Agricultural Expansion Alongside Productive Forests and Parks in the Southern Peruvian Amazon

Material Information

Agricultural Expansion Alongside Productive Forests and Parks in the Southern Peruvian Amazon
Rojas, Rafael O
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (155 p.)

Thesis/Dissertation Information

Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Interdisciplinary Ecology
Committee Chair:
Perz, Stephen George
Committee Members:
Southworth, Jane
Deere, Carmen D
Schmink, Marianne C
Kainer, Karen A
Graduation Date:


Subjects / Keywords:
Agricultural land ( jstor )
Agricultural land use ( jstor )
Agriculture ( jstor )
Community associations ( jstor )
Deforestation ( jstor )
Environmental conservation ( jstor )
Forests ( jstor )
Land cover ( jstor )
Land use ( jstor )
Social capital ( jstor )
Interdisciplinary Ecology -- Dissertations, Academic -- UF
agriculture -- amazon -- deforestation
City of Gainesville ( local )
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Interdisciplinary Ecology thesis, Ph.D.


Current studies of agriculture in the tropical forest usually consider the forest as a uniform entity and reduce the farmers’ role to one of only clearing forest. Agricultural land use by landholders in the Southern Peruvian Amazon calls into question this view, because the forest in this region is comprised of important non-timber forest products and protected areas, both of which influence land and forest use decisions. In addition, agricultural land use not only reflects decisions by individual farm families but also strong farmers’ organizations, which play a crucial role in defining how both cleared land and forests are used. This dissertation proposes a multi-scale approach that not only incorporates decision-making by individual landholders but also the larger sociopolitical context, particularly in terms of the influence of farmer organizations and the presence of protected areas and NTFPs in forests on and near farm properties. To pursue an analysis of agricultural expansion in its broader context in the Southern Peruvian Amazon a three-part method will be used. First, multivariate statistical models of agricultural and forest resource use will be developed in order to better understand differences in these key components of resource management among individual families. Second, an institutional and political analysis will assess the role of farmers' organizations and other social actors in defining sustainable forest use, and therefore on agricultural expansion, through the establishment of alliances with the government and other social actors. Third, land-use/land-cover analysis will be employed to evaluate the spatio-temporal dynamics of forest change in order to understand the relationship of agricultural expansion with the national economic context of rapid or slow growth. This approach will also allow the assessment of current proposals, such as productive conservation for conservation and development. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (Ph.D.)--University of Florida, 2013.
Adviser: Perz, Stephen George.
Electronic Access:
Statement of Responsibility:
by Rafael O Rojas.

Record Information

Source Institution:
Rights Management:
Applicable rights reserved.
Embargo Date:
Resource Identifier:
885021045 ( OCLC )
LD1780 2013 ( lcc )


This item has the following downloads:

Full Text




2 201 3 Rafael Oswaldo Rojas


3 To my parents


4 ACKNOWLEDGMENTS I would like to thank local people, rural leaders, and friends in Ma dre de Dios that helped on the fieldwork for this dissertation. And I should specially acknowledge my friends on Conservation International Peru at Puerto Maldonado. The financial support from Compton Foundation was also important for this dissertation. Fi nally, I have to mention the insights, support and patience of Stephen Perz that were the most crucial factors to finish this dissertation.


5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 2 AGRICULTURAL LAND USE AND DIVERSIFICATION ................................ ......... 16 Theoretical Frameworks for the Analysis of Land Use in Frontier Regions ............. 19 Non timber Forest Products and Forest Extractivism ................................ ....... 19 Agricultural Land Use ................................ ................................ ....................... 21 The Livelihood Strategies Approach: Capi tals and Capabilities ....................... 25 A Changing Frontier in the Southeastern Peruvian Amazon ................................ ... 30 Analyzing Land Uses Interaction under Transp ortation Infrastructure Improvement ................................ ................................ ................................ ........ 35 Methods of Data Collection ................................ ................................ .............. 35 Data for Analysis ................................ ................................ .............................. 37 Analytical Techniques ................................ ................................ ...................... 43 Findings ................................ ................................ ................................ ............ 44 Discussion ................................ ................................ ................................ .............. 49 3 INTERMEDIATE AND DISTANT DRIVERS AND LAND COVER CHANGE IN THE SOUTHEASTERN PERUVIAN AMAZON ................................ ....................... 58 Explanations for Deforestation in Frontier Areas ................................ .................... 62 Drivers and Dynamics of Land Cover Change in the Southeastern Peruvian Amazon ................................ ................................ ................................ ............... 69 Data and Methods for Analysis of Land cover Change Trajectories ....................... 75 Results ................................ ................................ ................................ .................... 80 Discussion ................................ ................................ ................................ .............. 83 4 SOCIAL MOVEMENT ORGANIZATIONS AND THE UNSUSTAINA BILITY OF SUSTAINABLE INITIATIVES ................................ ................................ .................. 92 Social Organizations and Sustainable Development ................................ .............. 96 Social Capital ................................ ................................ ................................ ... 98 Political Ecology ................................ ................................ ............................. 100


6 The Emergence of Grassroots Organizations in the Southeastern Peruvian Amazon ................................ ................................ ................................ ............. 103 Research Methods ................................ ................................ ................................ 107 The Emerging of a Pro ................................ 111 Sustainable Development I nitiatives ................................ ................................ ..... 117 The Sustainability of Social Movement Organizations ................................ .......... 122 Enablers and Constraints ................................ ................................ ...................... 130 5 CONCLUSIONS ................................ ................................ ................................ ... 135 Summary of Key Findings ................................ ................................ ..................... 135 Implications for Policies and Future Research ................................ ...................... 138 LIST OF REFERENCES ................................ ................................ ............................. 141 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 155


7 LIST OF TABLES Table page 2 1 Descriptive Statistics and Correlations with Land Uses ................................ ...... 55 2 2 Descriptive Statistics for Explanatory Variables a nd Correlations with Land Uses ................................ ................................ ................................ ................... 56 2 3 Three Stage Least Square Model for Land Uses ................................ ................ 57 3 1 Population growth in the Madre de Dios region ................................ .................. 88 3 2 Land cover change trajectories in the Madre de Dios region ............................. 90


8 LIST OF FIGURES Figure page 2 1 The Madre de Dios region ................................ ................................ .................. 54 3 1 Evolu tion of Peruvian Castaa exports ................................ ............................... 88 3 2 Evolution of castaa price in Peru ................................ ................................ ...... 89 3 3 Evolution of gold price at wor ld mar ket ................................ ............................... 89 3 4 Buffers for land cover change trajectories for Madre de Dios ............................. 91


9 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 AGRICULTURAL EXPANSION ALONGSIDE PRODUCTIVE FORESTS IN THE SOUTHER N PERUVIAN AMAZON By Rafael Oswaldo Rojas Ma y 201 3 Chair: Stephen G. Perz Major: Interdisciplinary Ecology There is increasing recognition of the importance of forest products in agricultural frontiers; grassroots organizations have emerged an d offer their own proposals for sustainable development of tropical forests; and a new generation of large scale infrastructure projects is going forward as prices of many commodities from tropical regions are rising. This dissertation takes up these issue s by focusing on southeastern Peru. The first part takes up the question of how non timber forest products (NTFPs) may influence agricultural land use. I sampled farms with and without access to forest concessions with NTFPs. I present the findings from m ultivariate statistical models of land use on agricultural properties, highlighting the effects of access to forest concessions. The findings show that households with access to castaa concessions have considerably more pasture than households without con cessions. This finding suggests a positive synergy between NTFPs and agricultural land use, which runs contrary to much previous work on this issue. Second, I focus on land cover trajectories in Madre de Dios over an extended period of time. I employ mult i step and multivariate theoretical frameworks from land


10 change science to highlight distant and intermediate determinants of land cover change. I draw on a time series of satellite imagery spanning the period from 1986 to 2010, during which the driving fo rces behind land cover change in Madre de Dios all exhibited substantial alterations. The findings show strong non linearities over time and substantial spatial differences across the study region. Finally, the third part of the analysis focuses on the eme rgence of grassroots organizations and the question of what factors affect their effectiveness in promoting sustainable development. I focus on a farmer federation that aggregated various community associations and engaged government, non governmental orga nizations and other social actors. I employ a two tier theoretical framework that highlights social capital within the federation and the broader political ecology that constituted the eness. The findings show that while the federation was initially very effective, it has more recently experienced difficulties. And, both social capital and political ecology are important to understanding federation effectiveness, and the lack thereof.


11 CHAPTER 1 INTRODUCTION Tropical frontier regions are changing very rapidly, with the consequence that the processes surrounding land cover change there are very different from those seen in the past. In particular, there is increasing recognition of the importance of forest products in agricultural frontiers; grassroots organizations have emerged and offer their own proposals for sustainable development of tropical forests; and a new generation of large scale infrastructure projects is going forward as pr ices of many commodities from tropical regions are rising. This dissertation takes up these issues by focusing on southeastern Peru, specifically the Region of Madre de Dios, which has become highly dynamic in recent years and is undergoing rapid change. M adre de Dios is thus a useful study case for understanding the shifting context of tropical land cover change. The complexities of agricultural expansion are clearly showed in the Southern Peruvian Amazon, which is subject to increasing levels of deforesta tion due to agriculture. Although to lesser extent in comparison with some other areas, this region has been undergoing permanent in migration during the last 50 years. However, this region is currently on the verge of experiencing many economic, social, a nd environmental changes due to the paving of the Inter Oceanic Highway, which will with ports on the Pacific Ocean. It is therefore important to understand the proc ess of agricultural expansion in order to elaborate proposals that could provide a sustainable future for the region. A key analytical strategy I pursue is to recognize the importance of factors operating on more than one scale. I will therefore discuss b oth macro and micro economic approaches to understanding agricultural change. But in addition, I will


12 also focus on meso level factors, namely community/farming associations, by drawing on social capital and political ecology. The dissertation is organiz ed in three sections that present the findings on the analysis conducted on these three levels of scale. The first level of analysis addresses the agricultural land use at household landholdings. I adopt a focus on household livelihoods and draw on theoret ical models that relate household capitals and capabilities to livelihood strategies. In particular, I use this framework to take up the question of whether and how non timber forest products (NTFPs) and forest extractivism may affect agricultural activiti es. Since forest extractivism requires standing forest and agriculture requires deforestation, the two types of activities have often been viewed as existing in tension, and have at times become flashpoints for social conflict. In Madre de Dios however, so me rural households have access to lands with different tenure rules, and can engage in forest extractivism in forest concessions along with agriculture on agricultural lands. I therefore focus on the question of the effects of access to forest concessions for NTFPs on agricultural land use, and ask whether NTFPs may in fact reduce the extent of agricultural areas held by the same households. The analysis draws on a household survey in southeastern Peru, where there are both forest concessions containing ca staa, an important NTFP, as well as an expanding agricultural frontier. The findings indicate that access to forest concessions results in larger areas of cattle pasture in agricultural landholdings. The findings imply that there are not necessarily trade offs between NTFPs and agriculture, and indeed to the contrary that NTFP income may encourage agricultural expansion.


13 The second part of the analysis focuses on the regional level of scale, in the broader context of national and international processes. I focus on land cover change trajectories in Madre de Dios over an extended period from 1986 to 2010, when the region incurred sweeping changes. I draw on theoretical frameworks from land change science that highlight multi step causation and multiple causat ion in explaining spatial and temporal variation in land cover change. Given the regional focus of the analysis, I highlight distant and intermediate drivers on cover change in the Peruvian Amazon, notably biophysical conditions, government policies, commo dity prices, new infrastructure, and land tenure rules. I draw on a satellite time series to observe different land cover trajectories in order to pursue a temporal analysis of land cover across multiple periods that differ in terms of the distant and prox imate determinants. The trajectories reveal substantial non linearities among the periods in the satellite time series, and show that new infrastructure and rising commodity prices are key explanations for the recent acceleration of land cover change in Ma dre de Dios. The analysis also differentiates among sub regions within Madre de Dios that differ in terms of the intermediate factors. Trajectories of the three sub regions differ, and can be accounted for in terms of proximity to markets, biophysical cond itions and land tenure rules. The presence of important areas containing non timber forest products as well as central regional market influence the pace and type of land cover change. All three sub regions exhibit accelerating land cover change, which rai ses serious questions and the sustainability of development in Madre de Dios going forward. The third and final level of the analysis combines the local and regional. I focus on the case of a farmer federation in the Southeastern Peruvian Amazon, which


14 eme rged out of a collection of local community associations. The federation became an important voice for proposals for conservation and development initiatives in Madre de Dios, and thus a promoter of sustainable development. The federation experienced a bre akthrough beginning and a period of strong conservation initiatives, followed by period of organizational breakdown and inactivity. In the broader context of work on bottom up approaches to sustainable development, I ask questions about the conditions that affect the effectiveness of grassroots organizations in pursuing conservation and development initiatives. I present a two tier theoretical framework that highlights both internal and external factors with respect to grassroots organizations as explanatio ns for their effectiveness. I deploy the concept of social capital to understand the internal capacities of grassroots organizations for collective action; and I draw on political ecology to account for the broader socio political context of a regional far mer federation. I use this framework to pursue a dynamic assessment of the interplay of internal and external factors to grassroots organizations to understand their evolution as manifest in their changing effectiveness in the pursuit of initiatives for su stainable development. The change from the first to the second period can be explained by both internal and external factors to the federation. The findings show that success of conservation and development initiatives by grassroots organizations requires a favorable conjunction of internal and external conditions, which carries implications for grassroots mobilization and conservation initiatives. The dissertation concludes with a final chapter that summarizes the findings and offers synthetic observations concerning their implications for several scholarly literatures. I address implications for the relationship of NTFPs and agriculture, the role


15 of grassroots organizations in bottom up approaches to sustainable development, and distant and intermediate dr iving forces behind land change.


16 CHAPTER 2 AGRICULTURAL LAND USE AND DIVERSIFICATION The land science literature emphasizes the importance of land tenure for land use and land cover change. Land tenure rules define possible land uses for a given social actor. At the same time, the livelihoods literature emphasizes that rural households engage in multiple productive activities. In this light, the land science literature recognizes tenure diversity. Less attention has however gone to landholders wit h access to lands with different use rules, and the consequent interactions of use of one type of land on use of another land type. In particular, there are by now large literatures on non timber forest products (NTFPs) as well as agricultural land use in developing regions. While there has been debate about the economic viability of NTFPs and emphasis on the economic importance of agricultural activities, there has been less attention to whether access to lands for NTFPs might reduce reliance on agricultur e. I therefore evaluate the effects of access to forest concessions on agricultural land use, asking whether NTFPs in fact reduce the extent of agricultural areas held by the same households. The analysis draws on a household survey in southeastern Peru, w here there are both forest concessions containing castaa, an important NTFP, as well as an expanding agricultural frontier. I model the effects of access to castaa concessions on agricultural land use. The findings indicate that access to forest concess ions results in larger areas of cattle pasture in agricultural landholdings. The findings imply that there are not necessarily tradeoffs between NTFPs and agriculture, and indeed to the contrary that NTFP income may encourage agricultural expansion. Land science literature emphasizes that causation behind land use and land cover change (LULCC) is complicated, involving many different factors (e. g., Geist &


17 Lambin, 2 006 ; Wood & Porro, 2002 ) Prominent explanatory frameworks highlight the distinction between proximate, intermediate and distant causes of LULCC as a means of sorting out the order in which processes gene rate LULCC over time and across space. One key underlying assumption in such frameworks is that proximate causation operates on a single land parcel, or at least parcels that are close to each other. This is a simplifying assumption that works in contexts where landholders hold one parcel or one type of land. At the same time, the livelihoods literature emphasizes the multifaceted nature of land use and other activities in households in developing regions (e.g. Ellis, 2000 ; Netting, 1993 ) In particular, diversified livelihoods may rely on different kinds of lands for different kinds of economic activities. This raises issues concerning diversity in land tenure as it influences livelihoo d activities, and how land use practices on different lands influence each other. Land science frameworks also recognize the importance of institutions as explanations for land use ( Ostrom, 1990 ) and diversity as a factor for land use diversity ( Ostrom, 2005 ) In many regions, land tenure is diversified insofar as different lands have different rules. As a result, the same social actor may have access to two or more types of lands and thus pursue different types of resource management. T enure diversity thus interacts with livelihood diversity to produce varying land use in different kinds of lands. In this case, because a single landholder controls multiple parcels, decisions made about one type of parcel may be influenced by or jointly t aken with decisions about another type of parcel. This situation applies to the case of households with access to both lands intended for forest extractivism and lands for agriculture. There are large literatures on extractivism and especially agriculture as key livelihood


18 activities in developing regions. By contrast, there are fewer treatments of the interactions and occasional conflicts between forest extractivism and agricultural land use. Previous work suggests that forest conservation for NTFPs confli cts with forest clearing for agriculture. However, in the Amazon, conflicts over natural resources have yielded a somewhat different situation where tenure diversity has emerged, as governments have variously recognized different kinds of land claims, and/ or engaged in zoning plans that in turn recognize different aptitudes for different lands. Under such circumstances, some landholders acquire access rights to multiple types of lands, and simultaneously use forest lands for NTFPs as well as for agriculture This leaves open questions of whether and how access to lands for NTFPs influences use of lands for agriculture. This paper takes up these issues by pursuing an analysis of the effect of access to lands designated for forest extractivism on agricultural land use. To pursue this analysis, I take up the case of southeastern Peru, specifically the department of Madre de Dios in the Amazonian lowlands. There, the Inter Oceanic Highway was recently paved through an area that has both forest concessions for ext raction of non timber forest products (NTFPs) as well as private landholdings for agricultural use. In this region, households may hold both a forest concession as well as an agricultural property. However, not all households with agricultural properties h ave access to forest concessions. This circumstance permits a comparison of agricultural land use among families at varying locations along the highway, where some have forest concessions while others do not. I therefore compare agricultural land use among families with and without forest concessions to see if access to concessions affects agriculture while controlling for other factors, notably market access in terms of distance


19 to markets. The analysis involves multivariate statistical models for a sample of households in MDD in which access to a castaa concession is a key explanatory variable for various agricultural land uses, and the findings for forest concessions on land use are strong. The results bear implications for theory on land use and livelih ood diversity as well as public policies concerning land tenure, infrastructure planning and regional development. Theoretical Frameworks for the Analysis of Land Use in Frontier Regions I begin with a review of issues concerning forest extractivism and n on timber forest products (NTFPs). I then discuss agricultural land use and cattle ranching, as well as their interactions with forest extractivism. I then bring these literatures together via a discussion of the effects of improvements in transportation infrastructure and land tenure regimes, and how the two in turn influence household decision making on resource management, including forest and non forest land use. N on timber F orest P roduct s and Forest Extractivism In the 1980s, NTFPs became the topic of debate over the rational use of tropical forestlands. High profile publications purported that NTFPs from standing forests in the Amazon could generate greater economic value than logging or agriculture ( Peters, Gentry, & Mendelson, 1989 ; Vasquez & Gentry, 1989 ) These findings highlighted NTFPs as a key element in reconciling forest conservation and sustainable development. NTFPs harvested from intact forests would enhance forest value by providing incomes for loc al populations while also encouraging forest conservation ( Arnold & Ruiz Perez, 2001 ; Assies, 1997 ) As a result, researchers proposed forest extractivism as the proper model f or sustainable development in the Amazon ( Anderson, 1990 )


20 In parallel, local peoples and their supporters articulated an alternative model to deforestation for agriculture as the mode for land occupation in the Amazon, drawing on tradition al practices of forest extractivism. In the Brazilian Amazon, supporters of rubber tappers proposed the extractive reserve model ( Allegretti, 1990 ; Schwartzman, 1992 ) In the Peruvian Amazon, the government recognized forest concessions for NTFP harvesting. In both cases, there was official re cognition of a new land tenure type that was based on traditional forest extractivism and exhibited different access and use rules from lands intended for agricultural production, including prohibitions or limitations on deforestation ( Ankersen & Barnes, 2004 ) More recent research has called into question the earlier findings about the economic value of NTFPs. Some researchers argued that the economic benefits from NTFPs were much less than previously alleged ( Sheil & Wunder, 2002 ) Others have noted that infrastructure, technical support, stable market demand, and well defined property rights among yet other factors are required for the success of NTFPs, but are o ften lacking in forest frontiers ( Belcher, Ruiz Perez, & Achdiawan, 2005 ; Ruiz Perez et al., 2004 ) Addi tionally, researchers have observed that the role of NTFPs also depends on household characteristics; households with more productive assets will have more possibilities of obtaining a bigger income share from NTFPs through commercialization ( Pyhala, Brown & Adger, 2006 ) Despite the shortcomings, NTFPs are unique resources that grant economic Southeastern Peruvian amazon, which enjoy most of the requirements f or viable NTFP markets noted above. Castaa has a well established international market ( Collinson,


21 Burnett, & Agreda, 2000 ) and the southern Peruvian Amazon has gained improved connectivity to markets via the paving of the Inter Oceanic Highway under the In itiative for the Integration of Regional Infrastructure in South American ( Santa Gadea & Wagner, 2002 ) encompasses access and use rights for har vesting of castaa in a sustainable fashion, with benefits from harvest sales accruing to the concessionaire ( Pea, 2010 ) There are also initiatives such as certification that seek to provide technical support for harvesting and processing in order to improve product quality for export. This favorable scenario highlights the viabi lity of at least some NTFPs. However, the case of forest concessions for castaa in southeastern Peru raises interesting questions for their ramifications for agricultural land use among castaa concessionaires. While studies have been conducted on the im portance and features of castaa harvesting in the Peruvian Amazon ( Escobal & Aldana, 2003 ; Lanao, 1998 ) none h as taken up the relationship with agriculture, despite the fact that many families rely on both. Similarly, although some authors have reviewed the issues concerning the forest concession system in Peru ( Escobal, Agreda, & Aldana, 2000 ) there remains as y et no analysis on the importance of households with access to different types of lands for their resource management. Agricultural Land Use Agriculture now prevails in many corners of the Amazon. Road corridors commonly have land settlements with land tenu re types emphasizing agricultural land uses. Agriculture in frontier areas of the Amazon generally features a combination of annual (food) and perennial (fruit) crops, along with livestock husbandry, often featuring cattle ranching ( Fearnside, 1989 )


22 Agricultural land use by households is usu ally carried out along with cattle ranching in frontier regions, as part of its diversification strategy. Cattle ranching has been profusely studied by scholars mainly because its impacts on deforestation at the early stages of colonization in the Amazon. Although the emphasize of initial studies was on large landholdings, more recent studies point out that cattle ranching is widespread among all sizes of landholdings ( Walker, Moran, & Anselin, 2000 ) More recently, other researchers are noting an increasing shift or diversification towards cattle ranching for smallholders previously engaged in extractivist strategies ( Gomes, Vadjunec, & Perz, 2012 ; Salisbury & Schmink, 2007 ) Perspectives on small scale agriculture highlight the importance of size and demographic composition of households. This appr oach is well suited for agricultural frontiers where farmers have land abundance and family members are the main source of labor ( Brondizio et al., 2002 ; Stephen G. Perz, Walker, & Caldas, 2006 ; Walker & Homma, 1996 ; Walker, Perz, Caldas, & Texeira, 2002 ) Beyo nd labor availability, landholding size and access to capital (often limited in frontier areas) are emphasized as key household assets that can influence their land use. Research on roadside farms in the Amazon also underscores the importance of factors b eyond household characteristics; among the most important is transportation infrastructure. The development of transportation infrastructure modifies the spatial economy and fosters market connections that make location increasingly important for land use ( Britaldo Soares Filho, 2004 ) New infrastr ucture facilitates market access and improves the profitability of land use by reducing transport costs. This can result in land use intensification close to towns and land use expansion farther out. Most of this


23 re is a potential rent associated with any land use of any land plot, and consequently a given land plot will be dedicated to the land use that generate the highest rent ( Walker, 2003 ) Therefore, agricultural profitability will be determined by distance to markets, so agricultural development first proceeds close to cities and roads that allows farmers to reach cities more easily ( Chomitz & Gray, 1996 ; Nelson & Hellerstein, 1997 ) While the theoretical effects of infrastructure improvements for land use are clear, conclusions drawn about the wisdom of infrastructure in the Amazon vary. Transportation infrastructure is often discussed as a limiting factor for land use in the Amazon, to be overcome by governments to achieve economic development ( Easterly & Serven, 2003 ) By contrast, Amazonian scholars have mostly emphasized the negative environmental impacts of roads in the Amazon, highlighting their role in facilitating deforestation and low productivity economic activiti es ( Coffin, 2007 ; Fearnside, 1987 ; Pfaff et al., 2007 ) Yet other authors have pointed out that improvements in transportation infrastructure can yield diverse outcomes for different interest groups, such that some benefit economically whereas others suffer from loss of livelihoods or income ( Shriar, 2006 ) Beyond infrastructure, land tenure is another important f actor affecting resource management. In parallel with forest extractivism, work on agricultural land use also highlights the importance of tenure security ( Alston, Libecap, & Mueller, 1999 ) Tenure i nsecurity has been held out as an explanation for rapid deforestation as competing stakeholders seek to clear forest as a means of establishing agricultural land claims. The implication is that secure tenure, as via definitive land titles, will yield slowe r forest


24 clearing and more sustainable land use via investments in long term production. This line of argumentation has however been called into question by other authors who have noted that secure tenure often proceeds alongside land use intensification a nd continued deforestation ( Angelsen & Kaimowitz, 2001 ) Yet other researchers point out that land tenure regimes should be considered in conjunction with biophysical and socio economic factors in order to understand its impact on land use ( Futemma & Brondizio, 2003 ) Such discussions presume a focus on private exclusive land rights. However, tenure r egimes have become considerably more diverse in the Amazon. In particular, use rules have proliferated among different kinds of lands, with many variations on regulations and prohibitions. Even with regard to agricultural lands on private properties, Brazi l and Peru differ in that the first has deforestation limits whereas the second does not. More generally, tenure regimes vary among the countries sharing the Amazon with regard to access, use rights, deforestation limits, fungibility and transferability, a mong other key characteristics of land tenure. Infrastructure projects and land tenure rules can in turn be viewed as one element of a broader suite of public policies that influence land use. Scholars consider public policies such as colonization project s, credit, or policies to foster expansion of cattle ranching by engaging the dynamic international and national demand for beef ( Bowman et al., 2012 ) Brazil has for many years had credit and monetary policies favorable to cattle ranching ( Binswanger, 1991 ; Walker et al., 2009 ) However, in the case of smallholders, some authors point out that ca ttle ranching does not respond to costs benefits logic focusing on short term gain, but instead to strategies premised on


25 accumulation as savings to serve an insurance function in case of household crises ( Siegmund Schultze, Rischkowsky, da Veiga, & King, 2007 ) Likewise, pasture can hav e more than one purpose: ranching may serve to grow forage for animals, appropriate land in the sense that it will help with land reclamation, and increase the value of land ( Fearnside, 2001 ; Mertens, Poccard Chapuis, Piketty, Lacques, & Venturieri, 2002 ) The analysis of agricultural land use is usually conducted by modeling household decision making concernin g land allocation with all factors mentioned above taken into account ( Angelsen & Kaimowitz, 1999 ; Caldas et al ., 2007 ; Singh, Squire, & John, 1986 ; Vosti, Witcover, & Carpentier, 2002 ) This draws on proximate intermediate distant determinants frameworks where p roximate factors such as household assets, intermediate factors such as infrastructure, and distant determinants such as public policies influence land use allocation. Most models assume that households follow a profit driven logic; however, such assumptio ns can be restrictive in the Amazon, where small producers combine subsistence and market oriented production The Livelihood Strategies Approach: Capitals and Capabilities I adopt a livelihood approach to evaluate the relationship of agriculture and NTFP. The livelihoods approach highlights the asset profile of households, including the possibility that they hold multiple kinds of land; it also views livelihoods as encompassing multiple types of activities, such as NTFP harvesting and agriculture, among ot hers. This approach allows for the consideration of diversification of productive activities within households, making it more suitable to address the relationship of NTFPs and agriculture. ( Chambers & Conway, 1991 ; Ellis, 2000 ; Netting, 1993 )


26 Livelihoods frameworks also relax certain assumptions in economic models of household production. It points out that households not only produce for market sales per standard economic assumptions about rationality as profit maxim ization; households also prioritize risk management by engaging in subsistence production for auto consumption. Hence, households may manage their land and other assets for multiple purposes. Indeed, this is a key reason for livelihood diversity: household s may pursue complementary activities to suffice both sustenance and marketing goals. A livelihoods approach thus not only affords evaluation of multiple livelihood activities, it also permits analyses that highlight the relationships between household ass ets and particular activities, as well as how pursuit of one type of activity influences other activities. This requires attention to various types of household assets, as well as the multiple possibilities households have for pursuing productive activitie s. Just as households must allocate scarce assets carefully to pursue a given activity, output from one activity may in turn generate assets that help support pursuit of other activities. Hence a livelihoods approach highlights the importance of assets for activities, as well as assets generated by one activity for another. The livelihoods approach takes a broad perspective on household assets as the ( Bebbington & Perreault, 1999 ) This fra mework goes beyond the typical focus on land, labor and (productive) capital inputs in standard household production models by considering the various types of capitals. Different types of capitals should be considered, such as natural capital (natural res ources), human capital (knowledge, skills and capabilities), produced capital (means of production produced by human beings) and social capital ( Bebbington,


27 1999 ) Notable in this rega rd is the addition of social capital ( Coleman, 1990 ; Putnam, 1993 ) as a household asset. Social capital refers to relations of trust and social support; social capital can thus be viewed as an asset of households, but accounting for so cial capital requires recognizing that household assets in turn stem from relationships with other households and cooperation. Social capital may be an intangible asset, but given scarcity of other forms of capital as in the Amazon, social capital can offs et other capital scarcities and play an important role in supporting rural livelihoods ( Uphoff & Wijayaratna, 2000 ) Also feasible in the livelihoods framework is recognition of different kinds of lands to which households may have access. This raises issues concerning assets tied to land tenure, especially for cases where households have access to lands with distinct tenure rules and thus enable different productive activities. Whereas standard househol d economic frameworks view land as a generic category, the livelihoods approach underscores the importance of distinguishing among lands with different tenure rules. Hence access to multiple types of lands becomes a key household asset that permits livelih ood diversity. Further, land assets tied to distinct livelihood activities allow identification of the importance of access to one land type for productive activities on other land types. I employ the capitals and capabilities framework to evaluate the imp ortance of various household assets for productive activities, focusing on NTFPs and agriculture. Within this framework, I feature the role of access to different kinds of lands, notably castaa concessions, as castaa harvesting may affect other livelihoo d activities, particularly agricultural land use. This raises theoretical questions about the likely


28 relationship between access to forestlands for NTFP harvesting and other lands for agricultural land use. Prior work strongly suggests a fundamental confl ict between the two. There is by now a large literature on frontier expansion in the Amazon that has often featured conflicts between indigenous peoples and forest extractivists who rely on standing forests and colonists and ranchers who seek to clear fore sts for agriculture (e.g. Schmink & Wood, 1992 ; Simmons, Walke r, Arima, Aldrich, & Caldas, 2007 ) In particular, much attention has gone to the conflicts between ranchers and rubber tappers in the southwestern Brazilian Amazon ( Allegretti, 1990 ; Ehringhaus, 2005 ) Theoretical appraisals of the viability of NTFPs and agriculture also casts doubt on the harmonious co existence of th e two. The scenario typically portrayed in regions where both are present is one where forest extractivism is replaced by agriculture as part of the process of regional development and land use intensification ( Homma, 1992 ) There is limited empirical work on these issues, but it also suggests an inverse rela tionship of forest extractivism and agriculture. Some scholars who have focused on the impact of agriculture on NTFPs and found a negative effect in terms of land use ( Pulido & Caballe ro, 2006 ) Such work leaves open important questions concerning NTFPs and agriculture, especially in light of the issue of access to multiple types of lands, which may make the two livelihood activities more consonant with each other. For one thing, acc ess to both forested lands for NTFPs and other lands for agriculture permits households to pursue both activities rather than favor one over the other. Of course, this also depends on household labor and capital availability. The case of castaa and agricu lture however


29 offers further reason to expect complementarity rather than tradeoffs. There is seasonal complementarity in labor requirements for castaa harvesting and agricultural activities. That is, engaging in castaa does not present a major opportuni ty cost for practicing are important distinctions, such as crop cultivation and cattle ranching. Crop cultivation is more labor intensive than cattle ranching, whic h may allow greater complementarity in labor scarce households between castaa and cattle. Given these considerations, castaa may exhibit synergies with agriculture via the effects of facilitated market insertion. Unlike other NTFPs, castaa is an eminen tly commercial commodity, intended primarily for sale in markets. And there is a dynamic regional market with local companies that compete among them to buy any amount of castaa they are offered. In this scenario castaa can be an important source of cash for rural household on the region, importance that increases when compare with other farm activities. Thus, market insertion via castaa can generate income for investment in other activities, including agriculture. Such investments in turn facilitate agr icultural expansion, whether via crops or cattle, and production of a surplus for additional marketing. Hence castaa may facilitate agricultural market insertion, and in the process expand agricultural land use. Hence the relationship of NTFPs with agric ultural land use depends on factors including whether the two productive activities occur on the same land, whether there are seasonal complementarities in peak labor demand, and the labor intensity of the agricultural activities in question. In this chapt er, I consider the relationship of access to castaa concessions with agricultural land use where the two occur on different lands,


30 exhibit seasonal complementarity in labor demand, and where agriculture involves both crop cultivation and cattle ranching. Under these circumstances, NTFPs and agriculture may not compete, and may not exhibit a negative (inverse) relationship. Indeed, to the extent that income from castaa provides investment capital, castaa may facilitate market insertion for agriculture, mo tivating expansion of agricultural land use, yielding a positive relationship between NTFPs and agriculture. A Changing Frontier in the Southeastern Peruvian Amazon To pursue this question, I focus on the Southeastern Peruvian Amazon. This area has been re garded as the most biodiverse area of the Amazon ( Myers, Mittermeier, Mittermeier, da Fonseca, & Kent, 2000 ; Sa ndoval & Wilson, 1996 ) Specifically, I focus on the region of Madre de Dios, which is made up of the provinces of Tambopata, Manu and Tahuamanu. Although the region has some highlands, most of its territory is lowland forest on the eastern side of the A ndes between 80 and 400 meters above sea level (IIAP 2001). One key reason for using Madre de Dios as the study area is because it has castaa stands, which are zoned for forest extractivism via The other key reason for selecting Madre de Dios is because it is experiencing rapid change via new infrastructure and agricultural expansion. While Madre de Dios is still largely forested ( Southworth et al., 2011 ) it is now incurring changes due to the Inter Oceanic Highway, which connects Madre de Dios to the rest of Peru. This highway was recently paved under the auspices of the Initiative for Integration of Regional Infrastructure of South America ( IIRSA, 2008 ; Santa Gadea & Wagner, 2002 ) The policy goal of paving the Inter Oceanic Highway was to improve market access from


31 places like Madre de Dios, thus fostering regional development via the export of natural resources. Figure 1 shows castaa concessions and the route of the Inter Oceanic Highway in Madre de Dios. Castaa concessions predominate in the eastern third of the region. The Inter Oceanic Highway is the central transportation corridor that links several small population centers from Iapari on the frontier with Brazil south to the regional capital of Puerto Maldonado and then west to Cusco and Puno and other routes to ports on the Pacific coast. Along the highway are lands zoned for agriculture. Southworth et. al. ( 2011 ) document that most deforestation in Madre de Dios has occurred in those agricultural lands alo ng the Inter Oceanic Highway. Figure 2. 1 also shows that the Inter Oceanic Highway passes thru the castaa zone. Hence it is possible to identify areas with households who have access only to agricultural lands, and other areas with households with access to both castaa concessions and agricultural lands. This permits comparisons of agricultural land use among households with and without castaa concessions. The economic history of Madre de Dios is intimately tied to the extraction of NTFPs as well as the route where the Inter Oceanic Highway now passes. The initial developments of what is now the highway date to the beginning of the twentieth century; a path was initiated in Cusco due to interest in reaching the gold mining areas on the eastern side of And es around the border with Madre de Dios. In the 1960s, the route was opened as a road from Cusco to Puerto Maldonado. Since then, the road has undergone continuous improvements.


32 Alongside these bo om bust cycles tied to extractive economies ( Barham & Coomes, 1996 ) The first dramatic change in the economy of the region was the rubber boom in the late nineteenth century. After the rubber boom crashed in the 1920s, Madre de Dios went into a period of economic stagnation. It w as not until after WWII that migrants again began to come to the region, driven in part by demographic expansion in the Andean highlands ( Lossio, 2001 ) By the late 1970s, castaa harvesting became rooted in the regional economy. Consequently, many households increasingly relied on forest extractivism. Government support played an important role through a variety of policies such as subsidized credit, price guarantees, health assistance, and of course road improvements. Migration at this time focused on acquiring land for agricultural production ( Rodriguez, 1991 ) As agriculture and castaa harvesting emerged in Madre de Dios, land tenure diversified. By the end of the 1990s, the Government of Pe ru, thru its environmental agency INRENA, established several protected areas (INRENA 2003). Notable among these are the Tambopata National Reserve (274,690 ha) and the Bahuaja Sonene National Park (1,091,416 ha). These and other protected areas in Madre d ( INRENA, 2003 ) By the 1980s, yet other activities were em erging in Madre de Dios. Notable here are gold mining and logging. The economic recession in the early 1980s led to a spike in gold prices, and discoveries of alluvial gold in Madre de Dios prompted establishment of mining camps. Similarly, the 2000s econo mic crisis catalyzed a rapid rise in gold


33 values, and stimulated a large gold boom in Madre de Dios ( Mosquera, Chavez, Pachas, & Moschela, 2009 ; Swenson, Carter, Domec, & Delgado, 2011 ) Madre de Dios also held stands of mahogany and other valuable tropical hardwo ods, and this attracted logging operations of varying size and capitalization. The logging sector was reformed in the 2000s under the New Forestry Law (Cossio 2009), which sought to require management plans as a step toward certified timber extraction, wit h higher market prices. However, the costs of securing management plans have proven high, and stocks of highly valuable wood have proven limited, yielding some concentration in the logging sector and limiting overall activity ( Cosso, 2009 ) Eco tourism has also e merged in Madre de Dios as an activity of importance, this taking advantage of the protected areas established in the area ( Stronza, 2000 ) All these economic changes have proceeded along side demographic expansion. The urban population comprised approximately 40% of the population in 1972 grew rapidly, reaching 73% in 2007 ( INEI, 2009 ) The expansion of urban population, especially in the capital of Madre de Dios, Puerto Maldonado, in turn fostered the growth of the service, financial and commercial sectors. In this context, the castaa sector has retained its importan ce, in part due to high if fluctuating castaa prices in the 2000s. Castaa concessions are thus a valuable asset for many people in Madre de Dios. There are around 1,000 castaa concessions in Madre de Dios, each with an average size of 1,000 hectares ( Lanao, 1998 ) The concessions are accessed through rivers such as the Lower Madre de D ios River and Piedras, as well as secondary roads, leading from the Inter Oceanic Highway. Castaa concessions are granted to individuals for a one year period, but concession contracts


34 are routinely renewed almost automatically, and concessionaries may ke ep their concession as long as they want with exclusive rights for castaa harvesting ( Pea, 2010 ) Although concession tenure was insecure due to unclearly defined boundaries, this problem is declining since the tools and methods to define boundaries have been improving in government agencies that handle the concessions. The key limitation of castaa concessions is rather that concession contracts only permit castaa harvesting; there is no right to develop agriculture in the concession areas, and logging is allowed only under specific permits and for limited amounts of wood. Henc e while castaa concessions are valuable and secure, they are limited use areas tied specifically to NTFP harvesting. Agriculture and cattle ranching in Madre de Dios have yet to reach the level of importance that they have achieved in the Brazilian Amazo n. Commercial agriculture remains limited due to little technical assistance and the presence of livelihood alternatives, notably informal commercial and service activities. Farmers have annuals, perennials, and pastures in their landholdings, and some sec ondary forest as result of agricultural lands that have been left to fallow. There is usually a mixture of crops in the farms; some are mostly for the market and others are primarily for household consumption. Both annual and perennial crops are put to bot h uses. Nonetheless, even among those crops that are sold in the market, the percentage sold out of total production is small. The use of fertilizers or pesticides is rare, which results in relatively low crop productivity.


35 Whereas NTFPs like castaa are e minently market commodities, agriculture in the Southeastern Amazon region is not fully integrated to market chains. Highway paving and urban population growth are likely to increase the importance of market distance and travel times for commercializing ag ricultural produce. Less clear is the question of whether having access to a castaa concession affects agricultural land use. Insofar as castaa harvesting complements agriculture in terms of the lands on which they occur and seasonal labor demand, there may be no relationship between the two. But since castaa is eminently a commercial product, income from castaa concessions may support investments in market oriented production on agricultural lands. Such investments would tighten the relationship betwee n NTFPs and agriculture and indeed make the relationship positive. That is, having a castaa concession may expand commercial agriculture via greater market engagement thru castaa. Analyzing Land Uses Interaction under Transportation Infrastructure I mpr ovement Methods of Data Collection In order to evaluate the importance of castaa concessions for agricultural land use in a capitals and capabilities framework at households, I conducted fieldwork with rural households in Madre de Dios in 2007 and 2008. I collected data for this study via a survey using a two stage cluster sampling design ( Bernard, 2002 ) In the first stage, I selected 17 communities based on their location along the Inter Oceanic Highway in Madre de Dios, to ensure variation in distance to the main market in the capital of Puerto Maldonado and avoid clustering. While this is not fully random, it is si milar to systematic random sampling. I focused the sampling on the Provinces of Tambopata and Tahuamanu, which contains most of the agricultural lands as well as castaa


36 concessions. 1 Within those provinces, I sampled communities in three different areas : communities close to (<30 km) from the regional capital of Puerto Maldonado, those west of Puerto Maldonado, and north of Puerto Maldonado (the castaa zone). The result is a sample of communities in various locations and with varying degrees of access t o castaa concessions. In the second stage, within a given community, I selected households using simple random sampling. Therefore, where feasible, I worked from a community association list of member families; in the other cases where such lists were ou tdated or unavailable, I worked from estimates of community size and sampled systematically. The second stage yielded a sample size of 306 households. For each household sampled, I conducted a face to face interview with a structured questionnaire. The qu estionnaire included items on the land, labor, and various capital assets of the households, including whether they had access to a castaa concession. The instrument also included information on household location necessary for calculating distance to mar ket. The questions yielded variables that serve as explanatory factors in the analysis, with a focus on castaa concessions. In addition, the questionnaire had a series of items on different land uses as well as additional economic activities carried out b y households. The agricultural land use variables serve as outcomes in the analysis. 1 It is im portant to note that agriculture is also organize around rivers, but I choose to focus on roads since I want to asses the role of transportation infrastructure.


37 Data for Analysis Table 2.1 shows the descriptive statistics for agricultural land uses on agricultural lots. I distinguish between four lands uses: crops (both annuals and perennials), pastures, secondary forest and mature forest. While there are important differences between them, I combined annual and perennial crops because both categories contain crops that may serve both subsistence and commercial purposes. Separati ng these categories, or even distinguishing individual crops, fails to consistently separate their very different uses among different households. For both mature and secondary forests, I let respondents estimate the respective land areas based on their un derstandings of these two categories. That is, I imposed no a priori definition for secondary forest. Forest regrowth can vary in characteristics because it remains ecologically distinct from mature forest for up to 20 years. In practice, respondents tende d to define secondary forest as regrowth on land that had previously been cut. Respondents generally categorized forest as mature if they had not cut it. Table 2. 1 shows that the agricultural lots in the households sampled average roughly 40 ha. Of that la nd, the first column in Table 1 indicates that households dedicate a small area to crops (3 ha) with somewhat more for pasture (8 ha). Most land is in forest, whether mature (20 ha) or secondary (9 ha). That said, the second column in Table 2. 1 indicates t hat there is considerable variation in land areas for these uses. As a result, the distributions are skewed (non normal). I therefore took the natural modeling. The r emaining columns in Table 2. 1 show bivariate correlations among the land uses. In general, correlations are positive, and stronger for the ln transformed variables than the raw values. The positive correlations among crops and pasture


38 indicate that househo lds have production systems of varying sizes, and those with more cultivation also have more livestock. The correlation between crops and secondary forests is not surprising, since larger cultivation systems in the Amazon necessitate larger fallow areas. W hat is surprising is that mature forest also exhibits positive correlations with the other land uses; this is evidence that lot sizes vary, and production systems as well as forest areas reflect land assets. This also suggests that production systems have not yet grown in size relative to properties where there are significant tradeoffs between uses. Table 2. 2 shows the statistics for explanatory variables 2 The first group of variables draws on the capitals and capabilities framework and lists indicators of household assets. For human capital, I consider the education attainment of the household through an index (HEI) that adds the level of education of all adults and divides the sum by the number of adults. The idea behind this index is to go beyond merel y the human capital of the household head(s) and consider the education level of all household adults since many times other household members participate in decision making concerning livelihoods. I also consider the number of adults in the households as an indicator of labor availability. Household land use models for developing areas highlight the crucial contribution of family labor (Chayanov 1966), hence greater labor availability should exert a positive effect on land use areas for crops and pasture. Some of the correlations show the usual signs while others may look at odds with what is expected. In the case of education, the correlations with land uses are non significant 2 Farm size was not considered as explanatory variable because it was not significant when intr oduced in the model. This can be related to the fact that it shows little variance (mean = 41.33, Std. Dev. = 30.81), which in turns indicates an agrarian structure very different from the minifundia latifundia paradigm usual in Latin America.


39 but negative which could be related with the fact that more education enhances the chances to obtain a non farm income. The correlations with number of adults in the household are also non significant; only crops show a negative sign which could also be related to having more chances of non farm income when there are more adults in t he household. The other household assets focus on various types of capital. First are forms of productive capital: possession of a chainsaw and access to bank credit. Theory emphasizes the positive contributions of these capitals to production. For credit I also take into account the access five years ago as well as in the past year, since past credit may entrain decisions about land uses that endure in subsequent years. The data indicate limited access to productive and financial capital, even more so 5 ye ars ago. The correlations of access to chainsaw correlations with pasture and mature forest are significant and positive, which can be explained by the fact that pastures usually requires the clearing of an extensive forest area, and because big land owner s tend to clear pattern, perhaps because credit is put to different purposes. The last asset variable is an indicator of social capital. I constructed this measure as th e number of organizations to which the household head belongs. The index takes higher values where household heads belong to more organizations, more different types of organizations (which serve as sources of different forms of information and support), t he number of years he/she has participated, and if he/she has held a leadership position.


40 Beyond the household capabilities variables, given the recent paving of the Inter Oceanic Highway under IIRSA, I also consider two measures of accessibility. First, I measure distance from the agricultural lot to the Inter Oceanic Highway. Secondary roads in Madre de Dios are unpaved and often in bad shape. Better maintained roads fo r motorcycles and pedestrians. Hence households with lots closer to the highway are more accessible, more valuable and thus likely to have more agricultural land use and less forest. The other accessibility variable refers to distance from the point where the unpaved road reaches the Inter Oceanic Highway to the region capital, Puerto Maldonado, which implies travel on the highway to the primary market in Madre de Dios. I keep these two measures separate since they are not highly inter correlated and they d istinguish the effects of accessibility along unpaved and paved roads. The correlations of distances with land uses are mostly negative which concur with theory, although there are some differences between types of distances considered. Distance to the hig hway shows bigger and more significant correlations, including a positive correlation with forest that can be explained by the fact that farms far from the road develop less agriculture. The correlations with distance to the region capital have the right s ign, only crops have an unexpected sign but it is non significant. Key to this analysis is the relationship between NTFPs in the form of castaa and agricultural land use. There are various ways to measure involvement with castaa, but pre existing theory as well as preliminary testing indicated that having a castaa concession is key. I therefore measure concession access as a binomial variable,


41 or via arrangements to harves t castaa from concessions held by others. Measuring involvement with castaa via a binomial variable for concession access carries the limitation that it does not distinguish among households with larger or smaller castaa harvests and incomes. I recogniz e that there are various other ways to measure involvement with castaa. One obvious candidate is the size of the castaa harvest (or by extension income from castaa). Previous experience in the Madre de Dios indicates that reporting of harvest is often n ot accurate, for various reasons. Moreover, there is considerable annual variation in the castaa harvest, which can potentially result in bias depending on the year for which information about the castaa harvest is requested. Given these limitations, and given that it is access to a castaa concession that is in any event key to benefitting from this NTFP, I therefore focus on access to castaa concessions. Per earlier discussion, concession access may exert different effects on agricultural land use, dep ending on how castaa influences other forms of market insertion. If castaa competes with agriculture or if income from castaa is invested in non agricultural activities, there is likely to be a negative relationship between access to castaa concessions and agricultural land use; but if there is complementarity and castaa income facilitates market insertion for agriculture, there is likely to be a positive relationship. Table 2. 2 shows negative correlations between castaa concession access and all four agricultural land use variables. This begs questions as to whether those correlations hold when controlling for the effects of other factors on agricultural land use. As an additional control variable, I also consider household involvement in off farm liv elihood activities, as an indicator of livelihood diversity. This is an index that is the


42 sum of the number of non agricultural income sources for a household (mining, timber, wages, retirement, rental income, etc.). The last group of variables serves as a dditional controls and focus on household background 3 theoretical framework, they are relevant to livelihoods and land use in Madre de Dios. The first background variable is age of househo ld head, intended to capture household life cycle effects, whereby younger families use land differently than older households ( Stephen G. Perz, 2001 ; Walker & Homma, 1996 ) Whereas younger households are more concerned with food security, they emphasize crops more than pasture and often have less secondary forest and more mature forest. I also include the origin of household head in order to account for differences between migrants and regional natives. The sample shows that 70% of household houses were born outside of Madre de Dios, most of them in the Andean region. These migrants may focus more on agriculture, wh ich may result in larger agricultural areas. Finally, I control for duration of residence on the agricultural lot. With time, land use shifts by reducing mature forest and expanding crops and pasture, which often then result in more secondary growth due to declining soil fertility and land degradation. Table 2. 2 confirms most of these expectations, though it also appears that older lots are also larger lots with more mature forest. 3 Sub region al dummy variables were considered on this group since castaales are not randomly because the sub regions are defined based in other explanatory variables such distance to Puerto Maldonado, the region capital.


43 Analytical Techniques Previous empirical work on land use has employed a va riety of statistical models to estimate the effects of explanatory variables. In this chapter, I employ a three stage least square (3SLS) model. This is because land use decisions are jointly made, such that on a given property, allocating more land to one use necessarily implies less land for others. As a result, land uses are mutually endogenous. 3SLS models account for endogeneity among land uses. If there is endogeneity as in the case of land uses, linear models that assume independence among outcomes w ill yield biased and inconsistent estimates of coefficients. 3SLS estimation therefore involves construction of a system of equations where the dependent variable in one equation is an endogenous regressor variable in one or more of the other equations. 3S LS estimation thus goes beyond other estimators such as two stage least squares (2SLS) and seemingly unrelated regression (SURE) ( Stephen G. Perz et al., 2006 ) Unlike SURE, 3SLS uses estimated (instrumented) values of each endogenous regressor from its equation in the other equati ons, which permits specification of effects among endogenous regressors; and unlike 2SLS, 3SLS permits simultaneous estimation of multiple equations for several endogenous variables. The model specification proposed diverges from other 3SLS models in the sense that not all land uses have the same explanatory variables. This is necessary in order to identify the system of equations. This also makes sense theoretically, as the land uses are different and thus different factors can be more important for one l and use than another. For one thing, the exogenous variables differ among the equations. Notably, the capitals and capabilities variables are highlighted for productive land uses as inputs for crops and pasture but not for mature or secondary forest By c ontrast, accessibility


44 is likely to affect all four land use outcomes. Further, given the focus here on NTFPs and agriculture, I included the castaa concession variable in all four equations for a full accounting of its relevance to land use. In addition, there are likely to be specific relationships among the land use outcomes. This is because the transitions from one land use to another tend to proceed in specific patterns; via deforestation, mature forest yields crops and pasture; older cropland becomes either pasture or secondary growth; and older secondary growth eventually becomes mature forest. Therefore, neither the exogenous nor the endogenous variables are identical for the four land uses. Model specification was thus driven by theory but also inf ormed by empirics. I ran a series of preliminary models with these and other similar variables in order to identify the most useful variables and to obtain robust findings. I modified model specifications by removing variables that were both theoretically unimportant and empirically weak. This permitted both testing of theory and simplification of the equations. However, I also often retained variables of theoretical importance in an equation even if they were empirically weak, to demonstrate those findings as in the case of access to castaa concessions. Findings Table 2. 3 shows the outcomes of the 3SLS estimations. I organize the table such that each column presents one of the four equations in the system. For each equation, I indicate the number of param eters, the chi square statistic for performance, and the coefficients for each variable along with an indication of its statistical significance. All equations show highly significant chi square values (0<0.001). The first equation is for annual and peren nial crops. Since crops are generally planted on lands previously forested, I consider secondary and mature forests as the


45 endogenous regressors. Among these, only primary forest is significant, but it is positive. This suggests that households with more f orest also have more crops. The coefficient for secondary forest is non significant and very small which can be related to the fact that secondary forest is not a restriction for the land allocation on crops, at least not at this point. Madre de Dios remai ns an agricultural frontier, and the secondary forest is not a constraint because it can be used without a long fallow since the nutrients are possibly not too depleted. The coefficients for the household assets behave as expected, such as the positive eff ect of adults (labor) and social capital. The exception is for education, which exhibits a negative effect although weakly significant; this can be explained by the fact that more educated households may participate in non farm activities instead of agricu lture. Both distance variables show the expected signs; farms farther from Puerto Maldonado, and farther from the Inter Oceanic Highway, have less land under crops. That said, the coefficients are weakly significant, which suggests that accessibility is no t (yet) a crucial variable for the development of agriculture. The key variable in this analysis concerns castaa concessions. The crops equation shows a negative but insignificant effect of concessions. The same holds for the non agricultural activities index. The non significant effects suggest diversified livelihoods, in that neither NTFPs (castaa) nor other activities (wages, etc.) impede crop cultivation. In particular, the insignificant effect for access to castaa concessions suggests that there is neither a tradeoff nor a synergy between NTFPs and agriculture among rural households in the study area.


46 As it happens, the household background variables are the most important explanatory factors for crops, beyond availability of forested land. As expec ted, older household heads planted fewer crops, and household heads originating in the Andes planted more crops. The second column in Table 2. 3 presents the pasture equation. Among the endogenous regressors, the strongest variable is mature forest, which exhibits a large, positive and significant effect. This seems to indicate that pasture is a land demanding activity, so the availability of big landholdings is important. Among household capabilities, only productive capital (chainsaw) is important, mainly because chainsaw is an important tool for clearing large amounts of forest in order to form pastures. The accessibility variables exhibit more significant coefficients for pasture than for crops, suggesting greater sensitivity to distance. The coefficien ts for distances are also negative as expected. However, only distance to Puerto Maldonado is significant; pastures are concentrated closer to the capital, regardless of distance from the highway. The key finding in the pasture equation is the large, posi tive, and highly significant coefficient for access to castaa concessions. That is, households with access to castaa concessions have larger areas in pasture. This finding suggests that the cash generating capacity of castaa is operating here; castaa i s after all a commercial NTFP with an established market. Hence households with castaa concessions appear to be investing castaa earnings in pasture. It is possible that households with concessions also have other cash income sources, but the coefficient for non farm income on pasture is insignificant. I therefore interpret these findings as


47 indicating that access to castaa concessions is generating castaa earnings and that those earnings are being invested in pastures for cattle. This interpretation co incides with other research on cattle among smallholders in the Amazon. Rural households often regard cattle as a safe investment, since pasture raises land values and cattle can be walked to market in case of crisis. In this sense, cattle serve an insuran ce function in frontier areas, providing a ready source of cash in the absence of other forms of insurance or savings ( Salisbury & Schmink, 2007 ; Walker et al., 2000 ) The pasture equation also shows a positive and significant effect for household cles in the Amazon ( Perz et al., 2006 ; Walker & Homma, 1996 ) The third equation, for secondary forest, includes endogenous regressors for crops, pastures and mature forest. The findings show a strong and significant positive coefficient for crops, which is not surprising since crop cultivation in the Amazon requires fallowing that leads to the emergence of secondar y forest. Interestingly, the coefficient for mature forest is negative, which indicates that there is a tradeoff between mature and secondary forest cover, and that secondary forests replace mature forests in agricultural properties. The coefficients for accessibility are positive as expected. That is, less accessible lands have more forest. Again, only the coefficient for distance to Puerto Maldonado is significant, implying that there is more secondary forest farther from the capital. This in turn sugges ts more intensive land use close to the main market, and more fallowing farther out.


48 The coefficient for access to castaa concession is negative but insignificant. One possible interpretation for this finding is evident by following the causal pathway vi a crops. Concessions have little impact on crops, and while crops are very important for secondary forests, the insignificance of the first relationship entrains a similarly insignificant effect of concessions on secondary forest. Among the household back ground variables, the duration of residence on the lot has a significant positive impact on secondary forests. This confirms prior work that emphasizes the importance of time for the emergence of secondary forests ( Perz & Walker, 2002 ) That is, in the absence of funds for fertilizers and other technological measures to sustain production, land use over many years in the Amazon eventually necessitates fallowing. The fourth and final equation in Table 2. 3 presents findings for mature forests. Here I included crops, pasture and secondary forests as endogenous regressors. For primary forest the coefficient for crops is positive and significant. Again, this suggests more forest and crops occurring on larger lots. In other words, the positive relationship of crops and mature forests is related to the availability of land among households and to the small land areas allo cated to crops. By contrast, secondary forest again shows a negative relationship with mature forest. As noted earlier, this suggests a tradeoff in forest covers, such that secondary forest replaces mature forest on the agricultural properties surveyed. A mong the accessibility variables, distance to Puerto Maldonado is again significant. As was the case for secondary forests, the effect of distance on mature forests is positive. The coefficient for distance to Puerto Maldonado and mature forest is


49 also lar ger than for secondary forests, which suggests a stronger distance gradient for mature forest. That is, on agricultural lots farther from Puerto Maldonado, there is a more rapid increment in mature than secondary forests. Hence landholdings located farthes t from the region capital keep more primary forest and less land under crops and pastures. The last finding of note for mature forest is that duration of residence on the lot exerts a positive and significant effect. While this is at first glance surprisi ng most research on family farms in the Amazon documents a decline in forest cover over time ( Perz et al., 2006 ) it can be explained by the fact that older landholdings in Madre de Dios tend to be bigger due to more land availability at the beginning of the settlement process. D iscussion Pr evious work on NTFPs and agriculture in the Amazon suggest a tradeoff between the two, such that the emergence of agriculture replaces reliance on NTFPs ( Pulido & Caballero, 2006 ) T he findings of this study show that net of the effects of other factors, NTFPs can foster expansion of agriculture. Key to this positive relationship of NTFPs and agriculture is access to lands with different tenure rules, which makes possible simultaneous NTFP harvesting and agricultural production. Where tenure and household livelihoods are diverse, this positive relationship between NTFP and agricultural activities should not be surprising. A key implication of the findings concerns the mixed ramificat ions of NTFPs such as castaa. While there is a large literature on the benefits of NTFPs for environmental conservation and livelihoods, that literature has emphasized the direct impacts of NTFP harvesting ( Hall, 1997 ; Kar & Jacobson, 2012 ; Manzi & Coomes, 2009 ) By contrast,


50 little has been said about indirect impact s of NTFPs on other landholdings also used by rural households that exploit NTFPs. For the case of castaa an NTFP with a well defined market, good accessibility, attractive prices, and other requirements for economic viability this study indicates a p on land use via cattle ranching. Hence while castaa concessions simultaneously provide for rural livelihoods and conservation of forest cover, access to them also permits families to invest castaa income in c attle pasture on their agricultural lots. While there has been debate about the potential and problems of NTFPs, it is important that evaluations of NTFPs takes into account the direct as well as indirect consequences. The link between castaa and cattle requires some additional comment for the case of Madre de Dios. Investment of castaa in cattle ranching can be interpreted in light of the lack investment opportunities in more ecologically benign activities, as well as the importance of cattle ranching a s a saving or insurance strategy ( Salisbury & Schmink, 2007 ) It is possible that households with castaa concessions invest th eir cash incomes in urban houses, education and other ends beyond pastures. Nonetheless, the findings show an important effect on pasture area, and the fact that pasture has been reported as an investment and/or saving strategy for smallholders in other re gions reinforce the findings on this research. Certainly the fact that castaa concessionaries have also access to farmlands where they can develop pastures facilitates the link between castaa and pasture. An implication here is that if funds from an e conomically viable NTFP are to be put to other sustainable purposes, there must be other ecologically sustainable livelihood options. In that light, from a capitals


51 and capabilities perspective, it is crucial to understand the tenure regimes and livelihood portfolios as key elements of the context of NTFP economies. If households have diverse land assets, investment may flow where value can accumulate, as in cattle ranching, rather than lands without accumulative possibilities such as forest concessions. Th is leaves open questions about how funds from castaa and other NTFPs can be invested in other lands where households have an incentive to conserve forest. These observations raise questions about the non significance of non agricultural activities for ca ttle pasture as well as the other land uses on agricultural lots. One interpretation is the lack of predominance of non agricultural incomes among the households surveyed. It is possible that non agricultural activities do not demand a great deal of resour ces from households so the effect on agriculture is not important. Another explanation is that funds from non agricultural incomes are invested for various purposes rather than concentrated on ranching. The findings for accessibility also require further comment. One might expect major investments in the Inter Oceanic Highway to show large effects on agricultural land use; similarly, one might expect unpaved secondary road to hamper land use farther from the highway ( Chomitz & Gray, 1996 ; Southworth & Tucker, 2001 ) The model shows that accessibility along the highway to have more significant effects than acc essibility to the highway itself. Further, the effects of accessibility along the highway are greater for some land uses (pasture) than others (crops). The somewhat limited effects of accessibility may reflect the fact that the Inter Oceanic Highway is not a new


52 or because there is more competition because products are being imported into the region ( Shriar, 2006 ) An additional explanation is simply that the effects of paving may take time to make themselves felt. There are two points more that can be considered, the first one is that despite the improved transport infrastructure the regions is still far away from the bigger and more active national and international markets. In this sense the benefits of paving the road wil l have only a regional impact, which should be only a moderate impact. The second point is that transportation infrastructure can help to overcome only natural obstacles such as rivers and mountains, other obstacles such as lack of technical assistance for productive initiative or affordable credits for farmers can hinder the advantages of having a paved road. The main implication of the findings for NTFPs, accessibility and agricultural land use is that in a frontier region characterized by household with diverse assets and land tenure, a given policy may generate little effect or unanticipated outcomes. If a policy (like paving a highway) is intended to stimulate agriculture, there is a need for complementary policies (e.g., technical support) to ensure th e desired outcome. Similarly, if tenure policy seeks to encourage sustainable use of forests alongside agriculture, but only permits private investments in agricultural and not sustainable use of forests, then it should not be surprising that income from N TFPs is put to cattle ranching. If in fact the policy goal of forest concessions was to expand sustainable use of forests, then there needs to be avenues for investing incomes from castaa harvests in the NTFP sector. While there are certification initiati ves, the castaa sector in Madre


53 de Dios remains highly concentrated in terms of buyers and processing plants. This then is a potential target for policy incentives.


54 Figure 2 1 The Madre de Dios region


55 Table 2 1 Descriptive Statistics and Correlations with Land Uses (n=306) Land Use Mean Std. Dev. Correlation with: Crops Pasture Secondary Forest Primary Forest Crops Has 3.08 3.13 1.00 Ln 1.14 0.76 1.00 Pastures Has 8.17 15.55 0.10+ 1.00 Ln 1.22 1.38 0.29*** 1.00 Secondary Forest Has 9.27 11.13 0.24*** 0.002 1.00 Ln 1.69 1.23 0.55*** 0.28*** 1.00 Primary Forest Has 20.77 21.27 0.31*** 0.10+ 0.11+ 1.00 Ln 2.24 1.58 0.60* ** 0.35*** 0.45*** 1.00 +p < 0.15, *p < 0.05, **p < 0.01, ***p < 0.001


56 Table 2 2 Descriptive Statistics for Explanatory Variables and Correlations with Land Uses (n=306) Explanatory variable Mean Std. Dev. Correlation with Crops Pastures Secondary forest Primary forest Household Capabilities Household education 4.27 1.85 0.062 0.001 0.07 0.02 Adults in household 2.60 1.36 0.02 0.07 0.07 0.006 Chainsaw now (0 = No, 1 = Yes) 0.6 0 0.49 0.03 0.15** 0.03 0.23*** Chainsaw 5 years ago (0 = No, 1 = Yes) 0.03 0.16** 0.09+ 0.17** Credit now (0 = No, 1 = Yes) 0.33 0.47 0.09+ 0.001 0.04 0.02 Credit 5 years ago (0 = No, 1 = Yes) 0.08 0.28 0.13* 0.03 0.06 0.15** Social capital 7.70 3.49 0.14* 0.05 0.17** 0.02 Accessibility Distance to main road (km.) 6.14 9.58 0.24*** 0.15** 0.28*** 0.25*** Distance to region capital (km.) 69.45 36.09 0.01 0.06 0.03 0.34*** Non agricultural activities Castaa concession (0 = No 1 = Yes) 0.29 0.46 0.36*** 0.12* 0.29*** 0.32*** Off farm activities 0.38 0.25 0.12* 0.03 0.04 0.06 Household background Household head age 47.36 14.35 0.13* 0.14* 0.05 0.04 Household head origin (0 = MDD, 1 = Non MDD) 0.71 0.45 0.19** 0.05 0.03 0.12* Landholding years of use 13.87 13.22 0.12* 0.26*** 0.35*** 0.22*** +p < 0.15, *p < 0.05, **p < 0.01, ***p < 0.001


57 Table 2 3 Three Stage Least Square Model for Land Uses (n=305) Explanatory Variables Crop s Pastures Secondary Forest Primary Forest Equation Parameters 13 13 8 8 Model Chi square 222.18*** 99.50*** 142.16*** 226.86*** Intercept 0.77** 1.48* 0.20 0.22 Endogenous Land Use variables Crops 2.15*** 2.21*** Pastures 0.09 0.12 Secon dary Forest 0.06 0.29 0.78* Primary Forest 0.42* 1.00*** 0.64** Household capabilities Household education 0.05+ 0.02 Adults in household 0.05+ 0.09 Chainsaw now (0 = No, 1 = Yes) 0.11 Chainsaw 5 years ago (0 = No, 1 = Yes) 0.64* ** Credit now (0 = No, 1 = Yes) 0.05 Credit 5 years ago (0 = No, 1 = Yes) 0.07 Social capital 0.02* 0.03 Accessibility Distance to main road (km.) 0.009+ 0.01 0.004 0.01 Distance to region capital (km.) Non agricultural activities 0.005+ 0.02** 0.009* 0.01*** Brazil nuts concession (0 = No, 1 = Yes) 0.10 1.31** 0.19 0.54 Off farm activities 0.03 0.10 Household background Household head age 0.004* 0.01* Household head origin (0 = MDD, 1 = Non MDD) 0.16* 0.03 0 .29+ 0.24 Landholding years of use 0.03** 0.03** +p < 0.15, *p < 0.05, **p < 0.01, ***p < 0.001


58 CHAPTER 3 INTERMEDIATE AND DISTANT DRIVERS AND LAND COVER CHANGE IN THE SOUTHEASTERN PERUVIAN AMAZON The Peruvian Amazon is a forested region that has been subject to public policies and infrastructure investments for regional development, which are likely to generate land cover change. However, current policies and investments are occurring in a heterogeneous space defined by a zoning plan based in part on spatial variation in forest resources. The Peruvian Amazon is thus a useful study case for examining distant and intermediate drivers on land cover change, as well as differences in biophysical characteristics and settlement history as they influence r egional deforestation patterns and land cover trajectories. This paper therefore evaluates deforestation and land cover trajectories in the southeastern Peruvian Amazon over a period of 25 years, emphasizing the most recent period when significant policy changes and infrastructure investments occurred. I then focus on different land cover change patterns in three sub regions that differ in terms of market proximity and forest resources. The trajectory analysis shows considerable acceleration in forest loss with the onset of infrastructure investments, and the sub regional analysis reveals large differences in land cover dynamics. These findings complement prior work in the Peruvian Amazon and bear implications for debates about public policy, infrastructure and environmental governance with respect to land cover change. The literature on land use and land cover change has been focused on the identification and analysis of drivers of change ( Geist & Lambin, 2006 ; Gutman, 2004 ) Although denominations differ, there is some agreement on grouping drivers on proximate, mediate and distant ( Geis t & Lambin, 2006 ) There is also some discussion


59 on the scale of analysis (Wood and Porro 2002), ranging from the international to the household levels. The choice of the scale of analysis is related to the relevant drivers. The present research is focu sed on assessing the role of intermediate and distant drivers on deforestation at regional and sub regional scale. In particular, I focus on public policies as distant drivers, emphasizing infrastructure investments, which result from public policies and i n turn serve as intermediate drivers. I focus on distant and intermediate drivers via a regional approach to understanding land cover change, (2005). The regional pers pective acquires importance when the region analyzed is also an area with specific policy and economic dynamism. I focus on the southeastern Peruvian amazon, which as part of the Peruvian lowlands has been targeted by specific public policies, and in rece nt years has received large scale infrastructure investments. The Peruvian Amazon is itself significant as a region characterized by the presence of important biodiversity spots and extensive protected areas ( IIAP, 2001 ; Myers, Mittermeier, Mittermeier, da Fonseca, & Kent, 2000 ) The Peruvian Amazon has incurred development initiatives via public policies, which have changed over time, with implications for land cover change ( Chavez, 2009 ) The forests of the southeastern Peruvian Amazon are currently under pressure due to improvements in transportation infrastructure through the paving of the Inter Oceanic Highway ( Santa Ga dea & Wagner, 2002 ) According to previous experiences elsewhere in the Amazon region, infrastructure upgrades can accelerate the process of land cover change ( Dourojeanni, Barandiarn, & Dourojeanni, 2010 ; Keller, 2009 ; Killeen, 2007 )


60 This raises questions as to whether the paving of the Inter Oceanic Highway has accelerated forest loss in the southeastern Peruvian Amazon. However, the Inter Oceanic Highway is being paved in a region with an established zoning plan that accounts for difference in forest resources, soils and past land uses ( IIAP, 2001 ) The southeastern Peruvian Amazon has many forest concessions for timber and non timber forest products (NTFPs) in some areas but not others. Forest concession s act somewhat like protected areas in imposing land use restrictions that can reduce deforestation. Protected areas, concessions and other institutional distinctions in land tenure regimes constitute a proximate factor that also influences land use and la nd cover change. This raises questions as to whether areas with forest concessions exhibit less forest loss than areas without such concessions in the same region receiving new infrastructure. The present research analyzes the change in land cover in the Madre de Dios region of southeastern Peru, where the Inter Oceanic Highway has recently been paved through areas with and without forest concessions. I draw on a time series of satellite imagery from 1986 to 2010. Specifically, I used forest / non forest c lassifications based on Landsat images for six time points from 1986, 1991, 1996, 2000, 2005 and 2010 to build a land cover change trajectory map. This lengthy period of analysis allows for the assessment of land cover trajectories during a period when dis tant drivers such as policies and intermediate drivers including the paving status of the Inter Oceanic Highway changed. In addition, I created three sub regional buffers within the land cover map, one around the regional capital of Madre de Dios and two others along other portions of the


61 Interoceanic Highway. Of those two areas, one of the buffers includes the section of the highway corridor that passes through a sub region with forest concessions for a key regional NTFP, and while the other buffer occurs in an area without such concessions. The comparison of trajectories among these three sub regions reveals differences related to their proximity to the capital and the presence of forest concessions. This permits comparative analysis of trajectories among sub regions with new infrastructure but differences in settlement history and the presence of forest concessions. This paper proceeds as follows. I first provide a review of relevant land science literature, focusing especially on discussions of distant, intermediate and proximate drivers. I then discuss the study case of the southeastern Peruvian amazon, highlighting the evolution of public policies for the area, along with changing demographic patterns and the effects of commodity prices on the regional economy. This motivates a discussion of rural settlements, zoning and forest reserves in the study region. I then discuss the satellite data and processing and classification, noting how the methods were standardized to permit comparisons both over time an d among images in mosaics. The analysis proceeds in two steps. I first focuses on the overall study region and emphasize non linearities in land cover trajectories from 1986 to 2010, notably the most recent period during pacing of the Inter Oceanic Highway The second highlights the comparative analysis of the three sub regions to identify differences in trajectories and deforestation under distinct circumstances of land settlement, tenure regimes, location in relation to the market and other factors. Both parts of the analysis document large differences, in forest loss among time periods and among the sub regions. These


62 findings bear implications for debates about public policy; infrastructure and environmental governance with respect to land cover change E xplanations for Deforestation in Frontier Areas The importance of land cover changes in forested areas is by now well established ( FAO, 2011 ) Land cover changes in the Amazon region, as well as the consequences for regional and global climate are also well understood ( Malhi et al., 2008 ; Shukla, Nobre, & Sellers, 1990 ; Vitousek, 1994 ) However, the discussions abou t drivers, agents, and processes of land cover are still unfolding. This reflects the complexities in regard to the diversity of drivers, especially across levels at which different drivers operate ( Wood & Porro, 2002 ) As a result, theoretical frameworks for explaining land cover change highlight multiple causation on different levels of scale ( Geist & Lambin, 2006 ) Earlier analyses of land cover change focused on deforestation with a highly aggregated approach ( Barbier & Burgess, 2001 ) This initial highly economic approach changed by incorporating demographic, institutional, social and biophysical factors. Land science also moved towards mo re spatially explicit approaches, frequently focused on smaller land units like pixels or properties belonging to households ( Br ondizio et al., 2002 ) As a result, there has been considerable emphasis on the spatial dimension of the study of land use and land cover change ( VanWey, Ostrom, & Meretsky, 2005 ) This perspective differentiates land covers in detail, which allows the researcher to deal with the impacts of various drivers on l and cover change at different scales. Some authors use spatially explicit approaches to evaluate land cover change at the household level to understand changes among farms in terms of household features ( Brondizio et al., 2002 ; McCracken et al., 1999 ) or the influence of institutional


63 arrangements on land use in communities and regions ( Southworth & Tucker, 2001 ) or the importance of macroeconomic policies on land uses and deforestation ( Mertens, Sunderlin, Ndoye, & Lambin, 2000 ) Additionally, multi scale analyses provide interesting insights about the role of drivers at different scales and different points in the time ( Mertens, Kaimowitz, Puntodewo, Vanclay, & Mendez, 2004 ) Nevertheless, the emphasis on methods in data demanding spatially explicit approaches can be also a weakness, since this approach has produced an increasing amount of empirical research and models with limited theoretical background ( Irwin & Geoghegan, 2001 ) An interesting path to address this issue is the systematization of theoretical explanations to account for patterns on various scales. Geist and Lambin ( 2001 ) provide an exhaustive meta analys is of explanations for deforestation, and they causes that include infrastructural demographic, economic, technological, policy or institutional and cultural and socio/political factors. Intermediate causes directly influence land uses that in turn produce deforestation. Finally, they note studies that also consider other factors that can be also regarded as drivers, such as land characteristics, biophysical drivers and social triggers. Wood ( 2002 ) provides a more developed categorization of explanat ions for land cover change. He presents a three tiered hierarchical approach to organize the socio economic and biophysical drivers that influence land use decisions make for nce. And drivers are organized taking into consideration three scales: proximate,


64 intermediate, and distant. Here, characteristics of land users constitute proximate factors; local and regional characteristics including infrastructure, population, land ten ure, technology and local prices comprise intermediate factors; and policies, commodity prices, and history constitute distant factors. This account separates intermediate factors like infrastructure and tenure from more distant factors like policy, which can influence infrastructure and land use rules. The frameworks offered by Geist and Lambin (2002) and Wood (2002) provide systematic accounts to organize explanations for land cover change, and permit several additional observations. The first is the dif ferentiation between proximate causes and underlying drivers; proximate causes are the land uses that produce land cover changes, while drivers are the factors fostering one or more land use. This issue seems trivial but it commonly confusion between land use and land cover change. Another important observation is raised by Wood ( 2002 ) who proposes biophysical factors as drivers along with socio economic factors. Geist and Lambin ( 2002 ) group biophysical ( Rudel & Roper, 1997 ) This claim makes sense because biophysical factors can be considered as given conditions, since fact ors other than soil fertility, like precipitation, topography, drainage, are not as dynamic as socio economic factors mentioned above. Nevertheless, it is important to consider both socio economic and biophysical factors as contributing explanations to lan d cover change, without disregarding their differences.


65 on land use and land cover ch ange is often considered only at the household level. Dynamics at larger scales are not merely aggregations of individual preferences; there are processes operating beyond the household level, which household models have difficulty accounting for. At the s ame time, it is also important to note that land cover changes at the household level is more complicated than merely considering it in the aggregate such as at the regional level. These considerations have driven many regional studies of land cover change accounting for processes on multiple scales while managing complexity by limiting the spatial extent of analyses. At the same time, structural frameworks for explaining land cover change tend to pay limited attention to temporal dynamics ( Perz & Almeyda, 2010 ) By definition land cover change is a dynamic process, and what is more, it may exhibit important non linearities over time. Land cover trajectories, defined by non linearities among different periods, are in turn likely to reflect changes in the importance and organization of explanatory fac tors over time. This remains a challenge for spatially explicit land cover research, which often has detailed data for a given time point but is constrained to cross sectional or single period analysis. But these considerations have also driven regional st udies with multi temporal data to account for complex (non linear) dynamics in land cover change over multiple time periods, using detailed change data for drivers, and managing the data demands by limiting the scope of the study to a single region. Given these general considerations, I now focus on a discussion of some key drivers of land cover change. I select drivers to highlight based on their relevance to my study region in the Amazon. I also feature drivers acting from the regional/local to global lev tier


66 proximate/ intermediate/ distant drivers scheme ( 2002 ) Specifically I will review the roles of population, infrastructure, land tenure, public policies, and commodity prices. The role of demographic growth is considered in almost any research on land use change. Initial research emphasized the relationship between deforestation and populat ion growth at national levels ( Mather & Needle, 2000 ; Rudel & Roper, 1997 ) the focus shifted toward households where issues such as age structure and household size gained importance ( McCracken et al., 1999 ; Perz, 2001 ) Others authors also emphasize the need to look on population not only as land users but also as demand for agricultural goods, especially when talking about urban population ( Faminow, 1997 ) Now it is accepted that population is not the main driver of deforestation but part of a set of drivers, and often it is pushed by sudden changes in frontier areas and public policies ( Lambin et al., 2001 ) In this sense it is posit that policies and contextual factors will define the final outcome on the population deforestation interaction. ( Perz, Arambur, & Bremner, 2005 ) Infrastructure, such as roads, has commonly been recognized as a key intermediate factor behind land cover change ( Coffin, 2007 ; Southworth et al., 2011 ) Roads make land s more accessible, raising land values and permitting access to resources for exploitation as well as to markets for commercialization ( Chomitz & Gray, 1996 ; Pfaff et al., 2007 ) Conversely, lack of access was considered as a limit for responses to markets and policies ( Schmook & Vance, 2009 ) In the Amazon and elsewhere, most deforestation occurs along lands more readily accessible by roads ( Keller, 2009 ) By improving access to lan d for land use, roads thus constitute an intermediate factor.


67 At the same time, land tenure rules by definition also influence land use. Use ( Ostrom, 1990 ) hence land tenure sets rules for land use. In the Amazon and elsewhere in Lat in America, there is by now a wide range of types of lands with distinct use rules ( Richards, 1997 ; Zoomers & Haar, 2000 ) which implies that land use and land cover change vary among different land tenure types. Hence in regions with land tenure diversity, there is good reason to expect different land cover change on different types of lands. In this sense, land tenure like infrastructure serves as ano ther intermediate factor that influences land use and thereby land cover change. Beyond land tenure specifically, state policies indirectly influence land cover change as distant factors. In frontier regions like the Amazon, policy initiatives like coloniz ation projects can influence infrastructure investments, migration, technical assistance, and other intermediate factors that influence land use. The projects developed in the Brazilian Amazon with the World Bank support are good examples ( Redwood, 2002 ) or the Special Development Projects carried out by the Peruvian government ( Rodrigue z, 1991 ) Other policies can also constitute tools that influence land use, with some governments emphasizing credits or tax exemptions ( Mahar, 1989 ) However, lately government role is changing and losing importance while private agents and markets forces gain significance ( Rudel, 2007 ) Thus, during late 1980s and early 1990s, in countries such as Peru and Bolivia the IMF fostered this process through programs of structural adjustment. These programs usually entailed cutbacks on public spending, land markets liberalization, opening to external commerce, etc. These


68 programs had mixed results in relations to land cover change, in some cases export oriented crops gained importance while those oriented to internal market declined ( Hecht, Kandel, Gomes, Cuellar, & Rosa, 200 6 ; Pacheco, 2002 ) This raises questions of the role of commodity prices. Prices on forest and non forest commodities should greatly influence land use decisions, to the extent that they are market or iented and thus made with relative profits in mind. To the extent that prices are low for forest products and high for agricultural products, land users should engage (Singh and Squire 1986, Ellis 1988). In a globally integrated economy, commodity prices on forest and non forest products are increasingly distant factors, determined as prices are by increasingly distant markets. Natural resources timber, non timber fore st products (NTFPs), as well as agricultural commodities ranging from crops to livestock, all have prices determined on national as well as international markets. In addition, factors affecting exchange rates can influence prices and thus land use decision s ( Redo, Millington, & Hindery, 2011 ) In this light, policies related to market prices are important to land cover change. Trade policies are a good example, specially the trade liberalizations implemented on the structural adjustment programs ( Redo et al., 2011 ) It is also important to note that trade policies affect land cover cha nges through agriculture by favoring export oriented crops ( Lopez & Galinato, 2005 ) Another important impact related to government policies is through economic growth, wh ich usually has a negative elasticity with forest cover, but growth reduces poverty that should have a positive impact on forest cover.


69 The drivers we just reviewed are important to understand land cover change, including in a temporal perspective, since t hey all vary over time. Infrastructure can be improved, land tenure rules can change, state policies shift with new administrations, and international commodities prices fluctuate. Because such changes can be non linear over time, including reversals as i n the cases of policies and prices, there is need for at least two periods to allow for observation of possible changes in trajectories in land cover change. Drivers and Dynamics of Land Cover Change in the Southeastern Peruvian Amazon In this section, I review the study region by focusing on key drivers of land cover change and how they themselves have changed over time. I focus on the region of Madre de Dios, located in southeastern Peru. This is a useful case for the study of land covers trajectories be cause Madre de Dios has considerable internal heterogeneity, which permits spatial comparisons of land cover change. In addition, Madre de Dios has incurred numerous shifts in key driving forces behind land use in recent years, with the consequence that th ere are indications of non linearities in land cover change. I begin by laying out the biophysical characteristics of this region. In general Madre de Dios can be characterized as lowland Amazon. most of its territory is below 400 meters above sea level ( Goulding, Canas, Barthem, Forsberg, & Ortega, 2003 ) The soils are typically Amazonian, with weathered topsoils that are not apt for intensive agriculture ( Martnez, 1998 ) Almost all the region falls within the Madeira River watershed, with the Tambopata and Madre de Dios rivers as the main tributaries to the Ma deira basin.


70 Maybe the most important biophysical feature of Madre de Dios is related to variation in the forests in terms of their economically significant species. Eastern Madre de Dios encompasses a significant extension of castaa forests. This is par t of a bigger Amazonian section that encompasses areas in Brazil and Bolivia. Castaa is an economically important NTFP, which is the focus of many forest concessions in Madre de Dios. Castaa forests thus provide an important component of local livelihood s, and thus sustain an important economic activity in the form of the annual castaa harvest ( Collinson, Burnett, & Agreda, 2000 ) The presence of castaa forests and castaa concessions in eastern Madre de Dios is thus arguably an important influence on land use and land cover change favoring forest retention. That said, Madre de Dios is experiencing rapid change due to a mix of public policies, new infrastructure, land tenure shifts, and changing commodity prices. Madre de Dios has incurred colonization proje cts since the 1980s, with the result that population in the region has grown over time, following the usual pattern observed in other frontier areas in the Amazon. Table 3.1 shows that population growth in Madre de Dios accelerated during the 1980s and 199 0s when most of the settlement process took place. The 2000s the growth rate in the region decreased, while growth in urban areas gained importance. This is a more or less a common situation on frontier regions, when final stages of population growth happ ening in the urban rural. It is possible to notice that population growth is bigger in the Tambopata province, which contains the regional capital of Puerto Maldonado. The capital not only contains the regional government, it serves as the main market cent er of Madre de Dios. Puerto Maldonado is also surrounded by areas of older rural settlement.


71 More generally, the provincial population data begin to reveal local differences in population. Among the three provinces in Madre de Dios, Tambopata (where Pue rto Maldonado is located) received the initial wave of colonists, and now has the largest population; somewhat later came migration into Manu province. Key to understanding demographic as well as land cover shifts in Madre de Dios are shifts in government policies ( Chavez, 2009 ) As in other countries on the Amazon, the Peruvian government put in place a set of policies aimed to foster settlement and agriculture in the Amazon. These have varied over time among different administrations (Chavez 2009). Under the Garcia administration in the late 1980s, colonization was promoted, supported by credit for agric ultural activities. These initiatives were supported by subsidized credit, fertilizers and pesticides as well as the establishment of ( Rodriguez, 1991 ) These initiatives also proceeded with the establishment of an extensive network of organizations to provide t echnical assistance and develop research applied to agriculture in the Amazon. Together, these policies prompted expanding agricultural development in Madre de Dios. It was an emerging agriculture with low productivity and few technical improvement but it implied a noteworthy land cover change. This entire scenario changed in the early 1990s due to economic crisis and the onset of the Fujimori administration. The new government discontinued credit and assistance programs. Under these conditions, the expans ion of agriculture in Madre de Dios was curtailed ( Alvarez & Naughton Treves, 2003 ) However, government policies still played a role. The regional government became more important in 1998 1 when it 1 Before this year the regional government in Madre de Dios was part of the Inca Region led by Cuzco regional government.


72 gained some independence as via the right to manage its own budget. These expenses were mostly focused on the regional capital, which experienced a noticeable improvement in urban services along the 2000s, and increa se its important as a market place. However, the regional government also began to invest in various productive sectors in Madre de Dios. Also notable is that during this period, land tenure in Madre de Dios diversified. The regional government pursued a zoning plan (IIAP 2001) that seeks to delineate use rules based on biophysical aptitudes of soils, forests and other natural resources. As a result, the national and regional governments began implementing protected areas such as parks and reserves, forest concessions for timber and castaa, and agricultural lands. The result has been the creation of land tenure mosaics, which vary across Madre de Dios. Along key roads, agricultural lands prevail; where there are significant forest resources, forest concess ions can be found; and in more remote areas with substantial remaining biodiversity, as in the Andes Amazon transition, there are indigenous lands and protected areas. Under the Toledo administration of the early 2000s, policies again changed, and this ti me infrastructure was prioritized. And under the second Garcia administration in the late 2000s, new development initiatives were prioritized via decree laws facilitating access to natural resources by private firms. During these two administrations, the m ost important government action in Madre de Dios was the paving of the Inter Oceanic Highway (IOH) ( Santa Gadea & Wagner, 2002 ) which was initiated under Toledo and implemented under Garcia. The IOH crosses the len gth of Madre de Dios from its border at Cusco and Puno east to Puerto Maldonado and from there north thru the


73 castaa zone up to the border with Brazil and Bolivia at Iapari. In addition to the well known effects on improving transportation infrastructure during construction, this project provided an important boost to regional economy by hiring labor and other project expenses. There has been much debate and polemic concerning the eventual impacts of the IOH in southeastern Peru ( Dourojeanni et al., 2010 ) But given that the I OH now permits one to cross Madre de Dios in one day, at any time of the year (including the rainy season), it is likely that the IOH has raised land values and impelled new economic activity and deforestation. While government policies have been important especially as they result in key intermediate drivers like infrastructure, land cover change in Madre de Dios also reflects fluctuations in commodity prices. Due to its importance for the economy region and the extension of castaa forest, castaa prices are important for land c over in Madre de Dios. Figures 3.1 and 3 .2 show castaa exports and prices, respectively. Both figures indicate growth in the importance of castaa during the 1990s and 2000s. This suggests that forest retention for the castaa har vest is likely to be an important factor for land cover change, at least in the castaa zone in Madre de Dios. However, other key regional commodities have also exhibited rising prices. International prices on logging were also important for land use in t he region, these prices experienced an important increase in the early 2000s, which generated illegal logging and violent social conflicts in the region ( Cosso, 2009 ) However, the impact of selective logging can be hard to detect in satellite images of land cove r. In any event, the most dynamic economic sector in Madre de Dios is withou t doubt in gold mining. Figure 3 .3 shows gold prices during the 1990s and 2000s. Gold


74 prices have risen far more than prices on other commodities in Madre de Dios. The frenzy on gold mining caused by these prices was developed mostly on the western area of Madre de Dios in Manu Province ( Mosquera, Chavez, Pachas, & Moschela, 2009 ; Swenson, Carter, Domec, & Delgado, 2011 ) Gold mining does yield land cover impacts, and significant areas ha ve been deforested in Madre de Dios for gold mining. However, gold mining remains a relatively minor explanation for land cover change compared to agriculture; this may yet change in the future as the gold sector continues to expand. For now, the demand fo r agricultural production continues to favor regional agriculture. The urban of Madre de Dios continues to grow rapidly, constituting demand for crops and livestock products. Overall, this discussion of distant and intermediate drivers of land cover chang e in Madre de Dios highlights spatial variation among local circumstances as well as significant temporal change over time. Further, while there are some factors that favor forest retention, as via the economic crisis of the 1990s, establishment of protect ed areas, and the growing importance of the castaa sector, there are also major forces favoring deforestation, notably population growth, large scale infrastructure projects, growing regional demand for agricultural commodities, and the gold boom. On bala nce, there is thus a tendency towards increasing pressure for land cover change, thought this is likely to be greater in some areas, like the lands surrounding Puerto Maldonado, than others, such as the castaa zone. I therefore pursue a spatial as well as temporal analysis of land cover change in Madre de Dios, considering both comparisons among distinct sub regions as well as multiple time periods with distinct sets of circumstances. I highlight not only distant drivers like prices and policies as they va ry over time for all of


75 Madre de Dios, but also intermediate drivers as they vary across sub regions within Madre de Dios. Data and Methods for Analysis of Land cover Change Trajectories Given the foregoing discussion, I focus on Madre de Dios and evaluat e land cover change there since the late 1980s. Specifically, I use a time series of land cover data for Madre de Dios that spans the period from 1986 until 2010. This time series covers the period from the first Garcia administration thru the 1990s and th e 2000s, a period of numerous shifts in policies and prices. This allows for temporal comparisons among periods of different circumstances to observe possible non linearities in land cover trajectories and to see whether those trajectories can be accounted for by circumstances during each period. The satellite imagery also permits delineation of different sub regions within Madre de Dios for spatial comparisons, as between the area around Puerto Maldonado and the castaa zone to the north. I pursue a two p art analysis. The first part is applied at a regional level and focuses on contrasting land cover change trajectories over multiple time periods. Here I highlight changes on public policies, prices of natural resources, demographic dynamics and infrastruct ure. The second part involves a spatial analysis that distinguishes three different sub regions with different locations in relation to the regional market, and biophysical features related to land tenure. The data used for the analysis consists of a fo rest non forest classification developed from mosaics of Landsat images that cover the study region for the years 1986, 1991, 1996, 2000, 2005 and 2010. The processed images are part of a broader analysis conducted for the tri the southwestern Amazon ( Marsik, Stevens, & Southworth, 2011 ; Southworth et al., 2011 ) An additional imag e for


76 2010 was added, which was processed and classified following the same protocol. All processing followed a common protocol for all images and time points to permit comparisons over time and among sub regions. I draw only on the imagery for Madre de Di os. The time series of images permit observation of land cover changes during each intervening period. There are five such periods: 1986 1991, 1991 1996, 1996 2000, 2000 2005, and 2005 2010. These periods correspond roughly to the various presidential adm inistrations in Peru discussed earlier; they also correspond roughly to e.g. the implementation of the Inter Oceanic Highway, which occurred during the last of these intervals. Hence the time series permits useful comparisons as they may relate to policy e nvironments and infrastructure implementation. I used the time series of satellite imagery to derive land cover change trajectories across the six periods for which land cover change can be observed. The land cover change trajectory was obtained through t he Spatial Modeler in ERDAS. Given that there ar e two land cover classes and six time periods, I obtained a total of 64 possible trajectories. I then mapped these trajectories in Madre de Dios to identify areas with distinct trajectories in land cover from 1986 to 2010. This provides the basis for the temporal analysis in the first part of my analysis. For the second part of the analysis, which features spatial comparisons among sub regions (see Fig. 3.4) I identified three areas with distinct situations i n terms of their characteristics defined by differences in intermediate drivers of land cover change. I focus on the three sub regions, which comprise a portion of the area of Madre de Dios. This complements a focus on all of Madre de Dios, which has been done elsewhere


77 (Southworth, et al. 2011). Specifically, the area for analysis is the path of the IOH thru Madre de Dios, with a 15 km buffer to either side of the highway. I employed a 15 km buffer for two reasons: first, the IOH runs along the border of M adre de Dios, and employing a larger buffer would not greatly increase the area covered in the analysis; and second, as the highway has only recently been paved, most forest close is very close to the road corridor. I focus on the portions of the highway c orridor that encompass the three sub southeastern portion of Tahuamanu province down into northeastern Tambopata Tambopata pr ovince), and the corridor west from the capital (in southwestern Tambopata province and the far southeastern corner of Manu province) in the area region is basically the area surrounding the city, to ens ure an adequate land area for analysis I use a buffer of 30 km for that sub region. I used a somewhat smaller buffer around the IOH in the other two sub regions because larger buffers would increasingly move away from the area of influence of the IOH and i n the Castaa area in particular, a wide buffer would include lands in Bolivia, since the IOH runs along the Bolivian border in the castaa zone. The result is a study region involving three buffers or areas with specific known characteristics tied to bio physical as well as socioeconomic factors theorized to be influential on land cover change, and of importance in the Peruvian context. In the following I discus the areas considered in terms of factors and drivers presented before: Settlement process: the Central area was settled by the first wave of migrants that came into the region. While the Pampa is the area where last migrants


78 Population: Although population has been grow ing in the regions during the entire period of analysis, the growth is more notorious in the Central area along with the increasing importance of urban population. Market location: undoubtedly the Central area is better located in relation to the market an d should be more impacted by market dynamisms. The Castaa and the Pampa are located in a similar position and the market impacts should be similar on both areas. Tenure regimes: the Castaa area comprehends a significant amount of lands under the concessi on regime which entails restriction for land use and should reduce the chances for land cover change. The other two areas present the same tenure regime with no restrictions. Other policies: Beyond tenure policies it is fair to assume that government polic ies affected the three areas in the same way. Biophysical features: the main difference among the sub regions is the presence of castaa forests in the Castaa area, which should reduce land cover change du e to the value of standing castaa forests. By fo cusing on the IOH corridor, we can evaluate the effects of paving on roadside land cover over time. Similarly, by focusing on the IOH corridor, we can evaluate the importance of policies from different periods, where many programs with land use incentives were largely implemented where land was accessible. And by comparing sub regions that differ in terms of their biophysical characteristics, settlement histories, land tenure, and proximity to markets, we have reason to expect to find different land cover c hange trajectories across space. Finally, the sub regions were defined to constitute land areas of similar sizes to facilitate robust comparisons without bias due to sample size; however the buffer around Puerto Maldonado and the one in the Pampa were curt ailed on the areas that overlap the Tambopata National Reserve.Overall, I expect two key findings. First, given the onset of paving of the IOH after 2005, I expect acceleration in forest loss after 2005 compared to previous periods. This reflects the large scale investments in highway paving and the improvements in the


79 accessibility and thus the value of lands near the highway corridor. While previous policies likely also had effects on land cover change, as noted by Chavez (2009), many were implemented for short time periods or had limited technical support, and so are likely to have had relatively modest impacts on land cover change. Second, given the differences among sub regions in the terms outlined above, I expect greater forest loss in the Central reg ion than elsewhere. Key to this expectation is the longer time of settlement and the proximity of land to the key regional market of Puerto Maldonado. Conversely, I expect less forest loss in the Castaa zone due to its biophysical characteristic of having castaa forests, which are protected via the tenure regime of forest concessions that prohibit forest clearing. Similarly, I expect less forest loss in the Pamp due to its relatively recent land settlement and greater distance from Puerto Maldonado. Afte r defining the three sub regions, I subset the map of land cover change trajectories using the mask function of ArcGIS. I then calculated the percentages of land areas in each of the 64 trajectories in each sub region. The first part of the analysis will t herefore evaluate the trajectories overall for the three sub regions combined to appraise the importance of distance factors like policies, prices and intermediate factor of the IOH, which passes thru the entire study area. The second part of the analysis will compare the sub regions to evaluate the effects of intermediate factors such as biophysical variation in forests tied to land tenure diversity in the form of castaa concessions, to see if the land cover trajectories vary among the sub regions as a re sult.


80 Results Table 3.2 presents the findings from the trajectory analysis of land cover in the study area in Madre de Dios. I begin by focusing on the overall findings to evaluate trajectories in light of distant and intermediate drivers affecting the ent ire study region. The 64 different change trajectories have been grouped into five classes to better read maintained its forest cover during all six years for which I ha ve land cover data. Overall, the study area retains most of its forest; nearly 83% of the land area originally with forest retained forest during 1986 which had been deforested by 1986. Very little land area had been deforested by 1986; the stable non forest area covered far less than 1% of the study area. These findings imply land cover dynamics during the period covered by the satellite image time ser ies. The third class includes trajectories where land cover shifts from forest to non forest at some time during the 1986 2010 period. The area considered as stable non forest is the complement on the previous category. The group of tending toward deforest ation encompasses those areas that at some point in time have been converted to non forest and kept in that condition. It is possible that these trajectories reflects mainly pastures since that is the land use that implies cutting forest cover and maintain it like that for a period up to 15 years. category indicates deforestation during 1986 1991 This percentage is smaller than the one for the next period (non forest since 1996, which refers to deforestation during


81 1991 1996), and that percentage is lower than the one for the next period. This trend continues until the last period, which has the higher percentage of all. Indeed, there is roughly a five fold jump in deforestation from the 2000 2005 period to the 2005 2010 period. At the same time, Table 3.2 also shows trajectories tending toward reforestation. And there we find a broadly opposite r esult: there is a decrease over time in reforestation, which implies a deceleration over time in this trend. Thus there is a large reduction (roughly four fold) in reforestation from the penultimate to the ultimate period considered. Therefore, the overall area deforested exceeds the area reforested. large number of trajectories where there is a shift from one land cover type to the other, and then back to the first. Becaus e the large majority of the 64 classes are rotational, and none of them is large, I have grouped them into on aggregated class for presentational purposes. Table 3.2 does however show that a considerable land area (nearly 4%) was in rotational land use fro m 1986 to 2010 in the study area. These findings together indicate accelerating land cover change in the study area, which implies growing impacts of drivers over time. In particular, it is hard to escape an interpretation that focuses on the paving of th e Inter Oceanic Highway, as well as relatively high prices for commodities, during the last time period observed (2005 2010). Land cover change mainly deforestation intensified during the paving of the IOH and rising prices for regional commodities. T able 3.2 also permits comparisons among the sub regions to evaluate spatial differences in land cover trajectories. The first sub region is built around Puerto


82 the capita the capital headed toward Cusco in areas of newer settlement. A comparison of land cover trajectories for these three sub regions reveals large differences in land cover dynam ics. First, stable forest percentages differ, being lowest in the Central area, somewhat higher in the Pampa, and highest in the Castaa zone. Second, stable non forest is also highest at the Central area. These findings correspond to expectations of great er non forest land cover and older non forest areas near the Capital. Perhaps most interesting however are the comparative land cover dynamics during the period for which I have satellite imagery. The trajectories tending toward deforestation exhibit the l argest percentages in the Central area, with the Pampa and the Castaa zone showing similar values. Further, the deforestation percentages for the Capital are larger for the earlier time periods, indicating the impacts of earlier policies and settlement th ere are bigger than elsewhere. The findings for trajectories tending toward reforestation show contrasts to those for deforestation. Whereas the Capital shows limited reforestation, there is considerably more in the Pampa, and rather little in the Casta a zone. Taken alongside the deforestation findings, the reforestation result suggest more intensive or at least sustained land use near the capital than elsewhere. The results for rotational use also suggest that there is more land in periodic use near the capital than in the other sub regions.


83 The sub regional analysis also confirms the importance of major changes in the drivers of deforestation over time, with much more land cover change in the most recent period. However, the importance of the most rece nt period varies among the sub regions. Whereas land cover change during the period since 2005 greater than all previous periods combined in the Pampa and the Castaa zone, with the bigger increase near Puerto Maldonado, where land cover change has occurre d over a longer period of time. Hence the Inter Oceanic Highway and price fluctuations are important everywhere, but especially around the regional capital. That said, there is evidence of other factors operating to differentiate land cover among the sub r egions. The castaa zone had the largest percentage of land in stable forest, which very likely reflects the presence of forest concessions there, where deforestation is prohibited. Meanwhile, the Pampa had more land in rotational land cover trajectories, which is suggestive of more extensive and older agricultural settlements than the castaa zone. Discussion The foregoing analysis of land cover trajectories in the southeastern Peruvian Amazon permitted both a temporal analysis to evaluate possible non l inearities in land cover dynamics, and a spatial analysis to identify local contrasts in land cover change among distinct sub regions. In turn, identification of non linearities and spatial differences can be interpreted in light of distant and intermediat e determinants of land cover change. In the case of Madre de Dios, we can confirm both non linear land cover trajectories due to distant and intermediate drivers as well as spatial differences due to intermediate factors. The acceleration in forest loss w ith the onset of paving of the IOH confirms the importance of infrastructure for land cover change. Hence the case of the IOH in Madre


84 de Dios provides a high profile example of how infrastructure projects lead to rapid land cover change. This has been see n elsewhere previously in the Amazon ( Arima, Walker, Perz, & Caldas, 2005 ; Wood & Porro, 2002 ) This finding raises questions about the wisdom of infrastructure in ecologically vulnerable regions, given the lessons from past projects. Madre de Dios is of particular ecological importance due to its exceptional biodiversity ( Myers et al., 2000 ) and the rapid expansion of forest loss since 2005 is cause for serious concern. That said, the findings also revealed substantial spatial differences in land cover change tied to intermediate factors like biophysical conditions tied to land tenure, and proximity to market centers. The presence of castaa forests and castaa concessions led to relatively less deforestation in the castaa zone. It seems the casta a forests operate as a buffer against pressures for more agriculture. These findings confirm the importance of biophysical factor for land use, including their influence via land tenure designations. Further, these findings confirm that even under conditi ons of substantial pressure for land cover change, as seen in Madre de Dios after 2005 with the IOH and rising commodity prices, land cover change does vary where land tenure rules and biophysical features vary. The area around Puerto Maldonado exhibited m uch more and older deforestation, reflective of market proximity and land values and thus relatively more intensive land use. This finding confirms that proximity to market is a dominant force for land cover change. Areas close to the main regional market place present the more lands involved on trajectories to permanent deforestation and transitional land cover. They are also responsive to changes in signals from intermediate and distant drivers.


85 While the findings provide evidence in support of some gove rnmental policies, such as via land tenure and highway paving, the temporal analysis also showed differences in the earlier time periods due to significant policy changes. In particular, the overall analysis as well as the sub regions did reveal large reve rsals in land cover change after the first Garcia administration to Fujimori, because the withdrawal of support for colonization. Hence policy changes become manifest in landscapes. This is an important finding since it allows comparing impacts of differen t government policies deployed in the study region. Chavez ( 2009 ) pointed out that policies b ased on maintained in time or complemented with other policies. Findings in this research add to that by showing that impacts of transportation improvement in the 2005 2 010 period are far more important than those on 1985 1990. The sub regional differences in turn confirm the importance of multiple causation behind land cover change ( Keller, 2009 ; Wood & Porro, 2002 ) While paving of the IOH is certainly important all three sub regions show a jump in forest loss after 2005 the size of the acceleration varied among the sub regions, due to other factors such as biophysical characteristics of forests as tied to land tenure ( Ostrom, 1990 ; Richards, 1997 ) Hence other factors also affect land cover change, though in the case of Madre de Dios they do not appear sufficient to fully offset the impacts of the IOH. What is the much more interesting and worrisome question now is whether the paving of the IOH and high commodity prices will continue to generate accelerating land cover change going forward. There is every indication that high gold prices will continue to yield expanding gold mining and thus deforestation in mining areas. At the same


86 time, the newly completed Intercontinental Bridge over the Madre de Dios River at Puerto Maldonado has raised land prices, resulting in land invasions, land speculation, and urbanization of peri urban rural properties. Farther north along the IOH, there are new plantations of semi perennials and perennials, a sign of land use intensification. Hence it is also possible that agricultural land use will expand outward from Puerto Maldonado, pushed by expanding urbanization and continued growth in urban demand for agricultural produce. A major question then becomes whether this pressure northward will in turn affect the castaa concessions. Also in play in Madre de Dios and Peru in general is the question of payments for ecosystem services (PES) programs notably REDD+ ( Armas et al., 2009 ; Fleck et al., 201 0 ) Peru is actively preparing for REDD+, and there have been numerous local initiatives for PES mechanisms. However there have also appeared reports of fraudulent scams where land titles are handed over rather than used as proof for participation in PES programs. This and the market mechanism behind REDD+ have engendered suspicion among many rural communities about PES. Hence the situation concerning land cover change in Madre de Dios is particularly fluid and uncertain, as the distant and intermediate d riving forces behind land cover change are multiplying even as they grow stronger. While it seems likely that forest loss will intensify, it is not clear whether that will be sustained, possibly resulting in reforestation or rotational patterns. Further tr ajectory analysis will thus be important for environmental monitoring and evaluation of the driving forces behind land cover change in Madre de Dios, and other areas where similar driving forces are at play.


87 All that said, the foregoing analysis focused o n geographic comparisons among sub regions that vary in terms of multiple characteristics. While the land science literature does highlight the importance of regional comparisons ( Lam bin et al., 2001 ) geographic comparisons have limitations in their ability to isolate causal factors as explanations for variation in land cover change. Hence while the sub regional comparisons in this chapter provide findings that implication highway p aving, distance to capital, time since settlement, biophysical characteristics and land tenure regimes in land cover change, comparisons fail to permit quantification of the relative importance of these factors. In that sense it is important to complement the present findings with additional research that allows micro level comparisons among many locations, which would permit statistical control and direct evaluation of the relative importance of different explanatory factors by conducting a using a tempora l perspective. This approach has been used before ( Mertens et al., 2004 ) but it can be improved by using an analysis unit that is also a decision making unit such as farms or concessions..


88 Table 3 1 Population growth in the Madre de Dios region 1961 1972 1981 1993 2007 Total Pop. Total Pop. Total Pop. Total Pop. Total Pop. Pop. Growth Pop. Growth Pop. Growth Pop. Growth Pop. Growth % % % % % Madre de Dios Region 14,890 4.8 21,304 3.2 33,007 4.8 67,008 5.7 109,555 3.4 Urban 15,960 38, 433 6.9 80,309 5.0 Rural 17,047 28,575 4.2 29,246 0.2 Tambopata 8,925 4.1 14,760 4.5 24,583 5.5 46,738 5.2 78,523 3.6 Urban 13,056 31,249 6.8 65,444 5.1 Rural 11,527 15,489 2.4 13,079 1.2 Manu 1,488 8.8 1,208 1.9 3,496 10.8 13,827 9.9 20,290 2.7 Urban 275 2,800 13.7 7,261 6.3 Rural 3,221 11,027 9.1 13,029 1.2 Tahuamanu 4,477 5.2 5,336 1.6 4,928 0.9 6,443 2.2 10,742 3.6 Urban 2,629 4,384 4.2 7,604 3.8 Rural 2,299 2,059 0. 9 3,138 3.0 Source: Instituto Nacional de Estadisticas Figure 3 1 Evolution of Peruvian Castaa exports (Source: PromPeru) 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Weight (Tons)


89 Figure 3 2 Evolution of castaa price in Peru (Sourc e: PromPeru) Figure 3 3 Evolution of gold price at world market (Source: 0.00 1.00 2.00 3.00 4.00 5.00 6.00 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Value (US $/Kg.) 0 200 400 600 800 1,000 1,200 1,400 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Gold price ($/oz.)


90 Table 3 2 Land cover change trajectories in the Madre de Dios region Study Central Pampa Ca staa Area Area Area Area % % % % Stable Forest (F F F F F F) 82.959 65.943 86.633 90.195 Stable Non forest (NF NF NF NF NF NF) 0.279 0.953 0.053 0.071 Tending toward Deforestation 7.515 16.483 4.304 4.967 Non forest since 1991 (F NF NF NF N F NF) 0.262 0.574 0.078 0.244 Non forest since 1996 (F F NF NF NF NF) 0.526 1.413 0.173 0.309 Non forest Since 2000 (F F F NF NF NF) 0.487 1.040 0.199 0.419 Non forest since 2005 (F F F F NF NF) 1.020 2.611 0.522 0.496 Non forest since 2010 (F F F F F NF) 5.220 10.846 3.332 3.498 Tending toward Reforestation 1.312 1.013 2.722 0.103 Forest since 1991 (NF F F F F F) 1.035 0.790 2.173 0.061 Forest since 1996 (NF NF F F F F) 0.223 0.076 0.533 0.010 Forest since 2000 (NF NF NF F F F) 0.014 0.046 0. 005 0.004 Forest since 2005 (NF NF NF NF F F) 0.014 0.042 0.006 0.005 Forest since 2010 (NF NF NF NF NF F) 0.025 0.058 0.006 0.023 Rotational 7.935 15.607 6.288 4.664 Sub region extension s (has.) 763,767 185,564 288,125 290,078


91 F i gure 3 4 Buffers for land cover change trajectories for Madre de Dios


92 CHAPTER 4 SOCIAL MOVEMENT ORGANIZATIONS AND THE UNSUSTAINABILITY OF SUSTAINABLE INITIATIVES This chapter takes up questions concerning grassroots soci al organizations such as farmer federations and the roles they can play in sustainable development. I deploy the concept of social capital to understand the capacities of grassroots organizations for collective action, and political ecology to account for the broader socio political context of a regional farmer federation. This multi level theoretical framework accounts for multiple processes to better understand how grassroots organizations can function as both enablers of and constraints on conservation i nitiatives. I use this framework to pursue a dynamic assessment of the interplay of internal and external factors to grassroots organizations to understand their evolution as manifest in their changing perspectives and initiatives concerning sustainable de velopment. I study the case of a farmer federation in the Southeastern Peruvian Amazon, which experienced a breakthrough beginning and a period of strong conservation initiatives, followed by period of organizational breakdown and inactivity. The change fr om the first to the second period can be explained by both internal and external factors to the federation. I argue that the success of conservation and development initiatives by grassroots organizations requires a favorable conjunction of external and in ternal conditions, which carries implications for grassroots mobilization and conservation initiatives. In the last few decades, conservation and development initiatives have highlighted the role of grassroots or local organizations. Such bottom up approac hes to conservation and development were a consequence of criticisms of top down approaches that emphasized the role of centralized authority of the state, which often failed to achieve stated goals and instead yielded negative environmental and social


93 imp acts ( Chambers, 1984 ) A case in point concerns the Amazon region, which in the 1960s and 1970s became the focus of many large scale and state led development initiatives that in turn resulted in deforestation and social conflicts ( Dourojeanni, 1990 ; Hall, 1989 ; Mahar, 1989 ) Developmen t projects designed without consulting local peoples led to conflicts and violence, and prompted mobilization of marginalized groups such as indigenous peoples and forest extractivists ( Hall, 1997 ; Schmink & Wood, 1992 ) Such mobilization in turn led to creation of grassroots associations who have sought to combine conservation with development ( Anderson, 1990 ; Hall, 1997 ; Kainer, Schmink, Leite, & Fadell, 2003 ) Grassroots initiatives in the Amazon have subsequently proliferated, and often aim to improv e local livelihoods while conserving natural resources. In this context, there is by now considerable literature on community based natural resource management (CBNRM), including criticisms thereof ( Agrawal & Gibson, 1999 ) Debate over CBNRM has documented the advantages as well as liabilities of bottom up approaches to conservation and development, and revealed enabling conditions as well as limitations of community based approaches. This in turn has drawn attention to formation of alliances among communities and between communities and outside organizations ( Manrin g, 2007 ; Poncelet, 2004 ; Wollenberg et al., 2007 ) However, there remains a need to more closely evaluate the effectiveness of federative organi zations that may span multiple communities and work with outside organizations. This paper therefore problematizes the capacities of regional grassroots organizations for conservation and development initiatives. Specifically, I address the question of the internal and external conditions that may serve as enablers and


94 constraints on the effectiveness of grassroots organizations to pursue conservation and development initiatives. I take up the case of the Amazon, which has been the subject of considerable prior work on conservation and development models, including the shift from top down to community based approaches. I focus on the southeastern Peruvian Amazon, which is a useful case study for present purposes. The southeastern Peruvian region of Madre de Dios incurred land settlement in the 1980s that in turn prompted the emergence of community associations and then regional grassroots organizations in the 1990s. Further, grassroots organizations in Madre de Dios have experienced rapidly changing fortunes ranging from some highly successful conservation and development initiatives as well as more difficult circumstances, which provide a range of experiences for comparisons over time. To understand the efficacy of grassroots organizations, I argue that it is necessary to account for both internal and external factors. I therefore present a theoretical framework that highlights factors within grassroots organizations as well as those outside of such organizations to serve as an interpretive framework for gr asping enablers and constraints on the effectiveness of organizational initiatives. For the internal factors, I focus on the social capital among individual members of grassroots organizations, on the argument that relations of trust are crucial for effect ive collective action ( Uphoff & Wijayaratna, 2000 ) For the external factors, I employ insights from political ecology that highlight the importance of relations with other interest groups and organizati ons, such as governmental entities, NGOs, aid agencies, and private firms ( Bryant, 1992 )


95 In the southeastern Peruvian Amazon, both internal and external fac tors have been highly dynamic and the effectiveness of grassroots organizations in pursuing conservation and development initiatives has varied over time. These observations imply that internal and/or external factors can serve as enablers or constraints o f organizational initiatives at different moments. The analysis therefore follows grassroots organizations over time. Following the evolution of an organization allows us to appreciate better the interplay among explanatory factors and organizational effec tiveness, and help identify the conditions necessary for grassroots organizations to efficaciously pursue conservation and development. The chapter is organized as follows. I first present the theoretical framework used in the analysis, via a discussion o f social capital and political ecology. The second section presents the study region of Madre de Dios, Peru by outlining the main characteristics of the area and its recent history. This discussion also introduces FADEMAD, a grassroots organization represe nting agricultural colonists in Madre de Dios, and its relationships with community associations, governmental organizations, and other social actors. The heart of the paper presents a multi temporal comparative analysis of three periods over the 20 year h istory of FADEMAD. The first period, 1990 its first initiatives. This period permits an account of the circumstances motivating the creation of FADEMAD. The second period, 1994 2000, encompasses several significant conservation and development projects carried out by FADEMAD. This period highlights the importance of both social capital and political ecology as enablers of successful grassroots organizational initiatives. The thir d period, 2001 2010, discusses a more


9 6 relations, and less effective initiatives. This period underscores the roles of social capital and political ecology as constraint s on grassroots initiatives. I conclude the paper by discussing the implications of the main findings from the analysis. Social Organizations and Sustainable Development ( 1993 ) and guidelines for norms and behaviors. I focus specifically on farmer fed erations, which are grassroots organizations that operate on a regional level, and thus encompass many grassroots community associations, which operate on a more local level. As such, regional federations constitute second level organizations that occupy r oles beyond those permitted by smaller community organizations, which have more limited capacity. Nonetheless, as grassroots organizations, farmer federations seek to respond to community concerns ( Bebbington, 1996 ) The emphasis on social organizations as for a focus for initiatives for sustainable development comes from interest in bottom up approaches that highlight incorporation of local peo ples in conservation and development initiatives. It is also the consequence of social conflicts in rural areas that stimulated social movements and social organizations ( Chambers, 1984 ; Uphoff, 1991 ) In the Amazon region, the emergence of social movements proceeded among diverse groups, largely seeking t o defend traditional and other informal land claims, notably as a means of protecting their livelihoods in ways that required conservation of forest resources ( Pace, 1992 ) Such movements include indigeno us peoples, forest extractivists, and even frontier agricultural colonists. These movements offered various proposals for sustainable


97 development such as via synergies between sustainable production forest conservation, ( Campos & Nepstad 2006 ; Hall, 1997 ) Such proposals various sought to combine forest extractivism and agricultural production, sometimes via agroforestry, in order to diversify production systems and income streams while conserving forest cover. Such initiatives were heavily influenced by bottom up perspectives on conservation and development notably CBNRM. This perspective emphasizes the advantages of communities and local people in natural resource management ( Western & Wright, 1994 ) The inclusion of local people in initiatives for conservation and development often rests on arguments such as detailed knowledge of local res ources, the interest communities have in their own development, and longstanding experience in local resource management ( Bailey, Headland, & Sponsel, 1996 ; Berkes, 1991 ; Western & Wright, 1994 ) CBNRM has received some criticism due to assumptions about community harmony and observations of limited community capa city ( Agrawal & Gibson, 1999 ) Nonetheless, CBNRM encompasses ideas and practices that have frequently been incorporated into conservation and development initiatives and appropriated by social movements and grassroots organizations. Attention to CBNRM and social movements in turn led to discussions of the importance and forms of grassroots organizations in sustainable development initiatives ( Uphoff, 1993 ) The community resource management, social mobilization and grassroots organizations all highlight questions concerning the viability of collective action as a means for c onservation and development ( Houtzager, 2001 ; White & Runge, 1995 ) At the same time, a large literature on social capital has emerged, and its


98 en counter with work on collective action provides a useful framework for considering the prospects for grassroots organizations and sustainable development initiatives. Social Capital This research focuses on social capital as a key element for understanding the effectiveness of grassroots organizations in the pursuit of sustainable development initiatives. A broad definition of social capital focuses on the relations of trust and the resulting civil society organizations or networks that generate economic e fficiency by reducing transaction costs and/or make the state more accountable ( Coleman, 1990 ; Putnam, 1993 ) Such broad definitions obscure various aspects of social capital, and have led to critiques and typologies in order to unpac k the concept for empirical analysis. Harriss and De Renzio ( 1997 ) propose one such typology: (1) family and kinship relations, (2) differences in sector and power, (4) political capital, (5) institutional and policy frameworks and (6) social norms and values. Uphoff and Wijayarat na ( 2000 ) also propose a set of categories of social capital: structural and cognitive. The structural forms of social capital are external or objectified such as social networks, rules and procedures. And cognit ive social capital consists of norms, values, beliefs and attitudes that predispose people to cooperate. The identification of varieties of social capital is crucial when one seeks to analyze social processes operating on multiple levels. Whereas some for mulations of social capital still highlight trust between individuals, social capital is increasingly studied at the level of local organizations or regional federations. Carrol and Bebbington ( 2000 ) draw on this distinction to analyze the role of federations or second level


99 organizations in sustainable development. Since their focus is on social organizations, they develop their discussion by focusing on structural social capital. These authors posit that there are six dimensions of structural social capital, which can be divided into two groups: internal relationships within a social organization and the external relationships involving alliances between the organization with other socia l actors. They show that both facets of social capital are important for understanding organizational effectiveness. This paper therefore draws on their multi level perspective, highlighting how social capital within an organization as well as relations of an organization with other collective entities can affect organizational effectiveness. There are nonetheless questions about the creation of and interactions among various forms of social capital. Portes ( 1998 ) highlights the contribution of Bourdieu by to an extent be transformed into othe r forms of capital. This fungibility can also be applied to different types of social capital. This implies that different forms of social capital can be transformed and allocated in order to achieve organizational goals. However, scholars have also been c areful to point out that different combinations of social capitals can themselves yield different outcomes in regards to development initiatives ( Woolcock, 1998 ) Further, given the important o f a multi level perspective on social capital, also it is also worth noting that other analyses have shown that policy interventions can also create social capital in local organizations, though there remain issues as to whether policies can be designed w ith that goal in mind ( Portes & Landolt, 2000 )

PAGE 100

100 These observations highlight the impo rtance of understanding social capital as it manifests in various forms in grassroots organizations. In the case of grassroots organizations, social capital is not only embodied in the social relations among individual members. At the organizational level, social capital is also evident in the ability of members to act collectively toward a shared goal. A good example comes from Bebbington and Perrault ( 1999 ) who focus on the role of structural social capital as a means for an organization to access new resources. This study, conducted in the Ecuadorian Andes, shows how increasing structural social capital, specifically involving alliances with external actors, allowed Andean communities to gain access to other types of capital such as natural resources, which proved cru cial for livelihoods. Recognition of structural social capital in turn implies the need to acknowledge the broader political context of a given organization. Most analyses of social capital focus on micro level social relations, and even structural account s tend to feature relations among a pair of organizations. Still needed is an account of the broader context in which these relations and organizations exist. Such perspectives understate the ways in which local, regional and national contexts can influenc e grassroots organizations and their effectiveness. Political Ecology Therefore, it is important to consider additional approaches that highlight social organizations in the context of other social actors, as well as the policies and markets. For such purp oses, political ecology is a useful approach due to its stress on the importance of contextual factors as they affect local social actors such as grassroots. Political ecology calls particular attention to interactions among social actors in light of their interests with regard to the use of natural resources. Political ecology also features

PAGE 101

101 cross scale processes arising from such interactions to underscore how action at different scales can influence actors on other scales, especially localities ( Bryant, 1992 ; Bryant & Bailey, 1997 ) Given the importance of the state, political ecology pays particular attention to actors at the national level ( Bryant & Bailey, 1997 ) National governments have the mandate and capacity to promulgate policies and programs as via interactions with many othe r organizations. In frontier areas, the state is a pre eminent actor due to geopolitical concerns about control of borders but also for promulgation of regional development policies to incorporate national peripheries into the national economy. Even if the state has a limited presence in frontier areas, the limited capacities of local governments often still imply that national governments are key actors in conservation and development initiatives. The Amazon region is a good example for although the Govern ment of Peru has long had centralist tendencies, it has nonetheless at times promoted frontier land settlements supported by policies to provide resources for technical and financial assistance ( Rodriguez, 1991 ) More recently, the Peruvian transp ortation infrastructure that facilitated market integration and the exploitation of natural resources ( Santa Gadea & Wagner, 2002 ) New infrastructure projects have proceeded alongside new legislation to improve gov ernance of natural resources, notably in the forestry sector for timber and non timber forest products ( Cosso, 2009 ) At the same time, the Government of Peru has pursued creation of protected areas in its Amazon territory. Hence the state serves multiple roles i n conservation and

PAGE 102

102 development, as both facilitator of natural resource exploitation as well as conservationist ( Bryant & Bailey, 1997 ) Political ecology extends its analytical focus beyond the nation state to the international (global) scale. International actors include transnational corporations, multilateral banks, aid agencies, and conservation organizations. Multilateral agencies have been especially important actors with regard to use of natural resources in frontier regions, first by funding development projects with negative im pacts and later by pushing policies such as structural adjustment programs that influenced actors at the national level ( Pacheco, 2006 ; Redwood, 2002 ) To an extent such actions have facilitated the entry of highly capitalized transnational corporations and led to conflicts of various kinds with local peoples over natural resources. The size of these actors increases their impacts on the exploitation of n atural resources as well as their capacity to influence national government policies ( Bryant & Bailey, 1997 ) Transnational corporations respond to international market dynamics, and sometimes these markets can exert influence through national or regional companies as the cases of beef o n Brazil ( Nepstad, Stickler, & Almeida, 2006 ) International markets can in turn consti tute new interest groups, such as when market prices change to motivate migration in the search for heretofore unattractive natural resources. One case in point in the Peruvian context is the recent boom in gold mining in Madre de Dios ( Mosquera, Chavez, Pachas, & Moschela, 2009 ) In the face of pressures for unsustainable development, international environmental and human rights organizat ions have become key players in the political ecology literature ( Bryant & Bailey, 1997 ) Political ecology highlights the emergence of alliances between local groups and international NGOs, including the

PAGE 103

103 formation of transnational advocacy networks ( Keck & Sikkink, 1999 ) Inte rnational environmental and human rights organizations have become close allies with local social movements, and together they have often advocated creation of sustainable use areas, recognition of traditional resource use, and sustainable alternatives to government policies ( Schwartzman, Alencar, Zarin, & Santos Souza, 2010 ; Schwartzman & Zimmerman, 2005 ) Together, social capital and political ecology constitute a theoretical framework that permits an evaluation of the effectiveness of grassroots organizations to contribute to conservation and development initiatives. Whereas social capital permits anal ysis of cognitive and structural aspects of social ties among members within grassroots organizations, political ecology complements social capital by going beyond relations among pairs of organizations to consider the broader context of a given organizati on, up to the international level. Social capital thus highlights internal factors affecting the effectiveness of a grassroots organization, while political ecology situates grassroots organizations in the context of other organizations, the state and its policies, and markets. It thus becomes possible to identify multiple reasons for the effectiveness (or lack thereof) of grassroots organizations. The Emergence of Grassroots Organizations in the Southeastern Peruvian Amazon To pursue an analysis of how soc ial organizations and their broader context can serve as both enablers and constraints on conservation and development initiatives, I focus on the southeastern Peruvian Amazon, specifically the department of Madre de Dios. This region has in recent years e xhibited rapid change. Until recently, the southeastern Peruvian Amazon exhibited limited growth. After significant migration

PAGE 104

104 during the rubber boom on the early 20th century, the region was depopulated. New ter ( Lossio, 2001 ) Transportation infrastructure in the region was not developed beyond horse paths until the 1960s, and gove rnment supported colonization was undertaken in a halting fashion through the 1980s ( Rodriguez, 1991 ) However, colonists kept coming into the region for agriculture. In addition, beginning in the 1980s, southeastern Peru exhibited extraction of natural gold ( Lanao, 1998 ) Agriculture in Madre de Dios is mainly for sub sistence, with a small portion of production sold in regional markets. Local productivity is low, and the percentage of land used in landholdings is small. Cattle ranching is also incipient in the region; with a few rare exceptions, most households have on ly a few heads of cattle, and ranching usually serves as a type of insurance ( GOREMAD, 2002 ) Despite limited agricultural dev elopment, agriculture is important in Madre de Dios because it provides food security and land for rural families and eventually will go to the next generation. Despite its relative isolation and sparse population, agriculture in Madre de Dios has been hea vily influenced by government policies. Until early 1990s, national governments pursued interventionist policies by deploying a variety of policy tools in order to support and modernize agriculture. Agrarian reform, programs of subsidized credit, and the d eployment of price guarantees, are some of the many policies that national governments employed to promote agriculture ( Arias, 1988 ) In addition, the Peruvian Amazon was the target of specific regional policies. Notable was the creation

PAGE 105

105 ales in Spanish), that provided technical assistance and developed infrastructure for agriculture ( Rodriguez, 1991 ) The emphasis on governmental interventionism changed in the 1990s due to economic crisis and fiscal austerity, which led to a structural adjustment programs in Peru ( Gonzales de Olarte, 1996 ) An economic shock exerted a negative impact on demand for agricu ltural products, sending prices lower. Following structural adjustment, agricultural credit was withdrawn along with price controls on agricultural inputs and products ( Figueroa, 1992 ) These neoliberal policies related to agriculture were maintained for most of the 1990s. While the re were also some isolated initiatives that aimed to promote agriculture, they were seen as populist gestures rather than serious policies. During the 2000s, the national government maintained a general policy of non intervention in agriculture, expecting that markets would provide incentives to those agricultural sectors with sufficient productivity to be competitive. One important difference for the Amazon region during this period was the increasing prices of key natural resources such as oil, timber and gold. This dynamic made the Peruvian Amazon more important for national government administrations. By the late 2000s, national governments were again pushing policies aimed to facilitate exploitation of these natural resources. Such policies generated im portant social conflicts over questions of conservation and development among social actors in Madre de Dios and elsewhere in the Peruvian Amazon. From the 1980s, the new influx of population and the rapid economic and policy dynamics combined to create co nditions favorable for the creation of grassroots

PAGE 106

106 organizations in Madre de Dios. Government led agrarian reform in the 1980s facilitated access to land among colonists, making government and farmers allies. Indeed, farmers organized associations precisely to better engage government and to take advantage of government policies. Farmer organizations thus initially focused their activities on securing lands granted by agrarian reform, and then sought policies promoting agriculture. When later administration s shifted to neoliberal policies in the 1990s, farmer organizations proved less effective in procuring favorable terms in markets for their products, and producer cooperatives failed. This circumstance produced negative perceptions about collective initiat ives among small farmers. It was under this form of federations of local associations. Second level organizations took a more political stand as they sought to confront go vernments with political and social demands as a means of reversing state withdrawal of support for agriculture. In the case of Madre de Dios, the path for farmer organizations was somewhat different than that seen elsewhere. The main organizations for fa rmers were the community associations. Colonists created such associations, although each colonist had private individual properties. Colonist neighbors got together by organizing associations around centralized community centers in order to deal with a d iversity of shared community level issues such as schools, health assistance, etc. While associations permitted collective action concerning community issues, they proved insufficient for larger political economic questions. In that context, the farmer fed eration of Madre de Dios, or FADEMAD, was created as a larger scale organization in order to

PAGE 107

107 address regional issues. In so doing, FADEMAD followed a path that entailed a new approach to understanding the role of social organizations. Research Methods This chapter focuses on FADEMAD and its history, with an emphasis on its effectiveness as a grassroots organization in pursuing conservation and development initiatives. FADEMAD serves as an instructive case study because it has experienced substantial changes in its fortunes, reflected in its ability to engage in collective action with successful outcomes. In particular, FADEMAD experiences some early successes, but in recent years has faced difficulties in more recent initiatives. I therefore employ the theor etical framework described earlier to explain these changing fortunes in order to identify whether internal factors tied to social capital or external conditions related to political ecology or both account for these changes in the effectiveness of FAD EMAD. A key design element is to follow the evolution of FADEMAD over time, in order to monitor the dynamics in its internal network and shifting external context. It is likely a priority on a methodology that permits collection of comparable data available over time. My methodology therefore focuses on data sources that permit review of the trajectory of FADEMAD for about a 20 year period, from its founding to 2010. I divided the trajectory of FADEMAD into three stages with regard to its effectiveness in pursuing conservation and development initiatives. The first phase involves the founding and emergence of FADEMAD and runs from 1990 to 1993. During this period, FADEMAD establ ished relations with other social actors, notably other community associations and governments on various levels. The second period ran

PAGE 108

108 from approximately 1993 to 2000. While the start and end dates should be viewed as rough approximations, these years bra cket a period when FADEMAD exhibited internal coherence in formulating its priorities and when FADEMAD was able to successfully pursue conservation and development initiatives by working with the Peruvian state and NGOs. Notable during this period was FADE and buffer zones in Madre de Dios. The third and final period in this analysis is the 2000s decade, from roughly 2001 to 2010. This period is marked by fragmentation within FADEMAD as manifest via internal conflict s and more limited participation, as well as a shifting political ecology in terms of state policies and market prices as well as To better grasp the internal and external factors that might account for Specifically, I draw from three main sources: archival work with FADEMAD documents, semi and participant observation of FADEMAD meetings (Bernard 2002). My goal in using these sources is to be able to gather complementary information about internal and external factors tied to FADEMAD and its conservation and development initiatives, especia lly over time from its founding. The archival work involved a review of documents, featuring project reports and Like other grassroots organizations, FADEMAD leadership was c areful to record agenda items, reports from committees, presentations by leaders and guests, and

PAGE 109

109 constitute a more or less standardized record of information concerning F ADEMAD activities over a long period of time, spanning its founding, its early successes and its especially concerning internal FADEMAD deliberations. In addition to the FADE MAD archives, I consulted various other secondary data sources including government documents, economic indicators, and published academic work on Peruvian development policies. These documents provide a picture of the broader political and economic contex t in which FADEMAD has operated since roughly 1990. I also conducted 19 semi structured interviews to complement the archival sources. The semi structured interviews involved current and former FADEMAD leaders as well as practitioners from NGOs who were a lso involved in FADEMAD projects. I secured many of these interviews as a result of having worked in Madre de Dios for several years, including with community associations that were in turn FADEMAD members. Interviews with FADEMAD leaders were important to provide additional information concerning the FADEMAD records by offering perspectives and reflections on e.g. why certain decisions were taken. In particular, these interviews permitted collection of information with reflections on why proposed initiativ es were pursued (or not) and why those pursued succeeded or failed. This provided additional information concerning issues related to social capital within FADEMAD. By contrast, interviews with personnel from NGOs that have worked with FADEMAD provided inf ormation on the broader regional context, thus offering the view of FADEMAD from the outside. In both cases, key informant interviews focused on people who had experience with FADEMAD at various points in its history. While relatively few interviewees coul d speak

PAGE 110

110 Finally, I conducted participant observation of FADEMAD meetings. Like the semi structured interviews, I built on my previous experience in Madre de Dios and with FADEMAD to be able to make such participant observations. Participation in FADEMAD meetings complements the archival work and interviews by affording direct observation of deliberatio ns, without reinterpretations by selected individuals. Direct observation in surprising given its earlier successes. Direct observation also allows for a direct appraisal of the social capital within FADEMAD as manifest in the tenor of meetings, both in terms of the extent of member attendance as well as active participation. Together, the archival work, in depth interviews and participant observations provide information about t he internal and external factors related to FADEMAD and its conservation and development initiatives. The archives provide data sources spanning the time frame under consideration, featuring internal FADEMAD deliberations but also allowing for FADEMAD to b e situated in its broader political and economic context. The interviews permit more informed interpretations of the records from the archival work, both from within and outside FADEMAD. And the participant observations complement the records and interview s by affording direct observation of contemporary FADEMAD proceedings. Each method permits a degree of triangulation with the others to evaluate how internal and external factors can enable or hamper conservation and development initiatives promoted at dif ferent moments in time by FADEMAD as a grassroots organization.

PAGE 111

111 The Emerging of a Pro In the remainder of this chapter, I provide an historical account of FADEMAD through the three periods noted above. In each, I draw from multiple data sources to describe and then evaluate FADEMAD and its effectiveness at pursuing conservation and development initiatives. I draw on the theoretical framework outlined earlier, organizing my interpretation in terms of internal factors ti ed to social capital and external factors as emphasized in political ecology. The historical organization of the discussion in turn permits identification of how social capital and political ecology may change; this permits an explanation for both successf ul initiatives, where internal and/or external factors serve as enablers of conservation and development initiatives, as well as difficulties, where explanatory factors become constraints. I begin with the 1990 1993 period. The early 1990s was a period of crucial changes in the organization of farmers in Madre de Dios. At this time Farmers in Madre de Dios were confronted by a complicated situation involving several policy changes being promulgated by the Peruvian state. The first of the changes was the pr oposed creation of protected areas in Madre de Dios. On January 26, 1990, the national government created the Tambopata Candamo Reserved Zone (ZRTC by its Spanish name) within the Madre de Dios region. The ZTRC was to be a new protected area that encompass ed 1,472,942 hectares. The de Dios rivers, along with numerous Brazil nuts concessions, and some indigenous communities. This government decision was taken without any consul tations with local peoples and thus came as a surprise for stakeholder groups in the region. Because most

PAGE 112

112 colonists in Madre de Dios still lacked clear titles to the lands they claimed, the proposed ZTRC was a frightening prospect as its establishment woul d threaten producer land claims and thus create a conflict between conservation and development initiatives. As a result, the proposal for the ZTRC without prior consultations stimulated a substantial outcry of protest among local people ( Moore, Chicchon, & Salas, 1995 ) But this was not all. At the same time, Peru wa s experiencing economic shocks and instability. to the Peruvian state, which in turn forced budget cuts and scaling back of policy programs on which many agricultural prod ucers relied. Of particular importance was the shutdown of the Agrarian Bank, a key source of credit for farmers in Madre de Dios and elsewhere. This left farmers unable to borrow for new investments, and inflation inflicted further problems by ballooning previous debts, further curtailing prospects for planting and production. Additionally, the activities of the government agency in charge of buying rice production (ECASA) were drastically diminished 1 Whereas rice had previously been bought at guaranteed prices, cuts at ECASA in effect removed a key buyer from the market. This was a key event since ECASA was the main buyer for rice production in Madre de Dios, where rice productivity is relatively low. As a result, farmers in Madre de Dios seeking to sell rice were suddenly uncompetitive in the market 2 ECASA was not the only such casualty; ENCI, the government agency in charge of selling inputs for agriculture with controlled prices, was also shut down. 1 Interview with former leader. Interview not to be released to the public. 2 Interview with practitioner. Interview not to be released to the public.

PAGE 113

113 Generalized economic instability, along with curtaile d state credit and purchasing programs, constituted a dire situation for farmers in Madre de Dios. In this circumstance, community associations mobilized in order to better organize at the regional scale in order to engage the state. A mobilization committ a period of a few months, the members of this committee were selected to lead FADEMAD. This new leadership included leaders from older settlements arou nd the regional capital of Puerto Maldonado as well as new settlements, some of them from the Andean region 3 The leadership of FADEMAD focused its attention on the changes in state policies and sought redress in order to improve the situation of farmers i n Madre de Dios. FADEMAD was able to counteract some of the actions against farmers that owed money to the Agrarian Bank. FADEMAD also demanded the reactivation of ECASA with greater control by farmers; the government promised to follow up, but this promis e was never fulfilled 4 Hence with regard to state institutions on questions related to production (such as credit and marketing), FADEMAD had a mixed record of success and failure. However, the most important result of farmer mobilization and organization via FADEMAD was related to conservation issues via the establishment of the ZRTC. Social mobilization to contest the establishment of this protected area led to organization of two public forums where civil society, government agencies, and NGOs spent sev eral days discussing the details and process for establishment of the ZRTC 5 3 Personal observation. 4 Interview with former leader. Interview not to be released to the public. 5 Interview with practitioner. Interview not to be released to the public.

PAGE 114

114 The public forums, held in 1991, were the first of this type implemented in the Peruvian Amazon. Although many issues were debated in this workshop, the key conclusion was to rec ognize the need to incorporate local peoples in the process of establishing new protected areas. Further, consultative processes required studies that account for the populations established in and around the proposed protected area. Aside from these proce dural conclusions, the forum was also a crucial encounter among farmers and NGOs involved in different ways in work eminently relevant to conservation and development 6 This initial encounter led to later such meetings that in turn generated concrete outco mes in the form of sustainable development projects in Madre de Dios. In 1992, another regionally important conservation and development issue arose, to which FADEMAD was able to mobilize a coherent response. This time it was the proposed paving of the roa d through Madre de Dios that would at long last break its isolation from the rest of Peru. According to regional people, it was a commonplace belief that the lack of reliable, efficient, year round access was the main cause of the regional economic stagnat ion. As a result, it was also a standard promise by politicians and governmental institutions to improve roads in Madre de Dios. When regional authorities refused to advance the building of roads in 1992, the local population under ertook the task. Farmers from different communities marched to the end of the road beyond Mazuko and worked on the road for several weeks, supported by the Catholic Church. This initiative forced regional authorities to reiterate their commitment to buildi ng the road 7 6 Interv iew with practitioner. Interview not to be released to the public. 7 Interview with former leaders. Interview no to be released to the public.

PAGE 115

115 government. These events consolidated the p articipation of many community associations under the FADEMAD umbrella. Further, this instance highlighted yet other social actors such as the Catholic Church in positive and collaborative initiatives. By this time, FADEMAD had begun to gain a reputation as an effective grassroots organization, which created new opportunities. In 1993, with support from USAID, the McArthur Foundation and Conservation International, FADEMAD launched a project to identify land use capacities in the farming communities. The project had a strong participatory approach and involved a significant amount of work by farmers, FADEMAD leaders, and NGO practitioners. The events during 1990 1993 thus e ncompass the formation of FADEMAD in response to desperate circumstances, followed by an initial series of positive experiences involving mobilization led by FADEMAD. The scenario described above is by no means unique to Madre de Dios; indeed, it replicate s similar processes involved in the emergence social organizations in Latin America since the 1980s. Grassroots organizations have often emerged out of social mobilization as engines for organized protest against unsympathetic government policies; grassroo ts organizations then garner the support from local and international NGOs ( Fow eraker, 2001 ) ; and alliances such as these are able to confront governments and win at least some concessions. Such outcomes have in other cases been held out as the result of a combination of

PAGE 116

116 circumstances tied to solidarity that supports collective act ion, complemented by external support. Key in many such accounts is the confluence among organizations around a set of shared interests. In the case of FADEMAD, this occurred with regard to concerns about a proposed protected area; it became clear to conse rvation NGOs that establishing the protected area would not be feasible without the support from local communities, represented by grassroots organizations such as FADEMAD. The social environmental NGOs to confront the state. In turn, the alliance was also important to FADEMAD because NGOs could bring in important technical capacities and financial resources. This in turn led to downstream activities where FADEMAD was able to leverage those resources and technical capacity to the benefit of its members. Hence there was a positive feedback from social capital supporting grassroots organization to attract the attention and then the support of external organizations, which then provided re sources and other benefits that validated initial mobilization and further strengthened social capital in FADEMAD. It is thus evident that the initial organizational tasks that set in motion constitution of FADEMAD relied heavily on social capital and requ ired that FADEMAD be responsive to the priorities of local communities. The creation of a grassroots organization can be considered as an indicator of significant structural social capital. That said, sustaining the legitimacy of FADEMAD as a grassroots or ganization also required expanding its network of support. This not only implied establishing new linkages and strengthening old ones at the local and regional level, meaning reaching new colonist communities and mobilizing others already incorporated to F ADEMAD. It

PAGE 117

117 also implied creation of external linkages with other local actors, as was the case of the local Catholic Church, and with yet others outside of Madre de Dios, such as conservation NGOs. While expanded local support was important, external supp ort was also crucial; FADEMAD thus relied on a favorable context where such support could be found and procured. The emergence of FADEMAD also highlights the important role played by the context, and this becomes clearer by following a political ecology ap proach that accounts for actors and processes operating on different scales. At the international level, there was increasing interest from international development agencies to support conservation initiatives that take into account local communities and community based projects Further, there was interest among international conservation NGOs specifically in Madre de Dios, and this facilitated the flow of financial reso urces for FADEMAD initiatives. On the national and regional level, government polici es played an important role due to their lack of support. The closure of the Agrarian Bank and other national agencies, and the refusal of the regional government to support farmers via road maintenance, proved crucial as flashpoints for collective mobiliz ation. Hence the lack of state support served to foment collective mobilization and solidarity, while the existence of international support promoted effective collaborative action. Sustainable Development Initiatives By the end of its first period, FADEM AD had shifted its focus from leading protests against state negligence to actively participating in collaborative projects with a positive agenda ( Stronza, 1996 ) However, that agenda increasingly took on a dual aspect. Indeed, some former FADEMAD leaders stress the idea of one organization

PAGE 118

118 with two arms, one being the usual federation activities wherein the regional razo via the formation and management of alliances for conservation and development projects. During the second period, which spans the late 1990s, the main efforts of FADEMAD increasingly sought to leverage its internal social capital by focusing on collaborative conservation and development projects. That is, FADEMAD became increasingly externally focused on alliances and projects. The main project carried out by FA DEMAD was a participatory project focusing on land use classification, for the sake of sustainable development in the buffer zone of Mayor de la Tierra y Desarrollo Sostenible en Areas de Influencia Humana de la Zona first, it sought to identify and select practitioners from participating communities in order to train them in land use evaluation and p lanning, and then, the community practitioners would work with communities to collect data about their lands in order to identify the most sustainably productive land uses. The data collected would then be returned to the participating communities in the f orm of maps that would guide land use decisions made by farmers ( Arce, Flores, Lossio, & Vilchez, 1996a ) The land us e project also included commercialization studies intended to help farmers become better integrated in markets ( A rce, Flores, Lossio, & Vilchez, 1996b ) This project was the first in a rapid succession of initiatives that not only responded to the priorities of community associations but took

PAGE 119

119 advantage of a favorable political ecology involving external funding for projects led by grassroots organizations. in Spanish). Like the land use initiative, this project was created from funds provided by international cooperation, and was intended to provide credit to farmers for developing agriculture and other activities in their farms 8 The logic of this undertaking was related to the experience of the closure of the Agrarian Bank and the ensuing difficulties for farmers, and the perception that agricultural stagnation in the region was due to the lack of financial support 9 This Pilot Fund project met with initial success, so FADEMAD continued to seek and obtain funding for small farmer credit. This positive experience in turn led to the implemen tation of CREDISMAD, a credit agency for farmers developed with the support of Canadian cooperation. CREDISMAD since grew in importance because during this period it was the only credit outlet for small farmers 10 It is important to stress that these projec ts also provided resources to build was always eager to connect with its members, and to this purpose FADEMAD also implemented a weekly radio program as well as a small monthly pamphlet that was distributed in communities 11 dvice with news about 8 Interview with former leader. Interview not to be released to the public. 9 Interview with practitioner. Interv iew not to be released to the public. 10 Personal observation. 11 Interview with practitioner. Interview not to be released to the public.

PAGE 120

120 FADEMAD activities plus other services to farmers such as messages and errands. The discussing issues related to land use and conservation, all in a language and presentation appropriate for farmers. Project investments and FADEMAD initiatives thus legitimacy among farmers. The organic activity of FADEMAD was also very dyn amic during the project implementation period. Most of the leadership kept its positions during this period; it was attained through reelections at General Assemblies almost without opposition. The participation of the 85 member communities of FADEMAD was regular and consistent. The visits of leaders to communities always gathered a majority of farmers in community assemblies. It was not uncommon to have arguments and complaints concerning some presentations by FADEMAD leaders, but the roles and the importa nce 12 While FADEMAD was managing successful projects via international cooperation, it also took on another role highlighting the favorable political ecology in the late 1990s, this t ime involving relations with the regional government. FADEMAD already occupied the role of regional representative for community associations. Increasingly, this role became one of mediating conflicts between farmers or associations and the regional govern ment. FADEMAD leaders increasingly spent time serving as brokers between farmers and government agencies, as by walking farmers 12

PAGE 121

121 to government offices in order to present a complaint 13 Similarly, it was possible for farmers to get assistance in FADEMAD offi ces to complete any paperwork required for action by government agencies. At the same time, FADEMAD served as a gateway for the regional government to reach out to farmers and associations. It was not uncommon for government officers consult FADEMAD leader s as intermediaries when they needed to reach farmers. This consultative role for FADEMAD also applied to national de Dios, it also consulted with FADEMAD 14 Hence the fa vorable political ecology surrounding FADEMAD involved not only cooperation and funding from international organizations, it also included constructive relationships with regional and national governmental agencies. Following Carrol and Bebbington ( 2000 ) it is clear how during this period there is a clear synergy between two types of structural social capital, namely the internal linkages built around FADEMAD among its member communities and the external alliances FADEMAD forged with other organizations. Indeed, per political ecology, those alliances involved relations not only with local and regional organizations but also national and international entities. The network of NGOs greatly contributed to projects carried out by FADEMAD, and those projects in turn improved the relations between FADEMAD and its member communities, especially by involving the communities in the projects. The resources brought by the projects also enhanced the r ole of FADEMAD 13 Personal observation. 14 Interview with practitioners. Interviews not to be released to the public.

PAGE 122

122 which in time made FADEMAD an interesting partner for external organizations and even for the government. Hence there were positive synergies between FADE legitimacy with the other. During the late 1990s, this permitted several successful conservation and development initiatives ranging from land use planning in the b uffer zone of a protected area to improved production and marketing practices for farmers. During the late 1990s, structural social capital and the broader political ecology constituted enabling factors for FADEMAD as a grassroots organization to pursue co nservation and development initiatives. This can be considered as the ideal scenario, where all actors involved win. The Sustainability of Social Movement Organizations If the 1990s constituted a period of implementation of successful projects that built and leadership. By the 2000s, FADEMAD had entered a period of difficulties arising from both internal fractures that undermined its social capital and thus its legitimacy, as well as external challenges concerning the broader political ecology of relationships with other organizations. The first indications of problems arose as early as 1997. The land use project began the process of returning results to communities. Despite the highly participatory process, the land use maps were scaled for entire communities but did not provide much detail on the scale of individual properties of community member s. As a result, debate arose over the project as individuals found the maps to be useless because they were interested in classifications for their individual landholdings. Practitioners and

PAGE 123

123 leaders involved in this project responded by pointing out that t always been a classification at the community level. They added that while it was certainly possible to refine the classification by conducting it at the level of individual landholdings, such classifications would require collec ting more field data and more processing time 15 objective of the land use classification project was never to conduct classifications for individual properties. According to b oth sources, the debate that unfolded after 1997 presenting the project in terms of community land use classifications, but also claiming that such classifications would help farmers to decide how to best use their lands. But since farmers decide their land uses on the scale of individual properties, they had understood that the classification would be conducted or at least applicable among individual landholdings. Althou gh it was stressed that studies on commercialization and production costs were also conducted in the project, there was a sense of offers unfulfilled among farmers. By the early 2000s, the difficulties seemed to be multiplying. For one thing, CREDISMAD, th e credit operator for FADEMAD, was also facing difficulties. CREDISMAD increased the volume of credit handed to farmers and local people, which farmers liked, but the increase in volume implied more credit extended to farmers unable to repay the loans ( FADEMAD, 2000 2012 ) As a result, by the 2000s, CREDISMAD faced a rise in unpaid loans and a prospective credit crisis. This situation 15 Interview with practitioner. Interview not to be released to the public.

PAGE 124

124 was exacerbated by a structural contradiction between CREDISMAD and FADEMAD. Since CREDISMAD was the financial branch of FADEMAD, CREDISMAD was ma ndated to support farmers; but due to its relationship with FADEMAD, CREDISMAD also faced constraints on compulsory measure to recover loans. This then led to discussions and conflicts among FADEMAD leaders over the primary role of CREDISMAD for FADEMAD me mandate in support of an advocacy role for CREDISMAD, others emphasized the threat and thus were inclined to take a more business like perspective. More broadly, many farmers started to question CREDISMAD under the assumption that its funds were donation rather than loans, so therefore there was no obligation to repay loans 16 This popular sentiment ensured that CREDISMAD was no longe r sustainable. Then its main funding source, the Canadian cooperation in Peru, was restructured, rendering CREDISMAD unviable, and it quickly shut down. The collapse of the relationship between FADEMAD and the Canadian cooperation in Peru was not the only external alliance facing difficulties for FADEMAD in the 2000s. In addition, the collaboration between FADEMAD and conservation NGOs underwent key structural changes. After the land use classification project and the other projects of the 1990s, there were no further important projects run by FADEMAD 17 This is not to say that FADEMAD had no further projects, or that collaboration with 16 17 Interview with practitioner. Interview not to be released to the pub lic.

PAGE 125

125 conservation NGOs ceased. Rather, collaborative projects were placed under NGOs control. One example concerns the consultat ive process surrounding the fate of the proposed protected areas in Madre de Dios. One part of the Tambopata Candamo Reserve Zone became the Bahuaja Sonene National Park in 1996, and the process was completed in 2000 with the creation of the Tambopata Nati onal Reserve. These two protected areas encompass most of the area on the initial proposal of the Tambopata Candamo Reserve Zone. These advanced stages of the process involved more conservation NGOs, but less participation by FADEMAD and its members 18 As a result, collaboration between FADEMAD and the conservation NGOs became more discussion among leaders about the nature of grassroots collaboration with NGOs. This became part o f a broader discussion about local grassroots leadership concerning rural issues in Madre de Dios. However, as collaboration with NGOs became nonexistent, this discussion never reached a conclusive ending 19 The problematic circumstances of the land use pro ject, the closure of CREDISMAD, and the growing distance to conservation NGOs collectively constituted a problematic political ecology for FADEMAD. The alliances, the collaborative learning, and the resources involved were thus curtailed. In turn, this mad e the pursuit of conservation and development initiatives problematic 20 Just as there had been positive 18 Interview with practitioner. Interview not to be released to the public. 19 20 Interview with current leader. Interview not to be released to the public.

PAGE 126

126 its members in the 1990s, the decline in external support led to a more negative synergy in the 2000s. As external support dropped, the means FADEMAD used to reach its bases were also discontinued in the 2000s. Whereas the lack of project funding led to a shortage of practitioners to design and work on the land use clas sification, the radio program was closed due to lack of funding. These difficult information to its base, and in turn for community associations to convey their needs to FADEMA D. This in turn led to yet other negative synergies such as the reduction in government. As a result, social capital within FADEMAD and among FADEMAD leaders and member communities also began to erode. There were several indications of this. For one thing, FADEMAD leaders engaged in increasingly heated debates over FADEMAD, which were complicated by a transition in leadership. The leaders who had served in the 1990s and initiated th e projects and consolidated relationships with external allies stepped down, as a consequence of the normal wearing out of a leadership that had been running the organization for 8 or 9 years, as well as the difficulties facing many activities. FADEMAD hel d elections in the early 2000s, and they were consequently hotly contested among aspiring new leaders. The elections were marked by numerous charges of mishandling of funds, nepotism, and other allegations 21 Once the elections were settled, a new leadershi p was established with many aspirations but with 21 Personal observation.

PAGE 127

127 diminished operational capacity and no external funding 22 In other words, the new leadership faced a difficult situation involving damage to its internal social capital and an external political ecology with limited opportunities for support. As a result, by the mid vident in participation in FADEMAD meetings and membership. Instead, some communities became more independent of FADEMAD, establishing their own direct relationships with governmental agencies and NGOs. This is not to say that in recent years, FADEMAD has lost all capacity to serve in its roles as a regional grassroots organization. Indeed, FADEMAD has maintained its capacity as a political operator in the region. In 2008, FADEMAD was an important actor in organizing major strikes against policies enacted by the Government of Peru, which sought to facilitate outside corporate investments in the exploitation of natural resources in the Amazon. The strikes were widespread, involving both indigenous groups and colonists all around the Peruvian Amazon region. F ADEMAD was instrumental in leading protests against the national government to demand action by the regional government of Madre de Dios. The details of the ensuing events go beyond the scope of this chapter; for present purposes, it is noteworthy that FAD EMAD still had the legitimacy in the late 2000s to organize a large strike on behalf of local people 23 22 Interview with current leader. Interview not to be released to the public. 23 Interview with current leader. Interview not to be released to the public.

PAGE 128

128 Analyzing this period using the theoretical framework outlined earlier reveals several key elements to the dynamics at play and their explanations. The first insight concerns the importance of the external political ecology for the structural social capital necessary for the maintenance of a grassroots organization. FADEMAD as a second level organization not only relies on the social capital stemming from a solidary support strengthened internal solidarity, in the 2000s the lack of external su pport undermined internal social capital. These observations raise important implications about the relationships of the dynamics of external political ecology and internal social capital in grassroots organizations. The history of FADEMAD casts doubt on notions that external support can offset internal fragilities. Whereas earlier projects made investments in organizational capacity of FADEMAD, it is slightly surprising that the resulting improvements in organizational capacity could not offset later diff iculties in procuring external support. This leaves open questions about the nature of those investments and whether they in fact are the best options to permit later functioning without external funding. One could also argue that external support creates dependency and hides fragility. In the case of FADEMAD, the solid external links in the 1990s allowed them not to focus on developing stronger relationships with communities. Therefore, the sudden withdrawal of external support can catalyze the collapse of a grassroots organization. Hence, external linkages can certainly have positive effects on grassroots organizations via projects and capacity building, but they also run significant risks of exerting negative

PAGE 129

129 influences on the development of linkages betw een grassroots organizations and their community constituencies. While it might seem easy to blame the grassroots organization for its failures, political ecology requires that they be seen in their broader context. It bears noting that shifts in markets, policies and external funding are hardly under the control of grassroots organizations. In the 2000s, at the international level, there was no longer an abundance of funding for initiatives of grassroots organizations involved in conservation and developme nt. This implied that competition for funding was more intense and the probability of securing support was lower than before. Meanwhile, other issues such as climate change and payments for ecosystem services became more important for funding agencies. At the same time, international market prices were rising for natural resources found in Madre de Dios, notably on timber, oil, and gold. These higher prices redirected the attention and interests of local peoples concerning the exploitation of natural resour ces. The resulting booms in these commodities advanced to the detriment of the importance of agriculture, and by extension the well being of agricultural organizations of FADEMAD. At the national level, the Government of Peru promulgated a series of decree laws that facilitated these market dynamics and put further pressure on FADEMAD and other local groups seeking sustainable development. This resulted in political battles over laws seeking to facilitate investments in the unsustainable exploitation of nat ural resources. Examples include the forestry laws that increased the upper limits on the size of timber concessions, the laws that facilitate the commercialization of communal lands, and the authorizing of oil concessions on protected areas or indigenous lands.

PAGE 130

130 Even setting aside the other difficulties regarding questions of capacity, policies prompted FADEMAD to return to a mode of protest rather than projects. Enablers and Constraints The problems of grassroots organization involved in the CBNRM are al ready identified ( Agrawal & Gibson, 1999 ; Charnley & Poe, 2007 ) However, a dynamic perspective on grassroot s organizations and their involvement in conservation and development initiatives provides useful insights concerning the multiple roles of social capital and political ecology. In each case, there is the potential for these factors to act as enablers as w ell as constraints on the effectiveness of grassroots organizations to pursue conservation and development initiatives at different moments in time. What is more, there are important synergies between them, such that a favorable political ecology that affo rds collaborative opportunities and external funding can foster greater solidarity among communities and grassroots organizations, whereas a reduction in external support can also erode social capital. Even following a period of capacity building and succe ss, the withdrawal of external support can still lead to internal fractures and a decline in solidarity. That said, external conditions are beyond the control of grassroots organizations, such that the external political ecology can worsen or improve rega rdless of the internal circumstances of a grassroots organization. This situation highlights the option for social organizations changing their orientation depending on the socio economic context and internal strengths. Thus social organizations could mov e from a development organization to a politicized one. Moreover, political ecology ( Bryant, 1992 ) point out not only the importance of the external context but also politicized nature of struggles for manage and access natural resources. Therefore for social

PAGE 131

131 organization moving between development and politics should be considered as a More specific insights arise from the case of the history of FADEMAD. One observation is that local linkages between grassroots organizations and community constituencies are crucially important during the initial stages of organization, when the focus is on collective mobilization for protest and making demands. However, when a social organization intends to engage in initiatives of sustainable development, external linkages then come to play a key role by channeling resources and permitting social lea rning. Once the activities are on going, internal networks are again key to implementation and success. And to sustain projects, investments in both external collaborative ties as well as reinforcement of local solidarities are absolutely crucial; for shou ld one fail, the other will be put at great risk. The FADEMAD experience raises questions about the interaction between internal social capital and the external political ecology. In this case study, the implementation of projects included the strengtheni ng of internal networks, since communities carried out collective tasks, and communities coordinated among themselves. So it can be said that internal networks were in place while an external network was being implemented; but by the same token, when the e xternal network disappeared, the local networks eroded also. While there are important synergies in their effects on grassroots organizational capacity to pursue projects, it seems that there was little interaction or influence between them. By that I mean that the development of external networks and alliances did not facilitate capacity building of durable internal networks. This situation indicates that building of social capital is not automatic. In this

PAGE 132

132 case study, the presence of external linkages did not result in the strengthening of internal linkages to the point where they became sustainable. It also indicates that external conditions should be considered in addition to internal strengths when policies are designed to foster social capital in socia l organizations ( Portes & Landolt, 2000 ) that some capacities attained do not disappear when external support evaporates. In particular, it appears that cognitive social capital was built up through the course of the externally funded projects. This is evident in in the roles that FADEMAD still plays in Madre de Dios. The experience of e.g. land use planning, consultations regarding parks and other activities did reinforce a sense of ownership of issues tied to sustainable development in the region. Despite the decline in external funding and str uctural social capital, FADEMAD was nonetheless able to mobilize communities in protest of an external threat perceived to be a shared threat. FADEMAD thus retained its mobilization capacity among farmers stemming from cognitive social capital tied to plac e and issues of sustainability. Likewise, FADEMAD leaders continue to intermediate in everyday conflicts, and famers and local and regional governmental agencies require FADEMAD experience to interact with communities. Former leaders are also important at other levels, as some of them assumed positions as mayors in small towns or as heads of local NGOs. The experience analyzed also indicates that dynamics of markets and government policies at different levels have strong influences on the success of grassr oots organizations pursuing conservation and development initiatives. This leaves open questions of future dynamics. Whereas international markets currently have high prices on natural resources in the study region, which influence governmental

PAGE 133

133 policies s eeking unsustainable exploitation of those resources, the prospects for grassroots organizations to pursue sustainable development appear dire. That said, past dynamics have been marked by their non linearity, and additional reversals are possible. The foc us on carbon markets and payments for ecosystem services is one such possibility; if this can be aligned with sustainable development interests in Madre de Dios, there is every possibility of future renewals in alliances between grassroots organizations li ke FADEMAD and external donors, conservation NGOs, governmental agencies, and local communities. In that context, there also remain questions about government neglect and policies that threaten the interests of groups seeking sustainable development. Gover nments have been largely absent from rural areas in Madre de Dios, and grassroots organizations like FADEMAD have often focused on engaging government via protest and demands, though this has varied through time. If government policies such as the decree l aws of 2008 are indeed set to change access rules for exploitation of natural resources, this challenges the access of local peoples already in the region. We can therefore expect ongoing social mobilization focusing on protest. That will require creation of additional social capital, especially in its structural form of networks that facilitate access to technical and financial resources. If future mobilization can foster engagement for collaboration, grassroots organizations can again operate as agents fo r sustainable development. But for that to happen, it will be crucial to assure the strengthening of internal networks in order to achieve organizational sustainability. Thus, the building of internal networks should be a specific objective on initiatives of sustainable development; it should be not only a means to other ends but also an

PAGE 134

134 objective its own right. Further, this goal needs to be pursued by considering negative scenarios with regard to external alliances and support; construction of structural social capital must account for external changes beyond the control of grassroots organizations.

PAGE 135

135 CHAPTER 5 CONCLUSIONS Summary of Key Findings This dissertation took up several issues related to recent changes to the context in which environmental chan ge is occurring in southeastern Peru, as well as other regions. In the process, I focused on different levels of scale, used different kinds of data and methods, and drew on theoretical frameworks from several different disciplines and scholarly communitie s. What unifies the analytical foci of this dissertation is the shared concern with evaluating the importance of different factors operating on different scales for the sustainability of natural resource use in tropical frontier regions. The first analytic al chapter addressed the relationship of non timber forest products ( NTFPs ) and agricultural land use. NTFPs emerged on the sustainable development agenda after earlier studies suggested that products from standing forest could be exploited sustainably and yield higher incomes than agriculture or ranching. This led to debate over questions of the conditions under which NTFPs might prevail over agriculture, which are premised on notions that NTFPs and agriculture are competing land uses. This dissertation re considers those assumptions by focusing on land use and NTFPs in southeastern Peru, where households can engage in both lands with different tenure categories for different land uses. Under such conditions, it is less clear that NTFPs compete with agricult ure, though it is still an open question of whether household reliance on NTFPs implies less capacity to pursue agriculture. The findings from the first analytical chapter make evident that access to forest concessions for NTFP harvesting actually has a p ositive effect on pasture areas in the study region of Madre de Dios, Peru. Income from castaa is evidently invested in

PAGE 136

136 pasture and cattle for ranching. This reflects the commercial orientation of the key NTFP, castaa, and the security offered by investi ng in pasture for cattle seem to explain this finding. As castaa prices have recently risen (and fluctuated), this finding implies that the future fate of both NTFPs and ranching seems interwoven for at least some households in the study region. Far from a tension point, NTFPs in turn subsidize ranching. This itself constitute another dilemma for sustainable development in the Amazon: NTFPs which require conservation of standing forest in some places like forest concessions in turn lead to deforestation in other places for cattle ranching. The second part of the analysis focused on a regional evaluation of land cover trajectories in Madre de Dios. This chapter drew on theoretical frameworks from the land science community that highlight multi level and mul tivariate causation behind land cover change. I used a time series of satellite imagery to identify land cover trajectories, featuring deforestation and reforestation. The temporal analysis thus permitted identification of non linearities in the trajectori es. The findings reveal a substantial acceleration in land cover change during the final period in the time series. This acceleration coincided with a distant factor, rising prices for regional commodities, and an intermediate factor, the paving of the Int er Oceanic Highway. The other part of the land cover trajectory analysis differentiated sub regions in Madre de Dios with distinct settlement histories, proximity to markets, biophysical conditions, and land tenure categories. Comparisons of the land cove r trajectories for the three sub regions revealed substantial differences that reflect those intermediate factors. However, all three sub regions still showed acceleration in land cover change in the final period.

PAGE 137

137 These findings confirm the importance of multiple explanatory factors for understanding land cover change, in terms of both its temporal dynamics and spatial variability. In particular, they highlight that driving forces often theorized as key factors behind land cover change are now operating mo re strongly than ever before in the study region. These factors, notably the highway and market demand, show no signs of near term abatement, which implies that land cover change is likely to continue to accelerate in the study region. This is likely to va ry across Madre de Dios, with a key sub region of interest in this context being the castaa zone. It is an open question as to whether pressures for land cover change will supersede the importance of castaa and forest concessions, given relatively high c astaa prices. What is perhaps more likely is that deforestation will expand in areas with gold mining, though given the findings from the first analytical chapter, if landholders make money in castaa and invest that in ranching, deforestation on agricult ural properties in the castaa zone may itself accelerate. The final analytical chapter took up the question of alternatives to top down initiatives for conservation and development by considering farmer federations as a key social actor in bottom up appro aches. The analysis drew on multiple data sources and followed the key farmer federation in Madre de Dios over a 20 year history. This chapter showed that the federation has experienced changing fortunes, from early successes to more recent difficulties. I n particular, the analysis drew on a multi level theoretical framework that highlighted social capital as an internal explanation for federation effectiveness, along with political ecology that highlighted external factors featuring relations between the f ederation and other social actors. The analysis showed that both

PAGE 138

138 internal and external factors are important for understanding both the early successes and the later difficulties. This chapter makes clear that while bottom up approaches to sustainable dev elopment certainly have value, they have their own perils in that internal and external conditions must be favorable for grassroots organizations to be effective. This bears implications for grassroots organizations as well as their allies, since both must exhibit the ability for collective action in order to be successful. This makes the creation of shared understandings and goals among federation members, as well as between federations and their external allies, a crucial requirement for effective action. A key open question that arises from this analysis is whether success followed by failure can in turn be followed by renewal and new success. Implications for Policies and Future Research The present research provided important insights on crucial issues. The analysis on land use at household level shows tenure diversity as key element on the decision making on agricultural land use. The tenure diversity is an outcome not only of natural resources diversity but also of policy decision aimed to protect tho se resources, or to foster the sustainable use of them. In this scenario, the analysis shows that having access to land with tenure restrictions can further land cover change in agricultural lands owned for the same household. This calls attention on po licies based on defining directly involved but they can have negative indirect impacts in areas beyond. This Thus, policies using conservationist or restricted tenure as a tool should also consider the actors and economic activities in the region, beyond the areas they intend to protect.

PAGE 139

139 This issue has special importance in th e southeastern Peruvian A mazon since similar policies are being proposed to contain gold mining in specific areas. This research also provides a new perspective on NTFPs as a tool for sustainable development. The positive roles are not straightforward evid ent when indirect impacts are considered. In this sense, if the NTFPs are part of a regional strategy the indirect negative impacts should be addressed. These negative impacts can be offset with regional policy that takes into account NTFPs exploitation a s part of a set of interrelated activities. There is a need for a set of sustainable economic activities that provide options to economic agents. Sustainable use forest is not only related to how forest is used but about all activities conducted by househo lds using the forest.. likely to change in a scenario with more instability due to climate change. It is also possible that castaa harvest and agricultural production are more subject to negative impacts because climate change. The role of roads i n land use and land cover change is noticeable in household and regional analysis. However, the temporal perspective in the regional analysis allows appreciating how improvem ents, such as the paving of the Interoceanic Highway, impact land cover change. It is also important to notice that impacts of transportation infrastructure could be fostered for a scenario signaled by economic growth, good prices on com modities, and impor tant growth i n urban population in the region. The fact that deforestation rate around the main regional m arket place is several times that observed in other areas within the study region underscore s the power of the road market combination. This situation adds challenges to sustainable development

PAGE 140

140 initiatives since market i s usually driven by changes in prices generated at another scales. Local and regional actors have no control over these dynamics and have no option but to adjust to them. The importance of drivers such as markets highlights the importance of local organizations as tools for sustainable development. However, unsustainability of social organization involved in sustainable d evelopment due to difficulties i n internal consolidation is trouble some On the other hand the lesson about the role of alliances is important since it indicates the poss ibilities of external support for the building of internal capacitie s. It is about focus not only o n the short term tasks such as development projects but also in the long term through the consolidation of social organization. In this context new actors with increasing importance such as local and regional governments provide grassroots organization with additional options for external alliances. Since t hese options are regional actors, alliances should have more stability and the objectives must be more compa tible. Additionally, the chance to establish and benefit from tools such as REDD that require extensive areas and solid institutions increase s the n eed for alliances within the regional realm. There is a clear need for additiona l research. R esearch with more details on It will be useful to understand the differ ent objectives or roles for economic activities making can also provide important insights. In the case of social organizations, research more focus ed s on their own organizations can be helpful on providing different options to consolidate social organization s

PAGE 141

141 LIST OF REFERENCES Agrawal, A., & Gibson, C. C. (1999). Enchantment and disenchantment: The role of community in natural resource conservation. World Deve lopment, 27 (4), 629 649. doi: 10.1016/s0305 750x(98)00161 2 Allegretti, M. H. (1990). Extractive reserves: an alternative for reconciling development and environmental conservation in Amazonia Alston, L. J., Libecap, G. D., & Mueller, B. (1999). Titles, C onflict, and Land Use: The Development of Property Rights and Land Reform on the Brazilian Frontiers Ann Arbor: The University of Michigan Press. Alvarez, N. L., & Naughton Treves, L. (2003). Linking national agrarian policy to deforestation in the Peruvi an Amazon: A case study of Tambopata, 1986 1997. Ambio, 32 (4), 269 274. Anderson, A. B. (1990). Alternatives to deforestation : steps toward sustainable use of the Amazon rain forest New York: Columbia University Press. Angelsen, A., & Kaimowitz, D. (199 9). Rethinking the causes of deforestation: Lessons from economic models. World Bank Research Observer, 14 (1), 73 98. Angelsen, A., & Kaimowitz, D. (2001). Forest cover and agricultural technology. In M. Palo, J. Uusivuori & G. Mery (Eds.), World forests, markets and olicies (pp. 231 238). Dordrecht: Kluwer Academic Publishers. Ankersen, T., & Barnes, G. (2004). Inside the Polygon: Emergin Community Tenure Systems and Forest Resource Extraction. In D. Zarin, J. Alavalapati, F. E. Putz & M. Schmink (Eds.), Working Forests in the Neotropics: Conservation thorugh Sustainable Management New York: Columbia University Press. Arce, R., Flores, C., Lossio, J., & Vilchez, H. (1996a). Informe Final Parte I: Proyecto de Clasificacion Participatoria de Uso Mayor de la Tierra y Desarrollo Sostenible en Areas de Influencia Humana de la Reserva Tambopata Candamo. Puerto Maldonado: FADEMAD. Arce, R., Flores, C., Lossio, J., & Vilchez, H. (1996b). Informe Final Parte II: Proyecto de Clasificacion Participatoria de Uso M ayor de la Tierra y Desarrollo Sostenible en Areas de Infuencia Humana de la Zona Reservada Tambopata Candamo. Puerto Maldonado: FADEMAD. Arias, C. (1988). La Politica Crediticia del Gobierno Aprista. Debate Agrario (2), 69 87. Arima, E. Y., Walker, R. T., Perz, S. G., & Caldas, M. (2005). Loggers and forest fragmentation: Behavioral models of road building in the Amazon basin. Annals of the Association of American Geographers, 95 (3), 525 541. doi: 10.1111/j.1467 8306.2005.00473.x

PAGE 142

142 Armas, A., Borner, J., Rug nit, M., Diaz, L., Tapia, S., Wunder, S., Nascimento, N. (2009). Pagos por Servicios Ambientales para la conservacion de bosques en la Amazonia Peruana Lima: SERNANP. Arnold, J. E. M., & Ruiz Perez, M. (2001). Can Non Timber Forest Product Match Tro pical Forest Conservation and Development Objective? Ecological Economics (39), 437 447. Assies, W. (1997). Going nuts for the rainforest: non timber forest products, forest conservation and sustainability in Amazonia. Going nuts for the rainforest: non ti mber forest products, forest conservation and sustainability in Amazonia. vii + 96 pp. Bailey, R. C., Headland, T. N., & Sponsel, L. E. (1996). Tropical deforestation : the human dimension New York: Columbia University Press. Barbier, E. B., & Burgess, J. C. (2001). The economics of tropical deforestation. Journal of Economic Surveys, 15 (3), 413 433. doi: 10.1111/1467 6419.00144 Barham, B. L., & Coomes, O. T. (1996). Prosperity's promise: the Amazon rubber boom and distorted economic development Boulder : Westview Press. Bebbington, A. (1996). Organizations and intensifications: Campesino federations, rural livelihoods and agricultural technology in the Andes and Amazonia. World Development, 24 (7), 1161 1177. doi: 10.1016/0305 750x(96)00028 9 Bebbington, A. (1999). Capitals and capabilities: A framework for analyzing peasant viability, rural livelihood and poverty. World Development, 27 (12), 2021 2044. Bebbington, A., & Perreault, T. (1999). Social capital, development, and access to resources in highland Ecuador. Economic Geography, 75 (4), 395 418. Belcher, B., Ruiz Perez, M., & Achdiawan, R. (2005). Global patterns and trends in the use and management of commercial NTFPs: Implications for livelihoods and conservation. World Development, 33 (9), 1435 1452 doi: DOI 10.1016/j.worlddev.2004.10.007 Berkes, F. (Ed.). (1991). Common property resources : ecology and community based sustainable development Dehra Dun: International Book Distributors. Bernard, H. R. (2002). Research Methods in Anthropology Walnut Creek: Altamira Press. Binswanger, H. P. (1991). Brazilan Policies that Encourage Deforestation in the Amazon. World Development, 19 (7), 821 829. doi: 10.1016/0305 750x(91)90135 5

PAGE 143

143 Bowman, M. S., Soares Filho, B. S., Merry, F. D., Nepstad, D. C., Rodrigues H., & Almeida, O. T. (2012). Persistence of cattle ranching in the Brazilian Amazon: A spatial analysis of the rationale for beef production. Land Use Policy, 29 (3), 558 568. doi: 10.1016/j.landusepol.2011.09.009 Britaldo Soares Filho, A. A., Daniel Neps tad, Gustavo Cerqueira, Maria Vera, Sergio Rivero, Luis Solorzano, Elaine Voll. (2004). Simulating the Response of Land cover Changes to Road Paving and Gobernance along a Major Amazon Highway: the Satarem Cuiba Corridor. Global Change Biology (10). Brondi zio, E. S., McCracken, S. D., Moran, E. F., Siqueira, A. D., Nelson, D. R., & Rodriguez Pedraza, C. (2002). The Colonist Footprint: Toward a Conceptual Framework of Land Use and Deforestation Trajectories among Small Farmers in the Amazon Frontier. In C. H Wood & R. Porro (Eds.), Deforestation and Land Use in the Amazon Gainesville: University Press of Florida. Bryant, R. L. (1992). Political Ecology: An emerging research agenda in Third World studies. Political Geography, 11 (1), 24. Bryant, R. L., & Bai ley, S. (1997). Third World political ecology London ; New York: Routledge. Caldas, M., Walker, R., Arima, E., Perz, S., Aldrich, S., & Simmons, C. (2007). Theorizing land cover and land use change: The peasant economy of Amazonian deforestation. Annals o f the Association of American Geographers, 97 (1), 86 110. Campos, M. T., & Nepstad, D. C. (2006). Smallholders, the Amazon's new conservationists. Conservation Biology, 20 (5), 1553 1556. doi: DOI 10.1111/j.1523 1739.2006.00546 Carroll, T. F., & Bebbington A. J. (2000). Peasant federations and rural development policies in the Andes. Policy Sciences, 33 (3 4), 435 457. doi: 10.1023/a:1004824803848 Chambers, R. (1984). Rural development : putting the last first / Robert Chambers London ; New York: Longman. Chambers, R., & Conway, G. (1991). Sustainable rural livelihoods: practical concepts for 21st century IDS Discussion Paper (Vol. 296). Brighton: Institute of Development Studies. Charnley, S., & Poe, M. R. (2007). Community forestry in theory and practice: Where are we now? Annual Review of Anthropology (Vol. 36, pp. 301 336). Chavez, A. B. (2009). Public policy and spatial variation in land use and land cover in the southeastern Peruvian Amazon Gainesville, Fla: University of Florida.

PAGE 144

144 Chomitz, K. M., & Gr ay, D. A. (1996). Roads, land use, and deforestation: A spatial model applied to belize. World Bank Economic Review, 10 (3), 487 512. Coffin, A. W. (2007). From roadkill to road ecology: A review of the ecological effects of roads. Journal of Transport Geo graphy (15), 396 407. Coleman, J. S. (1990). Foundations of Social Theory Cambridge, Mass.: Belknap Press of Harvard University Press. Collinson, C., Burnett, D., & Agreda, V. (2000). Economic Viability of Brazil Nut Trading in Peru. Kent: Natural Resourc e Institute. Cosso, R. E. (2009). Capacity for timber management among private small medium forest enterprises in Madre de Dios, Peru. (PhD), Univers ity of Florida, Gainesville Dourojeanni, M. J. (1990). La Amazona : qu hacer? Iquitos, Peru: Centro de Estudios Teolgicos de la Amazona. Dourojeanni, M. J., Barandiarn, A., & Dourojeanni, D. (2010). Amazona peruana en 2021 : explotacin de recursos naturales e infraestructura : qu est pasando? qu es lo que significa para el futuro? (2 ed.). Lima, P er: Sociedad Peruana de Derecho Ambiental, SPDA. Easterly, W., & Serven, L. (Eds.). (2003). The Limits of Stabilization: Infrastructure, Public Deficit, and Growth in Latin America Washington D.C.: The World Bank Stanforf University Press. Ehringhaus, C. (2005). Pos victory Dilemmas: Land Use, Development Policies, and Social Movements in Amazonian Extractive Reserves. (PhD), Yale University. Ellis, F. (2000). Rural livelihoods and diversity in developing countries Oxford: Oxford University Press. Esco bal, J., Agreda, V., & Aldana, U. (2000). Derechos de propiedad, regulacion de las concesiones y uso optimo de los recursos naturales. In J. Berdegue & G. Escobar (Eds.), Seguimiento y evaluacion del manejo de recursos naturales Santiago de Chile: Fundaci on de Comunicaciones en el Agro. Escobal, J., & Aldana, U. (2003). Are nontimber forest products the antidote to rainforest degradation? Brazil nut extraction in Madre De Dios, Peru. World Development, 31 (11), 1873 1887. doi: 10.1016/j.worlddev.2003.08.0 01|ISSN 0305 750X FADEMAD. (2000). Libro de Actas 1999 2000 Federacion Agraria de Madre de Dios, Puerto Maldonado.

PAGE 145

145 FADEMAD. (2012). Libro de Actas 2000 2010 Federacion Agraria de Madre de Dios, Puerto Maldonado. Faminow, M. D. (1997). Spatial economics o f local demand for cattle products in Amazon development. Agriculture Ecosystems & Environment, 62 (1), 1 11. doi: 10.1016/s0167 8809(96)01116 4 FAO. (2011). Global Forest Resource Assesment 2010 Forestry Paper (Vol. 163). Roma: Food and Agriculture Organiz ation. Fearnside, P. (1987). Deforestation and international economic development projects in Brazilian Amazonia. Conservation Biology, 1 (13). Fearnside, P. (1989). Deforestation and Agricultural Development in Brazilian Amazonia. Interciencia, 14 (6), 291 297. Fearnside, P. (2001). Land tenure issues as factors in environmental destruction in Brazilian Amazonia: The case of Southern Para. World Development, 29 (8), 1361 1372. doi: 10.1016/s0305 750x(01)00039 0 Figueroa, A. (1992). La Agricultura Peruana y el Ajuste. Debate Agrario (13), 35 48. Fleck, L., Vera Diaz, M., Borasino, E., Glave, M., Hak, J., & Josse, C. (2010). Estrategias de conservacion a lo largo de la carretera Interoceanica en Made de Dios, Peru Serie Tecnica Lima: Conservation Strategy Fun d. Foweraker, J. (2001). Grassroots movements and political activism in Latin America: A critical comparison of Chile and Brazil. Journal of Latin American Studies, 33 839 865. Futemma, C., & Brondizio, E. S. (2003). Land reform and land use changes in t he lower amazon: Implications for agricultural intensification. Human Ecology, 31 (3), 369 402. doi: 10.1023/a:1025067721480 Geist, H., & Lambin, E. F. (2006). Land use and land cover change : local processes and global impacts / Eric F. Lambin, Helmut Geis t (eds.) Berlin ; lNew York: Springer. Geist, H. J., & Lambin, E. F. (2001). What Drives Tropical Deforestation? LUCC Report Series (Vol. 4). Louvain: LUCC International Project. Geist, H. J., & Lambin, E. F. (2002). Proximate Causes and Underlying Drivin g Forces of Tropical Deforestation. Bioscience, 52 (2). Geist, H. J., & Lambin, E. F. (2006). Land use and land cover change : local processes and global impacts Berlin ; lNew York: Springer.

PAGE 146

146 Gomes, C. V. A., Vadjunec, J. M., & Perz, S. G. (2012). Rubber tapper identities: Political economic dynamics, livelihood shifts, and environmental implications in a changing Amazon. Geoforum, 43 (2), 260 271. doi: 10.1016/j.geoforum.2011.09.005 Gonzales de Olarte, E. (1996). The Peruvian economy and structural adjustm ent : past, present, and future Coral Gables, Fla: North South Center Press, University of Miami. GOREMAD. (2002). Plan Concertado del Departamento de Madre de Dios Puerto Maldonado. Goulding, M., Canas, C., Barthem, R., Forsberg, B., & Ortega, H. (2003 ). Las Fuentes del Amazonas: rios, vida y conservacion de la cuenca del Madre de Dios Lima, Peru: ACCA. Gutman, G. (2004). Land change science : observing, monitoring and understanding trajectories of change on the Earth's surface Dordrecht ; Boston ; Lo ndon: Kluwer Academic Publishers. Hall, A. L. (1989). Developing Amazonia : deforestation and social conflict in Brazils Carajas programme Manchester ; New York : New York, NY, USA: Manchester University Press. Hall, A. L. (1997). Sustaining Amazonia: Gra ssroots Action for Productive Conservation Manchester, UK; New York, USA: Manchester University Press. Harriss, J., & De Renzio, P. (1997). "Missing link" or Analytically missing?: The concept of Social Capital. Journal of International Development, 9 (7) 919 937. Hecht, S. B., Kandel, S., Gomes, I., Cuellar, N., & Rosa, H. (2006). Globalization, Forest Resurgence, and Environmental Politics in El Salvador. World Development, 34 (2), 308 323. doi: 10.1016/j.worlddev.2005.09.005 Homma, A. K. O. (1992). The dynamics of extraction in Amazonia: a historical perspective. In D. C. Nepstad & S. Schwartzman (Eds.), Advances in Economic Botany (Vol. 9, pp. 23 31). Houtzager, P. P. (2001). Collective action and political authority: Rural workers, church, and state i n Brazil. Theory and Society, 30 (1), 1 45. IIAP. (2001). Propuesta de Zonificacion Ecologica Economica Puerto Maldonado. IIRSA. (2008). Initiative for the Integration of Regional Infrastructure in South America, 2012 INEI. (2009). Perfil Sociodemografico del Departamento de Madre de Dios Lima: INEI.

PAGE 147

147 INRENA. (2003). Parque Nacional Bahuaja Sonene: Plan Maestro 2004 2008. Puerto Maldonado: INRENA. Irwin, E. G., & Geoghegan, J. (2001). Theory, data, methods: developing spatially explicit economic models o f land use change. Agriculture Ecosystems & Environment, 85 (1 3), 7 23. Kainer, K. A., Schmink, M., Leite, A. C. P., & Fadell, M. J. D. (2003). Experiments in forest based development inWestern Amazonia. Society & Natural Resources, 16 (10), 869 886. doi: 10.1080/08941920390231306 Kar, S. P., & Jacobson, M. G. (2012). NTFP income contribution to household economy and related socio economic factors: Lessons from Bangladesh. Forest Policy and Economics, 14 (1), 136 142. doi: 10.1016/j.forpol.2011.08.003 Keck, M. E., & Sikkink, K. (1999). Transnational advocacy networks in international and regional politics. International Social Science Journal, 51 (1), 89 +. doi: 10.1111/1468 2451.00179 Keller, M. (2009). Amazonia and global change Washington, DC: American Geo physical Union. Killeen, T. (2007). The Perfect Storm in the Amazon Wilderness Advances in Applied Biodiversity Science (Vol. 7). Washington: Conservation International. Lambin, E. F., Turner, B. L., Geist, H. J., Agbola, S. B., Angelsen, A., Bruce, J. W., Xu, J. C. (2001). The causes of land use and land cover change: moving beyond the myths. Global Environmental Change Human and Policy Dimensions, 11 (4), 261 269. doi: 10.1016/s0959 3780(01)00007 3 Lanao, M. (1998). Gendered political ecology of Braz il nuts in Madre de Dios, Peru : "no chorrea, pero gotea". (Master), University of Florida, Gainesville. Lopez, R., & Galinato, G. I. (2005). Trade Policies, Economic Growth, and the Direct Causes of Deforestation \ Land Economic, 81 (2), 25. Lossio, J. (2001). Plazas Centrales e Intermedias en Madre de Dios: del descubrimiento de Fitzcarrald (1897) al ultimo censo (1993). In M. Pulgar Vidal, E. Zegarra & J. Urrutia (Eds.), SEPIA IX Puno: SEPIA. Mahar, D. (1989). Goverment Policies and Deforestation in Brazil's Amazon Region Washington D.C.: World Bank. Malhi, Y., Roberts, J. T., Betts, R. A., Killeen, T. J., Li, W., & Nobre, C. A. (2008). Climate change, deforestation, and the fate of the Amazon. Science, 319 (5860), 169 172. doi: 10.1126/science.114696 1

PAGE 148

148 Manring, S. L. (2007). Creating and managing interorganizational learning networks to achieve sustainable ecosystem management. Organization & Environment, 20 (3), 325 346. doi: 10.1177/1086026607305738 Manzi, M., & Coomes, O. T. (2009). Managing Amazonia n palms for community use: A case of aguaje palm (Mauritia flexuosa) in Peru. Forest Ecology and Management, 257 (2), 510 517. doi: 10.1016/j.foreco.2008.09.038 Marsik, M., Stevens, F. R., & Southworth, J. (2011). Amazon deforestation: Rates and patterns of land cover change and fragmentation in Pando, northern Bolivia, 1986 to 2005. Progress in Physical Geography, 35 (3), 353 374. doi: 10.1177/0309133311399492 Martnez, L. J. (1998). Suelos de la Amazonia (2 ed.). Bogota: Ministerio de Educacin Nacional. Ma ther, A. S., & Needle, C. L. (2000). The relationships of population and forest trends. Geographical Journal, 166 2 13. doi: 10.1111/j.1475 4959.2000.tb00002.x McCracken, S. D., Brondizio, E. S., Nelson, D., Moran, E. F., Siqueira, A. D., & Rodriguez Pedr aza, C. (1999). Remote sensing and GIS at farm property level: Demography and deforestation in the Brazilian Amazon. Photogrammetric Engineering and Remote Sensing, 65 (11), 1311 1320. Mertens, B., Kaimowitz, D., Puntodewo, A., Vanclay, J., & Mendez, P. (2 004). Modeling deforestation at distinct geographic scales and time periods in Santa cruz, Bolivia. International Regional Science Review, 27 (3), 271 296. Mertens, B., Poccard Chapuis, R., Piketty, M. G., Lacques, A. E., & Venturieri, A. (2002). Crossing spatial analyses and livestock economics to understand deforestation processes in the Brazilian Amazon: the case of Sao Felix do Xingu in South Para. Agricultural Economics, 27 (3), 269 294. doi: 10.1111/j.1574 0862.2002.tb00121.x Mertens, B., Sunderlin, W. D., Ndoye, O., & Lambin, E. F. (2000). Impact of macroeconomic change on deforestation in South Cameroon: Integration of household survey and remotely sensed data. World Development, 28 (6), 983 999. Moore, T., Chicchon, A., & Salas, A. (1995). Estrategia Eco regional para el Desarrollo Sostenible en la Zona Reservada Tambopata Candamo (Peru). In A. Lopez (Ed.), Estrategias para el Desarrollo Sostenible: America Latina Gland, Zuitzerland: UICN. Mosquera, C., Chavez, M. L., Pachas, V., & Moschela, P. (2009 ). Estudio diagnostico de la actividad minera artesanal en Madre de Dios Lima: CooperAccion, Caritas, Conservacion Internacional.

PAGE 149

149 Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B., & Kent, J. (2000). Biodiversity hotspots for conserv ation priorities. Nature, 403 (6772), 853 858. doi: 10.1038/35002501 Nelson, G. C., & Hellerstein, D. (1997). Do roads cause deforestation? Using satellite images in econometric analysis of land use. American Journal of Agricultural Economics, 79 (1), 80 88. Nepstad, D. C., Stickler, C. M., & Almeida, O. T. (2006). Globalization of the Amazon soy and beef industries: Opportunities for conservation. Conservation Biology, 20 (6), 1595 1603. doi: 10.1111/j.1523 1739.2006.00510.x Netting, R. M. (1993). Smallholde rs, Householders: Farm Families and the Ecology of Intensive, Sustainable Agriculture Stanford: Stanford University Press. Ostrom, E. (1990). Governing the commons: the evolution of institutions for collective actions Cambridge, New York: Cambridge Unive rsity Press. Ostrom, E. (2005). Not just one best system: the diversity of institutions for coping with the commons Pace, R. (1992). Social Conflict and Political Activism in ethe Brazilian Amazon, A case study of Gurupa. American Ethnologist, 19 (4), 710 732. doi: 10.1525/ae.1992.19.4.02a00050 Pacheco, P. (2002). Deforestation and Forest Degradation in Lowland Bolivia. In C. H. Wood & R. Porro (Eds.), Deforestation and Land Use in the Amazon Gainesville: University Press of Floroda. Pacheco, P. (2006). Ag ricultural expansion and deforestation in lowland Bolivia: the import substitution versus the structural adjustment model. Land Use Policy, 23 (3), 205 225. doi: 10.1016/j.landusepol.2004.09.004 Pea, P. (2010). La Castana y la Shiringa en Madre de Dios Cua dernos de Investigacion Lima: SPDA. Perz, S. G. (2001). Household demographic factors as life cycle determinants of land use in the Amazon. Population Research and Policy Review, 20 (3), 159 186. doi: 10.1023/a:1010658719768 Perz, S. G., & Almeyda, A. M. ( 2010). A Tri Partite Framework of Forest Dynamics: Hierarchy, Panarchy, and Heterarchy in the Study of Secondary Growth. In H. Nagendra & J. Southworth (Eds.), Reforesting Landscapes:LINKING PATTERN AND PROCESS (Vol. 10, pp. 59 84).

PAGE 150

150 Perz, S. G., Arambur, C., & Bremner, J. (2005). Population, Land Use and Deforestation in the Pan Amazon Basin: a Comparison of Brazil, Bolivia, Colombia, Ecuador, Per and Venezuela. Environment, Development and Sustainability, 7 (1), 23 49. doi: 10.1007/s10668 003 6977 9 Perz, S. G., & Walker, R. (2002). Household life cycles and secondary forest cover among small farm colonists in the Amazon. World Development, 30 (6), 1009 1027. doi: 10.1016/s0305 750x(02)00024 4 Perz, S. G., Walker, R., & Caldas, M. (2006). Beyond population and environment: Household demographic life cycles and land use allocation among small farms in the Amazon. Human Ecology, 34 (6), 829 849. doi: DOI 10.1007/s10745 006 9039 8 Peters, C., Gentry, A., & Mendelson, R. (1989). Valuation of an Amazoniam rainfore st. Nature (339), 655 656. Pfaff, A., Robalino, J., Walker, R., Aldrich, S., Caldas, M., Reis, E., Kirby, K. (2007). Roads and deforestation in the Brazilian Amazon. Journal of Regional Science, 47 (1), 109 123. Poncelet, E. C. (2004). Partnering for the environment : multistakeholder collaboration in a changing world Lanham, MD: Rowman & Littlefield. Portes, A. (1998). Social Capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24 1 24. Portes, A., & Landolt, P. ( 2000). Social capital: Promise and pitfalls of its role in development. Journal of Latin American Studies, 32 529 547. Pulido, M. T., & Caballero, J. (2006). The impact of shifting agriculture on the availability of non timber forest products: the exampl e of Sabal yapa in the Maya lowlands of Mexico. Forest Ecology and Management, 222 (1 3), 399 409. doi: DOI 10.1016/j.foreco.2005.10.043 Putnam, R. D. (1993). Making democracy work: civic traditions in modern Italy Princeton, N. J.: Princeton University Pr ess. Pyhala, A., Brown, K., & Adger, W. N. (2006). Implications of livelihood dependence on non timber products in Peruvian Amazonia. Ecosystems, 9 (8), 1328 1341. doi: DOI 10.1007/s10021 005 0154 y Redo, D., Millington, A. C., & Hindery, D. (2011). Defores tation dynamics and policy changes in Bolivia's post neoliberal era. Land Use Policy, 28 (1), 227 241. doi: 10.1016/j.landusepol.2010.06.004 Redwood, J. (2002). World Bank Approaches to the Brazilian Amazon Sustainable Development Working Paper (Vol. 13). W ashington D.C.: World Bank.

PAGE 151

151 Richards, M. (1997). Common property resource institutions and forest management in Latin America. Development and Change, 28 (1), 95 117. doi: 10.1111/1467 7660.00036 Rodriguez, M. (1991). Proceso de Ocupacion y Construccion Soc ial del Espacio Amazonico. In F. Barclay, M. Rodriguez, F. Santos & M. Valcarcel (Eds.), Amazonia 1940 1990: el extravio de una ilusion Lima, Peru: Terra Nuova CISEPA. Rudel, T. (2007). Changing Agents of Deforestation: From State initiated to Enterpri se Driven Processes, 1970 2000. Land Use Policy, 24 (1), 35 41. doi: 10.1016/j.landusepol.2005.11.004 Rudel, T., & Roper, J. (1997). The paths to rain forest destruction: Crossnational patterns of tropical deforestation, 1975 90. World Development, 25 (1), 5 3 65. Ruiz Perez, M., Belcher, B., Achdiawan, M., Alexiades, M., Aubertin, J., Caballero, J., Youn, Y. (2004). Markets drive the specialization strategies of forest peoples. Ecology and Society, 9 (2). Salisbury, D. S., & Schmink, M. (2007). Cows ve rsus rubber: Changing livelihoods among Amazonian extractivists. Geoforum, 38 (6), 1233 1249. doi: 10.1016/j.geoforum.2007.03.005 Sandoval, A., & Wilson, D. E. (1996). Manu : the biodiversity of southeastern Peru Washington, D.C: Smithsonian Institution. S anta Gadea, R., & Wagner, A. (Eds.). (2002). La integracin regional entre Bolivia, Brasil y Per / editores, Allan Wagner Tizn, Rosario Santa Gadea Duarte Lima: CEPEI. Schmink, M., & Wood, C. H. (1992). Contested frontiers in Amazonia New York: Columbi a University Press. Schmook, B., & Vance, C. (2009). Agricultural Policy, Market Barriers, and World Development, 37 (5), 1015 1025. doi: 10.1016/j.worlddev.2008.09.006 Schwartzman, S. (1992). Land distr ibution and the social costs of frontier development in Brazil: social and historical context of extractive reserves. In D. C. Nepstad & S. Schwartzman (Eds.), Advances in Economic Botany (Vol. 9, pp. 51 66). Schwartzman, S., Alencar, A., Zarin, H., & Sant os Souza, A. P. (2010). Social Movements and Large Scale Tropical Forest Protection on the Amazon Frontier: Conservation from Chaos. Journal of Environment & Development, 19 (3), 274 299. doi: 10.1177/1070496510367627

PAGE 152

152 Schwartzman, S., & Zimmerman, B. (2005) Conservation alliances with indigenous peoples of the Amazon. Conservation Biology, 19 (3), 721 727. doi: 10.1111/j.1523 1739.2005.00695.x Sheil, D., & Wunder, S. (2002). The Value of Tropical Forest to Locla Communities: Complications, Caveats and Cautio ns. Conservation Ecology, 6 (2). Shriar, A. J. (2006). Regional integration or disintegration? Recent road improvements in Petn, Guatemala: A review of preliminary economic, agricultural, and environmental impacts. Geoforum, 37 (1), 104 112. doi: 10.1016/j .geoforum.2004.08.008 Shukla, J., Nobre, C., & Sellers, P. (1990). Amazon Deforestation and Climate Change. Science, 247 (4948), 1322 1325. doi: 10.1126/science.247.4948.1322 Siegmund Schultze, M., Rischkowsky, B., da Veiga, J. B., & King, J. M. (2007). Cat tle are cash generating assets farms in the Eastern for mixed smallholder Amazon. Agricultural Systems, 94 (3), 738 749. doi: 10.1016/j.agsy.2007.03.005 Simmons, C. S., Walker, R. T., Arima, E. Y., Aldrich, S. P., & Caldas, M. M. (2007). The amazon land war in the south of para. Annals of the Association of American Geographers, 97 (3), 567 592. doi: 10.1111/j.1467 8306.2007.00564.x Singh, I., Squire, L., & John, S. (Eds.). (1986). Agricultural Household Models: Extensions, Applications and Policy Baltimore and London: The World Bank and The Johns Hopkins University Press. Southworth, J., Marsik, M., Qiu, Y. L., Perz, S., Cumming, G., Stevens, F., Barnes, G. (2011). Roads as drivers of change: trajectories across the tri national frontier in MAP, the Southwestern Amazon. Remote Sensing, 3 (5), 1047 1066. doi: 10.3390/rs3051047 Southworth, J., & Tucker, C. (2001). The influence of accessibility, local institutions, and socioeconomic factors on forest cover change in the mountains of western Honduras. Mo untain Research and Development, 21 (3), 276 283. Stronza, A. (1996). Conservation and development at the grassroots: the challenges for a federation of colonist farmers in the Peruvian Amazon. (Master), University of Florida, Gainesville. Stronza, A. ( 2000). "Because it is ours" : community based ecotourism in the Peruvian Amazon. (PhD), University of Florida, Gainesville. Swenson, J. J., Carter, C. E., Domec, J. C., & Delgado, C. I. (2011). Gold Mining in the Peruvian Amazon: Global Prices, Deforest ation, and Mercury Imports. Plos One, 6 (4). doi: 10.1371/journal.pone.0018875

PAGE 153

153 Uphoff, N. (1991). Fitting projects to people. In M. M. Cernea (Ed.), Putting people first: sociological variables in rural development. (pp. 467 511). New York: Oxford Universit y Press. Uphoff, N. (1993). Grass roots Organizactions and NGOs in Rural Development: Oportunities with Diminishing States and Expanding Markets. World Development, 21 (4), 607 622. doi: 10.1016/0305 750x(93)90113 n Uphoff, N., & Wijayaratna, C. M. (2000). Demonstrated benefits from social capital: The productivity of farmer organizations in Gal Oya, Sri Lanka. World Development, 28 (11), 1875 1890. doi: 10.1016/s0305 750x(00)00063 2 VanWey, L. K., Ostrom, E., & Meretsky, V. (2005). Theories Underlying the S tudy of Human Environment Interactions. In E. F. Moran & E. Ostrom (Eds.), Seeing the Forest and the Trees Cambridge, Masachusetts; London, England: The MIT Press. Vasquez, R., & Gentry, A. H. (1989). Use and Misuse of Forest Harvested Fruits in the Iquit os Area. Conservatoin Biology, 3 (4), 350 361. Vitousek, P. M. (1994). Beyond Global Warming Ecology and Global Change. Ecology, 75 (7), 1861 1876. doi: 10.2307/1941591 Vosti, S. A., Witcover, J., & Carpentier, C. L. (2002). Agricultural intensification b y smallholders in the western Brazilian Amazon : from deforestation to sustainable land use Washington, D.C.: International Food Policy Research Institute. Walker, R. (2003). Mapping Process to Pattern in the Landscape Change of the Amazonian frontier. An nals of the Association of American Geographers, 93 (2). Walker, R., Browder, J., Arima, E., Simmons, C., Pereira, R., Caldas, M., de Zen, S. (2009). Ranching and the new global range: Amazonia in the 21st century. Geoforum, 40 (5), 732 745. doi: 10.1 016/j.geoforum.2008.10.009 Walker, R., & Homma, A. K. O. (1996). Land use and land cover dynamics in the Brazilian Amazon: An overview. Ecological Economics, 18 (1), 67 80. Walker, R., Moran, E., & Anselin, L. (2000). Deforestation and cattle ranching in t he Brazilian Amazon: External capital and household processes. World Development, 28 (4), 683 699. Walker, R., Perz, S., Caldas, M., & Texeira, l. (2002). Land Use and Land Cover Change in Forest Frontier: The Role of Household Life Cycles. International R egional Science Review, 25 (2), 169 199.

PAGE 154

154 Western, D., & Wright, R. M. (1994). The background to community based conservation. In D. Western & R. M. Wright (Eds.), Natural connections: Perspectives in community based conservation (pp. 1 12). Washington, D.C .: Island Press. White, T. A., & Runge, C. F. (1995). The Emergence and Evolution of Collective Action Lessons from Watershed Management in Haiti. World Development, 23 (10), 1683 1698. Wollenberg, E., Iwan, R., Limberg, G., Moeliono, M., Rhee, S., & Sud ana, M. (2007). Muddling towards cooperation: spontaneous orders and shared learning in Malinau District, Indonesia Wood, C. H. (2002). Land Use and Deforestation in the Amazon. In C. H. Wood & R. Porro (Eds.), Deforestation and Land Use in the Amazon Ga inesville: University Press of Florida. Wood, C. H., & Porro, R. (2002). Deforestation and Land Use in the Amazon Gainesville: University Press of Florida. Woolcock, M. (1998). Social capital and economic development: Toward a theoretical synthesis and po licy framework. Theory and Society, 27 (2), 151 208. doi: 10.1023/a:1006884930135 Zoomers, A., & Haar, G. v. d. (2000). Current land policy in Latin America: regulating land tenure under neo liberalism Amsterdam: Royal Tropical Institute.

PAGE 155

155 BIOGRAPHI CAL SKETCH Rafael Rojas was born in 1963 in Nasca, Peru. In 1982, he moved to Lima where he enrolled in the Catholic University, graduating in 19 social s cie nces, with a specialization in e conomics. He got his Masters of Arts in the Center for Latin American Studies at the University of Florida in 2004. His work experience involves researching sustainable development in the Peruvian Amazon and socio environmental assessment of infrastructure and investment projects.