Citation
Landownership turnover and family farm survival in an Amazon resettlement project

Material Information

Title:
Landownership turnover and family farm survival in an Amazon resettlement project
Alternate title:
Land ownership turnover and family farm survival in an Amazon resettlement project
Creator:
Cronkleton, Peter
Publication Date:
Language:
English
Physical Description:
xi, 175 leaves : ; 29 cm.

Subjects

Subjects / Keywords:
Agricultural land ( jstor )
Colonization ( jstor )
Community associations ( jstor )
Crops ( jstor )
Farms ( jstor )
Financial investments ( jstor )
Forests ( jstor )
Land use ( jstor )
Pastures ( jstor )
Rubber ( jstor )
Anthropology thesis, Ph. D ( lcsh )
Dissertations, Academic -- Anthropology -- UF ( lcsh )
Family farms ( fast )
Land settlement ( fast )
Land tenure ( fast )
Brazil -- Acre ( fast )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 168-174).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Peter Cronkleton.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029224530 ( ALEPH )
41352960 ( OCLC )

Downloads

This item has the following downloads:


Full Text










LANDOWNERSHIP TURNOVER
AND FAMILY FARM SURVIVAL IN AN
AMAZON RESETTLEMENT PROJECT














By

PETER CRONKLETON














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

UNIVERSITY OF FLORIDA

1998































To my parents, Howard and Mary T. Cronkleton.













ACKNOWLEDGMENTS

There are many people and institutions I would like thank for the help they

provided as I conducted the research described here.

First of all, I would like to thank Marianne Schmink, the chair of my doctoral

committee. The training, advice, and opportunity she gave me were crucial to my

development as an anthropologist. The example she has set as a scholar and mentor will

always be an inspiration to me. I would also like to thank the other members of my

doctoral committee: Peter Hildebrand, Maxine Margolis, Gerald Murray and Tony

Oliver-Smith. The training and insight I gained working with these scholars has been

very important in my development as a professional.

I would like to express my gratitude to my Brazilian counterparts in Acre. I

especially wish to thank the numerous members of PESACRE. Their logistical support

and expert advice helped make this research possible, and their hospitality made it

enjoyable. I would also like to thank the employees of EMATER who helped me with

their technical expertise and broad experience in Peixoto, as well as the INCRA

employees in Rio Branco, Campina, and Placido de Castro who gave me access to their

records and explained Peixoto's history and operation. I am also indebted to the many

families in Peixoto who took time from their daily routines to show me their farms,

answer my questions, and discuss their lives. Finally, I wish to thank Rivandro da Silva

Mota, my field assistant, whose hard work and friendship made long field trips possible.




iii








I also must thank the institutions that funded my doctoral research. Field work in

1994, 1995, and 1996 was funded by the North-South Center at the University of Miami,

the University of Florida's Tropical Conservation and Development Program, and

USAID. Without the support of these institutions this research would not have been

possible.

I would like to thank my parents, Howard and Mary T. Cronkleton, who gave me

the freedom and encouragement to pursue my interests. Finally, I want to thank my wife

Martha. Without her love, support, and advice I would not have finished this dissertation.




































iv














TABLE OF CONTENTS


Vage


A CK N O W LED G M EN TS .............................................................................................. iii

L IST O F T A B L E S ......................................................................................................... viii

LIST O F FIG U R E S ..................................................................................................... ix

A B ST R A C T ............................................................................................................ ............. x

CHAPTERS

1 RESETTLEMENT AND LANDOWNERSHIP TURNOVER ON BRAZIL'S
AMAZONIAN FRONTIER ..................................................................................

Frontier Transitions and Unintended Consequences in Amazonia..............................4
Purpose of Study ............................................................................................... ......... 1
O utline of D issertation........................................................................................... 12

2 THE AMAZON'S FRONTIER TRANSITION AND THE
MARGINALIZATION OF SMALLHOLDERS.................................................... 16

Government Policies and the Bias Toward Large-Scale Capitalist Enterprises ..........17
Am azonia's Insecure Land Tenure ................................... .......................................21
The Evolution of Amazonian Colonization Programs.................................................25
Rural Mobilization and International Alliances.......................................................34
Democratization and the Return to the Agrarian Reform Debate..............................36
C onclusions............................. ............................................................................... 42

3 RESETTLEMENT AND COLONIZATION IN ACRE ...........................................44

Acre Before the Developm ent Boom .............................................................. ........45
The Early Crisis and New Opportunities.................................................................48
The Development Boom Comes To Acre.................................................................51
Allies of the Poor and the Beginning of Rural Mobilization in Acre ....................56
Attempted Solutions Through Colonization and Resettlement ...........................60
The Federal Government Attempts Colonization Solutions................................61
Resettled and Abandoned ................................................................................. 65



v








Democratization, the Continuing Crisis and the Climax of Social
Mobilization ............................................................................................... 67
Emancipation From Unfulfilled Promises ..................... .. ........................69
C onclusions...................................... .......................... ..............................................7 1

4 LANDOWNERSHIP TURNOVER IN PEIXOTO .............................................73

P eixoto ................................................................................................................ . 75
Research Design And Sampling Methods Appropriate For Peixoto .........................80
Who Are Peixoto's Current Residents? ......................................................................86
O rigin ........................................ ............ ..... ......... ..................... ... ........... 87
Family Composition ........................................................................................ 88
E ducation ................... . ............... ...... ............................... .. .. ........... ..... ........89
H ealth Services ................................................................................................. 90
Transportation ...................................................................... ............................. 91
Farm Properties...................................................................................................92
Turnover Measures ................................................................................................94
Landownership Turnover.......................................................................................96
Why Is Turnover Taking Place in Peixoto?......................................... ............ 99
Current Settlement Patterns in Peixoto........................................ ..................... 103
Settlement Zones in Peixoto ................................................................................. ....107
C onclusions............................................................................................................ 112

5 RURAL MOBILIZATION AND FAMILY FARM SETTLEMENT
SUCCESS ............................................................................................................ 114

Souza Araujo's Settlement................................................................................... 115
Associagdo Rural Libertador ............. .......................................................117
Why Has the ARL Worked? .................. ................................................................ 121
Why Are Associations Like ARL So Rare?............................................................123
Appeasing Rural Resistance with Land Distribution...................................123
Forming Associations to Undercut Rural Mobilization................................125
Case Studies in Failed Organizations .....................................................................129
Extended Families and Community Cohesion...........................................................133
Conclusions..........................................................................................................135

6 LAND USE PATTERNS AND THE IMPACT OF LANDOWNERSHIP
TURNOVER........................................................................................................137

Colonist Farming Systems in Peixoto....................................................................138
Sm all Farm A ttrition ..................................................................................................140
Cash Income Alternatives .....................................................................................142
Typical Land Use in Peixoto................................. .............................................149
The Effects of Landownership Turnover on Small Farm Land Use........................153
D iscussion .................................................................................................................. 157
C onclusions............................................................................................................. 159



vi








7 C O N C LU SIO N S..................................................................................... ............. 161

APPENDIX A

EQUATIONS AND NOTATION USED FOR TWO-STAGE CLUSTER
SA M PLE ....................................................................................................... ............. 167

LIST OF REFEREN CES .................................................................................................168

BIO GRAPH ICAL SKETCH ................................................................................ ....... 175












































vii














LIST OF TABLES

Table page
Table 4.1 Origin of Selected Adults in Peixoto Household Survey ................................87

Table 4.2 Additional Properties Owned by Peixoto Residents........................................93

Table 4.3 Location of Additional Property ........................................ ....................... 93

Table 4.4 Location of Earlier Property ................................... ........................................94

Table 4.5 Number of Previous Owners.............................................................................. 97

Table 4.6 Ownership Status of Sub-Sample of Resident Owners......................................98

Table 4.7 Average Settlement Time by Ownership Status ..............................................99

Table 6.1 Distribution of Farms by Ownership Category..............................................154

Table 6.2 Mean Areas of Land Use Types by Ownership Category ...............................155

Table 6.3 Percentages of Cleared Land by Use Type and Ownership Category ..........156
























viii














LIST OF FIGURES

Figure page

Figure 4.1 Southeastern Acre and the Peixoto..............................................................76

Figure 4.2 Peixoto and Selected Field Sites................................................................. .... 85

Figure 4.3 Percentage of Original Residents in Sampled Clusters in 1995...................104

Figure 6.1 Land use on typical Peixoto farm................................................................. 151




































ix















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

LANDOWNERSHIP TURNOVER
AND FAMILY FARM SURVIVAL IN AN
AMAZON RESETTLEMENT PROJECT

By

Peter Cronkleton

May 1998


Chairman: Marianne Schmink
Major Department: Anthropology

In the Brazilian Amazon, attempts by the government to resettle landless and

displaced families have not successfully placed families for long-term settlement.

Instead, these colonization projects have been plagued by the phenomenon of

landownership turnover. This means that settled families abandon, sell, or trade their

land rather than staying and developing their farms. Compounding the failure to develop

long-term, permanent settlement, the resources spent to assist needy families do not

accomplish the desired goal, may increase environmental impacts, and actually benefit

wealthy and powerful interests who can take advantage of the situation.

Landownership turnover is one consequence of a complex process of change that

has swept over Amazonia in recent decades. The conditions that allowed unstable

settlement among smallholders to develop stem from government policies that ignore or




x








even undermine the land rights of the rural families. Official attempts to assist these

families have usually been halfhearted and often more interested in ending unrest rather

than assuring successful homesteading.

In order to better understand how this process has developed and may continue,

this research focuses on the Peixoto resettlement project in the Brazilian state of Acre.

By examining a single project, the research can examine detailed variation in the behavior

of residents and variation in the conditions that may drive turnover. The research focuses

on intermediate range processes at the community level without excluding attention to

individual, household, or regional level phenomena.

Landownership turnover in Peixoto is indeed very high and within the project it

has affected some communities more than others. A number of factors appear to produce

the conditions that discourage long-term settlement. Inadequate infrastructure and

services account for much of the difficulty faced by colonist families. Some exceptional

communities have organized and overcome the hardships encountered in Peixoto, but

most have not. Generally, the families living in Peixoto are able to meet their subsistence

needs but it is more difficult for them to acquire cash income. One result is that colonists

favor pasture formation and ranching. As pasture expands within the project, the area

becomes more attractive to expanding ranches around the project, and may lead to land

concentration. When colonist families hold large portions of their property in pasture it

decreases the area available for subsistence production. Eventually, as this land degrades,

it will be harder for later generations to support themselves on the land. If this trend

continues it could force the people now living in Peixoto to seek homesteads elsewhere in

Amazonia.




xi













CHAPTER 1
RESETTLEMENT AND LANDOWNERSHIP TURNOVER ON BRAZIL'S
AMAZONIAN FRONTIER

In the Brazilian Amazon, attempts by the government to resettle displaced

families have not resulted in successful long-term settlement by many smallholders.

Instead, these colonization projects have been plagued by landownership turnover, a

phenomenon in which settled families abandon, sell, or trade their land rather than

staying and developing their farms. It is not uncommon for Amazonian colonization

projects to quickly lose half the residents, and sometimes up to 80 percent leave after only

a few years (Millikan 1988; Martine 1990; Schneider 1995). Often, when these families

leave, they are replaced by others from the ranks of dispossessed people that fill Brazil's

rural zones and urban slums. At other times, as lots change hands, they are snatched up

by land speculators hoping to expand their holdings. As land concentration increases

there are fewer places for these families to go.

Land concentration in Brazil is severe. While 60 percent of the arable land is in

the possession of only 2 percent of the landowners, over half of Brazil's labor force is

landless (Diegues 1992). Amazonian colonization has been one of the few programs in

which Brazil has attempted to alleviate this disparity. Landownership turnover in

colonization projects is an important problem because the poor and displaced families

that were the targets of these projects are left without land and thus, without a place to

live and support themselves. As a result, the plight of impoverished families is

prolonged. Furthermore, resources and time spent developing land distribution projects


1





2


intended to assist the rural poor fall instead into the hands of individuals and groups with

greater economic and political power.

The cycle of landownership turnover is not just a rural problem but also impacts

the quality of life in the region's urban areas. Rural to urban migration by the region's

displaced smallholders, compounded by the migration into the Amazon by the poor from

other regions, has caused Amazonia's urban population to explode (Ozario de Almeida

1992; Schmink and Wood 1992; Browder and Godfrey 1997). In Acre's capital, Rio

Branco, the population jumped from 34,474 to 201,432 from 1970 to 1996 (IBGE cited

in Jomal Gazetta 1996). This represents a dramatic 6 percent annual growth rate. Much

of this growth resulted from governmental development initiatives that encouraged

investment and land sales that displaced rural families.

Amazonian colonization projects are an organized attempt to resettle the large

number of displaced people in Brazil. Most research on resettlement has focused on

involuntary migration: situations where the push factors are greater than the pull factors,

for example, the studies of the plight of people displaced by war, natural disaster, or even

large-scale development projects (see Oliver-Smith and Hansen 1982, or Guggenheim

and Cernea 1993). The smallholders that have migrated throughout Amazonia in recent

decades include both voluntary and involuntary migrants. Some migrating families had

little choice, for example those forcibly pushed off land by the hired gunmen of

competing landowners, or displaced by hydro-electric projects. Other families migrated

willingly, including small-scale landowners from southern Brazil who saw the

opportunity to exchange their small property for a large parcel in Amazonia, or former

sharecroppers looking to find land of their own. All of these diverse migrants are settled





3

together on colonization projects and they all bring different levels of resources and

social support networks. As a result there is a diversity of needs and abilities among this

population.

Colonization programs have an impact on the environment, and when turnover is

prevalent in these projects it can increase the destructive consequences for the region's

rainforest. Settlers clear land to provide for their livelihoods and to survive. If they

suspect that their homesteading attempt will fail, they prepare to move by converting

cleared land into pasture to sell to other landowners. When a family is displaced, it often

moves to new locations farther along the frontier and starts the cycle over again. As the

mass of landless families in the region increases, this cycle accelerates. This has led one

researcher to observe,

The issue is no longer how to prevent deforesting farmers from migrating to the
Amazon from the rest of the country, but how to ensure that farmers already in the
Amazon stay where they have already deforested, thus reducing migration (and
deforestation) farther inland. (Ozorio de Almeida 1995: 5)


Change in Amazonia in recent decades has been depicted as a frontier transition

where peasants act as the vanguard of capitalist penetration to an isolated frontier only to

be displaced with the capitalization of agriculture (Martins 1975; Foweraker 1981; Wood

1982; Sawyer 1984; Bakx 1986, 1988). This transition, sometimes called a moving

frontier (Margolis 1973), has brought drastic demographic, economic, political, and

environmental change to the regions it has crossed. A cynical view could see

landownership turnover and concentration as a product of an elaborate plan by elite

groups (wealthy investors, corporations, large landowners) to expand their holdings by

using and manipulating the poor. While this conspiratorial view is attractive--elite

classes have indeed manipulated the system to meet their goals, and have been the major





4


beneficiaries of change in Amazonia -- landownership turnover and concentration are not

the product of a deliberate, continuous strategy by the elite. Instead, the frontier

transition in Amazonia has been an extremely complex process, with shifting agendas, a

range of competing interest groups and varied outcomes. Colonization trends in Acre

must be understood within this frontier context.

Frontier Transitions and Unintended Consequences in Amazonia


This is a study of landownership turnover on the Peixoto colonization project in

the Brazilian state of Acre. The Peixoto project was first settled in the 1970s, so by the

mid-1990s when this research was conducted it should have been filled with well

established family farms. However, landownership turnover has taken its toll and

affected the pattern of smallholder settlement within the project. The conditions in

Peixoto illustrate examples of smallholder success and failure as they struggle to adapt to

change in the region. While it is one localized example, the outcomes in Peixoto are

emblematic of the end results of broader trends that have transformed Amazonia in recent

years. To explain a phenomenon like landownership turnover among colonist farmers it is

necessary to place this behavior into a broader context that explains the causal processes

at work. By describing specific conditions on the ground, this research allows detailed

examination of the end product of a complex change process.

This change process was described by Foweraker (1981) as an expanding pioneer

frontier that he broke into 3 economic stages (non-capitalist, pre-capitalist, and capitalist).

In Foweraker's account, the transition is set off as a function of authoritarian capitalism.

The intervention of the state is necessary to promote capitalist growth and especially to

guarantee the political subordination of the peasantry. The pioneer frontier expanded in





5

response to the national market, and functions to provide economic accumulation within

the national economy. This expansion began during the first decades of this century with

growing industrialization and urbanization in central and southern Brazil; and the

appearance of surplus labor due to immigration and demographic growth. Because of the

monopolization of land in the Brazilian center, surplus labor was pushed to urban areas or

to rural frontiers.

According to this model, authoritarian capitalism is distinguished by the

articulation of different modes of production (a capitalist mode and subordinate modes).

This is because a dependency relation develops between the frontier and the national

economy with value transferred from the periphery to the center. Economic

contradictions in the center are contained by directing their effects downward to the next

level. The peasants living on the frontier produce within one or more modes of

production and are directly subordinate to landlords who represent another mode of

production. This continued articulation allows expanded reproduction of authoritarian

capitalism.

In Foweraker's model, the peripheral region begins in the non-capitalist stage. In

this stage, the region is remote and isolated from the nation and national markets. Those

peasant producers that move into the area become reliant on extractive activities. There is

an emerging petty commodity sector, but at this time there is no market for land or labor.

The social relations of production are characterized by direct coercion of labor. Although

the region is sparsely populated, the key is that the peasantry starts to take possession and

develop the land.





6


The region undergoing this transition enters Foweraker's pre-capitalist stage when

migration increases and extractive activities intensify. The regular production of

commodities begins, and as this happens the land's value gradually begins to increase.

However, there is still little interest in investing in the land. Goods leave for markets

outside of the region, but little economic value remains in the region. Also, at this stage

there is no 'free' labor market. Because social controls are absent, violence characterizes

the struggle for land and control of labor.

The capitalist stage of the transition is reached when migration intensifies and

access is established to the national economy through road construction. In this stage

agriculture become the dominant activity and is increasingly capitalized. With increased

agricultural production there is a regular, diverse market for goods. Eventually, the value

of land increases and land markets are consolidated. This process began once peasants

started to clear the forest and accelerated with the growth of agricultural production. In

the capitalist stage of development, land has become an important commodity as legal

title is directly convertible to capital.

While peasants occupy the land in the first stages, they are eventually pushed off

in the final stage. In this way the expansion that had taken place during the pre-capitalist

stage created conditions for its substitution. As peasant migrants are drawn to the area,

they clear and occupy land. As they produce for the market, the land increases in value.

Eventually, the peasants are pushed off the land (either directly through violent coercion

or because the peasants cannot compete in the increasingly capitalized market). In the

long-term, the process tends towards the legally secure regime of private property but in

the short-term there is a lack of legal resolution and security in land tenure.





7


Foweraker's model explains the broad trends of the frontier transition process:

state intervention to facilitate capital expansion, the eventual dominance of the capitalist

mode in relation to the periphery, monopolization of land in the center, expulsion of

surplus labor, increased land value and capitalization of production on the frontier and

repeated expulsion of peasants. Although petty commodity production persists, capitalist

relations of production become dominant. He recognizes that the process varies both in

time and place, and it cannot account for all specific details. Foweraker states that even

in the final stage there is still diversity within the frontier because, in this system the

persistence of different modes is necessary to provide cheap seasonal labor for the larger

scale capitalist enterprises. In fact, Foweraker states that there is no reason to suppose

that the transition to capitalist production is ever complete (1981: 55).

Such a broad model of frontier change is less useful when attempting to account

for the wide variation in conditions found on the ground in Amazonia. With such models

it is difficult to explain why some small farmers persist on the land and others move on.

It also is hard to distinguish how and why the frontier transition process plays out in

specific cases.

Amazonia's frontier transition has not produced a smooth flow to a capitalized

economy. Bunker (1985) has called this lack of success of government initiatives the

"underdevelopment" of the Amazon. During the frontier transition, the region has seen

declining productivity, unrest, and environmental destruction as a result of state

intervention. Bunker argued that unequal power relations between Brazil's southern core

and northern periphery underdeveloped the Amazon and caused systematically irrational

results (1985: 239). The complex bureaucratic institutions the state used to promote





8


capital accumulation and social welfare "impeded and distorted the implementation of

development policies and compromised the state's own legitimacy, autonomy, and

bureaucratic authority" (1985: 238). Bunker continues,

The state's programs also intensified the ecological and demographic disruptions
which had limited the capacity of Amazonian social formations to respond in
sustainable and progressive ways to the changing world system's opportunities
and pressures. The resulting socio-economic simplifications and destabilizations
prevented the emergence of enduring or effective local organizations. The
absence of effective civil organizations at the local level reduced the state's
capacity to implement policy in the Amazon. Its complex bureaucracies were
obligated to act in an institutional vacuum, because correspondingly complex
institutional forms on which they were designed to act could not emerge in this
impoverished region." (1985: 239)


The federal agencies used to promote the government's development plans were imposed

on a region that lacked the corresponding institutions and organizations. Therefore,

federal agencies did not function properly. In fact, their inefficiency increased costs to the

state, to the region's inhabitants, and allowed corruption to flourish--ultimately

weakening the state's control.

The competition between different combinations of capital interests and rural

populations occurred not just in a regional context but nationally and internationally as

well (Schmink and Wood 1992) In fact, the complexity of the frontier transition in

Amazonia comes into focus when examining the specific occurrences on the ground. The

state's investment in the region; incentives offered for investment by other groups; and

increased access to the region's land and resources set off an intense competition among

interested parties. As a result, the frontier transition was not a process of linear change

but rather produced a diversity of contested frontiers with highly varied outcomes

(Schmink and Wood 1992: 13). The challenge to researchers in Amazonia is to explain

how specific outcomes from this transition fit into the overall process.





9


An outside observer might ask, why would a government start a transition that

would clearly marginalize a large segment of the population? Or more importantly, why

would large portions of the marginalized population continue to participate in the

transition process? The answer is that during periods of broad social and environmental

change, many specific outcomes are unexpected. Actors in the process exert limited

control in the overall process. Through time, interactions and feedback between

individual actors and between actors and the environment change the underlying

conditions and trajectory of the change process.

Amazonia's rural peoples have participated in the transition because they had no

choice or because they expected to benefit from the process. They quickly learned that

individually they exerted little influence over the changes that affected their ability to stay

on the land. While many have failed in their attempts to find a productive home, not all

have been so unfortunate. Sometimes, rural families mobilized and gained access to land.

However, gaining land was often a Pyrrhic victory when the family could not support

itself with that land.

In a macro-level phenomenon, such as the planned development and the frontier

transformation in the Brazilian Amazon, many specific results were not expected. In this

view, a phenomenon such as landownership turnover is an unforeseen product of

complex processes driven by human behavior but not actually controlled by individuals.

Sanderson (1995) uses the concept of unintended consequences to explain some events in

his model of historical social evolution. According to this model, individuals acting in

their own interest create social structures and systems that are the sum total of the actions

of many individuals. Normally, individuals lack complete information on which to base





10

their plans or actions. They do not know how conditions vary, what actions are being

taken by other individuals or institutions, or what influence other actions will have.

Interaction and feedback influence the change process and through time, the situation

itself may change. The resulting social structures and systems come together in ways that

individuals did not intend.

The individuals involved do not have unlimited choice in their actions but are

instead constrained by their earlier choices, and the choices of others. As the structures

and systems develop, they present constraints within which individual actions must

operate in the future (Sanderson 1995: 13). This view does not deny that human actors

act or shape their worlds, it just emphasizes the limited impact individuals exert over

broader phenomena. As Sanderson points out,

Social evolution represents the long-term consequences of the dialectical
interplay between human agency and social structure.... Human agency is never
something that occurs "freely"; all purposive human action is constrained by both
the biopsychological nature of human organisms and by the social structures that
previous generations of individual actors created through their agency. Agency is
therefore never to be taken as action that is "free" or "voluntary." (1995: 13)


There are many factors that influence landownership turnover among

smallholders in Amazonia, including among others endemic malaria, the use of

inappropriate agricultural technology, weak market infrastructure, debt, and land

speculation. Regardless of the proximal cause in individual incidences of turnover, these

factors are all part of a larger process--a complex chain of inter-related factors that create

conditions that marginalize small farms. This change process is a frontier transition in

which remote, isolated areas are integrated into the nation and the national economy.

Under the marginal conditions that result on the frontier, many rural families cannot (or





11


choose not to) remain on the land. Normally, the individuals with the greatest wealth and

power are able to take advantage of situations where smallholders vacate the land,

thereby concentrating holdings and pushing out the poor.

Purpose of Study


This dissertation will examine landownership turnover on one colonization project

in the Brazilian state of Acre. It will identify the principle causes of landownership

turnover, and describe its effects on the colonization project. Earlier studies of

Amazonian colonization have focused on case studies of individual families (Browder

1994) or on region wide phenomena (Smith 1980; Moran 1981; Schmink and Wood

1992; Ozario de Almeida 1992). The focus of this study fills an intermediate level that

has received less attention: a single project and the communities that make up the project.

This focus on an intermediate level of analysis is done without excluding the study of

individual cases or ignoring links to wider regional, national or international arenas.

The project studied in this research was the directed resettlement project Pedro

Peixoto in Acre, Brazil. Peixoto is the largest resettlement project in the state and was

intended to settle approximately 4,000 landless families. By 1995, when field work

began, Peixoto should have been a thriving autonomous area of self-sufficient small

farms. Instead, families continue to struggle; many have given way to ranches, and those

that remain are still largely dependent on the meager government assistance that is

offered.

This study will describe landownership turnover in the Peixoto project.

Landownership turnover, indicated by the presence or absence of original settlers, is very

high in Peixoto and varies widely from location to location within the project. While, the





12


forces that drive landownership turnover are complex and interrelated, key factors are the

lack of adequate infrastructure and services. In particular, the quality of transportation

infrastructure plays a dominant role since it determines, to a certain extent, a household's

access to other necessities (agricultural inputs, markets, credit, schools and health care).

This research will describe how variation in conditions within this project affects

different communities within the project. The conditions driving landownership turnover

affect the strategies farm families adopt to cope with the harsh conditions. This study

will identify how and when families respond collectively to common problems and why,

more often, they face hardship alone. In addition, this research will identify what impact

landownership turnover, and the conditions that cause turnover, have on land use within

the project.

Outline of Dissertation


In Chapter 2, I explain how the frontier transition that swept across Amazonia in

recent decades was set off by government initiatives intended to develop the region,

integrate it into the national economy, and relieve tensions in other regions caused by

poverty, landlessness, and unemployment. The government's conflicting goals of

stimulating large-scale capital investment and agrarian reform drew a wide range of

competing interests to the region, and generated conflict. Soon the change process was

out of control. One means used by the government to relieve rural tension was by

resettling the disenfranchised on land in Amazonia. However, these initiatives were

motivated more by the desire to alleviate tensions than to assist colonists to establish

successful homesteads. In fact, once families were settled they were left largely on their

own.





13

In Chapter 3, I describe how the frontier transition swept across the state of Acre.

In this chapter, I begin with a description of the historical setting in Acre to explain why

the development transition caused such turmoil in the state. Specifically, because of a

rural power vacuum, the population in some areas had developed systems of land rights

and production out of the ruins of an extractive economy. However, this placed these

people directly in the path of the government's development plans for the state. In the

1970s, the federal government's regional agenda brought about a series of changes that

drastically altered the property rights and agricultural production in the state. As families

were pushed off the land, rural unrest grew, which in turn brought about government

attempts to resolve the growing crisis. The government's reaction was to implement

directed colonization to resettle these people. When turmoil in the countryside continued,

government attention shifted to the new hot spots. As a result of this shift, the

colonization projects that had already been started became lower priorities, and the

families living on them were generally left on their own.

In Chapter 4, I focus on Acre's Peixoto resettlement project to estimate turnover

and identify some of the forces behind this phenomenon. I begin by describing the efforts

that were required to accurately describe turnover in Peixoto. I estimate that

landownership turnover has been very high in the project, and that the amount of turnover

varied from location to location within the project. I argue that the observed turnover can

be attributed, at least partially, to the poor transportation infrastructure and services

provided to residents. As a result of this variation, a distinct settlement pattern has

developed within the project. In some areas family farms have maintained their hold on

the land; however, in other broad areas large land holdings are encroaching on the project





14


and concentrating lands intended for smallholders. This pattern may foreshadow a

difficult future for smallholder settlement in the project.

In Chapter 5, I focus on the exceptional community at Souza Araujo with its low

incidence of landownership turnover. I argue that one of the reasons residents of this

community have stayed on the land is because of the activist organization formed by

community members. This group formed due to a combination of factors including their

settlement as a cohesive group, the presence of strong charismatic leadership, and the

establishment of linkages to non-governmental organizations that helped support the

community. This organization gave the residents the means to directly respond to the

hardship caused by isolation, and maintain their families and farms. However, Souza

Araujo also reveals an enigma: if activist organization helps families to stay on their land,

why have so few communities formed activist groups? There are two principal

explanations. First, the conditions in which Peixoto formed produced a social

environment that was not conducive to rural mobilization. Second, the government

implemented policies (whether intentional or not) that effectively undercut efforts to form

grassroots organizations. This was done by establishing weak associations that the

government could control. Both of these factors developed out of the government's

desire to avoid the mobilization of disgruntled masses in the countryside. The result was

that most colonist families ended up facing the hardship of frontier life on their own, with

the exception of Souza Araujo.

In Chapter 6, I discuss agricultural land use in Peixoto and how current systems

may have developed. Originally, Peixoto was expected to develop into a region of

diversified, self sufficient family farms. These families would rely on perennial tree





15


crops as their major income source. While farms are diversified in Peixoto they have not

developed as expected. Most families can produce enough to meet household subsistence

needs but it is more difficult to meet cash income needs. Cattle and pasture formation

play a much more important role in the colonist farming system than other crops.

Perennials play a very small role in most farm systems. The dominance of pasture in

these systems may contribute to the takeover of project lands by large scale ranches. In

this chapter, I show that an emphasis on pasture is more common on high turnover lots.

As continued efforts are made to reach equitable development policies for the

Amazon and decrease the environmental impacts these changes have brought to the

region, it is important to understand landownership turnover. When smallholders turn

over land in colonization projects, it means they have not been able to take advantage of

the opportunities the land offered and have accepted lesser benefits to cut their losses.

Furthermore, the investments that the state has made to help them have instead benefited

other segments of society. In this process environmental disturbances occur for short-

term gain. This research brings us another step closer to understanding how and why

landownership turnover occurs and what effects it has on people living in colonization

projects and their environment.















CHAPTER 2
THE AMAZON'S FRONTIER TRANSITION AND THE MARGINALIZATION OF
SMALLHOLDERS

During recent decades of rapid change in Amazonia, it has been difficult for rural

families to maintain their hold on the land. How and why this has occurred is complex.

To understand how this situation developed, it is necessary to briefly retrace some of the

trends that affected Amazonia's rural population over the last few decades. The dramatic

change began with federal government intervention in the region intended to reach two

general goals. First and foremost, the government wanted to develop the region and

integrate it into the national economy. Second, the government wanted to use Amazonia

to relieve tension in other regions due to the landlessness and unemployment that resulted

from earlier development. The government's policies and incentives drew diverse groups

of competing interests ranging from wealthy investors to poor migrants.

As this chapter will explain, early in the transition the government began to lose

control of the change taking place. Expulsions of rural people by development led to

rural unrest and, along with spontaneous settlement, threatened the government's

development agenda. In response, the government attempted to control settlement and

resolve conflicts by creating elaborate colonization schemes. However, these

colonization programs were unwieldy and ineffective. Through time, the initial grandiose

plans were continuously scaled back to simplify programs due to high cost, difficulty of

the endeavor, and political opposition. While intended as comprehensive solutions to



16





17


general problems faced by rural people, in reality the less ambitious colonization projects

amounted to little more than land distribution programs. With land in remote isolated

areas lacking services, it was difficult for families to meet more than very basic

subsistence needs. Consequently, many colonists either failed to support themselves or

chose to sell out when given the opportunity.

Government Policies and the Bias Toward Large-Scale Capitalist Enterprises


The Brazilian government initiated the development boom in Amazonia with two

primary goals: to increase the economic productivity of the region and to integrate it into

the national economy. While colonization has received much attention in the Amazonian

development literature, resettlement was not originally a priority of Brazilian

development planners. The government was consistently biased towards large-scale

capital investment. This bias was intensified by the economically and political powerful

(the landed elite, large corporations, and wealthy investors) who were able to influence

the institutions involved in development to meet their own needs.

From the point of view of government planners, Amazonia offered an attractive

opportunity for capital investment because of its vast mineral deposits, timber, and

unused agricultural lands. The government stimulated capitalist investment in Amazonia

three ways: 1) public financing and construction of infrastructure, 2) concessions of

resource rights to foreign companies, and 3) tax credits and fiscal incentives to Brazilian

companies (Bunker 1985). The government wanted capitalist enterprise to replace

traditional forms of rainforest extraction that dominated the region and were seen as

inefficient. The idea was to facilitate the entrance of corporations and investors by

making participation and investment more attractive.





18


The frontier transition process began to accelerate as the federal government

started building roads to integrate the region and tie it to the southern states by land. The

road building phase began in 1959, when SPVEA (Superintendencia do Piano de

Valorizagdo Economica da Amazonia) initiated a series of road building projects from the

Center-south to the Northeastern, Northern and Northwestern peripheries of the country

(Bunker 1985: 84). These intra-regional highways were intended to integrate isolated

regions into the nation and national economy. The eastern Amazon was opened by the

Belem-Brasilia highway which was completed in 1964. In the Northwest, the Cuiabi-

Acre road was gradually extended across Mato Grosso and Rond6nia. It was intended to

eventually reach across Acre through the city of Cruzeiro do Sul to the Peruvian border

(Nissely 1966). These roads would create ties to national industrial centers in the South

and open lands for investment.

The development policies encouraged the participation of diverse interest groups

from corporate investors to landless migrants, all of whom entered to compete for the

region's resources. As investors began buying up vast tracts of land, it required the

displacement of the region's smallholders. However, the roads not only attracted wealthy

entrepreneurs, but also allowed the entrance of poor migrants searching for land. That is,

the roads set the stage on which conflict would play out later and foreshadowed future

problems for the government's development plan.

After the 1964 coup, the new authoritarian government intensified the

transformation on the frontier. In 1966, the military government's Operation Amazonia

created a number of bureaucratic institutions like BASA (Banco da Amaz6nia) and

SUDAM (Superintendencia de Desenvolvimento da Amaz6nia) to grant concessions and





19


administer tax credit and fiscal programs. BASA provided financial incentives for the

new investors and at the same time cut credit for traditional extractive activities.

SUDAM, on the other hand, administered federal incentive programs that allowed

companies to reinvest taxable income in Amazonia. With its wide ranging power to

distribute incentives and certify projects for participation, SUDAM could channel

development in favor of capitalist enterprise.

Together, these two agencies facilitated the entrance of large scale enterprises

from outside the region, bringing dramatic change during the 1960s and 1970s. The

minimum project size to qualify for SUDAM's financial incentives was set at 25,000

hectares (Bakx 1986: 125). Initially, mining and timber extraction were the main

industrial investments that qualified. In 1968, SUDAM extended its fiscal incentives to

large ranching enterprises in Amazonia (Mahar 1979). Companies that had been

established in Amazonia before 1974, or that had been certified to be in the region's

economic interest, received a ten year tax holiday and attractive credit terms (Schmink

and Wood 1992: 60). Ranching received the largest share of tax breaks and along with

timber companies received most of the credits (Schmink and Wood 1992).

During the early 1970s, the conspicuous bias toward large-scale capitalist

enterprise was diminished briefly. In 1970, the Medici government announced the

Programa de Integragdo Nacional (PIN) which was to be the "cornerstone of efforts to

modernize Amazonia" (Schmink and Wood 1992:1). A major part of the PIN was the

Transamazon highway (BR-230) and the Santar6m-Cuiabi highway (BR-165), projects

that would further open up Amazonia for development. These roads would better

integrate the region into the nation and help with its defense. The road building plans





20


included colonization as a facet of the overall plan.1 The military government saw

colonization in Amazonia as a means to alleviate landlessness and unemployment in the

impoverished Northeast, lowering tension in that region without carrying out an actual

agrarian reform.

After 1973 there was a major shift in the government's Amazonian policy.

Colonization had been opposed by groups that had benefited from earlier subsidies.

These elite groups exerted political pressure and attempted to depict smallholders as a

predatory threat to Amazonia and, conversely, ranching as benevolent to the environment

(Wood and Schmink 1979). This lobbying was successful. After only a few years the

government de-emphasized colonization citing high costs and poor results in the first

years.

In 1974, as an alternative to the earlier program, the Geisel administration began

POLAMAZONIA which set out the new regional agenda, that was openly pro-big

business. With POLAMAZONIA, the government's objective was to designate growth

poles that would focus public and private investment on areas deemed to have economic

potential. Investments such as mining, timber extraction, and ranching qualified for the

program.

The efforts of government agencies did not produce the desired economic

development. SUDAM was supposed to supervise the implementation of these projects

but accounting was done from offices in the south making effective supervision "virtually

impossible" (Schmink and Wood 1992: 60). This allowed many investors to participate



1 Sawyer (1984) suggested that colonization along the Transamazon was not just to
occupy a vulnerable region but also to mask the military's motives and justify the
expense with a populist land distribution program.





21


as speculators or even invest their money outside the region. In 1978, critics of SUDAM

used a parliamentary commission to expose the lack of success of these programs. From

1966 to 1978 SUDAM spent $5.7 billion in the region (double the state of Para's annual

budget) and still Amazonia's economic position declined (Schmink and Wood 1992: 97).

Furthermore, benefits were not distributed equally. In Amazonia 85% of the properties

were owned by smallholders (holdings of less than 100 hectares) but only 3 small

properties had been supported with SUDAM's subsidies (Schmink and Wood 1992: 97).

BASA incentives from 1966 to 1979 went mostly to large enterprises based outside the

region. The irony is that much of the investment activity that took place in the region

was speculative and did not increase production or create jobs for those pushed off the

land (Goodland 1980; Hecht 1984).

Not only did these projects not produce the desired economic development, they

had a devastating impact on Amazonia's rural populations. The large projects SUDAM

encouraged required huge parcels of land. Most areas were occupied by rural people

without documented ownership. Because of this insecure tenure, many families were

displaced.

Amazonia's Insecure Land Tenure


Amazonia's land tenure system presented a bottleneck for development plans

from the outset. This system was a product of the earlier extractive mode of production

(Bunker 1985). When Amazonia's economy was dependent on forest extraction,

landholdings were characterized by poorly defined boundaries and unique systems of

traditional rights. The traditional systems defined access to resources, such as rubber

trees or Brazil nut groves, not to the land itself. When documents existed, they often





22


covered vague areas with indistinct boundaries. Poor documentation was further

confused by the existence of multiple, overlapping deeds issued by various government

agencies (see Bakx 1986; Schmink and Wood 1992). The lack of consistent records, and

the speed with which change was taking place led to widespread fraud. This situation

was further complicated by a Brazilian law which provided legal rights to land ownership

through occupation, or posse, which can be roughly translated into English as squatting.

According to Brazilian law, posseiros can establish posse on untitled, unoccupied land

(terra devoluta) if they live on, and cultivate the land for one year and a day.

After the collapse of the rubber economy, most large landowners did not actively

maintain use of their holdings and left rural people on the land on their own. Therefore, a

large portion of the rural population in Amazonia could claim legal rights to the land they

occupied even though they had no documentation. These posseiros presented a dilemma

for the government. Development plans required these people to move for project

implementation. However, the posseiros were not always willing to move, and their

resistance became a thorn in the side of government planners.

In 1970, in an effort to overhaul the tenure regime and facilitate the transition, the

government combined weaker titling agencies to form INCRA (Instituto National de

Colonizacdo e Reforma Agraria). INCRA was responsible for the administration of

federal lands, land titling, and the administration of colonization projects.2 The agency

was to serve as an arbiter between interests groups competing for land. Actually, much


2 INCRA's main tasks were: "1) land discrimination to separate public from private land
and clarify tenure status; 2) appropriation of unclaimed federal patrimony; 3) demarcation
and surveying; 4) land titling, including the regularization of past titles and occupations;
5) expropriation for social purposes and agrarian reform; 6) settlement and colonization;





23


like other government agencies, INCRA favored large investors over smallholders both

directly and indirectly.

Like SUDAM, INCRA was a powerful and autonomous agency because it

controlled public land and collected land taxes. The agency's incentive was to allocate

financial and administrative resources for large-scale land sales and titling rather than for

expensive and politically controversial colonization schemes (Bunker 1985: 110).

Furthermore, large-scale investors used their economic and political power to pressure

INCRA into favoring large capitalist interests in disputes over land. The pressure came

from many sources. For example, banks, aware of the demand for credit in the region,

pressured bureaucracies to recognize claims of large-scale capitalist enterprises so they

could use the titled property as collateral for loans (Bunker 1985: 90). The military

government even produced secret directives that allowed INCRA to accept fraudulent

claims that undercut those of smallholders if, in INCRA's view, the suspicious claims

"promoted the development of the region" (Schmink and Wood 1992: 64). Undercutting

legal procedures led to increased fraud.

While technically the posseiros had legal protection, in reality it was difficult for

them to claim their rights (Bunker 1985: 120). Isolated families found it difficult to resist

the pressure placed by gunmen hired by landowners, especially when supported by local

police and the unsympathetic judicial system (Americas Watch 1991). The costs of

documenting a claim often exceeded the ability of rural families to pay. The process

required that they travel to and stay in urban centers, often with unforeseen delays. Some

services, like land surveying, were very expensive. While these costs deterred most rural


7) maintenance of a cadastre with data on rural properties; 8) enforcing and collecting the





24


families, they did not inhibit large companies competing for the same land. Furthermore,

wealthier enterprises were more familiar with the system and could use their power to

influence the courts and police.

Across the region from Para in the east to the western states of Rond6nia and Acre

tension in rural areas increased. The new landlords were confronted with rural

populations that occupied the land. These confrontations had a variety of outcomes.

Sporadic violence broke out as owners' henchmen threatened and coerced the posseiros

to leave with virtual impunity. The rural poor often suffered the brunt of this violence.

Overall, hundreds ofposseiros were killed in land conflicts throughout the region in

recent decades (CPT 1989; Americas Watch 1991). Those pushed off the land either

moved to urban slums to look for employment or moved to other rural areas to start new

posse, to sharecrop, or to work on a ranch or placer mine. Some groups ofposseiros,

aware of their plight, stiffened their resistance to expulsions. Others agreed to accept

payment for "improvement" (i.e. land clearing) of the land. With few opportunities to

gain land some families realized they could invade land holdings and wait to be paid off.

As development efforts progressed, spontaneous settlement by migrants quickly

became a major problem for government institutions. Spontaneous settlement occurs

when families occupy land and begin homesteads without official recognition and

minimal government involvement. This is the dominant means of rural settlement in

rural Amazonia (Moran 1981, 1982). The newly created infrastructure in Amazonia

allowed a massive immigration by small holders who joined the region's rural residents

in claiming land through posse. Such settlement was unprecedented; millions of people



rural land tax; 9) sale of public land; 10) arbitrating disputes." (Mueller et. al. 1994: 269)





25


settled along the Belem-Brasilia road in less than a decade. After the initial occupation,

the smallholders who had settled near the road were either bought out by the ranchers or

eventually forced off the land (Sawyer 1984).

All these outcomes were sources of concern for government planners. Conflicts

increased the risk for the outbreak of wider rural unrest or uprising. The disputes delayed

implementation of what were seen as beneficial economic activities, brought by the new

investments. At the same time, the government was being criticized for its treatment of

rural populations in Amazonia, for the growing land concentration and unemployment in

other regions. The government needed to respond to these growing problems. Tension

and violence often broke out over land conflicts as posseiros and spontaneous colonists

attempted to defend their claim to the land. These conflicts were delaying or even

jeopardizing the government capitalist development agenda. Furthermore, the

government feared that this unrest could potentially become a more organized rural

resistance. Finding an adequate response would be tricky. The government needed to

alleviate this tension without alienating elite groups.

The Evolution of Amazonian Colonization Programs


The evolution of colonization programs in Amazonia must be understood to place

the marginalization of rural families in the proper context. The Brazilian government

responded to the growing rural unrest with colonization programs that were intended to

end the crises by giving rural people land and an opportunity for livelihood. The key here

is that the government took action and did not let the situation go too far. INCRA (and

other agencies) were usually one step behind problem situations and were continually

adapting colonization programs to evolving situations. After the initial failure to foresee





26


the needs of rural smallholders in Amazonia, the government has never been able to

move ahead of this problem. Originally, resettlement was a minor issue but turmoil

caused by displaced people and spontaneous settlement demanded a government

response. Federal efforts began as grandiose schemes with detailed plans to meet all

needs and guide the process to develop productive rural communities. The government

quickly began to retreat from ambitious plans due to difficulties, costs, and growing

opposition from elite interests. Programs gradually became simpler and required less

long-term responsibility on the part of federal institutions.

When the development transition began in Amazonia, government planners did

not foresee the need for comprehensive colonization policies. It was not until the

region's displaced rural people and migrant families began to create problems for

development plans that the government acted. As part of the PIN and Transamazon

programs, INCRA introduced a colonization model called the PIC (Projeto Integrado de

Colonizagdo). Along each side of the new roads the government set aside a 100

kilometer strip to be used for small farm colonization. The PICs were comprehensive

programs, where INCRA was not only responsible for survey and titling of farm lots but

also the construction of basic infrastructure, as well as for organizing all services needed

by the colonists. During the early years, PICs were also implemented in the western

Amazon.

The PIC design was intended to resolve recurring problems that had plagued

settlement in tropical areas such as the lack all-season roads; inadequate credit; poor

health care and educational facilities; unclear land titles; and isolation from markets

(Moran 1984). The PIC plan divided project areas into a grid of 100 hectare lots along





27


highways and feeder roads. It dictated colonist settlement patterns through the

construction of a series of planned population centers ranging in size from rural

population clusters (Agrovilas) to urban centers the (Agropolis, and Ruropolis).3 These

centers would facilitate the delivery of services to the rural population. The PICs were to

be implemented in stages. Eventually, after about 3 years, they would be turned over to

the states once project infrastructure (roads, schools, and health posts) was in place, land

was titled, and urban centers had formed (Mahar 1981: 50).

From the outset the PICs were plagued with problems. While plans went to

minute detail organizing the projects, the PICs were based on fundamentally flawed

assumptions. They were planned from Brasilia without the benefit of feasibility studies

in the field. As a result, the projects were superimposed on the landscape ignoring

topography and watersheds. Because of the rolling hills, streams, and swamps, roads

were hard to build and maintain. Additionally, farm lots were highly variable in

suitability for farming. The farmers' problems were compounded by their unfamiliarity

with the environment (Moran 1982) or the lack of effective technical advice by extension

agents familiar with the environment (Wood and Schmink 1979). Additionally, INCRA

and other agencies were not able to provide the necessary services in a timely manner

"leaving colonists attracted to an area to fend for themselves" (Moran 1982:10). INCRA

tried to select colonist families that were expected to survive, based on the amount of

household labor, capital, and agricultural experience. The irony of the selection process

is that it excluded many who needed assistance most. The rationale for the selection





3 See Smith 1980, Moran 1981 for in-depth discussion on the organizational plan.





28


process assumed that the colonist families settled earlier were largely responsible for

failure themselves, not the project organization (Bakx 1986; Wood and Schmink 1979).

A critical problem for colonists was the lack of services. Responsible institutions

could not maintain roads and provide transport to get crops to market (Wood and

Schmink 1979). Families were not provided adequate health care, schools, and

agricultural storage. Financing for small farmers was another major problem. Colonists

were caught in a catch-22: they needed title to get credit; to get title they needed to

improve their land, and to do that they needed credit. Credit for annual crops was risky

for colonist farmers. It was expensive (high interest rates and high transaction costs), and

if the family did not produce a good crop they fell into debt (Moran 1981). Often,

families fell into a cycle of debt and were forced to seek wage labor to make payments

(Collins 1985). As a result, they could not invest labor in their farms. Once wage

opportunities disappeared, they were forced to sell the land.

The goal of the PINs was to settle 100,000 families along the Transamazon during

the first five years. In 1976 only 7,128 had been settled (CEDEPLAR 1979) and by 1978

this figure had increased only to 7,900 families (Mahar 1981: 46). In Rondonia INCRA

had hoped to settle and title 23,439 families in PICs started in 1973; however, by 1978

only 6,929 had been titled (Mahar 1981: 51). It was not only hard to settle the families

but also difficult for the families to stay on the land. In 1974, 18 percent of families in

some areas had already abandoned their land (Martine 1980). By 1979 some areas had

lost up to 44 percent of original their original colonists (Smith 1982: 170)

While it was difficult to settle and maintain families on the land, the demand for

space in the PICs continued to expand. For example, the Ouro Preto PIC in Rondonia





29


was originally designed for 500 families but had to be expanded to accommodate 5000

families. By 1975, six more PICs had been installed in Rond6nia with the intention of

settling 22,700 families (Martine 1990). The PICs had been high profile projects and the

government had made a point of encouraging the landless and the unemployed to move to

land that was being made available. This triggered an increasing wave of migration, with

migrants moving in spontaneously as roads opened up. When they could not receive land

through official channels, they simply moved to unoccupied areas, invaded the land and

waited for INCRA to regularize their holdings. Spontaneous settlement was becoming a

greater problem. For example, by 1976 approximately 8 percent of the families in

Rond6nia were squatting on public or private lands waiting for titled lands in projects

(Mueller 1980: 149).

With the shift to the POLAMAZONIA program in 1974, development emphasis

switched again to large-scale capitalist investment. However, the federal government

could not abandon colonization efforts completely. There were too many families

waiting for land and the return of big investment projects would mean even more

displacement of people in the region.

The federal government's new agenda for colonization was to simplify projects

and shift responsibility to other agencies. Up to 1975, INCRA tried exerting

comprehensive control over settlement areas and until that year public services were

provided almost exclusively by INCRA (Mahar 1981: 49; Martine 1990: 27). In some

areas, the government experimented with private colonization projects. In these schemes,

the government sold large parcels of land to corporations that would develop the area and





30


sell lots to colonists.4 However, the government remained the primary colonizer in

Amazonia.

As part of the new agenda, government colonization programs switched to

directed settlement projects known as PADs (Projetos de Assentamento Dirigido). PADs

were similar to PICs, but were attempts to learn from mistakes and improve the system.

The major difference with the PADs was that hypothetically they would be easier and

faster to implement. This was because INCRA had a more limited role in supporting the

colonists once they were settled. These projects relied on greater participation by state

and municipal government agencies.

In the PADs, INCRA was responsible for 1) land distribution, 2) demarcation of

territory, 3) administration of the project, 4) organization of settlement process, and 5)

construction of the physical infrastructure. This meant that state agencies were

responsible for providing education, health facilities, housing, credit, and technical

services (Cavalcanti 1994). To provide this support, a wide variety of state organizations

were enlisted. However, these agencies often lacked the organization, resources, or the

will to adequately support colonization projects.5 Through poor planning, inadequate

funding, confused jurisdiction and political infighting, the crucial needs of colonists often

went unfulfilled.

INCRA estimated that it would take three years to implement the PADs, then five

more years for the projects to be consolidated and free from INCRA administration.


4 Private colonization projects are not within the scope of this work. These projects
generally attracted wealthier colonists who tended to do better than the colonists in public
projects. (See Moran 1990; Ozorio de Almeida 1992; Schmink and Wood 1992).
5 See Bunker 1985 description of the extension service's (EMATER) funding problems
that distorted services.





31


Eventually, after a total of seventeen years, INCRA expected the PADs to be fully

integrated into the economy and totally "emancipated" from INCRA's assistance

(Cavalcanti 1994). However, INCRA was unable to maintain these timetables. On huge

projects it took longer than expected to settle families and build infrastructure. The

projects were built in stages, and as each area (gleba) was finished, INCRA moved on to

another. While opening new glebas, INCRA had to spend time maintaining the old, plus

monitoring the colonists' progress toward earning the definitive title. It has not been easy

to emancipate the projects and the colonies have had a hard time integrating into local

economies. Many settled families were simply abandoned on their lots regardless of

INCRA's and other agencies' best intentions.

By the late 1970s immigration continued to overwhelm INCRAs attempts to settle

rural families. In 1977, the government set up road-blocks in Rond6nia to turn back

newly arriving migrants. Eventually, INCRA had to modify the colonization process

even further to respond to the waves of immigrants pouring into the region. In 1979 and

1980 INCRA streamlined the process.6 They brought in additional employees, cut down

on documentation required for colonists and even provided definitive titles immediately

to some qualified families (Mahar 1981). New settlement projects were to be started as

quickly and efficiently as possible. INCRA began "rapid settlement programs" (PARs,

Projecto de Assentimento Rapido) which legitimized spontaneous settlement along roads

or even invasions along rumored paths of future roads.





6 At this point Rond6nia's government switched its emphasis from discouraging to
encouraging immigrants to increase population to support efforts to gain statehood
(Martine 1990).





32

New settlement goals remained elusive. In 1982 in Rondonia, 15,000 families

were to be settled in 4 projects, but 4 years later only 5000 had been settled (Martine

1990). As attempts at organized settlement proved difficult, spontaneous settlement

continued at a high rate. Although not publicly abandoning the pretense of controlling

the settlement process, for all practical purposes at this point INCRA's emphasis

switched to simply documenting a process it could do little to stop. It was likely that

large portions of the farms which INCRA claimed to have settled through "rapid

settlement" or "land tenure projects" were actually the product of spontaneous occupation

(Martine 1990: 29).

INCRA's rapid settlement programs were responses to localized land conflicts.

While the government tried to control the process and impose order, its responses to

problems actually undermined its stabilization program. This is because whenever

INCRA regularized an invaded area, it encouraged other invasions. Eventually these ad-

hoc responses to land conflicts were institutionalized producing what has been called

"crisis colonization" (Schmink 1982, 1988, Schmink and Wood 1992).

The agency responsible for crisis colonization was GETAT--the Executive Group

for Lands in the Araguaia Tocantins Basin--a special institution with the power to

distribute land titles and thereby eliminate sources of unrest in that area of the eastern

Amazon (Schmink 1982; Schmink and Wood 1992). GETAT was created in 1980 by the

Brazilian federal government in response to the increasing violence and unrest in

southern Pard resulting from contested land claims and the inability of dispossessed

families to acquire land. GETAT had more political and economic power than preceding

institutions and more bureaucratic autonomy, which allowed it to cut through red tape





33

and place families on farm lots. The focus of GETAT's activities was to settle as many

families as possible as quickly as possible in contested areas. To that end, GETAT

placed titling quotas on its field offices and cut the minimum size of farm lots down to 50

hectares. Titles were often distributed during dramatic public ceremonies as part of

GETAT's strategy to "build grassroots political support in areas where violence and the

persistent conflict over land fostered growing opposition to the military regime"

(Schmink and Wood 1992: 176).

GETAT's agenda had little to do with the long-term agricultural sustainability of

the colonist families once they were settled. In fact, few of GETAT's activities focused

on providing the services that are crucial for establishing small farms, such as credit or

technical assistance. In 1981, GETAT regional personnel admitted that the fifty hectare

lots they distributed were incapable of sustaining viable small farms in the long run

(Schmink and Wood 1992: 178).

Crisis colonization led to greater turnover and land concentration. Land

consolidation was the result of two processes. First, once GETAT had titled a farm lot it

became an asset that easily could be turned into cash. On these small lots, many families

were in precarious economic situations. If an illness or emergency arose, the family often

had little choice but to sell out. The second process resulted from GETAT's policy

toward ranchers. When a land invasion took place, it was often easier to offer the rancher

a larger parcel elsewhere, rather than attempt to move all of the squatters off the contested

land. Sometimes ranchers manipulated the system to their advantage by encouraging

land invasion, knowing they would get better and larger parcels and leave the degraded

land to smallholders.





34


Ultimately, crisis colonization did not eliminate the types of social unrest that the

military government feared most. While GETAT was able to lower tensions and resolve

disputes in the short term, in the long-term these problems actually may have grown

worse (Schmink and Wood 1992). GETAT's autonomy caused it to undermine the

authority and operation of other institutions in the region. By siding with squatters to win

their grassroots support, GETAT demonstrated the government's fear of social unrest or

conflict that could lead into revolt. Squatters learned that they could get the government's

attention and often win disputes if they stood their ground, organized resistance, and

threatened social disruption. Consequently, the support that GETAT was able to generate

for the government through its actions was short lived.

Rural Mobilization and International Alliances


As rural people realized the adverse effects that were brought to their livelihood

and well-being by Amazonian development, they began a gradual process of rural

mobilization. Initially, local groups organized to resist expulsions from land. Later rural

unions, the Pastoral Land Commission (CPT) and the Catholic church began assisting

these people with their organizing efforts.

The rubber tapper movement provides an excellent example of how alliances can

be formed successfully to collectively fight for rights. Through time, groups of rural

families learned that together they could exert great influence on the change process. The

rubber tapper movement began with small groups of rubber tapper families forming

"empates," or strikes to halt the clearing of forests that were key to their livelihood.

Eventually, these groups coalesced into a national movement to attempt to have a greater

voice in development plans for the region (the National Council of Rubber Tappers,





35

CNS). The rubber tappers began forming alliances with other groups that had

corresponding agendas. In 1985, the CNS formed an alliance with indigenous people.

As these groups gained higher profiles, they formed international alliances with

environmental and human rights groups interested in monitoring conditions in Amazonia

(Schmink and Wood 1992).

The rural movements learned they could draw the attention of world press and

international institutions. The international alliance between rubber tappers and

environmental groups brought together organizations with political clout in lender

nations, with people that had the ability to monitor projects in the field and verify that the

government complied with environmental and social clauses in the loan agreements. The

growing influence of the alliance between rural peoples in Amazonia and international

interest groups is best illustrated in the conflict over the paving of the BR-364 from Porto

Velho, Rond6nia to Rio Branco, Acre.

The BR-364 highway was first paved as far a Porto Velho as part of the

POLONOROESTE project. This initiative began in 1982 with financing from the World

Bank, with the goal to pave the BR-364 highway into Mato Grosso and Rond6nia and to

develop sustainable farming systems for families migrating into the region.

In 1985, the Inter-American Development Bank (IDB) agreed to finance the

paving of the BR-364 from Porto Velho to Rio Branco. However, under pressure from

the CNS and international environmentalist groups, the IDB negotiated an agreement that

made the loan contingent on measures to protect the environment and indigenous peoples

in the area. This agreement was called the PMACI (Projeto de Protegdo do Meio

Ambiente e das Comunidades Indigenas) (Schwartzman 1989).





36

Through monitoring, international environmentalist groups with the assistance of

grassroots organizations in the region were able to document that this was not being done.

With this information, groups in the United States could put pressure on Congress to

withhold funding for the IDB. Chico Mendes, the leader of the rubber tappers movement

even traveled to Washington, D.C. to testify on behalf of Amazonian populations. In

1988 the IDB suspended funding for non-compliance with the PMACI, the first time ever

an international project was suspended on environmental grounds (Schwartzman 1989).

Eventually, in 1989 a new agreement was reached between the IDB and the Brazilian

government, and the BR-364 was finally paved as far as Rio Branco, Acre, in 1992.

Although the road was paved, in the end rural people were able to influence the process

and gain concessions.

The rubber tapper movement had been pushing for recognition of the land rights

of families living in deactivated rubber estates. In 1988, they won a significant

concession from the government when the first extractive reserve was formed. Extractive

reserves were a new type of ownership classification that recognized the usufruct rights of

forest populations. This brought land disputes for some of the region's smallholders full-

circle by recognizing theirposse claims and maintaining the status quo. Allowing them

to stay on their land was the first step toward recognizing traditional land rights.

Democratization and the Return to the Agrarian Reform Debate


After the mid-1980s, Amazon colonization and resettlement began to lose its

salience in the national debate. On the national stage, the battle returned to land

redistribution. The democratization process brought back the possibility of the national

agrarian reform debate that had ended with the 1964 coup. Much like the organized





37

movements that sprang up in Amazonia to influence policy, groups of both sides of this

volatile issue formed organized movements or pressure groups to influence the debate.

Politicians in the late 1980s and early 1990s struggled a balancing act between opposing

camps. With the political opening under democracy, special interest groups had

increased influence over government policies which had led to an "almost uncontrolled

expansion of fiscal incentive programs although programs had clearly failed in previous

decades" (Mueller et al. 1994: 270). As the debate over redistributing Brazil's highly

concentrated land holdings intensified, families that had gained rights to the land in

Amazonia were forgotten and there was little discussion of how to keep them on the land.

Examining the focus of agrarian reform illustrates a basic problem with the debate: since

it was almost solely concerned with landless families, it did not include programs to meet

the needs of families that had already received land. Neglecting to help those families

already on the land meant that settling additional families would not produce a net

improvement in land disparity.

Organized groups trying to influence the agrarian reform debate came from both

sides of the issue representing both the landed elite and the landless masses. A segment

of Brazil's land holding minority formed the UDR (Unido Democrdtico Rural) to defend

their interests even through armed force if necessary. The UDR gave landowners high

profile and a loud voice in the raging agrarian reform debate, especially so during the

1988 Constitutional Convention (Tavares 1995). On the other side of the issue was the

Landless Rural Workers Movement (Movimento dos Trabalhadores Rurais Sem Terra,

MST). The MST formed in 1984 in the Brazilian south but soon had spread to much of

the rest of the country. The MST used a confrontational style to force its hand in





38

implementing agrarian reform. The primary tactic was to organize and carry out mass

invasions and occupations of large, under-used estates. Once on the land, the MST

members would build small settlements and begin farming. By the mid-1990s the MST

had grown into a force to be reckoned with on the national stage, with widespread

popular support. However, before analyzing the efforts of these movements, it is

necessary to describe how the agrarian issue played out in the national political arena.

The return of the agrarian reform debate started when Tancredo Neves was elected

president in 1985. His objective was to "promote an austere economic policy so as to

tame the industrial sector and urban trade unions, while at the same time undertaking a

bold agrarian reform aimed at eradicating rural poverty" (Tavares 1995: 25). However,

before Neves could implement his ambitious plan he died only one month into his term.

Neves' vice president, Jos6 Sarney became president. Sarney was from the Northeastern

region, a region with the most severe land concentration in the country. He had been

included on the presidential ticket to provide balance. Sarey did not have as ambitious a

plan as Neves, but he could not fully ignore the agrarian reform issue. Sarney formed the

Ministry of Agrarian Reform (MIRAD) and placed pro-reform leaders in the Ministry of

Agriculture and in the Brazilian Association of Agrarian Reform. He also proposed the

first National Agrarian Reform Plan (PNRA), with the goal of expropriating 43 million

hectares and eradicating rural poverty in 20 years (Tavares 1995: 25).

However, the PNRA was never implemented and it did not expropriate land as

intended. Not only did PNRA lack broad support within the government, but the political

left was opposed to the program as long as it was supported by Sarney (Tavares 1995).

One of the criticisms was that it separated land policy from agricultural policy. This





39

meant that benefits were focused on those without land, leaving smallholders "totally

excluded" (Tavares 1995: 25).

The agrarian reform debate not only failed to address the problems of colonists

that had already received land in Amazonia, but the battle disrupted the remaining

programs that were serving smallholders. For example, in 1987, during the debate,

INCRA was dissolved and its duties and resources were transferred to the Ministry of

Agriculture. A year and a half later this decision was reversed, possibly because

powerful interest groups lost economic and political rents (Mueller et al. 1995). This

bureaucratic turf war caused much hardship for colonists waiting for land and title in

areas where INCRA was the major institution and service provider.

The organized landowners and their allies were able to paralyze the agricultural

reform debate and even turn it to their advantage. When the 1988 Constitution was

ratified, it banned the expropriation of "productive lands." Since then it has been difficult

for INCRA to expropriate land.7 When land was acquired, it was purchased, a type of

reform which avoids politically sensitive actions. (Mueller et al. 1995). Land purchases

for resettlement also provided large landowners a convenient escape from bad

investments. Sometimes large landowners encouraged invasions so they could be bought

out by the government (Schmink and Wood 1992).

After the initial fiery debates, movement toward agrarian reform lost momentum,

usually surfacing only as ambitious campaign promises. In the 1989 presidential



7 In the 1946 Constitution the state was responsible to see that land was used for social
welfare objectives, and obligated to expropriate idle lands. But article 141 required that
indemnification take place in cash; this limited the government's ability to expropriate
unused lands. A 1964 amendment removed the clause (Mont6iro 1980; 102). However,
with article 184 of the 1988 Constitution this clause was reintroduced.





40


election, agrarian reform was an important but not the central issue. In the 1994

presidential election, agrarian reform was not a prominent issue but was included in the

platforms of the principal contenders. Again, the programs focused on land distribution,

not on providing services to rural families with land.

The winner of the 1989 election, Fernando Collor de Mello made an incredible

pledge: he proposed agrarian reform projects that would resettle 500,000 families

(Tavares 1995). Once elected, he did not attempt to reach these amazing settlement

goals. In fact, Collor's policies caused greater hardship to those families already settled

because they disrupted the agencies that were still providing services in rural areas.

Collor shut down MIRAD, transferring its functions to the Ministry of Agriculture.

Then, during the 1990 economic crises, Collor cut employees and funds from federal

agencies as an anti-inflationary measure. Agencies like INCRA or SUCAM (the malaria

control service) were left with fewer employees, vehicles and equipment. These

measures introduced a period of hardship for colonists; many left their land and land

concentration accelerated.

The main opposition candidate in both the 1989 and 1994 elections, Luis Inacio

"Lula" da Silva tried to make agrarian reform a key point. However, his proposal

emphasized land distribution, not agricultural policies and rural development. In 1994

Lula's platform was to settle 800,000 families on 30 hectare plots in 4 years. To achieve

this goal, Lula proposed expropriating 24 million hectares of "unproductive land" at a

cost of 8 million dollars, excluding the price of land (Tavares 1995: 27).

The winner of the 1994 election, Fernando Henrique Cardoso ran on a platform

that pledged to adopt agrarian policy that would settle 40,000 families in the first year;





41


60,000 the second year; 80,000 the third year; and 100,000 the fourth year (Tavares

1995). However, Cardoso provided no details of how these numbers would be reached.

Considering that during the early 1970s the PIN failed to meet its promise of settling

100,000 families in a decade, one might question the sincerity of these campaign

promises.

Once Cardoso was elected, the MST began to exert increasing pressure for land.

The MST had gone from a secretive organization that carried out clandestine occupations

to a high profile group. They drew international attention by staging day time operations

with armies of landless workers parading down highways to occupy underused estates--

all announced in advance so television cameras could witness the events (The Economist

1996a). The MST claimed that between 1986 and 1995, it had settled 139,000 families

on 7.2 million hectares (Homewood 1996). In 1996, INCRA estimated that there were

22,000 families encamped along roads waiting for the opportunity to occupy land,

although the MST claimed this number to be even higher (the Economist 1996a).

Although the MST generally receive negative press coverage, they enjoyed

widespread popular support. Public support increased even more after TV cameras

filmed a massacre of 19 MST members who were killed by military police in April of

1996 (The Economist 1996b). Eventually, Cardoso had to give some concessions to this

popular confrontational group. In 1996, the budget for the agrarian reform was doubled

from $400,000 to $800,000 (New Scientist 1996). In April of 1997 a 2 month march by

2000 MST members converged on Brasilia where they were joined by 40,000

sympathizers (The Economist 1997). After this march Cardoso agreed to consult

regularly with the MST.





42


As the MST gains public support and Cardozo and other politicians try to seek

means to appease this group, the families that have already received land in Amazonia are

still being ignored. They have land and the implicit message is that they are now on their

own. However, if no efforts are made to assist families to stay on the land after

settlement, more are likely to fail in their homesteading attempts. If this happens, the

growth of movements like MST will only continue.

Conclusions


At the beginning of the Amazon development transition, the government did not

fully consider the impact of its actions on the region's rural families. It was only when

spontaneous colonization and conflict became a problem, that the government tried to

take control and guide the settlement process. With the realization that it could not

completely control the changing situation, the government retreated from colonization

responsibilities and focused more on eliminating conflict. With the advent of rapid

settlement and extractive reserves, the government did what it had neglected to do at the

outset in the 1960s and 1970s: provide a means for Amazonia's rural population to have

their land claims recognized.

In a sense, policies came full-circle from attempts to limit spontaneous settlement

to simply documenting it. As high rates of turnover demonstrate, colonization projects

have not led to long-term settlement, nor have they led to productive small farm

economies in most areas. Even though projects eliminated the threat of violent expulsion

for those included, they did not assure that they could survive on the land.

The democratization process shifted the focus from colonization to land reform.

Focusing the debate solely on further land distribution programs diverted attention from





43

rural support programs that would help newly settled colonists to establish their

homesteads. By not adequately assisting families already settled in colonization projects,

the government allowed progress made in land redistribution to slip away and the

problem of landlessness to be perpetuated.

For rural families, gaining land is a Pyrrhic victory if it is not accompanied by

other types of support that would allow them to establish their farms. In most cases,

resettlement in Amazonia provided access to land but little more. Resource poor farmers

attempting to establish homesteads on colonization projects were not supported and were

likely to be left to their own ends. Often they could not or chose not to stay on the land.

In some cases, official colonization projects sheltered areas from land concentration by

restricting land sales. As a result, enclaves developed where small farms persisted but

with high turnover among property owners. Do such colonization projects just delay the

inevitable land concentration or will they allow farmers to gain a toe hold and adapt to

harsh conditions as the frontier matures? Later chapters will take up this question by

examining one project in the western Amazon.













CHAPTER 3
RESETTLEMENT AND COLONIZATION IN ACRE

During the late 1960s and the early 1970s the Brazilian government's policies that

had set off Amazonia's frontier transition struck the eastern state of Acre with full force.

In a few short years the biased policies that favored large scale capitalist investment

transformed the state's land tenure regime, exacerbating a developing rural crises and

increasing expulsions of rural peoples from the land. The unrest caused by the

expulsions, the exodus to urban shantytowns, and the impending arrival in Acre of

migrants from the south, required the federal and local government to respond. The

government implemented a series of populist resettlement programs calculated to ease the

rural tension. However, there was less interest in assuring that once rural families were

settled, they received sufficient assistance to establish their homesteads. Almost all the

projects implemented during the 1980s suffered high turnover and gradual land

concentration. Rural violence and land conflict continued, and before the resettlement

projects were fully implemented, attention shifted to addressing land conflicts through

other types of projects that attempted to solve the emerging crisis.

Through political organizing, Acre's rural population eventually influenced

government land and settlement policies. Government agencies went from attempts to

dictate a solution to the crisis from above to accepting solutions advocated from below by

rural social movements. This meant the recognition of defacto land distribution carried

out by spontaneous settlers or the continuation of the status quo forposseiros living on



44





45

extractive reserves. However, while they may have gained lands, small producers in Acre

were marginalized in peripheral areas where they had little economic opportunity, and

virtually no access to the state's system of social support. The marginal situation made it

difficult for small rural producers to maintain their hold on the land.

This chapter will explore how the transition developed and how small holders

were repeatedly forced into marginal conditions. Before proceeding to this discussion

however, it is necessary to first review the conditions in Acre prior to the transition. After

all, it is the earlier economic and tenure systems that set the stage within which families

would have to absorb the brunt of change. Once this background has been laid out, the

focus of the chapter will then shift to examining how the initial development policies

ignored the rights and needs of Acre's rural population. It will become clear that, even

when families did win the right to land, it usually required the abandonment of traditional

rights. Moreover, policies did not assure that they would receive the assistance necessary

for establishing a successful homestead.

Acre Before the Development Boom



During the period before the development boom Acre was a peripheral sub-region

with a stagnant economy. Many of the trends that would later bring dramatic change to

the state, such as shifts in land rights and population, had already begun, but at this stage

they were not extreme. With the collapse of the extractive economy, the land tenure

system began to change. The problem of rural to urban migration existed but the

population shifts were largely the result of internal migration within the state.

Understanding the historical situation in Acre will help expose some the roots of later

conflict in the state.





46


In 1903, at the height of the rubber-boom in Amazonia, the area now known as

Acre was annexed by Brazil. By this time the forests of Acre were already divided into

large estates for collecting rubber. These forests were rich in naturally occurring rubber

trees (Hevea brasiliensis). On the estates, isolated families were assigned trails linking

the dispersed trees they cut for rubber. This type of forest extraction was the basis for a

thriving economy in Amazonia until the early decades of this century when rubber trees

exported to Southeast Asia began to produce (see Weinstein 1983 for a history of the

boom economy). In Southeast Asia, rubber could be grown in plantations, while in

Amazonia rubber plantations were not profitable because of pathogens occurring in that

environment (Dean 1987). When the inexpensive rubber from these plantations hit the

world market the boom economy collapsed. Amazonia's economy stagnated after the

collapse of the rubber economy and Acre suffered the same fate.


During World War II, Japan cut off the allies from the Southeast Asian rubber

supplies. In response, the United States pumped funds into Brazil as part of an agreement

to revitalize rubber production in Amazonia. Laborers were recruited in the Northeast to

be shipped off to the remote estates in Amazonia. The workers, called soldados da

boracha (rubber soldiers), brought a brief wave of immigration to Acre. However, with

the end of the war these programs ended and Acre returned to its depressed condition.


Much of Acre's population had migrated into the state drawn by the delusive

promise of wealth working on the rubber estates. With the exception of this brief

bonanza, and a brief flourish during W.W.II, Acre has spent much of its history as a

remote peripheral region with a stagnant economy. When the economy thrived, rural

families were held in debt peonage; when it did not, the impoverished families were





47

forgotten on the land. When extractive production was not generating a great deal of

wealth, rural families had two options: stay on the rubber estate (seringal) working rubber

trails largely autonomously, or move to urban areas in search of employment. Push and

pull factors for migration were there but they were not strong. Landowners wanted the

labor on the land but could not force the families to stay or supply their needs if they did.

On the other hand, urban areas were small, sleepy provincial outposts that did not offer

many employment opportunities. Rural families did not have much to gain by moving;

still many did abandon the forest.


The decline in the rubber economy changed the manner in which owners of rubber

estates (seringalistas) managed their lands (see Weinstein 1983, or Bakx 1986 for in-

depth description). The seringalistas were a small cog in the overall financing system for

rubber extraction known as the Aviamento system. In this system, foreign creditors lent

money to lenders in Amazonia (casas aviadores), which in turn offered credit and

material to the estate owners (seringalistas) farther down the system. Eventually, when

the system trickled down to individual estates, the seringalista used funds to supply dry

goods and tools at inflated prices to rubber tappers (seringuiros) in payment for their

labor. Rubber and other extracted products were passed back up through the system in

exchange for the credit. When the system functioned, the seringalistas dominated local

economic activity. Once the migrants agreed to work as rubber tappers, they were held in

debt peonage. They had to sell all their rubber to the seringalista's store where the price

paid for rubber was dictated by the estate owner and supplies were sold at inflated prices.

The rubber tappers had little hope of breaking even. Armed guards assured that tappers





48


did not leave, sell to others, or even plant subsistence crops which would divert labor

away from rubber production.


After the decline in rubber prices, the seringalistas loosened their grips over their

holdings and in some cases abandoned their estates. They did not enforce monopoly

control over markets, so rubber tappers could buy and sell to itinerant traders

(marreteiros). The tappers also could begin growing crops. This enabled them to escape

from the debt peonage that had kept them tied to the land, and to begin working the land

for themselves. If they stayed on the land, after paying their debt, they became

autonomous rubber tappers. If they had cultivated the land for a year, they had posse

rights to the land they occupied. With the stagnant regional economy, no one cared. Life

in the seringais was tough and isolated. Rubber tappers families eked out a meager

existence. They relied on subsistence farming and continued cutting rubber and

collecting Brazil nuts for cash income.

The Early Crisis and New Opportunities



Not all families chose to stay on the seringais. Some tappers left to migrate back

to the Northeast. Many others migrated to urban areas in the region attempting to find

employment and other opportunities. Although at this early stage rural to urban

migration was no comparison to recent decades, this population shift overwhelmed the

rudimentary urban infrastructure. Therefore, long before the current round of

development began, the rural-urban migration was a matter of concern for local

government.





49


Shortly after the rubber economy collapsed, Acre's government was obligated to

respond to the growth in shantytowns as people migrated to urban areas searching for

work. It responded with agricultural colonization programs around Rio Branco, the

state's capital, beginning with the first project in 1908 (called Gabino Besaouro), and two

more in 1912 and 1913 (named Deocleciano de Sousa, and Cunha Vasconcelos

respectively) (Guerra 1955). These colonies were located west of Rio Branco but little is

known about them. Evidently, small lots were distributed but little support or guidance

was given to the settlers. Other agricultural settlements sprang up spontaneously as

families settled small agricultural plots north of the city. By the 1950s, many of the lots

had been incorporated into the city or had become pasture with few occupants remaining

(Guerra 1955).


With the end of World War II, rural-urban migration began to increase. The war

had brought migrants from Brazil's northeast to Acre and had seen the seringalistas

reestablish control over the seringais. When the system collapsed, the rural population

was increasingly drawn to urban centers. This new phase of rural-urban migration

accelerated the growth of urban shantytowns. Migration was especially intense around

the departmental capital, Rio Branco, where government employment was possible. The

growth of the landless and unemployed population around the capital necessitated further

government action. As a result, in 1947, the government of Jos6 Guimard dos Santos

designated 80,000 hectares of the old Seringal Empresa north of Rio Branco for





50


colonies.1 This area was divided into three concentric zones for different types of

economic activities. It included a zonaflorestal consisting of 140 colocapoes for rubber

production, a zona agricola with 423 subsistence plots and a zona urbana with 388 lots

for urban workers. Although records are not clear, this was probably just official

recognition and demarcation of a pattern that had begun spontaneously.

Two years later, the state government began organizing additional areas devoted

specifically to small scale agriculture. The goal was to establish a colony in each of the 7

municipalities, but practically the only well organized attempts centered around Rio

Branco.2 Although these constituted the first real attempts to "fixar o homen ao solo"

("settle people on the land") in Acre, these projects were very poorly funded and in some

projects observers described the people as living in a "miserable state" (Guerra 1955:

122). The farm lots distributed in these colonies were only 12 to 24 hectares. To receive

title, the family had to plant 2 hectares of rubber on the land and conserve 25% of the

land's virgin forest (Guerra 1955). As with the other colonies, it was expected that

families would meet their own subsistence needs and sell surplus produce at urban

markets. The rubber, once it had matured, would provide an additional income source.



1 As early as 1942 the territorial governor Oscar Passos had authorized the purchase of
the remnants of Seringal Empresa, to create colonies for the migrants flooding into Rio
Branco (Guerra 1955). However, nothing was done with the land until after the war.
2 Colonies created at this time included Juarez Tdvara, Dias Martins, and Alberto T6res
in 1949, and in 1950 Daniel de Carvalho, and M.ncio Lima. Other colonies created
during this period included Sdo Francisco, Apol6nia Sales, Cecilia Parente, and Sousa
Ramos. Today most have been incorporated within Rio Branco. Those areas still on the
periphery of the city support few rural families, as much of the area has been
concentrated into small ranches. These were not the only colonies created. Near
Cruzeiro do Sul the government started the colony Vila Japiim which functioned well.
On the Placido de Castro highway the colony Guimard was created. This later developed
into the municipal capital Senador Guimard.





51


In the post war period, Acre was still an isolated peripheral region but it was

becoming increasingly integrated into the nation. A sign of this integration was a change

in the major transportation links to the rest of the country. For most of Acre's history the

primary transportation conduit to the rest of the nation had been to the east by river.

Gradually, overland connections to the south of Brazil began to integrate the area into the

nation. In the early 1950s Acre had few roads, and most of those linked the capital to the

surrounding colonies established to settle migrants. One road between Rio Branco and

the village Plicido de Castro accounted for 105 km of the state's 198 km of roads (Guerra

1954).3 Shortly after Acre became a state in 1962, the Rio Branco Placido de Castro

road went as far as Abund, Rond6nia and from there it was possible to reach Porto Velho

by boat (Nissely 1966). Finally, a land route from Porto Velho to Rio Branco arrived in

1968 (Valverde 1989: 159). These roads provided very tenuous connections and were

only open a few months each year. Nonetheless, they provided easier connection to

southern markets and set the stage for increased land values in the state, land sales, and

immigration into the state. After years of gradual change, Acre was about to shift from a

sleepy backwater to become the focus of intense investment and turmoil.

The Development Boom Comes To Acre



Initially, because the state was so remote, Acre was not affected by Amazonia's

dramatic transformation that began in the mid-1950s. However, during the 1960s things

changed. Acre became a state in 1962, and six years later it was linked to the south of

Brazil by land. When the development transition finally reached Acre in the late 1960s


3 When finished in 1952 the road was called the BR-29, but today it is known as the AC-





52


and early 1970s it brought dramatic change. Government policies in the state

simultaneously encouraged land sales, investment and immigration. The state's land

tenure regime was turned on its head and Acre's settlement landscape was transformed,

especially in the southeast portion of the state. The underlying premise of these policies

was to diversify the economy (Bakx 1986: 200). However, these policies did not have

the intended effect because speculative land purchases displaced the rural labor force, and

the displaced workers could not be absorbed into the urban labor force.

The transition began when the federal government delivered a devastating one-

two punch to Acre with the policies of BASA and SUDAM. BASA de-emphasized

extractive activities and restricted credit programs. This was a critical action because

BASA had monopoly control over commercialized rubber (Bakx 1986: 174 175). In the

late 1960s and early 1970s BASA began calling in the money owed by indebted

seringalistas. To cover their debt payments these landowners, who had been sitting on

deactivated rubber estates, were forced to sell off their holdings. While BASA was

forcing seringalistas to sell off their inactive estates, SUDAM was encouraging investors

to buy land in the state for ranching activities. Much of this new investment and land

purchases took place in the eastern half of Acre in the Purus valley, where the seringais

were larger and more in debt than in the Jurud valley. As a result, the Purus was the area

most affected by the BASA foreclosures and by the subsequent abandonment of land by

seringalistas (Bakx 1986: 193).

Government investments in transportation infrastructure made the Purus valley

more accessible. In 1968, Acre was linked to central and southern Brazil when the BR-



40.





53

364 highway reached Rio Branco from Porto Velho, Rond6nia. This highway was

originally projected to continue across the state through Cruzeiro do Sul and on to Peru,

but succeeded only in opening southeastern Acre. A second highway east of Rio Branco,

the BR 317, stretched north to Boca do Acre, Amazonas and to Xapuri and Brasileia in

the southwest. Access to ground transportation increased land values. Between 1972 to

1976 land prices along highways jumped 1000 to 2000 percent (Bakx 1986). The

highways not only allowed migrants and smaller investors to move into the state but also

linked the state to national markets, albeit tentatively.

The first investors to arrive in Acre were small and medium size ranchers who

located south and southeast of Rio Branco along the BR-364 and BR-317 roads and also

near the Bolivian border. Later, larger ranchers that entered the state were located

northwest of Rio Branco along the Iaco and Envira rivers near Taraucd. This location

was along the proposed route of the BR-364 to Cruzeiro do Sul on its way to Peru. At

these locations SUDAM initially offered fiscal incentives for areas up to 1 million ha

(Bakx 1986: 207). Land sales increased further during the administration of governor

Dantas (1970 to 1974), motivated by the state government's propaganda which

advertised: "Produce in Acre, Invest in Acre. Export via the Pacific" (Bakx 1990).

With the increasing land values and high debt, Acre's traditional land holding

elite began to sell off their abandoned seringais. Although these seringalistas had not

been actively managing their estates, the land was not unoccupied. Quite the contrary, it

was providing a livelihood to many rural families. Most of the deactivated seringais

located in the southeast, were occupied by autonomous seringueiros. Therefore, when the

new landowners took control of the land, they had to attempt to expel the rubber tappers





54


who had established rights of posse. If the posseiros would not agree to leave, the owners

began harassing them by blocking forest trails, and restricting the posseiros' ability to

market produce or purchase from marreteiros. (Cavalcanti 1983). The conflicts

escalated to threats of violence, and owners began clearing forest to destroy the rubber

tappers' livelihood. In response, spontaneous resistance developed as localized groups

tried to defend their claims to the land. As more people were expelled, the landless

traveled to urban centers, and with this rural-urban migration, the shanty towns exploded.

During these early stages, the federal and state governments did not pay attention

to the impending crisis in Acre. The state government, aware of thousands ofposseiros

living in the countryside, made no attempt to regularize their claims before inviting in

outside investors (Bakx 1986: 202). Little attention was paid to the posseiros occupying

the forest lands that were changing hands even though the claims of posseiros in Acre

were especially strong.4 In the early 1970s INCRA was preoccupied with problems in

Para and Rond6nia. In the case of Acre, INCRA initially focused on land tenure projects,

not colonization. Even though one of INCRA's primary goals was to eliminate the

latifundio/minifundio disparity, INCRA titled areas where there was no disputed

ownership, primarily a few large holdings and tiny colonies near urban centers (Bakx





4 Bakx (1986: 181) provides a good explanation of this. According to Law 601 from
1850, terras devolutas could only be titled if purchased from the state or through posse if
possessed and registered. The constitution of 1891 gave terras devolutas to the states but
Acre was only a territory until 1962. Therefore, most of Acre was terra devoluta until
that date. The only valid titles were from Peru or Bolivia. As a result, once Seringalistas
abandoned their land, the right to posse passed to the seringueiros occupying the land.
Furthermore, even if a legitimate title existed, posseiros could claim rights under the law
of usocapido, which provided a basis for land rights after ten years of occupancy of
private land.





55


1986). By focusing on less controversial areas, INCRA's land tenure project ignored the

vast majority ofposseiros who lived on disputed lands.

To make matters worse, by 1974 there was growing concern in the state

government, as well as in INCRA, that the migrating frontier passing through Rond6nia

would soon arrive in Acre. This would dramatically increase the demand for land, the

growth of urban shanty towns, and the problem of spontaneous settlement.

In 1974, a new national economic development policy, POLAMAZONIA,

explicitly switched focus away from colonization to large scale capital investment by

cutting national funding for colonization and forcing institutions like INCRA to cut

programs and personnel assigned to colonization (Bunker 1985). In the

POLAMAZONIA scheme Acre was designated a growth pole for ranching and rubber

production (Bakx 1986: 140). The proposed investment in rubber production would not

benefit traditional seringueiros. These programs envisioned a plantation system where

forests were clear cut and replaced with monoculture plantations of rubber trees.

At a time when colonization was being de-emphasized at a national level, federal

projects were just getting started in Acre. In 1974, INCRA's first tentative effort at

colonization started with the PIC Xapuri. The PIC Xapuri was a dismal failure. Only six

families were settled and they left after only 2 years (Bakx 1986: 308). INCRA did not

even list the PIC Xapuri in its 1996 list of colonization projects in Acre. Actually, this

PIC probably never got past the initial stages of implementation.

Pressure continued to build with rural crises and growing populations in Rio

Branco's slums. Within the political left and the Catholic church there was growing

concern for the poor families that had been pushed off the land. It was becoming





56

increasingly clear that something had to be done to accommodate landless families and

those who would be displaced. Because of the plight of these families, advocates began

programs to inform the rural peoples of their rights and help them organize to resist

injustice.

Allies of the Poor and the Beginning of Rural Mobilization in Acre

Acre's poor and disenfranchised were not without allies. The Catholic church's

Acre-Puris diocese where most of the conflicts were developing was run by an Italian

order (Os Servos da Maria) that followed liberation theology.5 In the early 1970s the

diocese's priests and lay workers focused primarily on urban areas, forming Christian

base communities (Communidades Eclesiais de Base, CEBs). Seeing the plight of the

rural population, the Catholic church became involved in the growing crises. In 1973,

Bishop Dom Giocondo issued A Cartilha do Seringueiro which described different types

of land rights and was distributed in rural areas (cited in Bakx 1986).

In 1976, Dom Moacyr Grecchi became the new Bishop in the Acre-Puris diocese.

Grecchi was the president of the Comissdo Pastoral da Terra (CPT) which had formed at

a national meeting of Bishops in Goais the year before. This was an important

development for the posseiros in Acre. Before continuing it would be helpful to digress

briefly to describe the institutions that were instrumental in assisting rural resistance in

the Amazon: comunidades eclesiais de base (base ecclesial communities, CEBs), and the

CPT.

Christian base communities and the Pastoral Land Commission (Comissdo

Pastoral da Terra, CPT) have played a crucial role in organizing, training and





57


supporting rural resistance in the Amazon, particularly when rural people are struggling

for land (Adriance 1994, 1995a, 1995b, also see Bemo de Almeida 1990: 226; Sousa

Martins 1990). Both institutions grew out of the activist work of members of the

Catholic church, but each was created with distinct purposes (one religious, the other

political/legal). They have converged to become players in peasant struggles in the

Amazon (Adriance 1995a), supporting possieros in their fights to defend land claims

throughout the region, from the east in the early 1970s (see Adriance 1995b), to the west

in 1980 (Bakx 1986). This assistance was not just moral or ideological. It has consisted

of organizational training and even material assistance. Both the CEBs and the CPT help

rural peoples understand their rights in a rapidly changing landscape. While the

organizers of CEBs encouraged members to join rural workers' unions, the CPT helped

document the struggle and channeled material support to activist groups. Moreover, this

support gave these conflicts a higher profile.

The CEBs are small Bible study groups made up of lay people. The CEBs were

first developed after the Brazilian Conference of Bishops' 1965 Unified Pastoral Plan

(Plano do Pastural do Conjunto) which was an attempt to follow up on the themes of the

Second Vatican Council. This document called for innovative forms of ministry, lay

leadership, and an orientation to human social development (Adriance 1995b). These

small Bible study groups would allow the church to evangelize as a response to

competition from evangelical and Pentecostal churches, even though the church suffered

from a lack of priests. Priests and nuns would train lay leaders with a new orientation to

apply the Bible to their own lives with a critical social perspective. Groups were formed


5 Acre's other diocese located in the Jurud valley was run by a German order (Espirito





58

throughout Brazil. They originated as religious lay organizations; it was only later that

they developed into rural social movements.

CEBs became vehicles for social change because their emphasis on lay ministers

provided a strong platform for the views of the poor majority. Group members were

encouraged to practice critical analysis of social reality and explicitly focused on social

problems as a response to growing competition from socialist movements (Adriance

1995a). When the posseiros were threatened by large-scale land sales, expulsions, and

violence, CEBs formed nuclei for rural resistance movements. Liberation theology played

a strong role in the ideology of some of these groups. This provided ideological support

to their cause, and faith that they would persevere. The CEBs sometimes organized the

occupation of unused land and defended claims to land after it had been occupied. As the

militancy of these groups grew, conservative Bishops attempted to censure the

movement. The church successfully pacified CEBs in urban areas and in the south.

However, in the rural north this was not as easy. Even when Bishops were opposed to the

stance of CEBs they could not exert complete control over remote areas because of

isolation from church authorities (See Adriance 1995b).

Not all CEBs became strong social movements in the north. Even though they

were grassroots movements, skilled leadership from religious and lay organizers was

crucial to this development. Furthermore, even those that did become activist were not

always successful in helping peasant members to win their struggle for land (see case

studies in Adriance 1995). Sometimes, changing leadership and turnover among





Santo) which focuses on spiritual guidance with minimal involvement of politics.





59


membership hampered the ability of a cohesive organization to develop that could

maintain itself through time.

The Pastoral Land Commission (CPT) was formed in 1975 by a group of Catholic

Bishops led by Dom Moacyr Grecchi but became an ecumenical institution with the

involvement of elements of the Lutheran and Catholic church. The CPT was an explicit

response to the land conflicts that were developing in the north. It was created to

document violence and support the weak in struggles for agrarian reform. To do this the

CPT documented and publicized rural violence (see Adriance, 1995:109 for list), provided

peasants legal support for land claims, and provided technical and material support for

farmers (Souza Martins 1990; World Watch 1991). Unlike the CEBs, the CPT was not

involved in organizing groups or land occupations; rather it assisted existing groups and

monitored rural conflicts.

In Acre, the church invited the Brazilian Confederation of Agricultural Workers

(CONTAG) to help organize the spontaneous resistance movements developing around

the CEBs into a network of rural unions (Sindicato dos Trabalhadors Rurais, STRs). The

church encouraged participation and provided its facilities to the union. In 1975, the first

STR was organized in Brasil6ia, and by 1980 STRs were formed in 7 of the 12

municipalities in Acre (Bakx 1986: 251).

In 1977, the state assembly formed a special commission to examine the land

sales taking place in Acre. Bishop Dom M. Grecchi testified that from 1969 to 1975,

80% of the state's land had changed hands. The new governor Mesquita testified that

from 1970 to 1975, 4,280,000 hectares out of the state's total area of 15,192,000 hectares

(38% of the state) had been sold to only 284 investors all from outside of the state





60

(Cavalcanti 1983: 70). These transactions took place rapidly and often without official

documentation of the transfer. Up to 1974, INCRA had issued only 81 titles in Acre

(Bakx 1990: 154). Furthermore, many land transactions were fraudulent. In 1981, Jornal

Gazeta published a Listdo de Grilagem no Acre ("Big List of Land Fraud in Acre") which

claimed that up to that point 3,600,000 hectares (24% of Acre's land area) had been the

target of land fraud, falsified titles or fictitious registration (cited in Cavalcanti 1983:

109).

Attempted Solutions Through Colonization and Resettlement

As the rural crisis in Acre grew, the state and federal governments implemented

programs to distribute land and settle families that had been displaced. These were

largely populist programs aimed at appeasing unrest and relieving tension in urban slums.

As will be seen, these programs lacked the long-term commitment to supporting those

that were already settled on the land. In 1975, Acre's state government, in a weak

attempt to appease smallholders and draw attention away from the social conflict,

increased investment in the coldnias agricolas that had been created after the war. The

colonies were renamed "Nucleos de Apoio Rural Integrado" (NARI) (Carvalho 1981).

The investment consisted of small increases in the project areas and creation of centers

within the colonies to concentrate social services. In each colony a center received

agricultural machinery, a health post, a school, and more frequent visits by technicians

and other state employees. These social programs were beneficial to colonists in these

areas. However, the underlying premise of these projects was misguided. These

programs were expected to affect the growing rural crises, but the NARIs focused on





61


people long resettled, not those on the verge of expulsion or those already living in Rio

Branco's shantytowns (Bakx 1988).

In the conflict between new landowners and smallholders occupying the land,

INCRA saw itself as a neutral arbiter. However, INCRA's policies were not favorable to

the smallholders. The policy of arbitration in conflicts focused on having the posseiros

accept alternative parcels of land, conceding their claim in return for parcels that were

often small and remote (Bakx 1986: 306). By encouraging posseiros to accept

resettlement, INCRA weakened their legal claim to land they held. As Bakx (1988) points

out, INCRA focused on a symptom rather than on the root of the problem. INCRA's

titling and settlement efforts were directed toward those smallholders who had already

been pushed from their land instead of focusing on those posseiros who were still

attempting to maintain the tenuous hold in the forests. Furthermore, these families could

not be assured of immediate land because the resettlement process was slow.

Resettlement also required families to shift from their traditional production system

because the small lots (maximum size 100 hectares) were too small to provide a

livelihood from extractive activities.

The Federal Government Attempts Colonization Solutions

In 1975, INCRA proposed the creation of Projetos de Assentamentos Dirigidos

(PAD's) in Acre. The PADs were intended to settle displaced families already living

around Rio Branco, regularize the demarcation of colonies, provide orderly occupation

and develop viable agricultural projects. In Acre, the long term aim of these projects was

not subsistence farming. Colonist families were settled on 100 hectare lots which were





62

intended for diversified agriculture. The families were expected to operate a 3 to 10 ha.

rubber plantation with 50% of the lot held as a forest reserve (Bakx 1990).

From the beginning, these projects were controversial and difficult to implement.

In 1976, the federal government tried to expropriate 700,000 hectares in Southeast Acre

to settle 8,000 families. This land was divided into two PADs: 408,000 hectares in the

Rio Branco municipality for PAD Pedro Peixoto and 292,000 hectares in the Sena

Madureira municipality for PAD Boa Esperanca (Bakx 1986: 308; 1990). Investors who

owned the expropriated land did not give up without a fight. In 1978, there was heated

debate in INCRA tribunals. Landowners opposed to expropriation called expert

witnesses who claimed that the areas were unsuitable for settlement because of infertility

and endemic malaria (local newspaper reports cited in Bakx 1986: 310). Another likely

factor for the opposition came from disputes over how much INCRA would pay

(Cavalcanti 1983: 93). As a result, by 1977 only 50,000 hectares were in INCRA's

possession, all in Boa Esperanga, where only 15 families were settled (Bakx 1990).

Some landowners successfully protected their land from expropriation. This

meant Peixoto would not cover a contiguous area because large ranches would be left

within the boundaries. Peixoto's area was further reduced by 150,000 hectares when

landless seringueiros from Rio Branco, Brasileia and Xapuri invaded the area and settled

spontaneously (Bakx 1986: 310). Peixoto's dimensions changed slightly through time as

new areas were added and settled. Today, Peixoto covers 317,588 hectares (INCRA

1996).

INCRA was not the only agency attempting to operate directed colonization

projects. In the late 1970s the state's colonization agency, COLONACRE, expropriated





63

land on the eastern edge of PAD Pedro Peixoto. This project, called Redeng o, was

modeled after Malaysian small scale rubber plantation projects (COLONACRE 1979).

Colonists in these projects were provided a small 50 -60 ha lot and each family was

expected to plant and maintain a small 3 to 10 ha rubber plantation with trees provided by

COLONACRE. Instead of each family maintaining 50 percent of their lot as forest

reserve, only 10 percent of each lot was to be held in forest reserve. A large general

reserve was set up within the project.6 The blueprint for Redencgo was based on an

elaborate grid with each square divided into pie shaped lots and a small communal area at

the center of each square. Redencgo included detailed rural and urban planning with

designated urban centers, commercial districts and public spaces. The plan for RendengAo

was strikingly similar to PIC plans used during Transamazon, and just as unrealistic.

Few of the urban features were actually implemented, and the forest reserve quickly

disappeared, divided into lots during a land rush as posseiros and Redencgo colonists

completed to stake their claims.

The resurgence of rural violence in 1980, which increased pressure for

government action, culminated with the assassination of the head of the Brasiliea STR

leader Wilson Pinheiro on July 21, 1980. A period of reprisals and assassinations

followed and tension increased (World Watch 1991).

INCRA settlement in Peixoto began officially in 1980. There were long delays

between the expropriation of land and the actual resettlement, which put added stress and

hardship on the resource poor families. By 1980 only 1,436 families had been settled in


6 This later developed into a classic "tragedy of the commons." COLONAC's weak
administration did not police the reserve. Eventually, the reserve was invaded by families





64

the original projects: 1,171 in Peixoto and 265 in Boa Esperanga. Three years later only

3,291 families had been settled in Peixoto (Bakx 1986). More land was expropriated in

1981 and 1982, totaling 187,072 hectares. This area was in three parcels that became the

Humaiti, Quixada and Santa Lucia PADs (Bakx 1986: 311). Once again, there were

difficulties implementing the projects INCRA blamed the state agencies for not fulfilling

their obligations and the STR's which were discouraging resettlement by advertising the

failure of the state consortium to keep its promises (Bakx 1986).

After 1982 there was a break in the collective resistance of rural movements. In

1980 a centralized state-level federation was installed making it difficult for activist

unions to control the federation. As a result, grassroots activism was restricted to the

municipalities of Xapuri, Brasileia, and Assis Brasil (Schwartzman 1989: 7). Later, the

Catholic church and some STRs (Rio Branco and Brasileia) adopted a reformist position

that allowed state agencies like INCRA to negotiate and resolve land conflicts, usually

through resettlement, not legitimization of existing posse. The STR in Xapuri refused to

enter in dialogue with government agencies, preferring to push ahead with its own

attempt at "agrarian reform from below" (Bakx 1986: 260). These STRs opposed

working with INCRA and the government, and advised members against agreeing to

resettlement.

A major reason settlement occurred slowly was that the process was convoluted

and complex. The selection criteria for distributing land in these projects actually

discouraged the selection of Acre's former posseiros. Many lacked personal documents

required to start the process. The criteria emphasized financial capabilities and


looking for land. This set off a massive land rush as colonists in RedenCio rushed to





65


experience with modem agricultural techniques, characteristics held by few of the ex-

rubber tappers (Bakx 1990). Therefore, INCRA looked to other regions for "suitable"

colonists.

INCRA recruited migrants from the south with slide show presentations that

depicted the projects in an overly rosy, if not fraudulent, light. Prospective settlers were

shown pictures of projects with paved roads, rural electricity, and they were promised

housing and cleared, fertile land for planting (colonist interviews 1996). Often they did

not know what to expect until they arrived on the projects. In field interviews, original

colonists told of being bussed or even flown to Acre only to be unloaded with their

possessions on the side of the road at the mouth of a forest trail. The new arrivals were

housed in tent cities and had to walk kilometers through the forest to the survey markers

that defined the front of their properties. If they did not give up immediately, the first

months consisted of back breaking labor of building a house and clearing forest for crops

(if they were settled during the season to cut and bum their plots). INCRA provided

some building materials for the first settlers and money or goods for the first months.

After that, new settlers were on their own (later owners received nothing). Sometimes,

families were settled months (or even years) before roads were constructed. Colonists

had to wait much longer for other infrastructure like schools and health posts. More often

than not, hard work and malaria took their tool and many families left.

Resettled and Abandoned

From the perspective of many smallholders in Amazonia getting a plot of land

from INCRA seemed to be a major victory. However, as they quickly learned, having the



seize plots to get a portion of what they felt was their reserve.





66

chance to own land did not mean their problems were over. Once settled, there were still

many problems facing the colonist families. While some endured the hardship to

establish successful homesteads, many more did not.

INCRA planners tried to learn from their mistakes and the PADs were supposed

to be improvements over the earlier PICs. Actually, the changes meant INCRA would

have fewer obligations to the families once they were settled. No efforts were made to

build urban centers, and the responsibility for providing basic services passed from

INCRA to other agencies. PAD projects were easier for INCRA to initiate but they were

not an improvement for colonists. The premise was that local and state agencies would

provide services to colonists once they had been settled by INCRA. This was an

unrealistic assumption. Local municipal governments varied widely in their ability and

willingness to provide rural services. State agencies were weak, underfunded and

inefficient. They were also highly politicized so that services were often doled out

selectively in return for political support from colonists during election years. INCRA

expected farmers to be supported by other agencies and when this did not occur, farmers

were left largely on their own. As a result, INCRA could not fully cut its obligations and

responsibilities in the projects.

Once the families were settled on the land, they found that the grid of property

lines had been laid out without concern for topography or access to water, thereby

repeating mistakes made earlier on the Transamazon colonization projects (Moran 1984).

The PADs were often remote and inaccessible, and INCRA was unable to open and

maintain all of the required feeder roads and bridges. By 1984, six years after the project

started, roads were far from complete in Peixoto (Bakx 1986). This lack of access made





67


it difficult for colonists to participate in the market economy, to sell crops or to purchase

necessary commodities. Furthermore, other basic social services were substandard. The

various agencies responsible for providing education and health facilities, as well as other

services, were unable to meet all of the needs of the thinly distributed population. The

level of support individual colonists received depended on their location within the

project. As a result of these obstacles, it was very difficult for many small holders to stay

on their land.

No new PADs were established after 1982, but the existing PADs continued to

expand. These projects were implemented slowly and settled in stages. As soon as new

areas, were surveyed, settlers were assigned to the lots. Only later were roads built or

schools and health posts constructed. Even though there was much hardship and

turnover, the demand for land continued to increase as migrants began arriving from

Rond6nia in the mid-1990s. Some of these migrants received land from INCRA; others

replaced those moving off the land. Migrants without land (and those who sold out)

either moved to urban shantytowns, invaded project reserves or other peripheral areas.

Democratization, the Continuing Crisis and the Climax of Social Mobilization

By the mid-1980s INCRA and other agencies were not having an easy time

dealing with Acre's smallholders and their settlement demands. Posseiros had realized

that they could defend their claims and resist efforts at relocation. The influx of new

migrants overwhelmed the capacity for orderly settlement in the PADs. For example,

according to INCRA's 1986 "5 year plan", INCRA expropriated an additional 49,000 ha

to settle 700 more families, but at that time, it was estimated that approximately 600

families were migrating into the state each dry season (Bakx 1988: 157). New settlement





68

models were used to accommodate posseiros that resisted moving and spontaneous

settlers invading available land. While these were immediately beneficial for the families

involved, this new focus distracted the attention from those who had been settled earlier.

The rubber tappers movement had learned to successfully defend their land rights

using empates, non-violent strikes in which forest residents marched to peacefully

confront workers cutting forests. In 1985, rubber tappers formed a national alliance with

other indigenous groups. Later, the forest peoples showed that they were increasingly

adept at drawing attention to their plight by forming international alliances with human

rights and environmentalist groups.

However, a major demand of the rubber tappers' movement was the formation of

extractive reserves, which would provide the tappers a collective concession rather than

dividing the area into small individual lots. This would not only protect their claim to the

land but would allow them to maintain the extractive production; that is, their main

means of support. Eventually the government gave in to pressure from the rubber tappers

and allowed reserves to be designated. From 1987 through 1991, six extractive reserves

were established in Acre. Whether or not the extractive reserves would prove to be a

long-term solution to the problems of the families living in the forests, the reserves at

least removed the threat of expulsion from these families.

Nevertheless, the problems of spontaneous settlement and land conflict did not

end with the implementation of the PADS. Therefore, INCRA once again changed its

focus. After the mid-eighties, the only new settlement projects implemented were rapid

settlement projects focused on disputed areas. The first rapid settlement project (PAR)

was started in 1986. A total of 5 were implemented before the end of the decade. These






69

projects were much smaller than the PADs. They could have as few as 24 families and

cover less than 900 hectares. This was in stark contrast to the PADs which were all

designed to accommodate more than 800 families over more than 60,000 hectares, some

substantially more.

Emancipation From Unfulfilled Promises

Change in Acre continued. In 1992, when the BR-364 was paved to Rio Branco,

it promised to make the state more accessible for migrants entering Acre. Not only were

the PARs drawing colonization resources to other areas but developments on the national

stage were also detrimental to the families that had already been settled earlier on the

PADs.

In 1990, the government of President Collar cut funds to federal agencies to fight

inflation. This measure resulted in cuts in personnel and services to most federal

institutions, and consequently increased hardship for the colonists that were already

settled. A case in point is SUCAM, the malaria control agency, one of the only

institutions that consistently reached most of the colonist households. With the cut in

personnel and lack of funds for gas and vehicle maintenance, SUCAM had to cut back on

fieldwork. This had a devastating impact as malaria was (and still is) a major problem

throughout the PADs if not controlled.

INCRA had been weakened in the late 1980s, and during Collar's program to

control inflation it virtually ceased to be a factor in rural Acre. When INCRA employees

were unable to oversee all of the projects they could not restrict land sales or land

concentration within the projects. Many professionals from nearby urban centers entered

the colonization projects as speculators and bought up the land of colonists.





70

Under President Cardoso, INCRA has regained some strength and the pace at

which PARs have been implemented has increased. In 1994 one PAR was implemented

in Acre, in 1995 three projects were started, and in 1996 seven more PARs were

implemented in Acre. INCRA personnel reported pressure from Brasilia to emphasize

settling people (instead of maximizing the number of titles issued). This appears to be a

response to the Movimento dos Trabalhadors Sem Terra (MST) which has grown in

strength. While members of the MST are less willing to move to Amazonia and are

pressuring the government for land distribution in the center and south, these new PARs

in Acre may be part of a policy to boost settlement statistics.

Today, those areas settled long ago are no longer the focus of attention and

receive only a small share of resources devoted to colonization in the state. Some

municipalities (i.e. Acrelandia) are trying to meet basic needs of the rural population;

others cannot or choose not to. State agencies continue to have problems with ineffective

agencies and programs distorted by local political agendas. A good example is provided

by a rural warehouse built by a state agency in Peixoto. In 1994, the warehouse had been

built with great fanfare. It was stocked with merchandise, a tractor, and a scale for

buying grain. Governor Magalhaes and other politicians arrived accompanied by local

news media to show their good work. When journalists and politicians left, the workers

loaded the stock and machinery onto the tractor's wagon, closed the warehouse, and left.

About a week later a truck arrived with workers to disassemble the warehouse. They said

they were going to move it to Xapuri. However, this misguided attempt to build popular

support through with bait and switch tactics backfired. The local farm association realized

what was about to happen and gathered to block the workers. The farmers were





71


successful in holding onto their warehouse but it did them little good. In 1996, the

warehouse was still sitting empty.

There are local initiatives that do seem to be helping Acre's rural families.

Recently, the municipal government in Rio Branco has begun an innovative, small-scale

land distribution and colonization project, called the PoloAgroFlorestal (Slinger 1996).

This project consists of formation of a small settlement near the capital. Colonist families

were given assistance to establish small agroforestry plots and the municipal government

provides necessary social services. This project is especially notable in the great lengths

taken by the municipal government to support the settlers and use the settlement as a

model for future development. While the project has had early success it was startlingly

similar to settlement attempts the state organized around Rio Branco in the 1950s.

Hopefully, these colonists will not have the same fate as their predecessors.

Conclusions


As they had throughout the region, government development policies brought

striking change to Acre. When the development transition arrived, it produced rapid and

dramatic change. The traditional economy with unique land use and tenure systems was

overturned as large scale capitalist investors gained the upper hand. The land sales

brought about by the policies of BASA and SUDAM caused widespread expulsions of

Acre's rural population, increasing landlessness, shantytowns and rural resistance, all of

which required a government response. The government implemented colonization

projects as an attempted solution to the rural crises. However, the projects were difficult

to establish. There was not a strong commitment to supporting families after they were

settled. Apparently, the primary goal was to lower rural tension. As a result, their projects





72


did not eliminate land disputes. With continued land conflict in other areas, government

attention focused on new problem areas. Colonists already settled were left to fend for

themselves.













CHAPTER 4
LANDOWNERSHIP TURNOVER IN PEIXOTO

In the preceding chapters, I have argued that colonization projects were primarily

populist vehicles for appeasing rural discontent. Brazil's government was determined to

facilitate capitalist investment in Acre, which continued to generate opposition from

smallholders who would be displaced. Through time, as the government attention shifted

to new centers of rural conflict, families that were already resettled were forgotten. These

developments diverted resources away from assistance programs that had been

overwhelmed from the start. As a result, those families that accepted farm lots found that

rather than the bonanza they expected, they were left in a marginal situation in a

peripheral area.

One way to gain a fuller understanding of landownership turnover is to examine a

mature colonization project and observe where turnover has occurred and how it has

affected settlement within the project. This research focuses on the PAD Pedro Peixoto,

the largest and oldest federal resettlement project in Acre. The goals of this chapter are to

describe landownership turnover in Peixoto, explain the complex factors contributing to

this phenomenon and identify the settlement pattern that has developed as a result. This

chapter combines quantitative and qualitative methods to reach these goals.

Peixoto is a natural experiment for studying the outcomes of directed resettlement

in the Amazon. As a natural experiment, there are limitations to the data that can be

gathered. In this case, I was studying change (landownership turnover) but could only



73





74


observe the aftermath of that change. Families that turned land over are gone and the

conditions that drove the process have changed. By observing current conditions, it is

possible to understand past change.

I begin this chapter by describing Peixoto and the methods and sampling design I

used to gather a representative unbiased sample of Peixoto's population. In the second

half of the chapter I will describe the state of landholdings in Peixoto. I will describe my

estimates of the level of turnover, and give explanations of why it has occurred. Finally, I

will describe the settlement pattern that resulted from turnover in Peixoto.

As I will show, there is a high incidence of landownership turnover in Peixoto.

The reasons families are not staying on the land are complex and include macro processes

outside the control of those involved. I argue that harsh environments created by

inadequate infrastructure and services contributed to this exodus from the project. This

isolation made it difficult for farm families to meet their needs. To support this

argument, I will show how turnover varies drastically from one area to another; a pattern

that reflects differences in infrastructure quality.

The difficult struggle to survive in Peixoto, compounded by a climate of

speculative investment in land, has meant that one result of ownership turnover is

increased land concentration. In some locations, many original settlers maintain control

of their land, while in others most have been replaced by incoming smallholder migrants,

or the land is consolidated into large holdings. The level of turnover and the resulting

settlement pattern indicates that the project has not given long-term land security to many

of the families resettled in Peixoto.





75

Peixoto


Peixoto is the federal government's oldest and largest directed resettlement

project in Acre. It was first proposed in response to land conflicts and the rural exodus

that plagued Acre during the early 1970s. Peixoto was a comprehensive settlement plan in

the tradition of the large colonization projects of the period. The government intended to

use Peixoto to relocate thousands of displaced people living in the shanty towns around

Rio Branco and redirect others away from the shanty towns. Peixoto was designed to

systematically settle displaced families, build up basic infrastructure (roads, bridges

services such as schools, health posts, and storage), and help families establish their

farms. The project's 317,588 hectares were originally divided into about 4,000 individual

farm plots. More than just a place for the poor to live and farm, planners hoped this

would be a long-term solution for the plight of the dispossessed.

Peixoto stretches across the state's southeastern corner, from Bolivia to the state

of Amazonas (see Figure 1.). The project dominates a highly populated part of the state;

however, when first visited, it seems oddly peripheral. While it is crossed by the state's

three major highways--the BR-364, the BR-317, and the AC-40, all of which are paved--

most families are settled back away from these roads in areas that are often remote and

inaccessible. Roads leading into the interior of the project are substandard and often

impassable.

Peixoto begins only 25 kilometers east of the capital Rio Branco, but this

proximity to urban services is deceptive because of the immense size of the project. Rio






76









Amazonas


Acre








Rio Branco 6 -Acrelad
^.^'\ I (Acrelancha


Senador a / c-
Guiomard




Placido de Castro


SBolivia


BR-317
-- ------ ---- --- Urban area

.'" Paved Highway

----' Unpaved Highway

Amazonas Peixoto



Acre -------

Pe|u
Bolivia


Figure 4.1 Southeastern Acre and the Peixoto





77

Branco is the principal urban center in the state: a hub where residents must go to carry

out official business with government agencies, access social services, and visit markets.

Peixoto may lie right outside the capital, but most residents travel much further (some

over 100 kilometers), often on substandard roads. Few residents can visit the capital, take

care of business and return home in the same day. There are other smaller urban centers

around and within the project that provide some services for the residents of Peixoto.

These include Senador Guimard, a large town that grew out of a state colonization

project from the 1950s; Placido de Castro, a village dating from the rubber boom;

Campinas, INCRA's administrative field headquarters in Peixoto, and Acrelandia, a small

town east of Peixoto that was formerly the administrative center of the defunct Redencgo

resettlement project. These smaller centers are mainly transit points to Rio Branco, the

dominant center.

A large project like Peixoto could only be implemented over time, so INCRA

settled it in stages. Unfortunately, it repeated mistakes from other projects: Peixoto was

planned in Brasilia and the network of roads and the grid of farm lots were superimposed

on the landscape ignoring topographical features, water courses and soil characteristics

of the area. As a result, roads would be hard to maintain and property lines would be

inappropriate. Many lots were left without access to water, while others were divided

several times by the same stream, as had been the case during the Transamazon project

(see Smith 1980). As INCRA teams surveyed sections of the expropriated land (glebas),

they started near highways closer to Rio Branco and the administrative centers. After the

land was surveyed, INCRA drew the names of qualifying families and began settling





78

them on these lots, often before roads were constructed. Once settled, INCRA moved on

to the next Gleba, each time moving farther into the project.

Peixoto continued to expand, even after the original area had been filled. INCRA

had left reserve areas within the project. These were places where INCRA did not

attempt settlement because they were deemed inappropriate for agriculture (due to the

presence of swamps, flood plains, poor soil) or unsuitable for road construction. Later, as

the project filled out and land was no longer available, these reserve areas were invaded

by new migrants and other landless families. Eventually, INCRA was forced to move in

and standardize the property lines in these invasions. As a result, Peixoto has continued

to grow. Today, some glebas date from the late 1970s while others were settled only in

the 1990s.

The glebas are still the administrative unit used by INCRA to run the project.

However, they do not constitute a very significant division on the ground. The principal

landmarks within the project are the primary and secondary access roads. A typical

primary road (called a ramal) has a specific name much like a city street. For example,

some are named after the construction company that built them, as in the case of the

Mendes Carlos 1 and 2 roads. Others are named after prominent people, as is the Nabor

Junior road, or they take the name of rubber estates that had been located nearby such as

the Granada and Sapucai roads. Secondary access roads, (called linhas) branch off some

of the larger primary roads. The linhas are numbered based on the distance in kilometers

from where they branched off to the start of the primary road. As the major landmark,

peoples' addresses and places are located using their position on the road (i.e. He lives at

kilometer 10 on Ramal Mendes Carlos 2).





79

Unfortunately, INCRA could survey and settle families faster than roads could be

built. In the early stages, families often were given lots that could be reached only by

long treks through the forest. It was not uncommon for the first families settled in a

gleba to wait for months or even years for a road to pass near their lot. By 1986, many

families were still enduring hardships and roads were far from complete (Bakx 1986).

Roads still vary widely in quality and maintenance. Since roads were superimposed on

the landscape without regard for topography or land forms, some are almost impossible to

keep open to vehicles. Others are never maintained because of unresponsive local

government or because residents had the misfortune of supporting a defeated political

party.

From my previous experience in Peixoto in 1992 and 1994, it was apparent that

conditions and settlement density varied from place to place in the project. Many areas

were sparsely populated, and some were clearly concentrated into large single holdings.

In other areas, family farms still predominated but many of the residents turned out to be

recent arrivals. I believe the settlement behavior was not coincidental. It is not random

chance that families stayed in some places and not in others.

I wanted to better understand the factors behind this apparently simple

phenomenon: Why families were not staying on their land. The first step to understand

landownership turnover and settlement variation is to accurately describe what is

happening. Obviously, before describing and explaining turnover, it is necessary to

gather data. Peixoto presents numerous constraints to data collection. Therefore, the next

section defines turnover, the difficulties measuring it in Peixoto and the methods I used to

overcome these constraints.





80

Research Design And Sampling Methods Appropriate For Peixoto


Whenever a landowner transfers land rights to another party, landownership

turnover takes place. Such a transfer can be through sale, trade, abandonment or even

expulsion. In theory, it should be easy to keep track of land ownership turnover, but it is

not. Quantifying the extent of landownership turnover and accurately describing the

resulting settlement is a difficult task because of conditions in the project, inadequate

record keeping and the nature of the turnover phenomenon itself. For research to be

productive, I had to overcome these problems and draw a unbiased representative sample

for the entire Peixoto project. This would allow a valid estimate of turnover in Peixoto

and allow me to go beyond discussion of individual cases.

The conditions that make life tough for the colonist farmers also make it difficult

to accurately describe the changing settlement patterns in Peixoto. Travel is difficult in

Peixoto: the project is huge, transportation is inadequate, and roads and bridges are

poorly maintained. Travel from one side of the project to the other can entail hundreds of

kilometers. It is not uncommon to find roads that can only be traveled with four wheel

drive vehicles, or roads that cannot be used because of fallen bridges. Some portions of

the project are only accessible on foot. Furthermore, because regular transport services

are inadequate, travel can be costly and time consuming. Regardless of these problems,

field work is a necessity because official records give little information about what has

happened or what can be expected on the ground.

Inaccurate documents and incomplete records compound the problems in

describing Peixoto. Once families have possession of the land it is difficult for

government institutions to monitor changes in ownership. In remote areas, like frontier





81

colonization projects, land transactions often take place without legal sanction or proper

documentation. INCRA technicians admitted they did not know who lived on many of

the lots or how often the land had changed hands. With limited personnel, vehicles and

funds, there were often long periods between INCRA field visits and it was difficult for

them to keep track of what was taking place in the field. INCRA's personnel complained

that it was not uncommon for lots to change hands several times in a single season. Not

only are poor records a problem with the field stations, but inconsistencies are found in

some of the basic documentation for the project. INCRA documents indicate that there

are 4,025 properties within the project; however, in a 1995 INCRA map of the project, I

counted only 3,960 farm lots. It is not clear where INCRA is counting the additional

properties. I know from experience that my count of farm lots exaggerates the number of

properties because in some areas land has been concentrated. For these reasons, there is a

lack of accurate statistics about what is taking place in Peixoto and no frame for easily

drawing a sample.

Because turnover may or may not be a documented process, it is difficult to know

the ownership status of a lot without visiting the area. Even then, past owners are gone,

so it is hard to know why they left or other details about past episodes. The researcher

can only talk to current residents, ask them about their experiences and what they know

of past events. Such questions would include: Were they the first family settled on the

lot? Do they know how many owners were there before them? How long have they

occupied the site? These are all measures of turnover. These measures are good

indicators of the stability of settlement patterns: they tell whether families have stayed on

the land and how long.





82

A valid description of Peixoto requires the use of a systematic random sample

where all combinations of sample units have the same probability of being drawn

(Scheafer, Mendenhall, and Ott 1990: 28). However, in a place like Peixoto, simple

random sampling would spread single sample units throughout the project. Logistically,

this would be impossible for my research. Instead, I chose a two-stage cluster sampling

method. To do this, I randomly selected areas (clusters), then within each selected cluster

I randomly selected sample units. This grouped my observations to ease data collection

without biasing my sample.

Choosing the appropriate method for drawing a sample in Peixoto was just the

first step. I then needed a sampling frame (or list) that could be used to accurately select

a sample. I was interested in farm properties and their owners. If an accurate list of

property owners existed, sampling would be a simple matter. However, there is no

accurate list available. Since INCRA maintained detailed maps of farm lots in the

project, I realized that I could use a 1995 INCRA map and sample lots as a proxy for

farm properties. Farm lots were a proxy for properties because some lots have been

concentrated into single holdings and, occasionally, single lots have been divided by

families into smaller parcels. The lots were numbered and could be located in the field

because SUCAM (Superintendencia do Combate da Maldria) painted these numbers on

the houses to identify the location for malaria control.

As a first stage of my sample design, (selection of the clusters), I decided to use

road segments. Roads are not just transportation corridors through the project. They are


1 Such samples have been done in Peixoto. Witcover and Vosti (1996) used simple
random samples to compare two Amazonian colonization projects (including Peixoto).





83

how locations are designated when traveling in the project, and they are corridors along

which social interaction among residents takes place. I could chose a ten kilometer road

segment (an area that could be covered on foot) and be reasonably sure of finding farm

lots that shared similar conditions and people that interacted to some extent and formed a

community. Conditions would be more alike among lots sharing the same road than

between lots on different roads.

I divided Peixoto into 82 sections each centered on a segment of access road. I

then randomly selected 10 of these areas as clusters (see Figure 3). This was the

maximum number I could visit and survey during the first six month field season. Within

each cluster I then selected 25 farm lots and 5 alternates. I selected the alternates as

replacements because it was likely that multiple lots from the same property would be

selected (farm lots have been concentrated into larger holdings and I was interested in

properties not lots). I wanted to get at least 25 lots from each cluster but would take 30 if

possible. When time was available, the households that occupied alternate lots were

included.

In 1995, I spent 6 months alternating week long visits to each of these clusters.

During these visits, an assistant and I visited the sampled lots to administer a

questionnaire survey. In 1996, I returned to four of the clusters for prolonged stays to

gather case study information by visiting the majority of occupied properties. These

visits enabled me to gather additional data on settlement history, community

organization, and agricultural production.




This was a multi-year study with a well-financed team of researchers. My research did
not have the financing, personnel or time.





84

The sample resulted in 223 cases where at least a minimum amount of

information could be determined.2 The 223 cases do not mean that there were 223

properties sampled. Sometimes multiple sample units appeared to be part of larger

properties with absentee owners. Even though they were likely to be one large property,

we could not tell who actually owned each area. Often, these units were within large

expanses of pasture. Because they were cleared and fenced within larger properties, we

could assume the original colonists no longer occupied or owned the lot. Although we

suspected these lots were part of larger properties, they were not grouped together

because we did not know the exact location of boundaries. Therefore, we chose to err by

counting too many cases. For example, out of 29 lots visited on one cluster (Nova

Aldeia), 26 appeared to be part of one large pasture. Each of these lots was counted

separately because we had no information about property lines.

In 146 cases, we interviewed families that owned and occupied their land.3 These

146 form a sub-sample that is used for much of the analysis. The response rate appears to







2 Out of a possible 300, we located and visited 262 sample units. However, not all of
these 262 units became individual cases included in the study. One family declined to
participate in the study. On ten lots no information about ownership or occupation status
could be gathered so these were missing. The remaining units were located in multi-lot
properties that had already been sampled and so were combined with the lot selected
earlier. Whenever we could verify that additional lots from a multi-lot property were
selected, and that these property was occupied by a colonist family, we combined the lots
and counted them as a single property.
3 Originally, we interviewed 153 families. During the course of the interviews we
learned that 8 of these families occupied lots they did not own. They were employees of
the owners. As a result, they were excluded from analysis because the surveys only
examined characteristics of landowning households.





85


Amazonas









Acre







OAcrelandia












Placido de Castro

Bolivia



[1 Peixoto B Field Sites:
A. Linha 24 F. Ramal do Bec
J Paved Highway B. Linha 8 G. Nova Aldeia 4
C. Bigode H. Triunfo
Unpaved Highway D. Iquiri I. Linha 16
E. Souza Araujo J. Linha 14



Figure 4.2 Peixoto and Selected Field Sites





86

be a very low: 146 out of 223. However, it is important to remember that my sampling

method included areas with very low population densities. Therefore, there was no one to

interview on many lots. In these cases, only a minimum amount of information was

available because the lots had absentee owners. They are officially included in Peixoto

but smallholders no longer occupy the land. If the sample had focused only on individual

households, these areas would not have been counted.

I used data from the cluster sample two ways. First, I used the complete sample

of all 223 cases where ownership could be determined to measure turnover throughout

Peixoto. Second, I used a sub-sample of 146 interviews with families that owned and

occupied the land. Because of the interviews, I have much more information on these

sites. This sub-sample provided turnover measures, allowed me to describe household

behavior, and to make comparisons within the population to examine why turnover is

taking place.

Who Are Peixoto's Current Residents?


Before analyzing turnover in Peixoto it will be helpful to describe the

characteristics of the population that occupies the project. The 146 families are a diverse

group. They come from 18 different states and primarily from rural backgrounds. For

most families, their single lot is the first land they have ever owned and it is a struggle to

maintain that land. Although conditions in the project have improved from the early

years, life on the project is still difficult. Infrastructure and services are poor throughout

the project: transportation is inadequate, schools are substandard, and health care services

are rudimentary. Much of the social infrastructure constructed during the 1980s has been

abandoned. Some INCRA schools and health posts literally have been stripped to the





87


foundation, with colonists even carrying off individual bricks. Often, where services still

exist they are provided by local municipal governments.

Origin

Peixoto's families migrated from all over the country, reflecting internal

migration within Acre, as well as immigration from other regions in the country. The

table below depicts the birth place of household heads and spouses,4 region where

household head lived the longest, and birthplace of the couples' parents. As can be seen,

the largest portion originated in South and Central Brazil. However, most household

heads have spent the longest period of their lives in Acre after a life of frequent moves.


Table 4.1 Origin of Selected Adults in Peixoto Household Survey
Acre and other regions % Acre % % % South / N
North Northeast Central
Birth place of Head and 28% 8% 16% 47% 265
Spouse
Location where head lived 49% 9% 7% 33% 146
longest
Birthplace of parents of 18% 10% 31% 41% 462
head and spouse.__ _
Note: I have divided this data into the state of Acre and three broad regional groups. The
North region includes Amapi, Amazonas, Park, Rond6nia, Roraima. The Northeast
region includes Alagoas, Bahia, Ceard, Maranhdo, Paraiba, Pernambuco, Piaui, Rio
Grande do Norte, Sergipe. The South / Central region includes Distrito Federal, Goids,
Mato Grosso do Sul, Mato Grosso, Tocantins, Espirito Santo, Minas Gerais, Parand, Rio
de Janeiro, Rio Grande do Sul, Santa Catarina, Sdo Paulo

Peixoto was initially planned to resettle Acre's displaced population. Initial INCRA

figures from 1983 indicated that 84 percent of colonists were from Acre. However, only

a couple of years later researchers were speculating that this figure was less than half that

amount (Bakx 1986). The higher percentage of southerners is a result of migrants





88

flooding into the state in the mid to late 1980s. Many of the Acreanos in my sample told

of being recruited out of the shantytowns around Rio Branco, but others told of being

evacuated by the air force from remote towns like Tarauaca for resettlement. Some of the

Southerners reported being recruited and flown or bused north by INCRA. However, it is

my impression from conversations with numerous settlers that most southerners entered

the state spontaneously. For many of these households, their arrival in Acre marked the

most recent stop in a life of continued migration. For household heads not born in Acre,

76 percent had lived in Parana in the past and 88 percent had lived in Rondonia at some

point.

The population comes from a primarily rural background. For the older

generation, 81 percent (out of 416) were born in rural areas which indicates that even if

sample members came from an urban background they are probably only one generation

removed from rural zones. Many of Peixoto's households had migrated from other rural

areas, and many who reported urban stays before Peixoto described these as brief

layovers. Furthermore, 36 percent of household heads had lived on a seringal in the past.

Life on a seringal would entail much more remote and rustic conditions than would be

found on a colonization project.

Family Composition

Peixoto has a young population; therefore many probably migrated with parents.

The average age of household heads is 43 years, but among all residents the average age

is only 24 years. Almost half (49 percent) of the population is 18 years old or less. The

average family has 5 members, and family size ranges from 1 to 15 members.


4 Head and spouse were self identified by informants. The majority of heads were male





89

Education

I expected that the quality of education and health care services would play a

significant role influencing differences in turnover rates between clusters. However,

these factors seemed to be uniformly poor throughout the project. I will use my data to

briefly describe this situation, but this survey did not gather detailed information to allow

a close examination of these factors.

Within the Peixoto project, 95 percent of the families interviewed said that a

school was available in the community. These schools are widely dispersed throughout

the project as 71 percent of the families had to travel less than 2 km to arrive at the

school. However, many of the school buildings in Peixoto have become run down,

especially those farther off main routes. These schools did function but at a basic level.

Most lacked basic teaching materials. The vast majority of schools only offered the

primary grades (89 percent up to the 4th grade). Most of the population has some

education but not much. Of the population over 15 years of age, 26.6 percent are

illiterate, 26 percent had some education but less than the 4th grade, 28.2 percent had

completed the 4th grade and only 18 percent had continued past that point.

Families that want their children to receive additional education are restricted by

the lack of schools providing the upper grades in rural areas. The lack of opportunity

means families with children that finish the primary grades must send them elsewhere to

continue their educations or do without. This was a common complaint of families that

have children at home. While sending children away to study may be difficult or

expensive, 23.3 percent of the families did have children that were studying somewhere



but not all.




Full Text

PAGE 1

LANDOWNERSHIP TURNOVER AND FAMILY FARM SURVIVAL IN AN AMAZON RESETTLEMENT PROJECT By PETER CRONKLETON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1998

PAGE 2

To my parents, Howard and Mary T. Cronkleton.

PAGE 3

ACKNOWLEDGMENTS There are many people and institutions I would like thank for the help they provided as I conducted the research described here. First of all, I would like to thank Marianne Schmink, the chair of my doctoral committee. The training, advice, and opportimity she gave me were crucial to my development as an anthropologist. The example she has set as a scholar and mentor will always be an inspiration to me. I would also like to thank the other members of my doctoral committee: Peter Hildebrand, Maxine Margolis, Gerald Murray and Tony Oliver-Smith. The training and insight I gained working with these scholars has been very important in my development as a professional. I would like to express my gratitude to my Brazilian counterparts in Acre. I especially wish to thank the numerous members of PES ACRE. Their logistical support and expert advice helped make this research possible, and their hospitality made it enjoyable. I would also like to thank the employees of EMATER who helped me with their technical expertise and broad experience in Peixoto, as well as the INCRA employees in Rio Branco, Campina, and Placido de Castro who gave me access to their records and explained Peixoto's history and operation. I am also indebted to the many families in Peixoto who took time from their daily routines to show me their farms, answer my questions, and discuss their lives. Finally, I wish to thank Rivandro da Silva Mota, my field assistant, whose hard work and friendship made long field trips possible. iii

PAGE 4

I also must thank the institutions that funded my doctoral research. Field work in 1994, 1995, and 1996 was funded by the North-South Center at the University of Miami, the University of Florida's Tropical Conservation and Development Program, and USAID. Without the support of these institutions this research would not have been possible. I would like to thank my parents, Howard and Mary T. Cronkleton, who gave me the freedom and encouragement to pursue my interests. Finally, I want to thank my wife Martha. Without her love, support, and advice I would not have finished this dissertation. IV

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS iii LIST OF TABLES viii LIST OF FIGURES ix ABSTRACT x CHAPTERS 1 RESETTLEMENT AND LANDOWNERSHIP TURNOVER ON BRAZIL'S AMAZONIAN FRONTIER 1 Frontier Transitions and Unintended Consequences in Amazonia 4 Purpose of Study 11 Outline of Dissertation 12 2 THE AMAZON' S FRONTIER TRANSITION AND THE MARGINALIZATION OF SMALLHOLDERS 16 Government Policies and the Bias Toward Large-Scale Capitalist Enterprises 17 Amazonia's Insecure Land Tenure 21 The Evolution of Amazonian Colonization Programs 25 Rural Mobilization and International Alliances 34 Democratization and the Return to the Agrarian Reform Debate 36 Conclusions 42 3 RESETTLEMENT AND COLONIZATION IN ACRE 44 Acre Before the Development Boom 45 The Early Crisis and New Opportunities 48 The Development Boom Comes To Acre 51 Allies of the Poor and the Beginning of Rural Mobilization in Acre 56 Attempted Solutions Through Colonization and Resettlement 60 The Federal Government Attempts Colonization Solutions 61 Resettled and Abandoned 65 V

PAGE 6

Democratization, the Continuing Crisis and the CUmax of Social Mobihzation 67 Emancipation From Unfulfilled Promises 69 Conclusions 71 4 LANDOWNERSHIP TURNOVER IN PEIXOTO 73 Peixoto 75 Research Design And Sampling Methods Appropriate For Peixoto 80 Who Are Peixoto 's Current Residents? 86 Origin 87 Family Composition 88 Education 89 Health Services 90 Transportation 91 Farm Properties 92 Turnover Measures 94 Landownership Turnover 96 Why Is Turnover Taking Place in Peixoto? 99 Current Settlement Patterns in Peixoto 103 Settlement Zones in Peixoto 107 Conclusions 112 5 RURAL MOBILIZATION AND FAMILY FARM SETTLEMENT SUCCESS 114 Souza Araujo's Settlement 115 Associa^ao Rural Libertador 1 1 7 Why Has the ARL Worked? 121 Why Are Associations Like ARL So Rare? 123 Appeasing Rural Resistance with Land Distribution 123 Forming Associations to Undercut Rural Mobilization 125 Case Studies in Failed Organizations 129 Extended Families and Community Cohesion 133 Conclusions 135 6 LAND USE PATTERNS AND THE IMPACT OF LANDOWNERSHIP TURNOVER 137 Colonist Farming Systems in Peixoto 138 Small Farm Attrition 140 Cash Income Alternatives 142 Typical Land Use in Peixoto 149 The Effects of Landownership Turnover on Small Farm Land Use 153 Discussion 157 Conclusions 159 vi

PAGE 7

7 CONCLUSIONS 161 APPENDIX A EQUATIONS AND NOTATION USED FOR TWO-STAGE CLUSTER SAMPLE 167 LIST OF REFERENCES 168 BIOGRAPHICAL SKETCH 175 vii

PAGE 8

LIST OF TABLES Table page Table 4.1 Origin of Selected Adults in Peixoto Household Survey 87 Table 4.2 Additional Properties Owned by Peixoto Residents 93 Table 4.3 Location of Additional Property 93 Table 4.4 Location of Earlier Property 94 Table 4.5 Number of Previous Owners 97 Table 4.6 Ownership Status of Sub-Sample of Resident Owners 98 Table 4.7 Average Settlement Time by Ownership Status 99 Table 6.1 Distribution of Farms by Ownership Category 154 Table 6.2 Mean Areas of Land Use Types by Ownership Category 155 Table 6.3 Percentages of Cleared Land by Use Type and Ownership Category 156 viii

PAGE 9

LIST OF FIGURES Figure page Figure 4.1 Southeastern Acre and the Peixoto 76 Figure 4.2 Peixoto and Selected Field Sites 85 Figure 4.3 Percentage of Original Residents in Sampled Clusters in 1995 104 Figure 6. 1 Land use on typical Peixoto farm 151 ix

PAGE 10

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 LANDOWNERSHIP TURNOVER AND FAMILY FARM SURVIVAL IN AN AMAZON RESETTLEMENT PROJECT By Peter Cronkleton May 1998 Chairman: Marianne Schmink Major Department: Anthropology In the Brazilian Amazon, attempts by the government to resettle landless and displaced families have not successfully placed families for long-term settlement. Instead, these colonization projects have been plagued by the phenomenon of landownership turnover. This means that settled families abandon, sell, or trade their land rather than staying and developing their farms. Compounding the failure to develop long-term, permanent settlement, the resources spent to assist needy families do not accomplish the desired goal, may increase environmental impacts, and actually benefit wealthy and powerful interests who can take advantage of the situation. Landownership turnover is one consequence of a complex process of change that has swept over Amazonia in recent decades. The conditions that allowed unstable settlement among smallholders to develop stem from government policies that ignore or X

PAGE 11

even undermine the land rights of the rural families. Official attempts to assist these families have usually been halfhearted and often more interested in ending unrest rather than assuring successfiil homesteading. In order to better understand how this process has developed and may continue, this research focuses on the Peixoto resettlement project in the Brazilian state of Acre. By examining a single project, the research can examine detailed variation in the behavior of residents and variation in the conditions that may drive turnover. The research focuses on intermediate range processes at the community level without excluding attention to individual, household, or regional level phenomena. Landownership turnover in Peixoto is indeed very high and within the project it has affected some communities more than others. A number of factors appear to produce the conditions that discourage long-term settlement. Inadequate infrastructure and services account for much of the difficulty faced by colonist families. Some exceptional communities have organized and overcome the hardships encountered in Peixoto, but most have not. Generally, the families living in Peixoto are able to meet their subsistence needs but it is more difficult for them to acquire cash income. One result is that colonists favor pasture formation and ranching. As pasture expands within the project, the area becomes more attractive to expanding ranches around the project, and may lead to land concentration. When colonist families hold large portions of their property in pasture it decreases the area available for subsistence production. Eventually, as this land degrades, it will be harder for later generations to support themselves on the land. If this trend continues it could force the people now living in Peixoto to seek homesteads elsewhere in Amazonia. xi

PAGE 12

CHAPTER 1 RESETTLEMENT AND LANDOWNERSHIP TURNOVER ON BRAZIL'S AMAZONIAN FRONTIER In the Brazilian Amazon, attempts by the government to resettle displaced families have not resulted in successful long-term settlement by many smallholders. Instead, these colonization projects have been plagued by landownership turnover, a phenomenon in which settled families abandon, sell, or trade their land rather than staying and developing their farms. It is not uncommon for Amazonian colonization projects to quickly lose half the residents, and sometimes up to 80 percent leave after only a few years (Millikan 1988; Martine 1990; Schneider 1995). Often, when these families leave, they are replaced by others fi-om the ranks of dispossessed people that fill Brazil's rural zones and urban slums. At other times, as lots change hands, they are snatched up by land speculators hoping to expand their holdings. As land concentration increases there are fewer places for these families to go. Land concentration in Brazil is severe. While 60 percent of the arable land is in the possession of only 2 percent of the landowners, over half of Brazil's labor force is landless (Diegues 1992). Amazonian colonization has been one of the few programs in which Brazil has attempted to alleviate this disparity. Landownership turnover in colonization projects is an important problem because the poor and displaced families that were the targets of these projects are left without land and thus, without a place to live and support themselves. As a result, the plight of impoverished families is prolonged. Furthermore, resources and time spent developing land distribution projects I

PAGE 13

2 intended to assist the rural poor fall instead into the hands of individuals and groups with greater economic and political power. The cycle of landownership turnover is not just a rural problem but also impacts the quality of life in the region's urban areas. Rural to urban migration by the region's displaced smallholders, compounded by the migration into the Amazon by the poor from other regions, has caused Amazonia's urban population to explode (Ozario de Almeida 1992; Schmink and Wood 1992; Browder and Godfrey 1997). In Acre's capital, Rio Branco, the population jumped from 34,474 to 201,432 from 1970 to 1996 (IBGE cited in Jomal Gazetta 1996). This represents a dramatic 6 percent annual growth rate. Much of this growth resulted from governmental development initiatives that encouraged investment and land sales that displaced rural families. Amazonian colonization projects are an organized attempt to resettle the large number of displaced people in Brazil. Most research on resettlement has focused on involuntary migration: situations where the push factors are greater than the pull factors, for example, the studies of the plight of people displaced by war, natural disaster, or even large-scale development projects (see Oliver-Smith and Hansen 1982, or Guggenheim and Cemea 1993). The smallholders that have migrated throughout Amazonia in recent decades include both voluntary and involuntary migrants. Some migrating families had little choice, for example those forcibly pushed off land by the hired gunmen of competing landowners, or displaced by hydro-electric projects. Other families migrated willingly, including small-scale landowners from southern Brazil who saw the opportunity to exchange their small property for a large parcel in Amazonia, or former sharecroppers looking to find land of their own. All of these diverse migrants are settled

PAGE 14

3 together on colonization projects and they all bring different levels of resources and social support networks. As a result there is a diversity of needs and abilities among this population. Colonization programs have an impact on the environment, and when turnover is prevalent in these projects it can increase the destructive consequences for the region's rainforest. Settlers clear land to provide for their livelihoods and to survive. If they suspect that their homesteading attempt will fail, they prepare to move by converting cleared land into pasture to sell to other landowners. When a family is displaced, it often moves to new locations farther along the frontier and starts the cycle over again. As the mass of landless families in the region increases, this cycle accelerates. This has led one researcher to observe. The issue is no longer how to prevent deforesting farmers from migrating to the Amazon from the rest of the country, but how to ensure that farmers already in the Amazon stay where they have already deforested, thus reducing migration (and deforestation) farther inland. (Ozorio de Almeida 1995: 5) Change in Amazonia in recent decades has been depicted as a frontier transition where peasants act as the vanguard of capitalist penetration to an isolated frontier only to be displaced with the capitalization of agriculture (Martins 1975; Foweraker 1981; Wood 1982; Sawyer 1984; Bakx 1986, 1988). This transition, sometimes called a moving frontier (Margolis 1973), has brought drastic demographic, economic, political, and environmental change to the regions it has crossed. A cynical view could see landownership turnover and concentration as a product of an elaborate plan by elite groups (wealthy investors, corporations, large landowners) to expand their holdings by using and manipulating the poor. While this conspiratorial view is attractive-elite classes have indeed manipulated the system to meet their goals, and have been the major

PAGE 15

4 beneficiaries of change in Amazonia landownership turnover and concentration are not the product of a dehberate, continuous strategy by the ehte. Instead, the frontier transition in Amazonia has been an extremely complex process, with shifting agendas, a range of competing interest groups and varied outcomes. Colonization trends in Acre must be understood within this frontier context. Frontier Transitions and Unintended Consequences in Amazonia This is a study of landownership turnover on the Peixoto colonization project in the Brazilian state of Acre. The Peixoto project was first settled in the 1970s, so by the mid-1990s when this research was conducted it should have been filled with well established family farms. However, landownership turnover has taken its toll and affected the pattern of smallholder settlement within the project. The conditions in Peixoto illustrate examples of smallholder success and failure as they struggle to adapt to change in the region. While it is one localized example, the outcomes in Peixoto are emblematic of the end results of broader trends that have transformed Amazonia in recent years. To explain a phenomenon like landownership turnover among colonist farmers it is necessary to place this behavior into a broader context that explains the causal processes at work. By describing specific conditions on the ground, this research allows detailed examination of the end product of a complex change process. This change process was described by Foweraker (1981) as an expanding pioneer fi-ontier that he broke into 3 economic stages (non-capitalist, pre-capitalist, and capitalist). In Foweraker's account, the transition is set off as a function of authoritarian capitalism. The intervention of the state is necessary to promote capitalist growth and especially to guarantee the political subordination of the peasantry. The pioneer fi-ontier expanded in

PAGE 16

5 response to the national market, and functions to provide economic accumulation within the national economy. This expansion began during the first decades of this century with growing industrialization and urbanization in central and southern Brazil; and the appearance of surplus labor due to immigration and demographic growth. Because of the monopolization of land in the Brazilian center, surplus labor was pushed to urban areas or to rural frontiers. According to this model, authoritarian capitalism is distinguished by the articulation of different modes of production (a capitalist mode and subordinate modes). This is because a dependency relation develops between the frontier and the national economy with value transferred from the periphery to the center. Economic contradictions in the center are contained by directing their effects downward to the next level. The peasants living on the frontier produce within one or more modes of production and are directly subordinate to landlords who represent another mode of production. This continued articulation allows expanded reproduction of authoritarian capitalism. In Foweraker's model, the peripheral region begins in the non-capitalist stage. In this stage, the region is remote and isolated from the nation and national markets. Those peasant producers that move into the area become reliant on extractive activities. There is an emerging petty commodity sector, but at this time there is no market for land or labor. The social relations of production are characterized by direct coercion of labor. Although the region is sparsely populated, the key is that the peasantry starts to take possession and develop the land.

PAGE 17

6 The region undergoing this transition enters Foweraker's pre-capitalist stage when migration increases and extractive activities intensify. The regular production of commodities begins, and as this happens the land's value gradually begins to increase. However, there is still little interest in investing in the land. Goods leave for markets outside of the region, but little economic value remains in the region. Also, at this stage there is no 'free' labor market. Because social controls are absent, violence characterizes the struggle for land and control of labor. The capitalist stage of the transition is reached when migration intensifies and access is established to the national economy through road construction. In this stage agriculture become the dominant activity and is increasingly capitalized. With increased agricultural production there is a regular, diverse market for goods. Eventually, the value of land increases and land markets are consolidated. This process began once peasants started to clear the forest and accelerated with the growth of agricultural production. In the capitalist stage of development, land has become an important commodity as legal title is directly convertible to capital. While peasants occupy the land in the first stages, they are eventually pushed off in the final stage. In this way the expansion that had taken place during the pre-capitalist stage created conditions for its substitution. As peasant migrants are drawn to the area, they clear and occupy land. As they produce for the market, the land increases in value. Eventually, the peasants are pushed off the land (either directly through violent coercion or because the peasants caimot compete in the increasingly capitalized market). In the long-term, the process tends towards the legally secure regime of private property but in the short-term there is a lack of legal resolution and security in land tenure.

PAGE 18

7 Foweraker's model explains the broad trends of the frontier transition process: state intervention to facilitate capital expansion, the eventual dominance of the capitalist mode in relation to the periphery, monopolization of land in the center, expulsion of surplus labor, increased land value and capitalization of production on the frontier and repeated expulsion of peasants. Although petty commodity production persists, capitalist relations of production become dominant. He recognizes that the process varies both in time and place, and it cannot account for all specific details. Foweraker states that even in the final stage there is still diversity within the frontier because, in this system the persistence of different modes is necessary to provide cheap seasonal labor for the larger scale capitalist enterprises. In fact, Foweraker states that there is no reason to suppose that the transition to capitalist production is ever complete (1981: 55). Such a broad model of frontier change is less useful when attempting to account for the wide variation in conditions found on the ground in Amazonia. With such models it is difficult to explain why some small farmers persist on the land and others move on. It also is hard to distinguish how and why the firontier transition process plays out in specific cases. Amazonia's fi-ontier transition has not produced a smooth flow to a capitalized economy. Bunker (1985) has called this lack of success of government initiatives the "underdevelopment" of the Amazon. During the fi-ontier transition, the region has seen declining productivity, unrest, and environmental destruction as a result of state intervention. Bunker argued that unequal power relations between Brazil's southern core and northern periphery underdeveloped the Amazon and caused systematically irrational results (1985: 239). The complex bureaucratic institutions the state used to promote

PAGE 19

8 capital accumulation and social welfare "impeded and distorted the implementation of development policies and compromised the state's own legitimacy, autonomy, and bureaucratic authority" (1985: 238). Bunker continues, The state's programs also intensified the ecological and demographic disruptions which had limited the capacity of Amazonian social formations to respond in sustainable and progressive ways to the changing world system's opportunities and pressures. The resulting socio-economic simplifications and destabilizations prevented the emergence of enduring or effective local organizations. The absence of effective civil organizations at the local level reduced the state's capacity to implement policy in the Amazon. Its complex bureaucracies were obligated to act in an institutional vacuum, because correspondingly complex institutional forms on which they were designed to act could not emerge in this impoverished region." (1985: 239) The federal agencies used to promote the government's development plans were imposed on a region that lacked the corresponding institutions and organizations. Therefore, federal agencies did not function properly. In fact, their inefficiency increased costs to the state, to the region's inhabitants, and allowed corruption to flourish— ultimately weakening the state's control. The competition between different combinations of capital interests and rural populations occurred not just in a regional context but nationally and internationally as well (Schmink and Wood 1992) In fact, the complexity of the fi-ontier transition in Amazonia comes into focus when examining the specific occurrences on the ground. The state's investment in the region; incentives offered for investment by other groups; and increased access to the region's land and resources set off an intense competition among interested parties. As a result, the frontier transition was not a process of linear change but rather produced a diversity of contested frontiers with highly varied outcomes (Schmink and Wood 1992: 13). The challenge to researchers in Amazonia is to explain how specific outcomes from this transition fit into the overall process.

PAGE 20

9 An outside observer might ask, why would a government start a transition that would clearly marginalize a large segment of the population? Or more importantly, why would large portions of the marginalized population continue to participate in the transition process? The answer is that during periods of broad social and environmental change, many specific outcomes are unexpected. Actors in the process exert limited control in the overall process. Through time, interactions and feedback between individual actors and between actors and the environment change the underlying conditions and trajectory of the change process. Amazonia's rural peoples have participated in the transition because they had no choice or because they expected to benefit from the process. They quickly learned that individually they exerted little influence over the changes that affected their ability to stay on the land. While many have failed in their attempts to find a productive home, not all have been so unfortunate. Sometimes, rural families mobilized and gained access to land. However, gaining land was often a Pyrrhic victory when the family could not support itself with that land. In a macro-level phenomenon, such as the planned development and the frontier transformation in the Brazilian Amazon, many specific results were not expected. In this view, a phenomenon such as landownership tiunover is an unforeseen product of complex processes driven by human behavior but not actually controlled by individuals. Sanderson (1995) uses the concept of unintended consequences to explain some events in his model of historical social evolution. According to this model, individuals acting in their own interest create social structures and systems that are the sum total of the actions of many individuals. Normally, individuals lack complete information on which to base

PAGE 21

10 their plans or actions. They do not know how conditions vary, what actions are being taken by other individuals or institutions, or what influence other actions will have. Interaction and feedback influence the change process and through time, the situation itself may change. The resulting social structures and systems come together in ways that individuals did not intend. The individuals involved do not have unlimited choice in their actions but are instead constrained by their earlier choices, and the choices of others. As the structures and systems develop, they present constraints within which individual actions must operate in the future (Sanderson 1995: 13). This view does not deny that human actors act or shape their worlds, it just emphasizes the limited impact individuals exert over broader phenomena. As Sanderson points out. Social evolution represents the long-term consequences of the dialectical interplay between human agency and social structure.. Human agency is never something that occurs "freely"; all purposive human action is constrained by both the biopsychological nature of human organisms and by the social structures that previous generations of individual actors created through their agency. Agency is therefore never to be taken as action that is "free" or "voluntary." (1995: 13) There are many factors that influence landownership turnover among smallholders in Amazonia, including among others endemic malaria, the use of inappropriate agricultural technology, weak market infrastructure, debt, and land speculation. Regardless of the proximal cause in individual incidences of turnover, these factors are all part of a larger process— a complex chain of inter-related factors that create conditions that marginalize small farms. This change process is a frontier transition in which remote, isolated areas are integrated into the nation and the national economy. Under the marginal conditions that result on the frontier, many rural families cannot (or

PAGE 22

11 choose not to) remain on the land. Normally, the individuals with the greatest wealth and power are able to take advantage of situations where smallholders vacate the land, thereby concentrating holdings and pushing out the poor. Purpose of Study This dissertation will examine landownership turnover on one colonization project in the Brazilian state of Acre. It will identify the principle causes of landownership turnover, and describe its effects on the colonization project. Earlier studies of Amazonian colonization have focused on case studies of individual families (Browder 1994) or on region wide phenomena (Smith 1980; Moran 1981; Schmink and Wood 1992; Ozario de Almeida 1992). The focus of this study fills an intermediate level that has received less attention: a single project and the communities that make up the project. This focus on an intermediate level of analysis is done without excluding the study of individual cases or ignoring links to wider regional, national or international arenas. The project studied in this research was the directed resettlement project Pedro Peixoto in Acre, Brazil. Peixoto is the largest resettlement project in the state and was intended to settle approximately 4,000 landless families. By 1995, when field work began, Peixoto should have been a thriving autonomous area of self-sufficient small farms. Instead, families continue to struggle; many have given way to ranches, and those that remain are still largely dependent on the meager government assistance that is offered. This study will describe landownership turnover in the Peixoto project. Landownership tiunover, indicated by the presence or absence of original settlers, is very high in Peixoto and varies widely from locaUon to location within the project. While, the

PAGE 23

12 forces that drive landownership turnover are complex and interrelated, key factors are the lack of adequate infrastructure and services. In particular, the quality of transportation infrastructure plays a dominant role since it determines, to a certain extent, a household's access to other necessities (agricultural inputs, markets, credit, schools and health care). This research will describe how variation in conditions within this project affects different communities within the project. The conditions driving landownership turnover affect the strategies farm families adopt to cope with the harsh conditions. This study will identify how and when families respond collectively to common problems and why, more often, they face hardship alone. In addition, this research will identify what impact landownership turnover, and the conditions that cause turnover, have on land use within the project. Outline of Dissertation In Chapter 2, 1 explain how the frontier transition that swept across Amazonia in recent decades was set off by government initiatives intended to develop the region, integrate it into the national economy, and relieve tensions in other regions caused by poverty, landlessness, and unemployment. The government's conflicting goals of stimulating large-scale capital investment and agrarian reform drew a wide range of competing interests to the region, and generated conflict. Soon the change process was out of control. One means used by the government to relieve rural tension was by resettling the disenfranchised on land in Amazonia. However, these initiatives were motivated more by the desire to alleviate tensions than to assist colonists to establish successfixl homesteads. In fact, once families were settled they were left largely on their own.

PAGE 24

13 In Chapter 3, 1 describe how the frontier transition swept across the state of Acre. In this chapter, I begin with a description of the historical setting in Acre to explain why the development transition caused such turmoil in the state. Specifically, because of a rural power vacuum, the population in some areas had developed systems of land rights and production out of the ruins of an extractive economy. However, this placed these people directly in the path of the govenmient's development plans for the state. In the 1970s, the federal government's regional agenda brought about a series of changes that drastically altered the property rights and agricultural production in the state. As families were pushed off the land, rural unrest grew, which in turn brought about government attempts to resolve the growing crisis. The government's reaction was to implement directed colonization to resettle these people. When turmoil in the countryside continued, government attention shifted to the new hot spots. As a result of this shift, the colonization projects that had already been started became lower priorities, and the families living on them were generally left on their own. In Chapter 4, 1 focus on Acre's Peixoto resettlement project to estimate turnover and identify some of the forces behind this phenomenon. I begin by describing the efforts that were required to accurately describe turnover in Peixoto. I estimate that landownership turnover has been very high in the project, and that the amount of turnover varied from location to location within the project. I argue that the observed turnover can be attributed, at least partially, to the poor transportation infrastructure and services provided to residents. As a result of this variation, a distinct settlement pattern has developed within the project. In some areas family farms have maintained their hold on the land; however, in other broad areas large land holdings are encroaching on the project

PAGE 25

14 and concentrating lands intended for smallholders. This pattern may foreshadow a difficult future for smallholder settlement in the project. In Chapter 5, 1 focus on the exceptional community at Souza Araujo with its low incidence of landownership turnover. I argue that one of the reasons residents of this community have stayed on the land is because of the activist organization formed by community members. This group formed due to a combination of factors including their settlement as a cohesive group, the presence of strong charismatic leadership, and the establishment of linkages to non-governmental organizations that helped support the community. This organization gave the residents the means to directly respond to the hardship caused by isolation, and maintain their families and farms. However, Souza Araujo also reveals an enigma: if activist organization helps families to stay on their land, why have so few communities formed activist groups? There are two principal explanations. First, the conditions in which Peixoto formed produced a social environment that was not conducive to rural mobilization. Second, the government implemented policies (whether intentional or not) that effectively undercut efforts to form grassroots organizations. This was done by establishing weak associations that the government could control. Both of these factors developed out of the government's desire to avoid the mobilization of disgruntled masses in the countryside. The result was that most colonist famiHes ended up facing the hardship of frontier life on their own, with the exception of Souza Araujo. In Chapter 6, 1 discuss agricultural land use in Peixoto and how current systems may have developed. Originally, Peixoto was expected to develop into a region of diversified, self sufficient family farms. These families would rely on perermial tree

PAGE 26

15 crops as their major income source. While farms are diversified in Peixoto they have not developed as expected. Most families can produce enough to meet household subsistence needs but it is more difficult to meet cash income needs. Cattle and pasture formation play a much more important role in the colonist farming system than other crops. Perennials play a very small role in most farm systems. The dominance of pasture in these systems may contribute to the takeover of project lands by large scale ranches. In this chapter, I show that an emphasis on pasture is more common on high turnover lots. As continued efforts are made to reach equitable development policies for the Amazon and decrease the environmental impacts these changes have brought to the region, it is important to understand landownership turnover. When smallholders turn over land in colonization projects, it means they have not been able to take advantage of the opportunities the land offered and have accepted lesser benefits to cut their losses. Furthermore, the investments that the state has made to help them have instead benefited other segments of society. In this process envirormiental disturbances occur for shortterm gain. This research brings us another step closer to understanding how and why landownership turnover occurs and what effects it has on people living in colonization projects and their environment.

PAGE 27

CHAPTER 2 THE AMAZON'S FRONTIER TRANSITION AND THE MARGINALIZATION OF SMALLHOLDERS During recent decades of rapid change in Amazonia, it has been difficult for rural families to maintain their hold on the land. How and why this has occurred is complex. To understand how this situation developed, it is necessary to briefly retrace some of the trends that affected Amazonia's rural population over the last few decades. The dramatic change began with federal govenmient intervention in the region intended to reach two general goals. First and foremost, the government wanted to develop the region and integrate it into the national economy. Second, the government wanted to use Amazonia to relieve tension in other regions due to the landlessness and unemployment that resulted from earlier development. The government's policies and incentives drew diverse groups of competing interests ranging from wealthy investors to poor migrants. As this chapter will explain, early in the transition the government began to lose control of the change taking place. Expulsions of rural people by development led to rural unrest and, along with spontaneous settlement, threatened the government's development agenda. In response, the government attempted to control settlement and resolve conflicts by creating elaborate colonization schemes. However, these colonization programs were unwieldy and ineffective. Through time, the initial grandiose plans were continuously scaled back to simplify programs due to high cost, difficulty of the endeavor, and political opposition. While intended as comprehensive solutions to 16

PAGE 28

17 general problems faced by rural people, in reality the less ambitious colonization projects amounted to little more than land distribution programs. With land in remote isolated areas lacking services, it was difficult for families to meet more than very basic subsistence needs. Consequently, many colonists either failed to support themselves or chose to sell out when given the opportunity. Government Policies and the Bias Toward Large-Scale Capitalist Enterprises The Brazilian government initiated the development boom in Amazonia with two primary goals: to increase the economic productivity of the region and to integrate it into the national economy. While colonization has received much attention in the Amazonian development literature, resettlement was not originally a priority of Brazilian development planners. The government was consistently biased towards large-scale capital investment. This bias was intensified by the economically and political powerful (the landed elite, large corporations, and wealthy investors) who were able to influence the institutions involved in development to meet their own needs. From the point of view of government planners, Amazonia offered an attractive opportunity for capital investment because of its vast mineral deposits, timber, and unused agricultural lands. The government stimulated capitalist investment in Amazonia three ways: 1) public financing and construction of infrastructure, 2) concessions of resource rights to foreign companies, and 3) tax credits and fiscal incentives to Brazilian companies (Bunker 1985). The government wanted capitahst enterprise to replace traditional forms of rainforest extraction that dominated the region and were seen as inefficient. The idea was to facilitate the entrance of corporations and investors by making participation and investment more attractive.

PAGE 29

18 The frontier transition process began to accelerate as the federal government started building roads to integrate the region and tie it to the southern states by land. The road building phase began in 1959, when SPVEA {Superintendencia do Piano de Valorizagao Economica da Amazonia) initiated a series of road building projects from the Center-south to the Northeastern, Northern and Northwestern peripheries of the country (Bunker 1985: 84). These intra-regional highways were intended to integrate isolated regions into the nation and national economy. The eastern Amazon was opened by the Belem-Brasilia highway which was completed in 1964. In the Northwest, the CuiabaAcre road was gradually extended across Mato Grosso and Rondonia. It was intended to eventually reach across Acre through the city of Cruzeiro do Sul to the Peruvian border (Nissely 1966). These roads would create ties to national industrial centers in the South and open lands for investment. The development policies encouraged the participation of diverse interest groups from corporate investors to landless migrants, all of whom entered to compete for the region's resources. As investors began buying up vast tracts of land, it required the displacement of the region's smallholders. However, the roads not only attracted wealthy entrepreneurs, but also allowed the entrance of poor migrants searching for land. That is, the roads set the stage on which conflict would play out later and foreshadowed fixture problems for the government's development plan. After the 1964 coup, the new authoritarian government intensified the transformation on the frontier. In 1966, the military government's Operation Amazonia created a number of bureaucratic institutions like BASA {Banco da Amazonia) and SUDAM {Superintendencia de Desenvolvimento da Amazonia) to grant concessions and

PAGE 30

19 administer tax credit and fiscal programs. BASA provided financial incentives for the new investors and at the same time cut credit for traditional extractive activities. SUDAM, on the other hand, administered federal incentive programs that allowed companies to reinvest taxable income in Amazonia. With its wide ranging power to distribute incentives and certify projects for participation, SUDAM could channel development in favor of capitalist enterprise. Together, these two agencies facilitated the entrance of large scale enterprises fi-om outside the region, bringing dramatic change during the 1960s and 1970s. The minimum project size to qualify for SUDAM's financial incentives was set at 25,000 hectares (Bakx 1986: 125). Initially, mining and timber extraction were the main industrial investments that qualified. In 1968, SUDAM extended its fiscal incentives to large ranching enterprises in Amazonia (Mahar 1979). Companies that had been established in Amazonia before 1974, or that had been certified to be in the region's economic interest, received a ten year tax holiday and attractive credit terms (Schmink and Wood 1992: 60). Ranching received the largest share of tax breaks and along with timber companies received most of the credits (Schmink and Wood 1992). During the early 1 970s, the conspicuous bias toward large-scale capitalist enterprise was diminished briefly. In 1970, the Medici government announced the Programa de Integragdo Nacional (PIN) which was to be the "cornerstone of efforts to modernize Amazonia" (Schmink and Wood 1992:1). A major part of the PIN was the Transamazon highway (BR-230) and the Santarem-Cuiaba highway (BR165), projects that would further open up Amazonia for development. These roads would better integrate the region into the nation and help with its defense. The road building plans

PAGE 31

20 included colonization as a facet of the overall plan. ^ The military government saw colonization in Amazonia as a means to alleviate landlessness and unemployment in the impoverished Northeast, lowering tension in that region without carrying out an actual agrarian reform. After 1973 there was a major shift in the government's Amazonian policy. Colonization had been opposed by groups that had benefited from earlier subsidies. These elite groups exerted political pressure and attempted to depict smallholders as a predatory threat to Amazonia and, conversely, ranching as benevolent to the environment (Wood and Schmink 1979). This lobbying was successftil. After only a few years the government de-emphasized colonization citing high costs and poor results in the first years. In 1974, as an alternative to the earlier program, the Geisel administration began POLAMAZONIA which set out the new regional agenda, that was openly pro-big business. With POLAMAZONIA, the government's objective was to designate growth poles that would focus public and private investment on areas deemed to have economic potential. Investments such as mining, timber extraction, and ranching qualified for the program. The efforts of government agencies did not produce the desired economic development. SUDAM was supposed to supervise the implementation of these projects but accounting was done from offices in the south making effective supervision "virtually impossible" (Schmink and Wood 1992: 60). This allowed many investors to participate 1 Sawyer (1984) suggested that colonization along the Transamazon was not just to occupy a vulnerable region but also to mask the military's motives and justify the expense with a populist land distribution program.

PAGE 32

21 as speculators or even invest their money outside the region. In 1978, critics of SUDAM used a parliamentary commission to expose the lack of success of these programs. From 1966 to 1978 SUDAM spent $5.7 billion in the region (double the state of Para's annual budget) and still Amazonia's economic position declined (Schmink and Wood 1992: 97). Furthermore, benefits were not distributed equally. In Amazonia 85% of the properties were owned by smallholders (holdings of less than 100 hectares) but only 3 small properties had been supported with SUDAM's subsidies (Schmink and Wood 1992: 97). BASA incentives from 1966 to 1979 went mostly to large enterprises based outside the region. The irony is that much of the investment activity that took place in the region was speculative and did not increase production or create jobs for those pushed off the land (Goodland 1980; Hecht 1984). Not only did these projects not produce the desired economic development, they had a devastating impact on Amazonia's rural populations. The large projects SUDAM encouraged required huge parcels of land. Most areas were occupied by rural people without documented ownership. Because of this insecure tenure, many families were displaced. Amazonia's iDsecure Land Tenure Amazonia's land tenure system presented a bottleneck for development plans from the outset. This system was a product of the earlier extractive mode of production (Bunker 1985). When Amazonia's economy was dependent on forest extraction, landholdings were characterized by poorly defined boundaries and unique systems of traditional rights. The traditional systems defined access to resources, such as rubber trees or Brazil nut groves, not to the land itself. When documents existed, they often

PAGE 33

22 covered vague areas with indistinct boundaries. Poor documentation was further confused by the existence of multiple, overlapping deeds issued by various government agencies (see Bakx 1986; Schmink and Wood 1992). The lack of consistent records, and the speed with which change was taking place led to widespread fraud. This situation was further complicated by a Brazilian law which provided legal rights to land ownership through occupation, or posse, which can be roughly translated into English as squatting. According to Brazilian law, posseiros can establish posse on untitled, unoccupied land {terra devoluta) if they live on, and cultivate the land for one year and a day. After the collapse of the rubber economy, most large landowners did not actively maintain use of their holdings and left rural people on the land on their own. Therefore, a large portion of the rural population in Amazonia could claim legal rights to the land they occupied even though they had no documentation. These posseiros presented a dilemma for the government. Development plans required these people to move for project implementation. However, the posseiros were not always willing to move, and their resistance became a thorn in the side of government planners. In 1970, in an effort to overhaul the tenure regime and facilitate the transition, the government combined weaker titling agencies to form INCRA (Instituto National de Colonizagdo e Reforma Agraria). INCRA was responsible for the administration of federal lands, land titling, and the administration of colonization projects.2 The agency was to serve as an arbiter between interests groups competing for land. Actually, much 2 INCRA's main tasks were: "1) land discrimination to separate public from private land and clarify tenure status; 2) appropriation of unclaimed federal patrimony; 3) demarcation and surveying; 4) land titling, including the regularization of past titles and occupations; 5) expropriation for social purposes and agrarian reform; 6) settlement and colonization;

PAGE 34

23 like other government agencies, INCRA favored large investors over smallholders both directly and indirectly. Like SUDAM, INCRA was a powerful and autonomous agency because it controlled public land and collected land taxes. The agency's incentive was to allocate financial and administrative resources for large-scale land sales and titling rather than for expensive and politically controversial colonization schemes (Bunker 1985: 1 10). Furthermore, large-scale investors used their economic and political power to pressure INCRA into favoring large capitalist interests in disputes over land. The pressure came from many sources. For example, banks, aware of the demand for credit in the region, pressured bureaucracies to recognize claims of large-scale capitalist enterprises so they could use the titled property as collateral for loans (Bunker 1985: 90). The military govenmient even produced secret directives that allowed INCRA to accept fi^audulent claims that undercut those of smallholders if, in INCRA' s view, the suspicious claims "promoted the development of the region" (Schmink and Wood 1992: 64). Undercutting legal procedures led to increased fi-aud. While technically the posseiros had legal protection, in reality it was difficult for them to claim their rights (Bunker 1985: 120). Isolated families found it difficult to resist the pressure placed by gunmen hired by landowners, especially when supported by local police and the unsympathetic judicial system (Americas Watch 1991). The costs of documenting a claim often exceeded the ability of rural families to pay. The process required that they travel to and stay in urban centers, often with unforeseen delays. Some services, like land surveying, were very expensive. While these costs deterred most rural 7) maintenance of a cadastre with data on rural properties; 8) enforcing and collecting the

PAGE 35

24 families, they did not inhibit large companies competing for the same land. Furthermore, wealthier enterprises were more familiar with the system and could use their power to influence the courts and police. Across the region from Para in the east to the western states of Rondonia and Acre tension in rural areas increased. The new landlords were confronted with rural populations that occupied the land. These confrontations had a variety of outcomes. Sporadic violence broke out as owners' henchmen threatened and coerced the posseiros to leave with virtual impunity. The rural poor often suffered the brunt of this violence. Overall, hundreds of posseiros were killed in land conflicts throughout the region in recent decades (CPT 1989; Americas Watch 1991). Those pushed off the land either moved to urban slums to look for employment or moved to other rural areas to start new posse, to sharecrop, or to work on a ranch or placer mine. Some groups of posseiros, aware of their plight, stiffened their resistance to expulsions. Others agreed to accept payment for "improvement" (i.e. land clearing) of the land. With few opportunities to gain land some families realized they could invade land holdings and wait to be paid off. As development efforts progressed, spontaneous settlement by migrants quickly became a major problem for government institutions. Spontaneous settlement occurs when families occupy land and begin homesteads without official recognition and minimal government involvement. This is the dominant means of rural settlement in rural Amazonia (Moran 1981, 1982). The newly created infrastructure in Amazonia allowed a massive immigration by small holders who joined the region's rural residents in claiming land through posse. Such settlement was unprecedented; millions of people rural land tax; 9) sale of public land; 10) arbitrating disputes." (Mueller et. al. 1994: 269)

PAGE 36

25 settled along the Belem-Brasilia road in less than a decade. After the initial occupation, the smallholders who had settled near the road were either bought out by the ranchers or eventually forced off the land (Sawyer 1984). All these outcomes were sources of concern for government plarmers. Conflicts increased the risk for the outbreak of wider rural unrest or uprising. The disputes delayed implementation of what were seen as beneficial economic activities, brought by the new investments. At the same time, the government was being criticized for its treatment of rural populations in Amazonia, for the growing land concentration and unemployment in other regions. The government needed to respond to these growing problems. Tension and violence often broke out over land conflicts as posseiros and spontaneous colonists attempted to defend their claim to the land. These conflicts were delaying or even jeopardizing the government capitalist development agenda. Furthermore, the government feared that this unrest could potentially become a more organized rural resistance. Finding an adequate response would be tricky. The government needed to alleviate this tension without alienating elite groups. The Evolution of Amazonian Colonization Programs The evolution of colonization programs in Amazonia must be understood to place the marginalization of rural families in the proper context. The Brazilian government responded to the growing rural unrest with colonization programs that were intended to end the crises by giving rural people land and an opportunity for livelihood. The key here is that the government took action and did not let the situation go too far. FNCRA (and other agencies) were usually one step behind problem situations and were continually adapting colonization programs to evolving situations. After the initial failure to foresee

PAGE 37

26 the needs of rural smallholders in Amazonia, the government has never been able to move ahead of this problem. Originally, resettlement was a minor issue but turmoil caused by displaced people and spontaneous settlement demanded a government response. Federal efforts began as grandiose schemes with detailed plans to meet all needs and guide the process to develop productive rural communities. The government quickly began to retreat from ambitious plans due to difficulties, costs, and growing opposition from elite interests. Programs gradually became simpler and required less long-term responsibility on the part of federal institutions. When the development transition began in Amazonia, government planners did not foresee the need for comprehensive colonization policies. It was not until the region's displaced rural people and migrant families began to create problems for development plans that the government acted. As part of the PIN and Transamazon programs, INCRA introduced a colonization model called the PIC {Projeto Integrado de Colonizagdo). Along each side of the new roads the government set aside a 100 kilometer strip to be used for small farm colonization. The PICs were comprehensive programs, where INCRA was not only responsible for survey and titling of farm lots but also the construction of basic infrastructure, as well as for organizing all services needed by the colonists. During the early years, PICs were also implemented in the western Amazon. The PIC design was intended to resolve recurring problems that had plagued settlement in tropical areas such as the lack all-season roads; inadequate credit; poor health care and educational facilities; unclear land titles; and isolation from markets (Moran 1984). The PIC plan divided project areas into a grid of 100 hectare lots along

PAGE 38

27 highways and feeder roads. It dictated colonist settlement patterns through the construction of a series of planned population centers ranging in size from rural population clusters (Agrovilas) to urban centers the (Agropolis, and Ruropolis).^ These centers would facilitate the delivery of services to the rural population. The PICs were to be implemented in stages. Eventually, after about 3 years, they would be turned over to the states once project infrastructure (roads, schools, and health posts) was in place, land was titled, and urban centers had formed (Mahar 1981: 50). From the outset the PICs were plagued with problems. While plans went to minute detail organizing the projects, the PICs were based on fundamentally flawed assimiptions. They were planned from Brasilia without the benefit of feasibility studies in the field. As a result, the projects were superimposed on the landscape ignoring topography and watersheds. Because of the rolling hills, streams, and swamps, roads were hard to build and maintain. Additionally, farm lots were highly variable in suitability for farming. The farmers' problems were compounded by their unfamiliarity with the enviromnent (Moran 1982) or the lack of effective technical advice by extension agents famihar with the environment (Wood and Schmink 1979). Additionally, INCRA and other agencies were not able to provide the necessary services in a timely manner "leaving colonists attracted to an area to fend for themselves" (Moran 1982:10). INCRA tried to select colonist families that were expected to survive, based on the amount of household labor, capital, and agricultural experience. The irony of the selection process is that it excluded many who needed assistance most. The rationale for the selection See Smith 1980, Moran 1981 for in-depth discussion on the organizational plan.

PAGE 39

28 process assumed that the colonist families settled earlier were largely responsible for failure themselves, not the project organization (Bakx 1986; Wood and Schmink 1979). A critical problem for colonists was the lack of services. Responsible institutions could not maintain roads and provide transport to get crops to market (Wood and Schmink 1979). Families were not provided adequate health care, schools, and agricultural storage. Financing for small farmers was another major problem. Colonists were caught in a catch-22: they needed title to get credit; to get title they needed to improve their land, and to do that they needed credit. Credit for aimual crops was risky for colonist farmers. It was expensive (high interest rates and high transaction costs), and if the family did not produce a good crop they fell into debt (Moran 1981). Often, families fell into a cycle of debt and were forced to seek wage labor to make payments (Collins 1985). As a result, they could not invest labor in their farms. Once wage opportunities disappeared, they were forced to sell the land. The goal of the PINs was to settle 100,000 families along the Transamazon during the first five years. In 1976 only 7,128 had been settled (CEDEPLAR 1979) and by 1978 this figure had increased only to 7,900 families (Mahar 1981: 46). In Rondonia INCRA had hoped to settle and title 23,439 families in PICs started in 1973; however, by 1978 only 6,929 had been titled (Mahar 1981: 51). It was not only hard to settle the families but also difficult for the families to stay on the land. In 1974, 18 percent of families in some areas had already abandoned their land (Martine 1980). By 1979 some areas had lost up to 44 percent of original their original colonists (Smith 1982: 170) While it was difficult to settle and maintain families on the land, the demand for space in the PICs continued to expand. For example, the Ouro Preto PIC in Rondonia

PAGE 40

29 was originally designed for 500 families but had to be expanded to accommodate 5000 families. By 1975, six more PICs had been installed in Rondonia with the intention of settling 22,700 families (Martine 1990). The PICs had been high profile projects and the government had made a point of encouraging the landless and the unemployed to move to land that was being made available. This triggered an increasing wave of migration, with migrants moving in spontaneously as roads opened up. When they could not receive land through official chaimels, they simply moved to unoccupied areas, invaded the land and waited for INCRA to regularize their holdings. Spontaneous settlement was becoming a greater problem. For example, by 1976 approximately 8 percent of the families in Rondonia were squatting on public or private lands waiting for titled lands in projects (Mueller 1980: 149). With the shift to the POLAMAZONIA program in 1974, development emphasis switched again to large-scale capitalist investment. However, the federal government could not abandon colonization efforts completely. There were too many families waiting for land and the return of big investment projects would mean even more displacement of people in the region. The federal goveniment's new agenda for colonization was to simplify projects and shift responsibility to other agencies. Up to 1975, INCRA tried exerting comprehensive control over settlement areas and until that year public services were provided almost exclusively by INCRA (Mahar 1981: 49; Martine 1990: 27). In some areas, the govenmient experimented with private colonization projects. In these schemes, the government sold large parcels of land to corporations that would develop the area and

PAGE 41

30 sell lots to colonists.4 However, the government remained the primary colonizer in Amazonia. As part of the new agenda, government colonization programs switched to directed settlement projects known as PADs {Projetos de Assentamento Dirigido). PADs were similar to PICs, but were attempts to learn from mistakes and improve the system. The major difference with the PADs was that hypothetically they would be easier and faster to implement. This was because INCRA had a more limited role in supporting the colonists once they were settled. These projects relied on greater participation by state and municipal government agencies. In the PADs, INCRA was responsible for 1) land distribution, 2) demarcation of territory, 3) administration of the project, 4) organization of settlement process, and 5) construction of the physical infrastructure. This meant that state agencies were responsible for providing education, health facilities, housing, credit, and technical services (Cavalcanti 1994). To provide this support, a wide variety of state organizations were enlisted. However, these agencies often lacked the organization, resources, or the will to adequately support colonization projects. ^ Through poor planning, inadequate fiinding, confused jurisdiction and political infighting, the crucial needs of colonists often went unfiilfilled. INCRA estimated that it would take three years to implement the PADs, then five more years for the projects to be consolidated and free from INCRA administration. 4 Private colonization projects are not within the scope of this work. These projects generally attracted wealthier colonists who tended to do better than the colonists in public projects. (See Moran 1990; Ozorio de Almeida 1992; Schmink and Wood 1992). 5 See Bunker 1985 description of the extension service's (EMATER) fiinding problems that distorted services.

PAGE 42

31 Eventually, after a total of seventeen years, INCRA expected the PADs to be fully integrated into the economy and totally "emancipated" from INCRA's assistance (Cavalcanti 1994). However, INCRA was unable to maintain these timetables. On huge projects it took longer than expected to settle families and build infrastructure. The projects were built in stages, and as each area (gleba) was finished, INCRA moved on to another. While opening new glebas, INCRA had to spend time maintaining the old, plus monitoring the colonists' progress toward earning the definitive title. It has not been easy to emancipate the projects and the colonies have had a hard time integrating into local economies. Many settled families were simply abandoned on their lots regardless of INCRA's and other agencies' best intentions. By the late 1970s immigration continued to overwhelm INCRAs attempts to settle rural families. In 1977, the government set up road-blocks in Rondonia to turn back newly arriving migrants. Eventually, INCRA had to modify the colonization process even fiarther to respond to the waves of immigrants pouring into the region. In 1979 and 1980 INCRA streamlined the process. ^ They brought in additional employees, cut down on documentation required for colonists and even provided definitive titles immediately to some qualified families (Mahar 1981). New settlement projects were to be started as quickly and efficiently as possible. INCRA began "rapid settlement programs" (PARs, Projecto de Assentimento Rdpido) which legitimized spontaneous settlement along roads or even invasions along rumored paths of future roads. At this point Rondonia' s government switched its emphasis from discouraging to encouraging immigrants to increase population to support efforts to gain statehood (Martine 1990).

PAGE 43

32 New settlement goals remained elusive. In 1982 in Rondonia, 15,000 families were to be settled in 4 projects, but 4 years later only 5000 had been settled (Martine 1990). As attempts at organized settlement proved difficult, spontaneous settlement continued at a high rate. Although not publicly abandoning the pretense of controlling the settlement process, for all practical purposes at this point INCRA's emphasis switched to simply documenting a process it could do little to stop. It was likely that large portions of the farms which INCRA claimed to have settled through "rapid settlement" or "land tenure projects" were actually the product of spontaneous occupation (Martine 1990: 29). INCRA's rapid settlement programs were responses to localized land conflicts. While the government tried to control the process and impose order, its responses to problems actually undermined its stabilization program. This is because whenever INCRA regularized an invaded area, it encouraged other invasions. Eventually these adhoc responses to land conflicts were institutionalized producing what has been called "crisis colonization" (Schmink 1982, 1988, Schmink and Wood 1992). The agency responsible for crisis colonization was GETAT-the Executive Group for Lands in the Araguaia Tocantins Basin~a special institution with the power to distribute land titles and thereby eliminate sources of unrest in that area of the eastern Amazon (Schmink 1982; Schmink and Wood 1992). GET AT was created in 1980 by the Brazilian federal government in response to the increasing violence and unrest in southern Para resulting from contested land claims and the inability of dispossessed families to acquire land. GET AT had more political and economic power than preceding institutions and more bureaucratic autonomy, which allowed it to cut through red tape

PAGE 44

33 and place families on farm lots. The focus of GETAT's activities was to settle as many families as possible as quickly as possible in contested areas. To that end, GET AT placed titling quotas on its field offices and cut the minimum size of farm lots down to 50 hectares. Titles were often distributed during dramatic public ceremonies as part of GETAT's strategy to "build grassroots political support in areas where violence and the persistent conflict over land fostered growing opposition to the military regime" (Schmink and Wood 1992: 176). GETAT's agenda had little to do with the long-term agricultural sustainability of the colonist families once they were settled. In fact, few of GETAT's activities focused on providing the services that are crucial for establishing small farms, such as credit or technical assistance. In 1981, GET AT regional personnel admitted that the fifty hectare lots they distributed were incapable of sustaining viable small farms in the long run (Schmink and Wood 1992: 178). Crisis colonization led to greater turnover and land concentration. Land consolidation was the result of two processes. First, once GET AT had titled a farm lot it became an asset that easily could be turned into cash. On these small lots, many families were in precarious economic situations. If an illness or emergency arose, the family often had little choice but to sell out. The second process resulted fi^om GETAT's policy toward ranchers. When a land invasion took place, it was often easier to offer the rancher a larger parcel elsewhere, rather than attempt to move all of the squatters off the contested land. Sometimes ranchers manipulated the system to their advantage by encouraging land invasion, knowing they would get better and larger parcels and leave the degraded land to smallholders.

PAGE 45

34 Ultimately, crisis colonization did not eliminate the types of social unrest that the military government feared most. While GET AT was able to lower tensions and resolve disputes in the short term, in the long-term these problems actually may have grown worse (Schmink and Wood 1992). GETAT's autonomy caused it to undermine the authority and operation of other institutions in the region. By siding with squatters to win their grassroots support, GET AT demonstrated the government's fear of social unrest or conflict that could lead into revolt. Squatters learned that they could get the government's attention and often win disputes if they stood their ground, organized resistance, and threatened social disruption. Consequently, the support that GET AT was able to generate for the government through its actions was short lived. Rural Mobilization and International Alliances As rural people realized the adverse effects that were brought to their livelihood and well-being by Amazonian development, they began a gradual process of rural mobilization. Initially, local groups organized to resist expulsions from land. Later rural unions, the Pastoral Land Commission (CPT) and the Catholic church began assisting these people with their organizing efforts. The rubber tapper movement provides an excellent example of how alliances can be formed successfully to collectively fight for rights. Through time, groups of rural families learned that together they could exert great influence on the change process. The rubber tapper movement began with small groups of rubber tapper families forming "empates," or strikes to halt the clearing of forests that were key to their Uvelihood. Eventually, these groups coalesced into a national movement to attempt to have a greater voice in development plans for the region (the National Council of Rubber Tappers,

PAGE 46

35 CNS). The rubber tappers began forming alliances with other groups that had corresponding agendas. In 1985, the CNS formed an alliance with indigenous people. As these groups gained higher profiles, they formed international alliances with environmental and human rights groups interested in monitoring conditions in Amazonia (Schmink and Wood 1992). The rural movements learned they could draw the attention of world press and international institutions. The international alliance between rubber tappers and environmental groups brought together organizations with political clout in lender nations, with people that had the ability to monitor projects in the field and verify that the goverrmient complied with environmental and social clauses in the loan agreements. The growing influence of the alliance between rural peoples in Amazonia and international interest groups is best illustrated in the conflict over the paving of the BR-364 from Porto Velho, Rondonia to Rio Branco, Acre. The BR-364 highway was first paved as far a Porto Velho as part of the POLONOROESTE project. This initiative began in 1982 with financing fi-om the World Bank, with the goal to pave the BR-364 highway into Mato Grosso and Rondonia and to develop sustainable farming systems for families migrating into the region. In 1985, the InterAmerican Development Bank (IDB) agreed to finance the paving of the BR-364 from Porto Velho to Rio Branco. However, under pressure fi-om the CNS and international environmentalist groups, the IDB negotiated an agreement that made the loan contingent on measures to protect the environment and indigenous peoples in the area. This agreement was called the PMACI (Projeto de Protegdo do Meio Ambiente e das Comunidades Indigenas) (Schwartzman 1989).

PAGE 47

36 Through monitoring, international environmentaUst groups with the assistance of grassroots organizations in the region were able to document that this was not being done. With this information, groups in the United States could put pressure on Congress to withhold funding for the IDB. Chico Mendes, the leader of the rubber tappers movement even traveled to Washington, D.C. to testify on behalf of Amazonian populations. In 1988 the IDB suspended fiinding for non-compliance with the PMACI, the first time ever an international project was suspended on environmental grounds (Schwartzman 1989). Eventually, in 1989 a new agreement was reached between the IDB and the Brazilian government, and the BR-364 was finally paved as far as Rio Branco, Acre, in 1992. Although the road was paved, in the end rural people were able to influence the process and gain concessions. The rubber tapper movement had been pushing for recognition of the land rights of families living in deactivated rubber estates. In 1988, they won a significant concession fi^om the government when the first extractive reserve was formed. Extractive reserves were a new type of ownership classification that recognized the usufioict rights of forest populations. This brought land disputes for some of the region's smallholders fullcircle by recognizing their posse claims and maintaining the status quo. Allowing them to stay on their land was the first step toward recognizing traditional land rights. Democratization and the Return to the Agrarian Reform Debate After the mid-1980s, Amazon colonization and resettlement began to lose its salience in the national debate. On the national stage, the battle returned to land redistribution. The democratization process brought back the possibility of the national agrarian reform debate that had ended with the 1964 coup. Much like the organized

PAGE 48

37 movements that sprang up in Amazonia to influence policy, groups of both sides of this volatile issue formed organized movements or pressure groups to influence the debate. PoUticians in the late 1980s and early 1990s struggled a balancing act between opposing camps. With the political opening under democracy, special interest groups had increased influence over government policies which had led to an "almost uncontrolled expansion of fiscal incentive programs although programs had clearly failed in previous decades" (Mueller et al. 1994: 270). As the debate over redistributing Brazil's highly concentrated land holdings intensified, families that had gained rights to the land in Amazonia were forgotten and there was little discussion of how to keep them on the land. Examining the focus of agrarian reform illustrates a basic problem with the debate: since it was almost solely concerned with landless families, it did not include programs to meet the needs of families that had already received land. Neglecting to help those families already on the land meant that settling addifional families would not produce a net improvement in land disparity. Organized groups trying to influence the agrarian reform debate came from both sides of the issue representing both the landed elite and the landless masses. A segment of Brazil's land holding minority formed the UDR {Unido Democrdtico Rural) to defend their interests even through armed force if necessary. The UDR gave landowners high profile and a loud voice in the raging agrarian reform debate, especially so during the 1988 Constitutional Convention (Tavares 1995). On the other side of the issue was the Landless Rural Workers Movement {Movimento dos Trabalhadores Rurais Sem Terra, MST). The MST formed in 1984 in the Brazilian south but soon had spread to much of the rest of the country. The MST used a confi-ontational style to force its hand in

PAGE 49

38 implementing agrarian reform. The primary tactic was to organize and carry out mass invasions and occupations of large, under-used estates. Once on the land, the MST members would build small settlements and begin farming. By the mid-1990s the MST had grown into a force to be reckoned with on the national stage, with widespread popular support. However, before analyzing the efforts of these movements, it is necessary to describe how the agrarian issue played out in the national political arena. The return of the agrarian reform debate started when Tancredo Neves was elected president in 1985. His objective was to "promote an austere economic policy so as to tame the industrial sector and urban trade unions, while at the same time undertaking a bold agrarian reform aimed at eradicating rural poverty" (Tavares 1995: 25). However, before Neves could implement his ambitious plan he died only one month into his term. Neves' vice president, Jose Samey became president. Samey was from the Northeastern region, a region with the most severe land concentration in the country. He had been included on the presidential ticket to provide balance. Samey did not have as ambitious a plan as Neves, but he could not fully ignore the agrarian reform issue. Samey formed the Ministry of Agrarian Reform (MIRAD) and placed pro-reform leaders in the Ministry of Agriculture and in the Brazilian Association of Agrarian Reform. He also proposed the first National Agrarian Reform Plan (PNRA), with the goal of expropriating 43 million hectares and eradicating mral poverty in 20 years (Tavares 1995: 25). However, the PNRA was never implemented and it did not expropriate land as intended. Not only did PNRA lack broad support within the government, but the political left was opposed to the program as long as it was supported by Samey (Tavares 1995). One of the criticisms was that it separated land policy from agricultural policy. This

PAGE 50

39 meant that benefits were focused on those without land, leaving smallholders "totally excluded" (Tavares 1995: 25). The agrarian reform debate not only failed to address the problems of colonists that had already received land in Amazonia, but the battle disrupted the remaining programs that were serving smallholders. For example, in 1987, during the debate, ENCRA was dissolved and its duties and resources were transferred to the Ministry of Agriculture. A year and a half later this decision was reversed, possibly because powerful interest groups lost economic and political rents (Mueller et al. 1995). This bureaucratic turf war caused much hardship for colonists waiting for land and title in areas where INCRA was the major institution and service provider. The organized landowners and their allies were able to paralyze the agricultural reform debate and even turn it to their advantage. When the 1988 Constitution was ratified, it banned the expropriation of "productive lands." Since then it has been difficult for INCRA to expropriate land.^ When land was acquired, it was purchased, a type of reform which avoids poUtically sensitive actions. (Mueller et al. 1995). Land purchases for resettlement also provided large landowners a convenient escape from bad investments. Sometimes large landowners encouraged invasions so they could be bought out by the government (Schmink and Wood 1992). After the initial fiery debates, movement toward agrarian reform lost momentum, usually surfacing only as ambitious campaign promises. In the 1989 presidential ^ In the 1946 Constitution the state was responsible to see that land was used for social welfare objectives, and obligated to expropriate idle lands. But article 141 required that indemnification take place in cash; this limited the government's ability to expropriate unused lands. A 1964 amendment removed the clause (Monteiro 1980; 102). However, with article 184 of the 1988 Constitution this clause was reintroduced.

PAGE 51

40 election, agrarian reform was an important but not the central issue. In the 1994 presidential election, agrarian reform was not a prominent issue but was included in the platforms of the principal contenders. Again, the programs focused on land distribution, not on providing services to rural families with land. The winner of the 1989 election, Fernando Collor de Mello made an incredible pledge: he proposed agrarian reform projects that would resettle 500,000 families (Tavares 1995). Once elected, he did not attempt to reach these amazing settlement goals. In fact, CoUor's policies caused greater hardship to those families already settled because they disrupted the agencies that were still providing services in rural areas. Collor shut down MIRAD, transferring its functions to the Ministry of Agriculture. Then, during the 1990 economic crises, Collor cut employees and funds from federal agencies as an anti-inflationary measure. Agencies like ENCRA or SUCAM (the malaria control service) were left with fewer employees, vehicles and equipment. These measures introduced a period of hardship for colonists; many left their land and land concentration accelerated. The main opposition candidate in both the 1989 and 1994 elections, Luis Inacio "Lula" da Silva tried to make agrarian reform a key point. However, his proposal emphasized land distribution, not agricultural policies and rural development. In 1994 Lula's platform was to settle 800,000 families on 30 hectare plots in 4 years. To achieve this goal, Lula proposed expropriating 24 million hectares of "unproductive land" at a cost of 8 million dollars, excluding the price of land (Tavares 1995: 27). The winner of the 1994 election, Fernando Henrique Cardoso ran on a platform that pledged to adopt agrarian policy that would settle 40,000 families in the first year;

PAGE 52

41 60,000 the second year; 80,000 the third year; and 100,000 the fourth year (Tavares 1995). However, Cardoso provided no details of how these numbers would be reached. Considering that during the early 1970s the PIN failed to meet its promise of settling 100,000 families in a decade, one might question the sincerity of these campaign promises. Once Cardoso was elected, the MST began to exert increasing pressure for land. The MST had gone from a secretive organization that carried out clandestine occupations to a high profile group. They drew international attention by staging day time operations with armies of landless workers parading down highways to occupy underused estatesall announced in advance so television cameras could witness the events (The Economist 1996a). The MST claimed that between 1986 and 1995, it had settled 139,000 families on 7.2 million hectares (Homewood 1996). In 1996, INCRA estimated that there were 22,000 families encamped along roads waiting for the opportunity to occupy land, although the MST claimed this number to be even higher (the Economist 1996a). Although the MST generally receive negative press coverage, they enjoyed widespread popular support. Public support increased even more after TV cameras filmed a massacre of 19 MST members who were killed by military police in April of 1996 (The Economist 1996b). Eventually, Cardoso had to give some concessions to this popular confrontational group. In 1996, the budget for the agrarian reform was doubled from $400,000 to $800,000 (New Scientist 1996). In April of 1997 a 2 month march by 2000 MST members converged on Brasilia where they were joined by 40,000 sympathizers (The Economist 1997). After this march Cardoso agreed to consult regularly with the MST.

PAGE 53

42 As the MST gains public support and Cardozo and other pohticians try to seek means to appease this group, the families that have already received land in Amazonia are still being ignored. They have land and the implicit message is that they are now on their own. However, if no efforts are made to assist families to stay on the land after settlement, more are likely to fail in their homesteading attempts. If this happens, the growth of movements like MST will only continue. Conclusions At the beginning of the Amazon development transition, the government did not fully consider the impact of its actions on the region's rural famihes. It was only when spontaneous colonization and conflict became a problem, that the government tried to take control and guide the settlement process. With the realization that it could not completely control the changing situation, the government retreated from colonization responsibilities and focused more on eliminating conflict. With the advent of rapid settlement and extractive reserves, the government did what it had neglected to do at the outset in the 1960s and 1970s: provide a means for Amazonia's rural population to have their land claims recognized. In a sense, policies came flill-circle from attempts to limit spontaneous settlement to simply documenting it. As high rates of turnover demonstrate, colonization projects have not led to long-term settlement, nor have they led to productive small farm economies in most areas. Even though projects eliminated the threat of violent expulsion for those included, they did not assure that they could survive on the land. The democratization process shifted the focus from colonization to land reform. Focusing the debate solely on further land distribution programs diverted attention from

PAGE 54

43 rural support programs that would help newly settled colonists to establish their homesteads. By not adequately assisting families already settled in colonization projects, the government allowed progress made in land redistribution to slip away and the problem of landlessness to be perpetuated. For rural families, gaining land is a Pyrrhic victory if it is not accompanied by other types of support that would allow them to establish their farms. In most cases, resettlement in Amazonia provided access to land but little more. Resource poor farmers attempting to establish homesteads on colonization projects were not supported and were likely to be left to their own ends. Often they could not or chose not to stay on the land. In some cases, official colonization projects sheltered areas from land concentration by restricting land sales. As a result, enclaves developed where small farms persisted but with high turnover among property owners. Do such colonization projects just delay the inevitable land concentration or will they allow farmers to gain a toe hold and adapt to harsh conditions as the frontier matures? Later chapters will take up this question by examining one project in the western Amazon.

PAGE 55

CHAPTER 3 RESETTLEMENT AND COLONIZATION IN ACRE During the late 1960s and the early 1970s the Brazilian government's policies that had set off Amazonia's frontier transition struck the eastern state of Acre with full force. In a few short years the biased policies that favored large scale capitalist investment transformed the state's land tenure regime, exacerbating a developing rural crises and increasing expulsions of rural peoples from the land. The unrest caused by the expulsions, the exodus to urban shantytowns, and the impending arrival in Acre of migrants from the south, required the federal and local government to respond. The government implemented a series of populist resettlement programs calculated to ease the rural tension. However, there was less interest in assuring that once rural families were settled, they received sufficient assistance to establish their homesteads. Almost all the projects implemented during the 1980s suffered high turnover and gradual land concentration. Rural violence and land conflict continued, and before the resettlement projects were frilly implemented, attention shifted to addressing land conflicts through other types of projects that attempted to solve the emerging crisis. Through political organizing, Acre's rural population eventually influenced government land and settlement policies. Goverrmient agencies went from attempts to dictate a solution to the crisis from above to accepting solutions advocated from below by rural social movements. This meant the recognition of de facto land distribution carried out by spontaneous settlers or the continuation of the status quo for posseiros living on 44

PAGE 56

45 extractive reserves. However, while they may have gained lands, small producers in Acre were marginalized in peripheral areas where they had little economic opportunity, and virtually no access to the state's system of social support. The marginal situation made it difficult for small rural producers to maintain their hold on the land. This chapter will explore how the transition developed and how small holders were repeatedly forced into marginal conditions. Before proceeding to this discussion however, it is necessary to first review the conditions in Acre prior to the transition. After all, it is the earlier economic and tenure systems that set the stage within which families would have to absorb the brunt of change. Once this background has been laid out, the focus of the chapter will then shift to examining how the initial development policies ignored the rights and needs of Acre's rural population. It will become clear that, even when families did win the right to land, it usually required the abandonment of traditional rights. Moreover, policies did not assure that they would receive the assistance necessary for establishing a successful homestead. Acre Before the Development Boom During the period before the development boom Acre was a peripheral sub-region with a stagnant economy. Many of the trends that would later bring dramatic change to the state, such as shifts in land rights and population, had akeady begun, but at this stage they were not extreme. With the collapse of the extractive economy, the land tenure system began to change. The problem of rural to urban migration existed but the population shifts were largely the result of internal migration within the state. Understanding the historical situation in Acre will help expose some the roots of later conflict in the state.

PAGE 57

46 In 1903, at the height of the rubber-boom in Amazonia, the area now known as Acre was annexed by Brazil. By this time the forests of Acre were already divided into large estates for collecting rubber. These forests were rich in naturally occurring rubber trees {Hevea brasiliensis). On the estates, isolated families were assigned trails linking the dispersed trees they cut for rubber. This type of forest extraction was the basis for a thriving economy in Amazonia until the early decades of this century when rubber trees exported to Southeast Asia began to produce (see Weinstein 1983 for a history of the boom economy). In Southeast Asia, rubber could be grown in plantations, while in Amazonia rubber plantations were not profitable because of pathogens occurring in that environment (Dean 1987). When the inexpensive rubber from these plantations hit the world market the boom economy collapsed. Amazonia's economy stagnated after the collapse of the rubber economy and Acre suffered the same fate. During World War II, Japan cut off the allies from the Southeast Asian rubber supplies. In response, the United States pumped ftinds into Brazil as part of an agreement to revitalize rubber production in Amazonia. Laborers were recruited in the Northeast to be shipped off to the remote estates in Amazonia. The workers, called soldados da boracha (rubber soldiers), brought a brief wave of immigration to Acre. However, with the end of the war these programs ended and Acre returned to its depressed condition. Much of Acre's population had migrated into the state dravm by the delusive promise of wealth working on the rubber estates. With the exception of this brief bonanza, and a brief flourish during W.W.II, Acre has spent much of its history as a remote peripheral region with a stagnant economy. When the economy thrived, rural families were held in debt peonage; when it did not, the impoverished families were

PAGE 58

47 forgotten on the land. When extractive production was not generating a great deal of wealth, rural families had two options: stay on the rubber estate {seringat) working rubber trails largely autonomously, or move to urban areas in search of employment. Push and pull factors for migration were there but they were not strong. Landowners wanted the labor on the land but could not force the families to stay or supply their needs if they did. On the other hand, urban areas were small, sleepy provincial outposts that did not offer many employment opportunities. Rural families did not have much to gain by moving; still many did abandon the forest. The decline in the rubber economy changed the maimer in which owners of rubber estates {seringalistas) managed their lands (see Weinstein 1983, or Bakx 1986 for indepth description). The seringalistas were a small cog in the overall financing system for rubber extraction known as the Aviamento system. In this system, foreign creditors lent money to lenders in Amazonia {casas aviadores), which in turn offered credit and material to the estate owners {seringalistas) farther down the system. Eventually, when the system trickled down to individual estates, the seringalista used funds to supply dry goods and tools at inflated prices to rubber tappers (seringuiros) in payment for their labor. Rubber and other extracted products were passed back up through the system in exchange for the credit. When the system functioned, the seringalistas dominated local economic activity. Once the migrants agreed to work as rubber tappers, they were held in debt peonage. They had to sell all their rubber to the seringalista" s store where the price paid for rubber was dictated by the estate owner and supplies were sold at inflated prices. The rubber tappers had little hope of breaking even. Armed guards assured that tappers

PAGE 59

48 did not leave, sell to others, or even plant subsistence crops which would divert labor away from rubber production. After the decline in rubber prices, the seringalistas loosened their grips over their holdings and in some cases abandoned their estates. They did not enforce monopoly control over markets, so rubber tappers could buy and sell to itinerant traders {marreteiros). The tappers also could begin growing crops. This enabled them to escape from the debt peonage that had kept them tied to the land, and to begin working the land for themselves. If they stayed on the land, after paying their debt, they became autonomous rubber tappers. If they had cultivated the land for a year, they had posse rights to the land they occupied. With the stagnant regional economy, no one cared. Life in the seringais was tough and isolated. Rubber tappers families eked out a meager existence. They relied on subsistence farming and continued cutting rubber and collecting Brazil nuts for cash income. The Early Crisis and New Opportunities Not all families chose to stay on the seringais. Some tappers left to migrate back to the Northeast. Many others migrated to urban areas in the region attempting to find employment and other opportunities. Although at this early stage rural to urban migration was no comparison to recent decades, this population shift overwhelmed the rudimentary urban infrastructure. Therefore, long before the current round of development began, the rural-urban migration was a matter of concern for local government.

PAGE 60

49 Shortly after the rubber economy collapsed, Acre's government was obligated to respond to the growth in shantytowns as people migrated to urban areas searching for work. It responded with agricultural colonization programs around Rio Branco, the state's capital, beginning with the first project in 1908 (called Gabino Besaouro), and two more in 1912 and 1913 (named Deocleciano de Sousa, and Cunha Vasconcelos respectively) (Guerra 1955). These colonies were located west of Rio Branco but little is known about them. Evidently, small lots were distributed but little support or guidance was given to the settlers. Other agricultural settlements sprang up spontaneously as families settled small agricultural plots north of the city. By the 1950s, many of the lots had been incorporated into the city or had become pasture with few occupants remaining (Guerra 1955). With the end of World War II, rural-urban migration began to increase. The war had brought migrants from Brazil's northeast to Acre and had seen the seringalistas reestablish control over the seringais. When the system collapsed, the rural population was increasingly drawn to urban centers. This new phase of rural-urban migration accelerated the growth of urban shantytowns. Migration was especially intense around the departmental capital, Rio Branco, where government employment was possible. The growth of the landless and unemployed population around the capital necessitated ftirther government action. As a result, in 1947, the government of Jose Guimard dos Santos designated 80,000 hectares of the old Seringal Empresa north of Rio Branco for

PAGE 61

50 colonies J This area was divided into three concentric zones for different types of economic activities. It included a zona florestal consisting of 140 colocagoes for rubber production, a zona agricola with 423 subsistence plots and a zona urbana with 388 lots for urban workers. Although records are not clear, this was probably just official recognition and demarcation of a pattern that had begun spontaneously. Two years later, the state government began organizing additional areas devoted specifically to small scale agriculture. The goal was to establish a colony in each of the 7 municipalities, but practically the only well organized attempts centered around Rio Branco.2 Although these constituted the first real attempts to ''fixar o homen ao solo" ("settle people on the land") in Acre, these projects were very poorly fianded and in some projects observers described the people as living in a "miserable state" (Guerra 1955: 122). The farm lots distributed in these colonies were only 12 to 24 hectares. To receive title, the family had to plant 2 hectares of rubber on the land and conserve 25% of the land's virgin forest (Guerra 1955). As with the other colonies, it was expected that families would meet their own subsistence needs and sell surplus produce at urban markets. The rubber, once it had matured, would provide an additional income source. ^ As early as 1942 the territorial governor Oscar Passos had authorized the purchase of the remnants of Seringal Empresa, to create colonies for the migrants flooding into Rio Branco (Guerra 1955). However, nothing was done with the land until after the war. 2 Colonies created at this time included Juarez Tavara, Dias Martins, and Alberto Tores in 1949, and in 1950 Daniel de Carvalho, and Mancio Lima. Other colonies created during this period included Sao Francisco, Apolonia Sales, Cecilia Parente, and Sousa Ramos. Today most have been incorporated within Rio Branco. Those areas still on the periphery of the city support few rural families, as much of the area has been concentrated into small ranches. These were not the only colonies created. Near Cruzeiro do Sul the government started the colony Vila Japiim which fianctioned well. On the Placido de Castro highway the colony Guimard was created. This later developed into the municipal capital Senador Guimard.

PAGE 62

51 In the post war period, Acre was still an isolated peripheral region but it was becoming increasingly integrated into the nation. A sign of this integration was a change in the major transportation links to the rest of the country. For most of Acre's history the primary transportation conduit to the rest of the nation had been to the east by river. Gradually, overland connections to the south of Brazil began to integrate the area into the nation. In the early 1950s Acre had few roads, and most of those linked the capital to the surrounding colonies established to settle migrants. One road between Rio Branco and the village Placido de Castro accounted for 105 km of the state's 198 km of roads (Guerra 1954).^ Shortly after Acre became a state in 1962, the Rio Branco Placido de Castro road went as far as Abuna, Rondonia and from there it was possible to reach Porto Velho by boat (Nissely 1966). Finally, a land route from Porto Velho to Rio Branco arrived in 1968 (Valverde 1989: 159). These roads provided very tenuous coimections and were only open a few months each year. Nonetheless, they provided easier connection to southern markets and set the stage for increased land values in the state, land sales, and immigration into the state. After years of gradual change. Acre was about to shift from a sleepy backwater to become the focus of intense investment and turmoil. The Development Boom Comes To Acre Initially, because the state was so remote, Acre was not affected by Amazonia's dramatic transformation that began in the mid-1950s. However, during the 1960s things changed. Acre became a state in 1962, and six years later it was linked to the south of Brazil by land. When the development transition finally reached Acre in the late 1960s 3 When finished in 1952 the road was called the BR-29, but today it is known as the AC-

PAGE 63

52 and early 1970s it brought dramatic change. Government policies in the state simultaneously encouraged land sales, investment and immigration. The state's land tenure regime was turned on its head and Acre's settlement landscape was transformed, especially in the southeast portion of the state. The underlying premise of these policies was to diversify the economy (Bakx 1986: 200). However, these policies did not have the intended effect because speculative land purchases displaced the rural labor force, and the displaced workers could not be absorbed into the urban labor force. The transition began when the federal government delivered a devastating onetwo punch to Acre with the policies of B AS A and SUDAM. RASA de-emphasized extractive activities and restricted credit programs. This was a critical action because BASA had monopoly control over commercialized rubber (Bakx 1986: 174 175). In the late 1960s and early 1970s BASA began calling in the money owed by indebted seringalistas To cover their debt payments these landowners, who had been sitting on deactivated rubber estates, were forced to sell off their holdings. While BASA was forcing seringalistas to sell off their inactive estates, SUDAM was encouraging investors to buy land in the state for ranching activities. Much of this new investment and land purchases took place in the eastern half of Acre in the Purus valley, where the seringais were larger and more in debt than in the Jurua valley. As a result, the Purus was the area most affected by the BASA foreclosures and by the subsequent abandonment of land by seringalistas 1986: 193). Government investments in transportation infrastructure made the Purus valley more accessible. In 1968, Acre was linked to central and southern Brazil when the BR40.

PAGE 64

53 364 highway reached Rio Branco from Porto Velho, Rondonia. This highway was originally projected to continue across the state through Cruzeiro do Sul and on to Peru, but succeeded only in opening southeastern Acre. A second highway east of Rio Branco, the BR 317, stretched north to Boca do Acre, Amazonas and to Xapuri and Brasileia in the southwest. Access to ground transportation increased land values. Between 1972 to 1976 land prices along highways jumped 1000 to 2000 percent (Bakx 1986). The highways not only allowed migrants and smaller investors to move into the state but also linked the state to national markets, albeit tentatively. The first investors to arrive in Acre were small and medium size ranchers who located south and southeast of Rio Branco along the BR-364 and BR-3 1 7 roads and also near the Bolivian border. Later, larger ranchers that entered the state were located northwest of Rio Branco along the laco and Envira rivers near Tarauca. This location was along the proposed route of the BR-364 to Cruzeiro do Sul on its way to Peru. At these locations SUDAM initially offered fiscal incentives for areas up to 1 million ha (Bakx 1986: 207). Land sales increased further during the administration of governor Dantas (1970 to 1974), motivated by the state government's propaganda which advertised: "Produce in Acre, Invest in Acre. Export via the Pacific" (Bakx 1990). With the increasing land values and high debt. Acre's traditional land holding elite began to sell off their abandoned seringais. Although these seringalistas had not been actively managing their estates, the land was not unoccupied. Quite the contrary, it was providing a livelihood to many rural families. Most of the deactivated seringais located in the southeast, were occupied by autonomous seringueiros. Therefore, when the new landowners took control of the land, they had to attempt to expel the rubber tappers

PAGE 65

54 who had established rights of posse. If the posseiros would not agree to leave, the owners began harassing them by blocking forest trails, and restricting the posseiros' ability to market produce or purchase from /warreto>o5. (Cavalcanti 1983). The conflicts escalated to threats of violence, and owners began clearing forest to destroy the rubber tappers' livelihood. In response, spontaneous resistance developed as localized groups tried to defend their claims to the land. As more people were expelled, the landless traveled to urban centers, and with this rural-urban migration, the shanty towns exploded. During these early stages, the federal and state governments did not pay attention to the impending crisis in Acre. The state government, aware of thousands of posseiros living in the countryside, made no attempt to regularize their claims before inviting in outside investors (Bakx 1986: 202). Little attention was paid to the posseiros occupying the forest lands that were changing hands even though the claims of posseiros in Acre were especially strong.^ In the early 1970s INCRA was preoccupied with problems in Para and Rondonia. In the case of Acre, INCRA initially focused on land tenure projects, not colonization. Even though one of INCRA 's primary goals was to eliminate the latifundio/minifundio disparity, INCRA titled areas where there was no disputed ownership, primarily a few large holdings and tiny colonies near urban centers (Bakx 4 Bakx (1986: 181) provides a good explanation of this. According to Law 601 from 1850, terras devolutas could only be titled if purchased from the state or through posse if possessed and registered. The constitution of 1891 gave terras devolutas to the states but Acre was only a territory until 1962. Therefore, most of Acre was terra devoluta until that date. The only valid titles were from Peru or Bolivia. As a result, once Seringalistas abandoned their land, the right to posse passed to the seringueiros occupying the land. Furthermore, even if a legitimate title existed, posseiros could claim rights under the law ofusocapiao, which provided a basis for land rights after ten years of occupancy of private land.

PAGE 66

55 1986). By focusing on less controversial areas, INCRA's land tenure project ignored the vast majority of posseiros who lived on disputed lands. To make matters worse, by 1974 there was growing concern in the state government, as well as in INCRA, that the migrating frontier passing through Rondonia would soon arrive in Acre. This would dramatically increase the demand for land, the growth of urban shanty towns, and the problem of spontaneous settlement. In 1974, a new national economic development policy, POLAMAZONIA, explicitly switched focus away from colonization to large scale capital investment by cutting national funding for colonization and forcing institutions like INCRA to cut programs and personnel assigned to colonization (Bunker 1985). In the POLAMAZONIA scheme Acre was designated a growth pole for ranching and rubber production (Bakx 1986: 140). The proposed investment in rubber production would not benefit traditional seringueiros These programs envisioned a plantation system where forests were clear cut and replaced with monoculture plantations of rubber trees. At a time when colonization was being de-emphasized at a national level, federal projects were just getting started in Acre. In 1974, INCRA's first tentative effort at colonization started with the PIC Xapuri. The PIC Xapuri was a dismal failure. Only six families were settled and they left after only 2 years (Bakx 1986: 308). INCRA did not even list the PIC Xapuri in its 1996 list of colonization projects in Acre. Actually, this PIC probably never got past the initial stages of implementation. Pressure continued to build with rural crises and growing populations in Rio Branco's slums. Within the political left and the Catholic church there was growing concern for the poor families that had been pushed off the land. It was becoming

PAGE 67

56 increasingly clear that something had to be done to accommodate landless families and those who would be displaced. Because of the plight of these families, advocates began programs to inform the rural peoples of their rights and help them organize to resist injustice. Allies of the Poor and the Beginning of Rural Mobilization in Acre Acre's poor and disenfranchised were not without allies. The Catholic church's Acre-Punis diocese where most of the conflicts were developing was run by an Italian order {Os Servos da Maria) that followed liberation theology.^ In the early 1970s the diocese's priests and lay workers focused primarily on urban areas, forming Christian base communities {Communidades Eclesiais de Base, CEBs). Seeing the plight of the rural population, the Catholic church became involved in the growing crises. In 1973, Bishop Dom Giocondo issued A Cartilha do Seringueiro which described different types of land rights and was distributed in rural areas (cited in Bakx 1986). In 1976, Dom Moacyr Grecchi became the new Bishop in the Acre-Punis diocese. Gracchi was the president of the Comissdo Pastoral da Terra (CPT) which had formed at a national meeting of Bishops in Goals the year before. This was an important development for the posseiros in Acre. Before continuing it would be helpful to digress briefly to describe the institutions that were instrumental in assisting rural resistance in the Amazon: comunidades eclesiais de base (base ecclesial communities, CEBs), and the CPT. Christian base communities and the Pastoral Land Commission {Comissdo Pastoral da Terra, CPT) have played a crucial role in organizing, training and

PAGE 68

57 supporting rural resistance in the Amazon, particularly when rural people are struggling for land (Adriance 1994, 1995a, 1995b, also see Bemo de Almeida 1990: 226; Sousa Martins 1990). Both institutions grew out of the activist work of members of the Catholic church, but each was created with distinct purposes (one religious, the other political/legal). They have converged to become players in peasant struggles in the Amazon (Adriance 1995a), supporting possieros in their fights to defend land claims throughout the region, from the east in the early 1970s (see Adriance 1995b), to the west in 1980 (Bakx 1986). This assistance was not just moral or ideological. It has consisted of organizational training and even material assistance. Both the CEBs and the CPT help rural peoples understand their rights in a rapidly changing landscape. While the organizers of CEBs encouraged members to join rural workers' unions, the CPT helped document the struggle and channeled material support to activist groups. Moreover, this support gave these conflicts a higher profile. The CEBs are small Bible study groups made up of lay people. The CEBs were first developed after the Brazilian Conference of Bishops' 1965 Unified Pastoral Plan {Piano do Pastural do Conjunto) which was an attempt to follow up on the themes of the Second Vatican Coimcil. This document called for innovative forms of ministry, lay leadership, and an orientation to human social development (Adriance 1995b). These small Bible study groups would allow the church to evangelize as a response to competition from evangelical and Pentecostal churches, even though the church suffered from a lack of priests. Priests and nuns would train lay leaders with a new orientation to apply the Bible to their own lives with a critical social perspective. Groups were formed Acre's other diocese located in the Jurua valley was run by a German order {Espirito

PAGE 69

58 throughout Brazil. They originated as rehgious lay organizations; it was only later that they developed into rural social movements. CEBs became vehicles for social change because their emphasis on lay ministers provided a strong platform for the views of the poor majority. Group members were encouraged to practice critical analysis of social reality and explicitly focused on social problems as a response to growing competition from socialist movements (Adriance 1995a). When the posseiros were threatened by large-scale land sales, expulsions, and violence, CEBs formed nuclei for rural resistance movements. Liberation theology played a strong role in the ideology of some of these groups. This provided ideological support to their cause, and faith that they would persevere. The CEBs sometimes organized the occupation of unused land and defended claims to land after it had been occupied. As the militancy of these groups grew, conservative Bishops attempted to censure the movement. The church successftilly pacified CEBs in urban areas and in the south. However, in the rural north this was not as easy. Even when Bishops were opposed to the stance of CEBs they could not exert complete control over remote areas because of isolation from church authorities (See Adriance 1995b). Not all CEBs became strong social movements in the north. Even though they were grassroots movements, skilled leadership from religious and lay organizers was crucial to this development. Furthermore, even those that did become activist were not always successfiil in helping peasant members to win their struggle for land (see case studies in Adriance 1995). Sometimes, changing leadership and turnover among Santo) which focuses on spiritual guidance with minimal involvement of politics.

PAGE 70

59 membership hampered the abihty of a cohesive organization to develop that could maintain itself through time. The Pastoral Land Commission (CPT) was formed in 1975 by a group of Catholic Bishops led by Dom Moacyr Grecchi but became an ecumenical institution with the involvement of elements of the Lutheran and Catholic church. The CPT was an explicit response to the land conflicts that were developing in the north. It was created to document violence and support the weak in struggles for agrarian reform. To do this the CPT documented and pubhcized rural violence (see Adriance, 1995: 109 for list), provided peasants legal support for land claims, and provided technical and material support for farmers (Souza Martins 1990; World Watch 1991). Unlike the CEBs, the CPT was not involved in organizing groups or land occupations; rather it assisted existing groups and monitored rural conflicts. In Acre, the church invited the Brazilian Confederation of Agricultural Workers (CONTAG) to help organize the spontaneous resistance movements developing around the CEBs into a network of rural unions (Sindicato dos Trabalhadors Rurais, STRs). The church encouraged participation and provided its facilities to the union. In 1975, the first STR was organized in Brasileia, and by 1980 STRs were formed in 7 of the 12 municipalities in Acre (Bakx 1986: 251). In 1977, the state assembly formed a special commission to examine the land sales taking place in Acre. Bishop Dom M. Grecchi testified that from 1969 to 1975, 80% of the state's land had changed hands. The new governor Mesquita testified that fi-om 1970 to 1975, 4,280,000 hectares out of the state's total area of 15,192,000 hectares (38% of the state) had been sold to only 284 investors all fi-om outside of the state

PAGE 71

60 (Cavalcanti 1983: 70). These transactions took place rapidly and often without official documentation of the transfer. Up to 1974, INCRA had issued only 81 titles in Acre (Bakx 1990: 154). Furthermore, many land transactions were fraudulent. In 1981. Jomal Gazeta published a Listao de Grilagem no Acre ("Big List of Land Fraud in Acre") which claimed that up to that point 3,600,000 hectares (24% of Acre's land area) had been the target of land fraud, falsified titles or fictitious registration (cited in Cavalcanti 1983: 109). Attempted Solutions Through Colonization and Resettlement As the rural crisis in Acre grew, the state and federal governments implemented programs to distribute land and settle families that had been displaced. These were largely populist programs aimed at appeasing unrest and relieving tension in urban slums. As will be seen, these programs lacked the long-term commitment to supporting those that were already settled on the land. In 1975, Acre's state government, in a weak attempt to appease smallholders and draw attention away from the social conflict, increased investment in the coldnias agricolas that had been created after the war. The colonies were renamed "Nucleos de Apoio Rural Integrado" (NARI) (Carvalho 1981). The investment consisted of small increases in the project areas and creation of centers within the colonies to concentrate social services. In each colony a center received agricultural machinery, a health post, a school, and more frequent visits by technicians and other state employees. These social programs were beneficial to colonists in these areas. However, the underlying premise of these projects was misguided. These programs were expected to affect the growing rural crises, but the NARIs focused on

PAGE 72

61 people long resettled, not those on the verge of expulsion or those already living in Rio Branco's shantytowns (Bakx 1988). In the conflict between new landowners and smallholders occupying the land, INCRA saw itself as a neutral arbiter. However, INCRA's policies were not favorable to the smallholders. The policy of arbitration in conflicts focused on having the posseiros accept alternative parcels of land, conceding their claim in return for parcels that were often small and remote (Bakx 1986: 306). By encouraging posseiros to accept resettlement, INCRA weakened their legal claim to land they held. As Bakx (1988) points out, INCRA focused on a symptom rather than on the root of the problem. INCRA's titling and settlement efforts were directed toward those smallholders who had already been pushed from their land instead of focusing on those posseiros who were still attempting to maintain the tenuous hold in the forests. Furthermore, these families could not be assured of immediate land because the resettlement process was slow. Resettlement also required families to shift from their traditional production system because the small lots (maximum size 100 hectares) were too small to provide a livelihood from extractive activities. The Federal Government Attempts Colonization Solutions In 1975, INCRA proposed the creation of Projetos de Assentamentos Dirigidos (PAD's) in Acre. The PADs were intended to settle displaced families already living around Rio Branco, regularize the demarcation of colonies, provide orderly occupation and develop viable agricultural projects. In Acre, the long term aim of these projects was not subsistence farming. Colonist families were settled on 100 hectare lots which were

PAGE 73

62 intended for diversified agriculture. The families were expected to operate a 3 to 10 ha. rubber plantation with 50% of the lot held as a forest reserve (Bakx 1990). From the beginning, these projects were controversial and difficult to implement. In 1976, the federal government tried to expropriate 700,000 hectares in Southeast Acre to settle 8,000 families. This land was divided into two PADs: 408,000 hectares in the Rio Branco municipality for PAD Pedro Peixoto and 292,000 hectares in the Sena Madureira municipality for PAD Boa Esperan9a (Bakx 1986: 308; 1990). Investors who owned the expropriated land did not give up without a fight. In 1978, there was heated debate in INCRA tribunals. Landowners opposed to expropriation called expert witnesses who claimed that the areas were unsuitable for settlement because of infertility and endemic malaria (local newspaper reports cited in Bakx 1986: 310). Another likely factor for the opposition came fi-om disputes over how much INCRA would pay (Cavalcanti 1983: 93). As a result, by 1977 only 50,000 hectares were in INCRA's possession, all in Boa Esperan9a, where only 15 families were settled (Bakx 1990). Some landowners successfully protected their land fi-om expropriation. This meant Peixoto would not cover a contiguous area because large ranches would be left within the boundaries. Peixoto 's area was further reduced by 150,000 hectares when landless seringueiros fi-om Rio Branco, Brasileia and Xapuri invaded the area and settled spontaneously (Bakx 1986: 310). Peixoto's dimensions changed slightly through time as new areas were added and settled. Today, Peixoto covers 317,588 hectares (INCRA 1996). INCRA was not the only agency attempting to operate directed colonization projects. In the late 1970s the state's colonization agency, COLONACRE, expropriated

PAGE 74

63 land on the eastern edge of PAD Pedro Peixoto. This project, called Reden^ao, was modeled after Malaysian small scale rubber plantation projects (COLONACRE 1979). Colonists in these projects were provided a small 50 -60 ha lot and each family was expected to plant and maintain a small 3 to 10 ha rubber plantation with trees provided by COLONACRE. Instead of each family maintaining 50 percent of their lot as forest reserve, only 10 percent of each lot was to be held in forest reserve. A large general reserve was set up within the project.^ The blueprint for Reden^ao was based on an elaborate grid with each square divided into pie shaped lots and a small communal area at the center of each square. Reden9ao included detailed rural and urban planning with designated urban centers, commercial districts and public spaces. The plan for Renden9ao was strikingly similar to PIC plans used during Transamazon, and just as unrealistic. Few of the urban features were actually implemented, and the forest reserve quickly disappeared, divided into lots during a land rush as posseiros and Reden9ao colonists completed to stake their claims. The resurgence of rural violence in 1980, which increased pressure for government action, culminated with the assassination of the head of the Brasiliea STR leader Wilson Pinheiro on July 21, 1980. A period of reprisals and assassinations followed and tension increased (World Watch 1991). INCRA settlement in Peixoto began officially in 1980. There were long delays between the expropriation of land and the actual resettlement, which put added stress and hardship on the resource poor families. By 1980 only 1,436 families had been settled in 6 This later developed into a classic "tragedy of the commons." COLONAC's weak administration did not police the reserve. Eventually, the reserve was invaded by families

PAGE 75

64 the original projects: 1,171 in Peixoto and 265 in Boa Esperan^a. Three years later only 3,291 families had been settled in Peixoto (Bakx 1986). More land was expropriated in 1981 and 1982, totaling 187,072 hectares. This area was in three parcels that became the Humaita, Quixada and Santa Lucia PADs (Bakx 1986: 311). Once again, there were difficulties implementing the projects INCRA blamed the state agencies for not fulfilling their obligations and the STR's which were discouraging resettlement by advertising the failure of the state consortium to keep its promises (Bakx 1986). After 1982 there was a break in the collective resistance of rural movements. In 1980 a centralized state-level federation was installed making it difficult for activist unions to control the federation. As a result, grassroots activism was restricted to the municipalities of Xapuri, Brasileia, and Assis Brasil (Schwartzman 1989: 7). Later, the Catholic church and some STRs (Rio Branco and Brasileia) adopted a reformist position that allowed state agencies like INCRA to negotiate and resolve land conflicts, usually through resettlement, not legitimization of existing posse. The STR in Xapuri refused to enter in dialogue with government agencies, preferring to push ahead with its own attempt at "agrarian reform fi-om below" (Bakx 1986: 260). These STRs opposed working with INCRA and the government, and advised members against agreeing to resettlement. A major reason settlement occurred slowly was that the process was convoluted and complex. The selection criteria for distributing land in these projects actually discouraged the selection of Acre's former posseiros. Many lacked personal documents required to start the process. The criteria emphasized financial capabilities and looking for land. This set off a massive land rush as colonists in Reden9ao rushed to

PAGE 76

65 experience with modem agricultural techniques, characteristics held by few of the exrubber tappers (Bakx 1990). Therefore, INCRA looked to other regions for "suitable" colonists. INCRA recruited migrants from the south with slide show presentations that depicted the projects in an overly rosy, if not fraudulent, light. Prospective settlers were shown pictures of projects with paved roads, rural electricity, and they were promised housing and cleared, fertile land for planting (colonist interviews 1996). Often they did not know what to expect until they arrived on the projects. In field interviews, original colonists told of being bussed or even flown to Acre only to be unloaded with their possessions on the side of the road at the mouth of a forest trail. The new arrivals were housed in tent cities and had to walk kilometers through the forest to the survey markers that defined the front of their properties. If they did not give up immediately, the first months consisted of back breaking labor of building a house and clearing forest for crops (if they were settled during the season to cut and bum their plots). INCRA provided some building materials for the first settlers and money or goods for the first months. After that, new settlers were on their own (later owners received nothing). Sometimes, families were settled months (or even years) before roads were constmcted. Colonists had to wait much longer for other infrastructure like schools and health posts. More often than not, hard work and malaria took their tool and many families left. Resettled and Abandoned From the perspective of many smallholders in Amazonia getting a plot of land from INCRA seemed to be a major victory. However, as they quickly learned, having the seize plots to get a portion of what they felt was their reserve.

PAGE 77

66 chance to own land did not mean their problems were over. Once settled, there were still many problems facing the colonist families. While some endured the hardship to establish successful homesteads, many more did not. INCRA planners tried to learn from their mistakes and the PADs were supposed to be improvements over the earlier PICs. Actually, the changes meant INCRA would have fewer obligations to the families once they were settled. No efforts were made to build urban centers, and the responsibility for providing basic services passed from INCRA to other agencies. PAD projects were easier for INCRA to initiate but they were not an improvement for colonists. The premise was that local and state agencies would provide services to colonists once they had been settled by INCRA. This was an unrealistic assumption. Local municipal governments varied widely in their ability and willingness to provide rural services. State agencies were weak, underfunded and inefficient. They were also highly politicized so that services were often doled out selectively in return for political support from colonists during election years. INCRA expected farmers to be supported by other agencies and when this did not occur, farmers were left largely on their own. As a result, INCRA could not flilly cut its obligations and responsibilities in the projects. Once the families were settled on the land, they found that the grid of property lines had been laid out without concern for topography or access to water, thereby repeating mistakes made earlier on the Transamazon colonization projects (Moran 1984). The PADs were often remote and inaccessible, and INCRA was unable to open and maintain all of the required feeder roads and bridges. By 1984, six years after the project started, roads were far from complete in Peixoto (Bakx 1986). This lack of access made

PAGE 78

67 it difficult for colonists to participate in the market economy, to sell crops or to purchase necessary commodities. Furthermore, other basic social services were substandard. The various agencies responsible for providing education and health facilities, as well as other services, were unable to meet all of the needs of the thinly distributed population. The level of support individual colonists received depended on their location within the project. As a result of these obstacles, it was very difficult for many small holders to stay on their land. No new PADs were established after 1982, but the existing PADs continued to expand. These projects were implemented slowly and settled in stages. As soon as new areas, were surveyed, settlers were assigned to the lots. Only later were roads built or schools and health posts constructed. Even though there was much hardship and turnover, the demand for land continued to increase as migrants began arriving from Rondonia in the mid-1990s. Some of these migrants received land from INCRA; others replaced those moving off the land. Migrants without land (and those who sold out) either moved to urban shantytowns, invaded project reserves or other peripheral areas. Democratization, the Continuing Crisis and the Climax of Social Mobilization By the mid-1980s INCRA and other agencies were not having an easy time dealing with Acre's smallholders and their settlement demands. Posseiros had realized that they could defend their claims and resist efforts at relocation. The influx of new migrants overwhelmed the capacity for orderly settlement in the PADs. For example, according to INCRA's 1986 "5 year plan", INCRA expropriated an additional 49,000 ha to settle 700 more families, but at that time, it was estimated that approximately 600 families were migrating into the state each dry season (Bakx 1988: 157). New settlement

PAGE 79

68 models were used to accommodate posseiros that resisted moving and spontaneous settlers invading available land. While these were immediately beneficial for the families involved, this new focus distracted the attention from those who had been settled earlier. The rubber tappers movement had learned to successfully defend their land rights using empates, non-violent strikes in which forest residents marched to peacefully confront workers cutting forests. In 1985, rubber tappers formed a national alliance with other indigenous groups. Later, the forest peoples showed that they were increasingly adept at drawing attention to their plight by forming international alliances with human rights and environmentalist groups. However, a major demand of the rubber tappers' movement was the formation of extractive reserves, which would provide the tappers a collective concession rather than dividing the area into small individual lots. This would not only protect their claim to the land but would allow them to maintain the extractive production; that is, their main means of support. Eventually the government gave in to pressure from the rubber tappers and allowed reserves to be designated. From 1987 through 1991, six extractive reserves were established in Acre. Whether or not the exfractive reserves would prove to be a long-term solution to the problems of the families living in the forests, the reserves at least removed the threat of expulsion from these families. Nevertheless, the problems of spontaneous settlement and land conflict did not end with the implementation of the PADS. Therefore, INCRA once again changed its focus. After the mid-eighties, the only new settlement projects implemented were rapid settlement projects focused on disputed areas. The first rapid settlement project (PAR) was started in 1986. A total of 5 were implemented before the end of the decade. These

PAGE 80

69 projects were much smaller than the PADs. They could have as few as 24 families and cover less than 900 hectares. This was in stark contrast to the PADs which were all designed to accommodate more than 800 families over more than 60,000 hectares, some substantially more. Emancipation From Unfulfllled Promises Change in Acre continued. In 1992, when the BR-364 was paved to Rio Branco, it promised to make the state more accessible for migrants entering Acre. Not only were the PARs drawing colonization resources to other areas but developments on the national stage were also detrimental to the families that had akeady been settled earlier on the PADs. In 1990, the government of President Collar cut funds to federal agencies to fight inflation. This measure resulted in cuts in personnel and services to most federal institutions, and consequently increased hardship for the colonists that were already settled. A case in point is SUCAM, the malaria control agency, one of the only institutions that consistently reached most of the colonist households. With the cut in personnel and lack of funds for gas and vehicle maintenance, SUCAM had to cut back on fieldwork. This had a devastating impact as malaria was (and still is) a major problem throughout the PADs if not controlled. INCRA had been weakened in the late 1980s, and during Collar's program to control inflation it virtually ceased to be a factor in rural Acre. When INCRA employees were unable to oversee all of the projects they could not restrict land sales or land concentration within the projects. Many professionals from nearby urban centers entered the colonization projects as speculators and bought up the land of colonists.

PAGE 81

70 Under President Cardoso, INCRA has regained some strength and the pace at which PARs have been implemented has increased. In 1994 one PAR was implemented in Acre, in 1995 three projects were started, and in 1996 seven more PARs were implemented in Acre. INCRA personnel reported pressure from Brasilia to emphasize settling people (instead of maximizing the number of titles issued). This appears to be a response to the Movimento dos Trabalhadors Sem Terra (MST) which has grown in strength. While members of the MST are less willing to move to Amazonia and are pressuring the government for land distribution in the center and south, these new PARs in Acre may be part of a policy to boost settlement statistics. Today, those areas settled long ago are no longer the focus of attention and receive only a small share of resources devoted to colonization in the state. Some municipalities (i.e. Acrelandia) are trying to meet basic needs of the rural population; others cannot or choose not to. State agencies continue to have problems with ineffective agencies and programs distorted by local political agendas. A good example is provided by a rural warehouse built by a state agency in Peixoto. In 1994, the warehouse had been built with great fanfare. It was stocked with merchandise, a tractor, and a scale for buying grain. Governor Magalhaes and other politicians arrived accompanied by local news media to show their good work. When journalists and politicians left, the workers loaded the stock and machinery onto the tractor's wagon, closed the warehouse, and left. About a week later a truck arrived with workers to disassemble the warehouse. They said they were going to move it to Xapuri. However, this misguided attempt to build popular support through with bait and switch tactics backfired. The local farm association realized what was about to happen and gathered to block the workers. The farmers were

PAGE 82

71 successful in holding onto their warehouse but it did them little good. In 1996, the warehouse was still sitting empty. There are local initiatives that do seem to be helping Acre's rural families. Recently, the municipal government in Rio Branco has begun an innovative, small-scale land distribution and colonization project, called the PoloAgroFlorestal (Slinger 1996). This project consists of formation of a small settlement near the capital. Colonist families were given assistance to establish small agroforestry plots and the municipal government provides necessary social services. This project is especially notable in the great lengths taken by the municipal government to support the settlers and use the settlement as a model for future development. While the project has had early success it was startlingly similar to settlement attempts the state organized around Rio Branco in the 1950s. Hopefully, these colonists will not have the same fate as their predecessors. Conclusions As they had throughout the region, government development policies brought striking change to Acre. When the development transition arrived, it produced rapid and dramatic change. The traditional economy with unique land use and tenure systems was overturned as large scale capitalist investors gained the upper hand. The land sales brought about by the policies of B AS A and SUDAM caused widespread expulsions of Acre's rural population, increasing landlessness, shantytowns and rural resistance, all of which required a government response. The government implemented colonization projects as an attempted solution to the rural crises. However, the projects were difficult to establish. There was not a strong commitment to supporting families after they were settled. Apparently, the primary goal was to lower rural tension. As a result, their projects

PAGE 83

72 did not eliminate land disputes. With continued land conflict in other areas, government attention focused on new problem areas. Colonists already settled were left to fend for themselves.

PAGE 84

CHAPTER 4 LANDOWNERSHIP TURNOVER IN PEIXOTO In the preceding chapters, I have argued that colonization projects were primarily populist vehicles for appeasing rural discontent. Brazil's government was determined to facilitate capitalist investment in Acre, which continued to generate opposition from smallholders who would be displaced. Through time, as the government attention shifted to new centers of rural conflict, families that were already resettled were forgotten. These developments diverted resources away from assistance programs that had been overwhelmed from the start. As a result, those families that accepted farm lots found that rather than the bonanza they expected, they were left in a marginal situation in a peripheral area. One way to gain a fuller understanding of landownership turnover is to examine a mature colonization project and observe where turnover has occurred and how it has affected settlement within the project. This research focuses on the PAD Pedro Peixoto, the largest and oldest federal resettlement project in Acre. The goals of this chapter are to describe landownership turnover in Peixoto, explain the complex factors contributing to this phenomenon and identify the settlement pattern that has developed as a result. This chapter combines quantitative and qualitative methods to reach these goals. Peixoto is a natural experiment for studying the outcomes of directed resettlement in the Amazon. As a natural experiment, there are limitations to the data that can be gathered. In this case, I was studying change (landownership turnover) but could only 73

PAGE 85

74 observe the aftermath of that change. Families that turned land over are gone and the conditions that drove the process have changed. By observing current conditions, it is possible to understand past change. I begin this chapter by describing Peixoto and the methods and sampling design I used to gather a representative unbiased sample of Peixoto 's population. In the second half of the chapter I will describe the state of landholdings in Peixoto. I will describe my estimates of the level of turnover, and give explanations of why it has occurred. Finally, I will describe the settlement pattern that resulted from turnover in Peixoto. As I will show, there is a high incidence of landownership turnover in Peixoto. The reasons families are not staying on the land are complex and include macro processes outside the control of those involved. I argue that harsh environments created by inadequate infrastructure and services contributed to this exodus from the project. This isolation made it difficult for farm families to meet their needs. To support this argument, I will show how turnover varies drastically from one area to another; a pattern that reflects differences in infrastructure quality. The difficult struggle to survive in Peixoto, compounded by a climate of speculative investment in land, has meant that one result of ownership turnover is increased land concentration. In some locations, many original settlers maintain control of their land, while in others most have been replaced by incoming smallholder migrants, or the land is consolidated into large holdings. The level of turnover and the resulting settlement pattern indicates that the project has not given long-term land security to many of the families resettled in Peixoto.

PAGE 86

75 Peixoto Peixoto is the federal government's oldest and largest directed resettlement project in Acre. It was first proposed in response to land conflicts and the rural exodus that plagued Acre during the early 1970s. Peixoto was a comprehensive settlement plan in the tradition of the large colonization projects of the period. The government intended to use Peixoto to relocate thousands of displaced people living in the shanty towns around Rio Branco and redirect others away from the shanty towns. Peixoto was designed to systematically settle displaced families, build up basic infi-astructure (roads, bridges services such as schools, health posts, and storage), and help families establish their farms. The project's 317,588 hectares were originally divided into about 4,000 individual farm plots. More than just a place for the poor to live and farm, plaimers hoped this would be a long-term solution for the plight of the dispossessed. Peixoto stretches across the state's southeastern comer, from Bolivia to the state of Amazonas (see Figure 1.). The project dominates a highly populated part of the state; however, when first visited, it seems oddly peripheral. While it is crossed by the state's three major highways-the BR-364, the BR-317, and the AC -40, all of which are pavedmost families are settled back away fi-om these roads in areas that are often remote and inaccessible. Roads leading into the interior of the project are substandard and often impassable. Peixoto begins only 25 kilometers east of the capital Rio Branco, but this proximity to urban services is deceptive because of the immense size of the project. Rio

PAGE 87

76

PAGE 88

77 Branco is the principal urban center in the state: a hub where residents must go to carry out official business with government agencies, access social services, and visit markets. Peixoto may lie right outside the capital, but most residents travel much further (some over 100 kilometers), often on substandard roads. Few residents can visit the capital, take care of business and return home in the same day. There are other smaller urban centers around and within the project that provide some services for the residents of Peixoto. These include Senador Guimard, a large town that grew out of a state colonization project from the 1950s; Placido de Castro, a village dating from the rubber boom; Campinas, INCRA's administrative field headquarters in Peixoto, and Acrelandia, a small town east of Peixoto that was formerly the administrative center of the defunct Reden^ao resettlement project. These smaller centers are mainly transit points to Rio Branco, the dominant center. A large project like Peixoto could only be implemented over time, so INCRA settled it in stages. Unfortunately, it repeated mistakes from other projects: Peixoto was planned in Brasilia and the network of roads and the grid of farm lots were superimposed on the landscape ignoring topographical features, water courses and soil characteristics of the area. As a result, roads would be hard to maintain and property lines would be inappropriate. Many lots were left without access to water, while others were divided several times by the same stream, as had been the case during the Transamazon project (see Smith 1980). As INCRA teams surveyed sections of the expropriated land (glebas), they started near highways closer to Rio Branco and the administrative centers. After the land was surveyed, INCRA drew the names of qualifying families and began settling

PAGE 89

78 them on these lots, often before roads were constructed. Once settled, INCRA moved on to the next Gleba, each time moving farther into the project. Peixoto continued to expand, even after the original area had been filled. INCRA had left reserve areas within the project. These were places where INCRA did not attempt settlement because they were deemed inappropriate for agriculture (due to the presence of swamps, flood plains, poor soil) or unsuitable for road construction. Later, as the project filled out and land was no longer available, these reserve areas were invaded by new migrants and other landless families. Eventually, INCRA was forced to move in and standardize the property lines in these invasions. As a result, Peixoto has continued to grow. Today, some glebas date ft-om the late 1970s while others were settled only in the 1990s. The glebas are still the administrative unit used by INCRA to run the project. However, they do not constitute a very significant division on the ground. The principal landmarks within the project are the primary and secondary access roads. A typical primary road (called a ramal) has a specific name much like a city street. For example, some are named after the construcfion company that built them, as in the case of the Mendes Carlos I and 2 roads. Others are named after prominent people, as is the Nabor Junior road, or they take the name of rubber estates that had been located nearby such as the Granada and Sapucai roads. Secondary access roads, (called linhas) branch off some of the larger primary roads. The linhas are numbered based on the distance in kilometers fi-om where they branched off to the start of the primary road. As the major landmark, peoples' addresses and places are located using their position on the road (i.e. He lives at kilometer 10 on Ramal Mendes Carlos 2).

PAGE 90

79 Unfortunately, INCRA could survey and settle families faster than roads could be built. In the early stages, families often were given lots that could be reached only by long treks through the forest. It was not uncommon for the first families settled in a gleba to wait for months or even years for a road to pass near their lot. By 1986, many families were still enduring hardships and roads were far from complete (Bakx 1986). Roads still vary widely in quality and maintenance. Since roads were superimposed on the landscape without regard for topography or land forms, some are almost impossible to keep open to vehicles. Others are never maintained because of unresponsive local government or because residents had the misfortune of supporting a defeated political party. From my previous experience in Peixoto in 1992 and 1994, it was apparent that conditions and settlement density varied from place to place in the project. Many areas were sparsely populated, and some were clearly concentrated into large single holdings. In other areas, family farms still predominated but many of the residents turned out to be recent arrivals. I believe the settlement behavior was not coincidental. It is not random chance that families stayed in some places and not in others. I wanted to better understand the factors behind this apparently simple phenomenon: Why families were not staying on their land. The first step to understand landownership turnover and settlement variation is to acciurately describe what is happening. Obviously, before describing and explaining timiover, it is necessary to gather data. Peixoto presents numerous constraints to data collection. Therefore, the next section defines turnover, the difficulties measuring it in Peixoto and the methods I used to overcome these constraints.

PAGE 91

80 Research Design And Sampling Methods Appropriate For Peixoto Whenever a landowner transfers land rights to another party, landownership turnover takes place. Such a transfer can be through sale, trade, abandonment or even expulsion. In theory, it should be easy to keep track of land ownership turnover, but it is not. Quantifying the extent of landownership turnover and accurately describing the resulting settlement is a difficult task because of conditions in the project, inadequate record keeping and the nature of the turnover phenomenon itself For research to be productive, I had to overcome these problems and draw a unbiased representative sample for the entire Peixoto project. This would allow a valid estimate of turnover in Peixoto and allow me to go beyond discussion of individual cases. The conditions that make life tough for the colonist farmers also make it difficult to accurately describe the changing settlement patterns in Peixoto. Travel is difficult in Peixoto: the project is huge, transportation is inadequate, and roads and bridges are poorly maintained. Travel fi-om one side of the project to the other can entail hundreds of kilometers. It is not uncommon to find roads that can only be traveled with four wheel drive vehicles, or roads that cannot be used because of fallen bridges. Some portions of the project are only accessible on foot. Furthermore, because regular transport services are inadequate, travel can be costly and time consuming. Regardless of these problems, field work is a necessity because official records give little information about what has happened or what can be expected on the ground. Inaccurate documents and incomplete records compound the problems in describing Peixoto. Once families have possession of the land it is difficult for government institutions to monitor changes in ownership. In remote areas, like fi-ontier

PAGE 92

81 colonization projects, land transactions often take place without legal sanction or proper documentation. INCRA technicians admitted they did not know who lived on many of the lots or how often the land had changed hands. With limited personnel, vehicles and funds, there were often long periods between INCRA field visits and it was difficult for them to keep track of what was taking place in the field. INCRA's personnel complained that it was not uncommon for lots to change hands several times in a single season. Not only are poor records a problem with the field stafions, but inconsistencies are found in some of the basic documentation for the project. INCRA documents indicate that there are 4,025 properties within the project; however, in a 1995 INCRA map of the project, I counted only 3,960 farm lots. It is not clear where INCRA is counting the additional properties. I know fi-om experience that my count of farm lots exaggerates the number of properties because in some areas land has been concentrated. For these reasons, there is a lack of accurate statistics about what is taking place in Peixoto and no fi-ame for easily drawing a sample. Because turnover may or may not be a documented process, it is difficult to know the ownership status of a lot without visiting the area. Even then, past owners are gone, so it is hard to know why they left or other details about past episodes. The researcher can only talk to current residents, ask them about their experiences and what they know of past events. Such questions would include: Were they the first family settled on the lot? Do they know how many owners were there before them? How long have they occupied the site? These are all measures of turnover. These measures are good indicators of the stability of settlement patterns: they tell whether families have stayed on the land and how long.

PAGE 93

82 A valid description of Peixoto requires the use of a systematic random sample where all combinations of sample units have the same probability of being drawn (Scheafer, Mendenhall, and Ott 1990: 28). However, in a place like Peixoto, simple random sampling would spread single sample units throughout the project. Logistically, this would be impossible for my research. ^ Instead, I chose a two-stage cluster sampling method. To do this, I randomly selected areas (clusters), then within each selected cluster I randomly selected sample units. This grouped my observations to ease data collection without biasing my sample. Choosing the appropriate method for drawing a sample in Peixoto was just the first step. I then needed a sampling frame (or list) that could be used to accurately select a sample. I was interested in farm properties and their owners. If an accurate list of property owners existed, sampling would be a simple matter. However, there is no accurate list available. Since INCRA maintained detailed maps of farm lots in the project, I realized that I could use a 1995 INCRA map and sample lots as a proxy for farm properties. Farm lots were a proxy for properties because some lots have been concentrated into single holdings and, occasionally, single lots have been divided by families into smaller parcels. The lots were numbered and could be located in the field because SUCAM (Superintendencia do Combate da Malaria) painted these numbers on the houses to identify the location for malaria control. As a first stage of my sample design, (selection of the clusters), I decided to use road segments. Roads are not just transportation corridors through the project. They are ^ Such samples have been done in Peixoto. Witcover and Vosti (1996) used simple random samples to compare two Amazonian colonization projects (including Peixoto).

PAGE 94

83 how locations are designated when travehng in the project, and they are corridors along which social interaction among residents takes place. I could chose a ten kilometer road segment (an area that could be covered on foot) and be reasonably sure of finding farm lots that shared similar conditions and people that interacted to some extent and formed a community. Conditions would be more alike among lots sharing the same road than between lots on different roads. I divided Peixoto into 82 sections each centered on a segment of access road. I then randomly selected 10 of these areas as clusters (see Figure 3). This was the maximum number I could visit and survey during the first six month field season. Within each cluster I then selected 25 farm lots and 5 alternates. I selected the alternates as replacements because it was likely that multiple lots fi-om the same property would be selected (farm lots have been concentrated into larger holdings and I was interested in properties not lots). I wanted to get at least 25 lots from each cluster but would take 30 if possible. When time was available, the households that occupied alternate lots were included. In 1995, 1 spent 6 months alternating week long visits to each of these clusters. During these visits, an assistant and I visited the sampled lots to administer a questionnaire survey. In 1996, 1 returned to four of the clusters for prolonged stays to gather case study information by visiting the majority of occupied properties. These visits enabled me to gather additional data on settlement history, community organization, and agricultural production. This was a multi-year study with a well-financed team of researchers. My research did not have the financing, personnel or time.

PAGE 95

84 The sample resulted in 223 cases where at least a minimum amount of information could be determined.2 The 223 cases do not mean that there were 223 properties sampled. Sometimes multiple sample units appeared to be part of larger properties with absentee owners. Even though they were likely to be one large property, we could not tell who actually owned each area. Often, these units were within large expanses of pasture. Because they were cleared and fenced within larger properties, we could assume the original colonists no longer occupied or owned the lot. Although we suspected these lots were part of larger properties, they were not grouped together because we did not know the exact location of boundaries. Therefore, we chose to err by counting too many cases. For example, out of 29 lots visited on one cluster (Nova Aldeia), 26 appeared to be part of one large pasture. Each of these lots was counted separately because we had no information about property lines. In 146 cases, we interviewed families that owned and occupied their land.^ These 146 form a sub-sample that is used for much of the analysis. The response rate appears to ^ Out of a possible 300, we located and visited 262 sample units. However, not all of these 262 units became individual cases included in the study. One family declined to participate in the study. On ten lots no information about ownership or occupation status could be gathered so these were missing. The remaining units were located in multi-lot properties that had already been sampled and so were combined with the lot selected earlier. Whenever we could verify that additional lots from a multi-lot property were selected, and that these property was occupied by a colonist family, we combined the lots and counted them as a single property. 3 Originally, we interviewed 153 families. During the course of the interviews we learned that 8 of these families occupied lots they did not own. They were employees of the owners. As a result, they were excluded from analysis because the surveys only examined characteristics of landowning households.

PAGE 96

85 1 1 Peixoto i H Field Sites: k A. Linha 24 F. Ramal do Bee ^^—^ Paved Highway B. LinhaS G. Nova Aldeia C. Bigode H. Triunfo '' Uiq)aved Highw^ D. Iquiri I. Linha 16 E. Souza Araujo J. Linha 14 Figure 4.2 Peixoto and Selected Field Sites

PAGE 97

86 be a very low: 146 out of 223. However, it is important to remember that my sampling method included areas with very low population densities. Therefore, there was no one to interview on many lots. In these cases, only a minimum amount of information was available because the lots had absentee ovmers. They are officially included in Peixoto but smallholders no longer occupy the land. If the sample had focused only on individual households, these areas would not have been counted. I used data from the cluster sample two ways. First, I used the complete sample of all 223 cases where ownership could be determined to measure turnover throughout Peixoto. Second, I used a sub-sample of 146 interviews with families that owned and occupied the land. Because of the interviews, I have much more information on these sites. This sub-sample provided turnover measures, allowed me to describe household behavior, and to make comparisons within the population to examine why turnover is taking place. Who Are Peixoto's Current Residents? Before analyzing turnover in Peixoto it will be helpful to describe the characteristics of the population that occupies the project. The 146 families are a diverse group. They come from 1 8 different states and primarily from rural backgrounds. For most families, their single lot is the first land they have ever owned and it is a struggle to maintain that land. Although conditions in the project have improved fi-om the early years, life on the project is still difficult. Infi-astructure and services are poor throughout the project: transportation is inadequate, schools are substandard, and health care services are rudimentary. Much of the social infi-astructure constructed during the 1980s has been abandoned. Some INCRA schools and health posts literally have been stripped to the

PAGE 98

87 foundation, with colonists even carrying off individual bricks. Often, where services still exist they are provided by local municipal governments. Origin Peixoto's families migrated from all over the country, reflecting internal migration within Acre, as well as immigration from other regions in the country. The table below depicts the birth place of household heads and spouses,^ region where household head lived the longest, and birthplace of the couples' parents. As can be seen, the largest portion originated in South and Central Brazil. However, most household heads have spent the longest period of their lives in Acre after a life of frequent moves. Table 4. 1 Origin of Selected Adults in Peixoto Household Survey Acre and other regions % Acre % North % Northeast % South / Central N Birth place of Head and Spouse 28% 8% 16% 47% 265 Location where head lived longest 49% 9% 7% 33% 146 Birthplace of parents of head and spouse. 18% 10% 31% 41% 462 Note: I have divided this data into the state of Acre and three broad regional groups. The North region includes Amapa, Amazonas, Para, Rondonia, Roraima. The Northeast region includes Alagoas, Bahia, Ceara, Maranhao, Paraiba, Pemambuco, Piaui, Rio Grande do Norte, Sergipe. The South / Central region includes Distrito Federal, Goias, Mato Grosso do Sul, Mato Grosso, Tocantins, Espirito Santo, Minas Gerais, Parana, Rio de Janeiro, Rio Grande do Sul, Santa Catarina, Sao Paulo Peixoto was initially planned to resettle Acre's displaced population. Initial INCRA figiires from 1983 indicated that 84 percent of colonists were from Acre. However, only a couple of years later researchers were speculating that this figure was less than half that amount (Bakx 1986). The higher percentage of southerners is a result of migrants

PAGE 99

88 flooding into the state in the mid to late 1980s. Many of the Acreanos in my sample told of being recruited out of the shantytowns around Rio Branco, but others told of being evacuated by the air force from remote towns like Tarauaca for resettlement. Some of the Southerners reported being recruited and flown or bused north by INCRA. However, it is my impression from conversations with numerous settlers that most southerners entered the state spontaneously. For many of these households, their arrival in Acre marked the most recent stop in a life of continued migration. For household heads not bom in Acre, 76 percent had lived in Parana in the past and 88 percent had lived in Rondonia at some point. The population comes from a primarily rural background. For the older generation, 81 percent (out of 416) were bom in rural areas which indicates that even if sample members came from an urban background they are probably only one generation removed from mral zones. Many of Peixoto's households had migrated from other rural areas, and many who reported urban stays before Peixoto described these as brief layovers. Furthermore, 36 percent of household heads had lived on a seringal in the past. Life on a seringal would entail much more remote and rustic conditions than would be found on a colonization project. Family Composition Peixoto has a young population; therefore many probably migrated with parents. The average age of household heads is 43 years, but among all residents the average age is only 24 years. Almost half (49 percent) of the population is 18 years old or less. The average family has 5 members, and family size ranges from 1 to 15 members. Head and spouse were self identified by informants. The majority of heads were male

PAGE 100

89 Education I expected that the quality of education and health care services would play a significant role influencing differences in turnover rates between clusters. However, these factors seemed to be uniformly poor throughout the project. I will use my data to briefly describe this situation, but this survey did not gather detailed information to allow a close examination of these factors. Within the Peixoto project, 95 percent of the families interviewed said that a school was available in the community. These schools are widely dispersed throughout the project as 71 percent of the famihes had to travel less than 2 km to arrive at the school. However, many of the school buildings in Peixoto have become run down, especially those farther off main routes. These schools did function but at a basic level. Most lacked basic teaching materials. The vast majority of schools only offered the primary grades (89 percent up to the 4th grade). Most of the population has some education but not much. Of the population over 15 years of age, 26.6 percent are illiterate, 26 percent had some education but less than the 4th grade, 28.2 percent had completed the 4th grade and only 18 percent had continued past that point. Families that want their children to receive additional education are restricted by the lack of schools providing the upper grades in rural areas. The lack of opportunity means families with children that finish the primary grades must send them elsewhere to continue their educations or do without. This was a common complaint of families that have children at home. While sending children away to study may be difficult or expensive, 23.3 percent of the families did have children that were studying somewhere but not all.

PAGE 101

90 else. Almost a third of these children are studying in Rio Branco (32.4 percent) with many also studying in Acrelandia and Placido de Castro. The local administration of Acrelandia is especially good at providing access to education for rural children. Acrelandia operates a rural school bus that travels surrounding roads to bring teenagers in for night classes. This was a very popular program among the families interviewed because the older children could still live (and work) at home. From the response we observed, many families around Acrelandia were taking advantage of the program. However, such programs are uncommon. Health Services Out of the ten areas in which the survey was conducted, only Triunfo did not have access to a local health post. The average distance families had to travel to arrive at local health posts was 10 km, with the maximum distance being 29 km. However, much of the time, these health posts barely functioned. They were poorly stocked, attendants often had minimal training, and services were provided infrequently. Because of the poor service, most residents treated family members at home, imless serious illnesses were involved. In cases of serious health problems or emergencies, the family had to make the difficult trip to hospitals located in urban centers. The hospitals most firequently used were located in Campinas, Rio Branco, and Placido de Castro. The mean distance traveled by families to arrive at the hospital was 41 km. Many families reported that they also relied on urban pharmacists as primary health care providers. They said pharmacies were easier to locate and use than public hospitals. This convenience stemmed from the fact that rather than having to visit several locations for treatment, medications could be purchased at the same time.

PAGE 102

91 Malaria is a major problem in some areas of Peixoto. Out of all the households I visited in Peixoto, 70 percent of the individuals claimed to have suffered from malaria in the past. SUCAM, the government's malarial control agency, was active in all the clusters and is often one of the only government agencies that regularly provides services to Peixoto's population. SUCAM must continuously battle to control malaria and even then outbreaks still occur. Transportation As I have mentioned earlier, many areas in Peixoto are remote and inaccessible due to the poor transportation infrastructure. For the residents of Peixoto, access depends on road quality (how often the roads and bridges are maintained and whether they can be used year round), the availability of transportation services for residents and their cargoes, and the cost and dependability of these sources of transport. This is an important variable because it determines how easily residents can leave the project to visit the market, take care of administrative business and use social services. Accessibility also determines how well the government is able to get services out to the residents. Out of the ten clusters I visited, four were based around roads that could not be used year round (Linha 24, Linha 16, Linha 14 and Triunfo). This is because the dirt roads became impassable quagmires during the rainy season. Other communities I have visited in the past have suffered when bridges washed out during the rainy season. For example, the cluster Linha 16 was divided in two when a bridge located about 10 kilometers from the mouth of the road washed out. This happened a couple of years before the 1995 survey. Families that lived east of the bridge walked the few kilometers east to the AC 401 to catch the bus to Placido de Castro. Families that lived west of the

PAGE 103

92 bridge usually walked over 20 kilometers west to the BR-364 highway. Only two clusters (Souza Araujo and Bigode) had buses that passed directly through the community. In the other communities, residents had to travel an average distance of 5 kilometers to get to buses, but some had to travel as far as 20 kilometers. Away from primary roads, bus service is only available once each day. Furthermore, transporting cargo can be a real problem. Some communities had access to market trucks (usually on a bi-weekly basis) provided by the government, but three clusters (Linha 8, Linha 16, and Linha 14) did not receive this basic service. Farm Properties Most of the residents sampled are living on the first and only land they have ever owned. The farm properties sampled ranged in size from 24 to 219 hectares with the average being 70 hectares. Most of these families have documented ownership of some sort. Among the households, about half have definitive titles (70 households) and most of the rest are somewhere along the INCRA process for gaining fiill ownership of their land. Only about 9 percent claim to have no documentation and only about 5 percent have only the equivalent of a sales receipt {contrato de compra e venda). These numbers appear to show fairly secure property rights; however, it must be remembered that this excludes the concentrated areas where no interviews were possible. Our data indicate that the land concentration taking place in Peixoto is not the result of expansion of smallholdings. Few small farmers own additional lots. In 1991, Cavalcanti estimated that only 18 percent of the families owned more than one lot (1994). In our sample only 32 families (21.9 percent) admitted that they owned additional property.

PAGE 104

93 Among the families that owned additional property, the majority have only one additional holding (see Table 2. below). Table 4.2 Additional Properties Owned by Peixoto Residents Number of additional Frequency Percent properties 1 26 81.3 2 5 15.6 3 1 3.1 Total 32 100 Most of these additional holdings are on the same feeder road. Often families buy these additional lots to provide land to dependent children once they marry. The additional property is usually nearby. Table 4.3 Location of Additional Property Frequency Percent Same road 18 56.3 Same glebal other road 3 9.4 Same project/ other gleba 3 9.4 Invasion 5 15.6 Urban 3 9.4 Total 32 100 Few of the colonist families interviewed had owned property before arriving on the current lot. Among interviewed families only 36 (24.7 percent) admitted to previously owning property. Of the small number that had previously owned land only 10 individuals (27.8 percent) had owned land in other states. The majority, 66.7 percent, said the earlier holding was located within Peixoto (see Table 4. below). Apparently, these families had exchanged lots and moved deeper into the project.

PAGE 105

94 Table 4.4 Location of Earlier Property Freauencv Percent Same road 6 16.7 Same glebal other road 11 30.6 Same project/other gleba 7 19.4 Other project 2 5.6 Other state 10 27.8 Total 36 Now that I have described the difficulties working within Peixoto and presented some of the basic characteristics of Peixoto 's current population, it is possible to discuss turnover. Turnover Measures There are different ways to measure turnover in Peixoto. I wanted to know whether the people originally settled remained on the land, and, if the lot had changed hands, how often this had happened. I used three different measures of landownership turnover: 1) Is the current resident the first settled by INCRA? 2) How many families have lived there since first settled? 3) How long has the current owner had possession of the lot? These measures would help me to define the extent of turnover. Once turnover behavior has been described, I can attempt to explain why this had happened. Explaining turnover is trickier than describing it. In this research, I was limited to people currently occupying the land. Since earlier residents were gone, I could not gather information about their specific reasons for leaving. 5 The first measure of turnover is the simplest to measure and to understand: whether the current resident was the family originally settled. This variable indicates I did track down some former residents, but not systematically.

PAGE 106

95 whether the families INCRA originally settled have remained on the land. The second measure--how many families have owned the lot-contains more information and indicates how frequently turnover is taking place. I will refer to this as the "previous owners" measure. Interviewed families could usually tell me how many individuals had owned the lot before them. While this does provide more information for each case than the simple measure, there are fewer cases. These data also are somewhat suspect because they rely on informant recall of events that they may not have witnessed. Farmers were usually confident in answering these questions, so I think it is worthwhile to consider their information. The final measure of turnover, the one that gets closest to describing the pattern of long-term settlement, is the length of time current residents have occupied the lot. In Peixoto this is not a straightforward measure because different areas in the project were settled at different times. There can be over a decade of difference in the settlement dates between some "original" colonists. However, there is a way to get around this. Instead of a raw score, I measured settlement time as a percentage of time the lot was available for occupation. In interviews, I learned when each family arrived. While I did not have the exact year that lots were surveyed and distributed, I did know the year each road was built. This was not exact, but in most cases these dates approximate (within a few months) the time of initial settlement. The period from the year of settlement to 1995 (when the survey was conducted) is the potential time the lot could be settled by any family. I call this measure "settlement time." By finding out when a family arrived, I could calculate the proportion of the time available for settlement that a family had occupied the lot (.00 to 1.0). For example, on a road built in 1985 a lot was occupied by

PAGE 107

96 a family that arrived in 1987. Since the survey was done in 1995, the family had been there eight years, or 80 percent of the time that the lot could potentially have been occupied. My observation for the settlement time variable for this family would be 0.8. I could only calculate this measure for the 146 families that were interviewed. With these 3 measures I will now describe the landownership turnover in Peixoto. Each measure produced similar (but not identical) results that indicate that most colonists settled in Peixoto have not held onto their land. Landownership Turnover Landownership turnover in Peixoto is very high. Few of the colonists originally settled during the past decade and a half are still on the land. Using the simple measure of turnover, I estimate that the portion of original residents still occupying their lots was only 22 percent (10 percent) (see Appendix A for equations used) Compared to earlier measures, landownership turnover has increased. In 1984, it was estimated that 92 percent of original families were still on their land; by 1991 this figure had decreased to 39 percent (Cavalcanti 1994). This sequence of estimates defines a downward slope of small farm survival as attrition and turnover increased through time. An INCRA employee estimated (off the record) that more than 8,000 families had probably settled in the project. This included both current residents and others who had failed in their homesteading attempts and moved off the land. It is obvious that Peixoto has not provided many of the original colonists the opportunity to stay on the land long-term and to develop viable farms. Some of the lots had changed hands more than once. While the average lot was on its second owner, some lots have had as many as five owners. Just under a third of the

PAGE 108

97 lots have changed hands more than once which means they have had three or more owners. Table 4.5 Number of Previous Owners Ownership Categories Cases percent First owner (original) 52 23 Second owner 62 28 Three or more owners 65 29 Missing 44 20 Total 223 On the 44 missing cases, I could only determine that they were not occupied by the original colonist family. Either the lot had been absorbed into a large ranch (so it could not be the original colonist), or neighbors confirmed that the current absentee owners were not those originally settled by INCRA. In these cases, I did not know how many times the land had changed hands before it was purchased by the absentee owner. This turnover measure might have been even higher if I could have gathered more information on the lots included in the concentrated areas. Many of the lots owned by absentee owners apparently turned over multiple times before being absorbed into ranches. The amoimt of time a family occupied their lot would add much to our description of turnover. Because a lot has turned over, it does not mean that long-term settlement has not developed with later owners. Some lots may have turned over early, but later residents were able to establish homesteads and remain on the land. However, information on settlement time could only be gathered for the lots where interviews were conducted. Therefore, to continue the discussion of turnover it is necessary to focus on

PAGE 109

98 the smaller data set consisting of the properties where households interviews could be conducted. The sub-sample of 146 interviewed families is not the same cross section of Peixoto population as the large sample of 223, but it is similar. This sample represents only the areas where farm families still live on their land. Because the absentee owners have been excluded, the turnover estimates are a little different. As can be seen in the table below, the proportion of original residents is somewhat higher. Table 4.6 Ownership Status of Sub-Sample of Resident Owners Ownership status N Percent Original Owner 49 34 Second Owner 40 27 Three or more owners 52 36 Unknown 5 3 Total 146 100 By looking at the potential settlement time data for these families it appears that settlement is a little more stable than the turnover figures imply. The average settlement time measure indicates that the typical families had occupied their lots 67 percent of the time ( 6 percent). The population ranged from scores of 0 for new arrivals to 1 for residents who had occupied lots from the beginning. From this smaller sample, 35 percent have been on their lot for all of the potential settlement time (and received scores of 1). In other words, according to this measure, 35 percent of the sub-sample are original residents. This slightly higher percentage may be due to the fact that roads were not built immediately some colonists gave up and were replaced by others before the road arrived. Not all original residents received a score of 1 because some glebas were not completely settled right away. Conversely, some second and third owners received high

PAGE 110

99 scores because they acquired lots that had turned over in the period before the road was built. However, as would be expected, when lots are grouped by ownership status (original settlers vs. later owners) the occupants of lots that had turned over in the past had lived in Peixoto for shorter periods of time. Table 4.7 Average Settlement Time by Ownership Status Ownership status Average settlement time (%) First owners (original) .95 Second owners .61 Third owners or higher .45 Why Is Turnover Taking Place in Peixoto? By all three measures, turnover was high in Peixoto. Moreover, the project had not led to long-term, stable occupation by most residents. The reasons underlying this situation are complex. One important factor is the isolation caused by inadequate infrastructure as well as poor access to markets, and social or government services. The level of isolation varied across the project due to differences in the quality of transportation networks. Still, my sample should capture this variation and allow comparisons of how differing conditions affect the ability of families to cope and stay on the land. It is not easy to study the effects of poor infrastructure and services on tiunover, since the project has existed 15 years and conditions have improved. During the early years, the project was much more remote. This fact probably accounts for much turnover. With the paving of the BR-364 and continued improvement of other primary roads, areas that had been very inaccessible were less so in 1995 and 1996. This meant that areas that suffered high turnover because of poor access could have had these past conditions

PAGE 111

100 masked by more recent improvements. However, even if the same conditions did not exist, there should have been remnants of the effects of earlier conditions in the patterns of settlement today. At the outset, I thought that level of access and quality of services provided to Peixoto's residents would play a major role in defining project conditions and families' decisions to stay on the land. Areas that were more accessible and that had better services would have less turnover. Resource-poor families lack the capital and surplus labor to dominate surroundings and provide all they need, especially on the fi-ontier. They are dependent on government services, transport, and social infrastructure for support while establishing their farms. Without the ability to get crops to market, deal with government institutions, or meet basic family needs, it was not possible for colonists to stay. As a result, they were forced to sell out or choose to move elsewhere. The quality of transportation is a strong determinant of whether they could meet these needs. To test this hypothesis, I decided to use the "settlement time" measure. This measure provides more information on individual cases as opposed to just knowing whether or not the lots had turned over because it indicates how long the families have been able to stay on the land. However, this measure can only be used on the sub-sample of 146 cases. The limitations of the smaller sample are compounded by the cluster sampling method which provides observations from a small number of locations. This method placed priority on accurately measuring turnover in the project. Important variation within the project may have been missed or underrepresented since only eight

PAGE 112

101 clusters were used.^ While the sample design may not have been ideal for testing relationships between lots, it will have to do. Even with these limitations, it is apparent that infrastructure and service conditions do affect turnover. I created a regression model to test my hypothesis that the isolation caused by poor quality of infrastructure and services impedes colonists' homesteading efforts. This model attempts to explain differences in "settlement time" (ST) observed for each case using three variables that described access to transportation, road quality, and familiarity with the rustic conditions on the frontier. The model is simple and does not attempt to include all variables driving landownership turnover. The first variable (access to fransportation) measured whether or not a market truck visited the ramal. Some parts of Peixoto are served by trucks sent by government agencies to haul cargo to market (either monthly or bi-weekly). Usually, these trucks are the only affordable way to get produce to market, especially for families far from the few bus routes in Peixoto. The families with market truck service were expected to be more likely to stay on their land. The second variable asked whether the road stayed open throughout the year. Some roads cannot be maintained during the rainy season, and for months the only access is by foot or animal. When this happens, it makes life more difficult for the struggling colonists. Families on roads that are open year round are expected to stay longer. o Nova Aldiea and Ramal do Bee were excluded from this analysis due to the small number of interviews in each cluster (3 and 5 interviews respectively). There were few interviews because many of the lots had been incorporated into large ranches.

PAGE 113

102 The final variable asked whether or not the head of household had ever lived in a seringa! on a rubber tapping homestead. I expected that those who had lived on a seringal would be more tolerant of the isolation found in colonization projects. Life on a seringal usually means living far from other families, with no roads, and infrequent, substandard government services. Coming from such a rustic background, life in Peixoto would not seem so bad. Conversely, for others with urban backgrounds and especially those from regions outside of Amazonia, Peixoto 's conditions would be more traumatic. The model is listed below. (p=.002) (p=.12) (p=.006) ST = .82 + .27(Market Truck) + .15(Year round access)+ .02(Lived on Seringal) R2 = .24 Signif F = .0028 In this model the first two variables were statistically significant ("market truck" and "lived on seringal"). The other variable (year round access) was of borderline significance. This model demonstrates that by knowing these three facts about household our prediction of settlement time is 24 percent better than a prediction based solely on the mean settlement time for the population. This is a reasonable result considering the limitations of the data I could collect. The ideal situation would have been to visit each lot early in the settlement and observe the conditions, then return later and see how many families remained. I also could have then designed a stratified sample that better represented variation in transportation infi-astructure and services in the project. Because the survey was only conducted in 1995, 1 was not able to accurately observe the effects of change, especially in

PAGE 114

103 transportation infrastructure in the state. With the paving of the BR-364 highway, areas on the far eastern half of the project that had been the most remote became much more accessible. After pavement, transportation opportunities increased, which improved the ability of remaining families to market and get other services they lacked. Families repeatedly told me that a major hardship came from the isolation in the project and the poor roads and infrequent transportation. This made it difficult for them to get in and out of the project, which in turn made it hard to sell crops, buy needed commodities, or get to education, healthcare or technical services in the community. I knew that conditions varied across the project some places are very remote while others can be visited easily. Differences in settlement also are apparent when one enters the project. Settlement is patchy: some areas are all pasture, other areas have prosperous farms, still other areas appear poor and have abandoned lots. The settlement pattern observed today in Peixoto is a direct result of the turnover that has taken place in Peixoto. Comparing differences in landownership turnover (and other conditions) between areas should provide insight into what is happening in Peixoto and how it has developed as a resettlement project. Current Settlement Patterns in Peixoto The high frequency of landownership turnover has affected settlement in the Peixoto project. Furthermore, there are dramatic variations in the level of turnover that has occurred at different areas within the project. As a result, a settlement pattern consisting of three distinct zones has developed. In this section, I will briefly describe the variation in landownership turnover, discuss the settlement pattern that has resulted, and examine the significance of each of the zones.

PAGE 115

104 As expected, turnover did not occur uniformly throughout Peixoto. There are dramatic differences in the proportion of original settlers that have remained on the land in each of the clusters. While one community had 56 and another 50 percent of the original occupants, the remaining clusters averaged only about 20 percent of the original occupants, with the lowest cluster having only 3 percent. The range of variation between clusters can be seen in the figure 3 below. Original Residents in Ciusters 60 50 S) 40 a I 30 S ^ 20 a. 10 o m GO (D (D n •c '5 c m ro CD O CO ^ > a> sq Dc z < OT < K o 1980 1981 1982 1982 1982 1984 1985 1987 1989 1993 Sampled Clusters Figure 4.3 Percentage of Original Residents in Sampled Clusters in 1995 Peixoto was settled in stages and has expanded through time as new glebas were surveyed and settled. When the clusters are arranged by the age of settlement, there is a clear pattern. In general, new areas have more original residents. Nova Aldeia, and Ramal do Bee two of the oldest, had only 3 and 5 percent of their original residents respectively. Triunfo, the newest area in the sample, had 56 percent of its original residents. Intuitively, it makes sense that newer areas have more original residents, while older areas have fewer. There simply has been less time for people to leave newer areas.

PAGE 116

105 However, there is more going on in Peixoto. Some communities like Souza Araujo have maintained stable settlements even though they were settled early. Souza Araujo was settled in 1981 but retains 50 percent of its original residents, the second highest percentage in the sample. Souza Araujo is clearly an exceptional case. The clusters where family farms persist^ fall into distinct groups of high and low turnover communities. When I compared the mean number of previous owners per lot for each cluster (using a one-way ANOVA and Duncan Test), the clusters form two distinct groups that are significantly different (p= .1). There is a low turnover group consisting of Triunfo and Souza Araujo and a larger high turnover group consisting of Bigode, Linha 16, Linha 14, Linha 8. Two other communities, Iquiri and Linha 24, were not significantly different from either group. The high turnover communities Bigode, Linha 16, Linha 14, and Linha 8 are located on the far western portion of the project and are not as accessible as Souza Araujo (even with improvements to BR-364). This grouping indicates that the variation that is observed is not just an impression, there is legitimate grounds for defining high and low turnover communities in the project. Low turnover is, at least partially, tied to the age of settlement, but this does not account for all the observed variation. Souza Araujo was one of the oldest in my sample. Triunfo is the newest area. The second and third newest areas, Iquiri and Linha 24, fall in a middle ground between the high and low turnover communities. These recent settlements are all inaccessible areas but because they are newer, many families have yet to turn over the land. Turnover will likely increase as these sites mature. Triunfo is a good example, considering the amount of turnover that has taken place in two years.

PAGE 117

106 While Triunfo is the community with the highest percentage of originals, it was only settled in 1993. This means that approximately 44 percent of colonists have left in just two years! This community is located in a former reserve area that had been swamp between the Seringal Triunfo and Peixoto. No roads enter this site; instead the colonists have cleared an 8 kilometer path through the forest into the community. A logger had agreed to improve this "road" in exchange for some of the timber from these colonists' lots. However, when I visited Triunfo it obviously had a very tenuous link to the outside world. Although the communities break into two distinct groups, this division is less clear when we compared the mean settlement times for current residents. This indicated that sometimes, even though turnover was high, later residents were able to establish themselves on the land. Again this is apparent in the data. When the clusters' estimated means for "settlement time" are compared (using an ANOVA again, two distinct groups were present (p= .05). However, this time more clusters fall into the gray area. Triunfo again differs from all the clusters except Souza Araujo. Souza Araujo is only different from Linha 8. This means that while clusters differ in how much turnover has taken place, in some of these places later residents have been able to stay on the land. After suffering through the initial periods of turnover, settlement has stabilized in these communities and later families have maintained their hold on the land. These tests demonstrate that there are high and low turnover areas within Peixoto. This variation shows that turnover behavior has created a distinct pattern in Peixoto. In These comparisons exclude Ramal do Bee and Novo Aldiea since land concentration

PAGE 118

107 some areas when turnover occurred it led to land concentration, as in the two clusters excluded. In a few areas families have held onto the land, like in Souza Araujo. In more remote areas, families have had a hard time remaining on the land, but as they turn it over they are replaced by other migrants. Settlement Zones in Peixoto Landownership turnover in Peixoto has resulted in a distinct settlement pattern. In some places few family farms remain, having been replaced by expanses of pasture. In other areas, family farms are doing well and are maintaining their hold on the land. In still other areas, many families who settled have left and others have taken their places. The pattern that has developed forms three zones: a ranch transition zone, and two family farm zones, one with stable settlement, and another with unstable settlement. These zones result from conditions that impoverished farms within the project, as well as macro economic trends that drove land speculation in Amazonia in recent years. In the ranch transition zone, family farms have practically disappeared, and have been replaced by pasture as land is concentrated into ranches. These areas have changed from a landscape of smallholdings to one of large properties. Generally, they are found in two types of locations: areas that border existing ranches around and within Peixoto, and areas along highways. Two clusters from my sample fit this description. Nova Aldeia is on the eastern edge of the project alongside a large ranch owned by a Brazilian cigarette manufacturer, and Ramal do Bee, which borders ranches left within Peixoto after the land dispute of the 1970s. In a single year, one of the ranchers near Ramal do has changed these areas considerably. They will be discussed later.

PAGE 119

108 Bee bought 16 lots along the road. Both of these were settled very early in Peixoto's history. Surprisingly, many of these areas are very accessible today. I had expected the most accessible areas to have the lowest turnover. However, they have become the most concentrated. One thing I did not fliUy consider when starting this research was how speculation during periods of high turnover would lead to land concentration. I did not ignore the fact that speculative investment and concentration were there, but 1 did not fully consider how one led to the other. Many of the concentrated areas in Peixoto were settled first, and because they were easy to get to, INCRA could document and title these families early. Because the land was titled and accessible, its value increased and it became desirable to large landowners hoping to expand their holdings. Therefore, it was tempting for colonists to sell out to gain capital for investment in land in other areas. More than once, I met individuals in Peixoto who told me they had owned land near the BR-364 but had sold it after receiving title. They said they knew it would be easy to purchase land from other colonists farther inside the project and had additional capital to invest in the new farm. I do not know with certainty that concentrated lands were first titled by smallholders but this is a reasonable assumption. Mahar (1981 : 51), in describing colonization projects in Rondonia, asserts that colonists settled early were the first to receive titles. The first colonists always settled in the most accessible areas. Later arrivals were settled away from primary roads and received their titles later, and so on. This created a concentric pattern around roads and administrative centers with the best documented land near the center and the less securely documented land out in the

PAGE 120

109 periphery. Once titled, land rights were easier to transfer and land increased in value, particularly because speculative investors wanted to buy titled land so their ownership could not be challenged. If a rancher bought undocumented land they could not be sure to hold it against the challenge from INCRA. In Peixoto, these same areas were attractive because they were easy to visit from Rio Branco, the home of many speculative investors. However, it is hard to gather information about this topic. Most large landowners are reluctant to talk about land they have acquired within Peixoto since such ownership may be of questionable legality. In the zone of stable family farm settlement turnover was low and original settlers have remained. This is how the project was supposed to work. However, such cases are rare in Peixoto. Souza Araujo is a great example. Families that were originally settled have maintained their hold on the land and have developed productive homesteads. Triunfo also is currently in this group but it is losing residents fast. I also would include in this group areas where turnover was high but settlement has since stabilized. I believe that in some high turnover areas conditions have improved enough that later arrivals have been able to remain on the land. Evidence for this can be seen in the inter-cluster comparisons (ANOVAs) of settlement time. When comparing the clusters by the frequency of turnover most fell into two distinct groups. However, when comparing clusters' mean value of the "settlement time" measure, more clusters fell into the gray area. This indicates that even though some areas have had much turnover, the amount of time current families have remained on land is not statistically different from low turnover communities. Clusters that may fit this description are Linha 14 and Linha 16. I will argue that these clusters have become part of the stable zone, because

PAGE 121

110 conditions have improved enough that families can meet subsistence and cash income needs and may develop long-term homesteads. The zone of unstable family farm settlement consists of high turnover areas that are still occupied predominately by family farms. The families struggle to meet both their subsistence and cash income needs. These areas often have poor access, making it difficult to sell crops and buy needed inputs. They are usually in remote areas or recently opened, undeveloped areas. Because they are usually very impoverished, farm families move out at high rate and are replaced by other families. These are areas that may be the most susceptible to future concentration. In 1996, 1 decided to revisit 4 communities in Peixoto to better understand the pattern of turnover and its causes. Since I am interested in family farms, I focused on areas where small farms are concentrated, and did not include any areas from the ranch transition zone. As an example of stable settlement, I chose Souza Araujo, the old community with a high percentage of original occupants. As contrasts, I chose three semi-stable or unstable areas: Linha 8, Linha 14 and Linha 24. Linha 8 and Linha 24 provide interesting contrasts because they are part of the same road network. Linha 8 is close to the asphalted BR-364 highway and Linha 24 is at the other end of the road network and more remote. Linha 14 is a poorly maintained feeder road between Placido de Castro and Acrelandia. It has suffered a great deal of turnover but family farms have continued to occupy the road. There are many differences between the selected communities besides their differing turnover rates. Differences in infrastructure and service conditions may account for some difference between groups, but this is unclear because conditions have changed

PAGE 122

Ill through time. As conceptualized, infrastructure and services encompassed health care, education, finance and technical assistance. However, in this discussion 1 will just focus on accessibility and transportation because they provide easily illustrated examples. The level of accessibility for each community is not the same as it was when the project began. However, it can be pieced together from what is known of the history of road building in Peixoto. In the early 80s, when Peixoto began, Souza Araujo may have been the most accessible of the 4 communities. This community was on a major thoroughfare, the BR-317 highway. The BR-317 was built through the forest in the 1950s, and although in 1995 it still was just a dirt lane, it received regular transit. This road had problems but it was definitely better than interior access-roads. Linha 14 and Linha 24 still have severe problems due to terrain that makes keeping the access road open during the rainy season exceedingly difficult. Sometimes even private transportation has trouble entering. Souza Araujo was also closer than any of the other communifies to Rio Branco, the state's principal administrative and commercial center. After 1989, when the BR-364 highway was paved, the situation in Peixoto took a dramatic turn. The interior of the project was now more accessible. This dramatically improved access to areas like Linha 8 that are close to this road. Conversely, the BR-317 highway is still a dirt road that closes periodically during the rains. In 1995, even though it was sfiU located on a dirt road, the transportation conditions on Souza Araujo were better than any of the other 3 sites. In Souza Araujo two buses come and go to Rio Branco through the community each day. There is also more truck traffic in general, and the local farm association has a vehicle. The other areas only have buses that pass near the communifies. Linha 8 and Linha 14 residents have to

PAGE 123

112 walk to the mouth of their roads to wait for the bus, usually less then 10 kilometers. No regular transport enters their road or picks up produce at their doorsteps. Linha 24 has no regular transport. Once a month, a government truck from INCRA comes but it is not dependable and stops during the rainy season. As a result, in the rainy season they must walk 10 to 20 km to reach transportation. The farm lots on Souza Araujo have remained in the hands of family farmers. Even when land does turn over, families are replaced by new owner occupants. Linha 24 and Linha 14 have suffered more turnover than Souza Araujo but also in these areas they generally were replaced by family farms. However, absentee owners have purchased lots in each community. Owners are merchants or public employees from urban areas. In most of these cases land is not intensively used by the absentee owners, just held as investment. On Linha 8 there are absentee owners and a growing trend that does not bode well for family farm settlements. Land concentration has already begun on Linha 8. Between Linha 8 and the BR-364 highway a small ranch has developed. This landowner bought 4 adjacent lots in the Linha 8 community a few years before my survey. The land is fenced off and is covered by expanses of pasture. Early in 1995 the rancher had bought 7 more lots, however INCRA claimed four of these before the rancher could consolidate his ownership. INCRA settled landless families. In 1996 only one of these families was still on the lot. If this process continues, Linha 8 could follow Ramal do Bee and Novo Aldeia in becoming highly concentrated. Conclusions In this chapter I set out to describe and explain landownership turnover in Peixoto. I was able to gather a representative, unbiased sample of the entire Peixoto

PAGE 124

113 project. As I had expected, turnover is very high in Peixoto. Only about 20 percent of the families first settled by INCRA remain on their land. While the average lot was on its second owner, almost a third of the lots have had three or more owners. Despite these indicators of high turnover, it does not mean that small farms have been completely unable to survive in the project. The measiu-e of settlement time indicates that many of the late arrivals have held onto the land for prolonged homesteading attempts. This turnover is driven (at least partially) by isolation and the difficult conditions found in the project due to poor transportation infrastructure. The level of turnover varies within the project. This has led to a distinct settlement pattern, characterized by sharp differences in settlement density. Turnover, combined with a speculative investment climate of the 1980s and 1990s, led to concentration of large portions of the project and the exclusion of family farms. If these areas continue to expand, it does not bode well for Peixoto as a settlement project. Few areas have provided conditions that allow long-term settlement (like Souza Araujo). I suspect that some low turnover areas like Triunfo will stay that way. More typical are areas where there is high turnover.

PAGE 125

CHAPTER 5 RURAL MOBILIZATION AND FAMILY FARM SETTLEMENT SUCCESS Souza Araujo is a unique community in Peixoto. It is one of the rare cases where Peixoto seems to have worked. Most of the colonists settled here have stayed on the land from the beginning, and therefore, the community has a low incidence of landownership turnover. One characteristic of this community that has helped these farmers maintain their hold on the land is their strong activist organization. This organization is the farmer's association called the Associagdo Rural Libertador (ARL). The organization has allowed these families to lessen the impact of poor access and to collectively respond to common constraints. The ARL has helped families maintain their hold on the land because it provided services that decreased the impact of isolation from urban markets, acted as a rallying point for collective action, and served as a mutual support network. This chapter will describe how the ARL developed and how it functions to help farmers stay on their land. Farm associations exist in other communities but they do not function as well. In communities lacking organizations, individual families bear the brunt of poor infrastructure and services by themselves. As a result, many families give up. However, establishing effective organizations is not easy. If activist organizations can be a successful strategy for colonist families to hold onto their land, why are such groups so uncommon? 114

PAGE 126

115 Effective local organizations have not developed for two general reasons. First, the Peixoto project developed in a way that dampened the tendency of rural people to mobilize in activist movements. The project's role was to placate growing unrest among displaced people. Second, once Peixoto was formed, the Brazilian government did not want the colonists to form a mobilized opposition to the state. Therefore, the government formed rural groups it could control and thereby undercut the ability of rural movements to get started. I will illustrate this argument with examples from my 1995 and 1996 research. Ultimately, both of these trends were the results of broader populist efforts by the state and federal governments to maintain control over rural peoples who were increasingly unwilling to accept the direction of development policies. Organizations can support rural families but the circumstances in which they form and actually fimction occur infrequently. Souza Araujo's organization developed because of the convergence of key factors: a strong cohesive community, the presence of a charismatic leader, and contact with outside organizers. In most areas of Peixoto one or all of these factors are missing. Souza Araujo's Settlement The area I call Souza Araujo is not a typical INCRA settlement area. It takes its name from the state's Souza Araujo leper colony located about 20 kilometers east of Rio Branco. Many of Souza Araujo's landowners lived in or near the colony. In the 1970s, a number of patients at the hospital no longer needed treatment and were free to go. However, they were poor and no land was available near the Souza Araujo colony. Landless families were squatting on this land. Some were families of Souza Araujo's patients who had moved in from surrounding regions to be closer to their loved ones.

PAGE 127

116 Others were families that had been pushed off land in the interior of the state, or had come to Acre in search of land. Some area had to be found to resettle these families. At this time, an Italian Catholic volunteer teacher, Pia, was working at the Souza Araujo leper colony. She was aware of the plight of the landless families around the colony and began looking for a solution. This was during the period in which INCRA was waging a battle to start the PAD Peixoto project. The project was generating publicity but had settled few families. Pia saw this as an opportunity to relocate patients and their families who were occupying land near the colony. She approached INCRA and proposed resettling these people in the new colonization project. She evidently had contacts within INCRA. Because of this, INCRA gave her the opportunity to resettle the families in Peixoto. INCRA designated areas at kilometers 58 and 52 of the BR-317 highway north of Rio Branco for this settlement. The area in my survey research was the community at kilometer 58. I call this area Souza Araujo because that is how it is listed on INCRA maps. However, residents referred to it as kilometer 58, probably to avoid confusion with the leper colony, and to avoid the stigma associated with leprosy. The families Pia selected for settlement were already a fairly cohesive group: they knew each other, worked together, and shared common experiences, especially from time spent around the leper colony. Since she knew the families, Pia probably selected people she believed had the best chance of remaining on the land. Many of the original settlers were ex-patients and their families, but not all. Some ex-patients did not want land, so they were replaced by other landless families living near Souza Araujo. Also, a few

PAGE 128

117 rubber tapper families that lived in the area expropriated for the settlement were given lots. Most of the families had come from seringais and had no official papers. Pia helped the families get the necessary documentation they lacked so they could meet INCRA's requirements. Pia organized the families and with INCRA's assistance, they spent two or three months of 1978 measuring the lots and cutting trails. The first families were settled in 1978 and 1979. INCRA helped the families move and provided staples to get them through the first year. However, Pia's involvement did not end there. She went a step further and moved out to the settlement with the families. By moving with the families, Pia supported their homesteading attempts in a variety of ways. She helped settlers directly as they worked to establish their farms. She had a vehicle and used it to carry supplies out to the communities. She was a great leader and encouraged families during the rough early years. Evidently, she was very enthusiastic and energetic; always ready with a pep talk or a scolding if she heard someone was considering moving. One of Pia's greatest legacies in this community is that she formed a Christian base community (CEB). The CEB Pia formed is important because it built up a unified organization that grew into an activist group. Later this group linked with NGOs like the CPT that helped the group develop projects. Associagdo Rural Libertador In addition to the CEB on kilometer 58, Pia worked with CEBs on neighboring roads. They all began to branch off into community development projects. It was difficult for colonist families to get all the commodities they needed on their farms. In

PAGE 129

118 the early years, round trip visits by colonists to Rio Branco could sometimes take days. Even if a family member made the trip, items such as gasoline or oil could not be carried on buses. The CEBs used the supplies Pia brought out and organized small communityrun stores (vendas comunitarios). The stores supplied colonists with staples that were hard to get in the project or expensive if bought from marreteiros (kerosene, sugar, salt). Members put in some capital to buy stock. All the residents could buy from the store but only members could get credit. At least four stores were formed, one each at kilometers 52, 58, 70 and 80. With Pia's help the residents organized a small association to manage the stores. Pia stayed with the settled families for 4 or 5 years through the early 1980s. She helped settle the farm families, helped them through tough early years and set in motion the process that built into a large producer group that helped the families stay on the land. In 1995 they still talk about her fondly as if she had just left. When Pia left, the CPT placed two Italian volunteers in the community. These volunteers were cooperative specialists who trained base community members to manage the small stores. The members of the CEBs and store associations formed a farm association called Associagdo Rural Libertador (ARL). It began in 1984 as a large group based at kilometer 52. In 1987, as the association developed, it broke into sub-sections with centers at kilometer 58 and 75. ^ Membership is open to anyone who can put in the required capital and attend the prescribed number of meetings. The group has no definitive religious or political affiliation, but most members are Catholic and members of the Partido de Trabalhadors (PT). The group is widely viewed in the community as a

PAGE 130

119 PT organization although members deny this (and to the consternation of some members the pubUc refers to their trucks as PTzinho and Ptzao--"little PT" and "big PT"). With the help of the volunteers and confident from their success, the ARL members gradually expanded their activities. The association members approached the CPT and with their assistance received a loan from the Catholic diocese to buy farm machinery. They bought the machinery and paid off these loans with produce. After that success, they received additional loans for a small Toyota pickup and later a larger Ford truck. Again, after a few years, the families paid off their debts. Because of the organization's strong structure and experience, they were able to manage these agreements. This process differentiates them from other government associations that I will describe later. The ARL built up to these projects gradually. Once they were organized they took on more responsibility. The association's trucks provide vital transportation service to the community (to both members and non-members). They worked hard for the trucks, and because they earned them, they have taken care of the trucks. The ARL has rules for the trucks use. These rules assure that the trucks continue to operate and do not serve a single clique. The larger truck travels on a regular schedule to each of the association's areas. The association charges fees that cover the operating costs and pay a driver, and they save money to cover ftature maintenance costs. They are proud of their vehicles and use them to provide important service to the community. During the rainy season these trucks are often the only form of fransportation available to local families. 1 Informants told me that the section at kilometer 75 had been the most active but it has died out since its primary leader has moved away.

PAGE 131

120 The ARL also has helped decrease isolation by lobbying government. For example, on two separate occasions, the lack of maintenance threatened the BR-317 highway. Each time, Souza Araujo's association drew residents from surrounding areas together and organized protests that blocked the BR-317 highway for days. As a result of these protests, they were able to keep traffic flowing to and from their community. The first time was during the term of Governor Edmundo Pinto in the early 1990s. Protests were organized again in July of 1996 while I was in Acre. These incidents illustrate the strength of this rural organization. The unpaved BR-317 provides access to the northwestern section of Peixoto that includes Souza Araujo, as well as numerous ranches. Furthermore, it is the only overland route to Boca do Acre, a town in Amazonas. If the road is not maintained, it quickly deteriorates. During the rainy season (December March) the road can close down for days and no access is possible. This happened during the early months of 1996. Something had to be done and soon, ARL members told me that even non-members started asking if and when the ARL would start another blockade. For the first protest, ARL members were the outspoken leaders and organizers of the protest, which drew a wide constituency of supporters from surrounding areas and gained statewide attention. However, informants said they were severely criticized from the political right which depicted them as PT instigators. This time, they said they would be willing to organize another, but other political parties had to get involved and they needed high profile spokespersons to serve as figureheads. The state representative (deputado) Sergio Petecao (PMN) agreed to act as intermediary between the protesters and the state. The first rally attracted two thousand protesters (Jomal Gazeta 1996).

PAGE 132

121 Representatives from the protesters met with the governor, but the government refused to release funds so they planned the strike for July. During the first week of July 1996 the protest started. They built a barricade at the first bridge north of the BR-364 highway and stopped all traffic on the road (A Gazeta 1996). This was during the dry season at the height of use. This also was a convenient time for the farmers because it was right after harvest but before major work of clearing and burning fields for the next year's crops. The protest was broadly supported by the public. Protesters included both ARL members and non-members who lived on the road. In 1996, the protesters even gained the support of ranchers that lived along the road, probably because the actions were not openly affiliated with the PT. The ranchers depend on the road too, so were not disinterested parties. Every day a different rancher donated a cow to feed the protesters. Informants said even Petecao manned the barricades and slept overnight with the protesters. After a few days, the state finally gave in and agreed to release funds to repair the road. Why Has the ARL Worked? In the case of Souza Araujo the farmers have been able to stay on the land because of their grassroots organization. Why did Souza Araujo's residents organize and how have they benefited from this organization? There are a number of factors of this group that contributed. First, idiosyncrasies of the community: they were cohesive and shared experiences so were likely to work together. Second this was combined with a charismatic leader, capable management, and strong outside organizers. Finally, they provided a means of responding to common problems (material benefits) such as the isolation of Peixoto.

PAGE 133

122 It is a cohesive group that provides members with mutual support and provides members and the wider community with services that decrease hardship and isolation. Such a group did not develop overnight. The households settled in Souza Araujo knew each other and were familiar with Acre. At the time of settlement residents were selected and moved as a cohesive group (however, not all residents participated in the ARL). When the ARL was form it benefited from a strong leader/organizer who started the CEB, helped organize the association, developed their management skills, and put them in contact with the CPT. Such circumstances did not occur in most places. Most locations in Peixoto lacked these advantages, and as I will describe shortly, government policies diminished the chances that such organizations could develop on their own. The ARL organization is important to the community for a number of reasons. Through the farm association the farmers collectively respond to common problems. The association's machinery allows members to process grain on site, saving labor and avoiding the need to send cereals elsewhere for processing. The association's stores lessened the impact of isolation by providing difficult to acquire staples and other goods at affordable prices. The store cut down on costs and isolation of the Souza Araujo residents, decreasing the need to travel to urban areas to buy necessities. When residents do have to leave the association provides important transportation services with their truck, primarily freight. During the rainy season the truck runs even when buses do not. This transportation has given the residents greater security because they know exactly when the truck will come and what it will cost. They know they can get to and from the market throughout the year.

PAGE 134

123 The organization also helped the community lobby for outside support. The group was able to lobby and attract needed support like loans and technical support. This assured their continued access to Rio Branco. By confronting common problems due to isolation and poor market access, Souza Araujo's association allowed more families to stay on their land. Why Are Associations Like ARL So Rare? Activist organizations are rare in Peixoto and the population is not mobilized. There are other CEBs, farm associations and even cooperatives in Peixoto but few have the same level of participation, range of activities or degree of effectiveness as the ARL. None of the other communities in my sample had farm associations that functioned in the same way as ARL. I knew effective grassroots organizations were not common from previous visits to Peixoto and from conversations with personnel from EMATER and workers from the CPT. The lack of grassroots activism in Peixoto has to do with two general factors: first, as I will explain shortly, Peixoto by its very nature appeased activist sentiments, thereby lessening motivation for collective action; and second, the government formed rural organizations that could be used to control rural resistance. Government groups not only undercut activist movements but usually were fiascoes that discredited collective action and organization as a viable strategy in the eyes of most residents. These factors are explored in more detail below. Appeasing Rural Resistance with Land Distribution As I have been arguing, land distribution programs in the Amazon were typically populist programs that attempted to cut down on organized opposition to development policies. Peixoto generally diminished the motivation for families to turn into rural

PAGE 135

124 activist groups. There are 3 reasons wliy this took place: 1) the distributed land eliminated the major demands that rallied rural movements in the Amazon, 2) the less militant tended to settle in Peixoto, and 3) the settlement process hampered the development of cohesive communities. By creating Peixoto the government gave in to the primary demand that had rallied rural people across Amazonia: the project distributed land. Throughout Brazil, a major catalyst for rural social movements has been the desire to own land. The Peixoto project was a direct response to mounting tension caused by large scale land sales and expulsions. Once families were settled they could feel some satisfaction that they had won a major concession from the government. Furthermore, The families that agreed to move into Peixoto were less militant. In Chapter 3, 1 discussed how radical STRs around Xapuri opposed cooperation with the government. The STRs discouraged members from accepting land in Peixoto because to do so weakened the claims of posseiros. Boycotting the project also was a way to protest the failure of state agencies to fulfill obligations to families already settled (Bakx 1986). Families that accepted land either were not involved in STRs, were not committed to the cause, or simply ignored the directives when offered a chance at land. Finally, the way families were settled in Peixoto diminished the formation of cohesive communities from the outset. Sometimes related families, or neighbors, were settled together, but more often they were dispersed across the project. Typically, families found themselves settled among strangers in groupings of families from different regions and backgrounds with little in common. Some communities were divided between Catholic and Evangelical factions that barely spoke, let alone worked together. Many

PAGE 136

125 Evangelical groups in Peixoto limited activist group formation because they prohibited church members from becoming involved in worldly distractions. The development of tight-knit communities was further hampered as turnover began to take its toll. Because of turnover, the composition of communities continued to change. It was difficult for families to work together when neighbors were continuously moving in and out of the area. For these reasons families usually found themselves confronting problems alone even though families that shared the area with others that had many of the same problems. Forming Associations to Undercut Rural Mobilization Why would there be a need to undercut activism in Peixoto? This is because the masses of unemployed landless around cities, and posseiros under siege in the countryside were becoming increasingly mobilized. During the 1970s and 1980s STRs and more specifically the rubber tappers' movement successfiilly challenged the government's policies in the state through local mobilization and international alliance building (Schwartzman 1989; Bakx 1990; Schmink and Wood 1992). Because of this threat, projects like Peixoto were developed to appease the militant rural population. To further assist farmers and diminish activist sentiment, the government organized alternative groups (associations) that they could control. As I will explain, these organizations normally were ineffective at providing social assistance but they did drain away support for the more activist organizations. Whether or not diminishing social activism was a deliberate intention, the government-controlled groups provided a seductive (yet flawed) alternative to colonists. These associations contributed to an apathetic or even distrustful view of collective action.

PAGE 137

126 In Peixoto, the government organized many farmer associations that it was able to control. These groups were formed everywhere by EMATER, INCRA, the Secretaria de Agricultura (SDA), and even municipal officials. However, Acre's state government formed most of the farm associations found in the project. At one time or another, most roads had associations but most did not function. These groups had little autonomy. They were controlled from above and created to serve the government's agenda. Often, they did not address specific needs of families living in the area. This was because government institutions found it easier to supply standardized technical support to groups of farmers rather than individuals. Government associations also were attractive to local politicians as a means of marshaling political support and buying votes. These associations did not function well as community organizations. They primarily served to disseminate information and handout agricultural inputs or machinery. Normally, little preparation was done to form these groups. In many cases informants recalled that government officials simply arrived in the community, called together residents, and told them if they formed an association they could receive assistance. To entice farmers to participate, the government offered material support including buildings, threshers, tractors, trucks and sometimes even complete saw mills. From the colonists perspective, the main benefit to joining these groups was this possible access to machinery and credit (which most farmers lacked). Once government officials had recorded the names of prospective members and had the group elect leaders, the officials would leave. No effort was made to train members to administer or manage the association. The resulting membership was usually haphazard with diverse people from

PAGE 138

127 broad areas. Such groups lacked cohesion, structure, or management abihty because they were established without the benefit of a preceding organization to serve as a foundation. The material incentives given to encourage participation rarely turned into longterm benefits. Repeatedly, farmers were attracted to government associations were attracted by the machinery and other goods. These offers distracted fanners fi-om the systematic failures of the project and the unfulfilled promises of roads, schools, and health posts. When material benefits were distributed, they did not last. Such fledgling organizations were not prepared to maintain the machinery or define rules for its use. They had no leadership, no cohesion, and mistrust was common. Members of these associations had invested little to receive the machinery and because these organizations were weak, there was no clear responsibility for taking care of the machines. Therefore equipment was used without maintenance until it broke down, then it was abandoned. Peixoto is littered with rusfing machinery distributed with no strings attached but also with no clear owner or means to provide for maintenance. When the machines disproportionately benefited one clique within the community, it produced conflicts that further undermined the group. Their poor performance probably does much to discredit these organizations. Because these associations did not grow out of local initiatives, or address local needs, they normally were inactive, had low participation, and soon disappeared. Members saw the association as a conduit for government aid, not as a local institution that would serve the interests of the community. An EMATER cooperative specialist estimated that hundreds of associations were formed in the manner described above but few actually functioned, maybe less than five. In many cases, they only existed on paper.

PAGE 139

128 In recent years there has been a resurgence of these groups because associations membership is required by BASA to quaUfy for some types of rural credit. The EMATER expert estimated that there were probably 100 to 150 associations like this in Peixoto. They exist solely for credit, and do not meet for other purposes and have no other activities. While the material benefits of these associations were attractive to Peixoto's residents, many families became disillusioned when the groups did not function well, or frequent conflicts erupted. Once machinery broke down, participation declined. As a result, residents doubted that such organizations could function and refused future participation without immediate material gain. Therefore, the formation of these government associations discourages fiirther grassroots organization. The government has little interest in nurturing activist groups that might demand promised support. By forming these weak groups the government is able to control the development of social movements within rural areas. Were these deliberate policies to undercut rural social movements? Leadership within Acre's CPT believes this to be true. They see the formation of government controlled associations as a deliberate strategy to undercut their work. Members of ARL also agree. When they were in the process of forming the ARL, the government started a well-stocked association closed to the community Dual membership was not allowed. Furthermore, police harassed the ARL truck: they stopped it frequently and fined the association for transporting people in the a truck without sides or roof (even though INCRA and SDA trucks lacked this amenities and still do). This competition ended after a few years when that government association collapsed due to an embezzlement scandal.

PAGE 140

129 Clearly, local and state politicians use these associations to organize support and buy votes. Furthermore, the federal government had controlled troublesome movements in the past by imposing vertical structures of leadership, in paternalistic unions (see Adriance 1995). A non-cynical explanation of why government associations exist and have failed would be because these were misguided, ill-conceived programs from the start. As described in Chapters 2 and 3, the lack of information and contact within Peixoto would make coordinating a comprehensive plan difficult if not impossible. While training or organizing are not enough to make an effective grassroots organization, distributing machinery cannot lead to successful development either. It is wasteful and can sow strife and conflict. Distributing machinery is easy; building organization and training members to assure that the machinery continues to function is a more difficult task. There are many committed employees of EMATER, INCRA, and other institutions who have a genuine concern for the people and wish to assist them, but they are seldom able to do this with the flawed government associations. While there were probably policies to undercut activism, there was probably not a long-term coordinated effort. Even if there were, it would have been difficult to guide such a policy in such a complex place as Peixoto. Case Studies in Failed Organizations The other communities surveyed in Peixoto were not as organized as Souza Araujo. Even though two had farm associations, these groups were struggling and were not as active or successful as the association in Souza Araujo. The lack of strong activist organization made it difficult for residents to respond to common problems. Each household had to bear the brunt of isolation individually.

PAGE 141

130 The other communities, because of their history and circumstances, are different from Souza Araujo. While none of the other clusters in my study had organizations as strong or active as the Souza Araujo's ARL, most have or had some type of association. Only one cluster, Linha 24, did not have any local associations. Most were inactive or did not function well. To explain this, I will discuss the circumstances of two other communities: Linha 8 and Linha 14. These other sites were not organized at the beginning of settlement, and turnover cut down on local cohesion. Furthermore, inefficient and failing associations discouraged organizing efforts. I will present two cases similar to Souza Araujo where organizations have not worked. Then I will discuss how some communities may pull together as long-term settlement develops. Linha 14 is a high turnover community located off the AC-401 road that links Placido de Castro to Acrelandia. This community has very poor access, and few transportation options. Those families that remain still struggle to stay on the land. It would appear to be a place ripe for organized movement to demand better support. This road has suffered from so much turnover, there was little chance for a strong community group to form. On Linha 14 there were two associations that divided the community, neither of which was as strong as ARL. One was a large government association, the other a small grassroots association that had developed out of a CEB. There was clear antagonism between members of each association. Some residents on the road wanted nothing to do with either group because they thought both were useless. The government association was based at the mouth of the Linha 14 on the AC401 road. Its membership was drawn from many of the adjacent linhas, and its

PAGE 142

131 membership had recently expanded in 1996, due mostly to farmers attempting to gain access to credit. When formed, the group received a number of buildings, machinery and even a sawmill and tractor. However, they had been unable to maintain this equipment and the mill and tractor no longer functioned. The association was controlled by government institutions, and it met whenever government representatives visited (EMATER, SDA, or representatives from the municipal government). Furthermore, many members appeared to support the political party that controlled the municipality. Informants told me the group was inactive and that when it was active, they did little but argue. On the other hand, members of the government associations referred to the small organization as a clandestine cell, not registered with the state. Most informants from my sample were members of the smaller association, concentrated at the end of the Linha 14. This was where access was most difficult, and the road's poor maintenance might be partly due to their support of the PT and opposition to local politicians. These farmers felt the state's association did not respond to their needs, so they organized this group. They were in the early stages of organizing, have some machinery and were expanding their association building. It was not clear how strong they will become. They were looking to establish links with outside institutions (the rural union and CPT). I will return to this group in a moment. On Linha 8, at the far northeastern edge of the project off the Granada road, there was no government association but there was an autonomous association. The association was called Capricho Ideal. It was similar to ARL but was not successful. This case shows that it was not just government associations that function poorly. There had been a CEB in the community but it served only as a Bible study group. Turnover

PAGE 143

132 among residents had kept the CEB from becoming a strong cohesive force in the community. Because of this, it had died out in the past because of lack of participation. In fact, during my visit the two most active families had been absent from the community. They had sold their land and moved away only to return later to live in the forest at the end of the road. The community lacked a charismatic leader. The Capricho Ideal association is an interesting case since a rural union organized the group. Its history is similar to Souza Araujo but, in this case, the organization has not developed the same way. In part this may be because turnover was already occurring when this group was organized. The union used a top-down strategy to organize and control its involvement that was similar to that used in government associations. Rather than focusing on training members to administer and manage the organization, the union's focus is on distributing aid. The union organized associations throughout Acre to provide and manage standardized development projects. Because of this, the group has fallen into a trap similar to that in government associations. They distribute expensive machinery and equipment as part of elaborate development projects. Because emphasis was not placed on organizational training, Linha 8's association members are apathetic and disorganized. As a result the organization functions poorly. Members wait for projects to be brought in from outside the community. A number of members could not provide details of how the group functioned (or even when it met) While I was there in 1 996 the union was building ponds for a complex fish farming project. This was a standardized project that it was implementing throughout the state, rather than a project that focused on specific needs of individual communities. All money for this project went to build fish tanks and buy equipment including a truck.

PAGE 144

133 Apparently, this was done without a feasibihty study or training of members to manage the project. The project had a vague plan in which members will pay back the investment, then use their income to expand the project. Few members understood how the project would function. They were also suspicious that some members were using the project's resources for their own benefit. Fish ponds were being built first on the land of principal leaders. Two years earlier these same people also had received similar fish ponds but the dams had not survived the rainy season. The colonist leader who was the union representative has moved into a nearby town and uses the association's truck as his own personal vehicle. It appeared that the project would not succeed and the association would not last. Extended Families and Community Cohesion In some locations where family farms remain, it appears that communities are becoming more cohesive. I believe this is because extended families are taking up land and they are providing unity that may allow community organization to grow. If, as the project matures, resettled families stay longer in some communities, the communities may develop organizations that will help them maintain their hold on the land. One way this may happen is when extended families move in to take over available lots. In some communities I visited in Peixoto, extended families seemed to be forming the nucleus of organizations. The expansion of extended families is one resuh of land ownership turnover. As lots are turned over, families that stay invite relatives and fiiends to migrate in and buy up the available lots. These extended family networks provide reciprocal support, and labor

PAGE 145

134 exchange between relatives and friends. These family groups provide a level of cohesion that makes collective action more likely. On Linha 14, while turnover is very high, many of the owners who migrated in later have been able to stay on their land longer. In the previous chapter, the average settlement time for residents of Linha 14 did not differentiate this community from others. This suggests that, after a period of high turnover later residents have been better at holding onto land. As I got to know the families living on Linha 14 I realized that three large extended families accounted for the majority of occupied lots. These three extended families formed the bulk of the CEB and the nucleus of the grassroots association. Leaders from these families were lobbying to get more resources for the group but it was developing slowly. They had formed links to the CPT and local STR but had not received aid while I was there. If these families succeed, this may help them to improve conditions in the community. Another case I am aware of is on Linha 12 on the Granada road. This community is not part of my sample but, being adjacent to Linha 8, it makes an interesting comparison. This road also suffered from high turnover but two large extended families have moved in and taken over many of the lots near this road. They come from two states in the south and center-south: Parana and Espirito Santo. The Espirito Santo families moved directly to Acre after one member first bought a lot on the road and saw opportunity for the others. The families formed a CEB. One member of these families had been active in union movements in the south and has become a charismatic leader of this group. They formed a farm association and were originally affiliated with the Capricho Ideal associafion (that also includes Linha 8). However, after a couple of years.

PAGE 146

135 they became disillusioned with participation in the large group. They said it was not responding to their needs so they pulled out and became an autonomous association. They formed their own group called Novo Ideal. They have maintained contacts with NGOs and EMATER as they seek out assistance. They have recently developed their own innovative agroforestry project and should be able to succeed in maintaining their hold on the land. Conclusions The Souza Araujo community is an example of an area in Peixoto where the goals of colonization worked; families were settled and were able to stay on the land and develop farms. One factor that helped these families stay was their strong farmers association. It gave the residents the means to collectively respond to isolation and poor access to outside world. It provided material support network and a means of lobbying the government. It is not easy to identify exactly why some communities organize and others do not. Key factors seem to be a cohesive community that can work together to solve problems common to all, and capable leadership to channel organization activities. Often, outside institutions play a role in assisting the development of activist groups by providing organization management training. The outside institutions are especially important if they can put the group members in contact with non-governmental organizations that can fulfill community needs. Institutions like CEBs and the CPT, have helped organize and support peasant resistance in the struggle for land. Conversely, the government through colonization projects and associations tried to appease the rural movements and even undercut them.

PAGE 147

136 As a result, resettled families typically did not unify against common constraints. Facing major obstacles alone, many families gave up their solitary battles within the project. While turnover can result from lack of cohesion in communities, it can also hamper the development of strong community organizations. However, this may not always be the case. Long-term settlement helps communities became more tight-knit. This is usually when extended families move in and take over land, and then support each other. If they can hold onto the land, they form a nucleus for a cohesive group. This starts a cycle of feedback that begins to stabilize settlement. While it is still a struggle to make successful farm, at least these families have a unified response to their problems.

PAGE 148

CHAPTER 6 LAND USE PATTERNS AND THE IMPACT OF LANDOWNERSHIP TURNOVER For many families that moved into Peixoto, long-term settlement has not been possible. In earlier chapters I have argued that a major reason families found it difficult to stay on the land was their marginalized socio-economic condition as well as the poor infi-astructure and services that characterized the settlement project. The families settled in these conditions have been able to meet their non-market subsistence needs on the farms, but they have had to struggle to find sufficient income to buy the necessary market goods. In this chapter I will discuss how the factors that made settlement difficult, and the shifting settlement pattern itself, have encouraged agricultural land use strategies that deviate fi-om the ones envisaged by government planners. Planners had anticipated that families would develop diversified farming systems emphasizing perennial tree crops as the primary source of income. This has not happened. In general, the pattern in Peixoto is for families to rely on cattle and pasture formation at the expense of fallow rotation. In fact, cattle and pasture formation play such an important role in these systems that some properties are becoming small and medium sized ranches. This type of land use could have deleterious effects on the productivity of the Peixoto environment and could jeopardize the project's long-term potential for supporting family farms. Even if land is not concentrated into ranches, it could become so degraded that it would be unable to support the resident families. 137

PAGE 149

138 This chapter will begin by describing colonist farming systems in Peixoto and why cash income opportunities are limited for the families living in the project. I begin by examining the range of possibilities open to families in search of cash income sources. Many of the problems associated with possible crops are exacerbated by the poor transportation infrastructure and services found in Peixoto. I will then describe the typical agricultural land use I observed in Peixoto. This will show that ranching and pasture formation have become the predominant investment and production strategy in Peixoto's farming systems. I will then discuss how the land use patterns are related to turnover. While pasture formation and ranching have provided a viable production option for some colonist families, the widespread expansion of pasture could ultimately limit the area open for colonist settlement. Colonist Farming Systems in Peixoto In colonization projects like Peixoto, the inadequate infrastructure, poor government services, and market isolation produce an insecure envirormient in which it is difficult for some colonists to meet their livelihood needs. These conditions hobble many of the resource poor famihes struggling to establish homesteads in harsh environments. All too often, colonization projects have attracted families to remote areas only to abandon them in the forest. They struggle to meet their needs and while they can produce enough for subsistence, cash income possibilities are limited. The difficulty colonists face meeting livelihood needs contributes to insecure land holdings. In these conditions many families can not, or chose not to, stay on their land because of difficulty meeting all their needs.

PAGE 150

139 In general, farmers in Peixoto practice slash and bum agriculture. They have abundant land but are usually resource poor, and have to contend with insufficient labor and capital. The farm lots in the project are large (average is 70 ha) and soils are relatively fertile (IMAC 1991). The problem is that to be used for agriculture the land must be clear of forest before the colonists can begin to establish their farms. Clearing the forest is hard, dangerous, and expensive work. It is labor intensive and many households suffer from a shortage of labor. Most farms rely almost exclusively on household labor, or exchange labor between families. Few of Peixoto's families can regularly pay wage labor. The difficulty establishing a colonist farm is compounded by the fact that many of the families have little start up capital or machinery and credit for small farmers is hard to get. When colonist families first settle in the rainforest, they must quickly clear land and plant the first year's crop. The timing of settlement is crucial because they must cut, dry, then bum the field during the dry months from June through September. Otherwise, they will not be able to plant crop until the following year. Once a plot is cleared, it can productively produce annual crops for two or three years. Then fertility declines and noxious weeds invade. A new plot must then be cleared for annual crops. The old plot can either go into fallow or can be converted to other agricultural uses. If the plot is allowed to go into second growth forest fallow {capoeira) eventually, after 10 or 15 years, the land would be suitable for cultivation again. However, as I will describe shortly, colonists in Peixoto rarely allowed fields to go immediately into fallow. Most families convert plots to other agricultural uses after the yield of annual crops declines. Sometimes if tree crops are intercropped with annuals the field gradually

PAGE 151

140 becomes an orchard or agroforestry plantation. More often the fields are seeded with grasses or legumes to produce pasture for livestock. The productivity of the plot continues to decline and, eventually, it is allowed to returned to forest fallow. Through time, a pattern of land use develops in which families maintain portions of their farm lot in forest, crop fields, pasture, perermial crops, and fallow. Every year or two the family must clear more land for annual crops. Old fields revert to other uses and at some point goes into fallow. The cycle eventually reaches the point where no primary forest is left and families then rely on fallow when they need land. A crucial management decision for the colonist farmers is deciding how and when to allocate land to each purpose. Because they have invested scarce labor in clearing the forest, it is in their short-term interest to attempt to maximize return to their investment. Forming pasture is a popular way to increase the time they can use the plot (Loker 1993). However, there is a critical drawback to this strategy: the longer the land is in pasture the more it degrades and the longer it must be in fallow to recover (Uhl et al. 1988). Hypothetically, if a family continues to clear more land each year, they eventually could reach a point where they have cleared all of their lot. Then, if the land had been used too long, no productive land would be available because it is recovering in fallow. Small Farm Attrition Most research has seen small farm attrition in Amazonia as the end result of a downward cycle of impoverishment. On colonization projects the lack of roads, affordable transportation and storage hamper the colonists' ability to establish viable farms (Wood and Schmink 1979). Households struggle to meet all their basic needs and as a result have a hard time staying on the land. This means they are not assured stable or

PAGE 152

141 secure land ownership. Furthermore, credit is rare for smallholders on the frontier. If it is available, credit terms are unfavorable to smallholders because it is expensive or has high transaction costs (Moran 1979, 1981). Credit problems also set off cycles of indebtedness forcing families to seek off farm wages. Later, as the area develops, the wage labor opportunities disappear and the cycle culminates in the indebted colonists losing their farms (Collins 1986). Even when farms do not fall into debt, frontier conditions often forced colonists to adopt short-term strategies such as the extensive reliance on annual crops which allow them to produce enough for immediate survival but not accumulate surplus or invest in their farms (Browder 1994). Many colonists realize the low chance of successfully establishing a viable farm and instead adopt an itinerant strategy (Ozorio de Almeida 1992). Typically itinerancy strategies are used by posseiro families that have insecure or undocumented land holdings. In an itinerancy strategy a families claims land on the frontier, lives off it for a few years, form pasture, and then sells the cleared land to others, or is reimbursed by the new owner. This strategy allows migrants to accumulate capital they can invest in later homesteading attempts. While the strategy is normally used by posseiros it also can be used by families on colonization projects. While colonists families do not suffer from the threat of expulsion, they may find it difficult to support their family because of conditions on the project and their lack of resources. The possibility of gaining capital and the chance to start over in another location make this strategy attractive to struggling families. Cycles of farm attrition are at work in Peixoto, possibly due to difficulty establishing viable farms that meet all household needs. A principal constraint for

PAGE 153

142 families in this project appears to be the lack of market access. In Peixoto the isolated and abandoned farmers have limited cash income alternatives. Even though families usually can produce enough to meet household food needs, it is harder to produce cash to meet market needs. This greatly influences the quality of life for many colonists. They depend on cash income to purchase materials they cannot manufacture themselves and they need cash to function and interact with government institutions that control their social, political and economic life. Therefore, when families lack viable cash crops, life in Peixoto becomes unacceptable if not impossible. Cash Income Alternatives As has ah-eady been discussed, residents in Peixoto struggle against isolation. Their lots form ribbons of cleared land stretching into the rainforest far from urban centers and markets. Many residents of Peixoto are unable to make a round trip visit to the capital in a single day. The problem is exacerbated when families live in areas that lack affordable or dependable transportation. As a result, the costs, the distance, and the general difficulty of travel mean colonist families have fewer options for cash income. In this section, I will briefly examine some major cash income alternatives and discuss the advantages as well as disadvantages of each alternative. I have grouped these alternatives into five sets: wage labor, traditional Amazon products, annual crops, perennial crops, and livestock. Since generating cash income is a constraint, seeking wage labor employment would be the most direct response. However, this is not easy. Most households do not have surplus labor, so any labor devoted to off-farm employment means less labor goes to developing farms. This is significant since many families survive on the frontier because

PAGE 154

143 they meet subsistence needs with farm production. Furthermore, in colonization projects there are few opportunities to earn wages. Manual labor to clear forest is probably one of the more common paid jobs. The problem is that most families cannot afford to pay for additional labor. There also are few opportunities for employment in rural areas around Peixoto. The ranches surrounding the project are all well established and no longer clear much land. As a result they need less labor. Even when such jobs are available they are highly seasonable. Practically the only regular wage in Peixoto is provided by the rural schools. Most schools usually employ one teacher and possibly an assistant to care for the building. Few of these jobs are available and those that exist require a certain level of training and possibly political ties. Sometimes, even these wage earners are not paid for months. Another source of cash income comes from the production of traditional Amazonian products. These include extractive forest products like rubber and Brazil nuts, and agricultural products like manioc flour (farinha). Historically, these products were the "bread and butter" of rural producers in the region. They are species adapted to the region, and historically production systems developed that were adapted to the isolation of the region. They require little investment to produce and prepare other than physical labor. These products are ideal in areas with poor market access because they can be stored with minimal spoilage. In the early stages of settlement families supplement their subsistence and income through extraction from the forest. Families reported that during the first years of settlement they cut rubber to provide income while waiting for the first harvests of other crops. Until recently, Brazil nuts also provided an important seasonal income. Almost

PAGE 155

144 80 percent of families reported that they had harvested nuts. The price had dechned by mid-1990s and most informant famihes had stopped gathering nuts for sale. Farinha continues to be an important staple among colonist families Manioc is easily cultivated in this environment and it is not a major investment to build a casa da farinha, a building used for preparing and roasting the farinha. The production of farinha is labor intensive but for the resource poor in need of cash it is sometimes the only option. A positive aspect of farinha is that it can be stored for extended periods of time. Because farinha can be stored, it is attractive to colonists that lack dependable sources of transportation. While it is not especially lucrative to produce, at least it provides some cash for families without other options. The use of traditional products has become less common among colonists as the project has matured. The major drawback to these products as a source of income is that they are low value products that require intensive labor. Extractive products like rubber and Brazil nuts occur in low density on the farm lots. The availability of these resources further declines as forest is cleared. They continue to be used but opportunistically and are sometimes relied upon during times of desperation and hardship. They are still used in remote areas where transport is infrequent or undependable. This is probably because they can be stored without spoilage between market trips and at least provide some cash for those with no other option. Another option for cash income is the sale of armual crops. In Peixoto, the principal staples are armual crops like rice, beans, and com. Generally, Peixoto's soils are fertile enough for families to meet their subsistence food needs. Whenever a surplus is produced it is sold, traded with neighbors, or bartered to cover costs of grain

PAGE 156

145 processing. When Acre's farm families sell annual crops they are not very lucrative. Storage facilities in the state are inadequate so annual crops are sold at harvest. Therefore, local markets are flooded and prices are low. Depending on annuals for all cash income can be risky. Given the poor roads and lack of dependable or affordable transportation, these investments could be lost. When a family depends solely on annual crops for cash income, it is because there are no other viable options. Perennials and agroforestry systems are potentially more lucrative for colonists but are not common. When Peixoto began, planners expected perennials to be the main income soiu-ce for colonists. This is because perennials could provide a stable, long-term income source. Furthermore, if used in agroforestry systems, they could provide an environmentally sound production alternative (see Nair 1989 for a description of agro forestry's benefits). The use of tree crops is not a foreign concept to most colonists. Many migrants had worked with perennials in coffee plantations in the south. Throughout Acre, most families incorporate agroforestry plots in home-gardens for household consumption and, occasionally, colonists sell surplus fruit from their gardens (Cartaxo Nobre 1998). These gardens are usually small areas around the home. Diversified agroforestry plantations, and monocultural plantations of perennial crops can provide major sources of income. They are used by some of Peixoto's most economically successftil family farms. Farmers report that coffee is the most lucrative crop.l but other crops including citrus, rubber (Hevea brasiliensis), and guarana 1 It was not always easy for farmers to profit from coffee. A number of producers reported that in the mid-1980s when they harvested their first coffee, they had a difficult time selling their crop because Rio Branco had no large coffee buyers. In 1996 there

PAGE 157

146 {Paullinia cupana) are also produced. Surprisingly, regional fruits that play an important part in agro forestry systems in other areas such as cupua9u {Theobroma grandiflorum), pupunha {Bactris gasipaes), and a9ai {Euterpe olerpe) were practically absent among sampled households (with the exception of small numbers of these trees in home gardens). While perennials and agroforestry appear to have good potential, large plantations are not common. There are often problems with agroforestry adoption and perennials in Amazonia related to credit availability and market systems for these products (Smith et al. 1995). Generally, affordable credit is not available for smallholders to invest in tree crops. Even if farmers have the resources and want to invest in trees, it is hard to find nurseries to supply seedlings to start plantations. Also, many agroforestry products lack developed markets. The absence of agro-industry in the region means that even if produced, these products are difficult to sell. Another constraint is that most regional fioiits, like cupuagu, have a short post harvest life. Given the transport problems and lack of agro-industry, farmers would be undertaking a considerable risk if they were to produce these fixiits in large quantities. Furthermore, many regional tree species also suffer from disease when planted as monocultural crops. This is especially true for rubber in Amazonia. Yields from rubber plantations in the region remain low due to South American leaf blight (Microcyclus ulei) (Dean 1987). Disease also can plague plantafions of cacao which can be devastated by the witches broom fungus {Crinipellis perniciosa). were a number of competing coffee buyers in Rio Branco as well as buyers in Acrelandia and Placido de Castro.

PAGE 158

147 Because many colonists are resource poor it is harder for them to commit the labor or capital that will not provide return for years investment. This is an important constraint since perennials and agroforestry plantations require long-term investments. Households must feel secure on the land and expect to remain on the land long enough to benefit if they are going to risk investment in tree crops. Many families probably do not feel the risk is reasonable. For most households, in Peixoto livestock provide a more lucrative source of income and savings than the other alternatives. While there are drawbacks, the benefits from livestock make them very attractive additions to most colonist farming systems. Small animals like pigs and chickens are important for these families because they can be sold easily and can be eaten if necessary. Chickens are sold most frequently, usually by women who raise and sell poultry to meet immediate household needs. Chickens are easily produced, transported, and usually can be sold in small quantities. Even if the total yearly income from poultry is not high compared to other seasonal sources, this income is spread throughout the year and is crucial for the day-to-day fiinctioning of the household. There also are a number of constraints that keep farmers from raising small animals in large numbers. For one thing, feed must be produced for chickens and pigs. While most families produce enough food for themselves, they report that there is often not a large surplus. Therefore, to supplement their animals' diet and save labor, colonists usually allow the animals to roam free. This saves food and labor but it has other drawbacks. The small animals fall prey to wild animals. In addition, disease takes a heavy toll since these families have almost no access to veterinary services and rarely

PAGE 159

148 have vaccines. Free roaming pigs can become nuisance animals destroying crops and angering neighbors. It is usually easier to have few or no hogs to preserve good relations with neighbors. Colonists consider cattle a very desirable investment. Research in other areas in the Amazon indicates that the most successful families are those that are able to invest in cattle and become small scale ranchers (Ozario de Almeida 1992; Browder 1994; Moran 1989). The importance of cattle production throughout Peixoto is apparent in the prevalence of cattle and the high percentage of land held in pasture. Why are cattle such a viable option for farmers on Amazonian frontiers? Loker (1993) points out that cattle serve multiple functions. First, cattle function as savings accounts or insurance. In areas without banks and with high inflation, holding cash can be risky. Cattle are an investment that retains value in high inflation economies. Second, cattle are very marketable. They generally have a high value and are easily sold. Farmers in Peixoto reported that cattle are one of the only commodities which the buyers will come to the farm to purchase. In a worst case scenario cattle can be walked to market by the owner. Third, if the family has dairy cattle, they provide dual purpose production with dairy products for both consumption and sale. Finally, cattle require low labor inputs relative to their high value. As Loker (1993: 18) points out, this is rather ironic: in general, Amazonia's rural people suffer when they are displaced by large-scale livestock operations that do not employ wage labor (Goodland 1980); however, the low labor required for cattle production also makes this an attractive investment for colonist families. The principal constraint to cattle is that for many colonist families it is difficult to gather surplus capital to invest.

PAGE 160

149 For frontier areas like those in the Brazilian Amazon, investment in cattle is favored because it is not tied to the land. Unlike other investments, such as tree crops, if a family leaves a farm lot they can take their investment in cattle with them. Pasture formation may also be tied to the cycle of turnover and not just as a means to raise cattle. In an environment where itinerant strategies are common, forming pasture makes sense because it raises the value of land for resale. Even when a family cannot afford cattle, this is still a good investment. For one thing, the family does not lose the labor they invested in clearing the land. But more importantly, by forming pasture they increase the value of land for resale. This allows them to accumulate more capital for later homesteading attempts. While pasture formation may make sense in terms of the immediate goals of a migrant family practicing an itinerant strategy, in the long run this type of strategy may have negative environmental consequences. I will return to this point later in the chapter. Typical Land Use in Peixoto So far I have described the difficulties and limitations of small farms in Peixoto. The project was intended as an opportunity for migrant farmers to develop diversified farms reliant on perennial crop production. This did not happen. In this section I will describe land-use on Peixoto 's family farms. During the survey, informants were asked to estimate the area covered by different categories of land use. The categories I used were: total lot size, annual crops, perennial crops, pasture, and second growth fallow.^ The category annual crops referred to fields of annual and semi-annual crops known as lavoura or lavoura branca by the colonists. The principal crops were rice, com, beans.

PAGE 161

150 and manioc. The category perennials referred to tree crops either in plantations or homegarden agroforestry systems. Principal crops included coffee, banana, and citrus. The pasture category referred to lands maintained in grasses (either improved or natural). The category second growth fallow referred to areas that had been cleared for agriculture but were allowed to revert to forest. There is much diversity in the farms found in Peixoto. Still, it is possible to describe the typical farm by using median and mean land use figures. Since there is wide variation from farm to farm, I will use the median areas to describe the typical farm unless otherwise noted. After describing the typical farm, I will then discuss variation in land use patterns and how they are related to turnover. Attempting to describe the typical farming system in Peixoto is difficult because there is such wide variation from farm to farm, especially in areas devoted to perennials, pasture and in the size of cattle herds. Among the colonist families sampled the median farm size is about 70 hectares, although farm properties ranged from 24 to 219 hectares. This size difference is largely the result of families dividing lots among family members or purchasing additional lots. Most of the farms (95 percent) are between 69 and 80 hectares. On a typical lot, the family has cleared and is using, about 30 percent of the land. The utilized portion consists of about 3 hectares of annual and semi-annual crops ^ I did not ask farmers directly how much of the forest had been cleared. This is a sensitive, politically charged question in the Amazon. Rather than losing the confidence of farmers, it was easier to calculate that figure indirectly from the other information.

PAGE 162

151 (rice, beans, com, manioc), 12 hectares of pasture, just over 1 hectare of second growth (capoeira), and less than one hectare of perennials (see Figure 4).3 Annual Crops 2.75 ha — Rasture 12 ha / P ^-sd f^ennlals 0.25 ha Rjrest 54.04 ha | L-*s^^ Second Growth I \ 1.2 ha Graph represents median land-use for a typical 70 hectare lot Figure 6.1 Land use on typical Peixoto farm The average property surveyed had 66 percent of forest cover remaining. The Brazilian government allows colonist families to clear 50 percent of the forest from their property. Most famiUes interviewed had not reached this limit. Only 17 percent of families surveyed had cleared more than 50 percent of their property. Nonetheless, there is a wide range of variability in each commimity. Forest remnants found on farm lots ranged from 90 percent on one property, to only 9 percent on another. For these famiUes perennials were not a common investment. While some farms have large plantations (the largest was 30 hectares), most have almost nothing. Out of the 137 sampled households that responded, 53 had no perennials at all. Of those that do invest in perennials, approximately 50 percent have 1 hectare or less. 3 Because of the wide range of variation and skewed distribution in some categories I am using the median value for all land use categories except perermials. For perennials the

PAGE 163

152 Cattle herds varied in size and the distribution was noticeably unequal. Some families had no cattle while the maximum herd size was 250 animals. Of the 146 families interviewed only 26 had no cattle. Four families had herds larger than 100 animals. Just 17 percent of the sampled families owned 45 percent of all cattle in my sample. Pasture formation was the dominant land use among the farms surveyed. On average, pasture made up about 60 percent of utilized land. This figure is slightly more than the area measured by other researchers in Peixoto (Witcover and Vosti 1996). Land area in pasture ranged from 0 to 120 hectares. Only one family claimed to have no pasture. Apparently, some families are forming pasture even though they have no cattle, a trend that has been noted elsewhere in Amazonia (Browder 1994: 51; Ellis 1995). Second grovv^h fallow did not play a major role in land-use among colonist farms in Peixoto. The area in second growth fallow ranged from 0 to 30 hectares but the median was only 1 .2 ha. Surprisingly, 42 percent of informants claimed to have no fallow on their land. This means they are prolonging the use of the land rather than letting it recuperate. When I began this research I expected there to be two extremes in land use strategies: one a diversified strategy with much use of agroforestry and fallow, and another strategy that would be less diversified but dominated by pasture formation and cattle. This did not turn out to be the case because while farms are diversified, perennials and fallow do not play a strong role. Conversely cattle and pasture formation are important across all farms. median value was zero. While the use of perennials is low, they are not absent.

PAGE 164

153 Some families are well on their way to becoming small or medium size ranchers. The colonists becoming ranchers attempt to continue expanding their pasture. Others are struggling to maintain their hold on the land. These households hedge their bets by forming pasture to improve resale value and buy cattle that are a portable investment. While many families still remain more diversified and fall between these two extremes, emphasis on pasture formation could eventually jeopardize the future viability of Peixoto as a project to resettle displaced families and provide them long-term access to land. Before continuing with this discussion it will be helpful to test whether land use patterns in Peixoto are related to landownership turnover. The Effects of Landownership Turnover on Small Farm Land Use How has land ownership turnover affected land use patterns in Peixoto? At the outset of this research I expected that there would be a close relationship between longterm settlement and land use. Families that had lived on the land for a long time and expected to continue living there would have developed farming systems that attempted to conserve the long-term productivity. Therefore, families that lived on farms that had not turned over and had been continuously settled would be more likely to emphasize long-term investments like perennials, and fallow rotations. On these farms, cattle and pasture would be present but would not have a dominant role in the production strategies. Conversely, in areas where land had changed hands frequently, there would be less reliance on fallow and perennials and greater emphasis on pasture and cattle. This is because past owners, once they realized they were not going to keep the land, would have been less likely to make long term investments. They would have been unlikely to invest in slow maturing tree crops nor would they be interested in letting the cleared land go

PAGE 165

154 into fallow to maintain fertility. Furthermore, if these families farmed pasture on old fields, they would not only maximize the returns from labor invested clearing the land, but they might increase the land's value for resale. Therefore, lots that turned over more often should have more pasture and less fallow and perennials. To test the relationship between turnover and land use I broke the sampled farms into three categories based on the number of previous owners. The categories are: farms occupied by the original owners (group #1), farms occupied by the second owner (group #2), and farms that have had three or more owners (group #3). The distribution of households is listed in Table 6.1. I compared differences in land use among these Table 6.1 Distribution of Farms by Owoiership Category Ownership Category Number of Farms Group #1 (original owners) 49 Group #2 (second owners) 40 Group #3 (three or more owners) 55 groups using a one-way ANOVA procedure and Duncan's multiple range test to identify differences between means. As I will explain, there were differences in land use behavior between high and low turnover areas. Farms that had turned over more frequently had higher percentages of cleared land in pasture and lower percentages in fallow. Property size varied between these three groups but the area cleared for use did not differ. The lots occupied by the original owners were significantly smaller (p=.05) than the average lots for the other two ownership categories (see Table 6.2). However, the area cleared did not differ significantly between the different groups."^ Even when '+ Although it does not show up when the lots are grouped by ownership category, in general, larger lots do have slightly larger areas cleared than smaller lots. However, there

PAGE 166

155 cleared area is compared as a percentage of total lot size to control variation in property size, there is no significant difference from group to group. This indicates that regardless of the number of previous owners, families have cleared roughly the same area for use on their land. When the area devoted to each land use type is compared across ownership categories there is some variation but there are no significant difference in the area of pasture, fallow or crops. In addition, there was no significant difference in the size of cattle herds between these groups. The only significant difference is in the area devoted to perermials. However this did not vary as I had expected. In this case, second owners (group #2) had slightly larger areas in perennials (p=.l) than the other two ownership categories. Table 6.2 Mean Areas of Land Use Types by Ownership Category Ownershi P Categories Mean area in Hectares Total Size Total used Pasture Fallow Crops Perennials Group #1 66.9* 20.3 12.6 3.2 3.5 1.0 Group #2 80.3 29.9 18.7 4.2 4.2 2.6$ Group #3 79.3 28.3 20.1 4.8 2.4 1.1 Significance: *(p=.05), $(p=.l) Since these farms were only using a portion of the entire lot, I decided to compare land use area as a percentage of the cleared area. This measured how colonists families allocate cleared land among the different land use types. With this measure there was significant differences in the manner in which families used their land. The difference is in the allocation of land to pasture and fallow between the ownership categories. is a wide range in these groups and the average portion of cleared land stays the same across the different sized lots.

PAGE 167

156 Table 6.3 Percentages of Cleared Land by Use Type and Ownership Category Ownershi P Categories Mean percent of used land area N % in use % crops % perennials % pasture % Fallow Group #1 .309 .20 .04 .54 .20 45 Group #2 .349 .19 .08 .56 .16 37 Group #3 .357 .20 .04 .67 .07 45 Significance: (less than p= 05) Group #3, which consists of the farms that suffered the highest rates of turnover, had a higher percentage of cleared land in pasture (67%) than either of the other two groups (p=.0169). Furthermore, the farms in this group also had a lower percentage of fallow than the other two groups (p=.0024). These figures indicate that turnover does influence land use. While comparison of actual areas does not indicate a dramatic disparity between the high turnover group (Group #3) and the other two categories, this examination of land use allocation within the farms indicates a trend. The allocation of land on the high turnover properties indicates that on average farms that have more turnover have a greater percentage of used land in pasture and a lower percentage in fallow. This is what I would have expected if past owners avoided long term investment and attempted to increase land value for resale. Maintaining land as pasture instead of fallow may indicate households that are trying to extract short-term gain at the expense of future productivity. Keeping land in pasture and out of a fallow rotation may increase the short-term production on a farm, but it also undermines the long-term productivity. Other aspects of my hypothesis were not supported. Surprisingly there was no difference in the size of the cattle herds among these three groups. Also, I thought that perennials would have been more common in low turnover areas. The area in perennials

PAGE 168

157 is very small among all three groups and, unexpectedly, the only significantly different group was the second owners who had larger areas in perennials than the other two ownership categories. Surprisingly, three of the largest perennial plantations were found on farms that had at least 100 head of cattle. Discussion Land use in Peixoto is turning predominantly towards pasture formation which could have negative connotations for the future of Peixoto for smallholder settlement. In Chapter 4, 1 pointed out that broad swaths of the project have already been concentrated by large landowners. Most of this land has been clear cut to form pasture for cattle. As smallholders convert land to pasture it could facilitate concentration by making land more attractive to ranchers. Even if the colonists do not intend to sell out, the reliance on pasture could degrade the farm lots to the point where the family will not meet all of their livelihood needs. Loker (1993) argued that when families use scarce labor to clear land, they maximize the return from their labor investment by prolonging their use of the land. Therefore, when crop yields decline on a plot, farmers seed pasture to use the land for a few more years. They do this even though the prolonged use ultimately lengthens the duration of fallow needed to regain soil fertility. Loker maintained that, in the long-term, these farmers could face a crisis because eventually they would run out of fresh crop land and pasture. The situation in Peixoto could be worse than the outcome Loker foresaw in Peru. Not only is pasture for ranching a principal investment among colonists, but fallow does not appear to play a major role in the farming systems of many families. I expected

PAGE 169

158 families to use secondary forest fallow as a rotation to conserve soil fertility. What I observed in Peixoto is that much of the pasture formation seems to be taking place at the expense of secondary forest fallow, capoeira. Actually, secondary forest made up a small percentage of land-use on most lots and many families did not even report this type of cover. Rather than allow crop lands to become secondary growth, most families attempted to maintain this land as pasture. This is especially true on high turnover lots in Peixoto. Instead of seeing capoeira as a positive means of regenerating soil, farmers had a negative view towards this type of fallow. Apparently, informants saw land reverting to capoeira as a sign of laziness and the inability to maintain the land's productivity. In group interviews, some individuals were teased for "only growing capoeira." Even the farmer with the most diversified, and largest agroforestry system I sampled said he wanted no capoeira. He intends to use cover crops and green manure whenever necessary to regenerate or rest soil. In his view, letting land return to forest meant losing the initial investment made to clear the area. Eventually, this trend could limit the area available for family settlement. Families in Peixoto depend on their farms to provide much of their subsistence. However, individual fields can only be used for a few years before they must go into fallow to be useful for crops again. Instead, the fields are usually converted into pasture. Pasture will eventually degrade and the longer it has been used the longer it will takes to recuperate. When wide areas are in pasture without regularly cycling into fallow, it means that at some point much of the land will be degraded and need longer fallow to recuperate. When this happens the family that owns the land will be without crop land or

PAGE 170

159 pasture. Peixoto's population is young and if these people cannot be supported in Peixoto, they will need to go somewhere. There are few open places in Acre. Ranches cover much of the area around the project and are already encroaching on the project. These large landowners exclude people from the land. Conclusions When Peixoto was originally planned, it was envisioned as a settlement of diversified family farms that were self sufficient through the production of food crops and prosperous through the cultivation of perennial tree crops. However, this did not occur. Peixoto has become an area dominated by cattle production and pasture formation. This is taking place on a number of fronts, from family farms within the project to ranchers who want to expand into the project. The small farms, while still diversified within the project, are investing heavily in pasture and cattle. A major reason this is happening is because there are few other cash income sources available to colonist families. Pasture makes up the largest areas of cleared land on most farm lots and almost all families are trying to buy cattle. Many of the more prosperous farms are well on their way to becoming small scale ranches. Other families that are struggling are investing in pasture even if they do not have cattle. On the other hand the use of perennials and second growth fallow is very low. These crops take up a minor portion of used land on most farms and on some, one or both of these land use types are absent. While high pasture and low fallow are the norm there is a relationship to turnover. The highest turnover farms have more pasture and less fallow.

PAGE 171

160 As the unstable settlement pattern persists within the project, it provides opportunities for large ranches to buy up farm lots when they are turned over. If this type of land use continues to expand, it may lead to decreased areas available for settlement by small farms. If this happens the families settled within Peixoto will once again join the ranks of dispossessed rural people in Brazil.

PAGE 172

CHAPTER 7 CONCLUSIONS Landownership turnover has plagued Amazon colonization projects from their inception. The high frequency of landownership turnover provides clear evidence that these projects do not provide a realistic possibility of long-term settlement or income for most resettled families. When land is tumed over on colonization projects, it means that rural families in Amazonia still have an uncertain future. Once they leave, their possibilities are limited. They could move to one of the region's growing urban slums to face probable unemployment. They could buy or invade land farther out in the forest and face continued hardship on the periphery. Or, they could remain in the area to work as a day laborer or sharecropper. These are the stark options that underlie the fact that rural peoples in Amazonia have little hope of escaping cycles of rural poverty and hardship. While many families use itinerancy strategies to try to survive in frontier areas, it is probably a small minority that finally become prosperous land owners. With the unstable settlement pattern in rural areas, it is the elite who benefit. This pattern gives land speculators access to cheap land that has been cleared with the labor of the poor. Elite groups have the capital to buy up the land and the political power to resist laws against the concenfration of land. As a result, the resources devoted to helping the poor open and develop land eventually help to expand the holdings of large landowners. This situation causes adverse environmental changes. Farmers turn cleared land into pasture. If the land is then concentrated by ranches, the remaining forest on the lot 161

PAGE 173

162 will be cleared. Once most of the land in Peixoto has been converted to pasture it will be difficult for farms families to support themselves from the land. At that point, the project will cease to be a possibility for rural people searching for land. This cycle of destruction continues to expand as failed colonists move further into the forest to start over. How could such a cycle develop? Even if it is assumed that the Brazilian system is undercut by corruption and that elite groups are motivated only by greed, would they have the foresight or ability to develop such a scheme? It is very unlikely. But then, why would smallholders agree to such a program? They would not, unless they had no choice or thought they could benefit. The fi-ontier transition that swept across Amazonia was a complex process. At most turns, competing interest groups contested the direction of development policies. The government's development agenda for Amazonia was generally biased toward the interests of elite groups. The goal was to encourage capital investment to develop in the region. Government was ambivalent to the situation of rural people and maintained contradictory policies for smallholders impacted by this change. When rural peoples resisted change, the government instituted populist programs like colonization and land distribution. The emphasis of these programs was in limiting rural unrest and there was less interest or concern for supporting these families. The underlying goal of programs like land distribution was to pacify the disenfi-anchised with seemingly unlimited land so that they would not impede the development agenda. In the case of Acre, the development transition brought rapid and dramatic change. The state's land tenure regime was rapidly transformed as old landowners were forced out and new investors were invited in to the state. Thousands of rural families

PAGE 174

163 were displaced and violence erupted. Instead of assuring that land rights were respected, the government expropriated land elsewhere to resettle those displaced by development. In this way the government responded to the symptoms rather than the cause of unrest. Through time, the government realized that the problems of rural people struggling for land would not go away. However, this recognition occurred gradually. As the government realized it could not control colonization and assure an orderly resettlement process, emphasis shifted through different settlement models, from PICs to PADs, then later to PARs and extractive reserves. Basically, this was a trend of decreasing government responsibility that eventually led to the acceptance of posseiro 's land rights. This occurred only when the rural people provided enough stiff resistance. Gaining land through a government land distribution program was often a Pyrrhic victory for families struggling to survive on the frontier. Having property in a remote area does not assure families much more than a subsistence livelihood. Families found it difficult to participate in the market and produce the cash income they needed. Many families won land only to learn later that they could not survive there. Peixoto is a prime example of what happens when families are resettled only to be abandoned in remote areas. Only a fraction of those families that were settled in the project have remained. Turnover in Peixoto is not only high but varies dramatically from place to place. Some areas no longer have family farms and have been transformed into ranches. In most areas of Peixoto the original colonists have left but were replaced by other migrants. In a few areas original owners have stayed, forming stable settlements. This pattern indicates that Peixoto has failed to provide long-term homesteading opportunities for most resettled families and may actually be opening up areas for

PAGE 175

164 consolidation by large landowners. As land rights were passed from owner to owner, large areas gradually fell into the hands of large landowners. The complex causes of turnover and the resulting settlement pattern include such things as age of settlement, resource quality and other institutional factors, but this research also shows the role played by infrastructural conditions and transportation services that define the level of isolation for each area. In some areas land is lucrative, because it is accessible and titled. The high price colonists can earn motivates farmers to sell out. In other areas, resettled families have poor access in and out of their properties which lowers their quality of life. Explaining turnover behavior is difficult because there are also exceptional pockets where families have overcome hardship and have held on to their land. The Souza Araujo community is a good example. Souza Araujo's infrastructure and service conditions were not perfect but they were relatively better than many other locations within Peixoto. Furthermore, when Souza Araujo's families were confi-onted with constraints like isolation, they were able to respond and overcome these problems. As Souza Araujo demonstrates, not all colonist families failed. In areas where families have been able to stay on the land, it is often because they have formed strong activist organizations. A number of factors must come together for these groups to form (for example cohesive communities, strong charismatic leadership, and developed management skills) and in many parts of Peixoto they were absent. When these organizations work, they limit the impact of isolation, and provide means for collectively responding to common constraints (by providing transportation, machinery and lobbying for road repair). Such groups are rare. Establishing organizations is not easy.

PAGE 176

165 Furthermore, efforts by the government to undercut and control rural organizations for fear of mobilized rural opposition effectively lessened the likelihood that groups would develop. In communities lacking organizations, individual families bear the brunt of poor infrastructure and services by themselves. As a result, many families give up. In Peixoto, most families meet their subsistence needs but have limited cash income options. The lack of cash crop alternatives is due largely to the difficulties family farms have interacting with urban markets. As a result, cattle and pasture formation have become dominant agricultural activities for most colonists since this type of production is less affected by market isolation. Consequently, fewer families invest in agroforestry or forest fallow rotation, activities that may be crucial to conservation and the long-term viability of small farms. Throughout its history, the harsh conditions and isolation have made it difficult -if not impossiblefor colonist families to remain on the land. In addition to frequent turnover in land ownership, this has also led to land-use strategies, like production cattle and extensive pasture formation, that allow colonists to cope with poor market access without tying their investments to insecure land holdings. These strategies may threaten the long-term productivity of the land, and the project's ability to support rural families in the flitiu-e. Among the families that remain, some are becoming small and medium ranchers. Others are positioning themselves to sell out to ranches, if their homesteading attempt fails. In fact ranches have already encroached on large portions of the project. The extensive transformation of forest to pasture and the prolonged use of pasture could produce a major land availability crisis in Acre. Peixoto was intended to alleviate the problems of displaced and landless people by providing long-term settlement and

PAGE 177

166 productive fanning opportunities for rural families. Past events have turned wide areas of the project over to landowners who maintain extensive pasture, land that no longer supports families. Colonist families that remain on the land devote much of their lots to pasture expecting that they can either continue to expand or that they will turn the land over and move on to a fresh lot. This cycle caimot go on indefinitely. It is already difficult to expand settlements into new areas in the state. The forests of Amazonia were never unoccupied. Even if conditions changed and turnover and itinerancy stopped today, it would be difficult to accommodate everyone, especially considering the young population and their future demands for land. In my sample, 50 percent of the population of Peixoto was less than 18 years old. In the fixture, these people will have to go somewhere else and will need to find some means of supporting themselves. Large landowners that now control wide tracks of land deny migrants new homesteads. Within a project like Peixoto there is no room to expand as all the reserved areas have been invaded by families that had lost their land earlier. The rubber tappers and indigenous groups that remain on the land in Acre will not move aside easily.

PAGE 178

APPENDIX A EQUATIONS AND NOTATION USED FOR TWO-STAGE CLUSTER SAMPLE N = the number of clusters in the population n = the number of clusters selected in a simple random sample M, = the number of elements in a simple random sample from cluster / m,= the number of elements selected in a simple random sample from cluster / Estimator of population proportion p : p = '^. 1=1 Estimator of Variance of p : \ N AuM-J nNM~ tT where TM-ip-pf s:=^ V M. P,Q, and Bound on the error of estimation: P2ylHf) Source: (Scheafer, Mendenhall, and Ott 1990). 167

PAGE 179

LIST OF REFERENCES Adriance, Madeleine Cousineau 1994 "Base communities and rural mobilization in northern Brazil." Sociology of Religion 55 (2): 163-78. 1995a "The Brazilian Catholic Church and the struggle for land in the Amazon." Journal for the Scientific Study of Religion 34(3): 377-84. 1995b Promised Land: Base Christian Communities and the Struggle for the Amazon. Albany: State University of New York Press. Bakx, Keith 1986 Peasant Formation and Capitalist Development: The Case of Acre. SouthWest Amazonia. Ph.D. Dissertation, University of Liverpool. 1988 "From proletarian to peasant: Rural transformation in the state of Acre, 1870-1986." The Journal of Development Studies 24(2): 141 -160. 1990 "The shanty town final stage of rural development? The case of Acre." In David Goodman and A. Hall eds.. The Future of Amazonia:Destruction or Sustainable Development. London: Macmillan Press. Bemo de Almeida, Alfredo Wagner 1990 "The state and land conflicts in Amazonia, 1964-88." In David Goodman and A. Hall eds.. The Future of Amazonia:Destruction or Sustainable Development. London: Macmillan Press. Browder, John 1994 "Surviving in Rondonia: The dynamics of colonist farming sfrategies in Brazil's northwest frontier." Studies in Comparative International Development 29(3): 4569. Browder, John and Brian Godfrey. 1997 Rainforest Cities: Urbanization. Development, and Globalization of the Brazilian Amazon. New York: Columbia University Press. Bunker, Steve G. 1985 Underdeveloping the Amazon: Extraction. Unequal Exchange, and the Failure of the Modem State. Urbana, IL: University of Illinois Press. Cartaxo Nobre, Francisco Rildo 1 998 Agroforestry Systems in Acre. Brazil: Variability in Local Perspectives. MA Thesis, University of Florida 168

PAGE 180

169 Carvalho, Lucas Araiijo 1981 Avaliapao do Projeto de Reestruturapao de Colonias Agricolas de Estado do Acre -NARI. Rio Branco, Acre: CEPA. Cavalcanti, Francisco Carlos S. 1983 O Processo de Ocupapao Recente das Terras do Acre. Disserta9ao de Mestrado, Belem: NAEA/UFPA Cavalcanti, Tristao Jose da Silveira 1994 Colonizacao No Acre: Uma Analise Socioeconomica do Projeto de Assentamento Dirigido Pedro Peixoto Disserta9ao de Mestrado, Universidade Federal do Ceara. CEDEPLAR 1979 Migra^oes Intemas na Regiao Norte: O Caso do Acre. Iniv. Federal de Minas Gerais Collins, Jane 1985 "Smallholder settlement of tropical South America: The social causes of ecological destruction." Human Organization 45(1) 1-10. CPT 1990 Conflitos no Campo/89. Goiania, Goias: CPT 1996 Conflitos no Campo. Goiania, Goias: CPT Cronkleton, Peter 1996 Variapao da Rotatividade da Posse da Terra no Projeto de Colonizacao Pedro Peixoto Acre. Brazil. Unpublished field report. Dean, Warren 1987 Brazil and the Struggle for Rubber. Cambridge: Cambridge University Press. Diegues, Antonio Carlos 1992 The Social Dvnamics of Deforestation in the Brazilian Amazon: An Overview Geneva: UNRISD Discussion Paper The Economist April 13, 1996 "Land for the landless: Brazil. (Movimento Sem Terra land-reform movement)." 38-39. May 18, 1996 "Human rights, political wrongs: Brazil. (President Fernando Cardoso issues human-rights reforms)." 43. May 3, 1997 "Mortal after all: Brazil. (President Fernando Henrique Cardoso encounters resistance to his modernization plan.)" 36 Ellis, Edward 1996 Land and Tree Use on Colonist Farms in the Aguarico Sector. Cavambe-Coca Ecological Reserve. Ecuador. MA Thesis, University of Florida.

PAGE 181

170 Foweraker, Joe 1981 The Struggle for Land: A political Economy of the Pioneer Frontier in Brazil from 1930 to the Present Day. Cambridge: Cambridge University Press. Goodland, R. J. 1980 "Environmental ranking of Amazonian development projects in Brazil." Environmental Conservation 7:9-26 Goodman, David, and Anthony Hall (eds.) 1990 The Future of Amazonia: Destruction or Sustainable Development? London: Macmillian Press. Guerra, A. T. 1955 Estudo Geografico do Territorio do Acre Rio de Janeiro: IBGE Guggenheim, Scott E. and Michael M. Cemea 1993 "Anthropological approaches to involuntary resettlement: Policy, practice, and theory." In Michael Cemea and Scott Guggenheim, eds., Anthropological Approaches to Resettlement: Policy. Practice and Theory. Boulder: Westview Press. Hecht, Susanna B. 1985 "Environment, development and politics: capital accumulation and the livestock sector in eastern Amazonia." World Development. 13(6):663-684. Homewood, Brian 1996 "Ploushares into swords." New Scientist 152(2059): 18-19. IBGE {Instituto Brasileiro de Geografia e Estadstica) 1988 Censo agropecuario Rio de Janeiro, Brasil: IBGE. 1991 Censo Demografico 1991. Rio de Janeiro, Brasil: IBGE INCRA 1996 Proietos de Colonizapao de Assentamento e Agroextrativista em Execupao. Unpublished document IMAC 1991 Atlas Geografico Ambiental do Acre. Rio Branco, Acre: IMAC Journal Gazetta July 1, 1996, July 9, 1996, Rio Branco, Acre Loker, William M. 1993 "The human ecology of cattle raising in the Peruvian Amazon: The view from the farm." Human Organization 52 (1): 14-24.

PAGE 182

171 Mahar, Dennis J. 1979 Frontier Development Policy in Brazil: A Study of Amazonia. New York: Praeger. 1981 Brazil: Integrated Development of the Northwest Frontier. Washington, D.C.: World Bank. Margolis, Maxine 1973 The Moving Frontier: Social and Economic Change in a Southern Brazilian Community. Gainesville: University of Florida Press. Martine, George 1990 "Rondonia and the fate of small producers", In D. Goodman, and A. Hall, eds., The Future of Amazonia: Destruction or Sustainable Development? London: Macmillan Press. Martins, J. de Souza 1975 Capitalismo e Tradicionalismo: Estudos Sobre as Contradicoes da Sociedade Agraria no Brasil. Sao Paulo: Pioneira. 1990 "The political impasses of rural social movements in Amazonia." In D. Goodman, and A. Hall eds. The Future of Amazonia: Destruction or Sustainable Development? London: Macmillan Press. Millikan, Brent H. 1988 "The dialectics of devastation: Tropical deforestation, land degradation and society in Rondonia, Brazil" Master's Thesis, Department of Geography, University of California, Berkeley Monteiro, Benedicto 1980 Direito Agrario e Processo Fundiario. Rio: PLG Comunica9ao. Moran, Emilio F. 1979 "The TransAmazonica: Coping with a new environment" In M. Margolis and W. Carter eds., Brazil. Anthropological Perspectives: Essays in Honor of Charles Wagley. New York: Columbia University Press. 133-159 1981 Developing the Amazon. Bloomington: Indiana University Press. 1984 "Colonization in the Transamazon and Rondonia," In M. Schmink and C. Wood eds. Frontier Expansion in Amazonia. Gainesville: University of Florida Press. 1989 "Adaptation and maladaptation in newly settled areas," In D. Schumann and W. Partridge eds., Human Ecology and Tropical Land Settlement in Latin America. Boulder: Westview. 1990 "Private and public colonization schemes in Amazonia." in D.Goodman, and A. Hall eds.. The Future of Amazonia: Destruction or Sustainable Development? London: Macmillian Press. Muller, Geraldo 1980 "Sao Paulo ~ o niicleo do padrao agrario modemo." In A Ouestao Agraria Hoje. 235-236

PAGE 183

172 Mueller, Bernardo, Lee Alston, Gary Libecap and Robert Schneider. 1994 "Land, property rights and privatization in Brazil." The Quarterly Review of Economics and Finance. 34, 261-280 Nair, P.K.R. (ed) 1989 Agroforestrv Systems in the Tropics. London: Kluwer Academic Publishers. Nissley, Charles 1966 Acre: An Amazonian Frontier of Brazil. Gainesville: Ph.D. Dissertation, University of Florida Oliver-Smith, Anthony and Art Hansen 1982 "Introduction: Involuntary migration and resettlement: Causes and context." In Art Hansen and Anthony Oliver-Smith, eds., Involuntary Migration and Resettlement: The Problems and Responses of Dislocated People. Boulder: Westview Press Ozorio de Almeida, 1992 The Colonization of the Amazon. Austin: University of Texas Press Ozorio de Almeida, Anna Luiza, and Joao S. Campari 1995 Sustainable Settlement in the Brazilian Amazon. Oxford: Oxford University Press. Sanderson, Stephen 1995 Social Transformations: A General Theory of Historical Development. Oxford, UK: Blackwell. Sawyer, Donald 1984 "Frontier expansion and retraction in Brazil." In M. Schmink and C. Wood, eds. Frontier Expansion in Amazonia. Gainesville: University of Florida Press. Schmink, Marianne 1988 "A case study of the closing frontier in Brazil," In D. Attwood, Bruneau, T.C., and Galaty, J.G, eds., Power and Poverty: Development and Development Projects in the Third World. Boulder: Westview. 1992 "Amazonian resistance movements and the international alliance." In L.A. Koinski ed., Ecological Disorder in Amazonia: Social Aspects. Rio de Janeiro: UNESCO/ISSC/Educam, 149-172. Scheaffer, Richard, William Mendenhall, and Lyman Ott 1990 Elementary Survey Sampling. Belmont, California: Duxbury Press Schmink, Marianne and Charles Wood 1992 Contested Frontiers in Amazonia. New York: Columbia.

PAGE 184

173 Schnieder, Robert 1993 Land Abandonment. Property Rights and Agricultural Sustainabilitv in the Amazon. LATEN Dissemination Note #3, World Bank, Washington, D.C. Schwartzman, Stephan 1989 "Deforestation and popular resistance in Acre: From local movements to global network" Paper presented at the American Anthropological Association meetings, Washington D.C. Slinger, Vanessa Anne Vere 1996 Analysis of a Planned Agroforestrv System in Amazon Urban Resettlement: A Case Study of the Pdlo Municipal de Produ(^ao Agroflorestal of Acre. Brazil. Gainesville: M.A. Thesis, University of Florida Smith, Nigel 1980 Rainforest Corridors: The Transamazon Colonization Scheme. Los Angeles: University of California Press. Smith, N., E.A.S. Serrao, P.T. Alvim, and I.C. Falesi 1995 Amazonia. Resiliency and Dynamism of the Land and Its People. New York: United Nations University Press. Tavares, Ricardo 1995 "Land and democracy: Reconsidering the agrarian question." NACLA Report on the Americas 28 (6): 23-43. Uhl, Christopher, Robert Buschbacher, and E. A. Serrao 1988 "Abandoned pasture in Amazonia, L Patterns of plant succession." Journal of Ecology 76:663-681. Valverde, Orlando; Ana Maria Teixeira Lopes, Angela de Moraes Neves, Eugenia Gon9alves Egler, Irio Barbosa de Costa, Irene Garrido Filha, Myriam Guiomar Gomes Coelho Mesquisa. 1989 A Organizapao de Espapo na Faixa da Transamazonica Volume 2: Acre e Regioes Vizinhas IBGE: Rio de Janeiro Weinstein, Barbara 1983 The Amazon Rubber Boom 1850-1920. Stanford: Stanford University Press. Witcover, Julie, and Stephen A. Vosti 1996 Alternatives to Slash-and Bum Agriculture (ASB): A Characterization of Brazilian Benchmark Sites of Pedro Peixoto and Theobroma. August/September 1994 Washington D.C. : International Food Policy, Technology, and Institute

PAGE 185

174 Wood, Charles 1983 "Peasant and capitalist production in the Brazilian Amazon: a conceptual framework for the study of frontier expansion." In E.F. Moran, ed., The Dilemma of Amazonian Development Boulder: Westview. Wood, Charles, and Marianne Schmink 1979 "Blaming the victim: small farmer production in an Amazon colonization project." Studies in Third World Societies 7:77-93.

PAGE 186

BIOGRAPHICAL SKETCH Peter Cronkleton was bom and raised in Iowa. In 1986, he received a Bachelor of Science degree from Iowa State University with a double major in anthropology and history. Shortly after graduating he entered the Peace Corps and spent nearly 3 years working in rural Haiti and Paraguay. During this time, Mr. Cronkleton developed a strong interest in rural development and agricultural extension. This interest brought him to the University of Florida in 1991 to study applied anthropology. After analyzing a beekeeping extension programs in the Brazilian state of Acre, he was awarded a Masters of Arts degree in 1993. He returned to Acre again in 1994 to lay the ground work for the study described in this dissertation. To complete this research, Mr. Cronkleton received funding from the NorthSouth Center at the University of Miami to cover the 1995 field work. His 1996 research was fiinded by the University of Florida's Tropical Conservation and Development Program, and USAID. 175

PAGE 187

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Marianne Schmink, Chair Professor of Anthropology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Maxine Margolis ^ Professor of Anthropology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Gerald Murray Associate Professor of Anthroj I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Anthon5^6liver-Smith Professor of Anthropology

PAGE 188

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fiilly adequate, in scope and quality, as a dissertation for the degree of Doctor of PhilosophiL_„^ •Petef Hildebrand Professor of Food and Resource Economics This dissertation was submitted to the Graduate Faculty of the Department of Anthropology in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May, 1998 Dean, Graduate School


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E4TWQW219_S8CNWF INGEST_TIME 2015-01-09T21:06:34Z PACKAGE AA00026592_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES