<%BANNER%>

Community-Based Forest Enterprises in Brazil and Mexico

Permanent Link: http://ufdc.ufl.edu/UFE0041510/00001

Material Information

Title: Community-Based Forest Enterprises in Brazil and Mexico Timber Production and Commercialization Models, Market Engagement, and Financial Viability
Physical Description: 1 online resource (146 p.)
Language: english
Creator: Humphries, Shoana
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: acre, amazonas, brazil, certification, community, enterprise, financial, forest, management, mexico, para, viability
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: COMMUNITY-BASED FOREST ENTERPRISES IN BRAZIL AND MEXICO: TIMBER PRODUCTION AND COMMERCIALIZATION MODELS, MARKET ENGAGEMENT, AND FINANCIAL VIABILITY By Shoana Humphries May 2010 Chair: Karen Kainer Major: Forest Resources and Conservation Community-based forest management (CFM) is an integral component of sustainable forest management planning and implementation in the Brazilian Amazon and Southeast Mexico, where communities control substantial forest areas. Successful implementation of CFM would provide multiple benefits to Brazilian and Mexican society including a higher standard of living for rural communities, improved governance of natural resources, and the conservation of biological diversity. However, after 25 years in Mexico and 15 years in Brazil of governmental and civil organization support for community-based forest enterprises (CFEs), surprisingly little is known about CFE financial viability and the factors influencing CFE market success. This study examined CFEs in the Brazilian Amazon and Southeast Mexico from three socio-economic perspectives: 1) a descriptive analysis of how and why CFEs were developed in each region and of their current models regarding administration, production, and commercialization, 2) an econometric analysis of the impacts of several internal and external factors on prices and market access for CFEs in Southeast Mexico; and 3) a profitability analysis for three CFEs in the Brazilian Amazon. The first set of results found that while community forestry was important to helping communities formalize their tenure rights in both regions, this process was realized in different contexts, especially regarding policies and markets, and resulted in distinct CFE models. Our analysis also identified general trends in CFE models and important lessons that each region has for the other. The second set of results demonstrated that product characteristics have the highest impact on price, but maintaining forest certification over time, gaining access to new markets, and developing relationships with buyers over time can also improve prices. The third set of results, for three CFEs in the Brazilian Amazon that differed in scale, intensity, and products, revealed that CFEs can be profitable, and highlighted the importance of maximizing economies of scale, especially through cost sharing among CFEs. These findings provide insight on the challenges and opportunities for improving CFE financial viability, and highlight the need for policies that support community-based forest management as an important land-use activity for forest conservation and local economic development.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Shoana Humphries.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Kainer, Karen A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041510:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041510/00001

Material Information

Title: Community-Based Forest Enterprises in Brazil and Mexico Timber Production and Commercialization Models, Market Engagement, and Financial Viability
Physical Description: 1 online resource (146 p.)
Language: english
Creator: Humphries, Shoana
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: acre, amazonas, brazil, certification, community, enterprise, financial, forest, management, mexico, para, viability
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Forest Resources and Conservation thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: COMMUNITY-BASED FOREST ENTERPRISES IN BRAZIL AND MEXICO: TIMBER PRODUCTION AND COMMERCIALIZATION MODELS, MARKET ENGAGEMENT, AND FINANCIAL VIABILITY By Shoana Humphries May 2010 Chair: Karen Kainer Major: Forest Resources and Conservation Community-based forest management (CFM) is an integral component of sustainable forest management planning and implementation in the Brazilian Amazon and Southeast Mexico, where communities control substantial forest areas. Successful implementation of CFM would provide multiple benefits to Brazilian and Mexican society including a higher standard of living for rural communities, improved governance of natural resources, and the conservation of biological diversity. However, after 25 years in Mexico and 15 years in Brazil of governmental and civil organization support for community-based forest enterprises (CFEs), surprisingly little is known about CFE financial viability and the factors influencing CFE market success. This study examined CFEs in the Brazilian Amazon and Southeast Mexico from three socio-economic perspectives: 1) a descriptive analysis of how and why CFEs were developed in each region and of their current models regarding administration, production, and commercialization, 2) an econometric analysis of the impacts of several internal and external factors on prices and market access for CFEs in Southeast Mexico; and 3) a profitability analysis for three CFEs in the Brazilian Amazon. The first set of results found that while community forestry was important to helping communities formalize their tenure rights in both regions, this process was realized in different contexts, especially regarding policies and markets, and resulted in distinct CFE models. Our analysis also identified general trends in CFE models and important lessons that each region has for the other. The second set of results demonstrated that product characteristics have the highest impact on price, but maintaining forest certification over time, gaining access to new markets, and developing relationships with buyers over time can also improve prices. The third set of results, for three CFEs in the Brazilian Amazon that differed in scale, intensity, and products, revealed that CFEs can be profitable, and highlighted the importance of maximizing economies of scale, especially through cost sharing among CFEs. These findings provide insight on the challenges and opportunities for improving CFE financial viability, and highlight the need for policies that support community-based forest management as an important land-use activity for forest conservation and local economic development.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Shoana Humphries.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Kainer, Karen A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041510:00001


This item has the following downloads:


Full Text

PAGE 1

1 C OMMUNITY BASED FOREST ENTERPRISES IN BRAZIL AND MEXICO : TIMBER PRODUCTION AND COMMERCIALIZATION MODELS, MARKET ENGAGE MENT, AND FINANCIAL VIABILITY By SHOANA HUMPHRIES 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 2010

PAGE 2

2 2010 Shoana Humphries

PAGE 3

3 To my family

PAGE 4

4 ACKNOWLEDGMENTS I will begin by thanking my committee members. I am especially grateful to my advisor, Karen Kainer, for her tireless support and guidance. I have enjoyed working with her immensely over the last seven years, and her high standards for research, writing, teaching, and collaboration have helped me become th e professional I am today. I am also deeply appreciative of Tom Holmes support and collaboration. I am grateful that our time and research interests in the Brazilian Amazon overlapped, and that we were able to join forces. Ronald Ward has also earned m y highest respect for his patience and tireless assistance in teaching me econometric analysis and new software. I am also very grateful to Janaki Alavalapati for his assistance in the early development of the research questions explored in this dissertat ion and for his continued support. I thank Christine Overdevest for her shared interest in forest certification as a research topic, and in helping me develop ways to investigate its impacts. Additionally, I will be forever grateful to Michelle Zweede and the USDA Forest Service Office of International Programs for the financial support of my field work and preparation of this manuscript; Michelles continued encouragement and enthusiasm is also much appreciated. This research would not have been possible without the collaboration of many people in both Brazil and Mexico. In Brazil, I am very grateful to Marcus Vinicio, of EmbrapaAcre, and Paulo Amaral, of IMAZON, for their assistance in making contacts with community based forest enterprises (CFEs) in t he country and for valuable discussions of my research topics. I am also very grateful to Johann Zweede of TFT for putting me in touch with Michelle Zweede and Tom Holmes that contact made a huge difference in my PhD trajectory and experience! I also w ant to thank Edson Cruz and

PAGE 5

5 the staff of of Amb, Gabriel Riva Koury, of IDESAM, and Rosana de Miranda Rocha, Miriam Marmontel, and the other staff of Mamirau Sustainable Development Institute, for their invaluable collaboration for the financial analys is study for this research. In Acre, I am very grateful for continued support of my research on community forestry in the state to: Henrique Araujo, Embrapa; Secretary Carlos Ovdio Duarte Rocha and Domingos Mesquita, SEF; Adriano Trentin Fassini, COOPERF LORESTA; and Nivea Marcondes Jefferson Amaro, and Evandro Araujo, CTA. In Par, I am very grateful to Toby McGrath, of INPA, for providing office support in Santarm and continued interest in my work, as well as to Marlon Menezes, Rossynara Aguiar, and C arlos Ramos for helping me organize my visits to CFEs along the Amazon River. In Amazonas, I want to thank Rubens Gomez, OELA, and Almerindo Jorge Silveira Mamede Neto for their support and help in making contact with community leaders, and to Leonardo Pacheco, Instituto Chico Mendes, for facilitating my contact with different government agencies in Manaus. Ricardo Ludke, of ProManejo, and Wolfram Maennling, of GTZ/ProManejo, were also very generous with their time and in sharing information with me about this instrumental program for supporting community forestry throughout Brazil. I am very grateful for their invitations to participate in ProManejo events. In Mexico, I am very grateful to Rosa Ledesma, of the Organization of Forestry Ejidos in the Zona Maya (OEPFZM), for serving as an early collaborator when I was first developing my research plan for Quintana Roo, and for her continued support in making contact with CFEs in the northern part of the state and in collecting data. Her colleague Ofelio Guate Biempica was also very helpful in visiting different ejidos To Hugo Galletti and Jesus Vargues, of the Society of Forest Producers in Quintana Roo, I am

PAGE 6

6 very appreciative for their generous assistance in helping me meet CFE leaders and collect dat a. I am also very grateful for the generous support and help from: Alfonso Arguelles, Tropica Rural Latino (TRL); Jose Areola Palacios and Sebastian Proust, Uyoolche; Jorge Sosa, Chactemal; and Escobar Ruiz, Kanan Kaax. Finally, I am very thankful to Li no Martinez, of SEMARNAT, for generously giving his time to discuss the status of community forestry in the state and providing access to CFE timber transportation records. I am also extremely grateful to all of the community leaders that took time to me et with me and share information and thoughts in both Brazil and Mexico! I hope to continue to work with these folks throughout my career. Next I would like to express my sincere appreciation to the WFT and TCD programs for financially supporting my academic development through many different fellowships and travel grants. I am also extremely grateful to these two programs for fomenting and supporting an amazing community of professors and students that I feel incredibly fortunate to have been a part of for the last seven years. Part of this community are dear f riends who also housed me during part of my dissertation field work : Christine Lucas, Leonardo Pacheco and Maria DiGiano Thank you for your warmth and generosity you made field work so muc h more enjoyable! I also want to offer my most heartfelt thanks to my closest friends whose company and support made graduate school such an incredible experience: Dave and Ellie Buck, Dave and Diana Castine, Alex Cheesman, JG Collomb, Maria DiGiano, Amy Duchelle, Anna Prizzia, Cara Rockwell Lisa Seals, and Claudia Stickler. And thank you to my

PAGE 7

7 d issertation support group -JG Collomb, Luke Rostant, and Lucas Fortini -for all your help in the early days of writing up Finally I want to express my heart felt gratitude to my family for their unwavering support: Travis and Laurie Humphries, and Tia, Stacy, Ashlin, Reagan, Grandma Shirley, and Grandma Billie. I really appreci ate your encouragement and love!

PAGE 8

8 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES .......................................................................................................... 11 LIST OF FIG URES ........................................................................................................ 12 LIST OF ABBREVIATIONS ........................................................................................... 13 ABSTRACT ................................................................................................................... 15 CHAPTER 1 INTRODUCTION .................................................................................................... 17 Community Based Forest Management and Ent erprises in the Tropics ................. 18 Community based forest enterprises in the Brazilian Amazon and Southeastern Mexico ................................................................................................................. 22 This Study ............................................................................................................... 23 2 COMMUNITY BASED FOREST ENTERPRISE MODELS IN THE BRAZILIAN AMAZON AND SOUTHEASTERN MEXICO .......................................................... 25 Introduction ............................................................................................................. 25 Methods .................................................................................................................. 26 Results and Discussion ........................................................................................... 28 Key Historical Events that Led to CFEs ............................................................ 28 CFE Models and Factors that Shaped Them ................................................... 33 Administration ............................................................................................ 33 Production and commercialization strategies ............................................. 36 Important Trends and Lessons ............................................................................... 41 3 HOW DO CERTIFICATION AND OTHER FACTORS AFFECT TIMBER PRICES AND MARKET ACCESS FOR COMMUNITY BASED FOREST ENTERPRISES? A QUANTITATIVE ANALYSIS IN SOUTHEasteRN MEXICO ... 52 Introduction ............................................................................................................. 52 C ommunity Based Forest Enterprises in Southern Mexico ..................................... 53 Methods .................................................................................................................. 57 Models Specification ........................................................................................ 58 Data Collection ................................................................................................. 63

PAGE 9

9 Results .................................................................................................................... 65 Independent Variable Correlations ................................................................... 66 Logs ........................................................................................................... 66 Boar ds ....................................................................................................... 67 Market access ............................................................................................ 68 Price Model Results ......................................................................................... 68 Logs ........................................................................................................... 68 Boards ....................................................................................................... 69 Market Access Model Results .......................................................................... 71 Discussion .............................................................................................................. 71 Factors Which CFEs Control ............................................................................ 72 Processing ................................................................................................. 72 Selling in non local markets ....................................................................... 73 Collaboration .............................................................................................. 74 Certification ................................................................................................ 75 Relationships with buyers .......................................................................... 77 Factors Over Which CFEs Hav e Less Control ................................................. 78 Conclusions ............................................................................................................ 79 4 ARE COMMUNITY BASED FOREST ENTERPRISES IN THE TROPICS FINANCIALLY VIABLE? CASE STUDIES FROM THE BRAZILIAN AMAZON ...... 9 2 Introduction ............................................................................................................. 92 Community Based Forest Enterprises in the Brazilian Amazon .............................. 94 Case Studies .......................................................................................................... 96 Methods .................................................................................................................. 99 Data Collection ................................................................................................. 99 Workshops ............................................................................................... 100 Case study specifications ........................................................................ 100 Data Analysis ................................................................................................. 103 Results .................................................................................................................. 105 Costs .............................................................................................................. 105 Income............................................................................................................ 105 Profit ............................................................................................................... 106 Discussion ............................................................................................................ 107 What Makes Community Timber Production So Expensive? ......................... 107 What Factors Affect Revenue? ....................................................................... 108 How Do These Results Comp are to Studies of Other Community Based Forest Enterprises? ..................................................................................... 111 Methodological Innovations and Limitations ................................................... 113 Future Funding for CFEs in Brazil .................................................................. 113 Key Lessons Learned ..................................................................................... 114 5 CONCLUSIONS ................................................................................................... 127

PAGE 10

10 APPENDIX A CFES IN THE BRAZILIAN AMAZON INCLUDED IN THE STUDY ....................... 131 B SAMPLING STRATEGY FOR MEXICAN CFES ................................................... 133 C CFES IN QUINTANA ROO INCLUDED IN THE STUDY ...................................... 135 D TREE SPECIES IN EACH VALUE CLASS BY CFE ............................................. 137 LIST OF REFERENCES ............................................................................................. 138 BIOGRAPHICAL SKETCH .......................................................................................... 146

PAGE 11

11 LIST OF TABLES Table page 2 1 Key historical events shaping CFEs in Southeastern Mexico and the Brazilian Amazon. ............................................................................................................. 46 2 2 Socio economic factors affecting CFE development in So utheastern Mexico and the Brazilian Amazon. .................................................................................. 48 2 3 Number of CFEs producing each primary and secondary timber product in Southeast ern Mexico and the Brazilian Amazon. ............................................... 49 3 1 Variables and expected signs for the price models for CFEs in Quintana Roo, Mexico. ............................................................................................................... 85 3 2 Variables and expected signs for the market access model for CFEs in Quintana Roo, Mexico. ....................................................................................... 87 3 3 Variables used to stratify the population of CFEs with permission to harvest timber in Quintana Roo, Mexico. ........................................................................ 87 3 4 Summary statistics for sales of logs and boards with price data (n=3,668 sales) in Quintana Roo, Mexico. ......................................................................... 88 3 5 Results of price models for logs (n=344) and boards (n=3314) in Quintana Roo, Mexico. ....................................................................................................... 90 3 6 Summary statistics and results of market access probit model (n=4709) in Quintana Roo, Mexico. ....................................................................................... 91 4 1 Summary characteristics of the three CFEs in the Brazilian Amazon. .............. 121 4 2 Input cost categories included for each of three CFEs in the Brazilian Amazon. ........................................................................................................... 122 4 3 Costs for each of three CFEs in the Brazilian Amazon by activity and type (USD, February 2008). ..................................................................................... 123 4 4 Income data for each of three CFEs in the Br azilian Amazon (USD, February 2008). ............................................................................................................... 123 4 5 Costs, income, and profit for the three CFEs in the Brazilian Amazon (USD, February 2008). ................................................................................................ 124 4 6 Financial case studies of CFEs. ....................................................................... 125

PAGE 12

12 LIST OF FIGURES Figure page 2 1 Division of responsibilities for forest management activities among CFE actors. ................................................................................................................. 50 2 2 Proportion of CFEs implementing each value added strategy ............................ 51 2 3 Proportion of CFEs selling each product ............................................................ 51 3 1 Simulations of average price across independent variables compared to average price across all variables ...................................................................... 81 3 2 Simulations of average board prices across indep endent variables compared to overall average price ...................................................................................... 82 3 3 Simulations of the probability of a CFE sale being in the local market across independent variables compared to average probability of a sale being in the local market ........................................................................................................ 84 4 1 Locations of the three CFE cases analyzed within Brazils legally defined Amazon region. ................................................................................................ 117 4 2 Cost by forest management activity (as proportion of total) for three CFEs in the Brazilian Amazon. ....................................................................................... 117 4 3 Cost by category (as proportion of total) for three CFEs in the Brazilian Amazon. ........................................................................................................... 118 4 4 Cost of production per cubic meter for three CFEs in the Brazilian Amazon. ... 118 4 5 Price received per product for three CFEs in the Brazilian Amazon. ................ 119 4 6 Total production costs and timber sale revenues for three CFEs in the Brazilian Amazon .............................................................................................. 119 4 7 Per unit production costs for ACAF and average weighted price received as a function of total volume produced, relative to actual volume produced (100%). ............................................................................................................. 120

PAGE 13

13 LIST OF ABBREVIATION S ACAF Community Agriculture and Forest Extraction Association bf board feet BOLFOR Bolivian Sustainable Forest Management BNDS National Bank for Sustainable Development CFE community based forest enterprise CFM community based forest management CIFOR Center for International Research COOMFLONA Mixed Cooperative o f Tapajos Green National Forest COOPERFLORESTA Cooperative of Community Forest ry Producers CTA Center for Amazonian Workers EMBRAPA Brazilian Agricultural Research Corporation FSC Forest Stewardship Council IEB International Institute for Education of Brazil ha hectares HPM Hedonic price method IBAMA Brazilian Institute of Environment and Renewable Natural Res ources IMAZON Amazon Institute of People and the Environment IPAM Amazon Environmental Research Institute ITTO International Tropical Timber Organization km kilometer(s) m3 cubic meters MSDI Mamirau Sustaina ble Development Institute NGO nongovernmental organization PFML Public Forest Management Law

PAGE 14

1 4 PPG 7 Pilot Project for Tropical Forest Protection ProManejo Project to Assist Sustainable Forest Management in the Amazon R$ Brazilian Reais (Brazilian currency) RIL Reduced impact logging RBM Mayan Biosphere Reserve SEMARNAT Secretary of Environment and Natural Resources SEF Secretary of the Forest SFB Brazilian Forest Service SO support organization TC Total cost TR Total revenue USAID United States Agency for International Development USD United States Doll ar WWF World Wildlife Fund IUCN International Union for Conservation of Nature

PAGE 15

15 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 C OMMUNITY BASED FOREST ENTERPRISES IN BRAZIL AND MEXICO : TIMBER PRODUCTION AND COMMERCIALIZATION MODELS, MARKET ENGAGE MENT, AND FINANCIAL VIABILITY By Shoana Humphries May 2010 Chair: Karen Kainer Major: Forest Resources and Conservation Com munity based forest management (CFM) is an integral component of sustainable forest management planning and implementation in the Brazilian Amazon and Southeast Mexico where communities control substantial forest areas. Successful implementation of CFM would provide multiple benefits to Brazilian and Mexican society including a higher standard of living for rural communities, improved governance of natural resources, and the conservation of biological diversity. However, after 25 years in Mexico and 15 years in Brazil of governmental and civil organization support for community based forest enterprise s (CFEs), surprisingly little is known about CFE financial viability and the factors influencing CFE market success. This study examined CFEs in the Brazilian Amazon and S outheast Mexico from three socio economic perspectives: 1) a descriptive analysis of how and why CFEs were developed in each region and of their current models regarding administration, production, and commercialization, 2) an econometric analys is of the impacts of several internal and external factors on prices and market access for CFEs in S outheast Mexico; and 3) a profitability analysis for three CFEs in the Brazilian Amazon. The first set of results

PAGE 16

16 found that while community forestr y was important to helping communities formalize their tenure rights in both regions this process was realized in different contexts, especially regarding policies and markets, and resulted in distinct CFE models. Our analysis also identified general trends in CFE models and important lessons that each region has for the other. The second set of results demonstrated that product characteristics have the highest impact on price, but maintaining forest certification over time, gaining access to new markets, and developing relationships with buyers over time can also improve prices. The third set of results, for three CFEs in the Brazilian Amazon that differed in scale, intensity, and products, revealed that CFEs can be profitable, and highlighted the import ance of maximizing economies of scale, especially through cost sharing among CFEs. These findings provide insight on the challenges and opportunities for improving CFE financial viability, and highlight the need for policies that support community based f orest management as an important landuse activity for forest conservation and local economic development.

PAGE 17

17 CHAPTER 1 INTRODUCTION Since the early 1980s community based forest management has been gaining in popularity as a strategy to conserve forests and improve local livelihoods, particularly in the tropics (Scherr et al., 2004; Amaral and Amaral Neto, 2005; Bra y et al., 2005; Sabogal et al., 2008) This strategy has coincided with rapid devolution of forested lands to communities (Agrawal, 1999; Stone and d' Andr ea, 2001) such that nearly onefourth of forests in developing countries is owned and/or controlled by low income forest communities (White and Martin, 2002) and this global trend continues (Sunderlin et al., 2008). In the 1970s, the first community forestry initiatives emerged as a way to meet peoples immediate subsistence needs on a local basis (Wiersom, 2009) India, Vietnam, and Nepal were early leaders in this area (Charnley and Poe, 2007) Community forestry has evolved, however, into a commercial activ ity in which communities around the world are competing with companies in local, national, and global timber markets (Wiersom, 2009) Latin America became a hub of community based forest management (CFM) in the mid 1980s, and Mexico and Brazil, the focal countries in this study, have emerged as leaders i n both the region and the tropics (Amaral and Amaral Neto, 2005) Southeastern Mexico and the Brazilian Amazon have large numbers of CFEs and many have been certified through Forest Stewardship Council system. Key events in each region fostered the growth of CFEs, and different CFE models emerged under various influences. This study aims to better understand the financial viability of CFEs

PAGE 18

18 by examining their origin in southeastern Mexico and the Brazilian Amazon, current CFE models, and factors affecting market engagement and profitability. Community Based Forest Management and Enterprises in the Tropics Community based forest enterprises (CFEs) are defined in this study as a group of people living relatively near each other that collaborate to produce and/or sell timber resources.1 They are usually linked to each other through legal agreements, such as membership in the same producers association or cooperative, which is typical of Brazil, or legally recognized membership in community, which is typical of the Mexican ejidos ( communities who received communal ownership of lands as part of agrarian reform in the 20th century ) C ommunity based forest management in the tropics has a long history (Arnold, 1998) Communities have been the only source of commercially important nontimber forest products, such as rubber in Brazil and chicle in Mexico in 19th and early 20th centuries (Keck, 1995; Merino Perez, 2004) and Brazil nuts in the Amazon basin up to the present (Kainer et al., 2003) However, especially in the cases of rubber and chicle, these were often under forced labor or debt slavery conditions (Dean, 1987; Merino et al., 1997) Communities didnt receive the rights t o own and/or control these resources, especially timber, until the latter part of the 20th century (White and Martin, 2002) Klooster and Aminakudige (2005) provide a very insightful overview of three interrelated transformations of theory and practice that led to the emergence of community forestry. First, the po litical pressure that led to a move away from authoritarian regimes to more 1 See Glasmeier and Farragan (2005) for a comprehensive review of definitions of community forestry.

PAGE 19

19 democratic systems in many countries in late 20th cen t tury also influenced perceptions of development strategies. This led aid agencies and development practitioners to question a nd abandon topdown approaches, and plac e more emphasis on decentralization, participation, empowerment, and grassroots action in development projects (Klooster and Ambinakudige, 20 05) with increased interest and attention to common property and collective resource management (Richards, 1997; Agrawal and Ostrom, 2001) Third, social movements were instrumental in supporting community forestry (Amaral and Amaral Neto, 2005; Klooster and Ambinakudige, 2005; Bray et al., 2006) ; civil societies in Brazil originally focused on health and social rights became major actors in community forestry projects in Brazil (pers. obs.). Now communities are increasingly important players in the global timber industry, especially f or tropical timber (Forster et al., 2003) as well as in efforts to conserve forests (Agrawal and Gibson, 1999) and prevent climate change (Murdiyarso and Skutsch, 2006) Nonetheles, the development of community forestry and community based forest enterprises (CFEs) has not advanced at the same r ate in all regions. In India, CFM, often in the context of joint forest management for timber between communities and government agencies, became an important strategy for forest resource management with a new forest policy in 1988, and in 2002 there were over 60,000 village committees participating in forest management all over India (Conroy et al., 2002; Klooster and Ambinakudige, 2005) In contrast, in Africa, although communities h ave a long history of managing resources collectively, they still seldom have the legal rights to the land they live on or the timber resources (Klooster and Ambinakudige, 2005; Molnar et al. 2006) This is changing though with new laws, e.g.

PAGE 20

20 in Cameroon (see Ezzine de Blas et al., 2009). The Americas, where devolution of land rights to communities has progressed significantly since the 1930s, have become the leaders in community forestry. In Mesoamerica, according to Forster et al. (2003), there were community forestry operations more than 50 years ago and their number grew to more than 150 CFEs in the region. In Honduras, community groups were given the right to extract logs in the mid 1970s, though the short term and bureaucratically complicated agreements have not changed much since (Forster et al., 2003) Community forestr y really began to flourish in the region in the mid 1980s when ejidos in the southeastern Mexico state of Quintana Roo, who owned 95% of the states forest cover, were given the legal right to harvest their timber resources and a large project was set up t o provide technical and financial assistance to these new CFEs (Galletti, 1998; Forster et al., 2003; Klooster and Ambinakudige, 2005) In the mid 1990s, communities in the Mayan Biosphere Reserve (RBM in Spanish) were given rights to l ong term forestry concessions for sizeable areas of tropical forests (Gretzinger, 1998) They also benefitted from a large USAID funded project aimed at building technical and marketing capacity of the new CFEs (pers. obs.). The Mexican ejidos (communally owned lands) and a CFE in Honduras (COATLAHL) were among the first forestry operations in the world to obtain certification in t he late 1980s, and the communities in the RBM were required to obtain certification to maintain their concessions (Gretzinger 1998; Molnar et al., 2006) In Costa Rica, community forest management was used as a way to secure land tenure and encourage social organization for approximately 1300 families (Mendes, 1999, in Amaral and Amaral Neto, 2005).

PAGE 21

21 In South America, Bolivi a redistributed large areas of forests to indigenous communities and changed its forestry law in the 1990s to allow communities to manage their timber resources (Nittler and Nash, 1999; de Jong et al., 2006) The BOLFOR project, funded by USAID, pu t substantial support towards these new CFEs and research on tropical silviculture techniques (Nittler and Nash, 1999; Fredericksen et al., 2003) In Brazil, since the late 1980s the land granted to communities under long term tenure rights has grown to 60% of public forests (almost all in the Amazon region). The ProManejo program supported the growth and consolidation of CFEs in the Amazon region from 1999 to 2007. Community forestry was the focus of two large pilot projects in the Peruvian Amazon and central highlands in the mid 1980s, but these initiatives were not able to overcome management and marketing problems (Klooster and Ambinakudige, 2005) While CFEs in Latin America enjoyed substantial technical and financial support, much less effort was put into the marketing side of operatio ns. Thus, the CFEs have faced the same operational challenges and market barriers of small timber enterprises around the globe. These include limited access to capital as well as technical and financial assistance (beyond short term projects), low volumes of high value species due to very biologically diverse forests, limited market intelligence and business experience, and poor transportation and communication infrastructure (Amaral and Amaral Neto, 2005; Bray et al., 2005; Molnar et al., 2007 ) These limitations result in difficulties maintaining product quality and quantities to the detriment of customer relations (Forster et al. 2003)

PAGE 22

22 The growth in importance of communities in control ling tropical forest resources, has warranted more in depth research on community forest management as a critically important landuse. This study will contribute to a growing body of literature based on case studies of specific CFEs, but also goes beyond by offering comparative analysis at an indepth as well as broader level, including all of the CFEs in a large portion of the Brazilian Amazon and a sizable sample in southeastern Mexico two hotspots for CFM in tropical Latin America. Community based fo rest enterprises in the Brazilian Amazon and S outhe astern Mexico Community forest management is promoted by a variety of actors for different reasons among diverse forestland holders in southeaste rn Mexico and the Brazilian Amazon. The context for CFEs in Quintana Roo, Mexico will be presented first, followed by a description of the Brazilian Amazon. In Quintana Roo, Mexico, timber extraction has a long history, dating back at least to the Caste Wars of the second half of the 19th century during which May an rebels, needing cash for weapons to fight the Mexican army, sold timber to the English. From the 1950s to the mid 1980s, private companies were awarded concessions to harvest timber in forests owned by communities living in ejidos (Merino et al., 1997; Galletti, 1998) In the 1980s the government gave in to political pressure and gave ejido communities the rights to manage their timber resources, and a large project financed by the Mexican government and foreign aid agencies helped build the ejido s capacity for timber production (Galletti, 1998; Bray et al., 2006) The government continued to offer smaller levels of financial assistance to the ejidos for timber operations, and local supp ort organizations provided technical support (Merino et al., 1997; Galletti, 1998)

PAGE 23

23 The level of vertical integration of the ejidos varied substantially : some s old standing trees while others operated their own tractors and sawmills. Most sales were in the local market, although some sold in national and international markets. In addition, some of the ejidos ha d an ethnic Mayan majority, while others were founded by immigrants from other parts of Mexico. Community forestry in the Brazilian Amazon began in the mid 1990s, and received substantial support from the ProManejo program, which provided funds for technical assistance, training, and equipment. Community forest management has been used as a tool by communities for establishing and demonstrating land tenure. The federal and state governments in Brazil have foment ed community forestry to encourage compliance with minimum forest cover laws, to support sustainable economic development of forest communities, and to encourage legal sources of wood products (see Amaral and Amaral Neto, 2005) Non governmental organizations (NGOs) also pl ayed important roles in promoting community forestry to further their conservation and/or sustainable developm ent objectives. CFEs operate in Brazils various types of extractive settlements and reserves in which families lived and engaged in diff erent productive activities depending on the type of reserve (see www.mma.gov.br ) and their CFE production and marketing strategies were also quite diverse Examples of varied CFE participants include ranchers living i n colonist settlement areas to fisherman who lived in flooded forests. This Study The overall aim of this study was to analyze the factors affecting financial viability for CFEs in the tropics. We did this through the examination of CFEs in the Brazilian Amazon and southeastern Mexico from three socioeconomic perspectives: (1) a

PAGE 24

24 descriptive analysis of how and why CFEs were developed in each region and of their current models regarding administr ation, production, and commercialization strategies, (2) an econometric analysis of the impacts of several internal and external factors on prices and market access for CFEs in southeastern Mexico; and (3) a profitability analysis for three CFEs in the Brazilian Amazon. The field component was implemented from July 2007 to February 2008 in the Brazilian Amazon, and April 2008 to July 2008 in Quintana Roo, Mexico. The methods used included surveys, review of forest management documents, econometric modeling, and workshops to collect and analyze CFE financial data. This dissertation is organized such that the second, third, and fourth chapters are three individual and fully structured papers. Chapter two provides key historical events in each region that led to the development of CFEs in Quintana Roo, Mexico and the Brazilian Amazon, describes and compares the contemporary CFE models, and identifies important factors that influenced the CFEs through time and across scales Chapter three presents statistical analyse s of factors affecting prices and market access for timber products in Quintana Roo, Mexico Chapter four presents profitability analysis results for three CFEs in Brazil, and explores challenges and opportunities for improving financial returns. Conclusions from the study are presented in Chapter 5.

PAGE 25

25 CHAPTER 2 COMMUNITY BASED FOREST ENTERPR ISE MODELS IN THE BR AZILIAN AMAZON AND SOUTHEASTERN MEXICO Introduction Communities around the globe are increasingly engaging in commercial forest management for timber products (Charnley and Poe, 2007; Sabogal et al., 2008) Governments, civil organizations, and donors are supporting community b ased forest enterprises (CFEs) to both improve local livelihoods and conserve forests (Sabogal et al., 2008) As communities adopt timber management as a livelihood strategy, they are both adapting conventional timber production methods to their local realities and developing innovative new models (Bray et al., 20 05; Molnar et al., 2007; Sabogal et al., 2008) CFEs in the tro pical regions of Mexico and Brazil have emerged as important references for community forestry globally, in terms of total number and areal extent of CFEs, quantity of CFEs with Forest Stewardship Council (FSC) certification, and, especially for Mexico, length of experience (Amaral and Amaral Neto, 2005; Klooster and Ambinakudige, 2005) We studied CFEs in southeastern Mexico and the Brazilian Amazon to better understand the historical events that supported the development of CFEs in these two regions, as well as in the contemporary factors that influence and challenge them. Our main research questions were: 1) What were the key historical events that led to emergence and evolution of CFEs in Southeastern Mexico and the Brazilian Amazon over time? 2) How did CFE administrative organization, production, and commercialization strategies vary within and between these two regions? 3) What were the main socioeconomic conditions (at international, national, regional levels) that

PAGE 26

26 shaped these contemporary CFEs? 4) What trends and lessons can be identified from t hese analyses? Methods W e used a political ecology framework to identify the key factors (e.g., policies, market demand) that have shaped CFEs through time. This is similar to the hierarchical approaches used by Schmink (1994) and Wood (2002) to examine drivers of tropical deforestation. This method is relevant to the forest managers in our research sites who develop and adapt their production and marketing strategies to manage their forests for timber in a context of changing policies markets and institutional relationships An extensive review of literature, field interviews, and the authors previous research in the study areas were used for the historical analysis for each region. Data for the current administration, production and commercialization strategie s came from field work in the Brazil ian Amazon (July 2007 to February 2008) and the state of Quintana Roo (QR), Mexico (April to July 2008) In Brazil, communities were identified through civil and government databases, documents, and consultations with governmental and civil organizations working in community forestry in the study area. We identified 16 CFEs that had sold timber products within the last three years, or planned to sell timber before of the end of 2008 all were included in our study. By state, these were: 9 in Acre (27 more were in development); 5 in Para ( 1 CFE had 6 member associations ; 5 others were in development); 2 in Amazonas ( 1 had 26 member associations and 4 more awaited management plan approval; one additional CFE was in development ) (App endix A ) These states were chosen because they had the highest numbers of CFEs in the region (IEB, 2006)

PAGE 27

27 In QR a list of CFEs with approved timber management plans in 20061 was obtained from state regulatory agency, SEMARNAT ( Natural Resource Harvesting and Restoration for Quintana Roo) A stratified sample of 25 ejidos was chosen b ased on three factors that appeared to differentiate the CFE models in the state: ( 1) ownership of or close proximity to a mill, (2 ) permission to harvest precious timber species and (3) FSC certifi cation (Appendi ces B and C ) The SEMARNAT Director estimated 56 of the 80 ejidos (or 70%) harvested timber in 2006 (Lino Martinez, pers. comm.), thus our sample represented 44.6% of the population of active CFEs. For each CFE in Brazil and Mexico, descriptive data was collected on CFE administration, and quantitative data was compiled on production and sales through documentation and/or interviews with CFE leaders, SO representatives (e.g., NGOs, government agencies) and secondary organizations (e.g., cooperatives, producers groups). In Brazil, sales d ata was available from contracts and/or databases ; in Mexico, copies of product transportation documents were used as sales contracts were uncommon. Based on the historical development of the CFEs and the current CFE models regarding administration, produc tion, and marketing strategies we identified key socioeconomic factors that shaped these models through time. Finally, the historical trajectory, current models, and main influencing factors identified were used to identify general trends in CFEs in bot h regions, and to flesh out lessons that CFEs in each re gion could learn from the other 1 The year 2006 was chosen to exclude ejido production in 2007. Many ejidos that do not typically engage in timber production received temporary permission to extract and transport logs after hurricane Dean struck in August 2007.

PAGE 28

28 Results and Discussion Key Historical Events that Led to CFEs International demand for nontimber forest products in the second half of the 19th and first half of the 20th centuries laid the groundwork for patterns of land settlement and forest resource use in the two regions (Table 2 1) In Brazil, the Amazon was subjected to two rubber booms: this vast region was the major source of native rubber (e.g., Hevea brasiliensis ) for tires and machinery for the industrial revolution until Southeast Asian plantation rubber began dominating the market (Weinstein, 1983; Barham and Coomes, 1994) (Table 2 2). Amazonian rubber became a critical source again during World War I I when A llied forces access to Asian plantations was cut off (Dean, 1987) In Mexico, soon after the Mexican army defeated the rebel Mayans in the Caste Wars in central and southern Q uintana R oo (QR) in 1915, the state became a major center for the production of chicle, the exudate from the chicazapote tree ( Manilkara acharas ) used to make chewing gum. In both cases, a mixture of indigenous peoples and migrants formed the work force, and to make production viable, each family need ed a large area: approximately 500 ha per family for chicle (Merino Perez, 2004) and approximately 300 to 450 ha per family for rubber (Allegretti, 1990) These householdlevel settlement patterns formed the basis for the creation of ejidos or communal land properties, in Mexico from 1935 to 1942 (Bray et al., 2006) and the formation of extractive and other sustainable use reserves in Brazil in the 1980s through the present (Allegretti, 1990; Kainer et al., 2003) In the 1950s and 1960s, the national governments in Mexico and Brazil turned their sights on tropical forests as sources of commercial wealth and import substitution. In Mexico, commercial concessions were granted on ejidal lands throughout the country

PAGE 29

29 (Merino Perez, 2004) In QR, the company MIQRO aggre ssively exploited mahogany ( Swietenia macrophyll a King) and Spanish cedar ( Cedr ela odorata L.) for plywood and other products for the international market (Galletti, 1998; Forster et al., 2003) In Brazil, a military dicta torship encouraged Amazonian development with investments in infrastructure, migratory settlement, and large scale production of cattle, agriculture, mining, and timber (Bunker, 1984; Schmink and Wood, 1987; Hecht and Cockburn, 1990) In both cases there was widespread environmental damage and litt le benefit for rural communities, discontent over which spawned grassroots social movements (Schmink and Wood, 1992; Hall, 1997; Bray et al., 2006) In the 1970s, these social movements in both countries began to call for improved land rights for communities. In Mexico, where ejidos had long term tenure to approximately 80% of the countrys forests and 95% of the forests in QR (Forster et al., 2003) communities sought right s to harvest the timber and other resources on their lands. Ejido dwellers and supporters led protests around the country, and the government finally agreed to programs to support nascent community forestry efforts (Bray et al., 2006) In Brazil, in contrast, families that had been living in rubber estates in many cases for generations, were not aware of and had not documented their squatters rights to the land, and began finding their forests and livelihoods under attack (H echt and Cockburn, 1990; Almeida, 2002) Frequently the forest ed land s were sold to ranchers, or speculators simply took possession of areas; in both cases the new owners were usually quick to clear and burn the forest, which resulted in violent conf licts with local communities (Bunker, 19 84; Almeida, 2002) A movement of rubber tappers emerged, led by Chico Mendes and others, and joined forces with

PAGE 30

30 environmental and social movements that portrayed the tappers as defenders of the forest with legitimate land rights (Hecht and Cockburn, 1990; Keck, 1995; Hall, 1997) A paradigm shift in the 1980s and 1990s away from central control and loosely regulated resource exploitation and towards local governance and environmental conservation -institutionalized in new legislation and supported by large foreign aid projects and domestic programs in both countries -paved the way for commu nity forestry. This transition happened more quickly in Mexico, where ejidos already had relatively greater land tenure security. The Plan Piloto Forestal (PPF) initiated in 1983 in QR as a joint federal and state effort dedicated substantial resources to wards preparing ejidos in QR to manage their forest resources, sell timber, and defend their rights ( Galletti, 1998) By the time the Mexican government passed a new forestry law in 1986 that ended concessions in ejidos and gave co mmunities the rights to forest resources, 10 CFEs were already harvesting and selling their own timber (Galletti, 1998; Wilshusen, 2006) ; three had purchased their own machinery and sawmills (Galletti 1998) Continued investments in the CFEs were made possible by several government funding programs initiated in the late 1990s (Bray et al., 2006) In contrast, in Brazil, it wasnt until the late 1980s/early 1990s that the first agroextractive settlements and extractive reserves were created (Allegretti, 1994) and the first few community forestry initiatives did not appear until the mid 1990s (Amaral and Amaral Neto, 2005) (Table 2 1). Most of the communities involved in the first CFEs had insecure tenure rights, and many had violent clas hes with illegal loggers and/or ranchers. The continued creation of multi use reserves by the government has given communities long term use rights to approximately 60% of Brazils public tropical forests

PAGE 31

31 (SFB and MMA, 2009) Community forestry was used by the early CFEs and continues to be used in the new reserves to help communities document and demonstrate tenure rights. A critical source of support for the expansion of community forestry was the ProManejo program (1999 2007), which supported pilot demonstration projects with training, equipment, and machinery across the region, and brought proj ect participants together to share challenges and lessons learned, and discuss policy proposals (Amaral Neto, 2002; ProManejo, 2007) The Brazilian government has been slow to adapt to the growth of community forestry: it did not recognize forest management plans at the community level or allow simplified documentation for these operations unt il 1998, and its first permanent federal program on community forestry was approved in 20092. This new community forestry program is expected to continue simplification of regulations for CFEs, which are quite onerous and time consuming (Hirakuri, 2003) especially compared to QR. In both countries, state governments were often quicker to lend support than the federal governments, and played important roles in supporti ng community forestry (Table 22 ). The governor of QR was very supportive of the states new CFEs, and even initiated a state forestry plan (PPFE) after the PPF ended (Galletti, 1998) Similarly, the Acre state government in Brazil aggress ively supported multi use community forestry in its pursuit of forest based development (Kainer et al. 2003). It played a very important role in channeling funds, technical assistance, and machinery to CFEs, and also lured timber product companies to the state that committed to 2 However, the 2006 public forests law explicitly detailed plans for increasing the area of public forest under community use, including timber production.

PAGE 32

32 purchasing certified wood from communities3. CFEs in Par and Amazonas had to rely more on short term projects with environmental and social civil organizations for technical and financial support although these groups were also very supportive of Acre CFEs (Amaral and Amaral Neto, 2005) Forest certification for CFEs was encouraged in both countries by federal and state governments and international and national NGOs; it was also encouraged by a few international buyers working in Mexico. The ejidos taking over of forest management in Mexico parralled the emergence of timber certification (see Conroy, 2007) and a few buyers from the US and Europe encouraged ejidos to get certified under the Rainforest Alliances SmartWood system, which several did in the late 1980s (Maynard and Robinson, 1998) After the Forest Stewardship Council (FSC) was established in 1993, CFEs renewed their certification under this system (Table 2 1). Only a few CFEs sold timber to buyers who requested certified wood (most buyers of Mexican timber did not care if it was certified) (Maynard and Robinson, 1998) However certification was later encouraged by national and state community for estry assistance programs as an activity eligible for funding, and as a criteria by which other funding requests were evaluated (i.e., in the evaluation of funding proposals, certification was an advantage). In Brazil, the World Wildlife Fund (WWF) condit ioned its substantial support for CFEs on obtaining FSC certification (N. Marcondes, pers. comm.), and in Amazonas, IMAFLORA (a Brazilian partner of SmartWood) helped initiate an early CFE with the objective of testing certification at the community forest ry level (see Koury, 3 The Acre state government pledged to encourage FSC certification as a way to make forest products more competitive in national and international markets (MMA (Minister of the Environment) and Government of Acre, 1999)

PAGE 33

33 2007) Certification for production of nontimber forest products was also encouraged and obtained by CFEs in both regions (e.g., chic le in QR, oils and seeds in Brazil). CFE Models and Factors that Shaped Them Three aspects of CFE models are presented here. First, the administration of the CFEs is described, including the institutional form of the main administrative body at the CFE le vel, the different CFE actors and their responsibilities, and CFE sources of support. Second, the strategies for producing timber and final products are categorized and compared. Finally, the commercialization strategies used are described and contrasted. Administration Community members administered their CFEs drawing on different institutional arrangements in each region. Mexican CFEs implemented timber management on forestland that belonged to the ejido Therefore, it was administered (in most cases) directly by the ejidal governance structure, and profits were equally distributed among all formal ejido members (usually male heads of households) (Bray et al., 2006) In some cases a committee was appointed to oversee day to day forest management and processing activities, and in some cases these roles were given to trained, permanent fulltime employees In contrast, in Brazil, CFEs were usually governed by an association of community members who chose to get involved in timber management. M anaged forests were frequently located on individual plots, and income went only to t hose involved ( although ther e was usually a provision for benefits at the community level ) In most cases in QR and half the cases in Brazil, CFEs took full responsibility for major timber production activities (Figure 2 1). Training in technical aspects of

PAGE 34

34 community forestry was st rongly encouraged in both regions (Galletti, 1998; Amaral and Amaral Neto, 2005) These CFEs implemented the various activities with technical help from a SO or through contracting service providers (e.g., renting machinery for skidding). In both regions, only a minority of CFEs that sold standing trees or logs yielded responsibilit y for harvesting and skidding to contractors. In contrast, the other half of the Brazil CFEs partnered with companies and/or cooperatives (Figure 1). In two cases, individual CFEs implemented all activities up to sales: one CFE passed sales responsibili ty to a cooperative (Oficinas Caboclas cooperative), while the other shared it (i.e., the Cooperfloresta case in which member CFEs sold some timber individually, and the rest through the cooperative). In one case a cooperative implemented all activities ( Cooperative Amb). Two CFEs had 15year partnerships with a company in which the company assumed responsibility for most of the activities, but the CFE still had a minor role in decision making (Associations Esperana and Virola Jatoba) Another CFE had a two year agreement with a local company for the provision of skidding and loading services for a share of the timber income (Association Arimum). A notable innovation in the administration of some CFEs in QR was the division of the ejido s annual approved harvest volume among work groups. Usually the harvest and sale of the total volume was overseen by the ejidal authorities. The use of work groups allowed each ejido member to allocate his share of the harvestable volume4 to the work group of his choice. The work group then appointed a group leader, who 4 Each formal ejido member is entitled to an equal share of the total annual approved timber volume. Thus, each group is entitled to the sum of shar es allotted to it. There was still only one management plan at the ejidal level, and the groups had to request timber transportation documents from the ejido president, who was legally responsible for their use.

PAGE 35

35 organized the harvest of the groups share of timber and found buyers. Permission to use work groups was written into the 1992 agrarian reform policy, which contained several neoliberal policy chang es (Bray et al., 2006) (Table 2 1). The usage of work groups was driven by different factors, including m istrust among ejido members and/or of ejido leadership, disagreement about how timber funds were managed, and/or strong family alliances (Table 22 ) Nine surveyed CFEs decided to either utilize work groups (eight), or work individually (one). Of the eight, one CFE regrouped to sell the timber at the ejidal level, four sold in the work groups and three work groups allowed their members to sell at the individual level. The one CFE that harvested timber at the individual level, regrouped to sell the ti mber a t the ejidal level An advantage of this model was that some groups invested in machinery, which they rented to other groups. Reinvestment of timber income in CFEs in QR was uncommon and problematic (Galletti, 1998) A disadvantage o f this option was that buyers faced higher transaction costs when they had to negotiate prices with individual groups instead of the ejido leadership. This was one reason some groups reverted to selling at the ejidal level (but continued conducting field activities in groups). As far as support organizations and partners, in QR, the CFEs received technical support and help in accessing state and federal financial assistance programs from 10 support organizations (SOs) that were established in the 1980s and 1990s specifically to work with CFEs (see Galletti, 1998) (Table 2 1). CFEs paid an annual membership fee to their SO based on how much timber they sold. None of the QR CFEs in our sampl e partnered with NGOs or companies, and none had received direct technical assistance for forestry from the state. However one SO was actively organizing its

PAGE 36

36 members to begin collective processing and sales, and other SOs were discussing similar possibilities. In Brazil, in contrast, state agencies, cooperatives, and companies played more important technical roles in CFEs; the agencies and cooperatives in Acre, and the latter more so in the state of Par. Company partners seemed to be more common for comm unities with little experience with forest products, e.g., settlements of migrants from southern Brazil. Funding for CFEs came mostly from short term grant money funneled to NGOs who worked directly with CFEs; the ProManejo program was also a critical source of funding (Amaral and Amaral Neto, 2005) Production and commercialization strategies Production. SOs and governments strongly enc ouraged CFEs to adopt valueadded strategies in both regions. We categorized our study CFEs into four main valueadded strategies based on the skills and equipment they had acquired5: (1) no valueadded (i.e., selling standing trees); (2) felling trees and selling logs either in the forest or skidded to a patio; (3) using a chainsaw to cut logs into rough boards (sawnwood); and (4) using a permanent or portable sawmill to cut fine or high quality boards (sawnwood) and/or carpentry tools to make finished pr oducts, such as furniture (this strategy is also referred to as full vertical integration). The proportion of CFEs who did not add any value to their standing trees was similarly small for both regions (Figure 2 2). Notably, although proportionally more CFEs in Mexico had physical capacity to produce sawnwood, more CFEs in Brazil had equipment (permanent sawmills or carpentry tools) to produce high quality sawnwood and finished products. 5 They did not always use these strategies, even though they had the capacity.

PAGE 37

37 CFEs value added strategies seemed to be based on a variety of fac tors, including SO goals, access to capital or grants, forest area, timber volumes, and ecosystems (e.g., upland versus flooded forests in Brazil) (Table 2 2). The strategies have adapted over time, as well, based on learning, and changes in these factor s. Perhaps due to the lack of diversity in models of CFEs implementing commercial timber harvests in natural forests in the 1980 and early 1990s6, especially in Latin America, the fully vertically integrated strategy of using a permanent sawmill to produce high quality sawnwood was encouraged by donors and adopted by some early CFEs in both regions. In QR, this was the model the CFEs essentially took over from MIQRO. This strategy was also adopted when two early CFEs in Brazil were formed: one small scale (Porto Dias) and one large scale (Amb). In QR, the CFEs were able to use government loan programs and grants to obtain the machinery and equipment they needed (Table 2 2); in Brazil a lot of machinery and equipment was donated (e.g., safety equipment, a sawmill) and otherwise the machinery was rented. CFE opportunities for credit in Brazil have been few and underutilized for many reasons (see Amaral and Amaral Neto, 2005 for a critique of credit programs) The CFEs who came after these pioneers generally had smaller areas and lower volumes of high value woods, and developed alternative valueadded strategies or simply sold standing trees. Many decided to utilize portable sawmills or chainsaws to cut logs into boards, especially since equipment was often donated in Brazil or attractively financed by SOs or government programs in Mexico This low tech option was also viewed positively where machinery (e.g., tractors) would be difficult and/or 6 Many early CFEs in Asia focused on small planted woodlots (Charnley and Poe, 2007)

PAGE 38

38 very damaging to use (e.g., flooded forests). However, difficulties maintaining machinery and obtaining replacement parts commonly caused production delays and compromised quality. The CFEs who sold standing trees or logs frequently were either located near cities and community members preferred to focus on other sources of income (2 cases in Mexico) and/or had little experience with forestry work (3 cases in Brazil). Some of the CFEs with processing capacity did not always choose to use it, especially in QR, and criteria for this decision varied based on factors at the state and local level (Table 2 2). In QR the m ain criteri on was species: CFE s with mills usually processed their precious woods (mahogany and Spanish cedar) or rented their mill to buyers who then processed the se logs; hard wood species were sold as logs because they were difficult to process and plyw ood mills wanted the soft wood species as logs7. Some of the CFEs that used a chainsaw to cut hardwood species into rough boards sold their precious species (too valuable to cut into rough boards) and soft woods (to plywood mills) as logs In a few other cases, CFEs with mills decided it was in their interest to sell only logs, citing as reasons : delay and difficulty in processing and drying wood, and finding buyers for processed wood. In Acre, Brazil, two pioneer CFEs, one with a permanent mill and one with a portable sawmill, ceased using this machinery when they joined a cooperative that bought and processed their logs. The younger CFEs in Brazil have not seen as many changes to their strategies. 7 These are three main categories of wood used in Quintana Roo and the surrounding region: precious wood, including mahogany and Spanish cedar, and hard and soft woods (based on wood density) (Forster et al., 2003) The terms hard wood and soft wood should not be confused with the meanings associated with these words in temperate zones.

PAGE 39

39 Based on the most recent sales for the CFEs, we found many sold a mixture of logs and finished products in both regions and relatively few sold standing trees (Table 2 3, Figure 2 3) CFEs in QR also sold as secondary products small poles for construction and decoration (for which demand grew with tourism in the 1980s and 1990s; in fact this was the sole product for many QR CFEs), and railroad ties (which were a more popular product in the 1990s). One of the sampled QR CFEs had stopped p roducing logs and boards, and was only selling small diameter poles. Commercialization Decisions on sales were us ually discussed among CFE participants and the n made by consensus (in Brazil and some QR ejidos ) or by appointed leaders (the case for most QR ejido presidents and some Brazilian CFEs with designated sales people). However, there were interesting cases of other actors helping CFEs with commercialization. In Brazil, a national certified producers group helped a few certified CFEs throughout the country sell in the national market, and the Group of Community Forestry Producers in Acre and its SOs organized a few sales for certified members in the early 2000s (Humphries and Kainer, 2006) This group also helped found a cooperative that sold members certified timber in different markets in 2006 and 2007. Cooperatives and joint negotiations have also been important i n selling CFE products in Brazil8 (Figure 2 1). In contrast, SOs in QR did not seem poised to assist certified CFEs access markets. While a SO helped facilitate joint sales among ejidos in the early years, (Galletti, 1998) CFEs later ch ose to sell individually (it is not clear why). Nonetheless, almost all of the SOs were considering the possibility of 8 The 26 CFEs selling timber in the Mamirau Sustainable Development Reserve used a strategy in which CFE representatives met one day to agree on minimum prices, and then jointly negotiate prices with buyers in a roundtable meeting the next day.

PAGE 40

40 creating central log yards for processing, drying, and selling timber for members; one had installed a sawmill and was working on sales logistics with members. CFE s in both regions sold m ost of their wood in the local market (see chapter 3 for data on QR CFEs ). Certified CFEs in both regions made sales in national markets, and QR CFEs even made a small number of sales to buyers in the US and Germany. Another exception was a cooperative of furniture makers in Par that sold to the national market. Many factors affected access to markets for CFEs in both regions (Table 2 2). Infrastructure in QR was generally good. CFEs that regularl y harvested timber had good access to roads and ports (although road quality varied), which facilitated transport to markets, electricity and water, which supported processing facilities, and basic telephone service (usually at least one working line). In contrast, for CFEs in the vast Brazilian Amazon, infrastructure was a problem: roads were difficult to maintain (although CFEs in flooded forests used rivers for log transport), few had electricity, and communications were often restricted to radio. Low levels of literacy in both regions, and low fluency in Spanish in some Mayan communities in QR, also limited business capacity; absence of written contracts in QR made it difficult for CFEs to hold buyers to agreed terms of sale. CFEs in both regions have faced stiff competition in the market. In QR, cheaper wood flooded the market after the elimination of trade barriers and modernization of national trade in the 1990s (Forster et al., 2003) Market competition came from tropical wood from Central and South America (including Brazil) and conifers from northern Mexico and elsewhere. In the Brazilian Amazon, the main competitors were

PAGE 41

41 illegal operations and legal, industrial scale operations (Table 2 2). Illegal operations do not bear the costs of preparing and implementing management plans, training workers, or paying a fair wage. Industrial scale operations have volumes that allow them to offer lower prices, and the experience and contacts to access different markets and negotiate good terms of sales. An advantage CFEs in this region have is demonstrated interest from buyers in the national and, in the case of Acre, local markets of purchasing wood that is certified and/or from communities This is not the case in QR, or at least buyers are not communicating this to CFEs, and the general consensus among CFEs and SOs was that certification had not brought the price premiums and market access they anticipated (but see Chapter 3) Important Trends and Lessons While obtaining the legal right to manage their forests for timber was important to formalize their tenure rights for communities in both the Brazilian Amazon and Southeastern Mexico, this process was realized in different contexts, especially regarding policies and markets, and has resulted in distinct CFE models. Our analysis identified some general trends in the development of CFE models, as well as some important lessons that each region could learn from the other. These should be of interest to CFEs, SOs and policy makers in other regions as well. The general tendencies regarding CFE models in the two regions was to move away from the model of each CFE having its own permanent sawmill and selling timber products individually, and to look to partnersh ips with other producers for commercialization and to companies to help new CFEs get established. Simpler valueadded processing with portable sawmills or chainsaws with guides has become more common than permanent sawmills, which are more expensive, need more energy, and

PAGE 42

42 require heavy machinery to move logs. Producer groups have provided opportunities to CFEs to share information, and/or negotiate sales jointly. Good examples include the state producers group in Acre and the 26 associations that worked w ith the Mamirau Sustainable Development Reserve, in Amazonas; members of the SOs in QR (which could be considered producer groups) shared information, but did not collaborate on sales. Cooperatives increasingly assumed responsibility in Brazil, for producing (e.g., Amb) and/or selling timber products (e.g., Oficinas Caboclas, Cooperfloresta), and were being considered in Mexico. Three CFEs established in recent years in Par were working with companies to establish timber operations infrastructure and t rain workers (2 cases) and to skid and load logs for a share of profits (1 case). It remains to be seen if FSC certification will continue to play an important role for CFEs in the two regions. The number of certified CFEs has increased in the Brazilian Amazon in recent years, while the number sharply dropped in QR in 2008. Two special policies under development by the Forest Stewardship Council for certified CFEs may help increase the benefits for and number of certified CFEs in the near future: one is dual certification for FSC and FairTrade (with the Fairtrade Labeling Organization), which is being tested for potential adoption, and another is a community origin label, which will be available in 2010. As far as lessons learned, each region provided im portant lessons for the other. The Mexican CFEs ten year advantage provided time for learning and consolidating models, yet the lessons they offer are not recent innovations. First, Mexico has a bureaucratic system for obtaining forest management plans that may not be perfect, but it has worked without major renovations for years. CFE actors at times complained it

PAGE 43

43 might take a few weeks to get their timber transportation documents. In contrast, it can take years to get a forest management plan and several months to get an annual operating plan approved in Brazil (see Koury, 2007 for an example) The documents and processes required to obtain permission to harvest and transport timber are onerous (100% inventory of trees in each harvest area), extremely complicated (requiring various licenses and fees), and timeconsuming (CFE representatives must visit the agencies in the c apital, and agency representatives must visit the CFE) (Amaral and Amaral Neto, 2005) The permission frequently comes after it is too late to harvest the wood due to annual changes in climatic conditions. The regulations are widely criticized as making it much easier to get a permit to deforest land in Brazil than to manage it for timber (Hummel, 2001; Hirakuri, 2003) The new Federal Program for Community Forestry will likely continue the process of simplifying the system for CFEs. Mexican CFEs have also enjoyed low interest government loan programs since their inception (Galletti, 1998) whereas it is very difficult for CFEs in Brazil to access credit (Amaral and Amaral Neto, 2005) Many Brazilian CFEs received donated machinery, especially processi ng equipment, through the ProManejo Program and some state programs; however it will be extremely difficult for them to maintain or replace it without access to credit (see Chapter 4). The new cooperatives may help channel funds from low interest loans t o CFEs. The Mexican model could provide lessons in how to make permanent credit programs readily available to CFEs in Brazil. Mexico also provides an important reference with its use of SOs as providers of fee based technical services. When the short term projects in the Brazilian Amazon that subsidized the salaries of technical staff for CFEs end, state extensiontype service

PAGE 44

44 or technical organizations, such as the ones working with CFEs in QR, will need to fill the void. Brazil presents interesting less ons and examples for QR CFEs with respect to marketing innovations. The producer groups and cooperatives in Brazil have been successful in gaining access for communities to new markets and improving prices received, especially for certified operations; however this strategy needs clusters of CFEs to be successful. Mexican CFEs expressed reluctance to partner with other CFEs for commercialization purposes, preferring to maintain their current models. Their SOs, however, expressed much interest in collabor ative processing and sales. It will be interesting to monitor this in coming years. Finally, these trends and lessons also provide valuable information and opportunities for reflection for CFEs and policy makers in other parts of Latin America, Africa, an d Asia. In Latin America, where 25% of the land is owned by communities (Larson et al., 2010) the other countries with the most advanced communit y forestry operations are Guatemala (Gretzinger, 1998; Carrera et al., 2006) and Honduras (Nygren, 2005) where successful cooperatives produce and export valueadded products, but also f ace many of the same bureaucratic and market challenges described in our study areas. For other countries in the region where communities have rights to large tracts of natural forest, but where community forest enterprises are not as developed, including Bolivia and Peru, our study provides potentially useful lessons in the importance of low interest loan programs and the benefits of collaboration among CFEs and with companies.

PAGE 45

45 For Africa, where by contrast governments control and manage 98% of forests (RRI, 2009) Mexico and Brazil serve as important examples of both devolution of significant areas of forests to communities through ownership ( ejidos in Mexico) or long term tenure rights (different types of reserves in Brazil), and successful community management of forests for forest conservation and income (Bray et al ., 2005; Nepstad et al., 2006; Duchelle, 2009) In the Congo Basin, Cameroon incorporated community forestry in its forest policy in the mid 1990s and is a leader in the region with several communities managing forests either on t heir own or in cooperation with companies (Ezzine de Blas et al., 2009) Ghana and the Democratic Republic of Congo have adopted similar legislation (Ezzine de Blas et al., 2009) but community forestry operations have not grown as quickly. In Asia, communities also own about 25% of the land (RRI, 2009) CFEs vary from small plantationbased production systems for fodder and firewood (more common) to larger natural forest concessions. In Nepal, FECOFUN, a large second ary organization, provides its 14,000 community forestry user groups with technical, learning, and political services It is an excellent example of the services secondary organizations can provide to CFEs. However, due to its unstable funding base (Paudel et al., 2010) it is also a candidate for learning from Mexicos support organization model of charging m embers fees for services. In the Philippines, the government began allowing access to 25year forest concessions for timber in the late 1980s, but communities still have difficulty complying with regulations due to long and expensive bureaucratic practice s (Cronkleton et al., 2010) similar to Brazil and Guatemala.

PAGE 46

46 Tabl e 2 1. Key historical events shaping CFEs in Southeastern Mexico and the Brazilian Amazon. Timeframe Quintana Roo, Mexico Brazilian Amazon 1840 1920 Caste Wars (1847 1901) First rubber boom Chicle production increases to make QR a main producer (1917) 1920 1950 First round of ejidos were formed around chiclebased livelihood (1935 1942) Second rubber boom (1939 1945) 1950 1970 MIQRO timber company operated large concession on many ejidal lands in QR (1950 1980) Forest Code (1965) established 50% legal forest reserve and required technical plans for logging Second round of ejidos were formed around agriculturebased livelihood (1960s 1970s) Government programs for infrastructure, and subsidies for agriculture, cattle, and timber 1970s Widespread deforestation for ranching and land speculation, poorly regulated logging and land grabbing, intense land conflicts Grassroots activism for community forestry national level Rubber tappers organize nationally and create alliances to call for land rights New government programs to support CFEs in ejidos 1980s New forestry law ends concessions and transfers timber rights to ejidos (1986) First extractive settlements and reserves created Plan Piloto Forestal (1983 1989) Ten CFEs formed; three obtain machinery and sawmills Decline in mahogany supply Increase in mahogany demand CFEs join forces to sell timber together Two support organizations formed (SPFQR and OEPFZM) Certification of CFEs under SmartWood

PAGE 47

47 Table 2 1 Continued. 1990s Agrarian Reform (1992) gave full ownership of land to ejidos and formalized role of work groups for timber production within ejidos UN Earth Summit in Rio de Janeiro (1992) NAFTA led to imports of cheap wood (temperate softwoods and tropical hardwoods) Pilot Program for the Conservation of the Brazilian Amazon (PPG7) provided much support for conservation and sustainable forest management UNOFOC (National Union of Community Forestry Organizations) and UNORCA (National Union of Autonomous Regional Farmers) founded Regulations require sustainable forest management plan (PMFS) for logging New CFEs produce railroad ties First pilot CFEs (dif models) Forest Stewardship Council (FSC) founded (1993) Legal forest reserve increased to 80% (1996) Several new government programs support CFM: PRODEPLAN, PRODEFOR, PROCYMAF Simplified PMFS allowed for community forest management (1998) Certification of CFEs under FSC New CFEs use chainsaws to cut logs into boards 2000s New CFEs sell mostly standing trees or logs ProManejo supports pilot CFEs and helps organize them across Amazon region (1999 2007) Acre Producers group formed Certification of CFEs in Acre and Amazonas Public Forest Law (2006) aims to increase forest area for community use Community Forest Management Working Group presents demands to Secretary of Environment (2007) Federal Community Forest Management Program initiated (2009)

PAGE 48

48 Table 22 Socio economic factors affecting CFE development in Southeastern Mexico and the Brazilian Amazon. Southeastern Mexico Brazilian Amazon Global context Markets Demand for chicle and other NTFPs Demand for high value species Demand for certified wood Policies Intl aid for conservation and local development Markets Demand for rubber and Brazil nuts Demand for certified wood Policies Intl aid for conservation and local development Legal reserve requirement National context Markets Demand for tropical hardwood Demand for conifers from plantations Policies Liberal trade policies Devolution of land to ejidos Devolution of rights to ejidos Federal funding programs Markets Demand for tropical hardwood Demand for certified wood Demand for community wood Policies Permanent forest reserve regulation (80%) Devolution of land to rural communities (creation of reserves and concessions) ProManejo Special regulations for CFEs State context Markets Competition from imports Demand for wood (especially for tourism) Policies Plan Piloto Forestal Plan Piloto Estatal (state program) Regulations for timber permits State funding programs for CFEs Access to credit Support organizations Civil societies focused on CFE Conflicts and alliances Conflicts with concession holders Grass roots moveme nt and protests Markets Competition from illegal wood Demand for wood Demand for certified wood (Acre) Policies State technical and financial assistance for CFEs Support organizations Social and environmental NGOs Cooperatives Producer organizations Conflicts and alliances Conflicts with ranchers and land speculators Rubber tapper movement

PAGE 49

49 Table 2 2. Continued. Local context Production strategy Volumes of mahogany and Spanish cedar Size of forest resources Secure tenure and access to forest resources Work groups Market strategy Certification Depend on local buyers Processing more sawnwood Migration (lack of workers) More alternative sources of income Infrastructure Good access to roads and ports Good ac cess to electricity and water Good access to telephone Production strategy Volumes of mahogany and Spanish cedar Size of forest resources Tenure and access to forest resources improving Market strategy Alliances with companies/cooperatives Certification Processing variety of products Infrastructure Poor access to roads and ports Poor road conditions Some have good access to riverways Poor access to electricity and water Poor access to telephone Poor access to formal education Table 2 3. Number of CFE s producing each primary and secondary timber product in Southeastern Mexico and the Brazilian Amazon. Main p roduct s Secondary products Stumpage Logs Mixed (Logs, Sawnwood) Sawn wood Only finished products Finished products Small poles Railroad ties Quintana Roo 3 4 14 3 0 4 19 4 Brazilian Amazon 3 4 6 1 2 0 0 0

PAGE 50

50 A. B. Figure 2 1. Division of responsibilities for f orest management activities among CFE actors. A) in Quintana Roo, Mexico. B) in the Brazilian Amazon. 0 5 10 15 20 25 30Pre harvest Harvest Skid & Load SaleNo. of CFEsActivity Company CFE Company Cooperative CFE Coop CFE 0 5 10 15 20Pre harvest Harvest Skid & Load SaleNo. of CFEsActivity Company CFE Company Cooperative CFE -Coop CFE

PAGE 51

51 A B Figure 2 2 Proportion of CFEs implementing each valueadded strategy. A) in Quintana Roo, Mexico. B) in the Brazilian Amazon. A B Figure 2 3. Proportion of CFEs selling each product. A) in Quintana Roo, Mexico. B) in the Brazilian Amazon. No valueadded Felled tree Rough Sawnwood Fine sawnwood or finished product No valueadded Felled tree Rough Sawnwood Fine sawnwood or finished product 12% 17% 58% 13% Stumpage Logs Mixed (Logs, Sawnwood) Sawn wood 19% 25% 37% 6% 13% Stumpage Logs Mixed (Logs, Sawnwood) Sawn wood Only finished products

PAGE 52

52 CHAPTER 3 HOW DO CERTIFICATION AND OTHER FACTORS AF FECT TIMBER PRICES AND MARKET ACCESS FOR CO MMUNITY BASED FOREST ENTERPR ISES? A QUANTITATIVE ANALYSI S IN SOUTHE ASTE RN MEXICO Introduction Since the early 1980s community based forest management has been a popul ar strategy to help local populations conserve forests and improve their livelihoods (Scherr et al., 2004; Amaral and Amaral Neto, 2005; Bray et al., 2005). In southeaste rn Mexico, community based forest enterprises (CFEs) have produc ed tropical timber fo r commercial sale since the mid 1980s and are considered references for community forestry in Latin America and around the globe (Galletti, 1998; Bray et al., 2006) O f the 80 CFEs with legal timber management plans in Quintana Roo, Mexico, eight had received FSC certification. These CFEs were encouraged to obtain certification by donors, technical assistance projects, government agencies, and buyers, to not only verify environmental and social management practices, but also to help them obtain b etter prices and access to highly competitive timber markets (Maynard and Robinson, 1998) Some have reportedly received these certification benefits, while others have not. However, the ability to obtain market and financial benefits of certification may be linked to more fundamental aspects of these CFEs (e.g., production volumes, distance to market ) or their products (quality, species) rath er than simply CFE negotiating capacity or buyer willingness to pay (Irvine, 1999; Bass et al., 2001; Carrera et al., 2006; Quevedo, 2006). Still lacking are quantitative comparisons among CFEs (and especially between those that are certified and not), th at link prices and market access to other potentially important independent variables. The high number of CFEs,

PAGE 53

53 especially certified CFEs, in Quintana Roo make it an appropriate location for this type of investigation. We used econometric methods to test : 1) the relationships between characteristics of products (e.g., species, quality grade), sales (e.g., year of sale, market in which product was sold), and producers (e.g., certification status, size of forest, level of vertical integration) with prices received by CFEs in Quintana Roo; and 2) the probability of the participation of these CFEs in local versus non local markets. The h edonic p rice m ethod guided the analysis of impacts of independent variables on price; a probit model was used to predict part icipation in markets. While our expectations regarding high impact of product characteristics on price were met, we also found important and somewhat surprising results regarding impacts on price and market access of factors such as : length of relationshi p with buyer ; CFE organization for sales ; and, indeed, certification. Community Based Forest Enterprises in Southern Mexico Extraction of timber and nontimber forest products has a long history in the tropical forests of Quintana Roo (QR) state in southeast Mexico. During the Caste wars (18471901), Mayans trying to maintain sovereignty over their lands in southern QR purchased weapons to fight the Mexican army with cash from timber sales to the English in Belize (Merino Perez, 2004). Settlers from other parts of southern Mexico migrated to the area in the early 1900s to harvest chicle, sap from the chico zapote tree used to manufacture chewing gum -a booming industry until the 1930s depression (Merino Perez, 2004) From the 1930s to the 1970s, l arge areas of forests were granted to communities in QR as part of land reform efforts (Bray et al., 2004; Merino Perez, 2004) These

PAGE 54

54 areas were designated ejidos ( the term is used to describe the areas and the communities themselves) and contained approximately 95% of the forests in QR (Forster et al., 2003) Ejidos were predominantly populated by indigenous Mayans in central part of the state, and by migrants from other parts of Mexico in the south. However, the ejidos land rights initially came with many restrictions a critical one being the federal government maintaine d the timber rights. From the 1950s to the mid 1980s, the private company MIQRO had a concession in many ejidal forests in QR, where it selectively harvested mahogany ( Swietenia macrophylla) and Spanish cedar ( Cedrela odorata) for plywood and veneer. In the early 198 0s, grass roots organizations mounted political pressure to give ejidos the right to harvest their timber resources (Bray et al., 2006) The concession system was discontinued in 1983, and the ambitious Plan Piloto Forestal was initiat ed the same year (Galletti, 1998) This joint initiative involving the federal and state governments and German assistance train ed the ejidos and helped set up not for profit civil societies h ereafter referred to as support organizations (SOs) (Ga lletti, 1998; Merino Perez, 2004) The SOs initially were used for negotiating sales, but soon their responsibilities shifted towards political representation of and technical assistance to ejidos (Galletti, 1998) In 1986, a new national forestry law transferred rights to ejidos to manage their forests for timber and other products with technical assistance from the societies (Galletti, 1998; Bray et al., 2006) Shortly thereafter, m any ejidos received management plan approval and began selling logs and, in some cases, boards with machinery and sawmills purchased through special governmental credit programs However, they soon found themselves

PAGE 55

55 in a changing market situation. One cause was that the CFEs, based on participatory forest inventories, reduced the amount of mahogany and cedar harvested, and n o other substitutes were found in the state. This prompted many buyers to begin looking elsewhere for these species. A second cause was that trade modernization and agreements (especially NAFTA) opened markets to a flood of cheaper domestic and internati onally produced tropical and temperate wood that posed fierce competition to the CFEs. The CFEs and their support organizations had to find new products and commercialization strategies. Several CFEs obtained certification, in part at the suggestion of foreign buyers who ended up making few purchases (Galletti, 1998; Maynard and Robinson, 1998) New markets were found for lesser known soft wood species with manufacturers of toothpicks, ice cream sticks, and pencils (Galletti, 1998) Also, in 1992 a flooring company began buying hardwoods from certified ejidos (Galletti, 1998) Ejidos also began to negotiate with buyers to take lesser known species as well as precious woods (pers. comm.). In addition, the growth of tourism i n the state (especially in Cancun) led to demand for small diameter poles (SEDARI, 2005) which is the only product for some CFEs (SEMARNAT unpublished data) Over time the number of CFEs and their support organizations increased to 80 and 10, respectively, and their respective roles and responsibilities changed. The SOs became important for not only technical services, but also accessing various government programs available to CFEs for activities (e.g., training, forest inventories, and plantation establishment) and investments (e.g., in equipment and machinery). The CFEs paid the SOs for their services based on timber volumes sold. Althought t he first

PAGE 56

56 few sales were negotiated jointly with the help of an SO (Galletti, 1998) the CFEs eventually chose to negotiate and sell individually. Nonetheless, the SOs did help set minimum prices among members. The mi nimum prices were based on species category1: precious wood ( madera preciosa ), hard wood ( madera dura), and soft wood ( madera blanda) 2; one SO set prices for individual species. Precious wood prices were set at approximately 1.3 to 1.8 times greater than hard and soft woods, which were usually priced similarly. Price was also broken down by management activity to account for average costs of each activity (per m3 of logs), so less vertically integrated CFEs (e.g., that sold standing trees) received less than more vertically integrated ones who incurred greater costs of production (e.g., transportation to the mill). While t he price list was intended to ensure a profit and to avoid competition among SO members, i t was potentially disadvantageous for small er CFEs (with higher per unit costs) and certified CFEs (the price list did not include certification as a cost incurring activit y or dictate a certification premium) The CFEs main market was local buyers, many of whom were intermediaries with sawmills wh o sent boards to regional and national markets (S EDARI, 2005; Wilshusen, 2006) Some certified ejidos infrequently exported timber products to the US or Europe (Maynard and Robinson, 1998) The main sources of competition were tropical wood 1 The same species cat egories were used by the ejidos and state agencies for planning and authorizing timber harvests. 2 The precious species, including mahogany and cedar, are typically medium density woods with specific gravities between 0.4 and 0.65, however some rare hard w oods are also sold in the precious wood price category, i.e., granadillo ( Dalbergia retusa) and ziricote ( Cordia dodecandra). Hard wood species are usually both darker in color and heavier, with a specific gravity above 0.65. Soft wood species are generally lighter in color and include light and medium weight species with a specific gravity below 0.65 (Forster et al., 2003) These terms should not be confused with nomenclature used in temperate zones.

PAGE 57

57 imports from the neighboring states of Campeche and Chiapas, as well as Chile, Brazil, and Guatemala, cheap plantationproduced pine from northern Mexico (Forster et al., 2003) Methods We applied the hedonic price m ethod (HPM) to test the impacts of independent variables on timber prices for CFEs in Quintana Roo (QR). HPM is based on the assumption that the many factors that influence the price of timber products make it a composite good, and the value at which it sells is its hedonic price (Garrod and Willis, 2000; King and Schreiner, 2004). Successful applications of HPM demonstrate how changes in each of a products attributes may affect an indi viduals utility for the product by decomposing the price of the product into the marginal value of each of its attributes (Garrod and Willis, 2000; King and Schreiner, 2004). This decomposition is achieved using regression analysis to model willingness t o pay for a good (using observed product prices) as a function of specific amounts of attributes or characteristics of the product (Rosen, 1974; Garrod and Willis, 2000). When a linear function is used, the coefficients of each attribute represent its cont ribution to the value of the composite good (King and Schreiner, 2004). For example, in a model to predict prices for certified wood, the coefficient for the variable certified would indicate the price premium associated with the certified variable (ass uming the certified variable is binary) The HPM will be used in this study to predict prices for CFE logs and boards based on several variables. HPM has been used frequently to model stumpage prices for temperate species (see Brannman et al., 1981; Holmes et al., 1990; Puttock et al., 1990) as well as for predicting price premiums for ecolabels (see Nimon and Beghin,

PAGE 58

58 1998; Galarraga and Markandya, 2004; Maguire et al., 2004) To our knowledge, thi s is the first time this method will be used in the context of tropical timber products eco labeled timber products, and community based forestry operations. Model s S pecification Based on a review of literature and previous research we dev el oped econometr ic regression models for price (PRICE) and a probit model for market access for logs and boards (sawnwood of different dimensions) as a function of three categories of independent variables: product, sale, and producer (Tables 3 1 and 32, Equations 3 1 3 2 and 3 3 ). The probit model was used to measure impact on the probability of a sale occurring in a local versus nonlocal market, since the goal of many CFE actors ( ejido members, support organizations, government agencies) was to reach nonlocal marke ts. These models were refined based on discussions with ejido leaders and SO representatives Although some of the sampled CFEs also sold other products, we chose logs and boards because these were the most widely comparable to CFEs in other parts of the world.3 Product variables, especially species (SG), were anticipated to have the strongest effect on price and market access. T he quality (QG), length (LENGTH) and thickness (THICK) of boards were also expected to positively affect price. It was also ex pected that boards would be more likely to be sold in nonlocal markets than logs due the formers relative ease of transport. 3 The two other most common products, furniture and small diameter poles, also presented challenges for modeling: for the former there were no detailed data (e.g., species or quality) on which to model prices or market access, and the latter were frequent ly sold in lump sales of different species and sizes.

PAGE 59

59 Sale variables were also expected to affect PRICE and market access Volume per sale (VOL) was expected to affect price positively as it can improve negotiating power. In the price model, MKT was defined based on where the buyer was located (local, regional, national, international), with nonlocal sales expected to result in higher prices than local market sales. Local market w as defined for each CFE as the area within which a buyer could expect to send a truck to retrieve the timber products and return within the same day. The regional market included buyers in bordering states that needed more than one day to transport timber products. The national market consisted of northern Mexico, as far as 2,700 kilometers away. While we expected to find that most sales were made in the local market as is common for CFEs (Amaral and Amaral Neto, 2005) we also expected to find sales in regional and national markets since QR is second largest tropical wood producer of three Mexican states (SEDARI, 2005) a s well as in international markets, at least for the certified CFEs Producer variables were characteristics of the CFEs (or ejidos ) e xpected to affect timber sales; in cases where these varied by year, they were measured for the year of the related sale. T hese variables were expected to provide insight on how differences in each CFEs natural resources, production organization, and sales experience affected prices and market access. Larger total timber volumes were expected to increase CFEs negotiating power and positively impact prices and access to nonlocal markets. Our proxies for this were total annual volume of logs and boar ds sold (VOLS), percentage of precious wood in total annual approved harvest volume (VOLAP), and total area of production forest (PERMF). The decision of ejido members to sell wood at the smaller group or individual level (COMFORM) (rather than all toget her) was

PAGE 60

60 expected to negatively affect prices and market access due to increased transaction costs for buyers and smaller volumes per group or individual (see Chapter 2 for a description of the use of work groups to produce and sell timber) The number o f families (NFAM) eligible to participate in forest management was used as a proxy for human resources. The level of vertical integration (VINTEG) varied greatly among CFEs and was expected to strongly influence prices for logs and boards. As previously mentioned, FSC forest certification (CERT) was obtained with the hope it would result in better market conditions for CFEs The fact the CFEs in QR were among the first communities certified in the world presented a unique opportunity to study the e ffect of time certified (CERTYRS) on prices. A recent compilation of CFE case studies found the number of years CFEs sold timber had a more important impact on the commercialization success than products or market access (Molnar et al., 2006); we measure this using the year the CFE management plan was approv ed (FMPYR). The total number of years the CFE had sold to the buyer as of the year of the sale (NSAL) was expected to positively influence prices as trust developed. Finally, b oth increased distance from m arkets (DIST) and roads (DISTR) were expected to have a negative relationship with PRICE and market access due to increased transportation costs (Bauch et al., 2007; Molnar et al., 2007; Ezzine de Blas et al., 2009) For some producer variables, there was not enough variability between the CFEs to detect an effect and they were excluded from the models, including: technical and financial assistance, which all ejidos had recently received; marketing and commercial i zation assistance, which none of the ejidos had received in the previous 5 years ; membership in secondary organization (all but five ejidos were members of

PAGE 61

61 SOs), the aggregation of volumes among ejidos for sales (none of the ejidos had done this in the previous five years ), and access to credit for operational cost s (only one ejido had received credit for forestry operations ; this variable does not refer to credit for machinery or related activities such as training ). Data for the variables market knowledg e and total number of known buyers were determined too subjective and unreliable After finding high correlations between many of the variables, we divided the variables into first tier and second tier groups, with first tier variables expected to have the clearest logical impact on price (Table 31) For tier 1 variables that were highly correlated with other tier1 variables we chose the variables that were least correlated with other variables. We included tier 2 variables that were not highly correlated with tier 1 variables. Linear models for prices were developed using multivariate regression analysis for logs ( Equation 31) and boards (Equati on 32). PRICE i 0 1 SG2i 2 SG3i 3 MKT3i 4 VOLi 5 CERTYRSi + 6 DISTRi + i (3 1) PRICE i 0 1 SG2i 2 SG3i 3 QG2i 4 QG3i 5 QG4i 6 QG6i 7 QG9i 8 LENGTH2i 9 THICK2i 1 0 THICK3i 1 1 VOLi 1 2 MKT2i 1 3 MKT3i 1 4 MKT4i 15 DISTRi 16 NSAL1i 17 NSAL2i 18 NSAL3i19 NSAL4 6i 2 0 CERTYRSi 2 1 VOLSi + i (3 2 ) i represents the ith observation, for 1 to n total observations i represents the stochastic error term A probability model for market access was developed (Equation 33). This model tested the probability of a CFE selling in the local market.

PAGE 62

62 LOCALMKT 0 1 DCERTi 2 PERMFi 3 DPRODi 4 DCOMFORMi + i (3 3) LOCALMKT = 0 non local market 1 local market only i represents the ith observation, for 1 to n total observations i represents the stochastic error term Heteroskedasticity was corrected for in the p rice models by using the ROBUST statement. Statist ical t ests of the price models included R squared, F test, and Ramsey Reset. To identify the effect of each variable on average price and market access while controlling for other variables, we ran a series of simulations for each model For each subset of dummy variables (e.g. SG1 SG2, SG3), we changed all observations for one of the variables to 1 (e.g., SG1) and 0 for the others (e.g., SG2, SG3) and calculated average price, while keeping the actual values of the other variables. We did this in turn for each of the dummy variables in the subset. For continuous variables, we simulated average price across the range of observed values. G raphs were generated to illustrate and compare each variables simulated impact on average pr ice. To further assist in comparing the relative impact of each variable, we standardized the coefficients for the variables in both the log and board price models (Equation 34). This allows direct comparison of model coefficients across units. = (3 4) is the sample standard deviation of the independent variable with coefficient is the standard deviation of the dependent variable Price is the standardized coefficient.

PAGE 63

63 Data C ollection A list of ejidos w ith approved timber management plans in QR in 20064 was obtained from the state regulatory agency, SEMARNAT. Of the 80 ejidos listed, two were removed for only selling small poles, which as previously mentioned were not used in our models. To facilitate a stratified sample, the remaining ejidos were divided into groups (Appendix 1) based on three factors that appeared to influence timber commercialization: ( 1) certification through the Forest Stewardship Council (FSC) system of the ejido s forest management, ( 2) close proximity to a mill, and ( 3) permission to harvest precious timber species (Table 33) Since so few ejidos were certified (8) and/or had mills (12), all of these were included in the sample (13 in total; 7 had both); another 18 were drawn from the remaining two groups (with and without precious wood) based on their proportion of the remaining 65 ejidos (Appendix B) Upon investigation of the 31 sampled ejidos six were removed due to either: insufficient information (two) or failure to complete a harvest season (four). The SEMARNAT Director estimated only approximately 56 of the 80 ejidos (or 70%) with permission actually harvested timber in 2006 (Lino Martinez, pers comm. ). Based on this information and time constraints it was decided a sample of 25 ejidos (44.6% of those that harvested timber in 2006) would be sufficient (Appendix B) Initially, due to the long history of community forestry in the region, we expected to find up to 20 years of timber sale data in the form of contracts and/or transportation documents. In contrast, we found few CFEs used written sales contracts or receipts, 4 The year 2006 was chosen to exclude ejido production in 2007. Many ejidos that do not typically engage in timber production received temporary permission to extract and transport logs after hurricane Dean struck in August 2007.

PAGE 64

64 and these documents were frequently disposed of with each change in ejido presidency (typically every one to three years). We t hus focuse d on recent sales (2002 2008) Price data were available for 21 of the 25 CFEs; data for 4 CFEs were not available.5 For boards, 94% of the price data came from receipts (as most of the data came from the few CFEs using receipts). In contrast, for logs, only 38% of the price data came from receipts, and 61% came from ejido president or treasurer recall, however we limited data collected this way to recent sales from 20052008 to avoid problems with recall bias .6 The price lists were consulted t o verify price ranges, but were not used as a price data source T he primary market data sourc e was transportation documents which noted product, species, volume, and buyer (but no prices) ; market data was available for all 25 CFEs Price d ata were recor ded for a total of 3,658 timber sale transactions7, comprising 3, 314 sales of boards (13 CFE s) and 34 4 sales of logs (15 CFEs ); seven CFEs had price data for sales of both products. Prices were recorded in Mexican Pesos (P$) and adjusted to 2008 to account for inflation. An additional 1,051 sales were recorded for which market data were available, but not price, for a total of 4, 709 observations for market access for 25 CFEs for both logs and boards We also interviewed ejido pr esidents and/or other officers, representatives of the SOs, and timber buyers on their perceptions of the i nternal and external challenges CFEs face d in commercializing wood, and opportunities for improving market access, prices, and total 5 Some ejido presidents did not remember specific prices for specific buyers or were not available to provide data. 6 As log prices did not vary much by year for the major species groups, and were not influenced by other factors as board prices were (e.g., quality and length), we are confident in the accuracy of the log prices obtained via ejido leadership recall.. 7 If more than one product was listed in a receipt or transportation document, each product was recorded in a unique transaction to capture its unique data (e.g., volume, price, quality).

PAGE 65

65 timber income. Their input was used to refine the models but was not measured quantitatively or qualitatively. Results A verage characteristics of sampled CFEs and their timber sales (Tables 3 4 and 3 6 ) we re influenced by the fact that two CFEs accounted for large proportions of the observations .8 The difference in number of observations among CFEs is due to differences in diversity of products, numbers of buyers, and number of categories related to product quality (e.g., some CFEs sorted boards by length and thickness while others did not). CFEs timber product s sales were as follows: t hree CFEs sold standing trees, four sold only logs, three sold only boards, 1 4 sold both logs and boards, and one sold only small poles9 (see Chapter 2) As secondary products, most CFEs also sold small diameter poles and some sold furniture and railroad ties ( although these secondary products were not included in our study ) The strategies for adding value also varied: 10 used chainsaws, and 7 used permanent sawmills Almost two thirds of the ejidos in the study internally divided themselves into work groups to better control timber production and sales (Wilshusen, 2006) (see Chapter 2) The CFEs with mills usually processed their precious species (SG1) or rented their mill to the buyer and charged a milling fee per board; the y typically sold their hard wood (SG2) and soft wood (SG3) as logs, citing insufficient mill capacity ,10 low prices for 8 Noh Bec accounted for 2,898 of 3,297 observations for boards, and Felipe Carrillo Puerto accounted for 143 of 354 observations for logs. 9 The CFE that sold only small poles was approved to sell logs, but we found they had opted not to. 10 Many of the mills had not been maintained or were simply outdated, and their production quality was low.

PAGE 66

66 harder species and/ or plywood factory demand for soft wood logs. Three CFEs that lacked permission to harvest precious species use d cha insaws to cut their hard wood logs into roughly finished boards. In contrast, some CFEs with sawmills chose to sell logs (even of precious species) rather than assume the processing related risks, including finding a buyer before the transportation documents expired, or buyer dissatisfaction with product quality and/ or noncompliance with payment agreements. Regarding sorting, while all CFEs sorted their products by species sorting by length, quality, and thickness were less common. The majority of logs were sold as noncertified (CERT0), while 92% of boards were sold as certified (CERT1). The ejidos selling boards were more often certified, certified for longer, and located closer to the road and the nearest market, but did not nece ssarily begin selling wood earlier (FMPYR). For both logs and boards, the number of years of previous sales to each buyer ( NSAL ) data showed that repeat sales to buyers were frequent; there were no first time buyers of logs reported for our study period. Boards were sold in a greater diversity of markets than logs, but the local market accounted for 9495% of sales for both products. Independent V ariable C orrelations Logs Several of the independent variables identified for the log price model were highl y correlated with each other. The variables certified (CERT) and number of years certified (CERTYRS) were highly correlated with variables for total forest area designated for timber (PERMF ; 0.75 ) and distance to nearest market (DIST ; 0.82 ), and moderatel y correlated with number of famil i es participating in timber production in each ejido (NFAM; 0.5 4 0.56), indicating the certified CFEs tended to have larger production

PAGE 67

67 forest areas, were located farther from roads than noncertified operations, and had gr eater numbers of families in the ejido Indeed, most of the certified CFEs were among the first CFEs in the state and generally had larger forests than newer CFEs Variable DIST was highly correlated with the form of organization for commercialization (C OMFORM ; 0.74 ), indicating the farther from the nearest market the ejido was the more likely it was to choose to sell wood in smaller groups or individually than together as one group. The level of vertical integration (VINTEG) was highly correlated with the proportion of precious species in the total approved harvest volume (VOLAP ; 0.72 ) and PERMF (0.85) and moderately correlated with FMPYR ( 0.65) indicating that the more vertically integrated CFEs were more likely to have higher proportion of precious wood to total approved harvest volume, larger forest areas, and more experience with forest management. Boards As with logs, m any of the independent variables identified for the board price model were highly correlated. There were moderately high correlations (0.500.68) between certification (CERT, positive correlations) and years certified (CERTYRS, negative correlations) with board LENGTH, board THICKness, organizational form of commercialization (COMFORM), and distance to nearest market (DIST); CERT and CERTYRS also both had high positive correlations with level of vertical integration (VINTEG; 0.67). These correlations indicate the certified operations tended to be vertically integrated and located farther from markets. VINTEG wa s also highly corre lated with VOLAP (0.85) indicating a positive relationship between vertical integration and percentage of total approved harvest volume comprised of precious wood species. Year forest management initiated (FMPYR) had a high negative

PAGE 68

68 correlation with VINT EG ( 0.8 3 ), CERT ( 0.66), area of permanent forest (PERMF, 0.63) and a moderately positive correlation with distance (0.57), indicating the longer a CFE had been harvesting timber, the more likel y it was to be vertically integrated and certified, and the greater its forest area and distance to market. Market access We also tested correlations for variables identified for the market access probit model which combined data for logs and boards. In this case, the high correlations remained between FMPYR with PERMF, VOLAP, and CERT, and between VOLAP and CERT, reaffirming the relationships previously described. Price M odel R esults Logs Our model explained 66% of the variability in prices for logs (Table 3 5 ). The average price was P$ 1,502 /m3,11 and the base price in the model was P$ 4,064/m3 for precious wood (SG1=1 ), local market (MKT 1 = 1 )12, zero volume (VOL), zero distance to the road (DISTR), and certified for zero years (CERTYRS). As many variables were hig hly correlated with each other, we gave model inclusion preference to those with the strongest logic for influencing price excluding other variables to avoid covariance problems. Product variables. Species grade had the largest impact on price Hard wo od (SG2) and soft wood species (SG3) were predicted to receive prices 70 to 71 % lower than precious species, controlling for other sources of variation (Figure 3 1.A). 11 Exchange rate: P$ 1.00 = 0.09 USD 12 The only two markets in which logs were sold were the local (MKT1) and national markets (MKT3).

PAGE 69

69 Sale variables. Selling in the national market increased predicted price by P$ 196/ m3, or approximately 5% (Table 35) Seven of the 13 CFEs sold both hard and soft wood species in the national market; local buyers purchased all of the precious wood logs. Volume of sale (VOL) was not significant in the regression model. Producer variables. The number of years certified (CERTYRS) had a large impact on price (Table 3 5 )13 and was significant based on a onesided t test (i.e., test if price was greater than the average) While the increase in price (or price premium) of P$ 20.50 Pes os /m3 per certified year is small (0.5%) the accumulated benefit for operations certified for 13 years is a 6.5% price premium (the majority of certified operations selling logs in the study had been certified for 12 or 13 years). Distance to road (DISTR ) had a negative and significant effect on price, as expected, of P$ 12.86/km, which would result in a price decrease of P$ 219 (or 5.3%) for an ejido located at the average distance to road (17.09 km). T he price simulation (Figure 3 1 D) clearly illustr ates the impact of DISTR on price especially near the upper limit of the observed distances. Boards Our model explained 6 3 % of the variability in price for boards (which were measured in board feet bf) (Table 3 5 ). The average price was P$ 13. 48/bf The 13 The way to interpret the standardized coefficient results for continuous variables is that a sample standard deviation change in the independent variable produces a change in average price equal to the product of the standardized coefficient and the standard deviation of average sample price. For example, a standard deviation change in distance to road (2.86 km), holding other variables constant, leads to a decrease in price of 252 Pesos ( .187 *1 ,352). The way to interpret the standardized coefficients for the dummy v ariables is that a one unit change in the dummy variable (0 to 1) produces a change in price equal to the product of the standardized coefficient and the standard deviation of price. For example, if the species grade is soft wood, holding all other variabl es constant, the average pri ce decreases by 2,914 Pesos ( 2.104* 1352), which is the same value as the coefficient.

PAGE 70

70 base price for the model was P$ 18.96/bf for boards of precious wood (SG1=1 ), first quality (QG1 =1 ), full length (LENGTH1 =1 ), less than 1.5 in thickness (THICK1=1 ), sold in the local market (MKT 1= 1) to a new buyer (NSAL0=1) by a noncertified CFE (CERTYR=0). Similar to the si tuation for the log price model several variables had to be excluded from the board price model due to covariance problems, including CERT1. Product variables. As in the price model for logs, species grade (SG) had the lar gest impact on price, with hard wood and soft wood species, respectively, obtaining prices 40 to 60% lower than precious species (Table 35) Board LENGTH and quality (QG) also significantly impacted price There was much less variability for board thick ness (THICK) ; t he negative sign for THICK 2 was unexpected, and is probably due to the need to re plane th e s e products to a more common one inch thickness. Sale variables. While t he coefficient for reg ional market was insignificant, t he average premiums for selling in the national and international markets were, respectively, P$ 0.91/ bf ( 4 8 %; similar to logs) and P$ 11.10/ bf (58.5%) (Table 35) This surprisingly large premium, however, was not enjoyed by many producers since the proportions of sales in t he national and international markets were very low: approximately 3% and 1%, respectively. Volume of sale (VOL) was significant; although it had a coefficient of 0.00, the simulated average price increased by 40% at the far end of the observed VOL values (Figure 3 2.E). Producer variables. The average price premium per c ertified year (CERTYRS) was P$ 0.25/bf (1.3% ) which is double that found for logs (Table 35) Thus, the accumulated price premium for CFEs certified 912 years (98% of the certified operations selling boards fell in this range) would be a 11.7 to 16.9 % premium For the

PAGE 71

71 number of previous years in which the CFE previously sold boards to the buyer (NSAL) only the results for two, three, and four years of previous sales were important ; the coefficient for one year was insignificant. Finally, distance to road (DISTR) was not significant for boards. Mark et A ccess M odel R esults This model made market access a function of the product with both logs and boards included in the same model. Our model had 95.9% correct predictions of a sale occurring in the local versus non local market (i.e., regional, national, or international) (Table 36 ) The area of permanent forest (PERMF), product and form of commercialization were significant and ha d negative coefficients, indicating that sales in nonlocal markets were more likely to occur for boards than logs, and by CFEs with large forest areas and that sell at the group or individual level The impact of certification was inconclusive. The simu lations revealed a minimal 1.5% difference in market access based on product, and larger differences for form of commercialization (9%) and range in permanent forest area size (Figure 33). Discussion For all CFEs adding value (through felling or processing), getting certified, selling in other markets, and maintaining repeat clients were the best ways to maximize the value of their timber products. Some of the factors we found to be significant are under the control of CFEs, such as processing quality and certification, and others are less so, such as distance to road and permanent forest area. We further explore here what our findings indicate regarding factors that could help CFEs maximize the value of their timber products, and make comparisons to other studies.

PAGE 72

72 Factors Which CFEs C ontrol Fortunately, CFEs have control over many more variables tested in our models than not. The most important factors here are processing, accessing other markets, buyer relations, and certification. Processing Proces sing added significant value to the wood, and t he impact on price of board qual ity and dimensions, especially length, underscored the importance of milling and sorting capacity to capture the products full value. The quality of milling definitely varied, and in some cases chainsaws produced better quality boards than mills; our model showed chainsawn hard woods (the only species group cut with a chainsaw) received prices similar to second grade boards cut with a sawmill. Some CFEs preferred to sell logs than assume the risks involved in processing (e.g., poor quality boards, poorly dried wood) (see Chapter 2). Indeed, several studies of CFEs have shown that primary processing is not always more profitable than selling logs due to high costs and low effic iencies (see Chapter 4) (Molnar e t al., 2007; Medina and Pokorny, 2008) Nonetheless, the potential for valueadded through processing warrants CFE consideration of how to access good quality sawmills, such as through improving their own mills, accessing neighboring CFEs mills, renting the service of a commercial mill, sharing a portable sawmill with other CFEs (see Chapter 4 for an example), or helping secure a central sawmill for a producers group or cooperative; the increase in value might offset transportation and other costs CFEs should also place importance on sorting skills and performance. While species type was always used and length was

PAGE 73

73 usually used for sorting, quality and thickness were used less -this indicates lost income considering the high impact of especiall y quality on board price. Selling in nonlocal markets While t he price premiums found in this study associated with market access clearly support the potential benefit for CFEs to sell in nonlocal markets especially for sawnwood, the percentage of sales in these markets was only 4 to 5%. Why? Unfortunately, our market access model results were not very informative. We found upon further examination of our data, though, that most sales in non local markets were for high value species and certified wood. However, the reasons behind CFE reliance on local markets may be more complex. Qualitative data from interviews suggest CFEs actually preferred to sell to local buyers for several reasons. For the vertically integrated CFEs with large volumes of precious woods, the reasons included: local buyers fully prefinanced harvesting expenses (sometimes on a weekly or biweekly basis); the buyers were willing to buy all of the species, not just the high value ones; and the CFE leaders knew and trusted the local b uyers more than buyers from other parts of the country or world. The smaller CFEs, who usually had no precious species, perceived their volumes were too small and/or they were too far from main markets to attract even local buyers, and they simply had no contact with buyers from other markets. What can be done to address this situation? It is known that very little of the wood is processed locally, except for soft woods that go to local plywood factories the other species go to northern Mexico or to int ernational markets. If CFEs want to capture the highest value for their wood, they need to find innovative ways to interact with buyers in

PAGE 74

74 a manner that satisfies the CFEs and the buyers needs. A key way to do this may be through collaboration with other CFEs and/or companies. Collaboration CFEs could collaborate with other CFEs, through producers groups or cooperatives, or companies to reach other markets and increase prices. Producer groups have been successful in Guatemala (Carrera et al., 2006) and Brazil in organizing CFEs for sales, and a cooperative has been making sales for CFEs in Acre, Brazil since 2006 (see chapter 2). The aggregation of different species could facilitate targeted marketing and sal es to end users of different species: e.g., precious woods to furniture and veneer companies and hard woods to flooring and furniture companies. Certification could continue to make the wood from QR attractive to these markets, especially with the growin g demand for wood that meets legality requirements (legality is a basic tenant of FSC certification). The challenges would be CFE leaders willingness to give up control of sales, start up funding for a producers group or cooperative, and controlling for corruption. At the time of this study several SOs were discussing initiating cooperatives to improve market access and prices for member CFEs, and one, whose members managed relatively smaller volumes of timber, had installed a mill and was working out the logistics of log transportation for members Another option being explored in a few of the QR ejidos was c ompany community partnerships {Merry, 2006 #767}. There was discussion of a possibility for one or two CFEs to collaborate with a company to instal l a carpentry factory that would purchase CFE wood and employ ejido residents. This was occurring in Brazil, where t hree recently formed CFEs in the Brazilian Amazon were working with companies in different

PAGE 75

75 capacities (see Chapter 2). CFEs stand to benef it from not only help with investment, but also capacity building and joint ventures in new products. Certification Many QR CFEs pursued certification with the expectation of price premium s (pers. comm.) and improved market access While we were not able to directly test the effect of certification in our models, we indeed found certification had important impacts on price and market access. We found number of years certified (CERTYRS) significantly impacted log and board prices Cumulative annual pri ce premium s were 4.5 to 11.7% for logs and boards, respectively, for CFEs certified at least 9 years (98% of ejidos with certification had been certified 913 years) The indication that it took a while for the CFEs to benefit from certification (i.e., the cumulative impact) makes sense as they were among the certified operations in the world, and it took a while for processing facilities to get certified (for chain o f custody) and for the market for certified products to develop. A price premium is generally very attractive to CFEs because their unfavorable economies of scale (i.e., limited resource base, limited access to capital, and lower processing capacity) ofte n result in higher cost s and lower quality products that must compete with both low cost illegal operations and more efficient industrial operations (Cashore et al., 2006; Molnar et al., 2006). In addition, inspection of our data revealed all international market board sales, 77% of national market board sales, and 40% of regional market log sales were made by certified CFEs. Th e se results contrast greatly with reports from ejido leaders and SO representatives who perceived that certification had few bene fits for CFEs in QR. Other case studies of certified CFEs have presented mixed results regarding market benefits. Many report no benefits (Madrid and Chapela, 2003; Bray et al., 2005;

PAGE 76

76 Fonseca, 2006) while Humphries and Kainer (2006) reported, in contrast, certified communities in Acre, Brazil, found buyers in the national market who paid a 400% price premium. Others have cast doubt on the basis of premiums, speculating they were based on species more than certification (Molnar, 2003) or that they would be temporary (Carrera et al., 2006) Several ejido leaders also attributed what they described as the rare price premium for certified wood to the fact buyers of certified wood typically demanded only high value species and the highest quality boards, and we re willing to pay more to get access to these products (pers. comm.). However in this study we isolated the effect of certification on price (specifically years certified) and found it was significant Still, buyers of certified wood we re mainly interested in mahogany, and, because CFEs frequently required buyers to also purchase other species (pers. comm.), the buyers may have paid higher prices for all species to access the mahogany. Nonetheless, higher prices for other species would still constitute an indirect benefit of certification. These findings corroborate results of other studies investigating price premiums. Studies compiled by Galarraga and Markandya (2004) estimated potenti al price premiums for certified wood ranging from 5 to 15%. Kollert and Pagan (Kollert and Pagan) used ANOVA to compare average sales prices over three years for certified and noncertified logs and found price premiums of 27 to 56% for high quality logs, and 2 to 30% for lower quality. The authors conceded, however, that sales procedures (auction for certified logs versus tender/quotation system for noncertified) may have been more highly correlated to the premiums than cer tification.

PAGE 77

77 Notably, in late 2007/2008 all but one of the certified CFEs in Quintana Roo lost their certificates due to noncompliance with certification standards. Ejido leaders reported they were more preoccupied with dealing with the devastating impact s of Hurricane Dean, which leveled large forest areas in the region in 2007, and made extractin g downed trees ( a serious fire hazard) more of a priority than proving compliance with certification standards (pers. comm.). The various financial and technical challenges f o r communities to obtain and maintain certification are well documented (Molnar, 2003; Cashore et al., 2006) Nonetheless, the perceived lack of price benefits of certification (contrary to evidence here) and the fact that most certified demand is from national or foreign buyers (whom ejido leaders assert are problem atic to deal wit h) may also have influenced the decision to let certificates lapse Relationships with b uyers The number of years that CFEs had previously sold to the same buyer significantly impacted prices. This underscores the importance of CFEs comply ing with sales agreements and maintaining good relationships with buyers. Forster et al (2003) cited QR CFEs relatively short relationships with buyers (typically less than ten years) as an impediment to market success. CFEs interactions with buyers could also be strengthened through written formal contracts which were rarely used. Forster et al. (2003) recommended CFEs seek legal advice to better understand their rights and obligations in contractual sales and suggested t he use of written contracts m ight entice new buyers from other marke ts who value formal agreements.

PAGE 78

78 Factors Over Which CFEs Have Less Control Species was the most important variable that explained log and board prices, with precious species long the focus of the timber industry in t he region obtaining prices 30 to 7 0% higher than hard and softwoods. Nonetheless, a small proportion of the CFEs in the state had permission to sell precious species (Appendix B ). Harvest permission was based on total commercial volume per species, and many ejidos did not have large enough forest areas or high enough densities of high value species to support harvests of precious woods. In some cases previous logging had deplet ed mahogany and cedar stocks (Galletti, 1998) and many CFEs were establishing plantations of precious species. Distance to road is another variable over which CFEs have little control, but which had significant impacts on prices. Distance t o road is frequently cited as negatively impacting CFE timber sales in general (Molnar et al., 2007) and our results highlight that it impacts CFEs differently dependi ng on the product sold. It was found to negatively impact log price, but not board price, which is logical given the relative ease to load and transport boards versus large, heavy logs. CFEs, SOs and policy makers might consider encouraging processing when CFEs are located farther from roads, but the decision should be informed by a financial viability analysis, given that processing is not always profitable. Access to loans to finance operating costs could help CFEs break their dependence on local buyers and obtain better prices from other buyers, but bank lending practices require ejidos to form separate companies to obtain a loan. This is a commonly cited problem for CFEs and other small producers (Molnar, 2003; Amaral and Amaral Neto, 2005; Macqueen et al., 2006) An innovative microcredit system was

PAGE 79

79 established by one of the SOs in QR (Forster et al., 2003) but it failed due to repayment and organizational problems. Conclusions This study clearly shows that certification, at least over time, has contributed to higher prices for CFEs in QR, and that the pricing system does differentiate according to species, quality, market, and other attributes of sales and producers. While the support CFEs need to maintain certification are not always present, the statistical evidence points to the long run gains associated with certification. Each community then has to balance those gains against certifications costs. In addition to certification, there are other ways that CFEs can attempt to maximize the value of their forest r esources including processing, sorting by species, quality, length, and thickness, accessing new markets, and maintaining good relationships with buyers. For CFEs with larger forest areas, transforming logs to boards and selling at the smaller group or i ndividual level may help obtain better access to nonlocal markets. For CFEs with smaller forest areas, t he best way for CFEs to do this may be through collaborative efforts with other CFEs through producer groups and/or cooperatives, and/or with companies. However, for this to work, the bonds that tie CFEs to local buyers (prefinancing, acceptance of verbal agreements based on trust, and abundance of lesser known species with low demand) will need to be broken. Alternative arrangements that address CFE s needs while also appealing to buyers in other markets will need to be negotiated. Another policy implication is that our models showed that sales by CFEs with larger forest areas dedic a ted to timber management we re less likely to occur in local markets, and that sales in nonlocal markets have significantly higher sales. Thus,

PAGE 80

80 when possible, CFEs and support organizations should lobby for larger forest areas for community forestry, and policies should favor maximizing forest areas for CFEs in order to m aximize the value CFEs can obtain from timber sales. These results should be used to inform production and marketing strategies for CFEs, as well as policies that support community based forest management and enterprises in QR and other tropical areas T raining in processing, marketing, and customer relations could help CFEs maximize values. In addition, technical and financial support to help the CFEs regain certification could prove immensely valuable in maintaining price premiums and access to national and international markets In addition, access to credit may be crucial in helping CFEs achieve financial independence from buyers and gain entry in new markets.

PAGE 81

81 A 0 1000 2000 3000 4000 Local (.94) National (.06) Mexican Pesos ($) Market (MKT) B 0 1000 2000 3000 4000 50% (.57) 100% (.02) 200% (.17) 300% (.24) Mexican Pesos ($) Percentage of mean distance to road (DISTR, avg=17.09km) C 0 1000 2000 3000 4000 0 (.69) 1 2 3 (.01) 4 5 6 7 8 (.02) 9 (.02) 10 (.01) 11 (.03) 12 (.12) 13 (.09) Mexican Pesos ($) Number of years certified (CERTYRS) D Figure 3 1. Simulations of average price across independent variables compared to average price across all variables (solid line). A) Species grade. B) Market. C) Distance to nearest road. D) Years certified. Note: the numbers in parentheses are the proportion of total observations that were 1 for dummy variables, and the proportion of total observations in the range between values for continuous variables. 0 1000 2000 3000 4000 Precious wood (.17) Hard wood (.73) *** Soft wood (.10) ***Mexican Pesos ($)Species Grade (SG)

PAGE 82

82 A 0 5 10 15 20 25 30 Mexican Pesos ($) Quality Grade (QG) B 0 5 10 15 20 25 30 Long (.62) Short (.38) Mexican Pesos ($) Length (LENGTH) C 0 5 10 15 20 25 30 < 1.49" (.46) 1.5 1.99" (.22) >2.0" (.32) Mexican Pesos ($) Thickness (THICK) D 0 5 10 15 20 25 30 25 (.72) 50 (.10) 100 (.07) 1000 (.07) 4000 (.04) Mexican Pesos ($) Percentage of mean volume sold (VOL, avg=757 bf) E 0 5 10 15 20 25 30 Local (.96) Regional (<.01) National (.03) Int'l (<.01) Mexican Pesos ($) Market (MKT) F Figure 32. Simulations of average board prices across independent variables compared to overall average price (solid line). A) Species grade. B) Quality grade. C) Length. D) Thickness. E) Volume sold. F) Market. G) Distance to nearest road. H) Number of previous year s selling to buyer. I) Years certified. Note: the numbers in parentheses are the proportion of total observations that were 1 for dummy variables, and the proportion of total observations in the range between values for continuous variables 0 5 10 15 20 25 30 Precious wood (.59)* Hard wood (.32) Soft wood (.09)Mexican Pesos ($)Species Grade (SG)

PAGE 83

83 0 5 10 15 20 25 30 25% (.03) 50% (<.01) 100% (.88) 500% (.06) 1000% (.03) Mexican Pesos ($) Percentage of mean distance to road (DISTR, avg=9.8 km) G 0 5 10 15 20 25 30 Mexican Pesos ($) Number of previous years selling to buyer (NSAL) H 0 5 10 15 20 0 (.08) 1 2 3 4 5 6 (.01) 7 8 9 (.21) 10 (.15) 11 (.36) 12 (.18) 13 (.01) Mexican Pesos ($) Number of years certified (CERTYRS) I Figure 32. Continued.

PAGE 84

84 A B C D Figure 33 Simulations of the probability of a CFE sale being in the local market across independent variables compared to average probability of a sale being in the local market (solid line). A) Certification. B) Area of permanent forest. C) Product. D) Organization for commercialization. Note: the numbers in parentheses are the proportion of total observations that were 1 for dummy variables, and the proportion of total observations in the range between values for continuous variables. 0.5 0.6 0.7 0.8 0.9 1 Not certified (.72) Certified (.28)Probability of selling in local market 0.5 0.6 0.7 0.8 0.9 1 25% 50% 75% 100% 125% 150% 175% 200%Probability of selling in local marketPercent of average permanent forest area (avg=17,308 ha) 0.5 0.6 0.7 0.8 0.9 1 Logs (.24) Boards (.76)Probability of selling in local marketProduct 0.5 0.6 0.7 0.8 0.9 1 Ejidal level (.83) Group or individual level (.17)Probability of selling in local market Organization for commericialization (COMFORM)

PAGE 85

85 Table 31. Variables and expected signs for the price models for CFEs in Qui ntana Roo, Mexico Dependent variables Description PRICE Price per unit sold in Mexican Pesos Independent Variables Description Tier for Logs Tier for Boards Expected Sign Product variables SG Species grade dummy variables SG1 Precious wood 1 1 0 SG2 Harder wood 1 1 SG3 Lighter wood 1 1 QG* Quality grade of boards dummy variables QG1* High quality 1 0 QG2* Second grade 1 QG3* Third grade 1 QG4* Mixed grades 1 QG6* Chainsawn 1 QG7 Not available 1 QG8* Heartwood 1 QG9* Sapwood 1 LENGTH* Length of boards in inches LENGTH1* Long 1 0 LENGTH2* Short 1 LENGTH7* Not available 1 THICK* Thickness of boards in inches THICK1* less than 1.49 1 0 THICK2* 1.5 1.99 1 + THICK3* 2 and greater 1 + THICK7 Not available Sale variables YEAR Year of sale, ranged from 2002 to 2008 + VOL Volume of product sold in corresponding unit (m3 for logs, bf for boards) 1 1 + MKT Market where buyer is located dummy variables MKT1 Local 1 1 0 MKT2 Regional 1 1 + MKT3 National 1 1 + MKT4 International 1 1 + MKT7 Not available Note s: Only applie d to boards ** Only applied to logs

PAGE 86

86 Table 31. Continued. Independent Variables Description Tier for Logs Tier for Boards Expected Sign Producer variables for year of sale CERT Certification dummy variables CERT0 Not certified 1 1 0 CERT1 Certified 1 1 + CERTYRS Number of years certified 1 1 + VINTEG Level of vertical integration of the CFE dummy variables VINTEG1** Stumpage (standing tree), log in the forest, or log in the patio (skidded) 1 0 VINTEG2** Log loaded onto buyers truck (loader capacity) 1 + VINTEG3* Board in the community/patio (cut with chainsaw) 1 0 VINTEG4* Board in the community/patio (cut with portable or permanent sawmill) 1 + DIST Distance from ejido population center by road (km) to nearest market 1 1 DISTR Distance from ejido population center by road (km) to the highway 1 1 FMPYR Year first management plan was approved 2 2 + VOLS Total volume sold in year of sale in cubic meters 1 1 + VOLAP Proportion of the total approved harvest volume comprised of precious wood species (%) 2 2 + PERMF Total permanent forest area designated for timber management 2 2 + NFAM Number of families participating in each CFE. 2 2 + NSAL Number of cumulative years CFE has sold to this buyer prior to the year of sale + NSAL0 Zero sales 2 2 0 NSAL1 One sale 2 2 + NSAL2 Two sales 2 2 + NSAL3 Three sales 2 2 + NSAL4 Four sales 2 2 + COMFORM Form of organization for commercialization COMFORM1 All ejido members together 2 2 0 COMFORM2 Groups of ejido members 2 2 COMFORM3 Individually 2 2

PAGE 87

87 Table 32. Variables and expected signs for the market access model for CFEs in Quintana Roo, Mexico Dependent Variable Description Expected Sign Localmkt Market dummy variable (1=local market) Independent Variables Description Expected Sign DCERT Certification dummy variable (1=certified) PERMF Total permanent forest area designated for timber management DPROD Product dummy variable (1=board) DCOMFORM Form of organization for commercialization dummy variable (1=at group or individual level) + Table 3 3 Variables used to stratify the population of CFEs with permission to harvest timber in Quintana Roo, Mexico. I ndependent v ariable Effect on d ependent v ariable Market a ccess Price Cert Certification can attract buyers from national and international markets. Certification can make a difference in the price buyers pay for sawnwood, especially first grade quality. Mill It is cheaper and more efficient to transport boards than logs. Thus, buyers, especially from national or international markets, may prefer CFE s in close proximity to a mill. Mills typically produce higher quality boards than chainsaws. Thus, CFE s sellin g millsawn boards may receive higher prices per board foot than other CFE s. Precious wood species CFE s with permission to harvest precious wood species may enjoy greater access to national and international markets than those without. CFE s with permission to harvest precious wood species may be able to negotiate higher prices for lesser known species with buyers who value access to the precious wood species.

PAGE 88

88 Table 34 Summary statistics for s ales of logs and boards with price data ( n=3,668 sales ) in Quintana Roo, Mexico Logs (n=344, for 13 ejidos ) Variable Description Mean Std Dev Min Max Mean Std Dev Min Max PRICE Price per unit sold 1,502.40 1,352.54 131.97 9,977.19 13.47 7.13 2.08 60.01 SG1 Precious wood (0,1) 0.17 0.37 0.59 0.49 SG2 Hardwood (0,1) 0.73 0.45 0.32 0.47 SG3 Softwood (0,1) 0.11 0.31 0.09 0.29 SG7 Not available (0,1) 0.00 0.02 QG1 High quality (0,1) 0.29 0.46 QG2 Second grade (0,1) 0.02 0.15 QG3 Third grade (0,1) 0.35 0.48 QG4 Mixed grades (0,1) 0.01 0.11 QG6 Chainsawn (0,1) 0.04 0.21 QG7 Not available (0,1) 0.04 0.19 QG8 Heartwood (0,1) 0.13 0.33 QG9 Sapwood (0,1) 0.11 0.32 LENGTH1 Long (0,1) 0.58 0.49 LENGTH2 Short (0,1) 0.36 0.48 LENGTH7 Not available (0,1) 0.06 0.23 THICK1 < 1.49 (0,1) 0.38 0.49 THICK2 1.5 1.99 (0,1) 0.18 0.39 THICK3 >= 2 (0,1) 0.27 0.44 THICK7 Not available (0,1) 0.16 0.37 VOL Volume of product sold in corresponding unit 93.73 647.27 0.98 11,576.00 756.88 2,499.87 0.75 47,673.00 MKT1 Local market (0,1) 0.95 0.23 0.95 0.21 MKT2 Regional market (0,1) 0.00 0.05 MKT3 National market (0,1) 0.06 0.23 0.03 0.17 MKT4 International market (0,1) 0.01 0.09 MKT7 Not available (0,1) 0.01 0.08 CERT0 Not certified (0,1) 0.69 0.46 0.08 0.26 CERT1 Certified (0,1) 0.31 0.46 0.92 0.26 CERTYRS No. of years certified 3.42 5.32 0.00 13.00 9.75 3.00 0 13 DIST Distance to nearest market (km) 57.76 52.88 7.00 150.00 14.11 44.64 0.00 260.00 DISTR Distance to the highway (km) 16.83 19.62 0.00 53.00 9.83 11.72 0.00 91.00 Boards (n=3,314, for 15 ejidos )

PAGE 89

89 Table 3 4. Continued. Logs (n=344, for 13 ejidos ) Variable Description Mean Std Dev Min Max Mean Std Dev Min Max VINTEG1 Stumpage, log in the forest or patio (0,1) 0.25 0.44 VINTEG2 Log loaded onto a buyers truck (0,1) 0.75 0.44 VINTEG3 Board in the community/patio (cut with chainsaw; 0,1) 0.04 0.21 VINTEG4 Board in community/patio (cut with portable or permanent sawmill; 0,1) 0.96 0.21 FMPYR Year first management plan was approved 1986 4.86 1983 2006 1985 2.97 1983 2006 VOLS Total volume sold in year of sale (m3) 1,699.74 2,941.02 105.84 14,264.93 1,058.92 4,985.37 17.60 76,318.42 VOLAP Proportion of the total approved harvest volume comprised of precious wood species (%) 0.06 0.03 0.00 0.16 0.08 0.02 0.00 0.11 PERMF Total area for timber management (ha) 19,556.91 10,108.00 2,000.00 32,500.00 17,039.42 4,575.95 1,000.00 32,500.00 NFAM Number of families participating in CFE 105.94 110.59 5.00 306.00 118.12 37.99 10.00 306.00 NSAL0 0 years of previous sales to buyer 0.25 0.44 NSAL1 1 year of previous sales to buyer 0.57 0.50 0.41 0.49 NSAL2 2 years of previous sales to buyer 0.33 0.47 0.17 0.37 NSAL3 3 years of previous sales to buyer 0.09 0.29 0.14 0.35 NSAL4 4 years of previous sales to buyer 0.02 0.12 0.02 0.15 COMFORM1 All ejido members sell timber together 0.60 0.49 0.90 0.30 COMFORM2 Small groups sell timber 0.37 0.48 0.06 0.23 COMFORM3 Individuals sell timber 0.04 0.18 0.04 0.20 Boards (n=3,314, for 15 ejidos )

PAGE 90

90 Table 3 5 Results of price models for logs (n=344) and boards (n=3314) in Quintana Roo, Mexico Independent Variables Dependent Variables Log Price (P$/m3) Board Price (P$/bf) Coefficient Std. Error tstatistic P-value Stand. Coef. Coefficient Std. Error tstatistic P-value Stand. Coef. Intercept Base price 4064.080 187.693 21.653 0.000 18.959 0.428 44.310 0.000 SG2 Hardwood (0,1) -2914.300 244.646 -11.912 0.000 -2.155 -7.904 0.261 -30.170 0.000 -1.107 SG3 Softwood (0,1) -2845.320 319.884 -8.895 0.000 -2.104 -11.368 0.261 -43.453 0.000 -1.592 QG2 Second grade (0,1) -2.760 0.492 -5.605 0.000 -0.387 QG3 Third grade (0,1) -5.733 0.220 -26.031 0.000 -0.803 QG6 Chainsawn (0,1) -2.493 0.460 -5.417 0.000 -0.349 QG8 Heartwood (0,1) 0.274 0.340 0.806 0.420 0.038 QG9 Sapwood (0,1) -1.678 0.336 -5.000 0.000 -0.235 LENGTH2 Short (0,1) -6.855 0.166 -41.408 0.000 -0.960 THICK2 1.5 1.99 (0,1) -0.748 0.208 -3.591 0.000 -0.105 THICK3 >= 2 (0,1) 0.502 0.202 2.488 0.013 0.070 VOL Volume of product sold in corresponding unit -0.022 0.024 -0.924 0.356 -0.010 0.000 0.000 5.223 0.000 0.075 MKT2 Regional market (0,1) 0.197 0.864 0.290 0.819 0.028 MKT3 National market (0,1) 195.944 82.513 2.375 0.018 0.145 0.912 0.431 2.119 0.034 0.128 MKT4 Int'l market (0,1) 11.097 1.610 6.891 0.000 1.554 DISTR Distance to the highway (km) -12.859 2.689 -4.782 0.000 -0.187 0.095 0.005 1.758 0.079 0.016 NSAL1 1 year of previous sales to buyer -0.305 0.201 -1.512 0.131 -0.043 NSAL2 2 years 0.483 0.251 1.926 0.054 0.068 NSAL3 3 years 0.721 0.226 3.185 0.001 0.101 NSAL4 4 years 1.660 0.486 3.417 0.001 0.233 CERTYRS No. of years certified 20.487 10.905 1.879 0.061 0.081 0.248 0.038 6.478 0.000 0.104 VOLS Total volume sold in year of sale 0.000 0.000 -1.336 0.182 -0.010 Adjusted R2 0.663 0.634 F-test 113.318 0.000 250.612 0.000 Ramsey Reset test 4.178 0.042 248.279 0.000

PAGE 91

91 Ta ble 36. Summary statistics and results of market access probit model (n=4709) in Quintana Roo, Mexico. Descriptive statistics Model results Variable Description Mean Std Dev Min Max Coefficient Std Error t statistic P value Intercept Base probability of sale occurring in local market 2.86 4 0.114 25.209 0.000 DCERT Certification dummy variable (1=certified) 0.72 0.45 --0.934 0.106 0.880 0.379 PERMF Total permanent forest area designated for timber management 17,308.38 5,341.25 1000 32,500 0.000 0.000 8.350 0.000 DPROD Product dummy variable (1=board) 0.76 0.43 --0.216 0.095 2.263 0.024 DCOMFORM Form of organization for commercialization dummy variable (1=at group or individual level) 0.18 0.39 --0.796 0.083 9.652 0.000 Fraction of correct predictions 0.960 R squared 0.141

PAGE 92

92 CHAPTER 4 ARE COMMUNITY BASED FOREST ENTERPR ISES IN THE TROPICS FINANCIALLY VIABLE? CASE STUDIES FROM THE BRAZILIAN AMAZON Introduction Since the early 1980s community based forest management has gained popularity as a strategy to conserve forests and improve local livelihoods, particularly in the tropics (2004; Amaral and Amaral Neto, 2005; Bray et al., 2005) This strategy has coincided with rapid devolution of forested lands to communities (Agrawal, 1999; Stone and d' Andrea, 2001) such that nearly onefourth of forests in developing countries is owned and/or controlled by low income forest communities (White and Martin, 2002) and this global trend continues (Sunderlin et al., 2008). Similarly, in recent decades, community forest management has become an integral component of sustainable forest management policy in the Brazilian Amazon, where communities control 59% of the public forests or 124 million hectares (SFB and MMA, 2009) The Brazilian Forest Service (SFB in Portuguese) would like to increase the area of forests managed by communities for timber and nontimber products by supporting community forestry on half of the public land designated for forest concessions, which in 2009 was almost 43 million ha (SFB and MMA, 2009) The International Institute of Education in Brazil (IEB, 2006) reported that a total of 176 plans for community forest management in the Brazilian Amazon had been presented for government approval, and Humphries (unpublished research) found approximately 65 of these were engaged in or planning timber h arvests in 2007. About onethird of these 65 received financial support from the government program Support for Sustainable Forest Management in the Amazon (ProManejo) as pilot projects from 1999 to 2007 (ProManejo, 2006b) This support was based on the hope that community forest

PAGE 93

93 management would provide a higher standard of living for rural communities, improved natural resource governance, biodiversity conservation, and sources of wood produced both legally and in a more environmentally sustainable way (Amaral and Amaral Neto, 2000) The expectation was that at least some of the pilot projects would develop into financially viable community based forest enterprises (CFEs) after a few years of support. There is no broad evidence to date to validate this hypothesis. The pilot projects have had difficulty getting management plans approved (due in large part to bureaucratic delays) and thus advancing to timber harvests and sales. Those pilot projects who have engaged the market (hereafter re ferred to as CFEs) have had difficulty competing with both industrial operations with large volumes and established market networks and illegal operations whose low costs and ubiquitous market presence keep prices artificially low. Profitability analyses of CFEs have been limited (Amaral and Amaral Neto, 2000) and comparisons of the few existing studies are difficult due to differences in methodology and reporting of results. As the Brazilian pilot projects/CFEs face reduced or discontinued subsidies, it is timely to assess profitability and consider if profits are sufficient for community based forest management to become a competitive and compelling landuse. As ProManejo ended in 2007, this question has become imperative. Our study addresses this information need through financial analysis of three CFEs in the Brazilian Amazon that differ in scale, products, and location: Amb, an industrial scale operation producing logs in Par state; Mamirau, a group of small scale operations producing logs and boards in Amazonas state; and ACAF, a small -

PAGE 94

94 scale operation producing boards in Amazonas state. We provide estimates of profitability for each, examine operating conditions that improve and hinder profitability, and identify key challenges and opportunities for these enterprises. Comm unity Based Forest Enterprises i n t he Brazilian Amazon Community based forest management pilot projects for timber production in Brazil began in the 1990s1. In 2000, Amaral and Amaral Neto (2000) reported 14 such projects in the Brazilian Amazon region; in 2006 the number of communities that had submitted forest management plans for timber in the legal Amazon dramatically rose to 176 (IEB, 2 006) (though only about onethird followed through with the plans). The ProManejo project, funded by the PPG 7 program (Pilot Program for Tropical Forest Protection, funded by Germany, European Union, UK, the Netherlands, USA, France, Italy and Russia) supported 17 community forestry pilot projects from 1999 to 2007 (ProManejo, 2006a) The projects varied greatly in participants and types of ownership, from ranchers on individual plots, to rubber tappers in federal agroextractive reserves, to riverside communities living in state sustainable development reserves. Forms of organization also varied, including associations, producer groups, and cooperatives. The scale of production ranged from 4 to 500 ha annual harvest units, with harvest rates between 1 and 15 m3 per ha (ProManejo, 2006a) At the time of our study, six had sought and achieved Forest Stewardship Council certification for their timber management practi ces (FSC, 2008) 1 While we recognize that community forest management is implemented throughout the region for multiple forest products and services, includi ng nontimber forest products, in this article we use the term to imply management for timber production.

PAGE 95

95 The community based forest enterprises (CFEs) in the region share many challenges, perhaps the most difficult bei ng secure land tenure; in 2000, 57% of the pilot CFE projects were on untitled land (Amaral and Amaral Neto, 2000) Another is the typical and restrictive 3year project funding cycle, at which point most CFEs are just getting established and facing many first time harvest and timber sale challenges. Amaral and Amaral Neto (2000) assert CFEs need a minimum of five years of funding to become established, in part because long delays in management plan approval and frequent changes to national and state regulations result in delayed harvests. Koury (2007) found the approval process for ACAF, one of our case studies, took 17 months. Given tight operational time constraints, delays have posed an enormous risk for communities and buyers. Perhaps because of preoccupations with these challenging conditions, no CFEs have accessed the few affordable credit programs available (Amaral and Amaral Neto, 2000; personal observations) which may be critical sources of capital when subsidies end. Finally, CFEs have difficulty staying informed of the market (products and prices demanded), and their small volumes, low quality, and isolation make it difficult to access attractive markets that pay higher prices for legal and managed wood. A notable innovation to overcome these challenges was the creation of COOPERFLORESTA in Acre, which was formed by local NGOs and government agencies to coordinate field operations, documentation, certification, and sales for the six member CFEs (Trentini, personal communication). The cooperative has secured sales in the national market to fine furniture makers and the construction sector at prices as much as four times those offered in the local market (Humphries and Kainer, 2006)

PAGE 96

96 Case Studies We chose three CFEs (Map 1) to capture diversity in scale, production, and location among CFEs in the region, and for cost and income data availability (Table 4 1). While they illustrate compelling differences among Brazilian CFEs, they are not necessarily representative of all CFEs in the region. Forest management for timber production was first implemented in the Tapajos National Forest from 1999 to 2003, as an experiment supported by the International Tropical Timber Organization (ITTO) to determine if industrial scale reducedimpact logging in the Amazon could be profitable (hereafter referred to as the ITTO project) (Caetano Bacha and Estraviz Rodriguez, 2007). In 2001, ProManejo representatives, after visiting industrial scale CFEs in Mexico, initiated discussions on a proposal for a community based timber management project in the national forest (Medina and Pokorny, 2009) In 2005, COOMFLONA (Mixed Cooperative of Tapajos Green National Forest) was founded and initiated new community based forest management activities under the auspices of the Amb project (Medina and Pokorny, 2009) The Amb project had a large staff and two offices (one i n the city of Santarem and a field camp that houses workers and machinery in the national forest 83 km from the city), which contributed to high overhead and transportation costs. Temporary workers were from local communities, and many received training t hrough the ITTO project, an asset for Amb. We examined Ambs second year of production on 300 ha. Annual harvest area size increased from 100 ha for the first harvest to 1,000 ha in 2009. Amb owned two trucks for transporting staff and workers and al so rented a tractor (D60F), skidder (R Miller), and loader (Clark R W 20), which came with operators. Amb used a closed bid system for timber sales, and sold logs to a local sawmill company.

PAGE 97

97 The Mamirau Sustainable Development Institute (MSDI) works wi th community associations in the Mamirau Sustainable Development Reserve, which covers 1,124,000 ha of flooded forest located at the confluence of Solimes, Japur e Auati Paran Rivers. It is home to approximately 6,000 residents who live in 63 settleme nts (Pires, 2005) The Reserve has a long history of small scale timber extraction for cash income, especially during the rainy season when agricultural land is inundated and fishing is less productive (Pires, 2005) Legal community forest management in the reserve began in 1993 with monitoring of traditional extraction methods and participatory mapping. In 1997, work with five community associations began (Pires, 2005) and in 2008, 30 associations were involved, each with its own management area. Previous experience with logging ranged among associations, with some older residents having engaged in it or having secondary knowledge of it as children of experienced loggers, and some with little to no previous experience. The MSDI had a professional forester and several forest technicians on staff to provide training and technical and commercialization assistance; funding was provided by various grants. Once the forest inventories were completed, association representatives met to set minimum prices and to collectively negotiate timber sales with buyers, mostly intermediaries for Manaus based companies. Each year as the river rises (four to five months after advances are received), the association members begin felling trees. Most associations wait for the water level to rise and then use motorized canoes and manual labor to move the logs into streams and float them to lakes where log rafts are formed. The buyer then floats the rafts to Manaus. At the time of our study, four associations had recently begun producing boards and other milled products with a

PAGE 98

98 portable saw mill (Lucas Mill 827), which was set up around each fel led tree in the forest. In these cases, milled products were manually transported to the river bank. Heavier items were left there for the buyer to retrieve; lighter ones were transported by canoe to the buyers rafts. ACAF (Community Agriculture and For est Extraction Association) was initiated by IMAFLORA (a Brazilian certifier) and the Boa Vista do Ramos municipal government as a pilot project for reduced impact logging, community development, and a test of forest certification for community enterprises (Koury, 2007) The municipality is located on a branch of the Amazon River 367 km downriver from the capital of Manaus. ACAF was founded in 1999, with 28 members from three c ommunities, and obtained a municipal timber concession. From 1999 to 2004, ACAF conducted 10 harvesting operations (i.e., harvested several times a year within the same harvest unit while respecting harvest volume limits); its first two annual harvest uni ts were 50 ha, which then increased to 80 ha. In 2005, ACAF obtained Forest Stewardship Council (FSC) certification from SCS (another certifier). ACAFs equipment included two chainsaws, a portable sawmill (Lucas Mill 827), a mini tractor (Agralle 4100) with an attachable cart, and a boat with a 114 horse power engine (Koury, 2007) In 2006, only 12 association members chose to work in forestry; other community members were hi red to meet labor requirements as needed. ACAF received free technical assistance from a professional forester from the Manaus Federal Agrotechnical School and a forest technician from OELA, a nonprofit organization. Usually the trees are felled in the forest and sawed with the portable sawmill, and then the boards are transported by the agricultural tractor to the river port (an average of 3 km), where they are transported by boat to the community, and then

PAGE 99

99 sent to the buyer. ACAF negotiated two timber sales in 2006 with buyers who wanted FSC certified wood one in the region and one in So Paulo. We chose to focus on one sale for this case, and used average cost and income data for the two sales. Methods Data Collection Data for the ACAF case were c ollected from March 2006 to February 2007 through frequent site visits using interviews, monitoring sheets, and cash box receipts (see Koury 2007 for more details). Data for the Amb and IDSM case studies were collected in November 2007 and February 2008, using an innovative workshop framework with stakeholders in each CFE. Workshops combined training in basic logging cost and profitability calculations with data collection and analysis (more workshop details provided below). For all three cases, machinery and equipment were purchased with funds from various donors (although Amb rented the large machinery it used); materials and supplies were purchased with advances from buyers. We included all costs, and focused analysis on one operational year or harv est cycle (which can span across two calendar years). We used monitoring data for actual costs when possible; in the absence of actual costs, we used conservative estimates, erring on the side of overestimation. Cost data was organized by type (labor, ma chinery and equipment, and materials and supplies) and activity (see Table 4 2). Cost types excluded in Table 4 2 for each CFE (e.g., sales negotiations) were either subsumed under other cost types, e.g., administration) or not incurred because they are not part of the CFE production model (e.g., skidding and transport). For income, we used actual timber sales for Amb and ACAF. For Mamirau, we estimated sales revenues based

PAGE 100

100 on average harvest volume by timber value class from 2000 to 2006, and anticipated 2008 prices. Workshops The two workshops were four days each, and aimed to train staff and community members in basic financial analysis and to collect and analyze case study data in a participatory manner. For Amb, five staff (from local communities ) and the projects head forester participated. For Mamirau, seven forestry and fisheries staff participated; transportation difficulties kept one community member from participating. We first presented basic financial concepts and reviewed regional cas e studies with participants. Next, participants described the history and future of their forest management operation. Then participants broke into small groups to organize and enter cost data for each forest management activity in Excel spreadsheets. F or Amb, monitoring sheets already in use for each activity were used as the basis for the spreadsheets; for Mamirau, spreadsheets were created. Finally groups came together to analyze the data and discuss results. Case study specifications We developed representative scenarios for each case for one year, using the best information available. We recognized that costs were expected to vary by year. Production. In the case of Amb, the study was based on 2006 2007 monitoring data. For the Mamirau associations, a typical scenario was developed based on average production across all associations between 2000 and 2006, and since a portable sawmill was recently introduced in four communities, we analyzed a scenario of producing a mix of b oards and logs. Since the Mamirau associations had limited cost monitoring data, the MSDI staff used discussions with association

PAGE 101

101 members, personal observations, and their best guesses to estimate many costs; when we refer to the Mamirau case, or simply Mamirau, in this study, we mean the one representative case we developed based on all of the associations. For ACAF, monitoring data was used for field operations and administrative costs were estimated; since the association conducted two harvests on half of the annual harvest unit and sold 20 m3 of boards for each, we calculated average costs for one harvest of 20 m3 of boards. For processing logs to boards, we assumed conversion and daily productivity rates based on field observations of MSDI staff for Mamirau and Kourys observations for ACAF (Koury, 2007) The conversion rates for Mamirau and ACAF were, respectively, 45% (i.e., 2.2 m3 of logs = 1 m3 of boards) and 39% (i .e., 2.6 m3 of logs = 1 m3 of boards); average daily productivity for each was, respectively, 3.0 m3 of boards and 1.2 m3 of boards. Labor. In reality, each case handled labor costs for temporary workers differently: Amb paid a monthly salary for seven months; in Mamirau, associations split profits and did not pay wages; and in ACAF, the association paid a daily wage. However, to facilitate comparison, we used a daily wage for each CFE based on local wages and days worked (in the case of Amb, there were many days that work was not performed due to weather and or machinery problems). Thus, for Amb we calculated average daily wage rates for team leaders and other workers (based on 25 work days/month), for Mamirau we used the minimum daily wage in the region, and for ACAF we used the daily wage for association members and nonmembers. Also, although ACAF received free technical assistance from a professional forester and a forest technician, we

PAGE 102

102 included salary costs for the forester and the technician (assuming a 20% and 50% time commitment to ACAF, respectively). Quantities of equipment and materials. Quantities of machinery and equipment (e.g., helmets, chainsaws) and materials and supplies (e.g., gasoline, oil) used in each harvest were estimated for most items since these costs were infrequently documented. Once items needed per activity were estimated, total quantity of machinery and equipment needed for the harvest cycle was calculated based on whether the different activities were carried out sim ultaneously or at different times. For Amb, since several work teams implemented different activities concurrently, and each team needed its own equipment, the quantities of equipment needed per activity were summed. For example, if 10 helmets were need ed for five different activities, then the total quantity of helmets was 50. For Mamirau and ACAF, one small team implemented all the activities sequentially, so only one set of each equipment item was needed to implement all activities (e.g., if two hel mets were needed for one activity and five were needed for another, the total quantity was five)2. The same rationale was used for materials and supplies. Cost sharing. For Mamirau, the technical assistance and administration costs of the MSDI were divided by 30 the number of forestry associations in the Reserve. Similarly, the cost of Mamiraus portable sawmill was divided by four to account for the number of associations that used it. Other considerations. We assumed the CFEs were responsible for all costs, unless otherwise noted. Stumpage fees were not considered since CFEs in our study 2 While this methodology gives a more realistic estimate of the quantity of items needed, it may overestimate depreciation (e.g., the five helmets wil l be depreciated at the same rate).

PAGE 103

103 do not pay them (they are subsumed by buyers), and we are not aware of any communities in the region that do. ACAF paid the out of state 12% sales tax (ICMS) rate on one sale and the instate rate of 17% on the other, and we used the average of these; the buyers of Mamiraus and Ambs wood were assumed to pay the tax, as t hey have in previous years. No study CFEs received interest bearing loans, and thus we did not need to include the time cost of money in our calculations. However, as an additional analysis, we calculated capital needed to meet cash flow requirements for one year, and included 5% annual interest on the cost of this capital. Finally, although the RILSIM data management program was designed to facilitate collection and analysis of data for industrial tropical forestry operations (see http://www.blueoxfores try.com/RILSIM/), it was not used for this study because its data requirements and assumptions (e.g., for depreciation) were ill suited to CFE models. Data Analysis Total costs were summed by type and activity. Total income was calculated by summing the i ncome per timber value class (three classes per product), which consisted of the volume sold per value class multiplied by the respective price. Net income (total costs minus total income) was divided by total costs and multiplied by 100 to determine rate of return, or profit rate. For ACAF and Amb, values in Brazilian Reais were adjusted for inflation to the year 2008 (Mamirau data was collected in 2008). Then values for the three cases were converted to the value of the US Dollar (USD) for February 2 008. Depreciation was the most complicated part of our analysis. Depreciation is an accounting charge for the wearing out of certain assets (Klemperer, 1996, p. 527) It is used to determine the contribution towards annual costs of assets that are used for

PAGE 104

104 more than one year. In our study, assets included machinery and equipment (e.g., boats, chainsaws). Because conditions in which our cases operated include d high humidity, inadequate storage infrastructure, and limited availability of maintenance skills and replacement parts, we decided to calculate deprecation differently than normal accounting techniques3 (Equation 4 1). We did not use a resale value (the only exceptions were two expensive vehicles for Amb that could be sold at least for parts at 10% of the total value in the nearby city) and did not graduate the costs. For the same reasons, we defined useful life as the number of harvest cycles each ass et would last as estimated by workshop participants for their CFE in general, rather than on the manufacturers estimates of productivity. Materials and services were usually not depreciated because they were consumed in one year. Exceptions to the methodologies for ACAF are described below. The following equation was used to determine the depreciated value of machinery : = Pi QiLi (4 1) Di is the annual depreciated value Pi is price Qi is quantity Li is the estimated useful life in terms of harvest cycles i indexes the specific item of machinery/equipment. = = 1 (4 2) Dij is t he depreciation value of each item by activity j indexes the activity n is the total number of activities for which item i is used aij is the number of days item i was used for activity j bi is the total number of days that item i was used for all activities 3 Depreciation per hour or production unit is usually calculated by taking the price of an asset, subtracting its resale value, and dividing the difference by the total number of hours it can be used or units it can pr oduce during its useful life. Then the depreciation cost for a given year is calculated by multiplying the per hour cost or per unit cost by the number of hours worked or units produced in a given year. These costs are often graduated for present value c alculations; for our study we did not graduate the costs.

PAGE 105

105 Results Costs Administrative costs, including technical assistance and licensing fees, were the highest expenditures for Amb and ACAF, and were second only to processing for Mamirau (Table 4 3, Figure 4 2 ). Processing, especially machinery costs, was the second highest expenditure for ACAF. Notably for Amb, administrative costs accounted for 71% of total costs, of which labor and materials/services (including seven months of food for workers and maintenance, insurance, and fuel for vehicles) comprised about 40% each For the three cases, just labor costs for administration (mainly technical assistance) comprised high proportions of total costs: 28% for Amb, 19% for Mamirau, and 18% for ACAF. Notably, Ambs administrative labor costs also included eleven Cooperat ive leaders that received a 12or 7 month salary. Cost allocation between labor, machinery, and materials/services were similar for the three cases (Figure 2). A notable difference was that Amb paid a rental fee for its large equipment (e.g., skidder) w hile the other two operations paid the depreciated annual value of their equipment/machinery. The cost per cubic meter of logs for Amb (USD 92) was almost four times as high as for Mamirau (USD 24), due to the latters cost sharing of technical assistance and reliance on manual transportation of logs via waterways; the cost per cubic meter of boards was USD 112 for Mamirau, and USD 1,023 for ACAF, due to the latters low harvest volume, lower efficiency, and greater use of expensive machinery (Figure 4 3). Income Differences in prices and volumes account for different CFE incomes (Table 4 4, Figures 4 5 and 4 6 ). Price differences are clearly related to products and value

PAGE 106

106 classes (based on tree species) (Table 4 4). The differences in class 2 log pric es between Amb and Mamirau (USD 105/m3 vs. 26/m3, respectively) and class 1 board prices between ACAF and Mamirau (USD 581/m3 vs. 112/m3, respectively) are related to higher market prices for upland forest species than flooded forest species, and ACAFs certified status. Price differentials for upland species class 1 logs for Amb (USD 163/m3) and class 1 boards for ACAF (USD 581/m3) reflect the value added through processing and certification for ACAF, although species sold were dissimilar. Mamirau made a small profit on both logs and boards. Amb and ACAF had products for which average production cost was greater than price received (Figures 4 3 and 4 5). For Amb, the lowest log class price of USD 58/m3 was much lower than the average costs of producing the logs, USD 91.56/m3. For ACAF, the price of producing boards was much higher than the highest board price received. P rofit Amb and Mamirau were profitable, while ACAF was not (Figure 4 6 Table 4 5). Beyond rate of return, labor income generated was a tangible benefit for the CFE participants. For Mamirau, producing logs generated 15 labor days per person and sawing generated 33 labor days, resulting in an average total income of USD 507. For Amb, the average pay per worker was USD 2,247 3,371 for eight months. For ACAF, because subsidies covered its total operating costs, workers received an average of USD 159 for two months.4 4 Our cases paid higher wages than the 2007 national minimum of R$ 380 (Anonymous, 2008) : Amb paid its workers a monthly salary 30% to 110% more, and Mamirau and ACAF, respectively, paid daily wages 10% higher and 74% higher than the national average based on a 22day work month (although ACAF paid nonmembers 13% less than the average).

PAGE 107

107 Discussion The findings that Amb and Mamirau were potentially financially viable, while ACAF was not illustrate that community based forest management is expensive, but can be profitable, especially when CFEs produce large volumes and/ or share costs. However, profitability is quite fragile for Amb and Mamirau. For Mamirau, a decrease in the number of associations that share the costs of technical services and/or the portable sawmill would jeopardize the models viability. For Amb, if we had included the monthly salaries for temporary workers, instead of wages based on days worked, the operations profitability would drop from 11 to 10%. Nonetheless, as Amb increases its harvest unit size from 300 to 1000 ha simultaneously increasing total harvest volume, while also improving efficiencies, its fixed costs per m3 will decrease and it should be more profitable, even while paying monthly salaries. ACAF, in comparison, will need major increases in volume and efficiencies to overcome its high fixed costs, especially for machinery5. Finally, all three cases will need new sources of subsidies or better access to credit if their operations are to continue. What Makes Community Timber Production So Expensive ? Administrative costs, including technical assistance, were the largest expenditure for Amb and ACAF, and second largest for Mamirau. This fixed cost would be more manageable for Amb and ACAF if it were spread over a larger product volume or among multiple CFEs, especially where s everal CFEs operate in geographic clusters, as for Mamirau. 5 ACAF could offset the depreciation cost of its machinery by renting it out.

PAGE 108

108 Processing was a large proportion of costs for Mamirau and ACAF. For Mamirau, the difference in income (USD 60) from selling one cubic meter of boards (USD 112) versus the equivalent in logs (USD 53, or 2.2 m3 of logs at USD 24/m3), is barely more than the processing cost, USD 59/m3. However, it tripled the amount of daily wages paid. For ACAF, in comparison, the difference in income (USD 136/m3) from selling one cubic meter of boards (USD 580/m3) versus the income for the equivalent in logs (USD 444, or 2.6 m3 of logs at USD 171/m3) is much less than the processing cost, USD 200/m3. The rate of processing also affected costs, with lower productivity per day resulting in higher total labor costs. The average productivity range for the operations Lucas Mill portable sawmills is 2 to 8 m3 of boards per day, depending on the size of the trees (Andreas Nagl, personal communication). Mamirau processed 3 m3 of boards per day, whereas ACAF processed half of that both on the low end of the range. Mamiraus processing efficiency was also higher than ACAFs (2.2 m3 versus 2.6 m3 of logs for 1 m3 of boards). The sales (ICMS) taxes were also burdensome for ACAF (buyers paid these costs for the o ther case studies). ACAF paid 17% for local sales and 12% for out of state sales (which favors larger producers with better market access). What F actors A ffect R evenue? Species were divided into three value classes (Table 4 4, Appendix D ), and class pri ces varied greatly among species within and across forest types, and even across states for the same species (Lentini et al., 2005) (although there was little overlap in species sold for our cases). For Amb, class 1 prices were 2.8 times class 3 prices; class 1 log price for Amb (upland species) was 4 times the class 1 log price for

PAGE 109

109 Mamirau (flooded forest species). Mamirau technical staff attributed lower prices for flooded forest species to a lack of commercial knowledge of these species; however, Arima and Verssimo (2002) attributed low timber prices near Tef to an abundance of timber resources and relatively long distances from processing facilities (400 to 850 fluvial km). The value added through processing is also reflected in the prices CFEs received, although the local prices for boards were the same whether produced with a chainsaw or a portable sawmill (Medina and Pokorny, 2009) The pr ice of class 1 boards for Mamirau was 2.7 times higher than for class 1 logs. However as mentioned above, due to small production volumes and low efficiency, the price differential of added value versus cost for processed boards was small for Mamirau and not nearly sufficient for ACAF. Nonetheless, the additional daily wages generated from processing could be argued to be sufficient reason to subsidize processing, or at least continue training to increase efficiencies and processing viability. Gretzing er (1998) found that a CFE in the Maya Biosphere Reserve increased profitability through processing with portable sawmills. While all three cases cited low prices as a problem, in general, they obtained prices much higher than recently reported state averages (Lentini et al., 2005) and the I TTO reported domestic prices for a few species (ITTO, 2008) perhaps due to their legal status and/or innovative commercialization strategies. Amb used a closed bid system (required for timber sales in national forests), which was intended to encourage higher prices and was uncommon in the industry. However, the bid system participation rules (e.g., no outstanding taxes or loan defaults) may have excluded some potential

PAGE 110

110 buyers; Amb received one bid for the first sale and two for the second. Mamirau associations collectively determined minimum prices and negot iated joint sales with buyers, and their prices have steadily increased annually (Kibler, 2008) ACAF used its certified producers group membership to access the national market and negotiate better prices. These case studies would benefit from local price reporting; data are available for some species in Par and So Paulo (CEPEA, 2009) but this provides limited reference for isolated CFEs. Location may also have influenced sales. Amb had the advantage of proximity to a major city (Santarem) and good access to roads and rivers. Associations in the Mamirau Reserve are on average five hours by boat from Tef, the nearest city, and quite distant from the closest major market in Manaus; ACAF is also quite distant from all major markets. In addition, the strong state pres ence in Santarem and in Tef may have encouraged buyers purchase of legal wood. In contrast, Koury (2007) reported that ACAF members perceived a lack of government oversight and control, and a plethora of regionally available illegal, cheap wood as major challenges for commercialization. Total revenue could have been much higher for all three case studies if they had harvested their total approved volumes and/or produced greater volume during an equivalent time period. Amb harvested 91% of its approved volume. ACAF harvested only eight percent of its approved volume per hectare (2.6 of 29.5 m3) due to long delays in the approval process and difficulty in finding buyers interested in their relat ively small quantity of certified wood. During our study period, ACAF, unable to find a buyer for its entire approved volume before the harvesting period, received orders at

PAGE 111

111 different times. This resulted in two harvests, which were more expensive and less efficient than a single harvest would have been, although this reduced risk and perhaps waste. ACAF would need to increase production of boards by 215% in one harvest (assuming the same productivity rates) to break even (Figure 6). In the case of Mami rau, both natural (low river level) and human factors (labor deficiencies) have led to harvests of only half the approved volume, even when the entire volume was contractually obligated to buyers. Medina and Pokorny (2009) found that, relative to industr ial operations, small scale and largescale CFEs were 75% and 25% less productive, respectively. They associated the lower productivity, especially for smaller operations, with a preference for maximizing labor income over profits, a high value for free t ime, and the importance of other productive activities such as agriculture. Yet, leaving trees in the forest as the CFEs gain experience in administration, improve technical skills, and become more business savvy, may allow CFEs to capture more value for these trees in future harvests. How Do These Results Compare t o Studies o f Other Community Based Forest Enterprises? In Brazil, Medina and Pokorny (2008) found similar results for ACAF ( Table 46). For Amb, they found a much higher rate of return, however they took average costs for a 100 ha harvest unit and scaled it up to 1000 ha (the expected total area in 2009), while we used actual costs to the extent possible for 300 ha, thus cost per m3 was much lower and total income much higher. For Mamirau, they studied two different cases for log and board production, and found only logs profitable due to the higher machinery and labor costs and much smaller volume produced for boards (i.e., higher fixed costs

PAGE 112

112 per unit); they also used a lower productivity rate for processing (1.5 m3/day of boards) (Table 4 5 ). Ambs profitability was lower than the ITTO demonstration projects, which ran from 1999 to 2003 in the Tapajos National Forest (Caetano Bacha and Estravi z Rodriguez, 2007) If costs are adjusted for inflation to February 2008, the ITTO projects total costs for its fourth operating year (2002) were almost double Ambs (USD 663,190 vs. 334,183). However, Ambs cost per m3 is three times as high due t o its lower harvest area and volume: USD 91.56/m3 for Amb vs. 32.38/m3) (Caetano Bacha and Estraviz Rodriguez, 2007) This is an example of the financial advantage of larger scale production. Pedro Peixoto, another small scale CFE in Brazil that produced boards with a portable sawmill, was found to be 63% profitable (Table 4 5 ) (Pinho de Sa and de Assis Correa Silva, 2004) however we suspect full costs of technical assistance, including transportation costs (e.g., vehicles, fuel, etc.) and infrastructure (e.g., c omputers, etc.), were not included in the analysis. Other financial case studies of CFEs in Mexico, Guatemala, and Cameroon, show profitability for a range of products (Table 4 6), However, the exclusion of potentially large costs (e.g., administrative c osts, machinery depreciation, which were among the highest costs for our cases) and differences in cost accounting methods, even within the (Antinori, 2005) study, make it difficult to compare these results and draw conclusions. Nonetheless, these c ases illustrate that community based forest enterprises are operating with the help of subsidies, managing forests, producing timber products, and generating wages and profits for local populations.

PAGE 113

113 Methodological I nnovations and L imitations Our innovativ e method of combining data collection and analysis with training allowed staff for the Amb and Mamirau cases to contribute more effectively to these studies, ensured greater accuracy of data and results, and facilitated valuable discussion of conclusions. It also prepared staff for continued monitoring and evaluation of costs and income. Overall, our study also had multiple limitations. First, it was based on one operating year, and costs and income will vary annually. A long term analysis of the net present value of community forestry for these CFEs and a comparison to competing landuses would be insightful. Finally, our analyses are based on first entries into each harvest stand, as are most studies (Boltz et al., 2003) It is doubtful futur e harvests in the same stands will produce similar rates of return due to slower than anticipated growth rates and/or the incursion of illegal logging. Future F unding for CFEs in Brazil Community forestry is an expensive endeavor, and it is unclear where r equired capital will come from to keep existing CFEs in operation post subsidies (i.e., cover operating costs and replace machinery and equipment as necessary) and/or to replicate these pilot projects. Medina and Pokorny (2008 ) estimated start up costs ra nging from USD 22,400 to 348,000 (not adjusted for inflation) for eight Brazilian Amazonian CFEs including the ones in this study. The annual operating costs for our study cases ranged from USD 13,176 to 334,183. Furthermore, all three cases had insuffi cient profits to cover subsequent year operating expenses, and depended on subsidies and buyers advance payments to cover costs. With the end of the ProManejo program in December 2007, it is uncertain how some CFEs will continue to operate. The new Fundo Amaznia, a Brazilian program whose goal is to compile funds and distribute

PAGE 114

114 them to projects aimed at reducing greenhouse gas emissions related to deforestation and forest degradation, represents a new source of project based funding support for community forestry that has yet to be tapped (Tazzo de Azevedo, personal communication). Utilizing loans to finance cash flow would increase total costs, but may be necessary if subsidies end. Although some credit lines were available to CFEs, including a new c redit line from Banco da Amazonia and federal and state credit programs (e.g., PRONAF) (see Amaral and Amaral Neto, 2005 for a critique of current credit programs available to CFEs) to our knowledge not a single CFE in the Brazilian Amazon accessed these for timber production in 2007. Mami rau and Amb would need loans to cover operating costs of USD 9,750 and USD 245,000, respectively, after including in cash flow the previous years profit and buyer cash advances (25% and 10%, respectively). These loan amounts represent 73% of operating c osts in both cases, and the five percent interest rate (offered by Banco da Amazonia) would reduce profitability to zero for Mamirau and 10 percent for Amb. Nonetheless, the financial freedom could help CFEs negotiate better prices and access different m arkets. Mamirau Sustainable Development Institute notably set up a microcredit loan program for the forestry associations in 2000, but in 2004 low river levels inhibited log removal, associations were unable to pay back loans, and the program was discontinued. Similar problems have led to recurring debt with buyers, complicating efforts to negotiate prices. Key Lessons L earned With a goal of increasing the area under community forest management to potentially 145.5 million ha, community forest managem ent is an important component

PAGE 115

115 of Brazils natural resource policy. Amb demonstrated the potential profitability of largescale operations that sell logs, as well as the value of renting rather than buying equipment. Mamirau illustrated that small, verti cally integrated operations can be profitable when the substantial costs of technical assistance and processing are shared with nearby CFEs. It also illustrated the value of joint negotiations and sales. Cooperatives and producers groups can facilitate c ost sharing and collective negotiations for CFEs that are geographically clustered; isolated, small CFEs will have a harder time benefitting from these groups. ACAF showed that small operations with very small production volumes cannot afford to pay full costs of technical assistance and machinery, and that processing is not always a profitable option. Wages for timber production are an important source of cash income for forest based communities where wage opportunities are rare; this is especially tr ue for ACAF and Mamirau. While low productivity and small volumes made the use of portable sawmills costly, the generation of daily wages could be sufficient justification to continue processing, certainly from the association members perspective. As the study CFEs continue operation in the face of decreased subsidies, their main challenges to financial viability will be paying for technical assistance and replacement parts for machinery and equipment, and improving productivity. Subsidizing technical assistance through government extension services would be a huge boon to CFE viability. In addition, improving access to low interest loans would help CFEs cover costs, make them less dependent on buyer financing, and perhaps improve prices received. Additional training or a change in production strategies is necessary to increase productivity. Other significant challenges include bureaucratic

PAGE 116

116 delays and expenses, which could be reduced by further decentralization and special tax rates for CFEs. Certif ication, while it helped ACAF obtain higher prices and access to the national market, also represented a significant cost. Finally, large volumes of illegally sourced timber continue to influence market prices in Brazil, despite government efforts to impr ove monitoring and sanctioning of illegal logging activities (Banjeree et al., 2009) Looking to the future, some might argue that CFEs should become more financially independent, and use this study to support that conclusion. Others might argue for continued financial support of CFEs given: (1) CFEs are at a disadvantage in the market competing agains t industrial and illegal timber; (2) other industries receive tax breaks and low int erest loans (e.g., ranching, agriculture), (3) community forest management provides important nonmarket benefits for the wider society (e.g., carbon storage, biodiversity conservation), and (4) the environmental and social costs of not subsidizing communi ty forestry would outweigh its financial costs (e.g., deforestation for other uses, migration of families with little formal education to cities already struggling with growth). We encourage continued research on the financial viability of community fores try to inform this discussion.

PAGE 117

117 Figure 4 1 Locations of the three CFE cases analyzed within Brazils legally defined Amazon region. ( Source: Adapted from IEB & IMAZON, 2007) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Amb Mamirau ACAF Inventory & planning Commercialization Felling Skidding Processing Product measurement Transport Permanent plots Administration Certification Figure 4 2 Cost by forest management activity (as proportion of total) for three CFEs in the Brazilian Amazon. Mamirau Amb ACAF

PAGE 118

118 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Amb Mamirau ACAF Materials Machinery Labor Figure 4 3 Cost by category (as proportion of total) for three CFEs in the Brazilian Amazon. 0 200 400 600 800 1,000 1,200 Amb Mamirau ACAF USD Cost USD/m3 logs Cost USD/m3 boards Figure 4 4 Cost of production per cubic meter for three CFEs in the Brazilian Amazon.

PAGE 119

119 0 100 200 300 400 500 600 700 Amb Mamirau ACAFUSD Class 1 log Class 2 log Class 3 log Class 1 board Class 2 board Class 3 board Figure 4 5. Price received per product for three CFEs in the Brazilian Amazon. 0 50 100 150 200 250 300 350 400 Amb Mamirau ACAF USD (in thousands) Costs Income Figure 4 6 Total producti on costs and timber sale revenues for three CFEs in the Brazilian Amazon. 0.00 5,000.00 10,000.00 15,000.00 20,000.00 25,000.00 30,000.00 Class 1 board (R$ 200/m3) Class 2 log (R$ 47/m3)

PAGE 120

120 0 500 1,000 1,500 2,000 2,500 50% 100% 125% 150% 175% 200% 225% 250% 300% USD Percent of total volume production Average production cost/m3 Average weighted price/m3 Figure 4 7 Per unit production costs for ACAF and average weighted price received as a function of total volume produced, relative to actual volume produced (100%).

PAGE 121

121 Table 4 1. Summary characteristics of the three CFEs in the Brazilian Amazon.* Characteristics Amb Mamirau ACAF Organizational structure A cooperative of 37 members from 18 communities in Tapajos National Forest 30 associations in the Mamirau Sustainable Development Reserve An association with 28 members in 3 communities in the Boa Vista do Ramos municipality Unit of analysis The cooperative One hypothetical association designed to represent the 30 The association Forest type Upland forest or terra firme Flooded forest or varzea Upland forest or terra firme Timber operation Industrial scale Small scale Small scale Annual harvest area 300 ha (30 yr cycle) 18 ha (25 yr cycle) 40 ha** (25 50 yr cycle) Total annual harvest volume per product 3,650 m 3 logs 93 m 3 logs & 97 m3 boards 20 m 3 boards Annual harvest volume of logs /ha 12.2 m 3 17.2 m 3 1.3 m 3 Type of extraction Mechanized (skidder) Non mechanized (manual labor) Mechanized (small tractor) Workers 7 permanent 40 temporary 7 permanent 5 temporary 21 temporary Daily wages (USD) 12 19 11 9 18 Forest certification No No Yes Values presented are based on operational characteristics in 2007 for Amb, 200708 for Mamirau, and 2006 for ACAF. **ACAF originally planned 80 ha annual harvest units to be managed on a 25year cycle, but over the last few years it has harvested only half of this area; the area harvested per y ear will determine the final rotation cycle.

PAGE 122

122 Table 4 2. Input cost categories included for each of three CFEs in the Braz i lian Amazon. Activity Amb Mamirau ACAF Inventory & planning Sales negotiations Tree harvest Skidding Product measurement Transport Permanent plots Administrative costs Certification

PAGE 123

123 Table 4 3. Costs for each of three CFE s in the Brazilian Amazon by activity and type (USD, February 2008). Activity Labor Machinery Materials Subtotals % Labor Machinery Materials Subtotals % Labor Machinery Materials Subtotals % Inventory & planning 5,513 20,261 1,830 27,604 8% 320 154 178 653 5% 647 708 125 1,481 7% Felling 10,042 2,447 1,213 13,702 4% 171 409 226 806 6% 854 596 366 1,816 9% Skidding 2,130 37,257 9,487 48,874 15% 256 190 206 652 5% 0% Processing 0% 1,735 1,653 2,375 5,763 44% 1,037 2,573 446 4,056 20% Product measurement 3,748 490 487 4,725 1% 160 234 311 705 5% 0% Commercialization 0% 811 62 873 7% 0% Transport 0% 0% 1,147 1,383 641 3,171 15% Permanent plots 1,098 244 1,343 0% 0% 0% Administration 94,581 35,594 107,759 237,934 71% 2,516 249 959 3,724 28% 3,661 292 4,459 8,412 41% Certification 0% 0% 1,530 1,530 7% Subtotals 117,113 96,293 120,776 334,183 5,970 2,889 4,317 13,176 7,346 5,553 7,567 20,465 % 35% 29% 36% 45% 22% 33% 36% 27% 37% Amb Mamirau ACAF Table 4 4. Income data for each of three CFEs in the Brazilian Amazon (USD, February 2008) Amb Mamirau ACAF Product prices (USD) Price Volume Value Price Volume Value Price Volume Value Logs Class 1 162.72 735.00 119,599.57 40.45 Class 2 104.61 1,747.00 182,746.76 26.40 93.00 2,455.62 Class 3 58.11 1,169.00 67,935.81 22.47 Boards Class 1 112.36 97.65 10,971.91 581.30 5.00 2,906.50 Class 2 489.52 7.50 3,671.37 Class 3 397.73 7.50 2,982.99 TOTAL 370,282.14 13,427.53 9,560.87

PAGE 124

124 Table 4 5. Costs, income, and profit for the three CFEs in the Brazilian Amazon (USD, February 2008). Amb Mamirau ACAF Total Costs 334,183 13,176 20,465 Income 370,282 13,428 10,593 Profit 36,108 252 9,872 Rate of return 11% 2% 48% This is the average income from the two sales (Total income/2); in Table 4 4 the total income is based on average prices and an estimate of volume per product based on the two sales.

PAGE 125

125 Table 4 6. Financial case studies of CFEs Author Location Product Size of operation Rate of return Study limitations Medina and Pokorny (2008) Mamirau, Amazonas State, Brazil Tropical natural forest logs 22 ha per year 55% Tropical natural forest boards sawn with a portable sawmill 10 ha per year 54% Medina and Pokorny (2008) Amb, Par State, Brazil Tropical natural forest logs 1000 ha per year 81% Took average costs for one 100 ha unit and scaled up Caetano Baucha and Estraviz Rodriguez (2007) ITTO project, Par State, Brazil Tropical natural forest logs 693 ha per year 36% Pinho de Sa and Assis Correa Silva (2004) Pedro Peixoto, Acre State, Brazil Tropical natural forest logs 4 ha per year 63% Torres Rojo et al. (2005) El Balcon, Guerrero State, Mexico Temperate natural forest of mixed conifers and hardwoods, and planted pine Sawn and dried boards Total area was 15,190 ha natural forest, 163 ha pine plantation 20 30% Did not include debt payments, depreciation or taxes Antinori (2005) 45 CFEs, Oaxaca State, Mexico Temperate natural forest Stumpage (standing trees) Size varied Averages: 39% Accounting methods differed among communities

PAGE 126

126 Table 4 6 Continued Author Location Product Size of operation Rate of return Study limitations Temperate natural forest Logs 48% Temperate natural forest Boards 54% Temperate natural forest Finished products 32% Gmez and Ramirez (1998) (see also Ammour et al., 1995 in Gretzinger, 1998) San Miguel, Petn State, Guatemala Tropical natural forest boards sawn with a portable sawmill 128 ha per year 46% Excluded concession fees and administrative costs Ezzine de Blas et al (2009) 20 communities, Cameroon Humid natural forest logs under company management28 General total range reported as 3000 ha to 4500 ha 63% Excluded demarcation of harvest units, forest inventories, and machinery Humid natural forest logs under self management 58% 29 28 The communities were given two timber production options: 1) enter int o an agreement with a company which would incur all costs; or 2) self management, in which the community incurred all costs. Four chose self management. 29 Ezzine de Blas et al. (2009) also report that the communities total benefits per harvested unit for the self management option (.2/m3) was on average twice the company agreement option (.4) due to the labor income for the former.

PAGE 127

127 CHAPTER 5 CONCLUSIONS These three studies provided a comprehensive look at CFEs in the Brazilian Amazon and in Quintana Roo, Mexico two hotspots of community based forest management in the tropics. We first provided an historical and contemporary overview of the models of co mmunity forestry for all CFEs in three states of the Brazilian Amazon and a large proportion of the CFEs harvesting timber in Quintana Roo. Then we analyzed over 4,000 timber sales for 25 ejidos in Quintana Roo to identify the factors that affected the pr ices CFEs received and their access to nonlocal markets. Finally, we performed detailed financial analysis for three CFEs in the Brazilian Amazon, and reflected on factors affecting profitability and potential ways to improve returns. In the first study we found that while in both study regions community forestry was important to helping communities formalize their tenure rights, this process was realized in different contexts, especially regarding policies and markets, and resulted in distinct CFE model s. Our analysis identified some general trends in the development of CFE models, as well as some important lessons that each region could learn from the other. The general tendencies regarding CFE models in the two regions was to move away from the earl y model of each CFE having its own permanent sawmill and selling timber products individually, and to look to new partners for producing and selling timber. Valueadded strategies began focusing on portable sawmills and chainsaws instead of expensive and difficult to maintain permanent sawmills. Also, CFE producer groups, timber cooperatives, and joint ventures with companies were becoming more common, especially in Brazil. The two former options were used where clusters of CFEs existed.

PAGE 128

1 28 Important lesson s in Mexico for Brazil included a simplified yet still rigorous process of obtaining forest management and operating plans, government programs for low interest loans for investment in equipment and machinery, and the provision by local civil organizations of technical assistance on a feefor service basis. Brazil offered several lessons in how to improve marketing and sales for CFEs. These included CFE joint negotiations with buyers and marketing through producer groups, the use of cooperatives for centr alized processing and/or sales, and joint ventures with companies. Further research on the innovations and adaptations of CFEs to local conditions is important for informing policy makers and donors. Results should also be actively shared with CFEs to as sist them in the constant adjustments needed for businesses to remain competitive. In the second study we looked at the role of forest certification and other factors on the CFE prices and market access. This study clearly shows that certification, at least over time, has contributed to higher prices for CFEs in QR, and that the pricing system does differentiate according to species, quality, market, and other attributes of sales and producers. In addition to certification, there are other ways that CFEs can attempt to maximize the value of their forest resources including processing, sorting by species, quality, length, and thickness, accessing new markets, and maintaining good relationships with buyers. The best way for CFEs to do this may be through collaborative efforts with other CFEs through producer groups and/or cooperatives, and/or with companies. However, for this to work, the bonds that tie CFEs to local buyers (prefinancing, acceptance of verbal agreements based on trust, and abundance of l esser known species with low demand) will need to be broken. Alternative

PAGE 129

129 arrangements that address CFEs needs while also appealing to buyers in other markets will need to be negotiated. In the third study, we evaluated the profitability of three CFEs i n the Brazilian Amazon. These were: AMBE, an industrial scale, upland forest operation producing logs in a national forest, in Par state; ACAF, a small scale operation in flooded forests producing boards with a portable sawmill in southern Amazonas state; and Mamirau, one of 30 CFEs in a reserve, in northern Amazonas state producing logs and boards in flooded forests. The most substantial costs were technical, especially technical assistance, and processing. In two cases, the additional cost of process ing was larger than the value added to the product. We found t he Amb and Mamirau cases were profitable, with respective rates of return of 11% and 2%. The Mamirau case benefitted from splitting technical assistance costs among 30 small CFEs and the co sts of a portable sawmill among four CFEs. The ACAF case was not profitable, due largely to the low volume of wood sold and the high fixed costs of machinery. This study illustrates the benefits of cost sharing among CFEs, and the potential return for in vestments in small and largescale community forestry. It also highlights the need for policies and programs that recognize the myriad challenges CFEs face and the opportunities that continued support for CFEs present for forest management and economic dev elopment. Overall, this research highlighted the important role forest policy has had in empowering communities to manage and harvest their timber resources. Substantial funding and effort have been invested to build capacity in communities to manage thei r forests, but many challenges remain. On one hand, communities have made good

PAGE 130

130 progress in adopting management techniques and will likely improve with continued training and experience, but on the other, little progress has been made in CFEs as businesses This study highlighted innovations in CFE models, identified factors affecting prices and suggested ways to use these to improve prices, and identified important ways in which communities are improving their economies of scale and could improve it more. Some CFEs will have the human capital and organizational capacity to advance as businesses, finding access to financial capital, making necessary investments, and developing marketing and sales skills over time. However, many others might fare better in cooperatives with professional staff. For others, the best option might be partnering with companies, a situation in which each partner could build upon their strengths. This research identified CFEs employing all of these models, and future research sh ould follow these and document the advantages, disadvantages, and benefits to communities of each. We are sure there is not one ideal model for all CFEs. Community forest management has many benefits. Not all CFEs will be profit maximizers; for some generating employment and cash income may be the most important objective. For others, it may be establishing control over forest resources and land. Nonetheless, timber management is an expensive and complex activity, and few CFEs will be financially viab le without subsidized costs, especially technical assistance.

PAGE 131

131 APPENDIX A CFES IN THE BRAZILI AN AMAZON INCLUDED I N THE STUDY

PAGE 132

132 CFE State Value added Strategy Product Works with Coop Works with Company Associao dos Moradores e Produtores do Projeto do Assentamento Agroextrativista Chico Mendes (Cachoeira) Acre fell trees and skid logs logs yes -Associao dos Produtores do Projeto de Assentamento Agroextrativista Seringal Equador Acre fell trees and skid logs logs yes -Associao Agroextractivista de So Jose (Mossor) Acre -standing trees --Associao de Produtores Rurais em Manejo Florestal e Agricultura (APRUMA) Acre portable sawmill logs and sawnwood yes -Associao de Serengueiros de Porto Dias (Palho) Acre permanent sawmill logs and sawnwood yes -Associao Vicente de Melo Acre -standing trees -Associao dos Produtores do Projeto de Assentamento Agroextractivista Santa Quiteria (AMPAESQ ) Acre -standing trees --Associao de Moradores do Remanso Capixaba Acre (AMARCA) Acre fell trees logs yes -AMBE (Cooperative FLONA Tapajos Verde) Para fell and skid trees logs yes -Oficinas Caboclas Cooperative Para end products finished products yes -Associao Esperanca Para permanent sawmill logs and sawnwood -yes Associao Virola Jatoba Para permanent sawmill logs and sawnwood -yes Associao Arimum Para portable sawmill logs and sawnwood -yes Associao Jucara no sales data Para end products finished products yes -Associao Comunitria Agrcola e de Extrao de Produtos da Floresta (ACAF) Amazonas portable sawmill sawnwood --Associaoes do RDS Mamirau (26 assoc) Amazonas portable sawmill logs and sawnwood --

PAGE 133

133 APPENDIX B SAMPLING STRATEGY FO R MEXICAN CFES

PAGE 134

134 A 7 8 CFEs Certified (8) Not Certified (70) With a mill (7) Without a mill (1) No Precious sp. (0 ) With a mill (5 ) Without a mill ( 65) 1) Caoba 2) Chacchoben 3) Naranjal Poniente 4) Noh Bec 5) Petcacab 6) Tres Garantias 7) X hazil 8 ) Laguna Kana 9) Bacalar 10) Felipe Carrillo Puerto 11) Laguna Om Plan de la Noria 1 7 ) Andres Q.R. 18) Candelaria II 19) Chancah Derrepente 20) Francisco Villa 21) Gavilanes 22) Piedras Negras 23) Puerto Arturo 24) Santa Isabel 25) Santa Elena 1 2 ) Chan Santa Cruz 13) Dzula 14) Francisco Botes 15) Rio Verde 16) Divorciados Precious sp. (20) No Precious sp. (45) Precious sp. ( 4 ) No Precious sp. (1 ) Precious sp. (1 ) No Precious sp. (0 ) Precious sp. (7) Sampled All Sampled All Sampled 3 Sampled 5 Sampled 9

PAGE 135

135 APPENDIX C CFES IN QUINTANA ROO INCLUDED IN THE STUDY

PAGE 136

136 CFE Value added strategy Primary product Secondary product Andres Quintana Roo chainsawn wood sawnwood small diameter logs Bacalar portable sawmill logs and sawnwood small diameter logs Candelaria II -standing trees railroad ties Caoba permanent sawmill logs and sawnwood small diameter logs, furniture Chacchoben fell trees logs -Chan Santa Cruz chainsawn wood logs and sawnwood small diameter logs Chancah Derrepente chainsawn wood logs and sawnwood small diameter logs Dzula chainsawn wood logs and sawnwood small diameter logs Felipe Carrillo Puerto permanent sawmill logs and sawnwood small diameter logs Francisco Botes portable sawmill logs and sawnwood -Gavilanes -standing trees small diameter logs, railroad ties Laguna Kana chainsawn wood logs and sawnwood small diameter logs, railroad ties Laguna Om fell trees and skid logs logs -Los Divorciados fell trees logs small diameter logs Naranjal Poniente permanent sawmill logs and sawnwood small diameter logs Noh Bec permanent sawmill sawnwood finished products Petcacab permanent sawmill logs and sawnwood small diameter logs, furniture Piedra Negra chainsawn wood logs and sawnwood small diameter logs Puerto Arturo -small diameter logs -Rio Verde chainsawn wood sawnwood small diameter logs Santa Elena fell trees logs -Santa Isabel -standing trees small diameter logs Tres Garantias permanent sawmill logs and sawnwood small diameter logs, furniture X hazil permanent sawmill logs and sawnwood small diameter logs Yodzonot Nuevo chainsawn wood logs and sawnwood small diameter logs

PAGE 137

137 APPENDIX D TREE SPECIES IN EACH VALUE CLASS BY CFE Value class CFEs Amb Mamirau ACAF High Class 1 Angelim pedra Capitari Ip Cedro rosa Faveira Muiracatiara Ip roxo Git Pau d'rco Ip amarelo Louro abacate Sucupira Jatoba Louro amarelo Louro p reto Louro chumbo Muiracatiara Louro inamu Muirapiranga Louro preto Maparajuba Mulateiro Perereca Pipinho Piranheira Tanimbuca Medium Class 2 Cedrorana Abiorana Pau ferro Currupix Arapari Piquiarana Garapeira Araparirana Itaba Assacu Maaranduba Biribarana Sapulcaia Castanharana Tatajuba Caxinguba Tauari Cedrinho Macacaricuia Mangarana Muiratinga Munguba Mungubarana Mututi Paricarana Tacacazeiro Ucuba Low Class 3 Jutai mirim Murupita Marup Fava vermelho Seringa barriguda Louro faia Goiabo Jarana Louro rosa Louro vermelh o Quaraba cedro Virola

PAGE 138

138 LIST OF REFERENCES Agrawal, A., 1999. "Community" and natural resource conservation. In: Gale, F., M'Gonigle, R.M. (Eds.), Nature, Production, Power: Towards an ecological political economy. Edward Elgar, London, UK, pp. 3555. Agrawal, A., Gibson, C., 1999. Enchantment and disenchantment: The role of community in natural resource conservation. Wor ld Development 27, 4, 629649. Agrawal, A., Ostrom, E., 2001. Collective action, property rights, and decentralization in resource use in India and Nepal. Politics & Society 29, 4, 485514. Allegretti, M., 1990. Extractive reserves: an alternative for reconciling development and environmental conservation in Amazonia. In: Anderson, A. (Ed.), Alternatives to Deforestation: Steps toward sustainable use of the Amazon Rain Forest. Columbia University Press, New York, pp. 252262. Allegretti, M.H., 1994. Reservas Extrativistas: Parmetros para uma poltica de desenvolvimento sustentvel na Amaznia. In: Arnt, R.A. (Ed.), O Destino da Floresta: Reservas extrativistas e desenvolvimento sustentvel na Amaznia. Relume Dumar, Rio de Janeiro, Brazil, pp. 1747. Almeida, M.B.d., 2002. The politics of Amazonian conservation: the struggles of the rubber tappers. Journal of Latin American Anthropology 7, 270219. Amaral Neto, M., 2002. VI Oficina de Manejo Florestal Comunitario da Amaznia Brasileira. Instituto Internacional de Educao do Brasil, Brasilia, Brazil, 24 p. Amaral, P., Amaral Neto, M., 2000. Manejo Florestal Comunitrio na Amaznia Brasileira: Situao atual, desafios e perspectivas. Instituto Internacional de Educao do Brasil, Brasilia. 57 p. Amaral, P., A maral Neto, M., 2005. Manejo Florestal Comunitrio: processos e aprendizagens na Amaznia brasileira e na Amrica Latina. IIEB and IMAZON, Belm, 84 p. Anonymous, 2008. Brazil raises minimum wage by nine per cent. In: The Star Online. March 2, 2008. Rio de Janeiro, Brazil; http://biz.thestar.com.my/news/story.asp?sec=business&file=/2008/3/2/business/ 20080302075622 Antinori, C., 2005. Vertical Integration in the Community Forestry Enterprises of Oaxaca. In: Bray, D.B., Merino Prez, L., Barry, D. (Eds.), The Community Forests of Mexico. University of Texas Press, Austin, pp. 241272. Arima, E., Verssimo, A., 2002. Preos da madeira em p em plos madereiros prximos de cinco florestas nacionais da Amaznia. Ministrio do Meio Ambiente, Braslia, 28 p.

PAGE 139

139 Arnold, J.E.M., 1998. Managing Forests as Common Property. FAO Forestry Paper 136. FAO, Rome, Italy, 69 p. Banjeree, O., Macpherson, A., alavalapati, J., 2009. Toard a policy of sustainble forest management in Brazil: A historical analysis. The Journal of Environment & Development 18, 2, 130153. Barham, B., Coomes, O., 1994. Wild rubber: industrial organization and the microeconomics of extraction during the Amazon rubber boom (1860 1920). Journal Of Latin American Studies 26, 1, 3772. Bauch, S.C., Amacher, G., Merry, F., 2007. Cost of harvesting, transportation, and milling in the Brazilian Amazon: Estimation and policy implications. Forest Policy And Economics 9, 903 915. Boltz, F., Holmes, T.P., Carter, D.R., 2003. Economic and environmental impacts of con ventional and reducedimpact logging in Tropical South America: a comparative review. Forest Policy and Economics 5, 1, 6981. Brannman, L., Buongiorno, J., Fight, R., 1981. Quality adjusted price indices for Douglas fir timber. Western Journal of Agricult ural Economics 6, 2z, 259272. Bray, D., Merino Prez, L., Barry, D. (Eds.), 2005. The Community Forests of Mexico: Managing for sustainable landscapes. University of Texas Press, Austin. 390 p. Bray, D.B., Antinori, C., Torres Rojo, J.M., 2006. The Mexica n model of community forest management: The role of agrarian policy, forest policy and entrepreneurial organization. Forest Policy and Economics 8, 4, 470484. Bray, D.B., Ellis, E.A., Armijo Canto, N., Beck, C.T., 2004. The institutional drivers of sustai nable landscapes: a case study of the 'Mayan Zone' in Quintana Roo, Mexico. Land Use Policy 21, 333346. Bunker, S.G., 1984. Modes of extraction, unequal exchange, and the progressive underdevelopment of an extreme periphery: the Brazilian Amazon, 1600198 0. The American Journal of Sociology 89, 5, 10171064. Caetano Bacha, C.J., Estraviz Rodriguez, L.C., 2007. Profitability and social impacts of reduced impact logging in the Tapajs National Forest, Brazil A case study. Ecological Economics 63, 7077. Ca rrera, F., Stoian, D., Campos, J.J., Morales, J., Pinelo, G., 2006. Forest certification in Guatemala. In: Cashore, B., Gale, F., Meidinger, E., Newsom, D. (Eds.), Confronting Sustainability: Forest certification in developing and transitioning countries. Yale School of Forestry and Environmental Studies Publication Series, New Haven, CT, pp. 363406.

PAGE 140

140 Cashore, B., Gale, F., Meidinger, E., Newsom, D., 2006. Forest certification in developing and transitioning countries: Part of a sustainable future? Envir. 4 8, 9, 6. CEPEA, 2009. Informativo CEPEA Setor Florestal. Universidade de So Paulo and Escola Superior de Agricultura "Luiz de Queiroz". Date accessed: June 16, 2009. http://www.cepea.esalq.usp.br/florestal/files/2009/05mai.pdf. Charnley, S., Poe, M., 2007. Community forestry in theory and practice: Where are we now? Annual Review of Anthroplogy 36, 301336. Conroy, C., Mishra, A., Rai, A., 2002. Learning from self initiated community forest management in Orissa, India. Forest Policy and Economics 4, 227237. Conroy, M., 2007. Branded!: How the 'certification revolution' is transforming global corporations. New Society Publishers, Gabriola Island, Canada. 335 p. de Jong, W., Ruiz, S., Becker, M., 2006. Conflicts and communal forest management in northern B olivia. For. Pol. and Econ. 8, 447457. Dean, W., 1987. Brazil and the Struggle for Rubber: A study in environmental history. Cambride University Press, Cambridge, UK. 228 p. Ezzine de Blas, D., Ruiz Prez, M., Sayer, J.A., Lescuyer, G., Nasi, R., Karsenty A., 2009. External Influences on and Conditions for Community Logging Management in Cameroon. World Development 37, 2, 445456 Forster, R., Albrecht, H., Belisle, M., Caballero, A., Galletti, H., Lacayo, O., Ortiz, S., Robinson, D., 2003. Forest Communi ties and the Marketing of Lesser used Tropical Hardwoods in Mesoamerica, DF, Mexico, 150 p. Fredericksen, T.S., Putz, F.E., Pattie, P., Pariona, W., PenaClaros, M., 2003. Sustainable forestry in Bolivia Beyond planning logging. Journal Of Forestry 101, 2, 37. FSC, 2008. FSC Certified Forests. Forest Stewardship Council (FSC) Accreditation Business Unit, Bonn, Germany. 2008, February 1, 2008. Date accessed: January 2008. http://www.fsc.org/en/whats_new/fsc_certificates/certificate_lists Galarraga, I., Ma rkandya, A., 2004. Economic techniques to estimate the demand for sustainable products: A case study for fair trade and organic coffee in the United Kingdom. Economia Agraria y Recursos Naturales 4, 7, 109134. Galletti, H., 1998. The Maya Forest of Quinta na Roo: Thirteen years of conservation and community development. In: Primack, R., Bray, D., Galletti, H., Ponciano, I. (Eds.), Timber, Tourists, and Temples: Conservation and development in the Maya Forest of Belize, Guatemala, and Mexico. Island Press, W ashington, DC, pp. 3346.

PAGE 141

141 Garrod, G., Willis, K., 2000. The Hedonic Price Method. In, Economic Valuation of the Environment: Methods and case studies. Edward Publishing, Inc., Northhampton, MA, pp. 87121. Glasmeier, A.K., Farrigan, T., 2005. Understanding community forestry: a qualitative metastudy of the concept, the process, and its potential for poverty alleviation in the United States case. The Geographical Journal 171, 1, 5669. Gretzinger, S., 1998. Community forest concessions: an economic alternati ve for the Maya Biosphere Reserve in the Petn, Guatemala. In: Primack, R., Bray, D., Galletti, H., Ponciano, I. (Eds.), Timber, Tourists, and Temples. Island Press, Washington, DC, pp. 111124. Hall, A.L., 1997. Sustaining Amazonia: grassroots action for productive conservation. Manchester University Press, Manchester, UK. 269 p. Hecht, S., Cockburn, A., 1990. The fate of the forest: Developers, destroyers and defenders of the Amazon. Harper Perennial, New York, NY. 357 p. Hirakuri, S.R., 2003. Can Law Save the Forest? Lessons from Finland and Brazil. CIFOR, Bogor, Indonesia, 120 p. Holmes, T.P., Bentley, W.R., Broderick, S.H., Hobson, T., 1990. Hardwood stumpage price trends and characteristics in Connecticut. Northern Journal of Applied Forestry 7, 1, 13 16. Hummel, A.C., 2001. Normas de Acesso ao Recurso Florestal na Amaznia Brasileira: O caso de manejo florestal madeireiro. MS Thesis. Instituto Nacional de Pesquisas da Amaznia. Universidade do Amazonas, Manaus, Brazil, 101 p. Humphries, S.S., Kain er, K.A., 2006. Local perceptions of forest certification for community based enterprises. Forest Ecology and Management 235, 3043. King, S.A., Schreiner, D.F., 2004. Hedonic Estimation of Southeastern Oklahoma Forestland Prices. Conference paper. Southe rn Agricultural Economics Association Annual Meeting, Tulsa, Oklahoma, February 1418, 2004. IEB, 2006. Iniciativas de Manejo Florestal Comunitario e de Pequena Escala na Amaznia Legal. IIEB, IMAZON, ALFA, USAID. http://www.iieb.org.br/diagnosticomfc/index.html, Date accessed: January 11, 2007 ITTO, 2008. Tropical Timber Market Report. International Tropical Timber Organization. 13, 4. Date accessed: May 23, 2009. http://www.itto.int/en/mis_detail/id=39010000.

PAGE 142

142 Kainer, K.A., Schmink, M., Leite, A.C.P., Fad ell, M.J.d.S., 2003. Experiments in forest based development in western Amaznia. Society and Natural Resources 16, 10, 869886. Keck, M.E., 1995. Social equity and environmental politics in Brazil: Lessons from the rubber tappers of Acre. Comparative Poli tics 27, 4, 409424. Kibler, J.F., 2008. Comercializao de madeira em tora na RDS Mamirau. Documento Interno., Manaus, Brazil, 11 p. Klemperer, W.D., 1996. Forest Resource Economics and Finance. McGraw Hill, Inc., New York. 551 p. Klooster, D., Ambinakudige, S., 2005. The Global Significance of Mexican Community Forestry. In: Bray, D.B., Merino Prez, L., Barry, D. (Eds.), The Community Forests of Mexico: Managing for sustainable landscapes. University of Texas Press, Austin, TX, pp. 305334. Kollert, W., Pagan, P., 2007. Do certified tropical logs fetch a market premium? A comparative price analysis from Sabah, Malaysia. Forest Policy and Economics 9, 7, 862868. Koury, C.G.G., 2007. Manejo Florestal Comunitrio em Terrafirme no Baixo Amazonas: Custos e Entraves da Produo Madeireira. Mestre em Cincias de Florestas Tropicais. Instituto Nacional de Pesquisas da Amazonia INPA, Universidade do Amazonas UFAM, Manaus, Amazonas. 119 p. Lentini, M., Verssimo, A., Sobral, L., 2005. Fatos Florestais da Ama znia. IMAZON, Belm, Brazil. 138 p. Maguire, K.B., Owens, N., Simon, N., 2004. The price premium for organic baby food: a hedonic analysis. Journal of Ag. and Res. Econ. 29, 1, 132149. Maynard, B., Robinson, D., 1998. Quintana Roo Forest Certification. Ethical Trade and Sustainable Rural Livelihoods Case studies. Natural Resources Institute and Natural Resources and Ethical Trade. 46. Date accessed: February 26, 2010. http://www.research4development.info/PDF/Outputs/Forestry/R7285Quintana_R oo_Forest_Certification.pdf. Medina, G., Pokorny, B., 2008. Avaliao financeira de sistemas de manejo florestal por produtores familiares apoiados pelo ProManejo: Verso preliminar. ForLive Consortium. http://www.waldbau.uni freiburg.de/forlive/06_Products/SciPub/Com pendio_Avalia_Financeira.pdf. Date accessed: March 1, 2009. http://www.waldbau.uni freiburg.de/forlive/06_Products/SciPub/Compendio_Avalia_Financeira.pdf.

PAGE 143

143 Medina, G., Pokorny, B., 2009. Avaliao financeira de sistemas de manejo florestal por produtores familiares apoiados pelo ProManejo: Verso preliminar. ForLive Consortium. http://www.waldbau.uni freiburg.de/forlive/06_Products/SciPub/Compendio_Avalia_Financeira.pdf. Date accessed: March 1, 2009. http://www.waldbau.uni freiburg.de/forlive/06_Products/Sci Pub/Compendio_Avalia_Financeira.pdf. Merino Perez, L., 2004. Las polticas forestales y de conservacin en Mxico y en Quintana Roo. In: Armijo, N., Llorens, C. (Eds.), Uso, Conservacin y Cambio en los Bosques de Quitana Roo. Universidad de Quitana Roo, C hetumal, Mexico, pp. 1542. Merino, L., Alatorre, G., Cabarle, B., Chapela, F., Madrid, S. (Eds.), 1997. El Manejo Forestal Comunitario en Mxico y Sus Perspectivas de Sustentabilidad. Centro Regional de Investigaciones Multidisciplinarias, Universidad Nac ional Autonoma de Mexico, Cuernavaca, Mexicop. MMA (Minister of the Environment), Government of Acre, 1999. Alternativas para o desenvolvimento de atividades sustentveis. Agenda Positiva do Estado do Acre. Government of Brazil, Brasilia. http://www.amazon ia.org.br/guia/detalhes.cfm?id=13324&tipo=6&cat_id=38&sub cat_id=505, Date accessed: June 3, 2005 Molnar, A., 2003. Forest Certification and Communities: Looking forward to the next decade. Forest Trends, Washington, D.C. Date accessed: March 1, 2004. http://www.forest trends.org/documents/publications/Forest%20Certification%20and%20Communit ies.pdf. Molnar, A., Liddle, M., Bracer, C., Khare, A., White, A., Bull, J., 2007. Community based Forest Enterprises in Tropical Countries: Status and Potential. Fores t Trends and Rights and Resources Group, for the International Tropical Timber Organization, Washington, DC, 86 p. Murdiyarso, D., Skutsch, M., 2006. Community Forest Management as a Carbon Mitigation Option: Case studies. Center for International Forestry Research (CIFOR), Bogor, Indonesia. 125 p. Nimon, W., Beghin, J., 1998. Are Eco labels Valuable? Evidence from the Apparel Industry. CARD Working Paper. Center for Agricultural and Rural Development, Iowa State Universityp. Nittler, J.B., Nash, D.W., 199 9. The certification model for forestry in Bolivia. Journal of Forestry 97, 3, 3236.

PAGE 144

144 Pinho de Sa, C., de Assis Correa Silva, F., 2004. Aspectos Financeiros e Gerenciais do Manejo Florestal para Produo de Madeira Certificada em reas de Reserva Legal em Pequenas Propiedades, no Acre. Comunicado Tcnico 161. Embrapa, Rio Branco, Acre, 4 p. Pires, A., 2005. Case study Mamiraua SDR Amazonas, Brazil. In: Molnar, A., Liddle, M., Bracer, C., Khare, A., White, A., Bull, J. (Eds.), Community based Forest Ente rprises in Tropical Forest Countries: Status and Potential Report to the ITTO. Forest Trends/Rights and Resources Group, Washington, DC, pp. 8695. ProManejo, 2006a. Iniciativas em manejo florestal na Amaznia: Relatos de oito experincias apoiadas pelo ProManejo. Periodo 2000 2004, Brasilia, Brazil, 66 p. ProManejo, 2006b. Relatorio de Atividades, Brasilia, Brazil, 80 p. ProManejo, 2007. Relatorio Final do Sistema de Monitoramento SM&A. IBAMA, Manaus, 70 p. Puttock, G.D., Prescott, D.M., Meilke, K.D ., 1990. Stumpage Prices In Southwestern Ontario A Hedonic FunctionApproach. Forest Science 36, 4, 11191132. Richards, M., 1997. Common property resource institutions and forest management in Latin America. Development and Change 28, 1, 95117. Sabogal C., de Jong, W., Pokorny, B., Louman, B. (Eds.), 2008. Manejo Forestal Comunitario en Amrica Latina: Experiencias, lecciones aprendidas, y retos para el futuro. CIFOR, Bogor, Indonesia. 294 p. Scherr, S., White, A., Kaimowitz, D., 2004. A New Agenda for Forest Conservation and Poverty Reduction: Making markets work for low income producers. Forest Trends, CIFOR, IUCN, Washington, DC. 160 p. Schmink, M., Wood, C., 1987. The "Political Ecology" of Amazonia. In: White, G.F. (Ed.), Lands at Risk in the Third World: Local level perspectives. Westview Press, Boulder, USA, pp. 3857. Schmink, M., Wood, C.H., 1992. Contested Frontiers in Amazonia. Columbia University Press, New Yorkp. SEDARI, 2005. Programa Sectoral de Desarrollo Forestal 20052011. SEDARI, Chetu mal, Mexico, 36 p. SFB and MMA, 2009. Plano Anual de Outorga Florestal, Brasilia, Brazil, 111 p. Stone, R.D., d' Andrea, C., 2001. Tropical Forests and the Human Spirit: Journeys to the brink of hope. University of California Press, Berkeley and Los Angel es, California and London, England. 315 p.

PAGE 145

145 Weinstein, B., 1983. The Amazon Rubber Boom, 18501920. Stanford University Press, Stanford, USA. 356 p. White, A., Martin, A., 2002. Who Owns the World's Forests? Forest tenure and public forests in transition. Forest Trends and the Center for International Environmental Law, Washington, DC, 30 p. Wiersom, K.F., 2009. Community forestry between local autonomy and global encapsulation: quo vadis with environmental and climate change payments? Conference paper. Thinking Globally -Acting Locally: Community forestry in the international arena, Pohkara, Nepala, September 1518, 2009. Wilshusen, P., 2006. ITTO Country Case Study: Petcacab. In: Molnar, A., Liddle, M., Bracer, C., Khare, A., White, A., Bull, J. (Eds. ), Community based Forest Enterprises in Tropical Forest Countries: Status and potential. ITTO, RRI and Forest Trends, Washington, DC, p. not available. Wood, C., 2002. Land use and deforestation in the Amazon. In: Wood, C., Porro, R. (Eds.), Deforestation and land use in the Amazon. University of Florida Press, Gainesville, USA, pp. 1 31.

PAGE 146

146 BIOGRAPHICAL SKETCH Shoana Humphries was born in Grass Valley, California, in 1974. She moved around quite a bit during her childhood, and graduated from high school in Anniston, Alabama. Shoana received her Bachelor of Science degree in natural resource conservation from the School of Forest Resources and Conservation at the University of Florida, in 1997. Shoana then worked for three years in forest certification in the Southeastern United States and three years in community development and conservation in Bolivia and Peru. In 2003, she returned to the University of Florida to complete a Master of Science degree in community forest management. The day after she finished her MS degree, she started her PhD program in forest resources and conservation. She is now working with an international conservation organization in Bon n, Germany, and looking forward to exploring Europe. However, her heart is in the tropics, and she continues her work there, as well.