fearnside, 2002a). Furthermore, due to the difficulty of

preventing natural succession and weed development, in situations of abundant inexpensive land

(a scenario that no longer applied in much of the Amazon and most of Alto Acre as of the mid-

00's), it has sometimes proven less expensive to simply clear new forest tracts for pasture than to

rehabilitate older degraded ones (Fearnside, 2001). Many authors further argue that land

speculation has been a major driver of pasture creation, with individuals anticipating rising land

prices as new roads gradually bring forested lands into greater connectivity with market centers

(Hecht and Cockburn, 1990).

However, other authors, while considering pasture expansion as an important driver of

deforestation, present a different and somewhat more complex picture of its social and ecological

implications. Faminow (1998) argues that, rather than being an outgrowth of governmental

subsidies, the expansion in cattle production in the 1990's stemmed primarily from the growth in

regional demand, associated with an expanding and increasingly affluent urban Amazonian

population (see also Browder and Godfrey, 1997). In rural Amazonian communities, in addition

to serving as a capital investment, cattle represent an easily transportable and liquidable

commodity, thus serving as an insurance substitute. That is, cattle represent a fairly secure form

of stored wealth that can be quickly transformed into cash in cases of medical or other

emergencies (Perz, 2001 la). Furthermore, unlike horticulture and rubber extraction, the labor

requirements of cattle ranching are relatively low and product transportation is generally simple









(Arima and Uhl, 1997). And, given appropriate agricultural practices, such as rotational stocking

(Rueda et al., 2003) and grass-legume mixes (Carneiro and Valentim, 1997, Burns et al., 2004)

productive pastures can be maintained for decades or indefinitely while chemical inputs are

reduced or eliminated.

In an extensive discussion of the relationship between cattle production and deforestation

in the Amazon, Faminow (1998) points to the fact that many of the dire predictions offered by

scholars in the 1980's regarding the impacts of cattle expansion assumed an exponential growth

in deforestation rates, which in most cases has not been realized. Furthermore, he finds fault

with arguments coupling land speculation and pasture degradation as drivers behind pasture

expansion. Were this train of thought correct, he argues, it would assume an infinite supply of

irrational land speculators willing to pay incrementally higher prices for increasingly degraded

land.

Faminow begins his argument by countering assertions that Amazonian soils are unfit for

long-term agriculture, pointing to the similarities between Amazonian soils and those of the

Southeastern U.S., many of which have been in continuous cultivation for more than 200 years.

Furthermore, in light of recent evidence from the Amazonian cattle industry, Faminow and other

researchers have questioned the assumed lack of viability of cattle ranching as an economic

endeavor-pointing out that, despite the reduction of government subsidies, cattle ranching has

generally been shown to be economically productive (Faminow, 1998, Mattos and Uhl, 1994,

Arima and Uhl, 1997; Machagata and Brown, 2003). Under such arguments, the spread of cattle

ranching arises less as a result of speculation and land degradation than as a function of

economic decision-making by property owners operating in the existing political and economic

system (Machagata and Brown, 2003).









Various researchers have further argued that it is less cattle production per se than

inefficient grazing practices that threaten Brazil's tropical forests (Arima and Uhl, 1997,

Faminow, 1998). Researchers have found that heavier stocking rates (stocking rate > 1

head/hectare) and better managed pasture rotation, by minimizing the accumulation of senescent

forage, actually improve long-term pasture quality (Machagata and Brown, 2003; Faminow,

1998). Arguing that cattle-related deforestation currently arises from herd expansion rather than

land speculation, Machagata and Brown assert that intensification of cattle production, in many

cases, will slow deforestation. Coupled with improved enforcement of existing environmental

laws limiting deforestation, such intensification should be able to improve local livelihoods

without raising deforestation rates (Arima and Uhl, 1997; Machagata and Brown, 2003).

Fire

Another land use practice given close attention in my research project is the use of fire. In

the Brazilian Amazon, fire is a commonly-used low-cost tool to establish and maintain pastures,

especially in extensive cattle production systems. As most cattle production in the Amazon-

especially at smaller scales-is based on continuous rather than rotational stocking, cattle

producers often rely upon fire as a strategy to remove the low-quality senescent plant material

that tends to develop in such systems. Fire can have positive impacts for pasture development

and maintenance, such as the elimination of invading plants, reduced shading, and improved

carbon:nitrogen ratios through the release of nitrogen and the volatilization of carbon in

senescent plant material (Heringer and Jacques, 2002). However, it also results in negative

effects as well-including the volatilization of nitrogen and potassium and increased

susceptibility to erosion (Almeida, n.d.). And while grass regrowth yields a short-term increase

in palatable dry matter production, animal weight gains are generally lower than those in similar

unburned pastures (Almeida, n.d.). Further negative impacts of fire include the loss of crops,









fodder, livestock, timber, fences and other infrastructure in the cases of wildfire (as occurred in

several communities in the field site in 2005).

From the farmer's perspective, fire can have further negative impacts when forest areas are

accidentally burned. For example, researchers have noted a general decline in tree fruiting and

in the populations of frugivorous vertebrates in burned forest in the Amazon (Barlow and Peres,

2006). This in turn can have negative impacts for local populations who depend on non-timber

forest products (NTFPs) and hunting for part of their livelihoods. The reduced production and

eventual destruction of Brazil nut trees that have been affected by fire is a case in point.

While these impacts are of great concern at the level of the household and local

community, the impacts of wildfire upon the forest and for the larger socio-ecological system are

of even greater importance at regional and societal levels. Of particular concern are the potential

impacts of fire in damaging standing forests, affecting regional rainfall and affecting the health

and general well-being of local populations.

Burning of pastures tends to kill existing trees and prevent regeneration of new recruits-

leading to the long-term replacement of native forest vegetation with exotic grasses and fire-

adapted weeds. Furthermore, wildfire entering surrounding forests can lead to positive local-

scale feedbacks setting the conditions for future fire (Cochrane et al., 2003; Hoffman et al., 2003;

Alencar et al., 2004). While an initial fire will usually remain in the understory, the increase in

solar radiation from canopy gaps and in fuel load from dead or damaged trees provides the

conditions for successively more intense future fires (Cochrane et al.,, 2003). Reinforcing this

positive feedback is the probable impact of fire upon regional climate. Several researchers have

drawn attention to the possibility that smoke, by augmenting the number of condensation points,









may inhibit the creation of large atmospheric water droplets and thereby reduce regional rainfall

(Maslin et al., 2006; Cochrane et al., 2003).

In addition to the impacts for rural livelihoods and the natural environment, fire and smoke

have immediate negative implications reaching both rural and urban populations, including a

notable rise in respiratory illnesses (Mendonga, 2004). Mendonga (2004) estimates that as much

as 8% of all hospital-treated respiratory cases in the Brazilian Amazon are directly attributable to

the burning of biomass; in 1998, an El Nifio year in which large areas of the Amazon succumbed

to wildfire, the number of cases reached 13,000.

Roads

Roads are generally recognized as the strongest single predictor of tropical deforestation

due to their numerous direct and indirect impacts upon land use (Kaimonwitz and Angelson,

1998; Geist and Lambin, 2001; 2002; Pfaff, 1997). Road construction in places such as the

Brazilian Amazon generally increases the profitability of timber and agricultural production and,

by improving access, facilitates land speculation and colonization (Soares-Filho et al., 2004;

Chomitz and Thomas, 2001). In an extensive cross-national review of prominent deforestation

literature, Geist and Lambin (2001) found that road construction or paving served as a driving

factor in two-thirds of the cases examined.

Various analytical approaches have been used in the study of roads and their impacts on

tropical deforestation. In a ground-breaking paper, Chomitz and Gray (1996) used a spatial

econometric analysis in their study of deforestation in Belize. Using this technique, they found

that roads were correlated with agricultural land use, with soil fertility acting as a mediating

variable. Roadside areas with good soils had an especially high (50%) probability of conversion

to agriculture. Common in the literature on roads and land use has been the extrapolation of

historical patterns to future projections within road buffers. This approach has been applied, for









instance, in recent studies of the probable impacts of current highway paving projects in the

Brazilian Amazon (e.g., Laurance et al., 2001). However, the extrapolation of past to current and

future trends is an inherently error-prone procedure; it requires that other factors contributing to

land use change (such as government policy) be held constant-a problem that has surfaced in

widely differing predictions for future deforestation due to the Avanga Brasil highway paving

program.

Historically, Amazonian roads have contributed to agricultural expansion and deforestation

by changing migration patterns and opening the region to populations from other portions of the

country in search of land for subsistence or for investment (Moran et al., 1980; Schmink and

Wood, 1992). In the case of Acre, the opening of roads in the 1970's and 1980's led to an

increase in in-migration, in turn leading to numerous land conflicts (Bakx, 1986). While

government-sponsored colonization of the region has largely ceased, the impact of the now-

paved road for migration-especially between rural and urban areas-has continued into the mid

2000's.

Various predictions have been made regarding the likely implications of the Avanga Brasil

project. Predicted deforestation resulting from the project varies widely-due largely to differing

assumptions underlying the predictive models used. In particular, some models extrapolate past

patterns of governmental unwillingness or inability to create and enforce conservation policy to

the future (e.g., Laurance et al., 2001, Fearnside 2003), while others (e.g., Nepstad et al., 2001,

Silveira, 2001, Soares-Filho et al., 2004) allow for the possibility that recent political changes in

Brazil indicate a greater likelihood that viable systems of forest governance can emerge to

control land use in roadside forests. This topic will be addressed in greater detail subsequently.









Public Policy

Public policy is a commonly cited factor affecting deforestation (Skole et al., 1994),

including both overtly pro-deforestation policy and policy that has deforestation as an

unintentional consequence. It also includes policies oriented toward forest conservation, as Acre

has witnessed in the late 1990's into the 2000's (Kainer et al., 2004). As with roadways, policy

rarely functions alone in its impacts on land use but rather most often occurs in conjunction with

other drivers. Geist and Lambin (2001; 2002) found policy and governance to be a factor in 78%

of the cases of deforestation they considered-most often operating in tandem with other drivers.

This tendency holds true in the case of recent literature addressing the combined effects of roads

and governance in the Brazilian Amazon (e.g., Nepstad et al., 2001, Laurance, 2001, Nepstad et

al., 2002; Fearnside, 2002, Soares-Filho et al., 2004).

Among the recent government initiatives intended to control land use change and

deforestation in the Amazon has been agro-ecological zoning. Dennis Mahar of the Economic

Development Institution and formerly a member of the World Bank's POLONOROESTE

project, describes the mixed successes encountered by an ambitious agro-ecological zoning

initiative in the Amazonian state of Rond6nia (Mahar, 2000). The prescriptive zoning program

was intended to impose order in a state that had witnessed the squandering of forest resources

and displacement of traditional peoples during a 1980's land rush accompanying the World

Bank-funded paving of the BR-364 highway through the state. Under the zoning plan, land use

was to be regulated based upon both current practices and ecological suitability. Despite the

program's relative successes in controlling deforestation, Mahar points to several limitations.

Among the criticisms include the fact that few incentives exist for local-level governments to

enforce zoning rules and that perverse incentives exist for land owners to deforest illegally in the

hope of changing their land's zoning status.









A major platform of the Acre State Governo da Floresta has been Ecological-Economic

Zoning (ZEE) (Kainer et al., 2004). The concept of ZEE stems from language in the Federal

Constitution promoting the decentralization of governmental decision making-especially in the

area of environmental protection-including Articles 21, 23, and 30. Presidential Decree

number 99.540 of 21 September, 1990 created a Coordinated Commission for Ecological and

Economic Zoning of the National Territory (Governo do Acre, 2001)-outlining a conceptual

and legal framework for ecological and economic zoning to be implemented at the state level-

with the Legal Amazon defined as a priority area. ZEE represents an instrument to guide

governmental investments in socio-economic development in a way that recognizes and builds

upon the natural comparative advantages of each region. Through ZEE, planners in new

agricultural and settlement frontiers (such as many parts of Acre) have hoped to avoid the

anarchy, environmental destruction and social injustice of past occupation processes in the

Amazon. Implementation has been underwritten in part by financing from the PPG7 program of

"integrated environmental management projects" (PGAI) (Governo do Acre, 2001), with further

assistance from the GTZ.

In the past, agrarian and social policy in Brazil, by subsidizing colonization and pasture

expansion in the Amazon, tended to directly stimulate deforestation (Schmink and Wood, 1987;

1992). Since democratization in the mid-1980's and the coinciding domestic and international

concern about the fate of the Amazon's forests, many of the policies promoting deforestation

have been changed and new environmental regulations (e.g., establishment of national parks and

80% forest reserve regulations on private land) have come onto the books. However, the ability

of the government to effectively enforce these regulations is still in question and much of the

land conversion occurring in the Brazilian Amazon today continues to be illegal (Carvalho et al.,









2002). Compounding this problem, under-staffed Brazilian authorities often have great

difficulty detecting and combating violations (Carpentier et al., 2000).

Perhaps the largest debate regarding road paving and deforestation in the Brazilian

Amazon centers upon the ability of frontier governance systems to mediate this relationship.

Several researchers have offered qualified optimism in their projections for Avanga Brasil's

impacts upon deforestation. Soares-Filho et al., (2004) estimate that deforestation in the

Brazilian Amazon may be reduced by 1.3 million hectares by 2050 given a governance scenario

in which current environmental regulations are enforced. Such predictions are often based upon

the shift at the federal and, in some cases, at the state level away from policies that have

historically promoted deforestation in the region's highway corridors. An early example of this

shift was the creation and implementation of the RESEX model due to political pressure from

social movements in the 1980's and early 1990's (Schmink and Wood, 1992). More recently, at

the local level, researchers have noted the growing capacity of local governments in

environmental and development planning and the overall increasing effectiveness of governance

as a means of minimizing deforestation (Nepstad et al., 2002; Silveira 2002).

Nepstad points to the Environmental Crimes bill of 1998, authorizing IBAMA to levy fines

and jail sentences for unauthorized fire and land clearing and the specific case of PROARCO-a

recent fire control program-as evidence of the potential effectiveness of governance in

positively affecting land use change. Subsequent to the implementation of PROARCO in 2000,

Nepstad notes a two to four-fold decrease in the number of fires in satellite images of heavily-

settled areas of the southern and eastern Amazon (Nepstad et al., 2002). Other researchers have

found corroborating evidence for the effectiveness of governance as a means of minimizing

deforestation in other parts of Brazil that have not necessarily been recently affected by road









paving. For example, Fearnside (2003) notes that in the state of Mato Grosso, a state-level

deforestation licensing and enforcement program has demonstrated positive results.

Furthermore, areas of the state with the strongest enforcement have witnessed the greatest

declines in deforestation rates.

A counter-argument asserts that Brazil's return to democracy and the introduction of

environmental legislation will not be sufficient to significantly slow deforestation along highway

corridors. While environmental regulation has generally increased since the mid-1980's,

Laurance et al., (2001) note that, despite a dip during the early 1990's, deforestation rates

actually increased during the late 1990's, largely due to improved economic stability resulting

from the federal government's Plano Real monetary reform program. Like the Plano Real

program, even various governmental incentives designed to be pro-forest or forest-neutral have

had perverse effects. For example, Almeida and Uhl (1995) illustrate how the Brazilian Rural

Land Tax-a tool intended to encourage efficient and productive land use-actually penalized

farmers who wished to invest in forest-conservation and intensification of cattle production.

In the case of Avanga Brasil, Laurance estimates that between 269,000 to 506,000 hectares

of forest land will be lost annually due to the program (Laurance et al., 2001). Non-enforcement

of existing laws has been a commonly cited concern; despite improvements in remote

monitoring, it nonetheless remains difficult to monitor land use and enforce regulations,

especially in remote areas. According to this perspective, the problem is exacerbated by

corruption at various levels. Researchers have also pointed to problems intrinsic in the political

process behind Avanga Brasil and similar projects that may contribute to deforestation.

Fearnside (2003) notes several specific problems in the environmental legislation process

common in Brazil. In many projects, such as Avanga Brasil, the project is proposed on its









own-without a suite of alternative projects. Furthermore, efforts are often made to secure

funding prior to impact assessments-consequently building political support that will help

ensure the project's ultimate vitality regardless of the environmental impacts. Finally, little

impact assessment comes from within the project programs, but as in the case of Avanga Brasil,

studies tend to come largely from independent research organizations outside of the project.

Political Ecology

My research draws heavily upon a political-ecological approach to land use and land cover

change in this region. A political-ecological perspective explores the relationships between

human society and natural resources, giving special attention to the role of politics as individuals

and groups contest rights to these resources (Bryant and Bailey, 1998). The term "political

ecology" can be traced to the early 1970's in Eric Wolf s call for an approach integrating land

use with a local-global political ecology (Wolf, 1972; Peet and Watts, 1996). In response to both

Neo-Malthusianism and closed-system ecological anthropology, early political ecology drew

upon a Neo-Marxist perspective to bring politics and class struggle into what previously had

been a largely apolitical view of human-environment interactions (Schmink and Wood, 1987;

Peet and Watts, 1997; Bryant and Bailey, 1998).

Informing the environmental movement of the 1960's and 1970's, Darwinism and Neo-

Malthusianism saw unregulated population growth as a key factor in global environmental

degradation (Peet and Watts, 1997). For example, the acceleration of tropical deforestation at

this time was, in both academic circles and in the popular media, often blamed on faceless rural

peoples whose unabated population growth drove them ever further into the forest in search of

cultivable land (Jarosz, 1996). Seeing these explanations as inadequate, early political ecologists

looked to political and social factors to better explain phenomena such as environmental

degradation and famine (Watts, 1983; Moore, 1996).









Systems-based ecological anthropology represented another key paradigm spurring the

early development of political ecology (Watts, 1983, Peet and Watts, 1997). In particular,

political ecologists drew exception to the homeostatic equilibrium-based models then dominant

in ecological anthropology (e.g. Rappaport, 1967, Orlove, 1980). Such work tended to focus

upon energy and resource flows in closed-system societies (Bryant and Bailey, 1998). For

instance, in the book Pigs for the Ancestors, Roy Rappaport (1967) used a systems ecology

approach to explain the livelihood system of the Maring-an isolated tribal group living in New

Guinea. As part of his argument, Rappaport claimed that ritual slaughtering of pigs served as a

means whereby the Maring maintained equilibrium within their environment and society. In

addition to his use of an equilibrium-based model, Rappaport was also criticized for his

inattention to the wider political-economic context in which the Maring lived. In contrast,

political ecologists argued that human-environment interactions could only be understood within

the context of local-global articulations and linkages among the local community, the nation-

state and international institutions (Biersack, 1999).

Political ecology emerged as a distinct theoretical framework in the mid-1980's through

analyses of the political-economy of environmental degradation. In a discussion of soil erosion

in the developing world, Piers Blaikie (1985) argued that environmental decision-making at a

given scale is best explained when contextualized into the political-economic context of the next

larger scale. Consequently, he called for a bottom-up research perspective which begins with the

smallest decision-making level and progressively scales upward to the level of state and

international political-economic structures in order to understand the forces shaping local level

land use.









In one of the first works to apply an explicitly political ecological approach to the Amazon,

Schmink and Wood (1987) discussed the ways in which the wider socioeconomic and political

context of the Brazilian Amazon have affected human use of natural resources while also looking

at the implications of this relationship for environmental policy in the region. Schmink and

Wood proposed a political-ecological model that links the Amazonian political-economy-

including, among other factors, the state, markets, and class structure-to patterns of resource

exploitation in the region. Like Blaikie, Schmink and Wood focused upon local scale land use

decisions, but saw them as being largely determined by larger-scale political-economic

phenomena like capitalist expansion.

In their study of frontier expansion in the Brazilian Amazon, Schmink and Wood (1992)

considered frontier expansion and land conflicts surrounding the town of Sao Felix do Xingu in

the state of Para. This discussion is couched within analysis of the regional, national and

international settings. The authors offered a temporal contextualization of the problem-placing

the processes occurring at these multiple levels within a broad historical context encompassing

nearly 500 years. In this study, the authors expanded beyond this largely top-down framework in

their earlier work. The authors addressed not only the ways in which political and economic

structures have affected land use in the Sao Felix do Xingu region. They also addressed the

ways in which individual actors, through their land use and through their contestation of rights to

land, shaped the wider political and economic environments. Hence, they brought a sense of

individual agency to questions of political-economy and land use that had largely been absent in

earlier political ecological work.

Wood and Porro (2002) offer an implicitly political-ecological framework for

conceptualizing the relationship between the socio-economic and biophysical drivers of









deforestation in the Amazon Basin. In this model, the ultimate agent of land use and land cover

change-the household or firm-is contextualized within increasingly larger-scale

socioeconomic and biophysical drivers. The household or firm's land use and land cover

outcomes, in turn, have feedback effects for all scales of socio-economic and biophysical drivers.

Hence, Wood and Porro, like Schmink and Wood (1992), introduced a multi-directional chain of

causality into their model of LULCC. However, the feedback effect occurs not through the

direct agency of the household (or firm) responding to and contesting the larger scale political-

economic context, but through the land use and land cover outcomes themselves.

Through the 1980's, political ecology tended to draw upon a neo-Marxist perspective to

look at local conflict largely in terms of capitalist penetration and class conflict (Bryant and

Bailey, 1997). Subsequently, however, political ecologists have begun to draw upon a wider

range of theory to explain human-environment relationships-including, among others, gender

analysis. Using a gender analysis approach, researchers have found that the traditional focus

upon social class in political ecology has overlooked other key variables affecting human-

environment relations (Sontheimer, 1991; Roucheleau, et al., 1996; Schmink, 2000).

Environmental policies may have very different impacts for men and women when they do not

account for gendered patterns of resource access and use (Schroeder and Suryantata, 1996). For

example, in her study of a government-sponsored agroforestry project in the Gambia, Carney

(1996) found that, by overlooking the role of gender in shaping local resource access and the

traditional male control of tree resources, the project effectively shifted the balance of resource

control toward men and encouraged the economic marginalization of women in the affected

communities.









While offering a gender analysis of governance and land use is not the primary purpose of

the dissertation, attention to gender provides an understanding of the variable ways in which

roads and governance affect households and individuals within households in the BR-317

highway corridor. Specifically, I address the variable ways in which the road has affected men

and women, addressing gender-specific migration patterns between rural and urban areas. By

tying this analysis with gendered land use practices (i.e. approaches to cattle production), I

demonstrate the way in which gender can serve as a mediating factor between road paving and

land use.

Governance and Institutional Design

Although it doesn'tWhile not falling directly within the rubric of political ecology, other

like-minded authors have addressed issues of governance and their implications for land use.

Complementing political ecology, a considerable body of literature addresses the questions of

institutions and institutional design and the implications for the management of collective

resources (Ostrom, 1990, Agrawal and Gibson, 1999, Ostrom, 2005). Ostrom defines

institutions as "the prescriptions that humans use to organize all forms of repeated and structured

interactions including those within families, neighborhoods, markets, firms, sports leagues,

churches, private associations, and governments at all scales" (Ostrom, 2005: 3; see also:

Agrawal and Gibson, 1999; Kiser and Ostrom, 1982; and Adger et al., 2003). Institutions may

represent mechanisms of self-organization by communities or may be created and imposed by

external entities such as state governments (Poteete and Ostrom, 2002). Institutions include both

the overall institutional framework and the specific institutional arrangements of particular sets

of rules-and often form nested structures in which larger-scale institutions limit the alternatives

available at lower levels (Adger et al. 2003).









Like political ecology, institutional theory developed in opposition to Neo-Malthusian

theories forecasting the inevitable demise of common pool resources not subject to privatization

or state control. For example, in his much-cited study, Garret Hardin (1968) argued that in

common pool resource systems-such as commonly held grazing pastures-the rational choice

of each participant is to overuse the resource as s/he appropriates full benefit from each marginal

unit of use while the impacts of resource degradation are distributed among all members. Such

authors recommended alternatively for government control or privatization (Demsetz, 1967;

Posner, 1977). Implicit in both arguments is an assumption that local resource users are

incapable of designing effective means of self-governance (Ostrom, 1997). While institutional

theorists generally accept that Hardin's "tragedy of the commons" may in fact apply to truly

unregulated open-access resources, in many cases, communities have been able to avoid this

situation through institutions of self-governance-allowing joint use that is not permitted to

exceed some threshold of extraction (Ostrom, 1990; Ostrom et al., 1999; Agrawal, 2001)

While institutions shape nearly every aspect of human life, a significant body of literature

focuses specifically upon the institutions of land use. According to Agrawal and Gibson,

institutions of land use face several shared tasks: they establish rules about the use, management

and conservation of resources, the implementation of the rules that are created and the resolution

of conflicts that consequently arise (Agrawal and Gibson, 1999). Ostrom (2005) points to

several elements of institutional design that are generally requisite for effective management of

shared resources. Among the elements of successful design are: clearly defined boundaries,

proportional equivalence between benefits and costs, graduated sanctions, conflict-resolution

mechanisms, collaboration between local and government authorities in the enforcement of rules









and trust between members-whether through shared cultural norms or repeated interactions, or

both (Ostrom, 2005).

Political Ecology in the Dissertation

In the dissertation, I have utilized a political-ecological approach to explore the

relationship between government agrarian, forest and infrastructure policies and household and

association-level land use decisions. In this process, while I have focused upon the household as

the unit of analysis, I have framed the study of a household's land management decisions within

the association and the administrative unit, through the state, federal, and international contexts

of environmental governance. While the dissertation does not focus exclusively upon political

drivers of land use change, it is grounded in an understanding that the multi-directional

relationship between a household and wider political arenas plays an important role in shaping

how land use decisions are made.

By considering local-level governance systems and the role of individuals in interpreting,

responding to and reshaping governance systems through their land use practices, I have

attempted to avoid deterministic interpretations based on top-down chains of causality between

large-scale government policies and an individual's land use decisions. In the dissertation, I use

a political-ecological approach to explore the relationship between centralized governance

systems and household-level land use decisions. Similarly, I explore the implications that a

household's elective decision to participate in or opt out of the various institutions of governance

available to them affects their land use. I also address the ways in which individuals and

households, through their participation in wider structures of governance, reshape the landscape

of environmental governance.

In a 1983 essay entitled "Against Political Ecology", (Vayda and Walters, 1999), Andrew

Vayda and Bradley Walters offer several critiques of political ecology, including the scant









attention it often gives to ecology. However, most important, I would argue, is their argument

that political ecology makes an a priori assumption that politics is always central to the use of

natural resource issues. In this article, they draw on a case study of mangrove deforestation in

the Philippines to show how other non-political factors, such as tidal action and local preferences

for certain timber species have been equally, if not more important than politics in shaping the

extent and composition of the mangroves. An exclusive focus on political contestations for

rights to resources would have overlooked other important explanations for deforestation

occurring in this region.

While not seeing it as a ground for the out-right rejection of political ecology, Vayda and

Walter's concern about the tendency in political ecology to prioritize politics as the a priori

explanation for any and all phenomena involving natural resources has guided my thinking as I

apply a political ecological perspective to the topic of roads, governance and land use. For

instance, in quantitative analyses, I recognize the importance of and control for largely apolitical

explanations of land use such as property size and years of residence. Similarly, in the final

chapter, when exploring the contradictions between the concern that people express about the

environment and their actual land use practices, I recognize that multiple factors, including, but

not limited to politics, are likely involved.

Another critique that has been leveled against much political ecological literature, even

that which addresses questions of agency and feedback loops between political economic

structure and local land use decisions, is the tendency to make an a priori judgment that political

decisions made at the local level are preferable to those made at larger scales. For example,

Brown and Purcell (2005) argue that the question of scale should be treated as an object of

inquiry in political-ecological studies. They note that much political-ecological work has fallen









into a "local trap," in which local scale actors and organizations are assumed to be more effective

in promoting positive outcomes such as environmental sustainability and social justice. They

argue that political-ecological research should carefully address the question of scale without

making a priori assumptions that one scale should be preferred over another.

The question of scale is treated implicitly throughout my research project-for example,

consideration is given to the land use implications of regional and national scale migration on

land use; similar consideration is given to the land use implications of the actions of

governmental agencies from municipal to national levels. Furthermore, scale is treated as an

explicit question of analysis in two discussions about the relationship between governance and

land use. First, addressing Brown and Purcell's concern about the "local trap" in political

ecology, I consider the effectiveness of associations operating at two distinct spatial scales: at the

level of the administrative unit and of the community. That is, in each hypothesis concerning

governance and land use, I use two measures of governance-one operating at the household

level (e.g., participation in courses, STR, etc) and one at the level of the administrative unit (e.g.,

deforestation limits, IBAMA presence, etc)-to assess at which scale (household or

administrative unit) the impacts of governance upon land use can be best measured.

In drawing upon political ecology, I have also recognized its limitations. While not

necessarily a weakness, due to its interdisciplinary nature and numerous theoretical influences,

political ecology has not developed into a coherent unified theory but represents a broadly-

defined approach and area of interest that can be integrated into a diverse array of perspectives

on human-environment interactions. Furthermore, as political-ecology tends to focus upon

narrative explanations of human-environment interaction it is rarely, in and of itself, well

adapted to rigorous hypothesis testing (Peet and Watts, 1996). In the case of my research, I have









sought to overcome this limitation by drawing both on political ecology and the broader LULCC

literature. While theoretically diffuse, the LULCC literature does provide fertile material for

developing testable hypotheses. A case in point is the argument between Laurance and Nepstad

about whether or not governance reduces predatory land use in road corridors. Hence, I have

drawn largely upon the LULCC literature to identify key variables and hypothesize causal

relationships between them. Political ecology and institutional theory perform an important

complementary role by focusing the analysis on key factors not commonly discussed in the

LULCC literature-including individual agency, contestation for the rights to resources and to

define the "rules of the game", and the implications of scale in affecting the role that governance

has for land use.

Variables Used in Analysis

Drawing upon the LULCC literature, especially that which focuses upon issues of roads

and governance in the Brazilian Amazon, I identified five land use and land cover variables

which are treated as dependent variables and 10 variables that can be considered drivers of

LULCC and are treated as independent variables in chapters 3 and 4.

Dependent Variables

Deforestation

Deforestation is a continuous variable which refers to the net area, in hectares, of mature or

secondary forest that had been removed by the family from the time of arrival on the property.

As interviews occurred over a one year period, I used 2005 as a base year for calculating this

variable. To obtain this information, I subtracted the amount reported as deforested at the time

of the household's arrival on the property from the amount deforested at the time of the

interview. Household-level deforestation represents the area in hectares of forest (primary and

secondary) that had been converted to non-forest during the family's tenure on the property. As









mentioned in chapter 1, this method can be subjected to under-reporting. However, given the

techniques available to minimize this bias (also addressed in chapter 1) and the difficulty of

determining the exact property boundaries in most portions of the field site, I elected to use

reported values rather than remote sensing.

Cattle herd size

Herd size reflects the total number of cattle owned by household members, including cattle

held both on and off-farm. Data were also collected for on-property cattle herd size (regardless

of owner). As my unit of analysis in the dissertation was the household and not the land parcel, I

elected to use the former measure rather than the latter.

Brazil nuts

Brazil nut harvest is a continuous variable measuring the amount in latas (1 lata = 18

kilograms) of Brazil nuts that were harvested by the household in 2004-including those

destined for consumption and for sale.

Annuals

Annuals is a continuous variable representing the area, in hectares, that the household had

dedicated to the production of annual crops (e.g., beans, rice, manioc, maize) at the time of the

interview.

Last burn

Years since last burn is a continuous variable referring to the number of years that had

passed since the last time the household had intentionally burned an area of established (as

opposed to new) pasture on their property Since interviews were conducted over a one year

period, I used the year 2005 as a baseline for calculating this variable. Years since burning

1 While great variability exists between families in terms of fire use as a pasture maintenance strategy, given the
general non-existence of technological alternatives to fire when forming pasture in the region, fire is used nearly
universally when new pasture is established.









established pasture reflects the number of years between 2005 and the last time fire was

intentionally used in established pasture on the property (excluding fire used to establish new

pasture).

Independent Variables

Place of origin

Place of origin was coded as a binomial variable. Households with both heads born in

Acre or neighboring regions of Bolivia or Peru2 were labeled "Acreano" and coded as 0.

Households that had at least one head that was born outside of Acre or neighboring areas of

Bolivia or Peru were labeled "Migrant" and were coded 1.

Distance from road

Distance from road is a continuous variable that measures the distance (in hours) that a

household lies from the Inter-Oceanic Highway. As an individual may use various forms of

transportation (walking, bicycle, horse, motorcycle, truck, etc) on different occasions, I

standardized this measure by using travel time by foot.

Bi-localism (male and female)

Male and female bi-localism are dichotomous variables and represent the frequency with

which male and female household heads, respectively, traveled to an urban area each month and

were coded as binomial variables (0=low and l=high)3.



2 All individuals in the sample who were born in Peru and Bolivia were ethnically Brazilian, most of whom traced
their ancestry to people from Brazil's Northeast who had come to Acre and neighboring regions of Peru and Bolivia
to tap rubber several generations earlier.

3 Male and female bi-localism were initially coded as four-category categorical variables (0=never/almost never,
l=less than monthly, 2=monthly, 3=more than monthly, but less than weekly and 4=weekly or more). Given the
lack of variance in the 0 category (n=l for both male and female bi-localism) and the lack of large differences
between some categories, I recorded the five categorical variables as binomial variables. For both male and female
bi-localism, the lower level (codes 0,1 and 2) was coded as 0 and the higher level (3 and 4) was coded as 1.









Adult labor (male and female)

Adult labor (male and female) are continuous variables and represent the number of

"working age" (15+ years of age) males and females in each household. I used household labor

as a proxy measure for out-migration. While the number of working age household members

does, of course, reflect other factors beyond out-migration (e.g. births, mortality), it is also

intuitive that the effects of fewer working-aged adults upon land use would parallel the effects of

adult out-migration.

Years of residence

Years of residence is a continuous variable measuring the number of years that a household

has resided on the property. As fieldwork was conducted over the course of a year, this measure

was standardized using 2005 as a base year (i.e. if a family arrived in 1985, the value would be

20). This variable serves as a control in multiple regression analysis.

Property size

Property size is a continuous variable measuring the area (in hectares) of the household's

land holding. This variable serves as a control in multiple regression analysis.

Governance (administrative unit)

Unit governance is an ordinal variable based on a four point index of the governance

characteristics of administrative units in the region. This index is explained in detail in chapter

4.

Participation in governance

Participation in governance is a continuous variable. It is a factor-weighted index based on

principle component factor analysis of 12 individual measures of participation in governance.

This index is explained in detail in chapter 4.









Hypotheses


Roads and Land Use

Within the first general research question about the impacts of the road for land use, I

address four specific issues and how they relate to land use. First, regarding the impacts of the

road by opening the region to colonization in the 1970's and 1980's, I posited the following four

hypotheses about the lasting impacts of place of origin and land use:

* H 3.1 4 Migrant households tend to exhibit higher rates of deforestation than those with
heads born in Acre.

* H 3.2 Migrant households tend to own more cattle than those with heads born in Acre

* H 3.3 Migrant households tend to produce more Brazil nuts than those with heads born in
Acre

* H 3.4 Migrant households tend to burn their pastures more frequently than those with
heads born in Acre

Secondly, to address the implications of rural-urban migration and bi-localism that have

been facilitated by the construction and especially the paving of the highway, I posited the

following eight hypotheses:

* H 3.5 Higher levels of female bi-localism are associated with less production of annuals.

* H 3.6 Higher levels female bi-localism areassociated with less production of Brazil nuts.

* H 3.7 Higher levels of female bi-localism are associated with larger cattle herds

* H 3.8 Higher levels of male bi-localism are associated with less production of annuals.

* H 3.9 Higher levels of male bi-localism are associated with less production of Brazil nuts.

* H 3.10 Higher levels of male bi-localism are associated with larger cattle herds.

* H 3.11 Lower numbers of working age females are associated with less production of
annuals.


4 Hypotheses starting with 3 are addressed in chapter 3. Hypotheses starting with 4 are addressed in chapter 4. For
example, H 3.1 is the first hypothesis addressed in chapter 3.









* H 3.12 Lower numbers of working age females are associated with less Brazil nut
production.

* H 3.13 Lower numbers of working age females are associated will larger cattle herds.

* H 3.14 Lower numbers of working age males are associated with less production of
annuals.

* H 3.15 Lower numbers of working age males are associated with less Brazil nut
production.

* H 3.16 Lower numbers of working age males are associated with larger cattle herds.

Thirdly, to address the implications of access to the highway for land use, I posit the

following two hypotheses::

* H 3.17 The greater the distance from the roadway, the lower the deforestation.

* H 3.18 The greater the distance from the roadway, the smaller the cattle herds.

To address the implications of the expanding cattle economy that has developed in the

wake of the road's construction and paving, I posited the following four hypotheses:

* H 3.19 The larger the cattle hers size, the less area dedicated to the production of annuals.

* H 3.20 The larger the cattle herd size, the lower the production of Brazil nuts.

* H 3.21 The larger the cattle herd size, the more frequently pasture is burned.

Governance and Land Use

To address the role of governance, at the level of the administrative unit, in affecting the

land use in the highway's area of influence, I posit the following three hypotheses:

* H 4.1 Higher levels of governance at the administrative unit level correspond with lower
deforestation.

* H 4.2 Higher levels of governance at the administrative unit level correspond with smaller
cattle herds.

* H 4.3 Higher levels of governance at the administrative unit level correspond with less
frequent pasture burning.









Finally, in order to test the implications of participation in governance for land use, I posit

the following three hypotheses:

* H 4.4 Higher levels of participation in governance correspond with lower deforestation.

* H 4.5 Higher levels of participation in governance correspond with smaller cattle herds.

* H 4.6 Higher levels of participation in governance correspond with less frequent pasture
burning.









CHAPTER 3
ROADS AND LAND USE CHANGE IN ALTO ACRE

A central issue in discussions about sustainable development in Brazil concerns the effects

of highway paving on social-ecological systems in the Amazon Basin. While road paving

promises to improve access to goods and services among previously isolated rural and urban

communities in Amazonia, it can also have negative environmental and social consequences,

including deforestation, widespread use of fire, and the displacement of rural smallholders.

Roads in tropical ecosystems have brought negative ecological impacts including fragmentation

of forest cover, the release of carbon into the atmosphere, increased fire frequency and species

extinctions (Cumming et al., 2005). In the case of Amazonia, this has been accompanied by

negative social impacts such as the increased violence and dispossession of traditional land

holders (Schmink and Wood, 1992). In academic and policy circles, a heated debate exists

regarding the land use changes brought by roadways and the capacity of governance to prevent

them (e.g., Nepstad et al., 2001, 2002; Laurance, 2001, Soares-Filho et al., 2004). Given that

households are often the ultimate agents of land use change in the Brazilian Amazon's highway

corridors, a household-level research approach provides an important perspective in

understanding the decision-making processes driving land use change in areas such as Alto

Acre's Inter-Oceanic highway corridor.

The first part of this chapter addresses the question of migration-specifically how land

use differs among migrants and natives, and among households of different age and gender

combinations. It also addressed the impacts of bi-localism (the division of residence between

rural and urban areas) for land use. The second portion of the chapter addresses the impacts that

roadway access and expanding herd sizes have for land use.









Descriptive statistics for the independent and dependent variables1 used in the analysis are

shown in Table 3-12. Of the sampled households, 38% had at least one head born outside of

Acre. On average, households had more male adults present (1.95) than female adults (1.38).

However, women scored slightly lower on the measure of bi-localism than did men. On average,

members of households included in the study had to walk 1.5 miles to reach the paved highway

and had been on their properties for nearly 20 years. Property size varied greatly-from 40 to

1200 hectares, with the average property having 254 hectares. The majority (59%) reported

using credit as of the time of the interview.

Brazil nut production was highly variable between properties, though most households

produced modest amounts. The mean Brazil nut harvest for 2004 was 26.5 latas. The average

household reported having cleared 24.24 hectares of land during their tenure on the property,

though this amount varied from 0 to 298 hectares. Of the land that had been cleared, households

reported an average of 3.35 hectares dedicated to the production of annuals as of the time of the

interview. Like property size and Brazil nut production, cattle herd size was also variable-

ranging from 0 to 800 head. The average household owned 58 head of cattle. On average,

households had spent 3.19 years since the last time they had last intentionally burned their cattle

pasture.

Roads and the "Occupation of the Amazon"

In the early 1960's Amazonia's long-standing isolation from the rest of Brazil came to an

end. While previous administrations-most notably that of President Gertulio Vargas several

decades earlier-spoke of the importance of integrating Amaz6nia into the Brazilian mainstream,

1 The variable "cattle" is used as a dependent variable when testing the impacts of migration for land use. It is used
as an independent variable when testing the relationship between expanding herd size and other land uses. Potential
issues of endogeneity in the use of cattle as an independent variable are discussed in that section.
2 Descriptive statistics for variables occurring in both chapters 3 and 4 are displayed in Tables 3.1 and 3.2.









by and large this did not result in concrete action until the 1960's. In the early part of the decade,

when the region's first major highway was constructed linking Belem with the new national

capital of Brasilia, Amaz6nia's historical isolation began to change rapidly. During the 1960's,

and even more so in the 1970's and 1980's, the federal government undertook an enormous

effort to encourage the colonization of the Amazon through propaganda, road construction, and

massive fiscal incentives (Moran, 1981; Schmink and Wood, 1992).

A number of explanations have been offered for this relatively sudden interest in the

colonization of the Amazon Basin on the part of the federal government. Subsequent to the

military coup in 1964, the government developed a xenophobic fear of the "internationalization

of the Amazon." Informed by scholars such as A.F. Reis (Reis, 1960), government officials

feared that a sparsely populated and poorly integrated Amaz6nia was vulnerable to threats from

neighboring and northern nations interested in the region's vast resources. This fear was

exacerbated by reports such as a 1971 United Nations Food and Agriculture Organization

document suggesting that, if intensively farmed, the Amazon Basin could feed a global

population of 36 billion (Pawley, 1971, Smith, 1982: 13).

Furthermore, the government's view of the Amazon as a "safety valve" for overpopulated

and impoverished regions of the nation was another important incentive behind programs to

construct roads and colonize the area (Bunker, 1985; Hall, 2000). In the Northeast, a series of

droughts led to waves of out-migration to other parts of the country, including Amazonia.

Exacerbating this problem, in both the Northeast and the South of Brazil, concentration of land

ownership and the growth of agro-industry was forcing smallholders from their land (Smith,

1982). While hindsight shows that implementation of agrarian reform would likely have been a

much more socially just and cost effective strategy to deal with these problems (Oz6rio de









Almeida, 1992), doing so would have run contrary to the right-wing dictatorship's philosophical

stance and, more importantly, by angering wealthy land owners, would have undercut a

fundamental component of the dictatorship's political base.

Combined with the desire to avoid the internationalization of the Amazon and to relieve

political pressure for agrarian reform was a desire on the behalf of the government to exploit

what was perceived as a vast untapped reserve of natural resources including timber, minerals,

and agricultural land (Bunker, 1985; Oz6rio de Almeyda, 1992). The development of this region

and its resources was intended to absorb investment capital and attract surplus labor from

overpopulated and impoverished areas of Brazil such as the Northeast (Schmink and Wood,

1992).

Highway construction, especially in sparsely-populated areas, often led to negative social

and environmental consequences including land speculation, rising land prices, population turn-

over and deforestation (Schmink and Wood, 1992; Hecht and Cockburn, 1990; Perz, 2001,

Nepstad, 2001, 2002; Laurance, 2001). Road construction and colonization generally went hand-

in-hand as large-scale colonization was only feasible given overland connectivity to regional and

national markets and population centers. Indeed, the creation of the Amazon's major

colonization poles (e.g., Trans-Amaz6nia and Rond6nia) was inextricably linked to the roadways

built through them (Browder and Godfrey, 1997).

Road Paving and Colonization in Alto Acre

In 1969, Acre's isolation from the rest of Brazil began to break, with the construction of

the unpaved BR-364 highway linking the capital city of Rio Branco with Rond6nia and the

remainder of the country (Bakx, 1986; Sawyer, 1984). The following year, Governor Wanderlei

Dantas, in a bid to stimulate outside investment in the state, offered a now infamous invitation to

an audience of Sdo Paulo ranchers and investors to "Produce in Acre, Invest in Acre, Export via









the Pacific." Many accepted the invitation and, by 1971, a full-scale land rush was well

underway (Bakx, 1986).

Initially, the roadway led to greater autonomy for rubber tappers, who consequently had

freer access to markets for selling rubber and purchasing other goods (Cardoso, 2002).

However, with the construction of the road and the arrival of Southerners came intense land

conflict. Rubber barons, many of whom had largely abandoned their estates in prior decades,

returned to reap the profits from soaring land prices, which had risen up to 2000% in highway

corridors (Bakx, 1986). State policies were enacted and land exchanged hands with little, if any,

recognition of the rubber tappers who continued to occupy the land. Indeed, many ill-informed

ranchers and investors were surprised and dismayed to find their land already occupied (Bakx,

1986). As rubber tappers rarely possessed legal title to their land, they were commonly evicted,

at times with threats of violence. Many crossed the border to Bolivia or relocated to the

burgeoning periphery of Rio Branco; others remained to mount a resistance movement to protect

their rights to the land (Bakx, 1986; Cardoso 2002).

By 1973, ranchers had begun resorting to violence as a means of removing rubber tappers

(Keck, 1995; Schmink 1992) and, as a consequence, with the assistance of political allies, rubber

tappers began to organize into a social and political movement. A principal technique

throughout the struggle was the use of empates-or non-violent stand-offs-when rubber tapper

land was threatened with clearing by ranchers (Keck, 1995; Barbosa, 2000; Kainer et al., 2003).

As in other parts of the Brazilian Amazon, grass-roots resistance movements emerged to

challenge the expropriation of land and destruction of rubber tapper livelihoods. However, the

movement in Acre was undoubtedly among the most dynamic, widely publicized, and ultimately

successful of the grassroots resistance movements of the Brazilian Amazon in this era. As it









gained in political power and publicity, the rubber tapper movement expanded into increasingly

broader political spheres (Keck, 1995)-from household-level conflicts to regional, state and

ultimately the international environmental movement (Schmink and Wood, 1992; Keck, 1995;

Kainer et al., 2003).

In the early 1970's the Alto Acre region became one of the first areas in the state to be

affected by the arrival of ranchers and speculators from Brazil's South (Bakx, 1988). It also

witnessed some of the most intense conflict between the newcomers and resident rubber tappers

defending their rights to the land they had occupied for generations. During the 1970s and

1980s, ranchers obtained and cleared large tracts of land in Alto Acre near the BR-317

highway-a spur from the BR-364 leading from Rio Branco to the Peruvian border. As a result,

the Alto Acre region, and my research site in particular, was a focus of conflict, especially in the

early years of the empate movement. Subsequent to the assassination of STR leader Wilson

Pinheiro in 1980 and the retaliatory killing of the ranch manager widely believed responsible,

the government saw the need to intervene before social unrest further destabilized the region.

These conflicts helped lead to the creation of the colonization area and, several years later in

1988, PAE Santa Quiteria-one of the first experiments in extractive reserves to be implemented

in the Brazilian Amazon. Ironically perhaps, some of the most out-spoken proponents of the PAE

and the preservation of rubber tapper livelihoods (including some who are widely believed to

have been actively involved in the violent stand-off with the rancher) have themselves become

some of the area's largest ranchers and vocal critics of environmental regulations in the PAE.

While a handful of the older ranches, including Fazenda Sta. Rita, have persisted to the

present, the majority were short-lived. In an effort to restore social calm in the region, many

ranches in the area were ultimately appropriated by INCRA and redistributed as 60-100 hectare









lots in Directed Settlement Project Quixada to both displaced rubber tappers and to landless

migrants arriving from outside of the state. During the early and mid-1980's, INCRA registered

and settled families onto 301 lots in PAD Quixada, Gleba 6 (the portion of the Quixada Project

included in my research project), offering land and occasionally building material, and

transportation to the region. While migrants came from many areas of the country, the majority

came from Brazil's south, especially the largely deforested former agricultural frontier of Parana.

Despite free land, the journey and subsequent fight for survival that migrants faced were

difficult, with many migrants abandoning their lots and returning to their homelands in Brazil's

South or relocating to the peripheries of neighboring towns after finding the extreme isolation

and the difficulties posed by the lack of basic social services and technical isolation

insupportable. This turnover is reflected in the fact that, though INCRA had discouraged the sale

of lots, only a third of interviewed families from PAD Quixada were either the original owners

or children of the original owners of the lot.

While most of the migrants in the region initially settled in the colonization area at the

highway's edge, as land became scarce in the colonization area, many subsequently moved into

PAE Santa Quiteria and, to some degree, to the RESEX Chico Mendes, both of which lie behind

the colonization area. While this has at times occurred through land invasion, the steady stream

of out-migrants from the RESEX and PAE willing to sell their coloca(oes for low prices has

provided a means of doing so without outright social conflict.

One family living immediately behind PAD Quixada in neighboring PAE Sta. Quiteria told

me that they had initially migrated in the 1980's from the state of Goais, along with the

husband's father, to a farm on the road's edge in the colonization area. As the then-young

couple began to start a family of their own, rather than purchasing a lot along the road, they









decided upon a relatively inexpensive and completely forested 200 hectare land holding nearby

in the PAE. While they would technically need community approval through AMPAESQ-an

association representing the PAE's residents-they were told by a contact in INCRA that they

could proceed with the purchase, but with no guarantee that they would not subsequently be

expelled by the community. Ultimately, the gamble paid off and they successfully received a

cartdo de assentamento ("settlement card") from INCRA, entitling them to live in the PAE.

While the majority of the residents of the PAE and RESEX were native-born Acreanos, such

stories from migrants and their children were nonetheless common.

At the time of research, cultural differences between native-born Acreanos and southern-

born migrants were reflected in stereotypes commonly held by each about the other group. On

one hand, migrants-especially older Southerners-were often held to be rigid and money-

oriented, with little concern for the forest, and at times insular and disinterested in participating

in community organizations, such as associations. On the other hand, while many migrants

boasted of their hard work in the hope of building a better future for themselves and their

children, they would at times contrast their own ambitiousness with the complacency of many

Acreanos who were perceived to be content to live on the threshold of poverty as long as the

immediate needs of the present were met. As evidence for such "preguiga" ("laziness"),

migrants would occasionally point out the modest homes, small clearings and lack of well-

maintained pasture of native-born Acreanos.

Despite initial conflicts and lingering stereotypes between these two groups, relations

generally improved over subsequent years and decades. Inter-marriage between native-born

Acreanos and migrants became commonplace, and many younger members of migrant families

had little or no recollection of the South, referring to themselves not by the place of their parents'









birth but simply as "Acreanos." This is reflected in a statement made by a daughter of

Paranaense migrants in her late 20's who told me that she took personal offense when people

spoke derogatorily about Acreanos-she scarcely remembered Parana and despite her blond hair

and blue eyes (commonly associated with southern migrants), she considered herself Acreana.

Intermarriage between Acreanos and migrants was common; in PAD Quixada, for example, 25%

of married couples had one spouse from Acre and the other from elsewhere (in all cases, the

husband being a migrant and the wife being Acreana). Through this process of inter-marriage

and the coming of age of Acre-born children of migrants, the cultural distinctiveness of

indigenous, northeastern and southern ethnic identities had largely melted into a shared Acreano

identity, especially among younger individuals. Children of migrants tended to refer to

themselves as Acreanos and not by their parents' state of origin; inter-marriage between migrants

and non-migrants was common (if not the norm); and, as shall be demonstrated, stereotypes

aside, after controlling for other factors, few distinctions persisted between these migrants and

native-born Acreanos in terms of their land use practices.

Testing the Relationship between Place of Origin and Land Use

The migrants coming to Acre in the 1980's differed from their Acreano neighbors in terms

of livelihood systems and history. Whereas Acreanos had a long tradition of forest-based

livelihoods (e.g., rubber tapping), most migrant livelihoods depended upon mixed subsistence

and market-oriented agriculture, including cattle production. Furthermore, it was almost entirely

native Acreanos who participated in the rubber tapper movement that tied land rights with forest

conservation, often in the form of extractive reserves. While many migrants had in fact been

involved in struggles for land rights, rarely were they tied to forest conservation. Due to these

historical differences, I anticipated a continuing differentiation between the two groups in terms

of land use. In order to test hypotheses 3.1, 3.2, 3.3, and 3.4 (listed in Chapter 2) about the









impacts of place of origin and land use, I first conducted bivariate analysis of the relationship

between place of origin and deforestation, cattle herd size, Brazil nut production, and pasture

burning. As each test involved a bivariate independent variable (Acreano or migrant) and a

continuous dependent variable, I relied upon t-tests.

Acreanos appeared to have a virtually equivalent mean level of deforestation (25.3

hectares) to migrants (22.3 hectares). An independent sample t-test of place of origin and

household deforestation (equal variance assumed) failed to reveal a significant difference

between native-born Acreanos and migrants (t=.360, df=84, p=.720). In contrast, migrants

appeared to have a slightly higher mean cattle herd size (64.5) than migrants (54). As with

deforestation, an independent sample t-test of place and origin and herd size (equal variance

assumed) failed to reveal a significant difference between native-born Acreanos and migrants in

terms of their land use (t=-.423, df=89, p=.673). Regarding Brazil nut production, Acreanos

appeared to have a much larger mean 2004 harvest (52.3 latas) than migrants (8.9 latas). Unlike

deforestation and herd size, an independent sample t-test (equal variance assumed) showed that

Brazil nut production was significantly higher for Acreanos than for migrants (t=3.592, df=83,

p=.001). Finally, regarding pasture burning, migrants appear to bum their pastures with slightly

less frequency (3.47 years) than Acreanos (3.02 years). An independent sample t-test (equal

variance assumed) failed to reveal that the difference was significant (t=-.895, df=78, p=.374).

Based upon bivariate analysis, I rejected hypotheses 3.1, 3.2, and 3.4. Place of origin did not

have a significant impact on deforestation, cattle herd size or use of fire in pastures. However,

supporting hypothesis 3.3, place of origin did appear to be significantly related with Brazil nut

production. Households originating in Acre produced significantly more Brazil nuts than

migrant households.









Modeling Place of Origin and Land Use

In order to further test the relationship between place of origin and Brazil nut production, I

subjected it to a multiple regression analysis, controlling for years of residence and property size.

The model indicated that Brazil nut production was not significantly affected by years of

residence. It was, however, significantly related to property size. Each additional hectare of

property size was associated with an additional .08 latas of Brazil nut production. Of key

concern to hypothesis 3.3, migrants were predicted to harvest 34.5 fewer latas of Brazil nuts than

native-born Acreanos (Table 3-2). I therefore accepted hypothesis 3.3, that households with

place of origin in Acre produce more Brazil nuts than migrant households.

Discussion

The stronger correlation between property size and Brazil nut production is intuitive, given

that larger properties can be expected to have more Brazil nut trees. The fact that, controlling for

covariates, Brazil nut harvests remain significantly related to place of origin lends a further

insight into the relationship between place of origin and land use change. On one hand, native

Acreanos seem to have readily adopted non-forest-based land use practices, such as cattle, as

they became viable in the region. On the other hand, even though Brazil nut production has

recently become a very lucrative activity, non-Acreanos have been less keen than their Acre-born

neighbors to begin exploiting this forest-based resource.

Thus, while the BR-317 has, and still does, affect land use in the region (as will be further

discussed in the next section), the cultural orientation of the migrants who followed the road to

this area appears to play a limited role in this relationship several decades later. Cultural

differences that once were very important in affecting land use between migrants and native

Acreanos seemed to be, with some exceptions, diminishing as a common Acreano identity









formed and all groups responded to shared market forces and governmental policies that have

followed in the highway's wake.

Intra-Regional Migration, Gender and Land Use

While two-thirds (66.3%) of households reported that they had modified their land use in

some way because of the paving of the roadway, all but one (who complained of increased

crime) indicated that their overall quality of life had improved because of the paving of the

road-many referring to it as an abengo (blessing). Travel between their homes and the

regional urban centers of Brasileia and Assis Brasil became much more feasible, putting markets,

schools, churches and medical services more easily within reach. A journey to town that

previously would take up to several days could be completed in as little as an hour, allowing the

majority of residents in the research site to travel to and from town in one day, even in the rainy

season.

This travel to and from town was generally motivated by one or two primary goals: the

economic transactions of selling farm products and purchasing food and other products

unavailable in the colonia, and accessing a myriad of urban services-including banks, hospitals,

government offices, secondary schools, brothels andforr6 dancehalls. Given the trade-offs

between rural and urban residence, the paved highway has facilitated temporary and permanent

micro-level migration between the two as well as an emerging phenomenon of bi-localism as

families split their residence between rural and urban areas. Most families made multiple-day

trips to town on a monthly or greater basis and while most stayed with family members while in

town, 17% of interviewed households maintained a second residence in town.

It is important to recognize that bi-localism generally operates at the intra-household level.

For various reasons, including the cost of transportation, on-farm responsibilities and fears of

break-ins, it was usually individuals rather than entire families who traveled between the farm









and city. For example, if a family needed to make a bank transaction or treat a sick child at the

hospital, one household head would usually go, leaving the other to care for the home and farm.

Similarly, in families with adolescent children and sufficient economic means, most would invest

in an urban secondary education for one or two children while the others remained to provide on-

farm labor.

A telling indicator of gender-based differences in rural-urban movement is reflected in the

numbers of men and women who had assumed longer term in-town residency. On average,

sampled households had sent .23 men and .37 women to the city in the prior four years. Given

the fact that men contributed the largest part of farm labor, the continuation of most on-farm

activities required either the presence of males or sufficient financial resources-such as salaries

or cattle-to contract male labor. Conversely, reproductive activities-including the care of

children and the sick-were dominated by women and were often more easily carried out in an

urban setting.

Girls and young women were more likely to pursue advanced studies in town than boys

and young men. Girls were often perceived as being both more adept at academics and less

prone to becoming side-tracked by the myriad temptations of city life than boys, making

investment in their education less risky than investment in boys' education. Furthermore, in a

growing service-based economy, women often found more opportunities for unskilled and semi-

skilled urban work than did men. And increasingly, in the field of skilled urban labor (generally

synonymous with the governmental sector in Alto Acre), a growing population of educated

women were finding in-roads into positions once dominated by men as evidenced, for example,

in the 2004 election of Brasileia's first female prefeita. And while unintended pregnancy was a

great concern for many parents as they decided whether to send their daughters to study in town









where they would be subject to less direct adult supervision, this risk was often accepted,

especially given the preponderance of the same issue in the rural area as well, even with closer

parental supervision.

Despite the advantages offered by urban residence, the possession of a rural land holding

remained very desirable for most residents, the outright sale of rural land being most common

among the elderly who desired to "cash-in" their accumulated farm investments and by the poor

who did so out of necessity. Many informants boasted of the peacefulness, quiet, healthfulness,

and relative safety from crime and pollution that they enjoyed in the countryside. Furthermore,

while economic opportunities were generally limited to the agricultural sector, these were

considered to be economically lower-risk activities-especially cattle raising-than urban

economic ventures. For wealthier families, possession of a rural land holding provided at least

three services: as a weekend "vacation" home, as a capital investment-especially when it

sustained a cattle herd-and as a hedge against economic failure in the city (e.g., lost job or

failed business venture). In such an event, possession of a rural land holding allowed the owner

to retreat to a relatively secure, if modest, on-farm subsistence.

Natural population growth, coupled with temporary and permanent rural-urban migration,

has contributed to rapid growth of the region's towns. As in many parts of the developing world

(Lambin et al., 2001) and in the Amazon in particular (Browder and Godfrey, 1997), the direct

impact of urbanization upon LULCC by replacing fields and forests with urban sprawl has been

negligible in Alto Acre. Urbanization, however, by changing the region's political, social and

economic landscapes, is a powerful indirect cause (and effect) of land use and land cover change

in rural areas. For example, wage labor, refrigeration and the proximity of meat processing

facilities contribute to a much larger per capital beef consumption in the urban versus rural areas;









hence as towns grow, so too does regional demand and price for beef. Also, the closely related

phenomena of bi-localism and rural exodus that are directly tied with the urbanization of the

region also have impacts upon land use in the countryside. In particular, one would expect that

the part-time and permanent relocation of rural residents to the city would result in shifts from

labor intensive land uses to activities that require less labor.

Testing the Relationship between Bi-localism, Out-Migration and Land Use

Most agricultural and extractivist activities conducted in the field site tended to be very

labor intensive. This was especially true for the production of annual crops and for the

extraction of Brazil nuts. In contrast, cattle ranching, while requiring a larger capital investment

(e.g., livestock, fencing, corrals, vaccinations), was a much less labor intensive activity,

especially once initial start-up labor investments such as the construction of fences and corrals

had been completed.

Hence, I posited that as rural household labor supply diminishes-through both out-right

migration to cities and through bi-localism, that households would reallocate the remaining

household labor from labor-intensive activities such as the production of annuals and Brazil nuts

to less labor intensive activities like cattle ranching. In order to test hypotheses 3.5 through 3.16

(listed in Chapter 2) about the impacts of household labor and bi-localism for land use, I first

conducted bivariate analysis of the relationship between male bi-localism and deforestation,

cattle herd size, Brazil nut production, and pasture burning. I repeated the tests using female bi-

localism as the independent variable. I then conducted bivariate analysis of the impacts of

household labor force for these same dependent variables. In the case of male and female bi-

localism and land use, as each test involved a bivariate independent variable (high bi-localism

and low bi-localism) and a continuous dependent variable, I relied upon t-tests. In the case of









household labor force and land use, as both the independent and dependent variables were

continuous, I used Pearson correlation analysis.

Female bi-localism and land use

Households with high levels of female bi-localism had virtually the same mean area in

hectares in annuals (3.2) as did households with low levels of bi-localism (3.3). An independent

sample t-test (equal variance assumed) did not reveal a significant difference between the two

groups in terms of their cattle herd size (t=.093, df=76, p=.926). Households with high levels of

female bi-localism had lower mean Brazil nut harvests (18 latas) than those with low levels of

female bi-localism (49 latas). An independent sample t-test (equal variance assumed) revealed a

significant difference between the two groups in terms of their Brazil nut production (t=2.274,

df=75, p=.026). Households with high levels of female bi-localism had smaller mean cattle

herds (37.69) than households with low levels of female bi-localism (75.26). An independent

sample t-test (equal variance assumed) revealed a significant difference between the two groups

in terms of their cattle herd size (t=1.345, df=77, p=.045).

Hence, bivariate analysis failed to support hypothesis 3.5 that households with higher

female bi-localism would be associated with less area in annuals. It also failed to support

hypothesis 3.7 that higher female bi-localism would be associated with larger cattle herds.

However, it did support hypothesis 3.6 that higher levels of female bi-localism would be

associated with smaller Brazil nut harvests.

Male bi-localism and land use

Households with high levels of male bi-localism had the same mean area in hectares in

annuals (3.4) as did households with low levels of male bi-localism. An independent sample t-

test (equal variance assumed) did not reveal a significant difference between the two groups in

terms of their cattle herd size (t=.015, df=77, p=.988). Households with high levels of male bi-









localism appeared to have lower mean Brazil nut harvests (24.8 latas) than those with low levels

of male bi-localism (46.5 latas). However, an independent sample t-test (equal variance

assumed) did not reveal a significant difference between the two groups in terms of their Brazil

nut production (t=1.613, df=75, p=.111). Households with high levels of male bi-localism had

smaller mean cattle herds (54.5) than households with low levels of male bi-localism (65.9). An

independent sample t-test (equal variance assumed) did not reveal a significant difference

between the two groups in terms of their cattle herd size (t=.419, df=78, p=.676)

Hence, bivariate analysis failed to support hypothesis 3.8 that high male bi-localism would

be associated with smaller area in annuals. It also failed to support hypothesis 3.9 that high male

bi-localism would be associated with smaller Brazil nut harvests. Neither did bivariate analysis

support hypothesis 3.10 that high male bi-localism would be associated with larger cattle herds.

Adult females and land use

Pearson correlations failed to show a significant relation between adult female adults and

area in annuals and Brazil nut production (Table 3-4). Hence, bivariate analysis failed to support

hypotheses 3.11 and 3.12 that lower female labor would be associated with less area dedicated to

annuals and lower Brazil nut production. A significant relationship did emerge between lower

levels of adult female labor and smaller cattle herds. However, the direction of the relationship

countered hypothesis 3.13 that lower levels of adult female labor would be associated with larger

cattle herds.

Adult males and land use

Pearson correlations showed a positive relationship between number of adult males and

Brazil nut production and cattle herd size. No relationship emerged in the relationship between

male labor and area dedicated to annuals. Hence, bivariate analysis failed to support hypothesis

3.14 that lower levels of male labor would correspond with a smaller area dedicated to annuals.









It did however support hypothesis 3.15 that lower levels of male labor would correspond with

less Brazil nut production. Finally, while adult male labor was significantly associated with

cattle herd size, the direction of the relationship countered hypothesis 3.16 that lower levels of

male labor would correspond with larger herd sizes.

Modeling Bi-localism, Household Composition and Land Use

I tested the first relationship between female bi-localism and Brazil nut production by

using multiple regression to further test these relationships while controlling for other variables

likely to affect land use (Table 3-5, column 1). Property size was significantly associated with

Brazil nut harvest. For every additional hectare of property size, Brazil nut production increased

by .084 latas. Of special concern to hypothesis 3.6, after controlling for years of residence and

property size, Brazil nut production was significantly associated with female bi-localism.

Female bi-localism explained 3.3% of the variation in Brazil nut production. Households with a

high level of female bi-localism produced 22.57 fewer latas of Brazil nuts in 2004 than did

households with a low level of female bi-localism.

I then replicated the test to determine the relationship between female bi-localism and herd

size (Table 3-5, column 2). In this model, the only significant predictor of herd size was years of

residence. For every additional year of residence, herd size increased by 3.392 head. After

controlling for years of residence and property size, the relationship between female bi-localism

and cattle herd size was no longer significant.

To test the relationship between adult males and Brazil nut production and between male

adult males and cattle herd size, I ran multiple regression analyses to control for the influence of

years of residence and property size in both relationships that proved significant in Pearson

correlation analysis. In the first model (Table 3-6, column 1), property size had a significant

impact on Brazil nut production, with each additional hectare corresponding with an increase of









0.104 in Brazil nut production. After controlling for years of residence and property size, the

relationship between adult male labor and Brazil nut production was no longer significant. In the

second model (Table 3-6, column 2), years of residence was the only significant predictor of

cattle herd size. For every additional year of residence, herd size is expected to increase by

2.821 head. In the third model (Table 3-7), as in the second model, years of residence had a

significant impact on herd size. For every additional year of residence, herd size is predicted to

increase by 2.847 head. Of special concern to the question of household composition and land

use, even after controlling for years of residence and property size, the number of adult females

remained significantly associated with larger cattle herds. For every additional adult female

present in the household, cattle herd size was predicted to increase by 34.327 head.

Discussion

Given the labor intensive nature of Brazil nut production and usual involvement of both

sexes in its production, the relationship between female bi-localism with lower production of

Brazil nuts may be due to labor shortfalls. Lower availability of female labor may cause families

to shift away from labor intensive activities such as Brazil nut production. It is also possible

that the relationship is spurious; for instance, greater possession of capital and assets in town

(something not directly measured in the study) might necessitate that women travel more

frequently to town while reducing the necessity of the family to focus on labor intensive farm

activities for subsistence.

The relationship between lower levels of adult female labor and smaller cattle herds

directly contradicts hypothesis 3.13 that lower levels of female, would be related with larger herd

sizes. The finding initially seems counter-intuitive, especially given that the care (though not the

ownership) of cattle, in this region, is an activity almost entirely dominated by men. In this case,

it seems unlikely that labor shortfalls are the primary explanation for the relationship. However,









it is possible that variance in cattle production systems plays a role. Producers generally practice

some combination of two general herd management schemes-selling off calves or keeping

them-especially salient in the case of female calves. One strategy is to sell off most or all

offspring-practiced both due to lack of additional pasture and desire for short-term income-

representing a subsistence-oriented herd management strategy in which both the size of the herd

and annual revenues from calf sales remain roughly constant. The other system, of keeping

heifers with reproductive potential allows, under certain conditions (e.g., unconstrained forage

availability), for exponential growth in herd size and, by extension, accumulated wealth in

exchange for lower short-term income. Furthermore, under such a system, the sale of the

majority of male calves not intended for reproduction can serve as a "dividend" payment while

having a minimal impact on the long-term reproductive potential of the herd.

While I did not measure the phenomenon directly, in the course of interviews, subtle

conflicts emerged between men and women in terms of their approach to cattle production.

Specifically, women seemed most likely to show caution in the management of their herds, being

careful to build herd size as a resource for family emergencies; men were often more prone to

"cash-in" portions of their herds to purchase vehicles and other consumer goods. For instance,

one elderly widower told me that his late wife had always preferred, when possible, to keep

calves rather than selling them. Upon her death, he told me, the herd had diminished as he

gradually sold it off. In another instance, when I asked a female informant how the family

(which seemed to be of modest means) had afforded to purchase a car, she replied that her

husband "Perdeu a cabeqa" ("lost his head") and sold off a large part of the family's herd to buy

the vehicle. This conclusion was corroborated by feedback from numerous community members

(though more often by women than by men) when I returned research results in 2006. Such









anecdotal evidence, combined with statistical analysis, suggests that households with more adult

females (and female decision-makers) present will tend to have the most rapidly expanding

herds. Hence, the disproportional tendency of women to migrate, permanently or temporarily,

may be having the effect of slowing the expansion of cattle herds in the region's rural areas.

Access to Roadway and Land Use

Two important dates stand out regarding the construction of the BR-317 roadway between

Brasileia and Assis Brasil. First was the construction of a dirt road in the mid-1970's which

launched an initial wave of in-migration, land use change and deforestation. Another milestone

occurred in 2002, when the long-awaited paving of this highway became a reality, and it became

possible to travel from Brazil's heartland through Alto Acre, to the Peruvian border without

leaving asphalt. While this had been a perennial campaign promise from numerous politicians, it

was ultimately realized by Governor Jorge Viana, thus ingratiating him to the majority of Alto

Acre residents, including many who were otherwise opposed to his Worker's Party.

With the scheduled paving of the Peruvian roadway stretching from Brazil to the Pacific

coast, much speculation has been offered about the likely economic and environmental

implications of this future asphalt link between Brazil's heartland, the Brazilian and Peruvian

Amazon, the populous Peruvian Andes, the Pacific and, ultimately, China and other rapidly

growing Asian economies. This prospect has raised fears of devastating social and ecological

changes, most notably an increased marketability of soy from the Amazon's southern fringes

(and resulting deforestation) in that region.

While the implications for Acre and the western Brazilian Amazon will undoubtedly be

considerable, the extent of larger-ranging ramifications in Brazil is doubtful. In the case of soy,

due to the distance involved and the formidable barrier posed by the Peruvian Andes-even on a

paved road-it is unlikely that exporters will prefer this route over shorter existing routes from









Central Brasil through Bolivia and Chile or cheaper water routes through the Amazon and

Atlantic Ocean. During a 2005 discussion with University of Florida students and faculty,

Governor Viana himself downplayed the wider-scale national and international consequences of

the highway-stressing instead that the major implications of paving to the Peruvian border and

beyond have occurred and will continue to occur at the regional level of Acre, the Bolivian state

of Pando and the Peruvian state of Madre de Dios.

As of 2005, Bolivia-especially the state of Pando and its capital Cobija-had become an

important part of the social and economic landscape for many residents of Alto Acre, including

my research site. For example, most of the Brazil nuts in the field site were sold to a Bolivian

firm, and clothes, electronics and gasoline used by interviewed households commonly came from

Cobija; many informants also spoke of family members who had gone to Bolivia searching for

jobs or land. With the completion of the paved highway between the field site, Brasileia, and the

neighboring city of Cobija, the city's free trade area had become an important source for

relatively inexpensive Chinese-made consumer goods not only for Acre's economic elite, but for

all but the poorest of the field site's rural residents. Additionally, for those with cattle or other

liquid assets, Cobij a's private hospital provided health services that, while expensive, were

universally considered superior to Brasileia's free public hospital.

Furthermore, the availability of inexpensive nationally-subsidized gasoline-

approximately one half of Brazilian prices-offered a considerable reduction in operating costs

to tractor and chainsaw operators and especially to those with automobiles, including the owners

of trucks running transportation routes between the region's rural and urban areas. In effect, as

in other Brazilian regions bordering Bolivia, the economic development of Alto Acre has been

able to free-ride upon Bolivia's politically-popular gasoline subsidies.









Additionally, the combination of improved transportation links between Brazil and Bolivia

and the growth of governance capacity in the former relative to the latter led Bolivia to be seen

as a new frontier for many Brazilians. Brazilians who had been displaced by the closing frontier

and rising land prices on the Brazilian side could, in exchange for compromised economic and

physical security, find opportunities unavailable in Brazil. Brazilian ranchers and loggers

frustrated with increased regulations in their home country had begun to view Bolivia as a low-

cost and lightly regulated arena in which to expand operations. Likewise, many landless

Brazilians had found that Bolivia offered an opportunity to find wage labor and land. However,

this came at a risk as no legal recourse existed for Brazilians in unsafe labor conditions and,

without legal tenure, any investments in land were subject to confiscation. It is possible, though

still unclear, that the recent election of Evo Morales of the Movement to Socialism Party as

Bolivia's president will change the generally lawless atmosphere that had prevailed in this region

of Bolivia previously.

As of the time of research, Peru had been less important than Bolivia in terms of its social

and economic impacts upon the people of Alto Acre. However, with the completion of a

permanent bridge between the two countries in late 2005, this could quickly change. Despite

considerable population in the neighboring Madre de Dios province of Peru, little of the state's

area had been converted to pasture, making the state a likely purchaser of Acre beef upon

completion of the bridge and import clearance. Similarly, should the paving of an asphalt link

between Acre and Peru's Pacific coast be achieved, the implications for Acre will likely be

considerable. While it is doubtful that the Brazilian soy industry, concentrated in the nation's

center, will find the Inter-Oceanic Highway to be a profitable export route, the road may,

however, provide an incentive for soy production in Acre itself. Likewise, rising demand might









also occur for products from Acre's forestry and cattle sectors with increased connectivity to

East Asia's expanding markets.

As of the mid-2000's, highway paving had led to increased market integration and

increased marketability for relatively high value and/or easily transportable farm products

including cattle, coffee, bananas and some NTFPs such as Brazil nuts. However, due to their

low value and the difficulty of transportation, most annual crops such as rice, cassava and beans

had been little affected by highway paving, as most of production was dedicated to household

subsistence. While most families reported selling surplus crop production, many claimed that

this tended to consist of relatively small quantities-as the majority of crops produced was

dedicated to subsistence-even with improved road access. Various farmers told me that the

paved highway had, in fact, hurt the profitability of some crops such as rice and beans as cheaper

produce from mechanized farms in Brazil's South began to flood the region's markets.

Despite the recent paving of the road, the tide of migrants to the state of Acre has slowed,

as less unoccupied land is now available for new colonization projects. Most of the region's land

is currently occupied by extractive reserves, colonization areas, indigenous reserves or titled

private property. In the occasional event of INCRA redistribution of unoccupied or expropriated

land, given the large number of landless families relative to available land, new lots tend to be

much smaller than in the past-tending to range from 5-30 hectares. However, the newly-paved

road does provide improved access to markets for families already living in the area. Roads

often serve as drivers of land use change by facilitating access to markets and promoting market-

oriented agricultural activities that might not be viable in more distant locations (Chomitz and

Gray, 1996). However, the benefits of road paving are not necessarily equal for all of the

residents living near it. Due to the poor state of many of the region's secondary roads, some









households remain fairly isolated from regional population centers and markets, especially

during the rainy season.

Testing the Relationship between Road Access and Land Use

One means of measuring the impacts of the highway upon land use is by looking at the

relationship between a household's access to the roadway and its land use decisions, while

controlling for other variables. Various techniques can be used to measure access to a roadway.

Possibilities include measuring the linear distance between the home and the roadway and

measuring the distance in terms of time required to travel from the home to the roadway. Due to

the differences between paths and roads in terms of their quality and due to the fact that on-the-

ground travel distances are often longer than linear distances, I chose the latter method. As area

residents tend to rely on various forms of transportation at different times-including foot,

bicycle, horses/mules/oxen, motorcycle and truck, I standardized the time-distance measurement

by obtaining walking time-distance from each household.

Bivariate analysis (Table 3-8) revealed that distance from the road showed no relationship

with deforestation (r=-.073; p=.506) and with herd size (r=-.104; p= .325). Even after

controlling for other variables-namely years of residence, years of cattle ownership, legal

deforestation limit and property size-no relationships emerged that could be clearly

differentiated from chance. Hence, bivariate analysis failed to support hypotheses 3.17 and 3.18

that increased distances from the road would correspond with lower deforestation and lower herd

sizes.

Discussion

In order to avoid overstating the lack of relationship between distance from the road and

deforestation and cattle, it is important to note that all sampled households lie within five hours

walking distance of a paved road. It is hence impossible, based upon the data, to draw a









conclusion regarding the relationship between distance to a roadway and deforestation for the

numerous families living even further from the roadway and outside the scope of this study. A

possible explanation for the lack of a relationship between distance to the road and deforestation

and with cattle herd size may be due to the fact that all households could feasibly travel to and

from urban market centers in one day if possible. This conclusion is supported by a study

conducted in Columbia (Feaster, 1970) that found that distance from markets only significantly

affected the land use among households where it was not possible to and from the market in a

single day. Due to the difficulties imposed by spending a night away from home, such families

enjoyed much less access to markets than those who could visit and return in the same day.

As seen in Figure 1-2, in the Alto Acre region as a whole, deforestation is most heavily

concentrated in areas nearest the BR-317 while forest cover tends to be more intact in more

remote areas. This conforms to various studies such as Moran (1982) who found that

deforestation had tended to follow road corridors, especially in areas that have undergone

colonization. While households living nearest the roadway certainly enjoyed easier market

access for farm goods, including cattle, the distances at the scale of this study did not appear to

be prohibitive in the production of cattle. Given the relative abundance of trails and secondary

(unpaved) roads in my research site and the fact that cattle, unlike most farm products, are able

to transport themselves even during the rainy season, most if not all families could reasonably

expect to transport cattle from farm to the edge of the roadway (which provided all-season access

for cattle trucks) within a day or less. Though not directly evaluated in the study, it is likely that

the opportunity costs of cattle ranching for families living several days or more into the forest

would manifest in lower herd sizes and less area dedicated to pasture.









Cattle and Land Use

In the Brazilian Amazon, cattle ranching has gained notoriety for replacing ecologically

important forest with low-biodiversity grassland. Many researchers point to cattle ranching as a

primary cause behind deforestation and social conflict in the Brazilian Amazon (Hecht, 1985;

Cardoso, 2002). Others, such as Faminow (1998) see a more complex relationship between

cattle and environmental degradation, with cattle production, in and of itself, being less

important than the specific types of production used.

It was clear, through both interviews and direct observation over the course of a year, that

many households' herds were rapidly increasing. While I did not systematically measure the

change, on return visits to households both during fieldwork and when returning research results

one year later, most households mentioned that their herds had increased since the time of the

initial interview. Given the tendency toward herd expansion that was occurring in the region,

especially given the facilitated market access that the paved road provided, I tested the

relationship between herd size and other land use practices, particularly the production of

annuals, Brazil nut harvesting, and the use of fire as a pasture maintenance mechanism.

Testing the Relationship between Herd Size and Other Land Use Practices

The strong correlation between herd size and deforestation seen in Table 3-8 is intuitive-

larger herds require larger areas of pasture to survive. However, the relationship between cattle

herd size and other land use practices is less obvious. In order to better understand the

relationship between the size of a household's cattle herd and other land uses, I tested the

relationship between herd size and the production of annuals and Brazil nuts3. I also tested the


3 The relationship between cattle and Brazil nuts and annuals should be interpreted with caution. Based on
qualitative observations of land use and of herd dynamics, it is arguable that cattle herd size functions as a driver of
changes in other land uses such as annuals and Brazil nuts. For example, I found numerous cases in which Brazil
nut production had been decimated as herds, and pasture expanded. I found no cases in which herd sizes had
expanded as households reallocated pasture land to Brazil nuts. With very rare exception, the same can be said for









relationship between herd size and use of fire as a pasture maintenance tool-a practice that can

be detrimental to neighboring forest, crops and to the pasture itself (Cochrane, 1998, Nepstad et

al.,, 2001, Almeida, n.d.).

As households have a finite supply of land and labor, I expected that cattle production

would result in an inevitable trade-off with other land uses and that larger herd sizes would

correspond with less production of annuals and Brazil nuts. I also expected that, due to their

lower relative dependence on crops and Brazil nuts, owners of large cattle herds would see

wildfire as a minimal risk to their livelihoods, consequently burning their pastures more

frequently than families with smaller herds.

In order to test hypotheses 3.19, 3.20 and 3.21, about the impacts of herd size for annual

production, Brazil nut production and use of fire, I first conducted bivariate analysis of the

relationship between herd size and the three dependent variables. As each test involved

continuous independent and dependent variables, in each case, I used Pearson correlations (Table

3-9).

Bivariate analysis revealed a significant relationship between cattle herd size and area

dedicated to annuals (r=.286, p=.007). However, the direction of the relationship contradicted

that which was posited in hypothesis 3.19 that larger herds would correspond with less area in

annuals. This suggested that suggesting that, rather than presenting competing land uses, cattle

ranching and annual production could be complementary. Bivariate analysis did not reveal a

significant relationship between cattle herd size and Brazil nut production. Hence, the analysis

failed to support hypothesis 3.20 that larger herds would be related with less Brazil nut

production. Finally, bivariate analysis revealed a significant positive relationship between size

annuals. However, as both the independent and dependent variables are land uses, the possibility for endogeneity
exists.









of cattle herd and the number of years since pasture was last burned (r=.322, p=.004). As the

relationship was positive rather than negative, it ran contrary to expectations stated in hypothesis

3.21. I had expected that households with larger herds would be less concerned about the

impacts of wildfire upon other land uses and therefore use fire more freely than those with

smaller herds. Instead, I found that households with larger herds had passed more years since the

last time they had used fire in their pastures.

Modeling Herd Size and Brazil Nut Production, Area in Annuals, and Use of Fire in
Pastures

In multiple regression models, I further tested the relationships between cattle herd size

and area in annuals and between cattle herd size and years since using fire in pasture. In both

cases, I controlled for years of residence and property size (Table 3-10).

After controlling for the effect of years of residence and property size, a small, though

significant, relationship did emerge between herd size and area in annuals. According to the

model, herd size explains 5% of the variation in area dedicated to annuals. For every positive

increment in the number of cattle owned, area in annuals increases by .007 hectares.

As previously mentioned, due to the possibility of endogeneity between the two land use

variables, it is difficult to ascertain the direction of causality between herd size and area in

annuals. However, the model does, in the least, suggest complementarity between the two

variables. Countering hypothesis 3.19 that larger herds would be associated with smaller area

dedicated to annuals, from this model we can conclude that cattle production does not, generally,

represent a competing land use with annual production, and may possibly be a complementary

activity. This conclusion conforms with interview data suggesting that production of annuals

subsidizes cattle production, and vice versa. On one hand, due to the low market value of most

annual crops, it is difficult for a household to recoup the large monetary and labor investments