(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Environmental Effects of the Coffee Crisis: a case study of land use and avian communities in Agua Buena, Costa Rica"

Environmental Effects of the Coffee Crisis: 

A Case Study of Land Use and Avian 

Communities in Agua Buena, Costa Rica 



Eve Rickert 

A Thesis 

Submitted in Partial Fulfillment of the Requirements for the 

Masters of Environmental Studies Program 

The Evergreen State College 

December 2005 



This Thesis for the Master of Environmental Studies Degree 

by 
Eve Rickert 

has been approved for 

The Evergreen State College 

by 




& 




Dr. Martha Rosemeyer 
Member of the Faculty 



Dr. Steven G. Herman 
Member of the Faculty 

Dr. Catherine Lindell 

Assistant Professor 

Department of Zoology 

Michigan State University 



I2/If/0S~ 



Date 



Abstract 

Environmental Effects of the Coffee Crisis: 

A Case Study of Land Use and Avian Communities 

in Agua Buena, Costa Rica 

Eve Rickert 

The last 15 years have seen the most dramatic and prolonged collapse in coffee prices in 

the history of the global coffee trade. In response to this "coffee crisis," many coffee 
farmers have been forced to abandon their land or change to other crops, including cattle 
pasture. This study looks at the environmental effects of these decisions at a local scale, 

focusing on the district of Agua Buena in southern Costa Rica. Interviews with 59 

farmers were used to assess the nature and extent of land-use change on coffee farms, and 

96 timed area-search bird surveys were used to assess the effects on avian communities 

of conversion of coffee to pasture. The interviews revealed that, in the study area, 

farmers were changing from coffee to other crops, with 93.3% of respondents listing 

coffee prices as a motivating factor in their decision. Pasture represented, by area, 64.3% 

of the land area converted, with other crops and land abandonment representing 15.3% 

and 14.0%, respectively. The bird surveys showed a reduced diversity at the species 

level, and reduced richness and diversity at the family level. The bird surveys also 

revealed changes in community composition with conversion to pasture - most notably 

the disappearance of four of five detected species of understory insectivores, a nine-fold 

increase in the relative number of detections of sparrows and finches (family 

Emberizidae), and a four-fold increase in the relative number of detections of granivores. 

The study concludes that, due to the scale of conversion of coffee to pasture revealed by 

the interviews, and the diminished habitat quality in pasture as indicated by the bird 

surveys, the coffee crisis has had negative effects on biodiversity at a local scale. The 

conclusions suggest that new price and supply controls are desirable to mitigate both the 

environmental and social effects of the coffee crisis. 



© Copyright by Eve Rickert 2005 



This document is licensed to the public under the Creative Commons Attribution- 
NonCommercial-NoDerivs 2.5 License: 

You are free: 

• To copy, distribute, display, and perform the work 
Under the following conditions: 

• Attribution. You must attribute the work in the manner specified by the 
author or licensor. 

• Noncommercial. You may not use this work for commercial purposes. 

• No Derivative Works. You may not alter, transform, or build upon this 
work. 

For any reuse or distribution, you must make clear to others the license terms of this work. 

Any of these conditions can be waived if you get permission from the copyright holder. 

Your fair use and other rights are in no way affected by the above. 

This is a human-readable summary of the legal code available at: 
http://creativecommons.Org/licenses/by-nc-nd/2.5 



Table of Contents 

ABSTRACT ii 

TABLE OF CONTENTS iv 

LIST OF FIGURES vi 

LIST OF TABLES viii 

ACKNOWLEDGEMENTS x 

I. INTRODUCTION AND BACKGROUND 1 

1. Executive summary 1 

2. Coffee and the Environment 2 

3. Anatomy of a crisis 16 

II. SOCIAL, ECONOMIC AND ENVIRONMENTAL EFFECTS OF THE CRISIS 
59 

1. International 60 

2. Costarica 67 

3. CotoBrus 71 

4. Agua Buena and Individual Households 79 

III. CONCLUSIONS AND ALTERNATIVES 147 

1. Environmental consequences 147 

2. Alternatives and solutions 148 

3. Conclusion 159 

GLOSSARY 161 

APPENDICES 163 

Appendix A. Additional information regarding interviews 163 

Appendix B. Additional information regarding bird surveys 173 

REFERENCES 184 



iv 



List of Figures 



Figure 1.2.1. Types of coffee production systems 6 

Figure 1.3.1. Total world production of coffee with average world export prices 1961-2002. 21 

Figure 1.3.2. Total world coffee trade 1961-2003 21 

Figure 1.3.3. Costa Rican production as a percentage of world production 31 

Figure 1.3.4. Coffee yields in kg/ha for Costa Rica and the world 1961-2003 31 

Figure 1.3.5. Brazil's effect on world coffee supplies 37 

Figure 1.3.6. World export prices for coffee, 45 

Figure 1.3.7. difference in world consumption and production, 1961-2002 45 

Figure 1.3.8. production in Brazil, Vietnam and Indonesia 1992-2004 51 

Figure 2.1.1. Area producing coffee worldwide from 1961-2004 65 

Figure 2.1.2. Area producing coffee by region, 1961-2004 65 

Figure 2.3.1. Gross sales of agrochemicals at a supply store in San Vito, Coto Brus, 1993-2002 

77 

Figure 2.4.1. Estimated land use of sample circa 1997 103 

Figure 2.4.2. Land use of sample in 2003 103 

Figure 2.4.3. Predicted future land use of sample 105 

Figure 2.4.5. Land use change as a percent of total original area 107 

Figure 2.4.6. Age of survey respondents compared to Coto Brus as a whole 109 

Figure 2.4.7. Age of survey respondents compared to Censo Cafetalero 109 

Figure 2.4.8. Educational level of survey respondents compared Censo Cafetalero and 2000 

Census Ill 

Figure 2.4.9. Example of coffee site 127 

Figure 2.4.10. Example of pasture site 127 

Figure 2.4.11. Map of survey sites 131 

Figure 2.4.12. Percent of all registrations by family for pasture and coffee 135 

Figure 2.4.13. Species richness estimates for pooled samples 137 

Figure 2.4.14. Family richness estimates for pooled samples 137 

Figure 2.4.15. Species diversity indexes for samples with all sites pooled 139 

Figure 2.4.16. Family diversity indexes for samples with all sites pooled 139 

Figure 2.4.17. Relative abundance of birds associated with forested vs. openhabhats 143 

Figure 2.4.18. Percent of bird species associated with forested vs. open habitats 143 

Figure 2.4.19. Percent of bird registrations in each feeding guild 145 

Figure 2.4.20. Percent of species in each feeding guild 145 

Figure B.l. Survey design 181 



VI 



List of Tables 



Table 2.4.1. Summary of interview responses on yes/no variables 95 

Table 2.4.2. Summary of interview responses for categorical variables 96 

Table 2.4.3. Summary of interview responses for quantitative variables 99 

Table 2.4.4. Summary of land use characteristics of the sample 101 

Table 2.4.5. species list 131 

Table 2.4.6. Results of ANOVA for diversity indexes for species 141 

Table 2.4.7. Results of ANOVA for diversity indexes for families 141 

Table A. 1 . Interview variables that showed a significant difference between interviewers. 1 69 
Table A.2. Animals for which respondents have perceived a change in abundance since 

ARRIVING IN THE STUDY AREA 170 

Table B.l. Schedule of bird surveys 182 

Table B.2. Similarity indexes for all site pairs, and for plot pairs within sites 183 

Table B.3. Pearson correlation coefficients comparing distance with similarity indexes 184 



Vlll 



Acknowledgements 

This work would not have been possible without the assistance of a great many 
individuals, only a few of whom I will be able to name here. My primary reader, Dr. 
Martha Rosemeyer at Evergreen, has provided constant support and assistance at a level 
that defies any superlatives that could be selected to describe it. I could not have hoped 
for a better advisor, and I feel unbelievably fortunate to have benefited for the past three 
years from her knowledge, wisdom, and compassion. 

My two other readers, Dr. Steve Herman at Evergreen and Dr. Catherine Lindell 
at Michigan State University, made valuable comments and suggestions during the 
project and provided feedback on the draft manuscript. Dr. Ted Whitesell at Evergreen, 
and Jai Ranganathan at Stanford University, also provided input and comments on 
portions of the manuscript. Dr. Gretchen Daily at Stanford University, and John 
Alexander at the Klamath Bird Observatory, assisted with suggestions during the project 
planning. Rip Heminway at Evergreen provided hours of valuable assistance with GIS. 

Dr. Cagan H. §ekercioglu at Stanford University graciously allowed me to 
accompany him on two mist-netting trips to help learn the birds of the study area, 
reviewed two early drafts of my prospectus, and provided suggestions and feedback 
throughout the project. Dr. Karen Holl at the University of California, Santa Cruz, 
reviewed two early prospectuses, provided aerial photos of the region, and made 
suggestions during the research. Both Jim R. Zook and Dionisio Castro took me birding 
in the study area on several occasions. Dr. Bruce Haines at the University of Georgia, 
Dr. Fred Werner at Cornell University, and Dr. Ken Orvis at the University of Tenessee 
all provided valuable comments and suggestions on various aspects of the project. The 
early guidance of Dr. Thomas W. Sherry at Tulane University helped me to begin to 
formulate a research question, and allowed me to acquire the background necessary to 
pursue this work. 

Darryl Cole-Christiansen at Finca Loma Linda provided invaluable logistical 
support by making the arrangements for a homestay, assisting in transportation, and 
allowing the use of his grounds for birding practice and his library for research. Rodolfo 



Quiros at the Wilson Botanical Garden generously allowed me to use the Garden's 
library and provided feedback, contacts and suggestions. The Mendez-Gutierrez family 
allowed me to stay in their home, assisted with the development of interview questions, 
and helped me to locate research sites. Javier Cedeno and Coopabuena allowed me to use 
their computer and telephone, and provided logistical support, suggestions and 
information throughout the project. The Cordero Campos family also allowed me to use 
their telephone frequently. Xinia Mesa and Luis Campbell provided a much-needed 
home base in San Jose. 

I am grateful to Omar Morera Barrantes, Carlos Jimenez Alvarado, Marco 
Salazar, Eladio Mendez, Henry Prendas Fernandez, Roberto Jimenez, the Corporation 
Cafetalera La Meseta, and two additional landowners for allowing me to use their farms 
for my bird surveys. I am also thankful to everyone who responded to questionnaires or 
allowed themselves to be interviewed during the course of the project. The managers of 
Finca Rio Negro and the Hartmann family allowed me to investigate their properties as 
potential research sites during project planning, and provided logistical support while I 
was doing so. This thesis received financial support from the Studebaker Foundation and 
the Evergreen State College Endowment. 

By far the biggest debt of gratitude I owe is to my husband, Luca Pellanda, who 
has allowed both our lives to be turned upside down in the pursuit of the completion of 
this thesis. Words fail in attempting to describe the value of his contributions to this 
work. For the unending and unconditional emotional and financial support, for believing 
in me when I myself couldn't do so, for being Helium to my Strindberg, and for the 
hundreds of small things that have kept me going for the past three years, this thesis is as 
much the fruit of his labor as it is mine. I am also thankful to the rest of my family for 
their emotional and financial support: my mother, Marilou Rickert, who also proofread 
the manuscript and helped "hold down the fort" during my time in Costa Rica, and my 
mother- and father-in-law, Bruna and Renato Pellanda. 

All of these individuals, and many others, have contributed to and improved this 
work, but the full responsibility for any errors it contains is, of course, my own. I have 
almost certainly unwittingly omitted a number of people to whom I owe thanks, and to 
these I offer my apologies. 



XI 



I. Introduction and Background 

/. Introduction and summary 

If you visited a coffee-growing region anywhere in Costa Rica in 2003, you 
would hear and see references to "the crisis." No one there needs it explained to them: in 
this small Central American nation, rapidly diversifying its economy but still heavily 
dependent on agricultural exports such as coffee and bananas, "the crisis" is in the 
newspapers, on the TV, on billboards, and the subject of books, but most importantly, an 
everyday reality for tens of thousands of people. The term "crisis" has been used 
throughout the history of coffee production whenever export prices take an inevitable 
nosedive. In this case, the term refers to the longest and deepest slump in coffee prices in 
the history of Latin American production of this commodity. 

Many of the brutal economic effects of the crisis on Costa Rica's 70,000 coffee 
farmers, who watched the price of their product collapse by 60% or more in just three 
years, are easy to anticipate. On the surface, the story of Costa Rican coffee farmers is 
another iteration of the same narrative, repeated throughout modern history, of all small- 
scale producers of raw materials struggling to survive a volatile world market. The 
peculiar dynamics of coffee production, described in Chapter 1.3, make the current crisis 
more complex and intractable than a typical downturn in the price of other types of 
commodities. 

Also not clear are the effects of the crisis on the Costa Rican environment, as 
farmers are forced to change to different products, or to leave the land entirely. It is these 
environmental questions, placed in the context of the economic reality of rural Costa 
Rica, which this study attempts to answer. In search of those answers, it will examine the 
effects of the coffee crisis, and responses to those effects, across scales: from individual 
households and small communities, to the national and international levels. 

This thesis first provides an overview of the world coffee market and the 
sequence of events leading up to and following the collapse of the International Coffee 
Agreement. Next, it looks at the effects of the price crisis, discussing social and 
economic effects where they interface with environmental issues, in order to understand 



the environmental effects of the crisis and their context. After a brief glimpse of these 
issues at an international scale, the study focuses in on the Central American country of 
Costa Rica, then on the canton of Coto Bras, a remote coffee-growing region of that 
country. 

This thesis examines in detail the effects of the crisis in the district of Agua 
Buena, a region of Coto Bras disproportionately affected by the crisis. At this "micro-" 
level, we are able to see a detailed picture of individual responses to the crisis, which then 
shed light on its large-scale effects. Two specific hypotheses are tested. The first 
hypothesis predicts that coffee farmers in Agua Buena are responding to the crisis by 
removing their coffee in favor of cattle pasture. The second predicts that the conversion 
from coffee to pasture will have a negative effect on bird communities that use the 
agricultural matrix in the region. The study concludes with a discussion of possible 
solutions to the crisis. 

2. Coffee and the Environment 

Agricultural areas take up a significant portion of the earth's surface, and that 
proportion is going to continue to increase as we address the need to provide food and 
goods to a world population that is growing in both size and consumptive ability. It is 
widely recognized that pristine areas, including protected areas such as reserves and 
national parks, will not be sufficient to preserve biodiversity in the future (Pimentel & 
Stachow 1992; Vandermeer & Perfecto 1997; Hughes et al. 2002). In order to ensure the 
greatest possible protection for wild species, attention needs to be given to the potential 
benefits and limitations of agroecosystems as wildlife habitat (Vandermeer & Perfecto 
1997; Hughes et al. 2002; Jules & Shahani 2003). In many tropical areas, permanent 
agroforestry systems such as coffee offer a more stable, and environmentally preferable, 
alternative to such common practices as slash-and-burn farming (Griffith 2000), and 
environmentalists have recently given a great deal of attention to coffee as an 
environmentally sustainable crop that, under the right circumstances, can provide 
important habitat for wildlife. This chapter will review the literature regarding the effects 
of coffee production on deforestation, biodiversity, and habitat connectivity. 



In the case of coffee, a great deal of biodiversity research has compared "shaded" 
Arabica farms in the Neotropics to primary forest and unshaded coffee monocultures 
(Greenberg et al. 1996; Perfecto et al. 1996; Greenberg et al. 1997; Johnson 2000; Sherry 
2000). Coffee evolved as an understory plant, and shade trees have been used for 
centuries in coffee production to control yields, coffee quality, the productive lifetime of 
individual coffee trees, protection from climatic extremes, and pest outbreaks (Beer 1987; 
Caramori et al. 1996; Beer et al. 1998; Muschler 1998; Soto-Pinto et al. 2000; Muschler 
2001). 

Unfortunately, while there is a niche market for "certified" shade-grown coffee, 
the certification process is expensive (Gobbi 2000) and the market for these products is 
small (Puelschen & Lutzyer 1994; Messer et al. 2000). Further, farmers may not actually 
receive a premium for their certified coffee if there is not also a certified processing plant 
available to take their product. Those plants that are certified can control the prices paid 
to farmers, and use regional monopolies to hold prices down (Lyngbaek et al. 1999). 
Additionally, in optimal growing zones, any shade at all may compromise yields, 
complicating the economic equation for farmers considering shade-grown coffee (Beer et 
al. 1998). 

Because of these factors, and of powerful pressures from government policies and 
funding agencies, much of the world's coffee is now produced on "technified" farms 
involving few shade trees and heavy agrochemical inputs (Rice 1999; Perfecto et al. 
1996; Borrero & Ignacio 1986). In Latin America, sun coffee farms make up about 40% 
of coffee area (Rice & Ward 1997; cited in Rappole et al. 2003a). Furthermore, only 
Arabica coffee is grown in shaded systems. Shade is completely inappropriate for 
Robusta (O'Brien & Kinnaird 2004), the coffee variety grown in lowland regions, 
including most coffee-growing areas of Asia and Africa. Much of the concern over 
deforestation arising from coffee production is due to the expansion of Robusta 
production. 

While the information on biodiversity in shade coffee is valuable, therefore, there 
is little information on other widespread methods of coffee cultivation in Latin American 
countries. The lack of information on non-shaded systems makes comparison of coffee 



farming in general (as oppsed to shade coffee farming) with other types of agricultural 
land use, such as pasture or vegetable crops, difficult. Because, however, of the large 
number of coffee farms in Latin America that are still planted with some level of shade, 
and because the research on shaded coffee makes up such a large portion of the 
ecological literature on coffee, this chapter will review what is known about the 
environmental costs and benefits of shaded coffee farms in comparison to other forms of 
coffee production. 

Because the purpose of this thesis, however, is to address the environmental 
consequences of the coffee crisis at multiple scales, and because the local area that serves 
as a case study for this thesis is dominated by coffee farms with sparse, low-diversity 
shade trees, this chapter also reviews the somewhat less abundant literature dealing with 
the impacts of typical coffee production and, where possible, makes comparisons 
between coffee and other agroecosystems suitable to the same conditions as coffee. 
While there are many areas of coffee's environmental impact that deserve discussion, 
such as agrochemical use, water and air pollution from coffee processing, and carbon 
storage, the focus of this work is biodiversity and associated variables, such as 
deforestation and habitat connectivity. 

Moguel and Toledo (1999) describe five categories of coffee production systems 
which, although broad, are useful as an aid in discussion of the ecological characteristics 
of different forms of coffee farming (Figure 1.2.1). "Rustic" plantations are primary 
forest where the understory has been removed and replaced with coffee plants. In a 
"traditional polyculture," some of the forest tree species have been replaced by other 
species, either of commercial value or of use to the farmer. A "commercial polyculture" 
has had the overstory removed, as well, and replaced it with a mixture of commercially 
useful trees - typically Musa spp., Citrus spp., and nitrogen fixing trees such as Erythrina 
spp. and Inga spp., and less often trees used for lumber. A "shaded monoculture" is a 
farm where the shade consists only of a single species, usually Erythrina spp. or Inga 
spp. Finally, "unshaded monoculture" is coffee grown in full sun, with no interplanted 
trees or crops. 



Where appropriate, this terminology will be used in discussions of different coffee 
production systems. The term "sun coffee" will be used interchangeably with "unshaded 
monoculture." Where the phrase "shade coffee" is used alone, it will generally mean 
either that additional information was not available about the level or type of shade, or 
that the system did not fit into one of Moguel and Toledo's four categories of shade 
coffee. 



f* * 



5 
I 



JS 
20 
13 

1U 
1 

if.- 
JO 

I'll 
in 




RinOc 





. 1 -.jfJ'-U.-UJOF t- ,i--.IvC6 



Traditional 
potyculture 
offaflflBrdan") 



13 

i a 



^ ^ 



«£ 



^W ' rt»$ 




ptlytgllure 



It 



1D 
5 




'"■jj> r -' ^ 



I UT TUl r t t j Tt'r-^r 1 , . 



Shitted 
monoculture 



Unshaded 
mnnccullurf 



Figure 1.2.1. Types of coffee production systems. Figure from Moguel and Toledo (1999). 



Deforestation 

In many regions, particularly where land is cheap and plentiful, deforestation is 
still a significant result of the expansion of coffee production. In addition to direct 
conversion of forests to coffee farms, firewood must be supplied to the driers at coffee 
processing plants, putting further strain on local forest resources (Pelupessy 1997). Even 
clearing the forest understory to plant shade-grown coffee can have important negative 
impacts on forest-dependent species (Rappole et al. 2003a), and some believe that 
promotion of shade coffee by conservation, aid and development organizations, and a 
lack of sufficient oversight of shade certification criteria, has created perverse financial 
incentives for farmers to do exactly that (Rappole et al. 2003b). 

In the Ivory Coast, about one million hectares were deforested for coffee 
production between 1960 and 1997 (Sendieu 1996 p. 13; cited in Pelupessy 1997 p. 28; 
ICO 2005e). Rappole et. al. (2003a) consider conversion of endangered Central 
American pine-oak forest to coffee farms to be an important threat to the endangered 
Golden-cheeked Warbler (Dendroica chrysoparia), a pine-oak obligate. O'Brien (2003) 
found a significant positive correlation between the price of Robusta coffee and the rate 
of deforestation in Indonesia's Bukit Barisan Selatan National Park, one of the last 
remaining habitats for the crtically endangered Sumatran tiger. 

According to Pelupessy (1997 p. 30), coffee production in Costa Rica has 
historically not been directly responsible for any significant amount of deforestation. For 
some this may be difficult to believe, given what is known about the role of coffee in the 
expansion of Costa Rica's agricultural frontier. The area planted in coffee in Costa Rica 
between those years of 1961 and 2000 for which data are available is highly negatively 
correlated with forest cover (r= -0.564, N=36, p=0.000) (data from UNFAO 2005; 1999; 
2001). Coffee area expanded during this time period by only 36,000 hectares, while 
forest cover was reduced by nearly 1.7 million hectares. Thus, expansion of coffee 
production accounted for, at most, two to three percent of deforestation in Costa Rica 
during those years. The agricultural area in Costa Rica increased by nearly 1.5 million 
hectares, representing 88% of forest loss, but coffee actually decreased as a percentage of 
agricultural area during this time period - from 5.2 to 3.9% (all data from UNFAO 2005). 



7 



The data, therefore, support Pelupessy on this point, at least on a national scale. 
While the expansion of coffee production is correlated with forest cover loss, this is due 
to the expansion of agriculture in Costa Rica overall in the latter half of the 20 th century, 
and other land use types, such as pasture (Hartshorn et al. 1982), make up the majority, 
by area, of uses replacing forest. Coffee has played an important role, however, at a local 
scale in some areas. For example, the deforestation of the southern canton or county of 
Coto Bras over the last 50 years has largely been the result of the expansion of coffee 
area, with the encouragement of government settlement programs (Manger 1992). 

Donald (2004) argues, without citing sources, that coffee expansion is directly 
responsible for "much" deforestation in Vietnam, where coffee expansion has exploded 
since the early 1990's. In fact, the data do not support this assertion. Data from the UN 
Food and Agriculture Organization show that between 1990 and 2000, the years of the 
most significant coffee expansion in Vietnam, forest cover in Vietnam actually increased 
by 52,000 hectares (0.5%) (UNFAO 2001 p. 155), while the area dedicated to coffee 
growing increased by 415,000 hectares, or 671%. 

During this time period, coffee area in Vietnam was not significantly correlated 
with forest area (r=0.361, N=8, p=0.380) (data from UNFAO 2005; 1999; 2001). It 
appears likely, therefore, that the expanded coffee area in Vietnam was converted from 
other agricultural uses rather than being carved out of forests, or was planted in areas 
defoliated during the Vietnam War (Wild 2004 p. 6). The expansion of forest area may 
be due to abandonment of other types of crops representing an area larger than the new 
area planted in coffee. More research on this point would be necessary before valid 
assertions could be made about the relationship of coffee to forest area in Vietnam. 

There is probably significant local variation, however, with coffee expansion 
clearly accounting for important forest losses in certain regions, such as Coto Brus in 
Costa Rica. Looking at forest losses on a national scale also disguises the effects of 
coffee expansion on specific types of ecosystems. For example, in a study of LANDSAT 
data for Costa Rica between 1986 and 1991, Sanchez-Azofeifa (2001) found that the 
premontane moist forest life zone, which provides the most optimal conditions for coffee 
production (Holdridge 1967 p. 16; Boucher 1983), was the most severely depleted of the 



life zones analyzed, with only 2% of primary forest remaining in this life zone. The same 
study found, however, that in the premontane rain forest life zone, which characterizes 
much of the local study area for this thesis (Hartshorn et al. 1982 p. 1 1) as well as many 
other coffee-growing areas in Costa Rica, 60% of forest cover remained. Roberts et al. 
(2000a) estimate that coffee plantations make up as much as 54% of the area that has 
replaced premontane and cloud forests in Mexico, Central America and Colombia. 

Coffee farms also affect the structure of secondary forests that replace them when 
the farms are abandoned. Marcano-Vega et al. (2002) studied forest regeneration in 
abandoned coffee farms and pastures in Puerto Rico, and found that the dominance of 
coffee in the landscape had significant effects on the species composition of secondary 
forests that regenerated in both coffee and pasture, with trees commonly used for shade in 
coffee plantations dominating the secondary forests. Species composition changed more 
rapidly in regenerating pastures than in regenerating coffee farms. This appeared to be a 
result of to the dominance of coffee plants and associated shade trees in the abandoned 
coffee farms, which competed with other secondary forest trees. In Puerto Rico, 
Thompson et al. (2002) were able to detect differences in the species composition of 
forest trees in regenerated forests that had previously been used for either selective 
logging or coffee production, some 40 to 60 years after abandonment. 

Biodiversity 

Agroforestry systems, such as coffee, may serve as refuges for wildlife in a 
landscape that is otherwise fragmented or degraded (Borrero & Ignacio 1986; Moguel & 
Toledo 1999; Perfecto et al. 1996; Perfecto et al. 1997; Komar 2000; Perfecto & 
Vandermeer 2002). For example, Puerto Rico has lost an estimated 99% of its forest 
cover in the 500 years since European colonization. Due in part, however, to canopy 
cover provided by shade coffee plantations, less than 12% of the native bird diversity was 
lost during this time period (Lugo 1988). In Costa Rica and Panama, the AMISCONDE 
Initiative of Conservation International and the Tropical Science Center has worked to 
encourage coffee production - primarily commercial polyculture or shaded monoculture 
- in the buffer communities surrounding the La Amistad Biosphere Reserve, with the 
belief that this will assist conservation efforts in the reserve (Young 2003). 



Clearly, the level of structural diversity and agro-biodiversity in different 
production systems for the same product will have an effect on the capacity of those 
production systems to support higher levels of associated (i.e., non-agricultural) 
biodiversity. In Central America, the diversity of vertebrates has consistently been found 
to be higher in shaded coffee than in sun coffee (Moguel & Toledo 1999; Perfecto et al. 
1996; Borrero & Ignacio 1986), with birds being the most frequently studied group. In 
Sulawesi, Indonesia, Siebert (2002) found bird diversity to be higher in rustic coffee than 
in shaded coffee monoculture. 

Studies in El Salvador found a steady increase in bird diversity as canopy cover 
increases, and a high level of correlation between tree species diversity and bird diversity 
(Komar 2000; 2001). This last finding is supported by research in other regions (Perfecto 
et al. 1996). Factors such as tree basal area, emergent trees, canopy height, and the 
abundance of flowering or fruiting trees were less important than canopy cover and tree 
species diversity, at least for birds (Komar 2000; 2001). 

Lindell et al. (2004) worked in Coto Bras in southern Costa Rica, where coffee is 
grown in a commercial polyculture where the "shade" consists primarily of Musa spp. 
and small (<7m) leguminous trees. They found a higher proportion of birds restricted to 
forest habitats (29.9% vs. 16.4%) than was found by Estrada et al. (Estrada et al. 1997; 
cited in Lindell et al. 2004) in coffee farms in Los Tuxtlas, Mexico, where coffee farms 
had shade cover consisting of forest trees over 15 m tall. Several species that Lindell et 
al. found restricted to forest were detected by Greenberg et al. (1997; cited in Lindell et 
al. 2004) in coffee plantations in Chiapas, Mexico that had a greater number of trees 
between 8-10 m tall than the Costa Rican sites. 

A study in Asia indicated that the bird species most impacted by conversion from 
primary forest to agroforestry systems, including shade coffee, were large insectivores 
(Thiollay 1995). Another study, in Jamaica, indicated that it was insectivores and 
nectarivores that showed the most activity in shade coffee plantations (Johnson 2000). 
From these apparently conflicting results, it would appear that the relationship between 
various types of coffee production and insectivore populations, and the ability for coffee 



10 



farms of various types to act as corridors for sensitive forest species, varies between 
localities, and needs to be further examined. 

Petit and Petit (2003) compared mid-elevation shaded coffee with pasture, fallow 
pasture, Caribbean pine plantations, residential areas, sugar cane, rice, gallery forest, 
lowland forest fragments, gallery forest, extensive lowland forest and extensive 
premontane forest in central Panama, with respect to their value as habitat to birds that 
are vulnerable to disturbance. They found that pasture and all crops except coffee were 
devoid of forest specialists, while coffee supported primarily forest generalist and species 
associated with open woods and scrubland. In terms of importance in the conservation of 
the most vulnerable species, they found that coffee was the most important of all non- 
forest habitats, with an "importance score" double that of the next-ranked habitat, gallery 
forest. The coffee in this study appears to have been traditional polyculture. 

In the Colombian Andes, Numa et al. (2005) found that bat species richness was 
not significantly different between shaded coffee farms and a commercial polyculture 
similar to that studied by Lindell et al. (2004). Both types of coffee had significantly 
lower species richness than nearby forest fragments, but the species composition in the 
shaded coffee was more similar to the forest than was the commercial polyculture, while 
the commercial polyculture showed higher ft diversity than the shade coffee. In 
Veracruz, Mexico, Pineda et al. (2004) found that bat species richness and composition 
was nearly the same in cloud forest and traditional polyculture coffee plantations. 

Arthropod diversity has also consistently been found to be higher in shade 
plantations than in technified plantations (Nestel et al. 1993; Perfecto & Vandermeer 
1996; Perfecto et al. 1997; Moguel & Toledo 1999; Roberts et al. 2000b; Rojas et al. 
2001; Perfecto & Vandermeer 2002). In Chiapas, Mexico, Perfecto et al. (2003) found 
that ant and butterfly diversity decreased along a gradient from forest to increasingly 
intensified coffee farms, with butterflies appearing to be more sensitive to the changes. 
In Sulawesi, Indonesia, Klein (2004) found that Chalcid wasps did not nest in traps 
outside of forest or rustic coffee farms, while Braconid wasps were found more often in 
intensively managed systems, and other families showed no differences across habitat 



11 



types. Species richness of social bees decreased with land use intensity, while species 
richness of solitary bees increased. 

Diversity of Coleoptera (beetles) has been found to increase with tree species 
diversity, up to at least 10 species of trees per hectare (Nestel et al. 1993). Research in 
Veracruz, Mexico found no significant differences in the diversity of Scarabid and 
Silphid beetles between disturbed tropical montane cloud forest and coffee with either 
monospecific or polyspecific shade cover (Arellano et al. 2005). In Veracruz, Mexico, 
Pineda et al. (2004) found a higher species richness of corpronecrophagous beetles in 
traditional polyculture coffee than in tropical montane cloud forest. 

Bandeira et al. (2005) studied tree species diversity on rustic coffee plantations in 
southern Mexico. They concluded that, while individual farms did not contain enough 
forest tree species diversity to contribute significantly to biodiversity protection, the high 
variability between farms, and the resulting ft diversity, made this production system an 
important component of tree diversity conservation in an agricultural landscape. This 
study found 27 wild species of trees and 12 cultivated species, which is substantially 
more species than are found in many other types of agricultural systems. Siebert (2002) 
compared diversity of trees, epiphytes and lianas in rustic and shaded monoculture 
coffee, and not surprisingly found diversity of all three to be higher in the rustic coffee, 
but unfortunately did not compare either to forest. 

Of course, despite shade coffee's benefits for biodiversity in comparison to sun 
coffee, even rustic coffee compares poorly to native forest. Rappole et al. (2003a) cite 
numerous findings of reduced biodiversity in shade coffee compared to forest, and point 
out that discussions of "biodiversity" can be deceptive when other variables are not 
considered, i.e., "[a]n agricultural site in which 30 forest bird species have been replaced 
by 30 open-country bird species can have the same avian diversity as a forest site." 
Forest-dependent specialist species are the most likely to disappear after conversion of 
forest to any type of coffee production. 

For example, in a traditional polyculture in Chiapas, ants showed decreased 
diversity in coffee as compared to primary forest, and diversity decreased further with 
distance from the forest. A shaded monoculture was, of course, less diverse still 



12 



(Armbrecht & Perfecto 2003). A study of army ants in Panama found that there was no 
significant difference in army ant populations between primary forest and shade coffee 
adjacent to primary forest (Roberts et al. 2000b), where the type of shade was remnant 
primary forest trees (pers. obs. 2003), but populations declined significantly with distance 
from forest. Army ants are considered top predators of the arthropod world, and are 
sensitive to habitat destruction (Lovejoy et al. 1986; Harper 1989; Roberts et al. 2000b). 

Lindell and Smith (2003) gathered information regarding bird nesting activity in 
pasture, abandoned sun coffee, and forest in the canton of Coto Brus in southern Costa 
Rica, near the local area studied in this thesis. They found that the family distributions of 
birds nesting in pasture and abandoned coffee were significantly different from the family 
distributions of birds found nesting in forest, but were not significantly different between 
pasture and abandoned coffee. Overall, they found that coffee and pasture were 
"functionally quite similar" for nesting birds. They also found, however, two species of 
forest-dependent birds, the Chestnut-capped Brush-finch (Buarremon brunneinucha), and 
the White-breasted Wood-wren, {Henicorhina leucostica), nesting in coffee but not in 
pasture. 

Certain groups of birds have been found to be highly sensitive to the conversion 
of forest to agricultural land uses, including coffee. During point counts in Coto Brus, 
Lindell et al. (2004) detected only one trogon (Trogonidae), and two antbirds 
(Formicariidae) in agricultural areas, compared to 80 and 543 detections, respectively, in 
forest areas. Lindell and Smith (2003) also found that, in the same region, antbirds were 
very unlikely to nest in coffee, even when the farms were adjacent to extensive tracts of 
primary forest. 

In Sumatra, Gillison et al. (2004) found that plant species diversity increased 
steadily from primary and late secondary forests through increasingly more structurally 
complex coffee farming systems. Hietz (2005) found reduced species richness of 
vascular epiphytes in commercial coffee polycultures compared to forests. Although 
species richness in traditional polycultures was similar to that found in forests, 
community heterogeneity was greater in forest. In Veracruz, Mexico, Pineda et al. 
(2005) found a reduction of 20% in species richness of frogs from tropical montane cloud 



13 



forest to traditional polyculture coffee, and only half of the species found in forest were 
also found in coffee. 

Habitat connectivity 

Patches of forest habitat in the tropics have become increasingly isolated over the 
past 20 years, and it is important for conservation goals to consider the matrix 
surrounding forest fragments (Defries et al. 2005; Jules & Shahani 2003). One ecological 
function that coffee farms may provide is to act as corridors between forest fragments 
(Perfecto & Vandermeer 1996; Roberts et al. 2000b; Sherry 2000). Additionally, coffee 
may be incorporated at a low density into a crop rotation system involving various stages 
of successional forest with its associated biodiversity, where coffee is used as a means to 
make these ecologically important rotations more economically productive (Darryl Cole- 
Christiansen, pers. comm. 2003) 

In the Lampung province of Sumatra, Indonesia, most of the nearly 20% of the 
land area that is protected are in small fragments. A significant part of the agricultural 
matrix between these fragments is made up of agroforestry systems, dominated by rubber 
but also including coffee. These lands appear to provide corridors for endangered 
Sumatran tigers, but because of the fact that they are used by both humans and tigers, 
they also serve as "hot spots" for encounters between the two species (Nyhus & Tilson 
2004). Woodroffe and Ginsburg (1998; cited in Nyhus & Tilson 2004) found that 
conflict with humans in such areas on the edges of reserves is a significant cause of 
mortality for large carnivores. Thus, in the case of animals that can be dangerous to 
humans, the very services that agroforestry systems provide in terms of habitat 
connectivity may also increase their risk of mortality from conflict with people. 

For birds on a global scale, the factors most affecting the risk of extinction due to 
habitat destruction have been shown to be small body size and habitat specialization 
(Owens & Bennett 2000). Insectivorous birds, particularly understory insectivores, are 
one of the guilds most affected by habitat fragmentation in the Neotropics (Stouffer & 
Bierregaard 1995; Stratford & Stouffer 1999; Sekercioglu et al. 2002). Studies of the 
responses of these birds to forest fragmentation indicate that food availability in 
fragments may not be, as previously believed, a limiting factor for population survival 



14 



(Burke & Nol 1998; Zanette et al. 2000; Sekercioglu et al. 2002). Rather, it appears that 
use of forest fragments by insectivorous birds is limited by a species' ability to disperse 
across the intervening, human-modified landscape (Sekercioglu 2002; Sekercioglu et al. 
2002). In a study of seven species of understory insectivorous birds in the Brazilian 
Amazon, Antongiovanni and Metzger (2005) found that the matrix habitat was important 
in determining whether sensitive species would be found in forest fragments. Thus, the 
matrix habitat surrounding and connecting forest fragments appears to be important in 
facilitating the continued survival of understory insectivores in fragmented landscapes. 

Birds that follow army ant swarms (many of which are understory insectivores) 
are among the most sensitive to habitat destruction and fragmentation, and are unwilling 
to cross open areas (Stouffer & Bierregaard 1995; Holmes 1996; Mason 1996; Stouffer 
1998; Renjifo 1999). Roberts et. al. (2000a) found army ants and ant-following birds on 
shaded coffee plantations adjacent to primary forest in Panama, though not on plantations 
distant from primary forest, indicating that shaded coffee farms, while in themselves not 
suitable habitat for army ants and ant followers, may provide connectivity between forest 
patches for these groups. 

Numa et al. (2005) studied elements of habitat connectivity provided by different 
types of Colombian coffee farms in which small forest remnants were embedded. Their 
results indicated that both shaded and commercial polyculture coffee provided habitat 
connectivity for bats in the region, but the shaded coffee appeared to be of higher 
importance in this respect. 

Summary 

Expansion of area for coffee production may a factor in local deforestation rates. 
Although deforestation for coffee production probably does not account for a significant 
portion of total deforestation, the suitability of coffee for production in biodiversity "hot 
spots" means that it may be a locally important factor in threatening certain imperiled 
ecosystems. Coffee farms, even those classifiable as "rustic," are species-poor in 
comparison to native forest, and many sensitive species are not found in coffee. 



15 



In comparison to other types of agricultural land use, however, coffee appears to 
have important value in providing wildlife habitat in countryside landscapes, and in 
providing some measure of connectivity between forest reserves. This is true even, in 
certain cases, of shaded monoculture coffee. Thus, from a conservation standpoint, 
coffee farms are a desirable component of tropical agricultural landscapes, and their 
conversion to other uses is likely to be cause for concern to those interested in tropical 
biodiversity conservation. 

3. Anatomy of a crisis 

Dynamics of the international coffee market 

The beverage known as coffee is an infusion of the dried, roasted and ground 
seeds of plants of the genus Coffea, in the family Rubiaceae. Nearly all the world's 
coffee is produced by the two species C. arabica and C. canephora, which produce 
Arabica and Robusta coffees, respectively. Coffee grows best in tropical or subtropical 
regions, between 800 and 1500m above sea level, where temperatures remain between 
23° and 28°C for most of the year and annual rainfall is between 1,500 to 2,000mm 
(Boucher 1983; Coste 2001). 

Coffee trees will begin to bear in about 3 years (Boucher 1983), take 6-8 years to 
reach maximum production (Pelupessy 1993 p. 25), and can remain productive for up to 
40 years (Aguilar 1992 p. 10). Thus, a mature coffee plantation represents a substantial 
financial investment that is not easily abandoned or exchanged for another product 
(Kutschbach 1994a p. 16). Coffee is often grown in steep areas, where there are few 
alternative uses for the land (Kutschbach 1994b p. 84). These constraints mean that 
supply and demand are rarely, if ever, in equilibrium (Aguilar 1992 p. 10), resulting in 
dramatic fluctuations in export prices (see Figure 1.3.1). 

The international coffee chain involves, in most cases, 6 actors. The farmer sells 
the fresh cherries to a processor, who removes the pulp and a thin membrane (known as 
the parchment) surrounding the bean, dries the beans and sells them to an exporter. The 
beans are next sold to a roaster, then to a retailer, and finally to the consumer. Because 



16 



roasted and ground coffee rapidly loses quality, those steps in the chain that add value to 
the raw material typically take place, by necessity, in the consuming country. 

Four primary types of coffee are recognized, and these, from lowest to highest 
quality, are: Robustas, unwashed Arabicas, and the two "washed" Arabicas, classified as 
Other Milds and Colombian Milds. Whether a coffee is classified as "washed" or 
"unwashed" is determined by the method used to process the whole cherries into green 
coffee (Aguilar 1992 p. 10), with the method used to produce "washed" Arabicas referred 
to as the "wet method." Within the category of the other milds, the coffees classified as 
"strictly hard bean," grown above 1200 m, are considered the highest quality (Varangis et 
al. 2003; cited in Rappole et al. 2003b). Historically, the price difference between 
Robusta and Arabica has been around 10%, rising as high as 30% (VNKT 1987; cited in 
Pelupessy & Tilburg 1994 p. 244), but currently the price of Arabica is more than double 
that for Robusta (Osorio 2005a). 

For coffee processed by the "wet" method, the first steps in the processing of the 
cherries tend to be heavily centralized, as they must be performed a very short time after 
harvesting (Pelupessy 1997). This severely limits the options available to farmers, who 
must rely on processing plants or collecting stations very close to their farms. In more 
remote regions, such as in Guatemala and the Mexican state of Chiapas, farmers sell their 
coffee to middlemen called coyotes, usually at steeply reduced prices, who transport the 
cherries to the processor, known in Latin America as a beneficio. The coyotes often 
provide credit to the farmers at high interest rates, to be paid off with the money from 
their harvests, keeping the farmers in an unending circle of indebtedness (Dicum & 
Luttinger 1999). 

Other Central American farmers take their coffee directly to either one of about 
400 beneficios in the region (Pelupessy & Tilburg 1994 p. 251), or, as in Costa Rica, to 
one of thousands of collecting stations, called recibidores, located within walking 
distance of the farms and operated by the beneficios. Regulations governing the prices 
paid by beneficios to farmers vary widely from country to country, with beneficios in 
some countries required to pass on a specified percentage of the export price (Pelupessy 
1997 p. 26), and others not subject to any regulation at all. In the early 1990's, 



17 



approximately 25% of Central America's 200,000 farmers, owning 15% of coffee land 
and producing 14% of Central America's coffee, belonged to farmer-owned processing 
cooperatives (Pelupessy & Tilburg 1994 p. 252). 

In the early 1990's, least 20 million families worldwide relied on the production 
of coffee for survival (Pelupessy 1993 p. 33). In certain years in the late 1980's, coffee 
was second only to oil in its share of the global commodity market (Portillo 1993), but 
that share has declined substantially since then (calculated from UNFAO 2005). During 
the decade ending in 2002, green coffee accounted for an annual average of $9.4 billion 
in international trade (UNFAO 2005). Approximately 90% of coffee is imported as 
dried, unroasted beans (Pelupessy 1993 p. 28), referred to in the industry as "green 
coffee" or, in Spanish, cafe oro or "golden coffee." In 2003, 5.4 million metric tons of 
green coffee were traded worldwide, valued at approximately $7.6 billion (ICO 2005b). 

Figure 1.3.2 shows the international trade in green coffee from 1961 to 2003. 
World exports by weight have nearly doubled during this 32-year period. Over the last 
100 years, world consumption has grown at a more or less steady rate of 1.8% per year, 
interrupted only by the two world wars (Daviron 1994 p. 38). Between 1961 and 2001, 
growth in coffee consumption slightly surpassed global population growth, averaging 
1.9% per year, while the population grew at 1.7% per year (UNFAO 2005). In the early 
1990's, roughly 20% of coffee produced was consumed in the producing countries 
(Pelupessy 1993 p. 21). 

Between 60-90% of the value of a pound of coffee is added in the consuming 
country (Pelupessy 1997 p. 9). This includes roasting, grinding, packaging and 
marketing. Stiff tariffs on value-added coffee products in producing nations pose a major 
barrier, hindering producing countries from expanding into these areas. Globally, coffee 
roasting is also dominated by a very small number of players. Five roasters - Kraft, 
Nestle, Procter & Gamble, Sara Lee, and the German roaster Tchibo - together control 
half of the world's coffee beans (Gresser & Tickell 2002 p. 2). Buyers for large roasters 
such as these often receive bonuses for purchasing cheap coffee (Dennis Macray, pers. 
comm. 2004). With both exports and sales of roasted coffee controlled by these 



Manager, Business Practices, Corporate Social Responsibility, Starbucks Coffee Company 
18 



international near-monopolies, a country such as Costa Rica, which produces only a few 
percent of total world supply, is in a comparatively weak bargaining position. 

Each of the four types of coffee (Robusta and the three types of Arabicas) 
represents about a quarter of the world supply (Pelupessy 1997 p. 8). In 2002, Arabica 
made up 68% of the world coffee market, with Other Milds making up 35% of the world 
trade in Arabica coffee (Fernandez et al. 2002; UNFAO 2005). Typically, each country 
only grows one type of coffee (Aguilar 1992 p. 24), although fourteen countries produce 
both (ICO 2005b). Although prices can be more than double for Arabicas (Osorio 
2005a), lower production costs mean that Robustas can give higher profits (Pelupessy 
1993 p. 20). Costa Rica is one of 22 countries, 16 of which are in Latin America and the 
Caribbean, that produce washed Other Milds (ICO 2005b), and in 2002 produced 10% of 
world supply of this type of coffee (Fernandez et al. 2002). In the early 1990's, Central 
American countries together accounted for 25% of production of Other Milds (Pelupessy 
&Tilburg 1994 p. 244). 

Arabica coffee is traded on the Coffee, Sugar and Cocoa exchange in New York 
City, and Robusta on the London Coffee Terminal Market. Speculation in the futures 
markets on both these exchanges tends to increase price volatility, and producing 
countries themselves have been known to buy and sell futures to influence prices 
(Pelupessy 1993 p. 33; 1997 p. 9). In the early 1990's, approximately 50 coffee trading 
firms controlled the world markets, with the largest eight controlling 50% (Pelupessy 

1993 p. 42). 

The history of coffee growing can be divided into two distinct periods (Daviron 

1994 p. 42). During the earlier period, countries increased production by expanding the 
amount of land used to grow coffee. In the 1960's, due to the dwindling availability of 
cheap land, the strategy began to change. Producers sought to increase production using 
technological advancements such as new, high-yielding coffee varieties, reduced shade 
cover, higher planting densities, and increased agrochemical inputs. This process is 
referred to as "technification" or "modernization" of holdings. Costa Rica and Kenya 
were the first countries to embrace this new model, followed quickly by others (Daviron 
1994 p. 45). In essence, the inputs required for coffee production - as with any 



19 



agricultural commodity - can be broken down into three categories: land, labor, and 
technology. The relative availability and cost of each input in a given region will 
determine the type of system used there (Kutschbach 1994b p. 87). 

Technified systems, while increasing income during good years, also make 
farmers more vulnerable to market risks (Kutschbach 1994a p. 19). Over the long term, 
producers with the lowest costs of production will always have the highest profits 
(Daviron 1994 p. 31), while medium-intensity farming appears to have the most 
flexibility in responding to price swings (Kutschbach 1994c p. 233), and overall financial 
risk is lower on more diversified, less technologically intensive coffee farms (Reeves & 
Lilieholm 1993; Gobbi 2000; Ramirez & Sosa 2000). Banks in some regions, however, 
may tie access to credit to technology use (Perfecto et al. 1996; Rice 1999), putting more 
traditional farmers at a competitive disadvantage. 



20 



Total Production and Export Prices 1961-2002 



c 
o 



c 
o 

3 
T3 
O 



9.00 



8.00 



7.00 



-.00 



5.00 



4.00 



3.00 




$6.00 



$5.00 



$1.00 



$0.00 



Figure 1.3.1. Total world production of coffee (all types) plotted with average world export prices 1961- 
2002. Export prices are in constant 2003 dollars. Data from UNFAO (2005). Inflation adjustments for all 
figures are calculated from U.S. Department of Labor (2004). 



World Exports 1961-2003 (millions of Mt) 



6.00 
5.50 
5.00 
4.50 




2.00 



T-tiii)NO)T-nmNO)T-n 

(00(000030)0)0)0)0)00 

oooooooooooooooooooooo 

1-1-T-T-T-T-1-1-1-CMCM 



i- n m n o) i- 

CO CO CO CO CO h- 



00 LD 

1^ h- 



N O) 



Figure 1.3.2. Total world coffee trade in millions of metric tons, 1961-2003. Data from ICO (2005b) and 
UNFAO (2005). 



21 



Coffee in Costa Rica 

In 2003, Costa Rica was a country of just over 4 million people spread over 
51,100 square kilometers (World Bank Group 2005b). Until the mid-1980s, Costa Rica 
had the highest per capita GDP of any Central American nation, only recently losing that 
place to Panama (Karnes 2001). Costa Rica scores well in most development indicators 
with, in 2003, a literacy rate of 96%, 90% of primary-aged children enrolled in school, 
89% of children immunized against measles, an infant mortality rate of eight children 
per thousand, life expectancy of 79 years, and population growth rate of 1.6%. Ninety- 
seven percent of Costa Ricans have access to potable water, and 92% have access to 
sanitation (World Bank Group 2005b). Costa Rica scores better than neighboring 
Panama in all these areas except population growth and primary enrollment, and better 
than Latin America as a whole in all areas except primary enrollment, immunization rates 
and population growth. Costa Rica's mean educational level in 2000 was 7.5 years, and 
46% of the population had a secondary education or higher (INEC 2001). 

Like most developing nations, Costa Rica is rapidly urbanizing. At the time of 
the 1973 population census (DGEC 1974), 41% of the population lived in areas defined 
by the census as urban, a number that had risen to 59% by 2000 (INEC 2001). By 2003, 
Costa Rica had seven telephones and four personal computers for every twenty people, 
compared to eight telephones and one computer per twenty people for both Panama and 
Latin America as a whole. In 2000, Costa Rica released 1.5 metric tons of carbon per 
capita annually, compared to 2.2 for Panama and 2.7 for all of Latin America (World 
Bank Group 2005b). 

The Costa Rican economy has always relied heavily upon exports, although that 
reliance appears to be decreasing. In 2004, exports accounted for 45.6% of GDP, down 
from 51.6% in 1999 (World Bank Group 2005b). In 2000, imports surpassed exports for 
the first time, and Costa Rica maintained a negative trade balance through at least 2002 
(World Bank Group 2003). The tourism industry is also an important part of the 
economy, with retail sales, restaurants and hotels accounting for 17.9% of GDP in 2001 
(SEPSA 2002). In 2003, Costa Rica's external debt was $5.8 billion, or $1,450 per capita 
(World Bank Group 2005b). 



23 



Agriculture, traditionally a main component of Costa Rica's GDP, has been 
gradually declining in relative importance, from 27.7% of GDP in 1982 (World Bank 
Group 2003), to 11.5% in 1995 (SEPSA 2002 p. 3) and 8.7% in 2004 (World Bank 
Group 2005b). Until recently, Costa Rica's most important agricultural export had been 
coffee. In 1984, coffee represented 27.6% ($491 million ) of Costa Rican exports, 
bananas 25.9% ($462 million), and manufactured goods 26.4% ($471 million). In 2003, 
they represented 3.2% ($194 million), 9.0% ($553 million), and 77.0% ($4.7 billion), 
respectively (World Bank Group 2005a). 

Coffee production took hold in Costa Rica in the early half of the 19 th century, 
decades before it became established in the rest of Central America (Williams 1994; cited 
in Rice 1999). By 1900, the coffee industry had driven population growth and the 
establishment of several urban centers. The absence of a large indigenous population 
created a chronic labor shortage in the coffee sector, a shortage that has never been fully 
resolved. Coffee represented nearly 100% of Costa Rican exports until 1900, when the 
country began to export bananas, and remained at around 50% until the 1930's (Winson 
1989 pp. 10-16). World War II dealt a serious blow to the Costa Rican coffee industry, 
further reducing its importance to the nation's economy (Winson 1989 p. 41) and 
increasing the need to diversify. 

Throughout most of the second half of the 20 th century, coffee was a motor of 
upward mobility for many in Costa Rica, increasing the wealth and consumption levels of 
most coffee growers (Hilje et al. 1994 p. 204). From the 1950s to 1980s, the social 
organization of coffee production in Costa Rica underwent a transformation, 
corresponding to the rise in technology use and the advent of farmer-owned cooperatives 
for processing and exporting coffee. Although state policies at the end of the 1950s 
aimed to encourage diversification and industrialization, the country maintained a 
dependency on agriculture, and coffee in particular, throughout this period (Vorst 1986). 

At the end of the 1980's, Central America accounted for 12% of world coffee 
production, with each country producing no more than 3% of world supply (Pelupessy 
1993 p. 13), and Costa Rica producing about 2%. Figure 1.3.3 shows Costa Rican 



In 2003 dollars 



24 



production as a percent of world production of all types of coffee from 1961-2003. Costa 
Rica's share of world production peaked in 1993 at 2.82%, and has declined fairly 
steadily since, to 1.6% in 2003. 

Despite this small portion of the world market, in the 1980s Costa Rica was one 
of four main centers of supply-side power in the International Coffee Organization, 
representing the producers of the Other Milds group of washed Arabicas (Aguilar 1992 p. 
26), of which it currently produces about 10% of world supply (Fernandez et al. 2002 
Anexo 5). Germany and the United States are the two largest importers of Costa Rican 
coffee, together accounting for 56% of exports in 2002 (Fernandez et al. 2002 p. 30). 

As of 1984, 30,706 km 2 of Costa Rica was farmland, or 60% of the total land 
area. About 900 km 2 of this (2.9%) was planted in coffee, 16,500 km 2 (53.7%) was 
pasture, and 14,600 km 2 (47.5%) was planted with other permanent crops, particularly 
bananas and sugar (DGEC 1987). Coffee area appears to have reached its peak in Costa 
Rica in 1989, at 1 15,000 hectares, after which it declined to 101,000 hectares in 1993-94, 
before increasing again to 113,000 hectares in 2001 (SEPSA 2002 p. 16; Pelupessy 1997 
p. 13). In Costa Rica, ripe coffee cherries are measured mfanegas, one of which equals 
400 liters, and cajuelas, twenty of which equals one fanega. Typically, one fanega of 
coffee cherries will yield about 100 pounds of green coffee (ICAFE 1993). 

In 2004, there were 95 beneficios in Costa Rica, 22 of which were farmer-owned 
cooperatives (Fernandez & Alvarez 2004). In the early 1990's, farmers who belonged to 
cooperatives received 67-69% of the export price of their coffee (Pelupessy & Tilburg 
1994 p. 253). In 2002, there were 34 roasters and 56 exporters in Costa Rica (Brenes 
2002). Most cooperative coffee was marketed directly by FEDECOOP, the Costa Rican 
Association of Coffee Cooperatives, circumventing exporters (Winson 1989 p. 108). In 
2001-02, 70,523 farmers and 303,766 paid workers were involved in the harvest (Brenes 
2002). 

Technology and input use in Costa Rican coffee production is among the highest 
in the world (Perfecto et al. 1996). Adoption of technology came early, when in the early 
1950's the Costa Rican government capped profits from fertilizer sales, lifted duties on 
imported fertilizers, and began to produce lime. In 1954 the Banco Anglo Costariccense 



25 



initiated a program to renovate 20% of the country's coffee farms that were larger than 
1.5 hectares. Cheap, easy credit and high labor costs also encouraged farmers to focus on 
inputs and technology (Winson 1989 p. 101-104). 

By 1963, a third of Costa Rica's coffee land was planted with new, high-yielding 
varieties (Winson 1989 p. 115). In 1978, the U.S. Agency for International Development 
(USAID) started a program, based in Costa Rica, to further encourage the adoption of 
new technologies in coffee production (Hernandez Navarro 1995; cited in Rice 1999). 
The scarcity of appropriate land for coffee production during the last few decades, and 
the high export prices of the late 1970's, encouraged monocultures in the most optimal 
coffee-growing regions. Throughout the 1970's and 1980's, the Ministry of Agriculture 
extension agencies promoted new technologies and coffee varieties, and the National 
Bank gave preferential interest rates for technification and renovation of coffee farms 
(Vorst 1986). 

By the early 1990's, an estimated 40% of Costa Rica's coffee-growing areas were 
"technified," representing the most intensive form of coffee production, while only 10% 
was considered "traditional" production. This represented the second highest level of 
technification in northern Latin America - after Colombia, with 69% of its coffee area 
technified (Rice 1999). 

As a result of the widespread incorporation of technology, Costa Rica has the 
highest coffee yields in Central America, where average yields in 1994 were around 800 
kg of green coffee per hectare, and as low as 500 kg/ha in Nicaragua (Pelupessy & 
Tilburg 1994 p. 251). From 1964 to 1993, average yields in Costa Rica more than 
tripled, from 561 kg/ha to 1585 kg/ha. During the same period, average yields worldwide 
went from 384 kg/ha to 546 kg/ha, an increase of 42%. By 2002, however, worldwide 
yields had reached 832 kg/ha, a doubling over 1964 levels (yields calculated from 
UNFAO 2005; ICAFE 2003). Some Costa Rican farmers can achieve yields of up to 70 
fanegas, or 3,175 kg, per hectare (Murillo 1996). Yields for farmers who belong to 
cooperatives is, on average, 22% lower than for farmers who sell to the private sector 
(Pelupessy & Tilburg 1994 p. 252). Figure 1.3.4 compares average yields for Costa Rica 
and the world from 1961 to 2003. 



26 



The Costa Rican Institute of Coffee is responsible for collecting taxes, 
establishing rules for coffee producers and processors, and providing technical and price 
information. When prices are above a pre-determined level, a tariff must be paid to the 
National Coffee Stabilization Fund (FONECAFE), which is used to compensate growers 
and processors when prices fall below costs of production. Processors' and exporters' 
profits are fixed at specific percentages, in an attempt to improve equity in the multi- 
tiered coffee export system (Pelupessy 1997 p. 26). For example, beneficios' profits were 
at 12% of the sale price from 1933 until at least the late 1980's (Winson 1989 p. 22). 
Enforcement of these caps appears to be strict, and the ICAFE collects and publishes 
detailed data about the prices received and paid by beneficios (pers. obs. 2003). 

While most coffee plantations in Latin America occupy 5-30 hectares of land 
(Pelupessy 1993 p. 23), the average size of a Costa Rican coffee farm in 1984 was 2.6 
hectares (Censo Agropecuario 1984). Because of the small average size of Costa Rican 
coffee farms, and the large number of farmers involved in coffee production, many 
writers have spoken of coffee as being a "democratizing" influence on the Costa Rican 
economy and culture. Even so, according to some analyses, land distribution in Costa 
Rica is the most unequal in Central America, if all types of land use are taken into 
account (Winson 1989 p. 96), although this could lend support to an argument that, 
compared to other products, coffee production is one of the more democratic sectors of 
Costa Rican agriculture. 

The historian Lowell Gudmundson argues that the "democratizing" influence is a 
myth, and that coffee has in fact served to consolidate power and influence in Costa Rica 
(Mora 2001a). Some have pointed out that the owners of many of the country's "small" 
coffee farms in fact belong to a few extended families (Stone 1976; cited in Boucher 
1983). Other authors have argued convincingly that the coffee "elite" in Costa Rica draw 
their power not from land ownership, as in other coffee-dependent countries, but from the 
control of processing, financing and exportation (Cardoso 1973; cited in Vorst 1986). 
Even so, the influence of the coffee elite appears to have peaked in the 1920's (Winson 
1989 p. 38), before the coffee crisis of the Great Depression. 



27 



Democratic or not, there are clearly many structural inequities in the Costa Rican 
coffee production chain. While more than 90% of farms occupy less than five hectares 
(Pelupessy 1997 p. 14), together these small farms make up only 40% of the total coffee- 
growing area in Costa Rica (Pelupessy 1993 p. 24). This is, however, still higher than the 
Central American average, where small farmers own 27% of land area and produce 21% 
of the coffee (Pelupessy & Tilburg 1994 p. 251). Small farmers in Costa Rica can afford 
fewer inputs (Winson 1989 p. 115), and thus have lower yields than large farmers. For 
example, in the 1991-92 harvest, farms with less than 5 hectares of land produced the 
equivalent of 1,300 kg of green coffee per hectare, while those with more than 5 hectares 
produced close to 2,000 kg per hectare (ICAFE 1993 p. 67). Almost 93% of Costa Rican 
farmers harvest less than a total of 4,500 kg of coffee per year, and together they account 
for only 45% of Costa Rica's coffee (Brenes 2002). 

Small farms are also the least likely to incorporate technology, and have the 
highest costs of production per pound if they do use a technology "package." Small-scale 
producers often can't take advantage of government renovation subsidies, because they 
cannot afford to wait for new varieties to come into production (Vorst 1986). The ICAFE 
calculated in 2002 that the average costs of production for a farmer using the 
recommended "technology package" and using only paid labor was $86.07 per fanega 
(Portilla R. & Araya M. 2002 p. 7), however many farmers are able to reduce costs of 
production by using family labor and reducing inputs. Three studies carried out between 
1999 and 2002 estimated production costs in Costa Rica at between $0.77-$ 1.00 per 
pound, the highest in Central America (Varangis et al. 2003 p. 20). 

Small-scale farmers in Costa Rica bring their harvests to one of thousands of 
collecting stations or recibidores around the countryside, each owned and operated by a 
processor or beneficio. Rather than being paid for the coffee on the spot, Costa Rican 
producers sell their coffee on consignment through the beneficio, receiving payment only 
when the green coffee has been sold. Thus, a farmer may not receive the final payment 
for his harvest, delivered in October or November, until July or August of the following 
year, in effect providing an interest-free loan to the beneficio for up to a year (Pelupessy 
1997 p. 24). This applies whether the beneficio is a cooperative or a private enterprise, 
and results in coffee farmers bearing most of the risk of the international market (Vorst 



28 



1986), while having little direct control over the variables affecting that risk (Pelupessy 
1997 p. 24). Nevertheless, this does guarantee that the farmers receive a fixed share of 
the real export price of their coffee, without relying on speculation at the time of harvest. 

Beneficios pay farmers based on the exchange rate at the time of harvest, while 
they sell the processed coffee at the rate in effect several months later. With the colon 
being devalued at upwards of 20% per year, this means that farmers receive substantially 
fewer colones, at a time when those colones are worth less, than the beneficio receives for 
selling their product. Beneficios are also required by law to sell a certain portion of their 
coffee (usually 10-15%) to the domestic market at a sharply discounted rate (Anonymous 
1992b). This discount, which in 2003 resulted in a $5.2 million loss to farmers (an 
average of $72 per farmer) and $2.3 million to processors (calculated from ICAFE 2003), 
is in effect a subsidy paid by Costa Rican coffee producers to the country's coffee 
drinkers. 

The introduction and wide incorporation of technology has resulted in making 
much of the employment in coffee production seasonal or temporary, and reduced the 
total labor needed overall while increasing labor needs during the harvest (Hidalgo 1983; 
cited in Vorst 1986). For example, herbicide use reduces the labor needed for weed 
control by 75%, and use of fertilizers and fungicides can triple yields (Vorst 1986). 

As a result, permanent employees only account for about one fifth of the harvest 
labor in Costa Rica: the others come from other sectors, the unemployed, or other 
geographical regions (Vorst 1986). According to studies by the Migrant Labor Division 
of the Costa Rican Employment Ministry, 13% of the country's coffee pickers are 
Panamanian Indians and 44% are Nicaraguans (Wa-chong 2001). Many coffee-growing 
families travel elsewhere to pick coffee after their own has been harvested. At least until 
the late 1980's, coffee workers did not have the right to strike because coffee production 
was considered a "national interest," and workers did not have representation in the 
ICAFE (Vorst 1986). 



29 



Figures 



Costa Rica Percent of World Production 



3.00% 
2.80% 
2.60% 
2.40% 
2.20% 
2.00% 



1.60% 



1.00% 



1993 High 2.82% 




1965 Low 1.17% 



T-comr^cDT-coLoi^coT-coioi^cnT-coioi^cn 
cococQcocoi^i^r^i^i^oooococococncncncncn 



o 
o 



o 
o 



Figure 1.3.3. Costa Rican production as a percentage of world production. Data from Fernandez et al. 
(2002) and UNFAO (2005). 



Coffee yields 1961-2003 



1600 
1400 
1200 



(0 


1000 


a. 


800 




600 



400 



200 




World Mean = 539 



■i- co lo K en t- 

CD CO CO CD CD r~- 



CO LO 

r^ r-~ 






CO LO I s - C73 i- CO LO 

i^oooococoooa>cno">co 



0)0)<J>0)0)0)<J>G><J>G>(J>G>0)0)<J>0)0)0)<J> 



cr> 
cn> 

CT> 



o 
o 



CO 

o 
o 



Figure 1.3.4. Coffee yields in kg/ha for Costa Rica and the world 1961-2003. Calculated from UNFAO 
(2005) and ICAFE (2003). 



31 



Chronic overproduction 

"Coffee gives you a jacket and it takes your shirt. " - Brazilian saying 

Coffee prices experience both short-term fluctuations and long-term cycles. The 
former are primarily influenced by speculative activity, with the latter being influenced 
by the relationship between world consumption and the amount of coffee available. The 
policies of the large trading houses also have a considerable impact on prices. Both 
supply and demand show low short-term price elasticity for coffee, meaning that prices 
have little short-term relationship to either the availability of coffee or consumers' 
interest in buying it (Portillo 1993). 

As mentioned above, because there is a lag of several years between a producer's 
decision to increase supply and the time that supply reaches the market, supply and 
demand are rarely in equilibrium. This is exacerbated by the fact that small producers, 
confronted by low prices, will often increase supply in order to continue to earn a 
subsistence income (Pelupessy 1993 p. 41), while the traditional economic model expects 
them to reduce supply. Because the initial costs of planting and maintaining coffee trees 
make up such a large share of total production costs, farmers will usually continue to 
harvest and sell beans while prices are well below total average costs, in order to recover 
at least some of their initial investment (Lindsey 2003 pp. 3-4), further reducing the 
ability of coffee supply to respond to low prices. 

Surplus beans from one harvest are carried over to the next in warehouses 
controlled primarily by large roasting companies or government entities in producing 
countries. The export price of coffee has a direct negative relationship to the size of these 
stockpiles (Daviron 1994 pp. 70-71), which continue to keep prices low even after 
production has decreased (Lindsey 2003 p. 4). "Boom" coffee cycles tend to be shorter 
than "bust" cycles, with low prices lasting an average of five years, and high prices only 
two (Morales 2000). As Wim Pelupessy (1993 p. 14) remarks, "We are not confronted 
by a fully competitive market, with transparency and free and equitable access by its 
participants." 



Quoted in Pendergrast (1999), p. 323 

33 



Brazil is the world's largest producer of unwashed Arabica and the third largest 
producer of Robusta (ICO 2005a; 2005e). With its large land area in production, 
susceptibility to natural climatic fluctuations such as drought and flood (Pelupessy 1991), 
and reports of annual government subsidies of up to $200 million to its producers 
(Anonymous 1995), Brazil's participation in the coffee market further magnifies the 
volatility of an already unstable product, making it by far the largest distorting factor in 
the global coffee market (Figure 1.3.5). 

The first major coffee overproduction crisis of the modern era was in 1896-1908, 
brought on by the vast expansion of coffee area in Brazil - which, at the turn of the 
century, supplied about 75% of the world's coffee (Daviron 1994 p. 43) - and by a global 
recession (Jimenez 1994; cited in Kutschbach 1994a p. 14). The low coffee prices 
persisted even after the North American and European countries had recovered from the 
recession (Kutschbach 1994b p. 93). Beginning in 1907, Brazil began the first program 
of coffee "valorization," an attempt to stabilize world prices by withholding stocks from 
the market (Brignoli 1994 p. 27). Prices recovered by 1910, stimulating another increase 
in the area under production (Kutschbach 1994b p. 96). 

As long as Brazil remained an important enough producer to be able to hold up 
prices on its own, other producing countries were unwilling to enter into multilateral 
agreements to control supply, and instead took advantage of the situation to increase their 
own market share. Ironically, by propping up the world's coffee prices, Brazil 
encouraged the development of coffee production in other countries, and by the 1920's 
Africa, and other countries of Latin America, had begun to emerge as important 
producers of the bean (Daviron 1994 p. 43, 49). From the 1930s to the 1970s, France 
financed the expansion of coffee production in its colonies and favored imports from 
those countries, with the result that coffee production in French colonies in Africa 
increased by a factor of fifteen (Daviron 1994 p. 51). 

The second crisis of the 20 th century began in 1927, as coffee from newly planted 
regions around the world entered the market. It was soon compounded by the beginning 
of the Depression in 1929, which affected all exports, imports and the international 
monetary system (Kutschbach 1994b p. 107). Tropical countries watched the prices of 



34 



all their commodities fall, and Central America was particularly hard hit by the 
simultaneous slide in coffee and banana prices (Kutschbach 1994b pp. 115-117). That 
year Brazil began destroying coffee stocks outright, burning 4.68 million Mt of green 
coffee between 1931 and 1942, the equivalent of three years of world consumption 
(Cordero et al. 1987; cited in Portillo 1993). 

The ability of farmers to survive the 1930's crisis had a great deal to do with 
diversification, particularly at the level of individual farms. Small-scale producers were 
less affected by this crisis than larger ones because of this diversification (Kutschbach 
1994b pp. 119, 122). Costa Rica was somewhat insulated from the crisis by their 
relatively privileged position in the European market - in 1936, Costa Rican coffee was 
fetching the second highest price in the world. German policy changes in 1934, however, 
pushed Costa Rica into a more competitive market (Winson 1989 pp. 37-39). 

In 1940, the United States and the Latin American producers signed the 
Interamerican Agreement, which initiated the world's first quota system (DiFulvio 1947; 
cited in Daviron 1994 p. 50). In the six months after signing the agreement, coffee prices 
jumped by 60%. After the U.S. declared war on Japan in 1941, the U.S. froze the prices 
of coffee imports until October 1946 (Daviron 1994 pp. 50-51). During the second 
World War, with Europe basically cut off from world markets, the United States was 
importing 80% of the world's coffee (Portillo 1993). 

After 1946, coffee traded in an essentially free market (Portillo 1993). The 
postwar economic revival in Europe, combined with a drought and subsequent frost in 
Brazil, led to a dramatic upsurge in prices (Kutschbach 1994a p. 18). World stocks were 
reduced from 1 to 0.39 million Mt, and the price of Colombian milds went from $0.19 to 
$0.78 per pound (Portillo 1993). Between 1954 and 1956, new plantations in Brazil, 
Africa, Central America and Mexico came into production, resulting in another glut of 
coffee, accumulation of stocks, and an accompanying fall in prices (Rowe 1963; cited in 
Portillo 1993; Daviron 1994 p. 53). 

In 1957 the Latin American producing countries signed the Mexico Accord, 
renewed a year later as the Latin America Accord. The goal of both agreements was to 
limit exports, but the strength of the agreements was limited without the support of the 



35 



African producers, who entered into the negotiations in 1959. An international accord 
was reached among producing countries in 1959, and renewed in 1961 and 1962 
(Eisemann 1982; cited in Daviron 1994 p. 53). 



36 



Figures 



c 
o 



O 

'_ 

Q. 

d> 
d> 

O 

o 



8 
7 
6 
5 
4 
3 
2 
1 



Brazil's Effect on World Coffee Supply 1961-2003 




World without 
Brazil 



T-comN-cn-i-comr^a>T-comr^a>T-comr^a)-i-co 

(D(DCD(Dtt)NSSNNCOCOCO(BCOO)0)0)0)0)00 
0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)0)00 

r-T-T-T-T-T-1-T-T-T-T-l-T-T-T-T-l-l-l-T-NW 



Figure 1.3.5. Brazil's effect on world coffee supplies. Data from (UNFAO 2005). 



37 



The International Coffee Agreement 

"It is plain that no program of economic development can be effected 
unless something is done to stabilize commodity prices." -U.S. President 
John F. Kennedy, introducing the Alliance for Progress on March 13, 1961 

The Cuban revolution in 1959 intensified fears that an economic crisis in 
developing countries could pave the way for the spread of Communism, and the United 
States, through the vehicle of the newly-created Alliance for Progress, began seeking 
ways to stabilize coffee prices (Pendergrast 1999 p. 276). In an effort to curb the 
volatility of the market and provide some security for both producing and consuming 
nations, the first International Coffee Agreement (ICA) was signed in 1962 (Kutschbach 
1994a pp. 18-19), with producers of 99%, and consumers of 85%, of the world's coffee 
as signatories (Daviron 1994 p. 57). The non-signatory nations were the countries of 
Eastern Europe and the Middle East, Argentina, Algeria, Chile, Magreb, New Zealand, 
South Africa, both Koreas and Hong Kong (Portillo 1993). The second ICA was signed 
in 1968, the third in 1976, and the fourth in 1983 (Pelupessy 1993 p. 33). 

The ICA had four main tools available to it to control the price of coffee: pricing, 
taxation, quotas, and storage. Each ICA producing country had an annual fixed and 
variable quota. The fixed quotas accounted for 70% of supply, were based on historical 
production levels, and often were politically influenced. Variable quotas were based on a 
country's available stocks in proportion to total world stocks, and accounted for 30% of 
supply (Aguilar 1992 pp. 16-17). 

Quotas would come into effect when an indicator price, calculated by averaging 
the price of Robusta and Arabica Other Milds, fell below a pre-specified price floor and 
remained there for at least 20 days. If the indicator price rose above a pre-set ceiling, 
quotas would be increased or dropped altogether. In the late eighties, this price band 
corresponded to around $1.20 to $1.40 per pound (Pelupessy 1993 p. 33). Quotas only 
applied to exports to other member nations, and did not apply to small-scale exporters 
(Aguilar 1992 p. 17). Countries that produced less than 6,000 metric tons of coffee were 



Quoted in Pendergrast (1999), p. 276. 

39 



exempt from the quota system, as their share of the total global quota was only around six 
percent (Pelupessy 1993 p. 33). 

With the exception of a price spike in the late 1970's, prices were relatively stable 
at around $2 a pound (in 2003 dollars) for the duration of the ICA (see Figure 1.3.6). 
Economic models indicate that the prices maintained by the ICA may have been as much 
as double the free-market price (Daviron 1994 p. 71). 

There were a number of weaknesses in the ICA system. Because the quotas were 
historically and politically based (Pelupessy 1993 p. 38), they did not allow flexibility in 
dealing with increases in production or meeting the changing demands of coffee 
consumers. Producers of Other Milds, in particular, wanted an expanded market share 
(Aguilar 1992 p. 18), and the United States wanted increased access to higher quality 
coffee. During the 1980's, the price difference between Robusta and the Arabicas grew 
from 12 cents to 50 cents per pound of green coffee. Since the ICA indicator price was 
based on an average of the price for Arabica and Robusta, this meant that quotas could 
not be lifted for Arabica even when prices for that type of bean increased, and could not 
be implemented for Robusta when the price of that bean fell (Daviron 1994 p. 62; Portillo 
1993). 

Many countries considered the quotas to be inequitable, as they often left smaller 
producers holding back as much as half of their production, while countries such as 
Brazil and Colombia were able to maintain stocks at low levels (Daviron & Lerin 1986; 
cited in Daviron 1994 p. 16). In the 1970's, Brazil had 30-40% of the world quota but 
maintained 80-90% of the world stocks, but by 1989 their share of world stocks was only 
eight percent (Portillo 1993). 

There was also contention over sales of coffee to non-member countries. When 
producing member countries had stocks available over their quota allocations, these could 
be sold to non-member countries for 40-50% below ICA prices (Daviron 1994 p. 59). 
The ICA permitted this, ostensibly, to encourage the growth of coffee consumption in 
non-traditional markets (Portillo 1993). Much of this coffee was then re-exported to 
member countries, earning it the nickname "tourist coffee" (Pelupessy 1993 p. 29). 



40 



While the non-member market served as an escape valve for producing countries to sell 
off excess stocks, it also undermined the system as a whole. 

While in 1971-73 exports to non-member countries accounted for only six percent 
of world exports, from 1982 to 1989 they represented 15% (Daviron 1994 p. 59; Portillo 
1993) and by 1990 they had expanded to 17% (Pelupessy 1993 p. 28). Indonesia was the 
biggest exporter to the non-member market, accounting for 20% of the total amount of 
coffee exported to non-member countries, but India, Mexico, Costa Rica and others were 
also very involved in this market (Portillo 1993). 

A final difficulty with the ICA was the inability of producing countries to control 
overproduction. In fact, the ICA's reliance on past production for setting quotas created 
strong incentives for countries to ramp up production in order to secure a larger quota in 
subsequent years (Rice 1999). The years 1965-1966 began the greatest coffee production 
in history, stimulating the accumulation of large stocks (Portillo 1993). Figure 1.3.7 
shows the difference between world production and consumption for 1961-2002. Up to 
40% of green coffee loses its export quality in storage (Pelupessy 1993 p. 40). The cost 
of maintaining - or destroying - these stocks must be deducted from the artificially high 
prices promoted by the ICA in determining the true costs and benefits of the ICA, and 
may have been the ultimate cause of the ICA's collapse (Pelupessy 1993 p. 34). 

Ironically, the artificially high prices maintained by the ICA stimulated expansion 
of growing areas, lack of diversification, and the use of higher inputs (Kutschbach 1994a 
p. 18; 1994b p. 120), contributing in the end to the decline of the ICA. For example, 
between the years of 1970 and 1992, coffee area in Mexico more than doubled (Bray et 
al. 2002). The second ICA included a "diversification fund," with the goal of encouraging 
production of alternative products in zones that were marginal for coffee (Daviron 1994 
p. 57), but this was dissolved in 1976 (Portillo 1993). Throughout the 1970's and 1980's, 
programs promoted by governments and international development agencies such as 
US AID encouraged "modernization" programs that increased yields (Rice 1999), raising 
the costs of production and compounding the problem of overproduction. 

In 1972, negotiations for the third ICA broke down because consuming countries 
refused to adjust the price bands to account for the devaluation of the dollar (Portillo 



41 



1993). The economic clauses of the ICA were suspended and the Agreement operated 
without quotas or a price band until 1980, although a third ICA was signed in 1976 
(Daviron 1994 p. 58). After record-high prices in the late 1970's due to a series of 
natural disasters (Pendergrast 1999), prices slumped again at the end of the decade, 
triggering the enactment in 1980 of the 1976 Agreement (Daviron 1994 p. 58). 

In 1983, a set of reforms were proposed for the ICA. The United States, now 
under the direction of the free-trading Reagan administration (Dicum & Luttinger 1999), 
suggested abandoning quotas in favor of management of stocks. The European Union 
suggested a global quota that included sales to non-member countries. The Central 
American countries wanted selective quotas based on coffee type. None of the suggested 
reforms were adopted. Two negotiating blocks formed: Central America, the United 
States, Canada and Australia favored a new, objective division of the quotas, which 
would favor Central America and Colombia. Brazil, Colombia, Africa, the European 
Union and the Phillipines wanted to maintain the old quotas but control sales to 
nonmembers (Portillo 1993). 

Producing nations had hoped that the 1986 Uruguay round of negotiations for the 
General Agreement on Tariffs and Trade (GATT, now the World Trade Organization) 
would improve their position in the market, lowering some of the barriers to importation 
of value-added goods. In fact, the Uruguay round had only a minimal effect on the coffee 
industry, with the EU agreeing to lower tariffs on green coffee from five to four percent 
(Pelupessy 1993 p. 38). The tariffs on value-added products remained. 

In 1985, a frost in Brazil led to another price spike, and quotas were suspended 
from February 1, 1986 to October 6, 1987 (Pelupessy 1993 p. 38; Daviron 1994 p. 58; 
Portillo 1993). Prices began to fall again in 1987, and quotas were re-established in 
February of that year (Pelupessy 1993 p. 39). The fourth ICA was set to expire in 1992. 
Negotiations for a new agreement stalled in 1989, when the two major negotiating blocks 
could not reach an agreement (Portillo 1993), and the agreement was suspended in July 
of that year. 

U.S. President George H.W. Bush, while an ardent free-trader, indicated in a 1990 
letter to Colombian President Virgilio Barco a willingness to negotiate a new ICA, with 



42 



the explicit goal of controlling drag trafficking. By September of 1992, all countries 
were finally able to agree on a global quota that included non-member countries and to 
deal with the issue of quota selectivity in relation to coffee types. The United States, 
however, insisted on a system that would more or less mimic the free market (Portillo 
1993). The final ICA expired in 1992, with no new agreement in sight. 



43 



Figures 



World Export Prices 1961-2002 



52 
re 

o 
a 

co 

o 
o 

CM 



$6.00 -r 



$5.00 



$4.00 



$3.00 




$2.00 



$1.00 



$0.00 

-T-cotfor^a>-!-ooLni"^a>-!-coLnr^a>-!-couor~a>-!- 

(O(O(O(O(0NNNKN<ttC0C0a}C0Ch<3>O><J)O)Q 

1-1-1-1-l-T-T-T-T-T-T-l-l-l-l-T-l-l-T-T-CM 

Figure 1.3.6. World export prices for coffee, all types, in constant 2003 dollars. Data from UNFAO 

(2005). 



Over/Under Production 1961-2002 



g 3 




/W/^ y 




T-coLnr^cn-i-coLnr^a)T-coLnh-cnT-coLnh-a» 
cD(DcDcDcDh-h-r^r^r^oooooocoooa)aiaia»a» 



Figure 1.3.7. Total world consumption and production, and difference in world consumption and 
production, from 1961-2002. Calculated from UNFAO (2005) and (ICO 2005b). 



45 



New Players on the International Stage 

Despite Brazil's market dominance, which continues to this day, other producing 
countries have gradually increased in importance. Between 1957 and 1972, Brazil and 
Colombia's share of the ICA quotas fell steadily, from a combined 54% to 46%, while 
countries in Africa and Asia increased their quota shares from 27% to 36% (Federacion 
de Cafeteros de Colombia 1978; cited in Daviron 1994 p. 55). At its peak, in 1961, Latin 
America accounted for more than 76% of world production. By 2004, that share had 
fallen to 59% (UNFAO 1995). 

In the early 1990's countries in Asia began to dramatically expand their coffee 
production. The UN Food and Agriculture Organization did not begin tracking aggregate 
data for Asian coffee production until 1992, but between 1992 and 2004, coffee 
production in Asia doubled from just under one million Mt to just over two, increasing 
from 16% to 27% of world supply. The area harvested went from 1.5 million to 2.2 
million hectares, and yields increased from 678 kg/ha to 935 kg/ha (UNFAO 2005). 

Asia's growth in this sector was driven largely by Vietnam and Indonesia. 
Between 1992 and 2004, Vietnam's production increased six-fold, from 119,000 to 
835,000 Mt of green coffee. Indonesia's production increased 60%, from 437,000 to 
702,000 Mt. In that same period, Brazil's production nearly doubled, from 1.3 to 2.5 
million Mt. While in 1989, Vietnam was 20 th in its share of world coffee exports, by 
2002 it was second, after Brazil (UNFAO 2005). In 1993, Vietnam accounted for only 
1.5% of world supplies. By 2002 its share was forecast to reach 8.0% (Franklin Lee* in 
U.S. House of Representatives 2002 p. 41) Figure 1.3.8 shows the growth in production 
in Brazil, Vietnam and Indonesia from 1992 to 2004. 

Many authors have argued that the World Bank was involved in the expansion of 
coffee production in Vietnam. Oxfam International (Gresser & Tickell 2002) criticizes 
World Bank and IMF policies that encouraged over-dependency of developing nations on 
export commodities, and link these policies to the overproduction of coffee. In his book 
Coffee: A Dark History , historian Antony Wild (2004) states: 



47 



Both [the World Bank and the Asian Development Bank] had lent heavily to 
Vietnam in the mid 1990s in line with their mandate to stimulate low-cost 
production and end market inefficiencies. Having massively defoliated the nation 
with Agent Orange during the Vietnam War, the USA promoted - through the 
World Bank, in which it has a controlling stake - the refoliation of Vietnam with 
low-grade Robusta coffee bushes . . ." (p. 6) 

The World Bank, however, has denied involvement in the expansion of Vietnam's 

coffee industry (O'Brien & Kinnaird 2004). According to Miriam Wasserman (2002 p. 

6), the World Bank ". . .says it resumed lending to Vietnam in 1994, after the country's 

coffee expansion was already under way and that, though $16 million from a loan to the 

Agricultural Bank was used to finance coffee farm rehabilitation, it has not lent directly 

to the coffee sector." Robert Rice (1999) argues that coffee was one of a number of 

"black-listed" commodities for which the World Bank historically would not provide 

funding, citing interviews with World Bank officials in 1993. 

The confusion, controversy and uncertainty over this issue may be a result of what 
Wild (2004) refers to as the "official veil of secrecy," with the World Bank "unwilling to 
shoulder the responsibility of having contributed substantially to the collapse in world 
coffee prices" (p. 295). Nevertheless, without more complete information, it is difficult 
to support conclusions regarding the World Bank's share of the responsibility for the 
coffee crisis. 

In 2000, what appears to be the first mention in the Costa Rican press was made 
referring to Vietnam and Indonesia, rather than Brazil, as the primary sources of 
overproduction (Rojas 2000). Concerns over Vietnam's growing role in the market 
increased as France announced plans to assist the Robusta-producing nation to begin 
growing Arabica (Ulett 2001). 

As coffee production expanded in Asia, and as Brazil continued to ramp up 
production, large coffee roasters began utilizing new processing techniques, such as 
steaming, that removed the bitter taste from lower-quality coffee beans, enabling them to 
decrease the quantity of higher-quality Arabicas in their blends (Lindsey 2003 p. 5; Inter- 
American Development Bank et al. 2002). Thus, while both Indonesia and Vietnam 
produced predominantly Robusta (ICO 2005b), these new technologies meant that 



48 



Robusta was increasingly able to compete directly with Arabica in certain low-cost 
sectors of the market. 

Despite the dramatic increase in Vietnam's production compared to earlier years, 
and despite the large amount of attention given to that increase, Brazil still remains the 
driving force in the world coffee market, accounting for most of the year-to-year 
variation (see Figure 1.3.5). In 2004, Brazil exported more than three times as much 
coffee as Vietnam, and nearly twice as much as Vietnam and Indonesia put together (ICO 
2005 e). Vietnam dominated the Robusta market, however, exporting more than four 
times as much Robusta as Brazil and twice as much as Indonesia (ICO 2005a; 2005e). 
Together, Brazil and Colombia exported nearly two-thirds of the world's Arabica beans 
(ICO 2005a; 2005e; 2005c). 



49 



Figures 



Coffee Production 1992-2004: Brazil, Vietnam, Indonesia 



3,000 



o 
tn 

C 
(0 

w 

3 
O 




1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 



Figure 1.3.8. Increases in production in Brazil, Vietnam and Indonesia from 1992-2004. Data from 
(UNFAO 2005). 



51 



The Current Crisis 

The 1980's have been called by some authors "the lost decade" in Latin America, 
characterized by economic stagnation, crippling foreign debt and new policies of 
"liberalization" that privatized many state functions (Kutschbach 1994b p. 124). The 
entrenchment of these policies make prospects dim for any future agreement resembling 
the ICA (Daviron 1994 p. 64), and the effects of the current crisis are so profound that 
they would be felt even long after such an agreement was implemented (Hilje et al. 1994 
p. 163). 

After the collapse of the fourth and final ICA in 1989, export prices fell by 40% 
in just a few months (Pelupessy 1993 p. 34), to the lowest level since the 1820's (Dumas 
1992; cited in Kutschbach 1994a p. 22). For most Latin American countries, which had 
not begun producing coffee until the mid- to late- 19 th century, these were the lowest 
prices in history. From 1989 to 1991, exports exceeded world consumption, stocks built 
up in consuming countries to double the usual level, and production continued to rise. By 
September of 1991, prices on the New York Coffee, Sugar and Cocoa exchange were at 
$0.50/lb, in real terms as low as during the Depression. Even as export prices of green 
coffee fell more than 40%, the price of roasted beans fell only 9.5% in the U.S. (Portillo 
1993). 

From 1990 to 1992, farmers worldwide earned $8 billion less than the previous 
period, while exporting 1.6 million Mt more (Gutierrez 1992; cited in Portillo 1993). 
Brazil alone lost an estimated $1.5 billion a year in coffee revenues, and Africa lost $3.2 
billion. For comparison, World Bank aid disbursements to Africa amounted to $2.9 
billion during the same period - or less than the total amount of coffee revenues lost. The 
price of cacao, another product important to many coffee-growing countries, also fell in 
the early 1990's, reaching a 14-year low (Portillo 1993). 

The situation was exacerbated by a concomitant collapse in banana prices. In the 
early 1990's, banana was the number one export from Costa Rica, accounting for 20% of 
export income. Coffee came in third, with 10% of export income. In Guatemala, coffee 
was first and bananas third, and in Honduras the two together accounted for 50% of 
export earnings. Panama, which depended on coffee for only 3% of export earnings, 



53 



received 45% from bananas, while in El Salvador, 45% of export earnings came from 
coffee. 

Between June 1991 and June 1992, banana prices had fallen by 60-65%. The 
European Union announced a plan to implement a quota system favoring the countries of 
Africa, the Caribbean and the Pacific, and a tariff of 20% on banana imports from Latin 
America. The Union of Banana Exporting Countries, which consisted of Colombia, 
Costa Rica, Guatemala, Honduras and Nicaragua, predicted that these moves would 
decrease Latin American exports by as much as 30%, with up to 400,000 families losing 
their livelihoods (Murillo 1992). 

By February of 1992, losses in coffee earnings to producing nations worldwide 
were $6-7 billion, with Costa Rica losing $100-150 million annually (Murillo 1992). The 
government of El Salvador had declared a state of emergency, and FEDECOOP 
requested that the Costa Rican government do the same (Vega M. 1992). 

In Costa Rica, coffee had ceased to be profitable for most farmers. The average 
Costa Rican farm yielded 3,295 pounds of coffee per hectare at a cost of $2,016 (in 1992 
dollars) or $0.61 per pound (Viquez 1992a). For the 1991-92 harvest, the average price 
paid to farmers was $0.51 per pound, or $1680 per hectare, for a net loss of $0.10 per 
pound or $336 per hectare (calculated from ICAFE data). One exporter prophesied that 
Costa Rica's coffee industry was "dying," and Mario Fernandez, director of FEDECOOP, 
predicted the disappearance of small producers (Vega M. 1992). An op-ed in the daily 
newspaper La Nation remarked, "The global crisis affects Costa Rican society so deeply, 
it is as though it were a crisis in our own democracy" (Corrales B. 1992). This author 
insisted that Costa Rican democracy would be threatened unless a way could be found to 
help small farmers keep their land. 

In 1993, a number of coffee producing countries agreed to implement a 
coordinated withholding of supply (Kutschbach 1994a pp. 23-24), which resulted in a 
small and temporary price recovery (see Figure 1.3.6). The average price of $0.77 per 
pound paid to growers during the first half of 1994 (ICO 2005b) was finally enough to 
cover production prices, but not enough to provide significant relief to farmers, and early 
in the year, the Central American countries considered another coordinated withholding if 



54 



the price did not rise above $0.79/lb (Murillo 1994), the price considered by the 
Association of Coffee Producing Countries to be the break-even point for farmers 
(Anonymous 1995). At the same time, the EU had still failed to reach an agreement with 
the Central American countries over bananas, further prolonging the uncertainty in that 
market (Associated Press Guatemala 1994). 

Between 1989 and 1994, the export price of Costa Rican coffee never rose above 
$0.84 a pound, making it the worst crisis in Costa Rican history (Munoz N. 1997). 
Thanks to a frost in Brazil (Anonymous 1995), average green coffee export prices 
reached another peak in 1995 at $1.31 a pound globally, then tumbled again, never to 
recover (data from UNFAO 2005). In Costa Rica, the price of coffee exports continued 
to rise for another three years, averaging $1.45 a pound in 1998 (Fernandez et al. 2000 p. 
26). Only $1.09, however, reached the farmer (Fernandez et al. 2000 p. 27), at a time 
when the cost to produce a pound of coffee averaged $0.92 (Barquero S. 1998). 

In late 1998, the Brazilian plantations that had been re-planted after the 1994 frost 
began to come into production, and Brazil announced a large harvest (Morales Ch. 1999), 
prompting prices to fall again (see Figure 1.3.6). While the value of Costa Rica's coffee 
exports for the first half of 1998 were $282 million, they fell to $193 million for the same 
period in 1999 (Morales Ch. 1999). By mid-August of 1999, the price of coffee on the 
New York exchange had fallen to $0.87, presumably on news that Brazil's 2000-01 
harvest would reach 2.4 million Mt, instead of the usual 1.8 million (Barquero S. 1999). 
In the same year, the value of Costa Rica's banana exports had fallen by half from two 
years previous (Murillo 1999). 

In 2000, consuming countries were maintaining stocks of green coffee equal to 
5.29 million Mt, while global consumption was only 4.97 million Mt. Between January 
and December of that year, the price of green coffee on the New York exchange went 
from $1.14 to $0.67 a pound (Morales 2000). This was a more severe price drop than the 
post-ICA collapse of 1989 (Prendas 2001b). In October of 2000, Costa Rica and other 
producing nations again sought to raise prices by withholding 20% of their coffee harvest 
(Rojas 2000). Another plan was discussed at the time to permanently remove 10% of 
Central America and Mexico's lowest quality coffee from the market (Morales 2000). 



55 



Also in 2000, the effect of falling coffee prices on national economies was further 
compounded by high oil prices, which had risen nearly 60% over the two previous years. 
The Costa Rican newspaper Al Dia reported that in January 1998, the proceeds from one 
quintal (46 kg) of green coffee would purchase 8.39 barrels of crude oil, but by October 
2000, 33 months later, one quintal of coffee would buy only 2.95 barrels of crude. Costa 
Rica received $19 million less for its coffee exports in 2000 than in 1999, while spending 
$140 million more on oil. Higher oil prices increased production costs and made 
improvements to infrastructure more costly (Rojas 2000). 

Meanwhile, the EU had still not reached an agreement with producing countries 
over bananas, and the devaluation of the Euro was further impacting banana prices, 
though the decline was not nearly as drastic as for coffee. In 1999 Costa Rica had 
exported 115 million 18 kg boxes of bananas at $5.45 a box; in 2000 they exported 100 
million boxes at $5.20 a box (Morales 2000). 

By 2001, it was clear that the 20% retention scheme had failed to affect prices, 
with only Costa Rica, Brazil and Colombia complying with its terms. Even the 
substantial quantities held by the latter two countries had little effect, in part because 
Brazil had undertaken a program to devalue its currency to increase exports, further 
depressing prices (Prendas 2001a). Oxfam International reported that in Vietnam's Dak 
Lak province, coffee prices only covered 60% of production costs (Gresser & Tickell 
2002 p. 2). 

Plans were made to destroy the lowest-quality five percent of Costa Rica's coffee, 
but this was not expected to have any effect on prices for up to 18 months (Prendas 
2001b). That year, Costa Rican coffee specialist Rodrigo Jimenez Robles estimated 
farmers' costs of production at $0.82 per pound (Mora 2001a), while the average price 
paid to growers that year was $0.48 per pound (Fernandez et al. 2002). Eugenio Porras 
Vargas, president of the Costa Rican College of Agronomic Engineers, predicted that in 
the 2001-02 harvest farmers would lose nearly $1,000 per hectare (Mora 2001a). 

In early 2002, new hopes for a price recovery were rekindled when Brazil 
predicted a small harvest for 2002-03, began paying its farmers to withhold a part of the 
harvest, and began to revalue its currency (Prendas 2002). That recovery appears to be 



56 



finally materializing, although it has taken longer than was hoped. While Brazil's 
production was down nearly 20% in 2003, by 2004 it had rebounded nearly to the record 
levels of 2002, and Vietnam and Indonesia have continued to maintain their high 
production levels (UNFAO 1995). Nevertheless, prices have improved. In 2004, the last 
year for which full data were available, the ICO indicator price averaged $0.62 a pound, 
while prices in 2005 averaged $0.90 a pound through October. 

It appears that the increase in prices has largely been driven by higher prices for 
Arabica beans. Prices for Other Milds averaged $1.16 a pound for January-October 
2005, while averaging $0.51 for Robustas. From January 2003 to October 2005, the price 
differential between Robusta and Other Mild Arabicas grew from 37.2% to 54.8% (ICO 
2005d), although a drought in Vietnam has begun pushing Robusta prices upwards again 
(Osorio 2005a). In September of 2005, Brazil announced the lowest level of government 
and private stocks ever recorded for that country. Catastrophic climatic events in the 
Gulf Coast region (where a quarter of U.S. green coffee stocks were stored), Mexico, and 
Central America are also likely to impact prices in the near future (Osorio 2005b). While 
the rebound in prices in 2005 certainly represents an improvement in the position of 
many coffee farmers, the ICO indicator price of $0.90 still represents only half of the 
historical average (UNFAO 2005), and the coffee crisis is far from over (Osorio 2005c). 

The remainder of this study will focus on the effects of the crisis through 2003, 
the year in which a local case study was conducted in southern Costa Rica. In that year, 
the ICO composite price averaged $0.52 per pound, with Other Milds, the type of coffee 
produced in Costa Rica, averaging $0.64 per pound (ICO 2005d). 



57 



II. Social, Economic and Environmental Effects of the Crisis 

To understand the effects of a prolonged price slump such as the one we are 
witnessing today, it is important to understand the responses of farmers and the 
governments of producing countries to price swings. The long-term development of a 
nation dependent on the production of a raw material depends on investing profits during 
good years in the diversification of the economy (Brignoli 1994 p. 30). The typical 
response to a spike in prices such as that which occurred in the late 1970s (see Figure 
1.3.6) is a syndrome known as "Dutch Disease" (Edwards 1984; Bevan et al. 1990; both 
cited in Pelupessy 1993 p. 26). Instead of using the windfall to improve infrastructure, 
diversify production or make other "rainy day" provisions, countries and individuals tend 
to respond by increasing consumption, especially of imports, which in turn drives 
increasing inflation, leaving producers even more dependent on coffee and more 
vulnerable to the inevitable downswings in the market. 

In turn, when prices drop, the first response of most farmers is to cut costs, 
starting with labor intensive activities and investments in fertilizers and pesticides, up to 
and including complete abandonment of the plantation except for the harvest (Pelupessy 
1993). In regions where most labor is performed by paid workers, this has immediate 
effects on employment levels and long-term effects on the productivity of the land. In El 
Salvador, for every one percent that costs are cut, approximately two percent of 
productivity will be lost (Pelupessy 1993 p. 26). This practice is most pronounced in 
countries where coffee dominates the economy, there are few alternative land uses, most 
coffee is produced by small farmers, or structural adjustment programs call for increases 
in the production of export products (Pelupessy 1991). In Central America, outright 
abandonment of the land by small producers in times of crisis is very common (Pelupessy 
1993 p. 13). 

Another, paradoxical effect in many countries is that, in times of low prices, small 
farmers will often attempt to increase production in order to meet their basic financial 
needs. Because of the combination of this factor with those in the above paragraph, 
coffee supply tends to be unresponsive to prices. For example, short-term (up to one 



59 



year) supply elasticity is only 0.03 in Central America, rising to between 0.14 and 0.77 at 
six years or more (Pelupessy 1991). 

This chapter will examine the social and economic effects of the most recent 
coffee crisis across scales, ranging from a global perspective to the level of individual 
households, with a focus on the district of Agua Buena, in the canton of Coto Brus in 
Costa Rica. 

/. International 

Low coffee prices hit the entire economies of coffee-producing nations hard. 
Most coffee-producing nations are deeply in debt, and rely on exports for badly needed 
foreign exchange. Export taxes are often heavily relied upon to fund government 
operations. In the three years after the suspension of the International Coffee 
Agreement, from 1989 to 1992, the Central American countries lost an estimated $3 
billion in export earnings (Pelupessy 1993 p. 13). The current crisis has coincided with a 
wave of privatization of state functions in producing countries, making state intervention 
to assist struggling farmers far less reliable (Kutschbach 1994b p. 120). 

Some countries' exchange positions have been hit harder than others. In Brazil, 
for example, while coffee accounted for 50% of foreign exchange earnings at the 
beginning of the 1970's, by the early 90's its share had shrunk to four percent, mainly 
due to industrialization (Portillo 1993). Thus, Brazil's economy was considerably more 
shielded from the crisis than other countries, for which coffee was more dominant in the 
export economy during the same period. 

Consolidation of the industry has continued throughout the 1990's, giving large 
traders and roasters even greater control over pricing and supply. At the end of the 
1980's, eight trading houses controlled over half of world supply (Pelupessy & Tilburg 
1994 p. 247; EIU 1991 p. 94). In the five years following the suspension of the ICA, five 
major trading houses closed and three merged (Daviron 1994 pp. 66-67). By 2002, half 
of the world's coffee beans were controlled by five roasters: Kraft, Nestle, Procter & 
Gamble, Sara Lee, and the German roaster Tchibo (Gresser & Tickell 2002 p. 2). 



60 



The percentage of the price of a pound of coffee that goes to farmers has shrunk 
throughout the 1990's. In the 1980s, coffee farmers received around 25% of the 
worldwide export price of coffee, with a net profit after expenses of approximately 15% 
of the end cost to the consumer (Pelupessy 1991). According to Gabriel Silva, General 
Manager of the National Federation of Coffee Growers of Colombia, in the early 1990's 
a total of about a third of the retail price remained in producing countries, but that share 
had fallen to less than 12% by 2002 (U.S. House of Representatives 2002 p. 62). That 
same year Nestor Osorio, Director of the International Coffee Organization, told the 
Costa Rican newspaper La Republica that the percentage had fallen to less than 9% 
worldwide (Anonymous 2002). 

Pelupessy and Tilburg (1994 p. 247-248) cite the specific example of Holland to 
illustrate this phenomenon. Prior to 1989, Central American farmers received 25% of the 
retail price of coffee in Holland, with a net profit to the farmer equal to 15% of the retail 
price. By 1994, those figures had fallen to 14% and 10%, respectively. Additionally, 
when prices fall, producing countries tend to lower export taxes, thus further reducing the 
percentage of the retail price that remains in the producing country (Pelupessy 1991). 

There are a number of factors influencing how farmers and their land will be 
affected by low prices. The most important of these can be loosely categorized as 
relating to the three types of inputs mentioned in Chapter 1.3: land, labor and technology. 
Other factors influencing the effects of a price crisis that do not fall neatly into these three 
categories include: the level of state intervention, the level of indebtedness of the 
producer, the amount of coffee stocks held by a country, and efficiency of production 
(Pelupessy 1991; Kutschbach 1994b p. 120). 

Land 

Coffee farmers who enjoy optimal environmental conditions are able to better 
survive market downturns: in sub-optimal areas, coffee is only profitable when it 
commands high prices (Kutschbach 1994c p. 225). Coffee grown at higher altitudes, 
while commanding a higher price, also offers lower profit margins. (Kutschbach 1994b p. 
85). Farmers with more diversified production systems will face fewer risks but often 



61 



forego potentially higher incomes (Reeves & Lilieholm 1993; Kutschbach 1994b p. 81; 
Perfecto et al. 1996; Moguel & Toledo 1999; Gobbi 2000; Ramirez & Sosa 2000). 

In Latin America, the crisis has had the effect of forcing farmers to diversify their 
production (Kutschbach 1994c p. 226). Colombia, Ecuador, El Salvador and Guatemala 
began diversification programs in the early 1990's and implemented incentives to reduce 
coffee production (Murillo 1992). In Minas Gerais, Brazil, farmers began changing their 
land use from coffee to eucalyptus, nuts, passionfruit, or cattle ranching. The number of 
coffee trees in Brazil fell from 4.2 billion to 3.2 billion from 1990 to 1992, and of those 
remaining, only 20% were being tended regularly (Portillo 1993). In more remote areas 
the production of illegal crops, such as cocaine, expanded (Wilson 2001; cited in Philpott 
& Dietsch 2003; Kutschbach 1994c p. 235). Some authors have expressed concern that 
the crops replacing coffee may be less appropriate to the steep slopes where coffee is 
frequently grown, resulting in increased soil erosion. Where shade coffee is grown, 
changing to other crops can result in the clearing of the shade overstory (Varangis et al. 
2003 p. 18). 

Farmers are also leaving the land altogether, either by choice or because of debt. 
In Guatemala, as many as 20% of farmers had lost their land by mid- 1992, and many 
Mexican farmers began migrating to the United States (Murillo 1992). In Peru, where the 
cost of production was $0.68 per pound and the price paid to farmers $0.27, an estimated 
25,000 hectares of coffee had been abandoned by 2001. Many of those farmers turned to 
coca, which paid three times as much as coffee. Governments in many Latin American 
countries have been intervening in a number of ways to support production (Prendas 
2001a). 

Productive coffee area dropped sharply after the Brazilian frost and other natural 
disasters of 1976, reaching a low of 7.9 million ha, then rebounded quickly (Figure 
2.1.1). Productive coffee area peaked in 1990 at 11.4 million ha, then fell 12.3% by 
1995, to a 15-year low of 10.0 million ha. After the price recovery of the late 90's, 
productive area increased slightly again, peaking at 10.6 million ha. As of 2004, it had 
dropped back to 10.1, just above the 1995 low (data from UNFAO 2005). 



62 



The total coffee-producing area for Asia, Africa, and Latin America and the 
Caribbean are shown in Figure 2.1.2. Productive area in Africa has been in steady 
decline since 1990, dropping by 42.6% by 2002, and area in Latin America declined 
17.6% from 1989-1995, then stabilized. Meanwhile, productive area in Asia increased by 
55.7% from 1992 (the earliest year for which the UNFAO offers data) to 2004. 

Labor 

Coffee farmers with only a few hectares of land generally rely entirely on family 
members for labor. Since family labor is unpaid, these farmers can tolerate lower prices 
and turn a profit even in marginal areas. In the case of larger farms that use paid laborers, 
lower coffee prices will tend to depress wages (Kutschbach 1994b pp. 89-90). Due to the 
seasonality of labor needs, even in good years the families of a large percentage of the 
day laborers on the world's coffee plantations lack basic nutritional requirements for five 
months out of every year (Pelupessy 1993 p. 17). Thus, the effects of a downturn in 
coffee prices are devastating for wage laborers in the coffee sector. 

By 2001, according to the Costa Rican Chamber of Coffee Growers, farmers in El 
Salvador, Honduras and Mexico were not harvesting their coffee because it cost more to 
pay pickers then they would receive from their coffee. In Nicaragua, only 1,800 of 
40,000 farmers could afford to finance their harvests (Prendas 2001a; 2002). By 2002, 
the Costa Rican daily La Republica reported that half a million workers in Central 
America and Mexico alone had lost jobs (Anonymous 2002). In 2003, the World Bank 
estimated that 190,000 permanent jobs and 350,000 seasonal jobs in coffee production in 
Guatemala, Honduras, El Salvador and Costa Rica had been lost (Varangis et al. 2003 p. 
17). 

Many people in coffee-growing areas have responded to the reduction in 
employment and the low coffee prices by migrating to urban areas and to more developed 
countries. In May of 2001, the bodies of fourteen migrants were found dead on the U.S.- 
Mexican border, six of whom were identified as coffee workers (U.S. Rep. William 
Delahunt, in U.S. House of Representatives 2002 p. 9). 



63 



Technology 

Of course, as a compromise between the typically intensive regime of inputs 
prominent during years of high prices, and complete abandonment or neglect of coffee 
farms, during price slumps a farmer can to continue to maintain and harvest from their 
coffee trees while lowering expensive inputs. At a Congressional hearing on the coffee 
crisis in 2002, Franklin Lee of the U.S. Department Of Agriculture testified that farmers 
worldwide were reducing fertilizer and pesticide inputs (U.S. House of Representatives 
2002 p. 43). 

Hard data on use of chemical inputs in coffee-growing regions worldwide are 
extremely limited, and do not allow for discussion on a global scale. Data from the UN 
Food and Agriculture Organization (2005), for example, show that fertilizer use in the 
countries that were the top 10 coffee producers in 1989 increased by 63% from 1989 to 
2003, but it is impossible to know on which crops these fertilizers were used. Changes in 
technology use are, therefore, reserved for discussion on a local scale. 



Deputy Administrator for Commodity and Marketing Programs, Foreign Agricultural Service, USDA 
64 



Figures 



12 





11.5 




11 
10.5 


u 


10 


o 


9.5 


w 

c 
o 


9 
8.5 


ei 


8 



7.5 



World Coffee Area 1961-2004 






CO 


m 


l^ 


CD 




CO 


m 


N. 


CD 




CO 


10 


l^ 


CD 




CO 


m 


r- 


CD 




CO 


«) 


CD 


CD 


CD 


CD 


r~- 


N. 


N. 


r^ 


N. 


ai 


co 


co 


co 


co 


CD 


CD 


CD 


CD 


CD 


O 


O 


0) 


CD 


CD 


CD 


CD 


CD 


CD 


o> 


o> 


CD 


CD 


CD 


CD 


CD 


CD 


CD 


CD 


CD 


CD 


CD 


O 
CM 


O 
CM 



Figure 2.1.1. Area producing coffee worldwide from 1961-2004, in millions of hectares. Data from 
UNFAO (2005). 



8.00 -i 



7.00 



(0 


6.00 


o 


5.00 


<*— 
O 


4.00 


c 


3.00 



2.00 



1.00 



0.00 



Coffee Area by Region 1961-2004 



Latin America & 
Caribbean 




Africa 




Figure 2.1.2. Area producing coffee by region, 1961-2004, in millions of hectares. Data from UNFAO 
(2005). 



65 



2. Costa Rica 

In Costa Rica, the coffee/banana crisis of the late 1920's and the Great Depression 
did not result, as in other Latin American countries, in violent upheavals. Instead, 
structural and institutional solutions were sought to stabilize the economy and preserve 
order. The result of this period was an enlargement and strengthening of the role of the 
Costa Rican state, a role which continued throughout the 20 th century (Kutschbach 1994b 
p. 117). This active participation by the state may have shielded Costa Rican producers 
from some of the worst blows of the current crisis, but farmers are still enduring great 
suffering and economic hardship. 

Institutional Reponses 

In early 1992 the Costa Rican government reduced the export tax on coffee by 
one percentage point, and reduced from 1% to 0.5% the tax paid by producers for the 
support of ICAFE (Briceno 1992). That May, the government suspended exports for ten 
days and called upon neighboring countries to join, but the move had no effect on prices 
(Murillo 1992). A $50 million National Coffee Stabilization Fund (FONECAFE) was 
created, and distributed to farmers as loans to be paid back when prices recovered 
(ICAFE 1993 p. 74). 

By 1997, the Costa Rican farmers had paid back the $50 million in loans, plus an 
additional $10 million, which was to be held in reserve for future crises (Munoz N. 
1997). In 2001, FONECAFE paid out another $75 million in loans to growers, leaving 
the fund depleted (Ulett 2001). The first $25 million disbursement amounted to a subsidy 
of about 6 cents per pound, with over 7,000 farmers receiving less than $10 each (Mora 
2001a). Two additional $25 million disbursements were made in 2001 (Ulett 2001). The 
loans were for a term of ten years, at eight percent interest, to be paid yearly beginning in 
2003 (Prendas 2001b). 

The Costa Rican government also began to push increased productivity as a way 
out of the crisis (Anonymous 1992a), because, when prices were lower, many farmers 
believed they could continue to meet their basic needs only by selling more coffee 
(Alfaro 1992). At 1993 prices, it was estimated that a farmer needed to yield 1800kg per 



67 



hectare to cover production costs, or about 300 kg more than the national average at the 
time. Programs were initiated to find ways to improve yields without raising input costs, 
such as through new pruning regimes (Vargas 1993). 

Squeezed by declining prices and harvests, and by the lack of available credit, 
four of Costa Rica's beneficios were forced to close between 2001 and 2002. Other, 
smaller beneficios had also opened to process specialty coffees (Barquero S. 2002). 
FEDECOOP purchased a roaster in Miami, in an effort to circumvent tariffs imposed on 
roasted coffee (Ulate 1997). To secure higher prices, some beneficios were selling 
directly to U.S. roasters (Ulett 2001). 

Farmer Responses 

Even in the first few years of the current crisis, Costa Rican farmers who had 
lived through previous ones reported that this one was the worst they had experienced 
(Hilje et al. 1994 pp. 165-166). The savings and loan crisis of the late 1980s had wiped 
out the life savings of many, leaving them with few resources to withstand low prices 
(Hilje et al. 1994 p. 182). While devaluation of the colon could have lessened the impact 
of falling prices to some degree, the policies of the Central Bank prevented the full 
benefits of this from being passed on to the producer (see also Chapter 1.3) (Vega M. 
1992). 

Land 

The crisis has, to some degree, encouraged diversification. The interplanting of 

citrus and plantain among the coffee bushes, which lost popularity as more farmers 
adopted the technified model, has re-emerged as low prices continue. In 1992, 
FEDECOOP proposed the creation of a credit program to help farmers below 900 meters 
to diversify (Viquez 1992a). Unfortunately, for many, "diversification" means a shift to 
cattle ranching (Dicum & Luttinger 1999, pers. obs. 2003). 

Other strategies used by farmers are to abandon coffee and plant something else 
or, when the farmer is indebted or without other options, to sell the land and seek work as 
a laborer (Hilje et al. 1994 p. 184), especially in banana plantations (Anonymous 1992a). 
By mid-2001, the Costa Rican National Chamber of Coffee predicted that 60-70% of 



68 



coffee farmers would eventually abandon the crop altogether. Abandonment of the land 
provokes migration to the cities, expanding what are known as "rings of misery" (Mora 
2001a). 

Labor 

During interviews of Costa Rican coffee farmers conducted in the early 1990's 

regarding the coffee crisis, most farmers agreed that the largest coffee farm one 
individual person can care for year-round, not including additional labor needed during 
peak periods such as the harvest, is about 3.5 hectares, beyond which it becomes 
necessary to hire additional permanent labor. Labor in Costa Rica is expensive in 
comparison to other coffee-producing nations. This is most keenly felt by small and 
medium producers, who cannot guarantee year-round work for laborers (Hilje et al. 1994 
pp. 168-171). As farmers cut costs, work has become more scarce in coffee regions, 
especially for pickers (Hilje et al. 1994 p. 185; Kutschbach 1994c p. 227). In some cases, 
farmers will not even harvest their own coffee, because they can't afford to pay the 
pickers, or they earn more in wages from picking others' coffee than from selling their 
own (Hilje et al. 1994 p. 196). 

By 2001, unemployment had become a major problem in coffee-growing areas 
(Mora 2001a). Many farmers stopped pruning shade trees because of labor costs (Brenes 
2002). Between the 1999-2000 and 2000-01 harvests, the number of permanent coffee 
workers fell 2.5-3.4% and the number of pickers by 18%. The total number of work days 
fell 12.3%, at an average value of $7.57 per day. The total reduction in employment was 
equivalent to an annual $84 per hectare (Valverde 2002). Between 1992 and 1996, the 
proportion of the Costa Rican workforce employed by coffee declined from 6.1% to 4.9% 
(ICAFE 1993 p. 91; 1997 p. 85). 

Wages for pickers stagnated at 300 colones per cajuela from 2000 through 2003 
(Prendas 2001b, pers. obs.), which actually represented a 22% drop in real terms (2003 
dollars). It takes about 90 minutes for an experienced picker to fill a cajuela (Brenes 
2002), making the hourly rate approximately $0.50 an hour in 2003. These low wages, 
and a lack of permanent employment in the coffee industry, created a paradoxical labor 
shortage during the harvest even as year-round unemployment increased. Farmers were 



69 



forced to allow pickers to include green berries in the harvest in order to attract workers, 
which lowered the quality of their coffee (Prendas 2001b). In 2001, the ICAFE asked 
parents to send their children to work in the fields after school. As many as 5,000 Mt of 
coffee, or 3.3% of Costa Rica's harvest, were lost that year due to labor shortages 
(Quesada & Marin 2001; Fernandez et al. 2002). 

Technology 

The level of technology used in Costa Rica is quite high relative to other coffee 

producing countries (Kutschbach 1994c p. 231). This has given Costa Rica the highest 
coffee yields in the world, but has also dramatically increased production costs, and thus 
made the current crisis more difficult to endure (Hilje et al. 1994 p. 167). In the late 
1990's, approximately 30. 6g of the herbicide Paraquat, 58. Og of the fertilizer ammonium 
nitrate, and 6.0g of the fungicide copper hydroxide were released into the environment 
for every pound of Costa Rican coffee produced (Pelupessy 1997 p. 38). Using an 
average yield of 3,000 pounds per hectare, this is the equivalent of 92kg of Paraquat, 
174kg of nitrate, and 18kg of copper per hectare. 

In light of this heavy use of inputs, therefore, the most immediate response of 
farmers has been the reduction in agrochemical use (Hilje et al. 1994 p. 183). Between 
1990 and 1995, the number of coffee producers in Costa Rica who applied fertilizer to 
their farms fell from 93.5% to 85.8% (ICAFE 1997). Despite this apparently small 
reduction, however, it is clear that many farmers who continued applying some fertilizers 
reduced the overall amount of nutrients applied. Between the 1988 and 1992 harvests, 
the average amount of nitrogen, phosphorus, potassium, magnesium and boron applied 
per hectare of coffee fell by almost half (ICAFE 1993 p. 66). As one farmer reported, "if 
we fertilize we don't eat, and if we eat we don't fertilize" (Hilje et al. 1994 p. 184). 

Overall, the consequences of the decline in fertilizer use on coffee yields were 
dramatic: some claimed that, in some areas, yields declined by nearly 90% (Viquez 
1992a). Many also contend that reduction in chemical inputs tends to reduce the quality 
of coffee. Eduardo Kooper of the Costa Rican Association of Fine Coffees reported in 
1997 that the quality of Costa Rican coffee had gradually declined over the previous 10 
years (Ulate 1997). In 1998, ICAFE began to implement a program encouraging farmers 



70 



not to neglect their fields (Barquero S. 1998). Reductions in inputs continued, however: 
between the 1999-2000 and 2000-01 harvests, total inputs declined 21.9%, or $101 per 
hectare (Valverde 2002). 

The reduction of chemical inputs in coffee does not seem to have had any 
important effect on Costa Rica's total chemical use. In 2001, a study by the Costa Rican 
Ministry of Agriculture (MAG) revealed that imports of agrochemicals and their active 
components had remained level for four years (1997-2000) (Mora 2001b). During the 
same period, the area of Costa Rica devoted to crops increased by 17,951 ha, or 4.1% of 
the area in 1996 (SEPSA 2002 p. 18). Thus, the total amount of chemicals used per 
hectare fell by approximately 5% during those four years. 

The combination of lower chemical and labor inputs, and a reduction in the land 
area devoted to the crop, have resulted in a nationwide decline in total coffee production. 
This decline has been most dramatic since the turn of the century. The 1999-2000 
harvest cycle saw the third largest harvest in Costa Rican history, at 162,600 Mt. By the 
2003-04 harvest, production had dropped to 110,136 Mt, the lowest since 1986 
(calculated from ICAFE 1993; 2005b; Fernandez et al. 2000). 

3. Coto Brus 

Coto Brus is the easternmost canton of the southern Costa Rican province of 
Puntarenas, with a land area of 935 km , divided into four districts (Rojas 1988). It is one 
of the most recently-settled cantons in Costa Rica (see Manger (1992) for a detailed 
description of the geography and historical background of Coto Brus). Until the early 
1990's, Coto Brus was one of the four fastest-growing cantons in the country (Viquez 
1992b), and its population more than doubled between the 1974 and 2000 censuses 
(DGEC 1974; INEC 2001). In the 2000 census, Coto Brus had a population of 40,082. 
The population was nine percent urban, with a literacy rate of 91.4%, a mean educational 
level of 5.5 years, and 22.9% of residents with a secondary education or higher, all lower 
than the nation as a whole (INEC 2001). 

Although there has recently been a relative increase in tourism and industry in 
Coto Brus (M. Rosemeyer, pers. comm.), the boom experienced by the rest of the country 



71 



over the last decade has, for the most part, not reached this remote area, which still 
remains relatively isolated and undeveloped. There is a sense among many of the 
population of neglect and abandonment by the politicians in the capital, giving Coto Brus 
the nickname "the forgotten canton" (Arguedas M. 1988, Leonel Sanchez,* pers. comm. 
2003). Farms in the region are less diversified and mechanized than the rest of the 
country. The canton receives fewer services, so farmers must do for themselves things, 
such as road maintenance, that would normally be provided by the government (L. 
Sanchez, pers. comm. 2003). In 2003 the Branca region of Costa Rica, which includes 
Coto Brus, had the highest number of poor households in the country, with 33.6% of 
homes in poverty, 21.5% unable to satisfy basic needs, and 12.1% in extreme poverty 
(INEC 2003b). 

The first coffee harvest took place in Coto Brus in 1949, and the first beneficio 
was constructed there in 1952 (Manger 1992 p. 70). As of 2001, Coto Brus had 11,633 
hectares of coffee and eight beneficios (Fernandez et al. 2002 p. 17, 56), three of which 
were cooperatives, the latter of which processed 60% of the canton's coffee (Rojas 
1988). By 2004, the canton had 11 beneficios, four of which were cooperatives 
(Fernandez & Alvarez 2004). The three new beneficios each processed less than 25,000 
fanegas of coffee, with two of them processing less than 1 ,000 fanegas each in the 2003- 
04 harvest (ICAFE 2005a). One of these latter two was a women's processing and 
roasting cooperative called the Asociacion de Mujeres Organizadas de Biolley (Montoya 
2003), and the other was an estate called Cafetales Lila, which processed and roasted its 
own coffee (ICAFE 2005c). 

According to Arnoldo Lopez, president of ICAFE in 1992, Coto Brus is unique in 
that its coffee is the first in the country to ripen, the industry is composed almost entirely 
of small farmers with an average of 4 ha of coffee each, and the canton is almost entirely 
dependent on coffee activity (Viquez 1992b). In 1984, more than 82% of the canton's 
3,179 farms grew coffee, and coffee accounted for 99.8% of the area planted in 
permanent crops (DGEC 1987). In the past, many children have had to leave school 



General Director of Coopabuena, a coffee processing cooperative in Agua Buena, 2003. 
72 



during the harvest, contributing to the area's relatively low educational level (Zeledon 
1987). 

Coffee farming in Coto Brus suffers from a high incidence of fungal diseases, and 
relatively poor dissemination of technical information to coffee growers (Rojas 1988). 
As a result, yields in Coto Brus are the second lowest in Costa Rica (after Perez Zeledon), 
averaging 24.8 fanegas per hectare from 2000-02, or 13.6% less than the national average 
of 28.72 fanegas per hectare (Fernandez et al. 2002 p. 22). Processing yields are the third 
lowest of Costa Rica's eight coffee regions, averaging 90.8 pounds of green coffee per 
fanega during the 2003-04 harvest, 4.1% lower than the national average of 94.81 
(calculated from Fernandez & Alvarez 2004). For the 2001-02 harvest, Coto Brus ranked 
third of seven coffee-growing regions in terms of total production (Fernandez et al. 2002 
p. 23). 

The price paid to farmers for ripe coffee cherries is also the second lowest in the 
country, averaging $48.35 per fanega for conventional coffee, and ranging from $34.06 
to $92.10, influenced in part by altitude and whether the coffee berries are ripe, 
(calculated from ICAFE 2005c), compared to a mean of $59.39 nationally (calculated 
from Fernandez & Alvarez 2004). The lower price can partly be explained by the low 
processing yields. While the price paid to farmers per fanega is 18.6% lower in Coto 
Brus than nationally, the price with processing yields taken into account is only 7.9% 
lower, or $0.58 per pound of green coffee in Coto Brus compared to $0.63 per pound 
nationally. Another reason for the lower prices is that the roads in the southern zone of 
Costa Rica are the worst in the country (Quesada & Marin 2001), making it more 
difficult, and expensive, to bring products to market (L. Sanchez, pers. comm.). 

Gilberto Gutierrez, director of FEDECOOP in 1992, reported that Coto Brus was 
one of four cantons in Costa Rica most hard hit by the current crisis (the other three were 
San Carlos, Sarapiqui and Turrialba), primarily as a result of the forgoing factors. 
Compounding matters, in October 2001, Hurricane Michelle made many roads 
impassable and caused recibidores to shut down, resulting in a loss of 30% of the harvest 
that year (Quesada & Marin 2001). 



73 



Land 

The reduction of coffee area by the turn of the century was most severe in Coto 
Brus, Perez Zeledon and Turrialba (Brenes 2002). Gilberto Gutierrez had estimated that, 
as early as July 1992, as many as half of Coto Brus coffee farmers had left their farms, 
primarily to migrate to banana-growing regions (Viquez 1992a). Ten years later Henry 
Rojas, director of ICAFE for the Branca region, reported that 1,089 farms in the southern 
zone had been abandoned, and 400 converted to other uses. Rojas said that most of these 
conversions had been to cattle pasture, as farmers had little money to invest (Valverde 
2002). Sugar was another popular choice for conversion (Brenes 2002). 

Labor 

Despite the abandonment of coffee farms and the changeover to other crops, the 
net rate of migration into Coto Brus remained positive throughout the 1990's, albeit at a 
much lower level than in earlier decades. The net annual rate of migration into Coto Brus 
was 10.3% for the period from 1968 and 1973, dropping to 2.4% from 1979 to 1984, and 
to 1.5% from 1995 to 2000. The population of the Southern Zone (the cantons of Perez 
Zeledon, Buenos Aires and Coto Brus) reached its peak as a percentage of Costa Rica's 
population in 1973, at 11%, before declining to 9% by 2000 (Barrantes & Pana 2005). 

Technology 

The reduction of inputs was more dramatic in Coto Bras than elsewhere. 
Between 1990 and 1995, the number of coffee producers in Coto Brus who applied 
fertilizer to their farms fell from 93.3% to 66.7%, a drop that was three times the national 
average (ICAFE 1993 p. 66). 

In 2003, 1 obtained a list of all major agrochemical suppliers was from the ICAFE 
office in San Vito, and provided each supplier with a questionnaire regarding 
agrochemical sales over the past 10 years. Only two suppliers responded to the 
questionnaire. Both suppliers who replied were subsequently interviewed in person to 
obtain clarification and elaboration of their answers. 

One supplier - the owner of a store in Canas Gordas, on the border with Panama, 
had only been in business for Wi years. He reported that the majority of his sales were to 



74 



farmers in Panama, and that sales had increased drastically after the mid-2003 closure of 
the supply store operated in nearby Coopabuena (two months before the interview). 
Locals reported that the Coopabuena supply store had closed due to financial insolvency. 
Because of the Canas Gordas store's unique situation, and the relatively short period of 
time that it had been open, the results of that store's questionnaire cannot be considered 
typical and are not included here. 

The second vendor owned a major agricultural supply store based in San Vito. 
He reported that the past several years had been "brutal." From gross sales of $1.4 
million in 1997 (reported as constant 2003 dollars), revenues dropped more than 74% 
over five years, to $352,000 in 2002. Figure 2.3.1 shows his reported gross chemical 
sales from 1993-2002, plotted with coffee prices over the same period. He reported that 
sales of those fungicides, fertilizers and nematicides normally used in coffee had declined 
dramatically, while fertilizers and fungicides normally associated with annual vegetable 
crops, such as green pepper and tomato, had been increasing. Sales of herbicides had 
remained steady, with glyphosate gradually replacing Paraquat, and sales of Tordon on 
the increase. Herbicides are used with both coffee and annual crops in Coto Brus, but are 
also used in cattle pastures. Glyphosate is often applied after land clearing, to prepare the 
ground for pasture grasses, and Tordon is used to eliminate broadleaf weeds. 

For this vendor at least, gross annual chemical sales were closely correlated with 
the price of coffee paid to farmers in the same year (N=10, r=0.719, p=0.019). Although 
no other suppliers responded to the questionnaire, anecdotal reports and personal 
observations support the inference that this San Vito supplier's experience is typical of 
what was happening to other agrochemical suppliers in the canton. No information was 
obtained regarding relative quantities of chemicals sold. 



2,4-D Picloram 



75 



Figures 



$1,500 




r $1.40 



$1.20 



CO 

o 
o 

CM 



$1.00 



$0.80 



$0.60 



$0.40 



Q. 
V) 

o 



$0.20 



Gross chemical 
sales 

-Coffee prices 
(producer) 



Figure 2.3.1. Gross sales, in 2003 US dollars, of agrochemicals at a major supply store in San Vito, Coto 
Brus, from 1993-2002, and average prices paid to farmers in Costa Rica, expressed as 2003 US dollars per 
pound of green coffee equivalent. Data from ICAFE (1997), ICAFE (2005c), Ferndandez et al. (2002), and 
ICO (2005b). Inflation data from U.S. Dept. of Labor (2004). 



77 



4. Agua Buena and Individual Households 

Agua Buena is the southernmost and smallest district in the canton of Coto Bras, 
encompassing 6,118 hectares (Sandoval 2002) on the border with Panama. The 
population of Agua Buena nearly doubled between 1973 and 1984, climbing 72.9% from 
3,729 to 6,446 residents, at which point the population nearly leveled out, rising only 8% 
between 1984 and 2000, reaching a population of 6,962 (DGEC 1974; MGP 1986; INEC 
2001). The district is considered by the Costa Rican Institute of Statistics and Censuses 
to be 100% rural (INEC 2001). 

Literacy and educational rates in Agua Buena have historically been slightly 
lower than the rest of the canton, but had caught up with or exceeded the canton-wide, 
average by 2000 (MGP 1986; DGEC 1974; INEC 2001). Comparative income data were 
only found for 1973, when the median annual income was 7.9% below the canton 
average, and 42.2% below the national average, at $3,350 (expressed as 2003 dollars) 
(DGEC 1974). In 2000, 54.75% of homes in Agua Buena had at least one area of critical 
shortage, compared to 36% for the country as a whole. Shortage areas, as defined by the 
Instituto Nacional de Estadistica y Censos, include access to adequate shelter, food, 
potable water, sanitation, clothing, and primary and secondary education (Solorzano & 
Fonseca 2005). 

The area has been converted in the past 50 years from mostly undisturbed forest 
to a patchwork of small farms (Manger 1992). Currently, small patches of forest remain 
throughout the area, primarily around streams, and living fences are the preferred method 
of marking boundaries between properties or cultivation areas. Like the rest of Coto 
Bras, Agua Buena has historically been economically dependent on coffee production. In 
1984, 1,134 hectares, or 28.6% of Agua Buena's 3,969 hectares of farmland, were 
devoted to coffee (DGEC 1987). In 1984, another 44.8%, or 1,778 hectares, of Agua 
Buena's farmland was devoted to cattle pasture, supporting 2,960 cows (DGEC 1987). 
By 2000, the area planted with coffee had decreased 15.7%, to 956 hectares (Sandoval 
2002).* 



No data on extent of cattle pastures are currently available for 2000. 

79 



In 2003, most of the recibidores in the District belonged to Paso Real, a private 
beneficio, or Cooperativa Agua Buena ("Coopabuena"), the area's coffee cooperative. 
During the 2003-04 harvest, Coopabuena paid an average of $53.70 per fanega; with 
processing yields of 89.57 pounds/fanega (Fernandez & Alvarez 2004). This represents 
$59.95 per 100 pounds of green coffee (known as a quintal), and is 13% above the 
canton-wide average of $52.93 per fanega, and 1% above the national average of $59.39. 
The higher prices paid to Coopabuena farmers may be a result of the cooperative's 
participation in two fair trade programs (Coocafe and a direct marketing program), 
through which, according to former U.S. interns at Coopabuena, Coopabuena marketed 
approximately 10% of its coffee from the 2003-04 harvest (Grosser et al. 2005). 

Staff at Coopabuena reported having repossessed approximately ten farms over 
the past ten years, of which seven were repossessed as a result of bad debt. When asked 
about changing population levels in Agua Buena, they responded that the population was 
"a little bit critical," due to the price of coffee and the devaluation of the national 
currency. The staff expressed fears that Coopabuena and environs could become a "ghost 
town" due to the declining interest in coffee farming. These fears appear to have been 
well-founded, as Coopabuena closed its doors after the 2003-04 harvest (Barquero S. 
2004e; 2005; ICAFE 2005c). 

Observations of the landscape in and around Agua Buena, and informal 
discussions with farmers, led to the formulation of two main hypotheses to be tested to 
examine the effects of the coffee crisis in Agua Buena from the socioeconomic and 
environmental perspectives: 

HI. Coffee farmers faced with low prices are choosing to convert their farms to 
other uses, and the predominant conversion choice is cattle pasture. 

H2. Coffee farms that had been converted to cattle pasture would show a lower 
bird diversity and richness than coffee farms that had not been converted. 

The methods and results for each hypothesis are explained and discussed 
separately below. 



80 



HI: Individual farmer decisions 

Hypothesis: Coffee farmers faced with low prices are choosing to convert their farms to 
other uses, and the predominant conversion choice is cattle pasture. 

Methods 

Sixty farmers in Agua Buena were interviewed to assess land use decision- 
making and related economic factors. The interviews were carried out in respondents' 
homes by one of two interviewers, with the exception of one respondent who, at the 
respondent's request, was given the question sheet for review prior to the interview, and 
returned later with written answers to the interview questions. A pilot survey instrument 
was administered to three respondents, then modified slightly, with additional questions, 
for the remainder of the respondents. Follow-up interviews were conducted with two of 
the three pilot respondents to collect additional data that had been added to the final 
survey (available in Appendix A). In addition to questions about land use and farm 
practices, respondents were asked questions about socio-economic background and 
environmental perceptions. After the main interview, farmers were also asked to answer 
additional questions regarding costs of production. Four farmers agreed to answer these 
additional questions. 

Interviews were conducted in the immediate environs of Agua Buena, in the 
center of town and following roads out as far as 2.5 km from the main plaza, in the 
directions of the towns of Campo Tres and Canas Gordas and the barrios of San Miguel 
and Santa Marta. Interviews were solicited by knocking on doors or approaching farmers 
in the field. The interviews were then either carried out immediately or scheduled for a 
time convenient to the farmer. If no one was home at the first visit, the house would be 
re-visited once to try and find someone at home. All houses in the survey area were 
approached, but only farmers who currently or formerly grew coffee and were the 
primary decision-makers for the land they cultivated were interviewed. No one who was 
found at home refused to answer these two initial qualifying questions. Three farmers 
who met the criteria for the interviews declined to be interviewed, for a response rate of 
95.2%. 

Excel 2003 was used to calculate totals, means, and ranges of data. SPSS 10.0.5 
was used to calculate medians and modes and to perform /-tests, Pearson's r, x an d 



81 



Fisher's Exact Test. For /-tests, where Levene's test for equality of variances was 
significant at p<0.05, the t statistic with equal variances not assumed was used. In all 
statistical tests, a was tested at 0.05. 

One interview was excluded from the analysis because the respondent was an 
employed caretaker for an absentee landowner, and was not the primary decision-maker 
for the farm. Additionally, six other questions are partially or wholly excluded from the 
analysis, due to possible interviewer bias or other confounding variables (for an 
explanation see Appendix A). 

Results and Discussion 

Demographic, social and economic background of sample 

See Table 2.4.1, Table 2.4.2, and Table 2.4.3 for summaries of the survey 

responses. The mean household size was 3.9 people. Seven (12%) of the respondents 

were female, and 52 (88%) were male. Only 6.8% of respondents were single, with 

81.4% married, and 11.9% widowed or divorced. On average, respondents had lived the 

region for 28.3 years, with 1975 the median year for arriving, by immigration or birth, in 

the region. The mean number of years of schooling was four, with the median being 

three years (the mean was influenced by one outlier, a single respondent who had 

attended university). 

All but one of the respondents owned their land outright. The total land area 
owned was 501.65 ha, or 99.9%, and the total area rented was 0.7 ha, or 0.1%. Two- 
thirds of survey respondents had purchased their land, while just over a quarter had 
inherited it. One respondent acquired their land by trade, one acquired it by marriage, 
and another acquired it through a government program to encourage settlement in the 
region. 

Most respondents did not know their annual income. To arrive at an estimate, the 
interviewer would ask questions regarding specific sources of income and amounts 
received from each, including government pensions and remittances from family 
members, and combine those data with information provided regarding the latest coffee 
harvest. The annual household income for the sample, calculated based on this data, 
averaged $4,133, or $1,060 per capita, with a median of $1,784 and a range from $360 to 



82 



$26,500. Respondents reported ten categories of additional, non-coffee income, of which 
six were on-farm and four were off-farm. In all, nineteen respondents (32.2%) cited 
additional on-farm sources of income, and 37 (62.7%) reported off-farm sources, with 
seven (11.9%) reporting both. Of those farmers who still grew it, coffee accounted for, 
on average, 52.9% of total income. Nine respondents, or 15.3%, of respondents reported 
using credit to pay the costs of production. 

Farming practices and coffee production 

The total land area managed by the respondents was 502 hectares. In total, the 

survey respondents reported producing a total of 39 different products on their farms, 

with a mean of 5.4 products per farm (a full list of farm products reported is available in 

Table A. 3 in Appendix A). Coffee growing was significantly correlated with the total 

number of products produced on the farm, with farmers who still grew coffee producing, 

on average, 6.1 products, and farmers who no longer had coffee producing an average of 

3.5 products (?=3.047, df=50.195, p=0.004). 

Most farmers did not belong to agricultural organizations, with 21 farmers 
(35.6%) reporting membership in a coffee cooperative and three (5.1%) reporting 
membership in another type of agricultural organization (a dairy marketing association, a 
plantain association, and a regional agricultural association). Of those who reported 
membership in an organization, the mean number of meetings attended per year was 2.63, 
with a median of one meeting per year (usually their cooperative's annual assembly). 

The mean number of workers on respondents' farms outside the harvest was 1.4, 
with a mean of 1 . 1 family member (usually the respondent him- or herself or a son) and 
0.3 paid workers. Most farms only had one worker: the respondent. The maximum 
number of year-round workers on a respondents' farm was six, with four paid workers. 

Thirty-three respondents (55.9%) reported currently using various agrochemicals. 
More information on the types of chemicals used is available in Table 2.4.2. Many of the 
other respondents reported that they used chemicals in the past. Of the 41 farmers who 



83 



still grew coffee, eight of them, or 19.5%, planned to convert to growing organic or 
"sustainable" coffee. 

The mean price received at the last coffee harvest was $35.59 per fanega, ranging 
as low as $26.88 and as high as $48.87 - all values well below the mean of $66.05 
reported as the price that coffee would need to reach to be profitable, and the mean of 
$82.03 that it would need to be worthwhile to grow. Yields ranged from 2.5 to 64.0 
fanegas per hectare, with a mean of 16.8 and a median of 15.0. The mean total harvest 
was 43.0 fanegas, resulting in a mean gross income from the last coffee harvest of 
$1,530. A plurality of sixteen respondents (27.1%) reported 1997 or 1998 as the year of 
the highest price they could remember. This corresponds with the peak prices for the 
1990's, but only six respondents (10.2%) reported a year prior to the expiration of the 
International Coffee Agreement, and only one respondent reported 1977, which saw the 
highest worldwide export prices in at least the past 40 years. 

Four farmers responded to an additional survey regarding costs of production. 
This survey included information about labor costs, input costs, replanting costs and 
property taxes, and did not assign a cost to labor provided by the respondent or his or her 
family - i.e., only paid labor was given a cost. Unlike the ICAFE costs of production 
model (Portilla R. & Araya M. 2002), the survey instrument did not take into account 
depreciation and replacement costs of items such as tools, equipment and buildings. 

Mean production costs for these four respondents were $18.53 per fane ga, or 
$280.09 per hectare. Given the prices that these respondents reported for their last 
harvest, the mean profit came to $18.34 per fanega. With the yields reported by these 
respondents, the mean profit per hectare was $276.96, for a total mean net income (taking 
into account farm size and total harvest) of $588.59 (gross income $1,425.63, expenses 
$837.04). The mean gross annual income for this group was $2,822.00, meaning that, on 
average, the respondents spent 29.7% of their annual income on the costs of production 
for their coffee. Gross revenues from coffee, in turn, provided an average of 50.5% of 



A category of coffee established by ICAFE, involving a flexible scoring system for environmental and 
social responsibility, based on the Starbucks "C.A.F.E. Practices" procurement standards. ICAFE has 
proposed a price premium of US$5 per fanega for sustainable coffee. 



84 



respondents' gross income. Because of the small sample size (N=4), care should be taken 
in making generalizations from these data. 

Coffee removal 

Forty-five respondents, or 76.3%, had removed at least some of their coffee in the 

preceding six years, with a total of 126.2 hectares, or 56.9%, of coffee removed. Of the 

59 respondents, 41 (69.5%) still grew at least some coffee, and eighteen farmers (30.5%) 

no longer grew any coffee at all. Of the 45 respondents who had removed some coffee, 

42, or 93.3%, listed low prices as a reason for their decision. Other factors listed were 

the high costs of production (15.6%) and large amount of labor (13.3%) required by 

coffee, problems with diseases or pests (6.7%), and the age of the coffee trees (6.7%). 

Of those who still grew coffee, thirteen (35.1%) of 37 respondents who answered 
said that they planned to reduce their coffee further in the near future, and eleven (29.7%) 
said that they would do so if the price of coffee did not increase. Only thirteen 
respondents (35.1%) said they definitely planned to continue growing coffee. 

The relationship between the decision to remove coffee, and the decision to keep 
some coffee, with other economic, social or demographic variables was tested using, as 
appropriate, / , Fisher's Exact Test, and t- tests. Several variables showed a significant 
relationship with either previous removal of coffee or whether the respondent still grew 
any coffee at all. 

The first of the associated variables was gender, with men almost twice as likely 
(80.8%) as women (42.8%) to remove coffee (Fisher's Exact Test, p=0.048). The next 
correlated variable was original extent of respondents' coffee area. Farmers who 
removed coffee started off with nearly twice as much coffee, on average, as those who 
did not (4.2 vs. 2.5 ha, t=2A56, df=51.172, p=0.017). Use of credit was significantly 
associated with whether respondents still grew some coffee, with 22.5% of coffee 
growers, and 0.0% of non-coffee growers, using credit (Fisher's Exact Test, p=0.048). 
Membership in an agricultural association was also significantly associated with whether 
a respondent still had coffee. Of those who still had coffee, 60.5% were members of an 
association, while only one respondent, or 5.6%, of those who did not have coffee 
belonged to one (Fisher's Exact Test, p=0.000). 



85 



For respondents interviewed by the Costa Rican interviewer, the price required to 
make coffee worthwhile to grow was significantly correlated with whether or not a 
respondent still grew at least some coffee. Respondents who still grew coffee desired a 
higher price ($&4.7l/fanega) than those who did not ($73. 3 llfanega, t=3A52, df=18, 
p=0.003). 

Farmers who used wood as a fuel source were significantly less likely to have 
removed their coffee than farmers who did not use wood (65.8% compared to 95.0%, 
Fisher's Exact Test, p=0.011). Whether or not respondents still had any coffee left was 
also associated with use of wood as a fuel source, with 77.5% of respondents who still 
had coffee using wood, and 38.9% of respondents who did not have coffee using wood 
(Fisher's Exact Test, p=0.007). 

The final variable correlated with coffee removal was respondents' belief 
regarding what percentage of the bird species on their farms lived solely in the forest, as 
opposed to cropland or a combination of cropland and forest. Farmers who had removed 
coffee believed that, on average, 56.0% of birds lived only in the forest, while farmers 
who had not removed coffee believed that no bird species lived only in the forest 
(1=3.162, df=8, p=0.013). 

Land use and population change 

The vast majority of respondents (86.4%) reported that at least one family 

member had left the community, and of those respondents, 82.4% listed economic 

reasons as being at least one factor in their family members' decisions to leave. Ten 

respondents (16.9%) reported that they planned to leave the community themselves, with 

six of those ten respondents (46.2%) saying they would sell their land, and four of the ten 

(30.8%) hoping to immigrate to the United States. 

One interview had to be excluded from the analysis of land use questions 
involving quantitative land areas, as the respondent did not provide data on the critical 
question of how much coffee he had prior to conversion to pasture. The total land area 
represented after this respondent was removed was 457.1 hectares. Land uses on the 
other respondents' farms were classified into several broad categories: coffee, pasture, 



86 



forest (primary and secondary), abandoned or charral, and other crops. All land use data 
for the respondents are summarized in Table 2.4.4. At the time of the survey, a total of 
20.9% of the land area managed by respondents was planted in coffee, 47.4% was used 
for pasture, 7.2% was planted in other crops, and 22.7% was not currently in use for 
agriculture (i.e. either abandoned, charral, or forest). 

Forty-five respondents, or 76.5%, had removed at least some of their coffee in the 
preceding six years, with a total of 126.2 hectares, or 56.9%, of coffee removed. At the 
same time, the total land area in pasture and other crops (such as plantain, tomato and 
peppers), increased by 81.2 and 19.3 hectares, respectively. These changes represented 
increases of 59.9% and 140.4% over their original extent. Seventeen hectares were 
abandoned, an increase of 186.3% in abandoned area. In all, the coffee removed 
represented 27.4% of the total land area managed by the respondents, with the 
replacement uses of pasture, other crops, and abandoned land representing 17.6%, 4.2%, 
and 3.8%, respectively, of the total land area, and 64.3%, 15.3%, and 14.0%, 
respectively, of the 126.2 hectares of coffee removed. Figure 2.4.4 shows the land uses, 
by percent of land area, to which respondents converted their coffee. 

Based on information given in the interviews about previous land use and future 
plans for the land, estimates were calculated for land use six years prior to the interviews, 
and for planned future land use. Figure 2.4.1, Figure 2.4.2, and Figure 2.4.3 show 
graphical depictions of these values, showing total land use at different points in time. 
Figure 2.4.5 shows land use changes as a percent of total area. 

Discussion 

The survey respondents appear to provide a representative sample for the area. 

By summing the reported household sizes of the respondents, it can be determined that 
the sample represents the households of 233 individuals living in Agua Buena, or 3.3% of 
the 6,962 people living in the district as of the 2000 census. The mean of 3.9 persons per 
household was very close to the district-wide average in 2000 of 4.0 INEC (2001). The 
gender composition of the sample is very close to the results of the Censo Cafetalero 



Pasture in a state of recent abandonment or overgrowth, characterized in the Agua Buena region by mostly 
woody vegetation averaging 1-2 meters in height; potentially an early stage of forest regeneration. 



87 



(ESTEC 2004), which found that 91.2% of coffee farmers in Coto Brus were male, and 
8.8% were female. 

The age profile for respondents was a great deal older than the census profile for 
Coto Brus as a whole (Figure 2.4.6), but much more closely resembled the profile from 
the Censo Cafetalero (Figure 2.4.7). Data from both censuses could not be plotted 
together, as the 2000 Census and 2003 Censo Cafetalero used different age groupings. 
Since many respondents noted that their sons helped them with the farms, this age 
distribution probably reflects land tenure practices, with land ownership and decision- 
making responsibilities resting with the older family members. The amount of time that 
respondents had lived in the region corresponded closely to the 1968-1973 peak of 
migration into Coto Brus (Barrantes & Pana 2005). 

When grouped by educational level (Figure 2.4.8, with respondents who had less 
than a year of schooling grouped with the "none or kindergarten only" group), the survey 
responses for education match closely those in the Censo Cafetalero. The Censo 
Cafetalero, which covered the cantons of Coto Brus and Turrialba, does not break down 
education data by region, and therefore those statistics are aggregates for both regions. 
When compared to the 2000 census data from Agua Buena, both the survey responses 
and the Censo Cafetalero show a greater percentage in the "primary" category, and fewer 
in the "secondary" and/or "post-secondary" category, possibly indicating that coffee 
farmers have had less access to education than the general population. This may also be 
affected by the comparatively greater percentage of coffee farmers in the "60 and up" age 
category, as educational opportunities improved for the country as a whole throughout 
the second half of the 20 th century. 

The 99.9% of land that was owned, rather than rented, by the respondents is 
similar to the Censo Cafetalero, in which 96.4 % of the surveyed land area in Coto Brus 
was owned. The mean income for the respondents of $4,163 is distorted by two high 
outliers of $26,500 and $18,790, thus the median of $1,784 may be more informative. 
This places at least half the sample at or below the mean income of the lowest income 
quintile nationwide (ESTEC 2003b). When the two outliers are removed, the mean 
household income for the sample drops to $2,976, or $714 per capita. Nationwide, as of 



88 



July 2003, the mean household income was $7,668 annually (ESTEC 2003b), or $1,987 per 
capita. 

Annual household income for respondents who reported other income besides 
coffee was higher, on average, than those who did not ($4,204 vs. $3,000), but due to the 
low number of respondents (two) that did not report other sources of income besides 
coffee, this difference could not be statistically tested. Furthermore, with the two outliers 
mentioned earlier removed from the analysis, the means for the two groups are nearly 
identical ($2,974 for respondents with other sources of income, $3,000 for respondents 
without). Due to the difficulty in assessing all income sources, it is likely that many 
respondents underreported income. 

Only a small percentage (15.3%) of respondents reported using credit to pay the 
costs of production. This is in stark contrast to the Censo Cafetalero (ESTEC 2003a), in 
which 72.4% of farmers in Coto Brus reported using credit. The interview responses 
showed no relationship to interviewer (Fisher's Exact Test, p=0.132), so the discrepancy 
with the census is probably not due to respondents being more or less willing to disclose 
credit use to a North American interviewer, or by a different phrasing of the question. 

The difference could be the result of differences in credit use by district not 
elucidated in the Censo Cafetalero 's aggregate data for Coto Brus, collected in the same 
year as the interviews. That is, it may indicate that farmers in Agua Buena have less 
access to credit than farmers in other parts of the canton, possibly due to being relatively 
further from banks and large, financially stable coffee processors. Many respondents 
reported that they used to use credit, but that it was no longer available. In August 2004, 
the Costa Rican daily newspaper La Nation reported that Coopabuena's critical financial 
situation was forcing it to restrict credit to its members (Barquero S. 2004a). 

Of those respondents who still grew coffee, only thirteen respondents (35.1%) 
said they definitely planned to continue growing it, with the rest saying that their decision 
was contingent on the future price of coffee. Therefore, given that the price of coffee has 
not substantially increased since the interviews were conducted in 2003, it is possible that 
as few as 13, or 22.0%, of the original 59 coffee farmers interviewed, are still growing 
coffee today. Despite the desire of these thirteen respondents to continue growing coffee, 



89 



this number could even be much lower, given the closure of Coopabuena after the 2003- 
04 harvest. 

The correlation between original extent of coffee and coffee removal may be 
explained by the fact that, because larger coffee farms require the use of paid labor, costs 
of production are often higher for larger farms. These higher costs may have impacted 
the decisions of respondents to either remove their coffee or decrease it to a more 
manageable area. In the costs of production survey, costs of production per hectare and 
per fanega were positively correlated with farm size, though not significantly so (per 
hectare: r=0.328, p=0.672; per fanega: r=0.594, p=0.406). 

Due to the small sample size (n=4) in the costs of production survey, however, the 
power to separate these groups is extremely low: given an expected r of 0.594 between 
farm size and costs of production per fanega, there is an approximately 10% chance of 
detecting a correlation at p<0.05 (Dunlap 1981). It may also be that respondents with 
more land perceived more options for diversification or changeover of crops: for 
example, it is easier to sustain cattle given a larger area of pasture. 

The correlation between credit use and whether respondents had coffee probably 
indicates that coffee growers either have a greater need for credit than non-coffee 
growers, or that they have greater access to credit, most likely through beneficios 
(respondents were not asked whether they used credit for coffee or for other crops). The 
correlation between maintaining coffee and membership in an agricultural organization is 
to be expected, since the cooperative coffee beneficios are the predominant agricultural 
associations in the region. 

The use of wood as a fuel was significantly correlated with coffee removal. 
Rather than use of wood as fuel being predictive of a propensity to keep growing coffee, 
it is likely that farmers who removed coffee subsequently had less access to fuel wood, 
and therefore had a greater incentive to switch to alternative fuels such as gas or 
electricity. This is supported by responses to other interview questions: of farmers who 
still used wood and still grew come coffee, 57.9% said that they did not expect to have 
enough wood for fuel if they removed more coffee, and 26.7% said that they would 
change to another fuel source. 



90 



Several of the statistically significant relationships between coffee removal and 
other variables are compromised by small sample size in one or more groups. These are 
the correlations between coffee removal and gender (n=7 for women), price needed for 
coffee to be worthwhile to grow (n=5 for those who no longer had coffee), and 
respondents' belief regarding what percentage of the bird species on their farms lived 
solely in the forest (n=2 for farmers who had not removed coffee). Caution must 
therefore be used in trying to draw generalizations from these data. The correlation 
between coffee removal and price desired for coffee, however, may indicate that farmers 
who maintained their coffee are willing to tolerate lower prices than those who removed 
it. 

Apart from the seven variables discussed above, no other significant relationships 
were found between coffee removal and any other variable tested in the interviews. It 
appears, therefore, that the only factor that could be predictive of a decision to change 
from coffee to another land use is the size of the original land area of coffee. 
Additionally, with the likely exceptions of fuel wood use, the data do not support any 
conclusion that removal of coffee has had any statistically significant effect on any of the 
other variables surveyed. Whether a farmer has retained some coffee, however, appears 
to have an effect on their use of both fuel wood and credit, on their membership in 
agricultural associations, and on the minimum price they believe they would need to 
receive for coffee production to be worthwhile. 

Planned future land uses (Table 2.4.4, Figure 2.4.3, Figure 2.4.5) should be 
viewed with skepticism. They do not take into account respondents' plans to leave the 
area or sell their land, unless the respondent mentioned this explicitly as part of their 
plans for their land. Respondents who expressed a desire to convert abandoned land into 
other crops, or to renovate charral into working pasture - resulting in the predicted drop in 
the "abandoned/charral" category - may not have actually had access to the resources to 
do so. Likewise, some farmers may not have had the resources to carry out a hoped-for 
conversion from coffee to another crop, and farmers who planned to sell their land may 
have had difficulty finding a buyer, resulting in land abandonment. 



91 



The "planned" land use category, therefore, only represents a desired or expected 
state of affairs in the minds of the respondents at the time of the interviews, and not 
necessarily a realistic future outcome. It does demonstrate, however, the inclination of 
respondents to reduce coffee production further, and expand pasture and other crops. 

Conclusions 

The results of this survey support HI, which states "Coffee farmers faced with 

low prices are choosing to convert their farms to other uses, and the predominant 
conversion choice is cattle pasture." The 76.3% of respondents who reported having 
removed coffee represent a clear majority of the sample (095%= 65.4%-87.2%), 
supporting the hypothesis that farmers are choosing to convert their farms to other uses. 
The hypothesis that conversion is a response to price is supported by the 93.3% of 
farmers reducing coffee who listed price as a motivating factor (095%= 86.9%-99.7%). 

Finally, the hypothesis that the predominant conversion choice was pasture is 
supported by the fact that 64.3% of the total land area converted from coffee was 
replaced by pasture. At the level of individual respondents, however, the trend is less 
clear-cut. The mean percent, by farm, of former coffee area converted to pasture was 
46.1% (CIg5%= 31.5%-60.7%). This shows that, while a majority of farmers may not be 
choosing pasture as the main replacement for coffee, those farmers who are represent a 
larger proportion of land ownership, with the net result that most of the lost coffee area 
has been converted to pasture. 

Fifteen percent of the land converted from coffee was planted with other crops, 
predominantly plantain, tomato and peppers. The vast majority of these other crops were 
more chemically intensive than coffee, based on reports from respondents themselves, 
neighbors, and agrochemical suppliers. A comparative analysis of the quantities and 
toxicities of chemicals used in conventional coffee and the crops that have replaced it 
could help elucidate the environmental impact of this aspect of land conversion. 

The conversion from coffee to other land uses was accompanied by a statistically 
significant decrease in agro-biodiversity, as represented by the reduced average number 
of products per farm (from 6.1 to 3.5). This can be accounted for, in part, by the fact that 
coffee represents one of the products lost. Additionally, low-growing crops, such as 



92 



yucca, beans and corn, which are frequently interplanted with coffee, are not appropriate 
for planting in pastures. Orange trees are frequently cut down during the conversion 
from coffee to pasture. It is also possible that the higher number of crops is indicative of 
diversification on the respondents' farms prior to conversion, which may indicate that 
more diversified farmers were more inclined, or more able, to keep growing coffee. 

The interview responses yield other important information regarding the direct 
and indirect environmental effects of the coffee crisis in the region. Twenty percent of 
respondents who used wood as a fuel source reported that, if they removed all their 
coffee, they would seek firewood elsewhere - either by collecting from other properties or 
by purchasing it. Another 40% reported that they did not know, or had not thought about, 
what they would do for fuel. These responses indicate that coffee removal is likely to 
lead to additional pressure on what remains of the area's forest fragments. 

The large numbers of family members reported as leaving the region, the ten 
respondents (16.9%) who planned to leave the community themselves, and the 14.0% of 
coffee land that was abandoned, indicate that land abandonment is another trend in the 
region that can be expected to grow in light of continued low prices. The rate of 
population increase in the area has been very low since the mid-1980's, coming close to 
leveling out by the turn of the century. Although many interview respondents had plans 
to expand crop production or renovate pasture, the simple fact is that many do not have 
the resources to do so. 

Some farmers who wanted to sell their land reported difficulty finding buyers, and 
farmers who wanted to stop farming in order to find paid work for themselves often could 
not find jobs. The closure of the Paso Real and Coopabuena beneficios almost certainly 
had dramatic effects on the ability of the area's population to sustain itself. While 
seeking interviews, interviewers heard stories of farmers borrowing large sums of money 
to pay coyotes to help them enter the United State for work, and when they could not pay 
back the loan, their land was repossessed by the lender (often the coyote himself). 
Frequently, repossessed land was simply abandoned. One woman described watching the 
coffee berries ripen and fall off the branches of the coffee trees on what used to be her 



Someone who transports or arranges transport of persons to the United States, usually illegally. 

93 



husband's farm, as the coyote who had repossessed the farm apparently had no interest in 
harvesting them. 

It may seem that land abandonment could have a positive effect on the 
biodiversity of the immediate area around Agua Buena, by allowing secondary 
regeneration of forests - and indeed, many interview respondents reported that, although 
animal populations had declined since they first moved to the region, many species were 
starting to rebound. It should also be considered, however, that migration to cities is 
accompanied by its own set of environmental problems. When the larger scale is 
considered, the environmental impacts of urbanization resulting from land abandonment 
could offset some benefits resulting from the regeneration of secondary forests on the 
abandoned land. Such a comparative analysis, however, is beyond the scope of this 
work. 

In short, three main responses to the coffee crisis revealed themselves during the 
course of the interviews: out-migration from the district, land abandonment, and 
conversion of land to pasture and other crops. Some of the environmental effects of 
conversion to pasture are examined in the following chapter. The effects of conversion to 
other crops is not examined in detail, though a hypothesis that this type of land use 
change is environmentally detrimental would be likely to be supported based on relative 
levels of agrochemical use and reduction in agro-biodiversity. Finally, out-migration and 
land abandonment is likely to have mostly positive environmental (though not social or 
economic) effects in the immediate area, but likely negative environmental effects on a 
larger scale, as emigrants move to cities. 



94 



Tables and Figures 



Yes/No Questions 


Yes 


No 


No answer/ 
Not applicable/ 
Don't Know 


Spouse participated in interview 


31 (52.5%) 


18(30.5%) 


10 (16.9%) 


Own their land 


58 (98.3%) 


1 (1.7%) 


(0%) 


Source of income besides coffee 


49 (83.1%) 


7(11.9%) 


3 (5.1%) 


Use credit to pay farm costs 


9 (15.3%) 


47 (79.7%) 


3 (5.1%) 


If uses credit, lender requires them to use 
specific practices 


9 (100.0%) 


(0.0%) 


(0.0%) 


Want children to be farmers as well 


18 (30.5%) 


31 (52.5%) 


10(16.9%) 


Member of agricultural association(s) 


24 (40.7%) 


32 (54.2%) 


3 (5.1%) 


Reported outside source of technical 
information 


17 (43.6%) 


20(51.3%) 


2(5.1%) 


Reported outside source of price 
information 


24(61.5%) 


13 (33.3%) 


2(5.1%) 


Has reduced coffee in the last 6 years 


45 (76.3%) 


14 (23.7%) 


(0.0%) 


Still grow coffee 


41 (69.5%) 


18 (30.5%) 


(0.0%) 


Plans to reduce coffee further (if they still 
grow coffee) 


13(31.7%) 


13(31.7%) 


15 (36.6%) 


Currently use agrochemicals 


33 (55.9%) 


21 (35.6%) 


5 (8.5%) 


Some family members have left the 
community 


51 (86.4%) 


7(11.9%) 


1 (1.7%) 


Planning to leave the community 
themselves 


10(16.9%) 


46 (78.0%) 


3(5.1%) 


Use wood as a fuel source 


38 (64.4%) 


20 (33.9%) 


1 (1.7%) 


If they use wood, expect to have enough 
fuel wood if they remove their coffee 


5 (13.2%) 


22 (57.9%) 


11 (28.9%) 


If continuing with coffee, plan to change 
from conventional to sustainable or 
organic coffee 


8 (19.5%) 


16(39.0%) 


17(41.5%) 



Table 2.4.1. Summary of interview responses on yes/no variables. 



95 



fa « 

a ^ 



CJQ T3 

< 8 

2 o 

3 ft 

PD CS 



fv <w 

IT c 

3 3 

ft 

£3" o 

C7- o 

n ??5 



rt 

c 

o 



c 
c 



o 

3 



o 



I 



o 

3 

ct 

3 



v. 

C. 
< 

ft 



o 
era 

o 



3 

3- 



3 

ft 



» 
S" 

4- 

t/3 

C 

3 

3 
p 



3 

Ft 

— . 

rt' 



O 
3 



O 

P 

Ft 

r/q 
O 

o' 



cr 



£ > 

It 

cr rj 

CD P 

^ pa 

p H 

5 o 

CD Cs 
ja OQ 

Z p 

3 B. 

2 N 

3 p 

» 2 



z 


O 


n 


Z 


n 


r^- 


o 


n 


3 


L> 


^ 




rt> 


l-h 


p 


>-t 


CD 


CS 




P 


O 






v> 


o 


rp 




O 


n 


»-j 




p' 


N3 


3 




r^- 


i-i 


n 




o 


P 


n 




CS 


< 


cr 






(U 


re 

3' 

o 

o 

3 


U> 


L»J 


(O 




on 


/ „ 


K^ 






On 


^ — s 




on 


• 


u> 




no 


1 


(A) 






w 


on 


o 

ON 






ON 

00 

*3 



o 

5= 

o 

CD 



3 
O 

o 
3 

CD 

cr 

CD 

t/3 
Pi 

CD 

o 

o 

B 

CD 
ft 



O 
s 

3 



C/3 

P 
r/Q 
p 
R 

n 

p 

CS 

re 



< 

CD 

r/Q 

re 

P 

cr 



no 



NO 



re 


P 


p 
CS 


R 


t/3 


P 


o 




H+! 




P 


P3 


o 

hO 


S" 


p 








R 








CS 




OQ 




i i 




P 




CS 




Q* 





o 

as 
<-► 
re 



O 

re 
CS 

rTRS 
o 

3 

© 

= 

re 

o" 

S3 



t/3 

re 



NO 



p 
re' 



re 



oo 



to 






00 
00 



Categorical Questions (continued) 


What was done with land that was 
formerly coffee 


Pasture 


25 


(55.6%) 


Abandoned 


10 


(22.2%) 


Reforested 


3 


(6.7%) 


Other crops 


17 


(37.8%) 


Sold 


2 


(4.4%) 


Not specified 


3 


(6.7%) 


Reasons for removal of coffee 


Prices 


42 


(93.3%) 


Labor 


6 


(13.3%) 


Input costs 


7 


(15.6%) 


Pests/disease 


3 


(6.7%) 


Age of coffee 


3 


(6.7%) 


Other 


6 


(13.3%) 


Not specified 


1 


(2.2%) 


Reasons for keeping coffee 


Protect investment 


3 


(7.3%) 


Personal values or tradition 


8 


(19.5%) 


Hope for improvement in prices 


12 


(29.3%) 


No money for conversion 


1 


(2.4%) 


Small amount of land 


3 


(7.3%) 


Other 


2 


(4.9%) 


Not specified 


15 


(36.6% 


Reasons family members have 
left 


Marriage 


7 


(13.7%) 


Economic reasons 


42 


(82.4%) 


School 


7 


(13.7%) 


Other 


3 


(5.9%) 


Not specified 


4 


(7.8%) 


If thinking about leaving the 
community, what they would do 


Sell land 


6 


(46.2%) 


Go to U.S. 


4 


(30.8%) 


Go to other part of Costa Rica 


1 


(7.7%) 


Don't know 


2 


(15.4%) 


Go wherever there is work 


1 


(7.7%) 


If wood is used as a fuel source, 
and they are removing coffee, 
what they will do for fuel 


Plant trees 


7 


(23.3%) 


Change fuel source 


8 


(26.7%) 


Buy wood 


1 


(3.3%) 


Collect wood elsewhere 


5 


(16.7%) 


Not specified/don't know 


12 


(40.0%) 



Table 2.4.2. Continued. 



97 



Categorical Questions Continued 


Classes of chemicals used by 
those who use agrochemicals 


Fertilizers 

Foliar fertilizers 

Fungicides 

Herbicides 

Nematicides 

Organic fungicides (incl. copper) 

Others 


9 

5 
11 
17 

1 

5 
7 


(27.3%) 


(15.2%) 


(33.3%) 


(51.5%) 


(3.0%) 


(15.2%) 


(21.2%) 


Specific chemicals used by those 
who use agrochemicals 


Atemi (Cyproconazol) 

Counter (Terbufos) 

Glyphosate 

Lime 

Mancozeb 

Others 

Paraquat 

Silvacur (Tebuconazol + 

Triadimenol) 

Tordon (2,4-D + Picloram) 


6 

1 

16 

3 
1 

7 
5 

4 

3 


(18.2%) 


(3.0%) 


(48.5%) 


(9.1%) 


(3.0%) 


(21.2%) 


(15.2%) 


(12.1%) 


(9.1%) 


Whether they believe there are 
fewer or more types and numbers 
of animals than when they first 
started with their farms 


Fewer 

More 

Same 

No answer 


20 

5 
12 

2 


(52.6%) 


(13.2%) 


(30.8%) 


(5.1%) 



Table 2.4.2. Continued. 



98 



a> o 


^ 


t£ 


^ 


^ 


^ 


^ 


£ 




# 




# 




^ 


^ 




# 


^ 




# 




# 


£ 


^ 


£ a 


r-» # 


r-_ 


^ 


>— ; 


O 


CN 


p 




— ; 




r- 




in 


p 




o 


p 




CN 




CN 


p 


ON 


V) M 


T—3 


T-H 


cn 


in 


o 


cn 


o 




in 




i> 




T—I 


o 




o 


o 




CN 




CN 


o 


T—3 


fi -M 














































^H 


No a 
Don' 


t-H 


t-H 


CN 


cn 


o 


^-^ 


o 




cn 




CN 




^ / 


o 




o 


o 




^H 




^H 


o 


^-^ 












in 














cn 




















t- 


















































VO 


CN 


-xt" 


ON 


VO 


-xt" 


l> 




On 




On 




o 


l> 




^f 


_ 




On 




2 


<S~) 


av 


VO 


en 


_; 


_; 


,-H* 


VO 
lO 


cn 




O 




On 




ON 




VO 


t~- 




r- 




00 


cn 




,—l 




, — 1 


, — 1 








^f 




it! 




CN 


o 




o 


o 




CN 




00 


,—l 


CN 














■6/5- 














cn 
















CN 






s 

3 


ON 


VO 


IT1 


CN 


_ 


O 


i> 




O 




o 




o 


p 




p 


o 




o 




O 


p 


av 


00 


,_H 


VO 


in 


'-' 


O 


,_H 




cn 




cn 




VD 




cn 


-*' 




^t-' 




oo 




s 












o 

o 














o 












"" ' 




o 






'H 












in 














o 
















o 






d 












vo" 




































s 












CN 

■6/5- 




































s 

9 


VO 


O 


cn 


cn 


^H 


O 


o 




o 




o 




# 


O 




o 


p 




CN 




^ 


p 


o 


CN 










o 














p 


O 




o 


o 




o 




in 


o 




£ 












o 

VO 














cn 
















CN 






3 












cn 

■6/5- 






























^ 






§ 
















































S3 


iri 


O 


o 


o 


o 


Tl" 


p 




o 




p 




^ 


O 




p 


p 




o 




^ 


p 


p 


OS 


"t" 


cn 


r-' 


CN 


-xt-' 


oo 


iri 








T— I 




o 


T— I 




T—I 


o 




CN 




cn 


cn 


CN 


'-3 


l/"1 




CN 


CN 
















-*' 
















cn 
















■6/5- 














in 
















00 






a 


cn 


o 


cn 


cn 


C7\ 


cn 


"vt 




VD 




cn 






^t 




1 — 1 


_ 




OV 






00 


OO 




-*' 


^c" 


00 


,—J 


cn 


cn 


in 




—<' 




VO 




^H 




^H 


cn 




CN 




CN 


CN 


in 




CN 


CN 




1 ^ 










CN 












o 
























■6/5- 














CN 

in 
















CN 






















>> 




>> 




















































o 








































"c0 




"c0 




4= 








































3 




3 




& 






s 

H 






CU 




CU 
























3 
3 

CO 




3 
3 
CO 




c/3 
I-H 

CU 










> 

CO 

4= 




> 

CO 

4= 






CU 
-3 
















e 


13 
cu 




13 
CU 




s 






3 

O 






O 
43 




O 
43 






CU 














c/3 
H 


S-H 

CO 

e 

o 


13 

3 

4— » 

^— > 

CO 




13 
3 

CU 

-4-» 
-4-» 

CO 




CO 

"3 
o 






o 






CU 

c/3 
O 




CU 

C/3 
O 




13 

CU 

> 

o 


CU 

> 

CO 
-3 












2 


CO 


,(-H 

<4-H 


c/3 
00 




00 








o 

4= 







-3 

H— » 




-3 




E 


O 
-3 










13 
O 


O 

-a 

eu 

C/3 


'o 
3 

CU 


13 
(U 
O 
3 


_3 

*4— » 

CU 
CU 




_3 

*-4-» 

CU 
CU 




CU 

o 




| 


C/3 
— 

CU 




CO 

3 
O 


>> 

43 

13 
CU 




>> 

43 

13 
CU 




CU 

'-H 

CU 

VJ— 1 


C/3 

t— 

CU 


73 








Oh 


3 


g 


13 

O 

S-H 

Oh 

C/3 

-4— » 


E 




a 




CU 




CO 


43 




C/3 


> 


c/3 


> 


/ ^ 


« 


43 


S3 

C/3 








13 

a 

CO 


O 

-a 
_3 


o 
o 

_3 


3 

*4— » 




3 

*-4-» 


— 

CU 


s 

■2 




3 
O 

C/3 


s 

CU 

s 




'-H 

CU 

o 


o 
S 

CU 


CU 
— 

CO 

H— » 

CU 


o 
S 

CU 


H— » 

3 
CU 
CU 

^H 


o 
o 

o 

00 


E 

CU 

E 


3 




C/3 

5-H 

3 


3 


-4—* 

a 

eu 

Id 


(U 

"Eh 

o 


2 

O 


o 

13 

O 


CO 

'o 
o 

C/3 




CO 

'o 

o 

C/3 


43 

s 

CU 

s 

3 

*-4-» 

CO 


CU 

s 

o 


CU 


'— i 

CU 

o 
o 

— 


>> 


cr 


13 


CU 

=4^ 


CU 

4=^ 

CU 


CU 
t4H 


CU 
Oh 

CU 


CO 

C/3 

'— 

CO 


s 




C/3 

— 

co 
u 


CU 

>> 

a 


o 
'5b 

CU 

a 


3 
o 


(U 

Oh 
<+H 

o 

S-H 


CU 

c/3 

o 

-a 


S-H 

Oh 

o 

S-H 


c/3 

CO 

<4^ 

o 


CU 

ft 

s 

Cv3 

C/3 


C/3 

CO 

o 

•— 


CU 

_3 

<4^ 

o 


^4-H 

o 
o 

CU 

> 

CO 
-3 


CO 

o 

■— 


c/3 

00 

_3 


'3 

Oh 
=h-h 

o 

— 


o 

CU 

o 

H— » 


s 

o 

c/3 

13 


o 

CU 

o 

H— » 


s 

o 

c/3 

13 


CU 

o 

— 


CO 

=4-H 

o .3 

^ 3 


+1 


v^ 




(U 


'cO 

3 


(U 


CU 




CU 


-4— » 

3 

CU 
CU 


CU 


CU 


13 


CU 


3 


CU 


3 


CU 


CU 


cu 3 


s 


CO 


C/3 


C/3 


43 


43 


42 


g 


42 


'o 


43 


43 


3 


43 


3 


> 


3 


> 


43 


3 E 

z 3 


« 


u 


3 


H 

CO 


H 

CO 


E 


E 


s 


*-4-» 


s 


o 

C/3 


S 


s 


CJ 


S 


O 

S 


O 

s 


O 

S 


O 

s 


E 


3 


00 


13 


0) 


CD 


3 


a 


3 


3 


3 


3 


03 


CU 

Clh 


7^ 


3 


3 


a 


3 


3 


a 


< 


PJ 


>H 


>H 


z 


< 


Z 


Z 


CU 


z 


CO 


-4—* 


z 


z 




z 


< 


CU 

•— 


< 


CU 

•— 


z 



.3 

- 
> 

CU 

> 






C 

o 

C3h 



o 

J-H 

-I 
E 
E 



— 

2 

CS 



99 



& * 

a er 

c <T> 

en ft 



o 
< 

c 

v. 






c 



33 
o 

o 

3 



3 

3- 
o 

i 

o 



3 

cr 
o 

? 

o 

3 

3- 
a 

n 

o 

x 
73 



(2 
3 

z 

o 
3- 

> 

3 

•". 
^_ 

o' 

3 

3' 

<. 

o' 



c 



H 

a: 

S" 

J- 
1*1 

n 

o 

3 

3' 

c 



s: I 

g- ^ 

~ cd 

Hj '-t 

S o 

hj h+! 

CD ^ 

3 p 

I 3 

Si £L 

p: k 

^ S! 

CD o 

rf O 

O OQ 

tt 3. 

rt CO 

cr a- 

^ o 

3 CJ 

£L 3 



o 



Percen 
foreste- 


3 


CD 

p 
R 


Price 


"s 5 


0' 

CD 


n 




c 
3 


cr 

CD 


o 


as CS 


3 


=3 


CD 
CD 


cr 

CD 


CL "■ 


Hj 


i-h 


Cr<5 CD 


>Q 


CD 


"hI- 

cd' 


HJ 


h-J Htl 


O 


CT 


a s. 


C^ 


CD 


O 


cd a" 
p Js. 


a; 


OQ 


* CD 

a- 


* 


CD 


& 


Ht; 

s 

C6 


o & 


i-j' 


cd 


c? 




3> 


p 


species believ 
f farms 


en 

CD 


O 

o 

Hb 









HI 
O 

O 


P 4 


&0 

C<5 


o_ 


CD 


3? 




^ 


cr 


cr 


CD' 

cr 

CD 

CD' 
< 

CD 

a- 


CD 

o' 

CD 


CD 
CD 
r^- 
O 

cr 

CD 




CD 
CD 

O 

cr 

CD 


p 

Hj 
< 

CD 


p 

CD 
CD 

a- 


CD 

a- 


3' 

3 

CD 


3 

O 




Hj 

O 




p 


r- f 

o 


r^- 


HI 
r-t- 




Htj 




P 


CD 


o 


3 
o 


cr 
cr 




P 

cr 




cr 


CD 


*< 






cd" 




p 

H| 


3' 


o 

3 




CD 








< 

CD 


o 






O 








Gn 


HJ 


Hi 












r+ 




p 




OQ 










*<f 


3 




H( 

O 










4^ 






■^5- 


■£/> 








^/i 






00 


ON 










to 


VO 
VO 


to 


On 




H^ 


4^ 




<-»J 


O 


O 




ON 


LO 


to 


Oi 


u> 


on 




OO 


O 



o 
o 



o 

CD 

Hj 

CD 

o 

CD 

<' 
CD 



o 

c 

S3 

B 



CD 

o 

c 

CD 
CB 

o" 

B 

VI 

n 

o 
B 

3' 
c 

CD 

a 



(X) 














■^5- 


■^5- 













. 


00 


ON 














IM 






to 


^0 





-j 


Lfi 


O 


Ln 





Lfi 


lyi 


^1 


to 


VO 


O 












Vi 


■^5- 











,.._. a 


~-i 


L»J 












VO 

^1 




On 
On 


to 


H- 





^1 


H 


'^ 


Lfl 


O 













Vi 








O 






&> 


H^ 








O 


to 


to 





NO 


K^ 




to 


on 







^) 


to 


On 


O 


H- 


O 





'-J 


■-1^ 


-P^ 


O 


O 





to 


-p- 





O 


O 




Of) 






r *> 


V5 







to 

ON 

00 
00 

■£/> 
4^ 

00 

00 



NO 



CD 

a 
5' 

B 



B 

3 

c 

3 



S3 

3 

c 

3 





















































ON 




& 


£ 




£ 


o 
o 




¥ 


£ 
-* 


& 


^ 


& 
-* 


^ 


o 

o 




CS 


oo 


^ 


^ 


O 


^ 


o 
o 




o 


P 


CS 


p 


vd 


00 




oo 


ON 


p 


>-; 


^d 


>-; 




io 


CS 


^ 


oo 


P 


oo 




CS 


-3- 


P 


MD 










Tt- 


CS 


en 


CS 












U~i 


ON 


en 








es 


MD 


00 


^H 


CS 


in 


p 


_H 




oo 


NO 


oo 


VO 


vo 


p 


^H 




CS 


p 


O 


vri 


VO 


p 


^H 


■M 


P 


vd 


P 


P 


^d 


oo 


p 




_' 


p 


en 


On 


>r-i 


iri 


_' 




ON 


— ' 


en 


P 


P 


oo 


^d 


o 


ON 


^H 


en 


CS 


r- 




in 




CS 


en 


^H 




f~~ 




^D 




MD 


^ 


-* 


^H 


r^ 




in 


H 




CS 










-xl" 




CS 












-xt" 






CS 










■xt 


Q 


CS 


r- # 


en 


en 


On 


in 






en 


oo 




t~-; 


On 


en 






^H 


oo 


t~- # 


^H 


On 


in 




in 


CS 


o 






no 


o 


P 




P 


oo 




o 


M2 


o 






CS 


o 






MD 


d> 


P 


S 

s 


O 


p 


ir> 


p 


p 


p 


V~i 




p 


p 


VO 


p 


O 


p 


VO 




p 


p 


p 


O 


O 


p 


vo 


o 


o 


P 


vd 


o 


en 


CS 




P 


\6 


I~-' 


iri 


O 


CS 


CS 




o 


o 


o 


vd 


O 


P 


CS 


I 


^H 


r- 






in 




^H 




^H 


in 






V~i 




^H 




^H 


r^ 


^H 




iri 




t-H 














t-H 
















t-H 
















t-h 


















































H 
















































« 
















































S 
















































s 

i 


O 


p 


p 


p 


p 


p 


CS 




O 


p 


p 


p 


p 


p 


CS 




p 


p 


O 


p 


p 


p 


O 


o 


o 


o 


o 


o 


o 


o 




o 


o 


o 


o 


o 


o 


o 




o 


o 


d> 


o 


d> 


d> 


o 
















































3 
































































































s 
















































S3 


p 


in 


p 


p 


p 


p 


p 




p 


p 


p 


p 


p 


p 


en 




p 


^H 


p 


p 


p 


p 


p 


2 


P 


o 


o 


o 


o 


o 


P 




en 


o 


o 


o 


o 


o 


en 




o 


P 


d> 


o 


C3 


C3 


en 


'-3 
















































0) 
















































S 
















































S3 


MD 


r- # 


•o 


IT) 


en 




ON 




oo 


en 


CS 


CS 


en 


^ 


p 




CS 


CS 


t> 


en 


en 


^H 


ON 


es 

0) 


P 


en 


o 


o 


— < 


o 


P 




en 


CS 


o 


o 


— i 


o 


oo 




p 


p 


d> 


C3 


P 


o 


P 


s 
















S3 
































^— V 
















# © 
































C/3 














































0) 










>> 






°C/3 










>. 
















>> 






S- 
53 










5-H 

CO 






S- 

0) 










CO 






0) 
73 










CO 
















13 






>• 

S3 
O 

U 










13 






P 










13 






0) 










O 

o 














O 

o 






T5 
S3 
53 

P 










o 
o 






0) 








^ H 


C/3 






s 








^ H 


CZl 












^ H 


M 






C/3 








CO 


5-H 












CO 


— 












CO 


— 






P 








id 


o 






•_ 








Id 


O 






0> 








K 


o 






T5 








CO 

43 


>> 




CO 


© 








CO 

43 


>> 

— 




CO 


a 








CO 

43 


— 




CO 


S3 








O 


CO 


13 


CD 


*S 








O 


CO 


13 


CD 

l_4 


-tf 








O 


CO 


13 


2 


C3 

P 






C/3 

Oh 

O 


13 


J 


0) 


CO 
13 


0) 






CZ1 

O 


13 




0) 

yd 


CO 
13 


s 






Oh 
O 


■v3 




0) 

yd 


CO 
13 


■M 






fl 


Ph 


'o 


a 


73 
P 






e 


Gh 


'o 


fl 


'O 






fl 


CP 


'o 


a 


S3 




D 


o 


O 


•* — "' 


u 


CO 




QJ 


o 


o 


^ — ^ 


u 


CO 


0) 




D 


CJ 


O 


^ ' 


u 


CO 


0) 


<u 


C 




13 


-4-* 


On 


^^ 




o 


'— 




13 


-^^ 


On 


^^ 


S3 
S3 
53 


o 


s5 




13 


-4-> 


On 


^^ 


s- 
3 


O 


3 

CO 


CD 

43 

-4-> 


e 

CO 
4= 


CO 

o 


C/3 

O 


'cO 

o 


T3 

S3 
C3 


O 


3 

CO 


5h 

4^ 

■4-» 


a 

CO 
43 


— 

o 


CZ1 

O 


'cO 

o 


o 


3 

CO 


-4— » 


CO 
43 


CO 

o 


CZ1 

O 


'cO 

o 


U 


U 


Ph 


o 


< 


Ph 


£ 


H 


P 


U 


Ph 


o 


< 


Ph 


2; 


H 


s 


u 


Oh 


o 


< 


Ph 


2 


H 



T3 
C 



C3 



3 

O 



-a 



T3 
C 

- 



3 

-a 
-j 
a 
c 



& 




H 


v^ 


-: 


u 


r/^ 


CJ 






(1) 


Lh 


-C 


ti 


q-n 


5 


r) 


f3 










o 


T1 




3 


'C 


a 








r- 


- 

*— 


o 


- 


^-( 


-C 


n 


o 




1) 


j= 


■yi 


on 


3 


^3 


-a 


ri 


c 




tfl 








- 


u 


o 


c 


>■> 


<u 


- 


> 


fc 


CD 


H 


C 


3 

CO 


cu 






Tf 


o 


^r 


X) 


r-i 


cu 


— 

2 


"a 


03 


^ 


H 


T3 



101 



102 



Reported Land Use of Respondents - ca. 1997 



Not specified 
1.1% 



Forest 16.4% 

Abandoned/ 
charral 2.1% 




Coffee 48.1% 



Other crops 
3.0% 



Pasture 29.4% 



Figure 2.4.1. Estimated land use of sample circa 1997, based on reports of land use 
changes in the six years prior to interviews. Total land area represented is 461 hectares. 



Reported Land Use of Respondents - 2003 



Not specified 
1 .8% 



Forest 16.7%-. 



Abandoned/ 
charral 6.0% 




Coffee 20.9% 



Other crops 
7.2% 



Pasture 47.4% 



Figure 2.4.2. Land use of sample in 2003, at the time that interviews were conducted. Total 
land area represented is 457 hectares (reduction from 1997 is due to farmers selling land). 



103 



Land Use of Respondents - Reported Future Plans 

Not specified 

1-8% .^^ks^ — Coffee 15.2% 



Forest 17.0%— 



Abandoned/ 
charral 3.8% 



Other crops 
9.4% 




-Pasture 52.8% 



Figure 2.4.3. Planned future land use of sample based on respondents' descriptions of 
their plans for their land, not taking into account respondents' plans to leave the region. 
Total land area represented is 456 hectares (reduction from 2003 is due to farmers 
planning to sell land). 



105 



Land Uses Replacing Coffee 1997-2003, by Area 



Sold 3.2% 



Not specified 2.4%— 



Forest 0.8%— 



Abandoned 14.0%- 




— Pasture 64.3% 



Other crops 15.3%— 



Figure 2.4.4. Land use, by area, replacing coffee that was removed. 



Land Use Change (Percent of total area) 




20% - 










15% - 














10% - 














5% - 
0% - 
















■n 








— , 




■ Coffee 
□ Pasture 


I 






I ! ' 






-5% - 
















■ Other Crops 


-10% - 
-15% - 
-20% - 


L 


inge 1997-2003 Planned change post- 
2003 




Total change 




□ Abandoned/Charral 
■ Forest 












-25% - 


L 










-30% - 





























Figure 2.4.5. Land use change as a percent of the total original area represented by respondents' farms. 



107 



Age of Sample vs. Census 2000 




15-29 

30-59 

Percent of Adults (Over 15) 



Census 2000 
Sample 



60 and up 



Figure 2.4.6. Age of survey respondents compared to ages of adults in Coto Brus as a whole. Census 
2000 data from (INEC 2001). 



Age of Sample vs. Censo Cafetalero 2003 




15-24 

25-59 

Percent of Adults (Over 1 5) 



Censo Cafetalero 



Sample 



60 and up 



Figure 2.4.7. Age of survey respondents compared to ages of respondents in the 2003 Censo Cafetalero 
(INEC 2004). 



109 



Educational Level 



Hone 



or 




20.0% 

1 0.0% 

0.0% 

Kinder* onW P ntn ag 



Census 2000 
Censo Cafetalero 
Respondents 



Figure 2.4.8. Educational level of survey respondents compared to results of the Censo Cafetalero for 
Coto Brus and Turrialba, and the 2000 Census for Agua Buena. 



Ill 



H2: Effects of conversion on bird life 

Hypothesis: Coffee farms that have been converted to cattle pasture will show a lower 
bird diversity and richness than coffee farms that have not been converted. 

Introduction 

The previous section found support for the hypothesis that, presented with the low 

prices of the past decade, farmers in Agua Buena were converting their coffee to other 
land uses, and that the predominant new land use by area was cattle pasture. Given the 
apparent scale of this conversion, and the fact that the agricultural matrix - with its 
various components such as pasture, coffee and vegetable crops - is largely the only 
habitat remaining in the study area, it is important to examine whether the conversion 
from coffee to pasture will result in a loss of habitat for wildlife. 

As many as 380 species of birds, or 32-45% of Costa Rica's 840 species (Stiles 
& Skutch 1989), are believed to have once inhabited the southern half of Coto Bras 
(Hughes et al. 2002), which includes the district of Agua Buena. The Las Cruces 
Biological Station bird checklist (Stiles et al. 1996) is considered a good baseline for the 
species that once occurred in the area (Daily et al. 2001). Dramatic forest fragmentation 
and loss due to clearing for agriculture has occurred in this portion of the canton over the 
past 50 years, with the largest remaining fragment represented by the 227-hectare forest 
at the Las Cruces Biological Station. 

Surveys conducted in 1995 found 272 species living in both forested and open 
habitats in the area, with 123 species observed at least part of the time in open habitats, 
while over 50% of species, or 149, were observed only in fragments of primary forest 
(Daily et al. 2001). In 1998, Hughes et al. (2002) observed 144 bird species utilizing 
agricultural habitat in Coto Brus. Another set of surveys, carried out by Lindell et al. 
(2004) from 1998-2000, detected 271 species in both forested and agricultural habitats 
(including coffee), of which 81 species, or 29.9%, were only found in primary or mature 
secondary forest. Characteristics of the birds utilizing the Coto Brus agricultural 
landscape are discussed in detail in Lindell et al. (2004), and both Daily et al. (2001) and 



113 



Lindell et al. (2004) provide detailed information on characteristics of birds utilizing 
agricultural and forest habitats in the area. 

Methods 

Study area and site selection 

The study area is located between 8°40' N and 8°50' N, and between 82°55' W 

and 83°00' W (IGNCR 1975). The area is classified as a premontane rainforest life zone 

(Holdridge 1971; cited in Rosemeyer et al. 2000), where precipitation averages 3600- 

4000mm per year. The wet season is considered to be April through November, when it 

rains almost daily, usually in the afternoons. During the dry season, December through 

March, it rains approximately once every two weeks (Rosemeyer et al. 2000). 

Eight sites were selected for surveys: four working coffee farms, and four 
pastures that had been converted from coffee within the past six years but still contained 
many of the overs to ry trees. All sites were between 1000 and 1200m in elevation and 
were between 1.3 and 2.5 hectares in size, as estimated by GPS data (see below). Figure 
2.4.1 1 shows a map of the study area, with survey sites marked. 

Sites of similar habitat were no less than 500m from each other, and sites of 
differing habitat were no less than 250m of each other, with the exception of pasture sites 
3 and 4, which were only 200 m from each other, separated by a 100m strip of gallery 
forest, a 75 m strip of pasture, and a dirt road. Pasture and coffee sites were selected to 
represent management styles typical for the area based upon personal observations. Site 
selection was influenced by transportation constraints and the availability of patches of 
homogenous habitat large enough to contain a survey site. Information regarding land 
management practices was obtained through questioning landowners, neighbors and farm 
workers. Detailed site descriptions are provided in Appendix B. 

Pasture sites (Figure 2.4.9) had originally been coffee farms with mixed shade 
cover of Erythrina spp. ("poro"), Musa spp. (banana/plantain) and/or Citrus spp. 
(primarily oranges). The sites had been cleared of all coffee trees and some shade trees 
within the past six years. None of the sites were currently under treatment with any 
114 



agrochemicals, with the exception of Site 4, which was treated with the herbicide mixture 
Tordon ( 2,4-D picloram) during the survey period. 

Coffee sites (Figure 2.4.10) were coffee farms in active production. Heights and 
distances for vegetation cover are estimates based on visual observation. All sites had a 
mixed shade cover of Erythrina spp., Musa spp. and/or Citrus spp., at a distance of 
approximately one tree per 7-10 meters, now a typical management system for the Agua 
Buena region. All shade trees on the sites were 5-10 m tall and in good condition with 
wide, leafy crowns, unless otherwise specified. Coffee bushes were between 1 and 2 m 
high, at a density of 3,000-4,000 bushes per hectare. All sites had approximately a 5 cm 
layer of organic mulch covering the ground, and a dense layer of herbaceous weeds less 
than 1 m high. No chemical fertilizers, fungicides or insecticides had been applied to any 
of the farms during the preceding year, but one site was treated with Roundup 
(glyphosate) during the survey period. 

GPS data 

Data points for plot boundaries were taken in July 2003. ESRI ArcMap 9.0 was 

used to generate polygon coverage using these data points, then to obtain the area for 

each plot and to determine the minimum distances between sites. The polygons were 

projected onto an aerial photo of the study area (IGNCR 1998) provided by Dr. Karen 

Holl, Department of Environmental Studies, University of California, Santa Cruz, to 

create a visual map of the survey sites (Figure 2.4.1 1). 

Bird survey method 

The survey method used to compare the bird fauna of the sites was a modified 

version of an area search bird census protocol provided by John Alexander of the 

Klamath Bird Observatory (Geupel et al. 2003, J. Alexander pers. comm. 2003). This 

method was chosen because it does not require the observer to be able to immediately 

identify all the birds of an area by sight and call, but produces relatively complete, 

statistically comparable samples (Dieni & Jones 2002). It appears that all of the 

published work using this method has, to date, been conducted in temperate regions. A 

115 



further discussion of the relative effectiveness of the area-search protocol compared to 
other methods is available in Appendix B. 

Each site was divided into two survey plots of approximately equal size, based 
upon natural or constructed boundaries such as paths, fences, or hills, for a total of 16 
plots. Pilot surveys determined that an appropriate period of time for these searches was 
15 minutes, allowing complete coverage of both coffee and pasture plots, with minimal 
backtracking. The observer started at one end of the plot and moved slowly in a zigzag 
or C-shape (depending on the shape of the plot and visibility), in such a way as to survey 
all parts of the plot, finishing the 15 minute period at the other end of the plot. If the 
observer neared the other side of a plot before the survey time was up, they would return 
to an area near the center of the plot, with high visibility, and remain there watching for 
new birds until the time was up. Notes were taken using a hand-held tape recorder, and 
transcribed at the end of each day. 

All birds seen, and birds heard that could be reliably identified and which the 
observer was sure were inside the plot, were counted. Care was taken not to double- 
count individuals. With the exception of hummingbirds, only birds that were observed 
perching or foraging on vegetation in the plot, or on the ground, were counted (i.e. birds 
that flew overhead were not included). Hummingbirds were counted when they were 
observed foraging on vegetation within the plot. Birds located in living fence trees 
bordering the plots were not counted unless they were also observed inside the plot. The 
habitat where a bird was situated was recorded if the bird was seen, or if a bird identified 
by voice could be reliably located. Breeding activity was noted. 

Surveys were carried out during the wet season, from August 18 through 

September 14. The plots were each sampled six times over this one-month period, 

between the hours of 5:15 a.m. and 9 a.m. To avoid any biases based on the time in the 

morning that the sites were sampled, searches were scheduled so that each plot was 

searched once each between 5:15 and 5:59 a.m., 6 and 6:59 a.m., 7 and 7:59 a.m., and 8 

and 8:59 a.m., and once each between 5:15 and 6:59 a.m. and 7 and 8:59 a.m. The time 

of sampling refers to the time at which the survey was begun. Figure B.l in Appendix B 
116 



gives a visual explanation of the site and survey design, and Table B.l gives the schedule 
on which the surveys were conducted. 

It is conventional to express the start times of these types of surveys as time from 
sunrise (Steven G. Herman, Evergreen State College, pers. comm. 2005). The proximity 
of the survey area to the equator meant that, between the start and end dates of the 
surveys, the sunrise time only ranged from 5:22 a.m. to 5:24 a.m. (U.S. Naval 
Observatory 2005). Thus, the fact that the rotation of survey times was not based on 
sunrise time most likely did not have a significant effect on the results. 

Most of the counts took place outside the migratory season, but a few migrants 
began to arrive during the last few days of counting. Because of this slight overlap, 
counting migrants would inflate species diversity for the last few surveys done. 
Therefore, while migrants were noted during the surveys, they are not included in the 
data analysis. 

Data analysis 

Information on family for each species were obtained from Stiles and Skutch 

(1989) and verified with the American Ornithologists' Union checklist (American 

Ornithologists Union 2005). Habitat associations were taken from Hughes et. al. (2002). 

When a species was not listed in this source, habitat association was assessed using 

information from Stiles and Skutch. Birds were categorized as being primarily associated 

with forest or open areas, or as being normally present in both, and relative percentages 

of both individuals and species associated with each habitat type were compared for 

pasture and coffee. The feeding guild for each species was determined using information 

from Stiles and Skutch (1989). 

Survey data were analyzed using Estimates software (Colwell 2005b). Estimates 
uses field data to compute randomized species accumulation curves, eight statistical 
estimators of species richness, and three diversity indexes. These statistics were 
computed for each individual plot and for each site (pooling the samples from the two 
plots that made up each site, for a total of 12 samples per site). According to Engstrom 

117 



and James (1981; cited in Gotelli & Colwell 2001), a set of replicated samples can be 
pooled and treated as a single data set. Richness estimators and diversity indexes were 
therefore also calculated with all surveys pooled for each habitat type (48 surveys per 
habitat type). 

Pielou (1975 pp. 17-18) discussed the idea of hierarchical diversity, which 
examines the diversity of a community at taxonomic levels above that of species. Later 
work expanding on this concept is described in Magurran (2004 pp. 121-123). 
Communities with similar numbers of species are considered more diverse if those 
species belong to a greater number of higher taxonomic groups. In other words, a 
community with two species is more diverse if those two species belong to separate 
genera than if they are congeneric, more diverse still if the two species belong to separate 
families, and so on. To examine this aspect of diversity, richness and diversity indexes 
were computed at the family as well as the species level. 

Statistical analyses were performed using SPSS 10.0.5 and Microsoft Excel 2003. 
See Appendix B for a discussion of non-sampling error sources and their importance to 
the results. 

Richness refers to the actual number of types of organisms present at a site, at a 
pre-determined taxonomic level (such as species, genus or family), and is the simplest 
and most intuitive way of quantifying biodiversity. Observed species richness was 
compared using a 1 -way, 2-group ANO VA, with surveys for plots A and B pooled for 
each site, and each site treated as a single subject, resulting in four subjects per group. 

In general, it is not possible to sample all the species that may use a given site, or 
be certain that all species have been sampled (Chao 2004). Instead, it is necessary to use 
statistical estimators of true species richness. In addition to observed species richness, I 
selected six richness estimators, based on those used in Hughes et al. (2002): 

1. Incidence-Based Coverage Estimator (ICE) of species richness 
(Chazdon et al. 1998; Lee & Chao 1994; cited in Colwell 2001) 

2. Chao 2 richness estimator (Chao2) (Chao 1987; cited in Colwell 2001) 
118 



3. First-order Jackknife richness estimator (Jackl) (Burnham & Overton 
1978; 1979; Smith & Belle 1984; Heltshe & Forrester 1983; all cited in 
Colwell2001) 

4. Second-order Jackknife richness estimator (Jack2) (Palmer 1991; 
Smith & Belle 1984; all cited in Colwell 2001) 

5. Bootstrap richness estimator (Bootstrap) (Smith & Belle 1984; cited in 
Colwell 2001) 

6. Michaelis-Menten richness estimator (MMMean) (Raaijmakers 1987; 
cited in Colwell 2001) 

While richness is a simple count of the total number of taxonomic units at a site, 
diversity takes into account both the richness of a site and the relative abundance of the 
different taxonomic units. A site dominated by one or a few types of organism is 
considered to be less diverse than a site with equal numbers of different types of 
organisms, even if both sites have the same richness. 

Estimates computes three diversity indices based upon richness and relative 
abundance: 

1. Fisher's Alpha diversity index (Hayek & Buzas 1996; cited in Colwell 2001; 
Magurran 1988) 

2. Shannon diversity index (Hayek & Buzas 1996; cited in Colwell 2001; 
Magurran 1988) 

3. Simpson reciprocal diversity index (Hayek & Buzas 1996; cited in Colwell 
2001; Magurran 1988) 

The three diversity estimators were compared using a 1-way, 2-group ANOVA, 
with surveys for plots A and B pooled for each site, and each site treated as a single 
subject. According to Sokal and Rohlf (1995; cited inMagurran 2004 p. 151), ANOVA is 
an acceptable method for comparing diversity indexes when multiple samples have been 
taken. 

119 



Results 

A total of 1,896 registrations of 62 species from 25 families were made during the 

surveys, with 1,078 registrations of 45 species from 17 families in pasture, and 764 
registrations of 48 species from 22 families in coffee. Thirty-two species and fifteen 
families were shared by both habitat types, with fourteen species and three families 
detected only in pasture, and sixteen species and seven families detected only in coffee. 
Table 2.4.5 lists all species and families observed, total counts of each species observed 
in each habitat type, and habitat association and feeding guild of each species. 

In pasture, fourteen species and two families were represented by only one 
individual. In coffee, this was the case for twelve species and three families. Rare 
species, defined for this study as species represented by two or fewer individuals across 
all samples, accounted for 31.3% of coffee species and 33.3% of pasture species. As 
many as five near-threatened or vulnerable resident species may reside, or have resided at 
one time, in the study area (IUCN 2004; Stiles & Skutch 1989). None of these species 
were detected during the surveys. Furthermore, none of the four species of understory 
insectivores identified by §ekercioglu et al (2002) as particularly sensitive to forest 
fragmentation in the study region were detected. 

Despite the relatively equal numbers of rare species in coffee and pasture, pasture 
was markedly dominated by birds of a few species from the family Emberizidae 
(sparrows and finches), whose members made up 64% of all registrations, with all other 
families having 7% or fewer registrations each. There did not appear to be any one 
dominant family in coffee. The most abundant family in coffee was the Thraupidae, with 
22% of registrations, followed by the Troglodytidae at 16%, and the Parulidae, 
Trochilidae, and Tyrannidae, at 9-10% each. These observations indicate higher diversity, 
defined as evenness of species composition, in coffee at the family level. The percent of 
registrations for species belonging to each family detected is shown in Figure 2.4.12. 
These qualitative observations were supported by the quantitative analysis below. 

Total observed richness was higher in coffee for both species and families. Figure 
2.4.13 and Figure 2.4.14 show richness estimates for bird species and families, 
120 



respectively, derived when all samples are pooled for each habitat type (e.g., 48 samples 
per habitat type), as well as the observed richness of species and families at the pooled 
sites. No consistent pattern emerges for richness at the level of species. The six richness 
estimators differed in which habitat type they predicted would have a greater total 
richness. The ICE, first-order Jackknife, Bootstrap and Michelis-Menten estimates 
predict higher species richness in coffee, while the Chao2 estimate and the second-order 
Jackknife estimate predict higher species richness in pasture. 

Observed species richness did not differ between habitats (one-way ANOVA, 
F=4.573, df=7, p=0.76). These results, combined with the lack of consistency among the 
other estimators, means that it cannot be concluded that there is a significant difference 
between the species richness of the coffee and pasture sites. At the family level, 
however, coffee sites consistently showed higher predicted richness for all estimators. 
Observed family richness between habitat types was significantly different (one-way 
ANOVA, F= 6.750, df=7, p=0.041). 

Diversity was higher in coffee for both species and families, as measured by 
Fisher's Alpha, Shannon and Simpson Diversity indexes (Figure 2.4.15 and Figure 
2.4.16). For species, the Shannon and Simpson indexes were significantly higher 
(p<0.05) for coffee than for pasture, but the difference in the Alpha index was not 
significant. For families, all three indexes were significant at the p<0.05 level (see Table 
2.4.6 and Table 2.4.7 for summaries of F-values and significance levels for these three 
tests). 

Not surprisingly, pasture showed a higher percentage, in terms of both numbers of 
registrations and number of species, of birds associated with open habitats, while coffee 
showed a higher proportion of birds associated with forested habitats and with both 
forested and open habitats (Figure 2.4.17 and Figure 2.4.18). Number of registrations 
was significantly different for habitat association between pasture and coffee (% =366.58, 
df=2, p=0.000). 



121 



This same pattern held for the habitat associations of species (as opposed to 
numbers of registrations). A/ test could not be performed on the numbers of species by 
habitat association, because the expected values for forest species were below 5 for both 
pasture and coffee. To enable a Fisher's Exact Test, which requires a 2x2 table, the 
number of open-country species found in coffee and pasture was compared to the number 
of forest species in each, with generalist (species associated with both habitat types) 
excluded. This comparison revealed no significant differences between the habitat types 
(N=57, p=0.408). Of the fourteen species found only in pasture, one (7%) was associated 
with forest, ten (71%) with open areas, and three (21%) with both. Of the sixteen species 
that were found only in coffee, three (19%) were associated with forest, four (25%) with 
open areas, and nine (56%) with both. 

Percent of registrations from each habitat type by guild is shown in Figure .4.19, 
and Figure 2.4.20 shows percent of species by guild. Pasture shows a striking dominance 
by granivores, while coffee has twice the proportion of insectivores and frugivores, and 
five times the proportion of nectarivores, as pasture (% , df=5, p=0.001). Species 
composition by guild appears to be fairly similar between habitat types, although small 
expected values preclude testing with a/ test. 

Of those species that were detected only in pasture, one (7%) was a frugivore, one 
(7%) was a granivore, ten (71%) were insectivores and two (14%) were omnivores. Of 
those species that were only found in coffee, one (6%) was a carnivore, five (31%) were 
frugivores, one (6%) was a granivore, seven (44%) were insectivores, and two (13%) 
were omnivores. Four of the seven species of insectivores found only in coffee, none of 
the ten found only in pasture, and one species found in both habitats, could be classified 
as understory insectivores. 

Discussion and Conclusions 

The lower diversity in converted coffee (pasture), the lower taxonomic richness 

(family richness) and greater proportion of open-habitat birds in pasture, the higher 
proportion of insectivores in coffee, and the presence in coffee of understory insectivores 

122 



not found in pasture support the argument that conversion of coffee farms to pasture will 
have a negative impact on biodiversity and habitat availability in the region. Based on 
these results, H2, which states that "coffee farms that have been converted to cattle 
pasture will show a lower bird diversity and richness than coffee farms that had not been 
converted," is partially supported. 

Species richness did not differ significantly between the two habitat types. Initial 
power analysis prior to the surveys indicated that four sites per habitat type was an 
adequate sample to test for species richness, but this power analysis relied on an estimate 
for the within-group variation that was too low, and an estimate for the difference 
between means that was too high. A post-hoc power analysis for the ANOVA of 
observed species richness reveals that the probability of detecting a real difference 
between means at the 0.05 significance level is less than 50% in this case. To have an 
approximately 80% chance of detecting a true difference in observed species richness 
between the two treatment types, a sample size of 8 sites per group would be required 
(Dunlap 1981). 

A number of characteristics of the bird communities in the study sites are of 
interest from a conservation standpoint. A significantly higher proportion of registrations 
in coffee were forest-dependent species, while pasture was dominated by open-habitat 
species. Additionally, coffee had twice the proportion of registrations that were 
insectivores, four species of understory insectivores were detected in coffee and not 
pasture, and only one species of understory insectivore was detected in pasture. This 
species was also detected in coffee, and no species of understory insectivore was found 
only in pasture. Insectivores, particularly understory insectivores, are the guild of birds 
shown to be most sensitive to forest fragmentation in the study region (Sekercioglu et al. 
2002) as well as other parts of the Neotropics (discussed further in Chapter 1.2). 

Many of the study sites were adjacent to forest patches. During the surveys, many 
birds, particularly those that were only detected once or twice, were observed flying from 
forest patches briefly into the survey plots, then returning to the forest. Thus, it is likely 

123 



that many of the birds relied upon these patches as their primary habitat, visiting the 
agricultural matrix to forage or to move between patches. 

These observations, combined with the difference in guild composition and 
habitat association discussed above, indicate that a changeover from coffee to pasture is 
likely to reduce important matrix habitat for forest-dependent species, and possibly 
reduce the mobility of birds between forest patches in this highly fragmented landscape. 
Other authors, discussed in Chapter 1.2, have shown that shade coffee can help provide 
connectivity for forest-dependent animals. The results of this study indicate that this 
could also be true of the relatively unshaded systems that predominate in Agua Buena. 

This study, however, does not provide information on a number of other 
variables. For example, as described in the previous section, most coffee farmers in the 
area had reduced or eliminated their pesticide use due to low coffee prices, but in times of 
higher prices, coffee is typically a high-input crop. During interviews, some farmers 
mentioned finding dead birds in their fields after applications of the nematicide terbufos, 
which all respondents but one had subsequently stopped applying - some due to an 
inability to continue to pay the high price of the chemical, others due to its perceived 
toxicity. Thus, the impacts on bird populations from lower pesticide use could be 
significant. 

Additionally, these surveys only looked at one alternative to coffee. Other land 
uses were not examined. For example, some farmers reported abandoning their land 
rather than converting it to an alternative use. This land abandonment, leading to 
regeneration of secondary forest, is likely to have a positive environmental effect. 
Alternatively, some farmers were changing to vegetable crops such as tomatoes and 
green peppers. These crops are reported by farmers to be more chemically intensive than 
coffee, and the soil where they are grown is typically kept bare, potentially leading to 
higher erosion than in coffee. Thus, conversion of coffee to these types of vegetable 
crops can be expected to have negative environmental effects. 



S-tert-butylthiomethyl 0,0-diethyl phosphorodithioate 
124 



While it is, therefore, not possible from these data to construct a complete picture 
of how the varied choices made by farmers in the Agua Buena region in response to low 
coffee prices are affecting biological diversity in the area, this research does illuminate an 
important component of land-use change. Interviews with farmers showed that pasture 
was by far the most popular alternative to coffee. Given the scale of conversion to 
pasture identified by the interviews, and the lower family-level taxonomic diversity 
observed in converted farms, the apparent reduction of habitat resulting from conversion 
should be cause for concern. 



125 



Figures 




Figure 2.4.9. Example of coffee site (site shown is Coffee 3). 




Figure 2.4.10. Example of pasture site (site shown is Pasture 2). 



127 



'■ ; : 0*. 






Pasture sites 



Coffee sites 




j 







<VS 






N? 



%\ 



'tiu *-M. 



■. 







>1 
u 

.-■ 






•- 

3 

WO 



I) 
M 

cd 
a. 
co 

3 

o 

'> 
u 





0) 
0) 




4l- 




Cm 




O 




u 




0) 




H 




s 




■** 




t/> 




« 




Ph 




S3 


■*J 


O 


« 


■M 


■** 


« 


A 


u 


C8 


o 


X 


C/3 
C/3 




«< 


ox 




= 


T5 


T3 




a* 


U 


a* 


6J 


ta 





a> 

s 

a 

S3 

S3 
O 

s 
s 

o 

u 



0) 

s 

s 

S3 

C/3 

.2- 

0) 

& 

in 



55 



Ph 



CQ 



Ph 



Co 

u 

eg 
"eg 

-a 
u 

CO 

-a 
U 

13 
03 

13 
eg 
03 

43 

I 

eg 
— 

a 






00 



CQ 



2 
o 



u 



u 



o 



<N 



Ph 



03 
> 

O 
13 

I 

13 

a 

o 
•— 

a 

13 
13 



o 
a 

+■4 

•S 



a 



•2 
ft, 

2 



CQ 



a 



CQ 



Ph 



> 

o 

13 

13 
03 
&! 
CM 



U 



ft 



o 
s 

s 



CQ 



On 



O 

o 
u 

U 

"3 

'3. 
ai 

CO' 



-s; 

2 



CN 



On 



1) 

ai 

eg" 

•— 

3 

eg 

HH 

e 

o 

E 

E 
o 
U 



Ph 



5 

o 

-s; 



CN 



CQ 



00! 



CQ 



00! 
C 

s 

3 

X 

13 
CU 



CO 

a 
o 



CQ 



CQ 



— 
03 

o 
03 

Oh 

13 
O 
O 



a, 

•2 
'C 
-a 
s 

2- 

c 



CQ 



u 

u 

& 

O 
O 



>5 

s 

o 



























a 






































eg 














0) 






u 


eg 
13 

•a 


u 

u 

eg 
13 


u 

eg 


u 

eg 

2 


eg 

3 


u 

eg 
13 


u 

eg 
13 


<D 
eg 
13 


"2 
E 


eg 

3 


eg 

3 


eg 

3 


u 

eg 

3 


03 

eg 


03 ! 
eg 


g 


13 


-a 


H 


&H 


o 
o 

r f3 


E 


E 


E 


o 

eg 


3 

o 

3 


3 

o 

3 


43 


43 


43 


43 


c« 


•— 


13 
— 


CO 


o 
o 




J3 

'o 


'o 


■4-» 


•— 

Oh 

eg 


o 

— 


o 


o 

S-H 


o 

— 


(J 


U | 


H 


U 


<l 


U 


< 


Ph 


U 


U 


u 


Ph 


u 


u 


u 


H 


H 


H 


H 


Ch 


Cm 



c 
o 



.o -S 



cd 

-^ Ph ^ 

cd , t3 

-*! S JS 

S m S 

«o« 

4-1 „ U 

■30 a 

l> M3 GO 

^ cd 

J O o 



cd o 



3 — 

43 43 a 

P tH O 



o 
> n 

Ph -S 



cd o 

S 



-3 a 



T3 



O C 

43 II 



3 

c 

U .Hj 



4H 



T3 
O 

Cd 

E 

CD 

O ^-, 

-S ON 

^ °° 

T3 On 
cd ^ 

Sa 

3 3 

cd CO 



- cSh 



"H cd 

jj> CO 

C CJ 

O -Q 

3 E 
O O 



CO 



a 

o 

cd 

E 

Uh 

eg 

C 

a 

o 



,-H °° 

Id 2« 

s -M 

I) ^ 

CO *H 

CO Ph 

'" _• 

co t3 

c^ .„ 

C/5 3 

iri 00 

rr .S 

u c2 
«■§ 



cd T3 

p CO 

C co 

co cd 

CO 43 

'O co" 

CO u 

^ is 

co O 



-3 .CO C 

O 



■ CO 

4h t3 



II £ 



131 



03 


o 


IT] 


^H 


^H 


W1 


00 


^H 


tI- 


o 


o 


wi 


o 


o 


^ 


CN 


vri 


vri 


CN 


CN 


^H 


^ 


o 


t- 


Si 




,—l 












-* 
















^ 


CN 


iO 


CN 


cn 


CN 


m 


^f 


O 

U 
















































03 


00 


^t 


o 


o 


ON 


O 


o 


CN 


v£> 


^H 


^H 


CN 


^H 


oo 


O 


oo 


o 


CN 


CN 


o 


-* 


o 


^D 


'— 

3 

73 


m 


































Tl- 












S3 
















































*- 2 
















































.-§ « 
















































* "3 


O 


O 


m 


O 


O 


PQ 


Ph 


O 


o 


Ph 


pq 


O 


m 


O 


Ph 


O 


m 


o 


m 


m 


m 


O 


m 


55 O 
















































K % 
















































< 
















































ex 
















































■9 2 
















































-o ... 

0) S 










































o 


HH 


z 


a) ex 
















































ta 






j_, 


































43 

s 

43 








0) 


'3 

H— » 

03 

'Eh 




03 

Oh 
03 

03 

tH 

C3 

13 
O 
O 


03 

3 


CO 

'3 

03 

CO 

w 


03 


CO 

1 

E 

00 


03 
43 

O 
^-^ 

CO 

u 

1 


03 
43 

O 
^-^ 

CO 

o 
>> 


03 

43 
O 

-^ 
CO 

o 

E 


S-H 

03 
43 
O 

CO 

o 


•— 

03 


tH 

03 
43 
O 

CO 

u 

>> 


T3 
•— 


-4— » 




s3 
03 

S-H 

13 




s^ 
03 
43 
O 

CO 


1 

03 

lo 

00 

_s3 
'h-j 
43 
00 

Z 

13 

03 

42 
i 

03 
00 


_S3 

42 
O 


S-H 

03 
CO 




E 

es 

S3 

S3 
O 

s 


00 

13 
03 

H— » 

CO 

03 
— 

42 


'3 

H— » 

13 

_s3 

'el 

00 


13 

03 
13 
CO 
03 

43 

1 

44 


$3 

CO 
— 
>> 

H 


13 

_03 

13 

i 


3 
S3 

e 

CO 

>> 

H 
>> 


13 
03 
-^ 

03 
'— 
O 
i 


13 

O 

H 

c 
o 

i 


E 

13 

03 

o 
o 


13 
03 
Oh 
Oh 
CO 
O 
i 
>> 


E 

13 

03 


43 
O 

CO 

o 

>> 

E 

lo 

'o 

o 


E 

13 
03 
Oh 
Oh 
CO 
O 
i 


42 
00 

S3 

'cO 


03 

"E 

03 
03 

S-H 

o 

S-H 

03 

CZl 

c« 

03 


s3 
03 

S-H 


03 

CO 

03 

S-H 

42 

1 


s3 
03 

S-H 

03 


o 

CO 

S3 

o 

r c0 


Pi 

13 

03 

S-H 

_o 

O 

o 


13 

03 

CO 

o 

1 


'3 

CO 

S3 


s 

o 


i 

03 

r co 


>> 

H— » 

CO 


CO 
03 
'— 

-4— > 


o 
13 


o 

13 


H 

r c0 


03 

r c0 
C3 


E 
o 


CO 

S-H 




H— » 

CO 
O 


>> 

a 


'Eh 

o 

S-H 


s3 
'3 




o 


'Eh 
O 

S-H 


s3 
CO 

S-H 


CO 




CO 
S3 
CO 


u 


Ph 


00 


00 
* 


>" 


>H 


Ph 


00 


U 


m 


Q 


m 


in 


O 


H 


hJ 


E 


rt 


X 


H 


o 

* 


U 


rt 


m 


a 

a 

a 

73 
03 

u 

03 

in 


e 

' 

' — - 

s 


s 

' — - 

e 


' — > 

s 

era 
crj 
Su 

■K» 

' — ^ 
-J 

03 
CO 

"2 

Oh 


o 

I 

a. 


•2 
2 
*> 

"■—a 


3 

£ 

! — * 

•2 

£ 
N 


* 

&a 
S 

■Kit 

'5, 

&a 
S 


S 
s 

2i 

•S 

s 

s 

*^ 
<< 

2 

e5 


en 
S 

•2 

en 

"a 
S 

t 


1 

' — - 

s 

2 

S 

&a 
S 

t 


s 
e 

'a, 

&a 
S 
-Si 

e 

1 


&a 
+»it 

& 


c 

s 

&a 

c 

1 


S 

' — > 

s 

' — > 

S 

&a 
S 

e 

s 

2 


H-* 

s 
"a 

&a 
S 


S 

"a 
S 

&a 
S 

03 
CO 


* 

&a 
S 
! — * 

■Ka 

s 

en 
S 

03 
CO 


s. 

•% 

to 

03 
CO 


£ 
s 

t 


*^ 

e 

2 

s 

&a 
S 

Sk 


'1 
k. 

S 


&a 
S 

!-. 

S 

■Kit 

s 


o 




43 
CO 


03 

CO 

"2 


CO 


03 


03 


03 


03 


03 


03 


03 


03 


03 


03 


03 


03 


"2 


"2 


"2 












>* 


O 

o 

a 


CO 

"2 
'3 


CO 

"2 

'3 


CO 

"2 

'3 


CO 

"2 

'£ 


CO 

"2 


CO 

"2 

'B 


CO 

"2 
'B 


CO 

"2 
'B 


CO 

"2 

"B 


CO 

"2 
"B 


CO 

"2 


CO 

"2 

'S 


13 


13 

O 


"-4— » 

13 

O 


03 
CO 
13 


03 
CO 
13 


03 
CO 
13 


03 
CO 
13 


03 
CO 
13 


• PN 


CO 


CO 


13 


CO 


CO 


S3 


CO 


Ch 


CO 


CO 


CO 


CO 


CO 


CO 


o 








!^^ 


•i-H 


•i-H 


J_^ 


J_^ 


a 


e 


e 


H 


H 


CO 


H 


CO 


H 


H 


H 


H 


H 


S-H 


03 


r ob 


r ob 


r ob 


> 


13 


13 


s 


s 


S-H 


S-H 


«H 


C 


S-H 


S-H 


«H 


«H 


«H 


C 


C 


C 


«H 


C 


o 


o 


o 




S-H 


S-H 


03 


=3 


j3 


03 


>> 


>> 


>> 


>. 


>. 


>. 


>, 


>. 


>. 


>. 


>. 




S-H 


S-H 


S-H 


^ 


=S 


=S 


CO 


CO 


fe 


e 


e 


Q 


H 


H 


H 


H 


H 


H 


H 


H 


H 


H 


H 


> 


H 


H 


H 


00 


H 


H 


Oh 


Ph 



3 

5 

a 
o 
U 

iri 

— 

S 
cs 

H 



132 



03 


•o 


ON 


cn 


o 


en 


^H 


^ 


CN 


CN 


CN 


On 


CN 


t- 


o 


o 


o 


o 


^_, 


CN 






-xt" 


^o 




m 


CN 




- H 






^ H 


^ H 






^ H 








CN 


t4H 








































o 








































U 








































<u 


o 


_H 


"3- 


,_, 


r- 


00 


MD 


t- 


_H 


O 


^D 


CN 


IT) 


_H 


o 


m 


00 


o 


00 


'— 




m 


CN 








^O 


^H 


C3N 




O 


CN 


00 




^H 










3 














CN 


"- ' 






"- ' 


















73 








































« 








































CM 








































S3 








































*- © 








































55 fa 








































s « 








































-2 "3 


m 


O 


O 


O 


PQ 


fa 


o 


o 


o 


O 


o 


o 


o 


O 


PQ 


O 


O 


fa 


o 


C5 O 








































W g 








































^ 








































ex 








































■9 2 








































-c .-a 

0) S 


fa 


fa 


fa 


fa 


fa 


fa 


o 


o 


o 


O 


o 


o 


O 


^^ 


O 


O 


O 


fa 


a 


a) ex 








































fa 








































03 

a 

C3 


a 

00 

CO 

$3 

CO 

H 


— 

u 

00 

CO 

e 

cO 

H 

13 
03 

Oh 


— 

03 
00 

CO 

S3 
CO 


— 

03 


— 

03 
00 
CO 

a 

CO 

H 

13 

03 


— 

03 
00 

CO 
S3 
CO 

H 

13 

03 
13 


'3 

CO 
— 


— 

03 
-^ 
CO 
03 
13 
03 


— 

03 

CO 
03 
13 

03 
03 

in 

13 
_03 


43 
O 

a 
fa 

I 

13 

03 
03 


'3 

CO 

o 

13 


o 

id 

CO 

Oh 

00 
13 


o 

g 

CO 

Oh 

in 

13 

03 

CO 


o 

CO 


— 

o 

CO 

lo 

00 
13 


03 

o 

CO 
— 


13 
— 

42 


CO 

'3 
o 

43 

Oh 

3 
fa 


43 
O 

S3 
cfi 


S3 

S3 
O 


13 
03 

_$3 

1 

co 

-4— » 


E 

— 
i 

-^ 

03 


H 
>> 

CO 

'— 
00 
i 
03 


00 

CO 

a 

CO 

H 
E 

'cO 


CO 
O 
— 

-^ 
i 
t— 

03 


O 
O 

43 

I 

a 

03 

2 


M 
o 

CO 

i 

03 


03 

on 

03 
CO 


13 

42 

1 

O 


13 

03 

42 

1 

M 
o 


03 
O 

1 

o 


03 

Oh 

■c 

-4-* 
CO 

1 

M 
o 


o 

1 

3 


'cO 

m 

13 

03 
M 
CO 


03 
CO 
O 
— 

43 

-^^ 

i 


13 

03 

'3 

-^^ 

i 

CO 


O 

U 

13 

03 
N 
S3 


13 

03 

42 

1 

44 
O 


2 
'o 
O 
— 

03 


O 


S 

£ 


3 


J3 


> 


o 


J3 


"h 

CO 


13 


43 


13 


CO 


4-H 

3 


03 
— 

H i 


3 


03 
— 


O 
— 


43 


03 


U 


oo 


m 


fa 


OO 


O 


m 


> 


Jh 


H 


>H 


2 


fa 


00 


fa 


o 


PQ 


H 


fa 






■5 






' — - 






G 


.=n 


u 




'C 








a 




o 




a 

53 
S3 

73 
.2- 


u 

co 
3 
S 

o 

-s; 


co 

s 

' — > 

Co 


&9 

s 

to 

5. 

•2 


s 

s 

s 

' — - 

a, 
.j2 


-s: 

o, 

5r 
co 
o 

cu 
.Co 


5s 


•5 
Co 

•2 


s 
'C 

s 

* — » 


'C 

•2? 
S 


s 

S 


S 

o 
'C 


s 
s 

o 

to 

e 


e 

^H 

•2 


Oh 

H-Hi 

cu 

■S- 

•2 

en 


S 

S 

s 


S 

c 

£ 

S 


to 
S 

e 

to 
S 


# Co 

'C 

H-Hi 

to 
O 
>. 

'S 

•2 
S 
o 


CO 

a, 

' — ^ 


t 


1 


5. 




60 




•1 






o 

In? 




'C 

H-H. 

o 






to 


-s: 

H-H. 


s 

13 


03 


co 


s^ 


k. 


e 


e 


' — - 


o 


o 




i. 


e 




' — - 


s 




S- 


s^ 


in 


£ 




g 


g 


e2 


e2 


^ 


CXh 

CO 


^H 


i. 

^ 


(3 


CO 


CO 




5s 
















03 


03 


03 


03 


03 


03 


03 


03 


03 












03 


03 


03 


03 


03 


03 


CO 


CO 


CO 


CO 


CO 


CO 


CO 


CO 


CO 






03 


03 




cO 


cO 


CO 


CO 


CO 


CO 


3 


3 


3 


3 


3 


3 


3 


3 


3 






CO 


CO 


>* 


3 

'Eh 


3 

'E 


3 

'E 


3 

'E 


3 

'E 


3 

'E 
















r c0 

S3 


'cO 

S3 


Cfa 
03 
CO 

3 


03 
CO 

3 


13 


13 


rs 


3 


3 


3 


3 


3 


3 


03 


03 


03 


03 


03 


03 


03 






'5b 

_S3 


'5b 

_S3 


a 

S 


03 
— 

43 


CO 

'— 

-a 


CO 
'— 

-a 


CO 

S-H 

43 


CO 

S-H 

43 


CO 

S-H 

43 


43 

E 


42 
E 


42 
E 


42 
E 


42 
E 


42 
E 


42 
E 


■3 

H 
CO 


■3 

H 
CO 


03 


03 


fa 


H 


H 


H 


H 


H 


H 


fa 


fa 


fa 


fa 


fa 


fa 


fa 


u 


u 






fa 


fa 



3 

C 

c 
o 
U 

iri 

— 

cs 

H 



133 



CD 






CI) 


3 


CD 


CO 


^ 


ctf 


o 


Q_ 


O 


□ 


■ 



3 

o 



o 




^ £ o g- 

^ o J= 2 

m o "> en 

10 o o c 

Si- c o 



I 

J: 



(M 



o 

(M 



o 



o 
o 



o 
in 



suojiB-usiBay jo weojed 



5°^ ^ 



*r 



SU 






3> 









Q 



«>. 



o 
o 



\ 
\ X 

w 
X 



c 



o 






-J 

C3 
T3 



1) 






a. 

o 

f- 



T3 

C 



a 
a. 



s 




cl 








>1 




J-l 




an 




e 




o 












-. 












M 




on 


OJ 


u 


-O 




I-i 


■ — | 


T3 


^ 


::", 


4h 


O 


o 


c 




3 


c 
■J 


EM 


■-• 


rri 


0) 


X) 


0. 


J) 






«N 


rl 


l—l 


P 


T 


GO 



"21 

fa 6 



135 



80 - 
75 - 
70 - 

CO 
CO 

CD 

c 65- 

u 

DC 60 - 
CO 
CD 

'8 55 - 

Q. 

CO 

50 - 

45 - 
40 - 
















O- 

— •— 


• Pasture 
- Coffee 










.O. 




,o 










'*. 














// * \ 


// ' : ^S\ 


ci "0 




Observed 


ICE 


Chao2 


Jackl 


Jack2 


Bootstrap 


MMMean 



Figure 2.4.13. Species richness estimates for pooled samples. "Observed" refers to the total number of 
species actually observed during surveys. See text for discussion of richness estimators denoted in legend. 





35 




33 




31 




29 


(/> 




(A 




<1> 
C 


27 


£ 




o 




cr 


25 


>. 




F 


23 


ro 




u. 






21 




19 




17 




15 




... - -o... 



rvO^^ 



O^ 



...-o ...... 



O" 



"O- 



-o 



-O - - Pasture 
-• — Coffee 



Observed ICE Chao2 Jackl Jack2 Bootstrap MM 



Figure 2.4.14. Family richness estimates for pooled samples. "Observed" refers to the total number of 
families actually observed during the surveys. See text for discussion of richness estimators denoted in 
legend 



137 











Diversity Index 

3 Ul O Ul O t 






23.02 




■ Pasture 

■ Coffee 
















11.37 






1 




9.49 






9.49 












■ 








2.75 3.33 








i I 


I 




Alpha Shannon 


Simpson 





Figure 2.4.15. Species diversity indexes for samples with all sites pooled. 



10 



x 

C 



<2 

CD 

> 

Q 2 



8.76 



4.23 



2.87 



1.49 



2.38 



□ 



2.37 



Alpha 



Shannon 



Simpson 



□ Pasture 
■ Coffee 



Figure 2.4.16. Family diversity indexes for samples with all sites pooled. 



139 



Index 


Mean- 
Pasture 


Mean - Coffee 


dftot 


F 


Sig. 


Alpha 


7.8000 


10.8775 


7 


3.300 


.119 


Shannon 


2.5675 


3.1050 


7 


*14.919 


.008 


Simpson 


9.1650 


19.8150 


7 


*32.964 


.001 



Table 2.4.6. Results of ANOVA for the Alpha, Shannon and Simpson diversity indexes for species in each 
site. F values with an asterisk are significant at the p<0.05 level. 



Index 


Mean- 
Pasture 


Mean - Coffee 


dftot 


F 


Sig. 


Alpha 


2.7825 


4.0025 


7 


*10.376 


0.018 


Shannon 


1.4100 


2.2800 


7 


*23.877 


0.003 


Simpson 


2.7350 


8.0325 


7 


*44.018 


0.001 



Table 2.4.7. Results of ANOVA for the Alpha, Shannon and Simpson diversity indexes for families in each 
site. F values with an asterisk are significant at the p<0.05 level. 



141 



Percent of Registrations by Habitat Association 



1 00% 



V) 

80% 

S 

.S2 60% 
<u 

DC 

•5 40% 

1 20% 

Q> 
Q. 



0% 



91.7% 



52.9% 



44.1% D Pasture 

■ Coffee 



0.7% 



3.0% 



7.5% 



Open 



Forest 



Both 



Figure 2.4.17. Relative abundance of birds recorded in survey plots that, according to Hughes et al. (2002) 
and Stiles and Skutch (1989), are normally associated with forested vs. open habitats. 



Percent of Species by Habitat Association 



/u/o 




62.2°/c 


> 










cn°/ 










DU /o 
CO 

.9} crw 




51.1°/c 


46.7% 


J3 oUvo 














CD 








(/) 4U7o 
O 






33.3% 






■g OU70 
CD 








£ 20/ ° 
D. 




8.9% 




lUvo 
0% 




4.4% I 





□ Pasture 
■ Coffee 



Open 



Forest 



Both 



Figure 2.4.18. Percent of bird species recorded in survey plots that, according to Hughes et al. (2002) and 
Stiles and Skutch (1989), are associated with forested vs. open habitats. 



143 



Percent of Registrations by Guild 



70% 
60% 
50% 
40% 
30% 
20% 
10% 
0% 



63% 



27% 



1 3% 



29% 



15% I 
1 2% i — M 



25% 



5% 



4% 



7% 



□ Pasture 
■ Coffee 



Frugivore Granivore Insectivore Nectarivore Omnivore 



Figure 2.4.19. Percent of bird registrations in each feeding guild. Carnivores are not shown because their 
bars are too small to see. A total of three carnivores were observed in all survey sites. 



Percent of Species by Guild 



50% 



40% 



30% 



20% 



10% 



0% 




□ Pasture 
■ Coffee 



Carnivore Frugivore Granivore Insectivore Nectarivore Omnivore 



Figure 2.4.20. Percent of species in each feeding guild. 



145 



III. Conclusions and Alternatives 

/. Environmental consequences 

There are three main trends that reveal themselves across scales when we examine 
the responses of farmers to low coffee prices: reduction of chemical and labor inputs, 
changing to different crops, and land abandonment. No one of these factors has easily- 
quantifiable environmental consequences, although this thesis has partially illuminated 
one aspect of the last. 

Land abandonment may lead to secondary forest regeneration on the abandoned 
plot, but the abandoned land can just as easily be sold cheaply for conversion to pasture 
or, in population-dense areas such as the Central Valley of Costa Rica, for development. 
Land abandonment and the reduction of labor inputs also contribute to urbanization and 
its accompanying environmental impacts. According to Varangis et. al. (2003 p. 18), 
abandonment of coffee trees or the elimination all chemical controls (when not replaced 
by appropriate, non-chemical technologies) can lead to the spread or worsening of pests 
and diseases, which could further lower profitability for neighboring farmers, or force 
them to further increase their own chemical applications. 

The reduction of chemical inputs has clearly positive environmental effects. This 
strategy, however, may be a short-term one, serving as a transition to land abandonment 
or a changeover to another crop, as the yields typically obtained without chemicals and 
without an increase in labor will not provide a sufficient income for most farmers. 
According to Anthony Winson (1989 p. 115), a Costa Rican farmer needed 4 hectares of 
coffee or more at Costa Rica's high average yields and 1980's prices to provide a 
subsistence income for a family. With the average coffee farm size in Coto Brus being 
3.9 ha (Sandoval 2002 p. 4), and the median size from the surveyed Agua Buena 
households at 3.0 ha, it can be anticipated that most farmers will not be able to continue 
living off income from coffee in the long term at post- 19 89 price levels. This could lead, 
in turn, to more land abandonment or a changeover to other, less sustainable crops, such 
as pasture, tomato and green pepper. 



147 



Conversion to pasture, at least in the Agua Buena region, has clear negative 
effects on the availability of habitat for bird communities, as demonstrated in the 
previous chapter. Pasture may still, however, be a preferable alternative to other more 
chemically intensive land uses, such as the production of tomatoes, green peppers, and 
other commercial vegetable crops. Given that all pasture sites in this study retained many 
of the shade trees originally planted over the coffee, it could be illuminating to conduct a 
comparison of pastures that had never been coffee, to assess whether tree cover was 
similar, and whether a difference in tree cover affected habitat availability. Luck and 
Daily (2003) found that isolated fruiting trees were important resources for avian 
frugivores in Coto Bras. Encouraging farmers who are changing from coffee to leave the 
shade trees may help to lessen the impact of conversion. 

2. Alternatives and solutions 

Since farmers can respond rapidly to lower prices by lowering agrochemical 
inputs, which will have impacts on production by the next harvest, the boom-and-bust 
cycles of coffee production have become shorter since the advent of the intensification 
era (Daviron 1994 p. 48). Thus, the extremely long duration of the current cycle of low 
prices is cause for concern, as it indicates that there is more at work that another iteration 
of the normal economic coffee cycle. Many coffee experts do not expect prices to ever 
fully recover to their pre- 19 89 levels. Nevertheless, volumes have been written about the 
coffee crisis, and many solutions have been proposed from the perspectives of numerous 
social, political and economic belief systems. Some of these are in the early stages of 
implementation. It remains to be seen which will be the most effective, but it seems clear 
that none of these solutions alone will be enough to solve the seemingly intractable 
problem of low coffee prices or their accompanying environmental impacts. 

"Sustainable" coffees 

A number of special certification programs have become popular in recent years, 
as consumers become more aware of the social and environmental effects of their 
morning coffee. These certification programs allow farmers to receive a price premium 
for their coffee, in exchange for meeting certain social and/or environmental standards. 



148 



Three labels dominate the certified coffee market: organic, fair trade, and shade grown. 
In addition, a number of non-certified coffees, such as "bird-friendly" coffee, are 
marketed to socially conscious consumers, and the Starbucks Coffee Company has 
implemented a set of procurement standards for all its coffee based on environmental and 
social criteria. This section will briefly discuss the two most important sectors of the 
sustainable coffee market: organic and fair trade. 

Organic coffee is a set of mostly production-based (as opposed to socially-based) 
certification criteria, based on a principle of minimizing or eliminating the environmental 
and health effects that are frequently the result of chemically-intensive agriculture. 
Standards for organic certification can vary widely, depending on the certifying 
organization, but there are some areas of fairly universal agreement. All organic 
standards prohibit the use of synthetic chemicals, and require measures for the protection 
of soil and human health (Rice 2001). 

Although many farmers appreciate the health and environmental benefits of 
organic coffee production, the premiums paid for organic coffee are the main motivating 
factor (Rice 2001). TransFairUSA pays a minimum of $1.41 per pound for certified 
organic coffee (TransFairUSA 2004a), or $0.15 above the fair trade floor price 
(TransFairUSA 2004b). The Costa Rican daily La Nation reports that Starbucks pays 
Costa Rican producers of organic coffee a premium of $0.50 per pound (Anonymous 
2005). Ultimately, the premium received depends upon the quality of the coffee and the 
marketing skills of the producing organization (Rice 2001). 

Reliably estimating the size of the current organic market, or of growth trends, is 
difficult. Rice (2001) estimated that, in the late 1990's, organic coffee represented 
approximately 3% of U.S. specialty coffee imports. Likewise, figures on the number of 
organic producers are difficult to locate (Rice 2001). According to Bray et al. (2002), 
Mexico is the world's leading producer of certified organic coffee. In 2002, they 
estimated that 4% of Mexico's coffee farmers, cultivating 2% of the country's coffee 
area, currently sell to the certified organic market, with a total production of more than 
6,000 Mt. 



149 



Unlike organic or shade-grown coffees, "fair trade" coffees address only the 
social and economic effects of coffee, by providing farmers with a guaranteed minimum 
price and long-term contracts. According to Oxfam International, the first fair trade 
coffee was from Guatemala, and was sold in Holland in 1973 (Gresser & Tickell 2002 p. 
40). The first initiative to bring fair trade into the mainstream was begun in Holland in 
1988, as a partnership with the private Max Havelaar Foundation and the Dutch 
government. The goal of the project was to expand from 0.3% to 10% the share of 
Holland's coffee supply that was purchased from cooperatives, with a premium of $0.90 
per pound to be paid to the producer or used for microcredit and development projects. 
The country's major coffee roasters responded with their own program to include in their 
blends 2-3% of coffee purchased directly from small producers, without raising prices or 
paying premiums to farmers (Pelupessy 1993 pp. 49-50). 

In 1989, data showed that producers of Max Havelaar coffee were receiving $0.09 
more per pound, while consumers were paying $0.50 more, with the extra profits being 
absorbed by the roasters, making "fair trade" a lucrative business for small roasters. 
After 1989, when free market prices fell 54%, the retail price of coffee in Holland fell by 
23%, and by 9% for the Max Havelaar coffee. Although the roasters continued to take a 
large cut of the price premium, they were still earning less than they could if they sold 
"free market" coffee (Pelupessy 1993). 

Pelupessy and Tilburg (1994 p. 253) calculated that in Costa Rica, farmers who 
sold coffee under the Max Havelaar program enjoyed an increase of about 13% in their 
total income from all sources. In El Salvador and Guatemala the increase was calculated 
at 14% and 10%, respectively. The price differential was clearly shielding some coffee 
farmers from the price collapse to some degree, but only a handful: even if all the Max 
Havelaar coffee sold in Holland in 1994 had been purchased from Central American 
countries, it would only account for 2% of that region's production (Pelupessy & Tilburg 
1994 pp. 249-251). In 1989, Max Havelaar bought 800 Mt of coffee from 1,400 farmers 
in Costa Rica and Guatemala (Wattel et al. 1991; cited in Pelupessy & Tilburg 1994 p. 
253), or 0.2% of the exports of both countries (UNFAO 2005). Max Havelaar is still an 
important fair trade certifying organization for Western Europe (Gresser & Tickell 2002 
p. 41). 



150 



Since the Max Havelaar initiative, the fair trade industry has grown and matured 
significantly, and now incorporates a specific set of standards and a fixed floor price 
based on costs of living and production (TransFairUSA 2004a). The fair trade market 
and fair trade certification is overseen by Fair Trade Labeling Organizations International 
(FLO), an umbrella organization in Bonn, Germany (TransFairUSA 2005). Certification 
standards are developed in concert with producers, traders, and the FLO's member 
organizations, and include a minimum floor price and numerous social criteria (FLO 
2005). Currently, the fair trade floor price for export is $1.26 per pound, paid to the 
cooperatives that process the coffee (TransFairUSA 2004b). 

There is debate over when the market for fair trade and other "sustainable" 
coffees will reach saturation. According to Oxfam (2002 p. 41), in 2002 "Fair Trade 
roast and ground coffee . . . accounted] for over seven percent of the UK roast and 
ground market and about two percent of the total coffee market." In 2001, while coffee 
sales grew by 1.5% overall, sales of fair trade brands grew 12% worldwide, and by 36% 
in the United States. TransFairUSA (2004a) claims that "81% of Americans say they are 
likely to switch brands to help support a cause, when price and quality are equal." While 
they report that in 2003, 18.5 million pounds of fair trade coffee was imported into the 
U.S. - nearly doubling over the previous year (TransFairUSA 2004b), this represents 
approximately two-thirds of one percent of all U.S. imports of green and roasted coffee 
that year (UNFAO 2005). Even Oxfam agrees that fair trade alone cannot solve the 
coffee crisis or the problem of coffee price volatility, and that wider-ranging solutions are 
needed (Gresser & Tickell 2002 p. 42). 

Additionally, fair trade currently is only available to small farmers who belong to 
processing cooperatives, and does nothing to address working conditions on larger coffee 
farms, or prices paid to farmers who sell to private processors. Fair trade may also 
encourage farmers to continue producing coffee who would be better off, over the long- 
term, changing to other products that they can produce more competitively. 

Another weakness of fair trade may be that it can, in effect, subsidize buyers who 
pay lower prices. Because cooperatives do not always sell all their coffee to fair trade 
buyers, the coffee sold at fair trade prices may simply raise the average price given to 



151 



farmers, without raising that price to the floor level that consumers believe farmers are 
receiving. For example, in 2003-04, Coopabuena sold 10% of its coffee through fair 
trade or direct marketing channels (Grosser et al. 2005). This helped to raise the final 
price paid to farmers above the average paid by cooperatives in Coto Bras, but only by 
about 9%, to approximately $0.60 per pound (Fernandez & Alvarez 2004). 

Dennis Macray at the Starbucks Coffee Company (pers. comm. 2004) says that 
the "subsidy" issue is one faced by anyone paying higher-than-average prices for coffee, 
including specialty roasters. When farmers are receiving high prices for some of their 
coffee, they can afford to accept lower prices for the rest of it, or to continue producing 
coffee to sell at those lower prices. In the case of Starbucks, this can happen when a 
farmer or beneficio produces coffee grown at different elevations (as is often the case in 
areas with steep slopes), which is sold at different prices. In order, therefore, to truly 
address the structural problems in the coffee market, it is necessary to implement changes 
that involve the large roasters, who currently buy on the commodities exchanges at the 
lowest available price. 

Specialty coffees 

There are essentially two areas in which a coffee producer can compete in the 
world market: cost or quality. With its relatively high labor costs and living standards 
compared to other coffee-producing countries, Costa Rica is not in a position to compete 
on price, and must turn its attention to quality if it wishes to continue producing coffee. 
While low production costs are of initial advantage to a country entering the coffee 
market, the best long-term competitive advantages are provided by technology, 
infrastructure and the availability of trained, knowledgeable people (Pelupessy 1997 p. 
7), all of which Costa Rica possesses in abundance. 

The market for higher-quality coffee is expanding, and many small coffee- 
producing nations, including Costa Rica, have pinned hopes on this market. In the late 
1990's, high-quality milds represented 30% of demand in the United States, and this 
demand could grow to 70% over the next century (Pelupessy 1997 p. 7). The Specialty 



Manager, Business Practices, Corporate Social Responsibility, Starbucks Coffee Company 
152 



Coffee Association of America reports that, in 2003, specialty coffee accounted for $1.7 
billion, or a third of the $5 billion U.S. coffee market (TransFairUSA 2004b). 

In 2001, the Costa Rican daily La Repiiblica reported that 20% of Costa Rica's 
coffee was sold through direct contracts (as opposed to the commodities exchanges), at 
prices nearly double the national average (Prendas 2001a). In 2004, La Nation reported 
that 60% of Costa Rica's higher-quality coffee was sold to the specialty market, receiving 
average prices of $1.20 per pound, compared to $0.80 per pound on the New York 
Exchange (Barquero S. 2004d). 

In July of 1993, the Costa Rican Coffee Institute (ICAFE), the National 
Autonomous University and the University of Costa Rica sponsored a symposium 
entitled "Modernization, Technology, Social Change and the Coffee Crisis," to seek 
public policy solutions to the crisis. For what appeared to be the first time, a long-term 
reduction in output was discussed (Guevara M. 1993). It was not until 1999, however, 
after the second price collapse in less than ten years, that the focus of government policy 
began to seriously shift from quantity to quality. It was in that year that ICAFE began to 
sign "quality agreements" with various coffee-producing zones (Barquero S. 1999). 

In 2001, while the average price was about $0.70 per pound for Costa Rican 
farmers, producers of specialty coffees were receiving a minimum of $1.00. Coffee 
specialists were arguing by this time that the salvation for Costa Rica's coffee sector 
would be high-quality specialty coffees, since lower-quality coffees could be produced 
much more cheaply in other countries. Juan Bautista Moya, director of ICAFE, began 
arguing that growers who could not find a niche in the specialty market would have to 
stop growing coffee. A program was begun to promote a Costa Rican brand, much like 
Colombia's successful "Juan Valdez" campaign, and to eliminate the production of the 
lowest quality 5% of Costa Rica's coffee (Mora 2001a). 

In 2002 Costa Rica, along with the other Central American countries and the 
Dominican Republic, signed a "Quality Coffee" agreement with the U.S. Agency for 
International Development. Through this agreement, USAID would provide funding for: 

"a market-based program to assist small and medium coffee producers to improve 
coffee quality, form new business linkages, secure longer-term contracts with the 



153 



specialty coffee industry, and identify and implement diversification options for 
producers that cannot be competitive." 

Similar programs were being implemented in Africa and Asia (U.S. Agency for 

International Development 2003). 

In 2003, Starbucks sparked excitement in Costa Rica when they announced a plan 
to increase their purchases of Costa Rican coffee by 150% over four years, to 46,000 Mt, 
or 42% of Costa Rica's 2003 production. Starbucks - which purchases 30% of its beans 
from Costa Rica - was already the country's largest buyer, paid prices nearly double what 
is paid on the New York exchange (Rogers 2003), and purchases all its coffee through 
direct, long-term contracts with growers and processors (Dennis Macray, pers. comm. 
2004). Juan Bautista Moya, executive director of ICAFE, remarked that due to 
Starbucks' plan "there will no longer be a crisis" (Rogers 2003). 

One option for producers who want to enter the specialty coffee market is to work 
together with their neighbors to develop an "appellation," or a brand of coffee based on a 
particular growing region. Coffees have distinctive characteristics based upon where 
they were grown and processed, much like wines. Many coffee connoisseurs claim to 
have preferences for coffees from certain regions, and regions that produce distinctive, 
high-quality coffee are beginning to capitalize on this growing interest in coffee of 
specific origins. Some well-known appellations that command a price premium include 
Guatemala Antigua, Hawaiian Kona, and Jamacian Blue Mountain coffee. 

While appellation-based brands may be more difficult to establish, due to the high 
degree of coordination and cooperation required among growers, processors and 
exporters in a region, they are more stable, because no one outside the appellation region 
can take over the region's appellation (Varangis et al. 2003), although counterfeiting is 
rumored to occur, most notably with Kona and Blue Mountain coffee. 

The Costa Rican ICAFE is working to establish appellation brands for its 
country's coffees, labeling coffee as belonging to one of seven regions: Tarrazu, Branca, 
Orosi, Tres Rios, Turrialba, Valle Central and Valle Occidental (The Branca region 
includes Coto Bras). According to the ICAFE' s web site (www.icafe.go.cr), "All the 
coffee-producing regions of Costa Rica have joined in signing a quality improvement 



154 



agreement, whereby the owners of processing plants in each region undertake to produce 
coffee of the highest quality." To qualify to carry one of the Costa Rican appellations, 
the area where a coffee is grown is required to meet strict requirements for altitude, soils, 
and precipitation, and the beans must meet standards for hardness, body and acidity 
(Barquero S. 2004c). 

The Costa Rican coffee region Tarrazu provides an example of the benefits that 
can be realized from a high-quality coffee region establishing its brand. In 2004, coffee 
carrying the Tarrazu appellation could receive a price premium of $0.25 to $0.40 per 
pound, and a bitter argument was taking place over alleged unauthorized use of the 
Tarrazu brand name (Barquero S. 2004b). In April 2005, the estate coffee La Candelilla, 
from Tarrazu, received Starbucks' "Black Apron Exclusives" quality award, which paid a 
premium (above the price of the coffee) of $15,000 for that year's harvest (Associated 
Press Costa Rica 2005). In December 2005, La Candelilla coffee was retailing for $22 
per pound at Starbucks' online store. 

Costa-Rica is well-positioned to expand its share of the specialty coffee market, 
and appears to be reaping benefits for its farmers by doing so. This solution is not an 
option for many other countries, or indeed for all Costa Rican producers. Areas such as 
Agua Buena, which have less-than-ideal conditions for coffee, are unlikely to find a place 
for their coffee in the relatively small specialty market. Additionally, as more producers 
of high-quality coffee attempt to enter the specialty sector, prices will be driven down. 
This "fallacy of composition," or what Oxfam refers to as "running for the same exit," 
(Gresser & Tickell 2002 p. 42), is being encouraged by government agencies in 
producing countries. Although there is awareness of the problem, it seems that, rather 
than look for other exits, everyone seems to rely instead on the hope that they can get to 
the exit first. 

Supply controls 

Pelupessy (1991) stresses the importance of controlling supply, possibly through 
programs that reduce inputs. Such an approach would have the added benefits of 
lowering farmers' costs, thereby insulating them from price swings, and reducing the 
environmental harm done by coffee production. He also suggests that there is a need to 



155 



seriously question whether coffee is an appropriate crop for small producers, given its 
high capital needs and price volatility. 

The question of whether some coffee producers should get out of the market 
entirely is an important one. Free-market promoters point to the crisis as a market signal 
to "supply a higher-value product or exit the market" (Lindsey 2003). Orin Smith, 
former CEO of the Starbucks Coffee Company (which supports an International Coffee 
Agreement-style quota and price agreement), has been quoted as saying that coffee below 
3,500 feet (1,050 meters) in elevation may not be economically viable over the long term, 
and those farmers should consider changing to other products (Dennis Macray, pers. 
comm. 2005). As this study has shown, many producers are already taking this step as a 
logical response to the crisis. This study has also illustrated some of the environmental 
consequences of such a change, especially when farmers do not receive adequate 
technical or financial support to change to environmentally sustainable products. 

In the era of "free trade," the prospects for controlling exports through another 
quota agreement such as the ICA look dim. Even if a new agreement could be reached, it 
is unlikely that it would be entirely free of the problems that undermined the first four 
agreements. The experience with ICA, however, points out many ways in which a new 
agreement could be improved. Most importantly, any new agreement needs to address 
the problem of over-supply - not simply by limiting exports, which under ICA resulted in 
the destabilizing accumulation of large stocks, but by addressing supply at the farm level. 
According to Oxfam, Nestle stands alone among the top five coffee roasters in supporting 
supply management, including an ICA-style agreement (Gresser & Tickell 2002 p. 43). 

In a Congressional hearing on the coffee crisis in July of 2002, a number of 
speakers recommended increasing the "purity" standards for coffee imports to the U.S. 
from an allowed level of 25% adulterants (bad beans, sticks, rocks etc.) to 5%, the level 
set by the ICO and European countries. U.S. Rep. Sam Farr claimed this would remove 
8-10 million bags of coffee from the market, and further encourage a shift to production 
of higher-quality coffees (U.S. House of Representatives 2002 p. 15). 

Of course, plans to limit supply are, in essence, plans to remove some proportion 
of coffee farmers from the business. When, as is often the case, supply management is 



156 



addressed as a quality control issue aimed at removing some percentage of the lowest- 
quality coffee from the market, it is the farmers in the most marginal areas (where prices 
are already likely to be the lowest) who will no longer have buyers for their product. As 
painful as it is, such a result is unavoidable when the problem is, at root, too much coffee. 

This thesis has shown some of the environmental results when farmers are driven 
out of the coffee market, and other authors have discussed the same issue for different 
localities (Varangis et al. 2003 pp. 54-63). It is therefore essential, from both an 
environmental and social standpoint, that programs of supply control address the 
provision of environmentally sustainable alternatives to coffee, and of the financial 
assistance needed to help with the transition to such alternatives. To address this issue, 
Oxfam recommends the creation of a diversification fund, to assist producers of lower- 
quality coffee in finding alternative livelihoods (Gresser & Tickell 2002 p. 49). 

Diversification 

Diversification is clearly an important component of any plan to address the 
coffee crisis. Just as investors with a balanced asset portfolio are less vulnerable to stock 
market swings, a farmer with several crops, at least some subsistence production, and/or 
off-farm sources of income, will be more shielded from price volatility in the 
commodities they produce. Unfortunately, when prices of a particular commodity are 
high (just as when a particular stock market sector is going up), the temptation is great to 
consolidate resources in that area, rather than forego some temporary profits in the 
interest of longer-term stability. 

The national governments of producing countries, which could play a role in 
encouraging diversification, are prone to the same financial enticements as farmers. 
Coffee exports are often an important source of foreign exchange for debt-burdened 
governments of developing countries (Pelupessy 1993), and it can be difficult to forego 
some of the additional export tax revenues that can be obtained during periods of high 
coffee prices. Furthermore, when government agencies do introduce diversification 
programs, poor planning or lack of ongoing support may undermine the efforts, resulting 
in reluctance by farmers to work with those agencies in the future. 



157 



Some alternative exports that grow well in coffee-producing regions are mango, 
citrus, and macadamia, but these crops, like coffee, take several years to begin producing, 
and require intensive marketing to break into the world trade (Viquez 1992a). In Coto 
Brus, both citrus and macadamia were introduced by the Ministry of Agriculture (MAG) 
as alternatives or supplements to coffee. Although citrus has been successful for some 
farmers, and plantain has also provided an alternative income for many, neither yields the 
same income that coffee once did. One farmer said that the orange cultivar sold to many 
farmers by MAG was not of export quality, resulting in substantial losses to those who 
planted it. 

Macadamia has been more or less a failure in the region, as sufficient processing 
and transportation infrastructure and access to markets were not available to farmers in 
the area, who found themselves unable to sell their harvest. None of the farmers 
interviewed listed macadamia as a current crop, although several talked about the 
attempts to introduce it in the canton. The experience of Coto Brus with macadamia 
underscores the importance of government agencies' working with the private sector to 
ensure that markets are available for new products. 

Vegetable crops, such as tomato and green pepper, require extremely high 
chemical and labor inputs, making these an unsustainable and environmentally 
detrimental alternative. Nonetheless, many farmers have been drawn to them for the high 
profits they offer. Although this was not mentioned during interviews, the annual growth 
cycle of vegetables, resulting in a short period from planting to sale, is most likely 
another attraction of these types of crops. 

There is also the need to consider that, when introducing any new crop, saturation 
of the market may result from many producers switching to that product at the same time, 
driving prices down. The coffee crisis itself is, to some extent, a result of this force - 
Vietnam introduced coffee to generate foreign exchange for the country, and was 
successful enough at it to drive coffee producers around the world into bankruptcy. 
Farmers in Agua Buena frequently mentioned fears that the prices of plantain, citrus, 
vegetables or beef would fall if more farmers switched to producing these products. 



158 



Despite these hurdles, diversification remains key to both addressing the current 
crisis and insuring against future ones. Many national governments, as well as 
international agencies, recognize this, although diversification does not appear to be a 
high priority in any response plan for Central American countries. The U.S. Agency for 
International Development claims to have included diversification as part of its response 
to the coffee crisis (US AID 2003; 2004). In 2001, Guatemala established a trust fund to 
support coffee growers which was intended, in part, to encourage diversification. The 
other Central American countries, however, appear to have focused their efforts almost 
entirely on the short-term goals of keeping farmers afloat by supplementing prices 
(Varangis et al. 2003 p. 27). 

3. Conclusion 

The current crisis facing coffee producers worldwide is structural in nature, the 
result of significant changes in world markets and production dynamics. In response to 
the crisis, coffee farmers who have not been able to either sufficiently cut costs or access 
higher-value sectors of the market are being forced to abandon coffee production. In the 
district of Agua Buena, Costa Rica, farmers are either abandoning their land or changing 
to other crops, with the predominant conversion choice being pasture. The conversion 
from coffee to pasture is having significant effects on the richness, diversity and 
community composition of resident birds that utilize the agricultural matrix. The 
relationship between the coffee crisis and environmental effects is complex, but the 
apparent habitat loss and urbanization associated with the crisis are causes for concern. 

A number of responses to the crisis are available at the international level. None 
of them is simple and none of them, taken in isolation, will solve all the problems created 
by the crisis. Improvement of coffee quality and expansion of the specialty market; the 
varieties of certified coffees, particularly fair trade; price and quota agreements 
resembling the International Coffee Agreement; and diversification are all components of 
a strategy to deal with the economic and environmental impacts of the crisis. 

Perhaps the most important, and the most overlooked, element of an effective 
response to the crisis is the provision of economically and environmentally sustainable 
alternatives for those producers in geographical areas that are not ideally suited for 



159 



growing coffee. Programs of supply reduction or quality improvement that leave farmers 
to "sink or swim" without transitional assistance are certain to have negative effects on 
the environment, as well as the social fabric, of the areas that are cut out of the coffee 
market. Conversely, fair trade programs need to address the appropriateness and 
competitiveness of coffee as a product for a given area, and address the issue of 
diversification for long-term financial stability. 

Finally, conservation groups focused on coffee-producing countries need to 
understand the importance of focusing attention on this issue, and cooperating with 
governments and NGO's to find viable solutions that address the environmental 
implications of coffee price instability. Resolution of the coffee crisis is of 
environmental as well as social and economic importance. 



160 



Glossary 

Beneficio: A facility that processes fresh coffee cherries into green coffee. 

Cajuela: Twenty liters of fresh coffee cherries. Coffee pickers are paid by the number of 
cajuelas they collect. Twenty cajuelas equal one fanega. 

Charral: Pasture in a state of early abandonment or overgrowth, characterized in the 
Agua Buena region by mostly woody vegetation averaging 1-2 meters in height; 
potentially an early stage of forest regeneration. 

Colon: The unit of currency in Costa Rica. In 2003 one dollar was equal, on average, to 
410 colones. 

Coyote: 1 . A middleman who purchases raw coffee cherries from farmers and transports 
them to a beneficio. 2. Someone who transports or arranges transport of persons 
to the United States, usually illegally. 

Fanega: Four hundred liters of fresh coffee cherries. Coffee growers are paid by the 
number of fanegas they bring to a processor. In Costa Rica, processing yields 
average around 100 pounds of dry green coffee from each fanega of fresh 
cherries. Processing yields in Coto Brus average 96 pounds. (ICAFE 1993) 

Gallery forest: A forest along a river or stream. 

Green coffee: Dried coffee beans that have had the pulp and "silver skin" removed. 
Most coffee is exported as green coffee. 

ICA: The International Coffee Agreement. 

ICAFE: The Costa Rican Coffee Institute, which regulates the coffee trade within the 
country. 

ICO: The International Coffee Organization. 

Parchment: The thin membrane, papery when dry, that separates the coffee bean from 
the pulp of the coffee cherry. 

Quintal: Forty-six kilograms, or 100 pounds, of dry green coffee. The standard unit of 
measure for exports from Costa Rica. (ICO uses sacks of 60 kg). 



161 



Recibidor: One of many small collecting stations for ripe coffee cherries stationed 
around the Costa Rican countryside. 



162 



Appendices 

Appendix A. Additional information regarding interviews 
Questionnaire for landowners 

1 . How old are you? 

2. How many years of education do you have? 

3. How many years have you had your farm? 

4. How many years have you lived in the area? 

5. How many hectares do you have? 

6. How many of those are yours, and how many do you rent? 

7. How did you acquire your land? 

8. How many hectares are in coffee, pasture, forest, annual crops, or other uses? 

9. What do you grow on your farm (include items for personal consumption). 

10. What is your income per year? (If they don't know, ask questions about sources 
of income and amount from each source, to allow for an estimation of annual 
income). 

1 1 . Do you have other sources of income besides coffee? What are they, and how 
much to you earn from each? 

12. Do you want your children to also be farmers? 

13. Do you belong to any agricultural organizations? Which ones? 

14. Do you attend meetings of any agricultural organizations? How many per year? 

15. Have you cut out any of your coffee in the last 6 years? 

16. When did you cut it out? 

17. How much did you cut out? 

1 8. Why did you cut it out, or if you kept it, why did you keep it? 

19. What did you do with the land where the coffee was? What are your plans with 
the land? Why? 

20. How many people work on your land? How many family members work on your 
land? Do the workers live on or off the farm? 

21. How many family members live with you? 

22. Have some children or other family members left the community? Why? 

23. How much did you receive per fanega for the last coffee harvest? 

24. How many fanegas did you sell during the last harvest? 

25. What price would you need for coffee to be profitable for you? (Also asked: 
What coffee price would you need for it to be worthwhile to grow coffee?) 

26. What is the highest price you remember getting for your coffee? What year was 
that? 

27. Do you use any agrochemicals? What types? 

28. Are you going to stay on your land, or do you have plans to leave? Why? If you 
leave, what do you plan to do? What are your plans if you do stay? 

29. What do you cook with? 

30. Do you think there will be enough fuel wood if you convert your land from coffee 
to another use? What will you do if there isn't enough? 



163 



3 1 . If you have changed from coffee to another land use, do you use fewer or more 
agrochemicals? Does your land require more or less work? 

32. If you are going to keep growing coffee, will you continue with conventional 
coffee or do you have plans to change to other cultivation methods (like 
sustainable or organic). Why, or why not? 

33. Where or from whom do you obtain technical advice on methods of growing? 

34. Where or from whom do you obtain information about prices and marketing? 

35. Do you use credit or borrowed money to finance your production? 

36. If you use credit, does your lender require any special practices by you before 
making the loan? 

37. What types of birds do you recognize in your farm? 

38. Have the types of numbers of wild animals on your farm changed over the years? 
How have they changed? If they have changed, do you have an idea as to why? 

Excluded data 

One interviewer, the author of this thesis, was a North American female, fluent in 
Spanish, who had lived for a year in Costa Rica. The other interviewer was a Costa 
Rican female, born and raised in the Agua Buena area. Both interviewers conducted 
surveys in the same neighborhoods, with the North American interviewer conducting 
surveys of farmers on one side of a street and the Costa Rican conducting surveys on the 
other side. The North American interviewer completed 38 interviews and the Costa 
Rican interviewer conducted 21. All answers were grouped by interviewer and tested 
using Mests or / tests to determine if any bias was introduced to the question by the 
interviewer. 

Table A.l summarizes the variables that showed a significant difference in 
responses based on interviewer. The fourteen variables that showed a significant 
difference between interviewers can be broken into two groups. The first group consists 
of variables, such as age or number of hectares of a specific land use, that are unlikely to 
be subject to cultural influence or other forms of interviewer bias; i.e., differences in 
these variables are probably due to chance. The second group consists of qualitative or 
opinion-based questions that are more likely to be subject to bias. 

The questions regarding age, number of hectares abandoned, and number of 
family members working on the farm are qualitative questions not likely to be subject to 
interviewer bias. For example, there is no readily apparent reason why respondents 
would over- or under-report their age to either interviewer. Additionally, the differences 



164 



for the latter two variables, while statistically significant, are small. It is therefore likely 
that the differences in these variables are due to chance, and none of these data are 
excluded from the analysis. 

Number of years of education showed a significant difference between 
interviewers, with respondents interviewed by the North American reporting, on average, 
an additional 1 .5 years of education. While it is conceivable to imagine that respondents 
might be over-reporting education to the North American interviewer, educational level 
was also significantly negatively correlated with respondent age (r= -0.524, N=58, 
p=0.000). The average respondent age was 10 years higher for the Costa Rican 
interviewer (f=-2.270, df=55, p=0.027), a result that is probably due to chance. For the 
purposes of this study, it will be assumed that the educational differences are a result of 
age, and not of interviewer bias. 

The respondent's spouse participated in 100% of interviewers conducted by the 
Costa Rican, and only 53.8% of those conducted by the North American (/2=14.400, 
df=l, p=0.000). Spouse participation, however, was not significantly correlated with any 
other interview variable. This difference is therefore not important for the purposes of 
this study. 

Nearly all (91.3%) of the respondents reported an annual income to the North 
American interviewer, while only one third (33.3%) reported an annual income to the 
Costa Rican interviewer (% =21.564, df=l, p=0.000). This difference may be due to 
respondents' discomfort in discussing income with a community member, but also 
probably has to do with relative interviewer effort. When respondents told the North 
American interviewer they didn't know their income, the interviewer would review their 
various income sources and question respondents about amounts received from each. 
Discussions with the Costa Rican interviewer revealed that in most cases, when a 
respondent told her they didn't know their income, she accepted the answer and did not 
ask follow-up questions. 

While total annual income for respondents who reported it was not significantly 
different between interviewers (£=1.207, df=38, p=0.235), significantly fewer (73.7% vs. 
94.6%, Fisher's Exact Test, p=0.038) reported sources of income besides coffee to the 



165 



Costa Rican interviewer. This is probably due to the difference in interviewer effort 
described above. For example, many respondents did not report remittances or pensions 
until specifically asked about them, and it was not clear whether the Costa Rican 
interviewer consistently asked about specific income sources. Therefore, despite the fact 
that income numbers themselves are not significantly different between interviewers, the 
income data from the North American interviewer are likely to be more complete, and 
data from the Costa Rican interviewer are excluded from the analysis. Information on 
sources of income is included, however, as this is qualitative information that may be of 
interest despite being potentially incomplete. 

The question of what price coffee would need to reach in order to be worthwhile 
was actually worded in two separate ways by the two interviewers. While the Costa 
Rican interviewer, who received higher values on average, asked the question as written: 
"What coffee price would you need for it to be worthwhile to grow coffee," due to 
language issues the North American interviewer actually phrased the question as, "What 
price would you need for coffee to be profitable for you." Thus, the difference in these 
responses may indicate that a price that would allow farmers to turn a profit may not be 
enough to make coffee worthwhile for them to grow. The responses for each interviewer 
are treated separately in this case, but neither set was discarded. 

The question concerning the year of the highest coffee prices is probably related, 
again, to the average age of the sample, as the mean year reported to the Costa Rican 
interviewer (1992) was significantly lower than that reported to the North American 
interviewer (1997, ;=3.429, df=28.245, p=0.002), and age was significantly negatively 
correlated with the highest price year in memory (r= -0.392, N=54, p=0.003). The mean 
for highest price in memory was much larger in the Costa Rican interviewer's group, 
though not significantly so ($230 vs. $132, r=-1.965, df=27.427, p=0.060). The highest 
prices in respondents' memories, however, were also closely negatively correlated with 
the year in which respondents reported receiving that price (r=-0.890, N=50, p=0.000), 
but not significantly correlated with age (r=0.171, N=49, p=0.239). 

Additionally, some values reported during interviews were exceptionally high, 
reaching over $550/fanega (over $5/pound) for both interviewers, for years in which the 



166 



world export price for other milds was only around $2/pound. The responses to this 
question, in order to enable comparison, had been converted based on the exchange rates 
and inflation factors for the year in which the price was received, but it is not clear 
whether all respondents reported prices in the currency actually received at the time of 
sale, or mentally "converted" the prices into contemporary equivalents. It appears that a 
mixture of these two options occurred, but there is no way to reliably separate the 
responses based on this factor. Thus, these data are not reliable and are excluded from 
analysis. 

The remaining four variables in Table A.l are qualitative questions that could 
conceivably show a cultural bias. Discussions with the Costa Rican interviewer 
regarding her expectations for these questions, and her belief as to the likely responses, 
and the overall lack of variation in responses recorded by the Costa Rican vs. the North 
American interviewer, lead to the conclusion that the Costa Rican interviewer may have 
been coaching respondents or otherwise biasing the responses to these four questions. 
Therefore, data from the Costa Rican interviewer are excluded from analysis of these 
variables. 

In short, six variables are wholly or partially excluded from analysis, as follows: 

1 . Annual Income (Costa Rican interviewer) 

2. Highest coffee price in memory (both interviewers) 

3. Outside sources of information on technology and farming practices (Costa 
Rican interviewer) 

4. Outside sources of information on price (Costa Rican interviewer) 

5. Change in numbers or species of animals over time (Costa Rican interviewer) 

6. Reason for perceived changes in fauna (Costa Rican interviewer) 



167 



Figures and Tables 



Variable 


US 
Interviewer 


CR 

Interviewer 


df 


Statistic 


Sig. 


Spouse participated in 
interview 


53.8% 


100% 


1 


x 2 = 14.400 


0.000 


Age of interviewee 


50.8(mean) 


60.81(mean) 


55 


t= -2.270 


0.027 


Years of education 


4.5 (mean) 
6.0(mode) 


3.0(mean) 
3.0(mode) 


54.9 


*;=2.127 


0.038 


Did not report annual income 


8.3% 


66.7% 


1 


X 2 =2 1.564 


0.000 


Reported other sources of 
income besides coffee 


94.6% 


73.7% 


N/A 


Fisher's 
Exact Test 


0.038 


Number of hectares of coffee 
abandoned 


0.7 


0.1 


38.4 


*;=2.483 


0.018 


Number of family members 
working on farm 


l.O(mean) 
l.O(mode) 


1.4(mean) 
l.O(mode) 


56 


t= -2.414 


0.019 


Price that coffee would have 
to reach to make it 
worthwhile to grow 


$66.19 


$82.03 


51.7 


*t= -5.054 


0.000 


Year of highest price in 
memory 


1997(mean) 
1998(mode) 


1992(mean) 
1997(mode) 


28.3 


*t= 3.429 


0.002 


Reported no outside source of 
technical information 


35.1% 


90.5% 


1 


X 2 =8.068 


0.005 


Reported no outside source of 
price information 


35.1% 


100% 


1 


X 2 =23.237 


0.000 


Believe there are fewer 
species/individuals of animals 
now than before 


51.4% 


100% 


1 


X 2 = 14.060 


0.000 


Listed hunting as primary 
reason for faunal change 


17.6% 


93.3% 


1 


/=18.331 


0.000 



Table A.l. Interview variables that showed a significant difference between interviewers. Where 
Levene's test for equality of variances was significant at p<0.05, the t statistic with equal variances not 
assumed was used. These t values are marked by an asterisk. Fisher's Exact Test was used where/ was 
not appropriate due to low expected values. 



169 



Perceived changes in fauna 


Animals of which respondents Armadillos 2 (3.4%) 


have observed more since arriving Birds 2 (3.4%) 


in the area Clay-colored Robins 2 (3.4%) 


Guans 2 (3.4%) 


Agoutis 1 (1.7%) 


Chachalacas 1 (1.7%) 


Doves 1 (1.7%) 


Grackles 1 (1.7%) 


Gray-necked Wood-rail 1 (1.7%) 


Hummingbirds 1 (1.7%) 


Monkeys 1 (1.7%) 


Mot-mots 1 (1.7%) 


Sloths 1 (1.7%) 


Squirrels 1 (1.7%) 


Tanagers 1 (1.7%) 


Toucans 1 (1.7%) 


Animals of which respondents Agouti 7 (11.9%) 


have observed fewer since Armadillos 3 (5.1%) 


arriving in the area Clay-colored Robins 3 (5.1%) 


Oropendulas 3 (5.1%) 


Peccaries 3 (5.1%) 


Coati 2 (3.4%) 


Doves 2 (3.4%) 


Flycatchers 2 (3.4%) 


Jays 2 (3.4%) 


Monkeys 2 (3.4%) 


Parrots/parakeets 2 (3.4%) 


Racoons 2 (3.4%) 


Squirrels 2 (3.4%) 


Tapirs 2 (3.4%) 


Toucans/toucanets 2 (3.4%) 


Anis 1 (1.7%) 


Chachalacas 1 (1.7%) 


Deer 1 (1.7%) 


Foxes 1 (1.7%) 


Guans 1 (1.7%) 


Insects 1 (1.7%) 


Jaguars 1 (1.7%) 


Macaws 1 (1.7%) 



Table A.2. Animals for which respondents have perceived a change in abundance since arriving in the 
study area. 



170 



Perceived changes in fauna (cont.) 


Animals of which respondents Mammals 1 (1.7%) 


have observed fewer since Ocelots 1 (1.7%) 


arriving in the area (cont.) Owls 1 (1.7%) 


Paraques 1 (1.7%) 


Snakes 1 (1.7%) 


Tanagers 1 (1.7%) 


Tinamous 1 (1.7%) 


Woodpeckers 1 (1.7%) 



Table A.2. Continued. 



171 



All farm products reported 


Coffee 41 (69.5%) 


Plantain 36 (61.0%) 


Corn 24 (40.7%) 


Yucca 23 (39.0%) 


Beans 22 (37.3%) 


Orange 21 (35.6%) 


Milk 17 (28.8%) 


Squash 15 (25.4%) 


Beef 14 (23.7%) 


Banana 12 (20.3%) 


Tequisque 12 (20.3%) 


Sugarcane 9 (15.3%) 


Greens 8 (13.6%) 


Pigs 5 (8.5%) 


Chicken 3 (5.1%) 


Chili 3 (5.1%) 


Lime 3 (5.1%) 


Other crops 3 (5.1%) 


Tomato 3 (5.1%) 


Calves 2 (3.4%) 


Cheese 2 (3.4%) 


Cilantro 2 (3.4%) 


Lumber 2 (3.4%) 


Nampi 2 (3.4%) 


None 2 (3.4%) 


Radishes 2 (3.4%) 


String beans 2 (3.4%) 


Cana india 1 (1.7%) 


Cas 1 (1.7%) 


Inga spp. 1 (1.7%) 


Jocote 1 (1.7%) 


Lemon 1 (1.7%) 


Mamones 1 (1.7%) 


Mango 1 (1.7%) 


Medlar 1 (1.7%) 


Pineapple 1 (1.7%) 


Sweet potato 1 (1.7%) 


Tilapia 1 (1.7%) 


Water apple 1 (1.7%) 


Zacate 1 (1.7%) 



Table A.3. All products reported by farmers, with number and percentage of farmers who 
reported each product. 



172 



Appendix B. Additional information regarding bird surveys 
Site Descriptions 

Pasture 1. 

This site belonged to the Corporation Cafetalera La Meseta, and was part of a 

several hundred-hectare coffee plantation that was being converted to pasture. The site 

was maintained by paid workers and a paid manager. According to field workers at the 

site, the coffee had been removed following the 2001-2002 harvest and the site was 

treated with the herbicide glyphosate, then planted with high-yielding pasture grasses. At 

the time of the surveys, the pasture grasses were mixed with weeds at an estimated cover 

of 10-40%. The original shade cover was Erythrina spp. The trees had been pollarded to 

an approximate height of l-2m following clearing, but had re-sprouted to form dense, 

leafy bushes or trees ranging from 2-10 m tall. No other trees inhabited the site. The site 

was bordered on one side by coffee fields that had been completely cleared (cut down 

and treated with herbicides) less than a month before, on one side by habitat similar to the 

site itself, and on two sides by gallery forest. No agrochemicals were used on the site 

during the year preceding the surveys. No cattle were present on the site during the 

survey period. 

Pasture 2. 

This site was managed by the owner and one paid worker, and was part of 70 ha 

belonging to the landowner. It was cut following the 1997-98 harvest and low-growing, 

tangling grasses covered the site at the time of the surveys. The original shade cover was 

poro and Musa spp., and most trees were left when the coffee was cleared. The 

remaining poro trees were in poor condition, most being l-2m tall and resembling sparse 

bushes. A very narrow (less than Vi m wide) stream ran through the site, and areas near 

the stream were swampy. Aim strip along the stream was inhabited by shrubby (2-3m) 

vegetation. One tall (~5m) orange tree, one small (2-3m) orange tree and one tall (7- 

10m) Inga spp. tree grew on the site. The site was bordered on one side by a dirt road, on 

one side by houses, on one side by a row of pine trees next to a dirt road, and on one side 

by gallery forest. There were two downed pine boughs on the site, near the pine trees. 

Approximately a dozen cattle were present on the site during the final survey. According 



173 



to the site's owner, no agrochemicals had been used at the site during the year preceding 
the surveys. 

Pasture 3. 

This site was managed by the brother of the owner, and was part of a 3 -ha farm. 

It was cleared of coffee following the 2001-2002 harvest, planted with a mixture of the 

hybrid grasses "Bombasa," "Brisanta" and "Bracaria," and treated with the broad-leaved 

herbicide Tordon. The owner had not used any agrochemicals since this initial treatment. 

The original shade cover had consisted of orange trees, Musa spp. and Erythrina spp. 

The Musa spp. trees remained and were in good condition. Most of the orange trees had 

been cleared, with only a small grove of 4 trees remaining, and the Erythrina spp. trees 

had been thinned to approximately half the original density. The remaining Erythrina 

spp. trees had been damaged by cows browsing bark from the lower trunks, and had very 

little foliage left, though many had grown quite tall (5 -10m). No other trees inhabited the 

site. The site was bordered on one side by gallery forest, on one side by a cana india 

(Dracaena fragrans) fence next to a dirt road, on one side by a Eucalyptus plantation and 

on one side by pasture. Twelve cattle were contained in this pasture during the survey 

period, and were present during every survey. 

Pasture 4. 

This site was managed by the sons of the owner, and was part of an 8-ha farm. It 

was cleared of coffee following the 2001-2002 harvest, planted with high-yielding 
pasture grasses and treated with Tordon. The ground was covered by an estimated 70% 
pasture grasses and 10% weeds with approximately 20% bare soil. The original shade 
cover had consisted of Erythrina spp. and Musa spp. A cluster of three tall second- 
growth trees inhabited the site, including one Inga spp. The poro trees remained on half 
the site and were about 3 m tall with wide, branching crowns and sparse foliage. The 
Musa spp. trees remained on the other half of the site and were in good condition. This 
5-sided site was bordered on two sides by tall (-10 m), mixed species living fences next 
to dirt roads, on one side by a cana india fence next to a dirt road, and on two sides by 
pasture separated by a fence of widely-spaced, very small and sparse Erythrina spp. 



174 



Cattle were present during two surveys. Tordon was sprayed once during the survey 
period. 

Coffee 1. 

This site was managed by paid workers and was part of a 30-hectare farm. The 

site had a mixed shade cover of Erythrina spp., Musa spp. and Citrus spp. Rows of 

Dracaena fragrans were planted along the contour of the hill and along the sides of the 

two paths that ran through the plantation. One tall (over 30 m), primary forest tree 

inhabited the site with four smaller (~10m) secondary forest trees. The primary forest 

tree supported lianas and other epiphytes. The site was bordered on one side by swamp 

and gallery forest, on one side by a D. fragrans fence next to a dirt road, on one side by a 

tall (-10 m), mixed species living fence next to pasture, and on one side by a D. fragrans 

fence next to pasture, vegetable crops and a coffee nursery. These trees were in poor 

condition but still alive, with sparse foliage. According to the spouse of the site's owner, 

no agrochemicals had been used at the site during the year preceding the surveys. 

Coffee 2. 

This site was managed by the owner and was part of a 6 ha farm. The site had a 

mixed shade cover of Erythrina spp. and Musa spp. The poro had been heavily pollarded 

and had small, dense heads of foliage. One orange tree was present at the site. Four tall 

(approximately 10 m), secondary forest trees inhabited the site. Three of these were in 

good condition with dense foliage, while the fourth was in poor condition with sparse 

foliage. The site was bordered on two sides by gallery forest, on one side by houses and 

gardens, and on one side by a house and a dirt road. No agrochemicals had been used at 

the site during the year preceding the surveys. 

Coffee 3. 

This site consisted of plots from two adjacent farms. Both farms were managed 

by the owners. One of the owners owned a 6 ha farm, and the other owned a 2 ha farm. 

Both sites had a mixed shade cover of Erythrina spp. and Musa spp. One plot was 

planted with 4 Terminalia amazonia trees at the top of a slope, and the other was planted 

with D. fragrans, mixed with the other shade cover. One plot was bordered on one side 

by gallery forest, on one side by a tall (-10 m), mixed species living fence next to 



175 



pasture, on one side by a garden, cornfield and house, and on two sides by coffee plots 
that had been cut and burned after the 2002-2003 harvest. The second plot was bordered 
on one side by a poro fence next to a dirt road, on one side by a tall (-10 m), mixed 
species living fence next to a driveway and pasture, on one side by a tall (-15 m), mixed 
species living fence next to another coffee plot, and on one side by a dense mixed-species 
fence covered in a tangle of vines. The ground of the first plot was treated with 
glyphosate during the survey period. The second plot was partially cleared of weeds by 
machete during the survey period. 

Coffee 4. 

This site was managed by paid workers and was part of the same 30-hectare farm 

as Coffee 1. The site had a mixed shade cover of Erythrina spp., Musa spp. and Citrus 

spp. Rows of D. fragrans were planted along the sides of the path that bisected the 

plantation. Three tall (over 30 m), forest trees inhabited the site. These trees were in 

very good condition with large, leafy crowns. The site was bordered on one side by 

gallery forest, on one side by a D. fragrans fence next to a dirt road, on one side by a D. 

fragrans fence next to a former coffee plot that had been cleared for pasture following the 

2002-2003 harvest, and on one side by a tall (-15 m), mixed species living fence next to 

pasture. The site was partially cleared of weeds by machete during the survey period. 

Error Sources 

The observe (this author) had 3 months previous birding experience in Costa 
Rica, and spent an additional 6 weeks prior to the start of the surveys learning to 
recognize the birds in the survey area by sight and sound, including attending two days of 
mist-netting and spending three days in the field with regionally-experienced birders. The 
learning process continued, however, over the course of the surveys, as the observer's 
ability to identify birds continued to improve. There are, therefore, likely to be more 
misidentified and unidentified birds in the earlier surveys than in the later ones. This 
source of error was minimized by scheduling the surveys in cycles, so that the six surveys 
of each plot were all spread out equally over the 1 -month survey period (Table B.l). 
Thus, any bias will apply equally to all sites and to both habitat types. 



176 



Working in the study area from 1998-2000, Lindell et al. (2004) found that point 
counts were likely to underestimate manakins (Pipridae) and hummingbirds (Trochilidae) 
compared to mist-netting, while point counts were more likely than mist-netting to detect 
flycatchers (Tyrranidae). The biases inherent in point counts are likely to apply to the 
area-search method, as well. Lindell et al.'s study was a comparison between the results 
of two techniques, not between either technique and the actual abundance of any species 
group. Ideally, more than one method would be used, but time and resources made this 
not possible for this study. Point-counts are more accurate at censusing vocal and 
conspicuous birds rather than smaller, less conspicuous birds. 

While the area-search method has the benefit of the observer being able to seek 
out birds and flush them when necessary, it is still likely to, like point counts, 
underestimate more cryptic birds. Dieni and Jones (2002) found that the area-search 
method provided less complete information about species composition in more complex 
habitat. Since coffee is more structurally complex than pasture, it is likely that the area 
searches underestimated species richness to a greater degree in coffee than in pasture, 
lowering power for any tests of differences between the two habitat types. Dieni and 
Jones, however, spent only 60 minutes in each of their plots, compared to 180 minutes in 
each of the sites for this study. Thus, these surveys likely provide more complete 
coverage and a better characterization of the species composition. 

There was a large discrepancy between plot sizes, with areas, as calculated by 
GPS data, ranging from 5637 m to 18957 m . These area calculations, while the most 
accurate possible, do not take into account the steep slopes at many of the plots, some of 
which were estimated as up to 25%. To test for the effect of plot size on the data, 
Pearson's r was calculated between plot size and observed species richness, and was 
found to be not significant at the p<0.05 level (N=8, r= -0.124, p=0.647). This was true 
also when the habitat types were analyzed separately (N=4; Pasture r= -0.438, p=0.278; 
Coffee r=0.546, p=0. 161). This is probably because, while it would be expected that 
more species would be seen in a larger plot, the survey effort was kept constant across all 
plots by using timed searches. 



177 



In addition, to ensure that any bias arising from varying plot sizes not detected in 
the above test would be distributed evenly between habitat types, a 2-tailed Mest was 
performed for site sizes and plot sizes vs. habitat type. Neither test showed a significant 
association at p<0.05 between habitat type and plot size (t=0.667, df=14, p=0.515) or site 
size (/=-0.138, df=14, p=0.895). 

John D. Alexander (pers. comm., 2003) recommends that sites for bird surveys 
be chosen at least 250m apart from one another to avoid double-counting of individuals. 
This goal was not achievable for all sites, due to logistical constraints and the difficulty in 
finding homogeneous patches of habitat large enough to survey. Pasture 4 was less than 
250m from Coffee 3, Coffee 4 and Pasture 3, and Pasture 3 was less than 250m from 
Coffee 3. 

The effect of double-counting individuals that moved between sites would be to 
under-estimate any differences between sites. In the case of different habitats (coffee and 
pasture) this would tend to result in an increased Type II error rate, but also a decrease in 
the Type I error rate. Thus, the proximity of Pasture 3 to Coffee 3, and of Pasture 4 to 
Coffee 3 and 4, means that any statistical tests comparing pasture to coffee will be less 
likely to detect a real change in fauna, but also less likely to report a false result, making 
the tests more robust. In other words, a statistical test showing a significant difference 
between sites would have a higher probability of reflecting a real difference between sites 
than that reflected by the p value. 

The proximity of Pasture 3 to Pasture 4 will have the opposite effect: if 
individuals moving between the two sites are double-counted, it would result in less 
variation among pasture sites than if the sites were truly independent. This would 
increase the Type I error rate and make the statistical tests less accurate. 

To determine the importance of these effects, the Classic Sorensen similarity 
index and Chao-Sorensen similarity estimator were calculated for each pair of pasture 
sites (Table B.2). To calculate any similarity index, all 12 samples from each site had to 
be pooled, which would result in double-counting of many individuals during separate 
surveys if abundance data were to be used. The Classic Sorensen index and Chao- 



Director, Klamath Bird Observatory 
178 



Sorenson Incidence-Based Estimator were therefore chosen because they are independent 
of abundance data (Colwell 2005a), and rely only on presence or absence of each species 
in each survey. The Classic Sorensen similarity index relies only on species actually 
observed in the sample, while the Chao-Sorensen similarity estimator relies on a 
calculated estimate of all species expected to be in the sample based on observed species. 

Pearson's r was calculated using the minimum distance between sites (i.e., the 
distance between the two nearest edges of the sites as determined by GPS data) as the 
independent variable and the Sorensen Classic index and Chao-Sorensen estimator as 
dependent variables. The test was repeated for all site pairs, for site pairs of differing 
habitat types, for pairs of similar habitat types, and for all site pairs plus within-site plot 
pairs (i.e., plots A and B for each site, with distance set at zero). See Table B.3 for 
results. 

The Sorensen Classic index was significantly correlated with distance only for all 
sites combined, with within-site similarity included. The Chao-Sorensen estimator was 
also significantly correlated with distance in this case, as well as for site pairs of differing 
habitats. These results indicate that distance between sites will bias the results, but that 
the effect of this bias will be to lower the Type I error rate and increase the Type II error 
rate. Thus, the surveys are less likely to detect a real difference between habitat types 
(i.e., power is lowered), but any statistically significant difference detected can be 
considered to be robust. 



179 



180 






>- w n *mcD 


CVI 


i- (M CO "J in CD 


"S 





"S 




□_ 


£ 


□_ 


£ 




3 




3 




CO 




co 



43 

fco 





t-cm n *ma) 


CVI 


*- (M CO "J IT) CD 


"5 





~S 




□_ 


£ 


□_ 


£ 




3 




3 




CO 




CO 













i- CM CO * U) (D 


CVI 


i- CM CO "J LO CD 


"S 





"3 




□_ 


£ 


□_ 


£ 




3 




3 




CO 




CO 



gj 

^ 



fco 





t- CM CO ■* LO CD 


C-vl 


i- cmco * m co 


"5 


>. 


"S 


>■ 










Q_ 


£ 


□_ 


£ 




3 




3 




CO 




CO 



fcO 



c/l 
CU 

3 

es 

H 

t3 
S 
« 

«! 
CU 
•- 

■ FN 

fa 






i- CM CO f LO CD 


CVI 


t- CM CO ■* LO CD 


1 ■ 








o 





o 





□_ 


£ 


Q_ 


£ 




3 




3 




CO 




CO 





i- CM CO "J LO CD 


rvi 


i- CM CO -*■ lo CD 


"S 


>■ 


"5 


>. 










□_ 


£ 


□_ 


£ 




3 




3 




CO 




CO 



bo 



fco 



i- CM CO "J LO CD 



t- CM CO ■* LO CD 



CO 



bo 





i- CM CO f CO CD 


CM 


i- CM CO ■* LO CD 


"3 





"3 




□_ 


£ 


a_ 


£ 




3 




3 




CO 




CO 



22 



c 
i) 

T3 

I) 
> 



O 
•- 

3 

015 



181 



H 
as 

S" 



X/i 

o 

a 

o 

o 



re 



n 
4^ 

co- 


n 


n 


n 

U> 

pa 


n 

to 

a- 


n 

to 
pa 


n 

5- 


n 

pa 


4^ 


T3 
4^ 
pa 


T3 
U> 

a" 


■o 
pa 


-a 

to 


T3 
to 
pa 


5* 


— 

pa 














^1 






















00 










LA 

-J 
















00 










la ! Ion !^i ! 








00 


^i 




00 


00 










LA 

to 




IO 

00 


o 
4^ 












LA 

o 




to 
to 


to 






























^1 




00 






























4- 
la 




to 

LA 




0C 


LA 


LA 










^1 


OC 














00 




to 
-J 


4^ 
U> 


to 

o 










o 


o 
o 














to 
-^1 






0C 




























00 






LA 

ON 




























to 

00 






























On 


LA 


00 






























O 
NO 


4^ 


o 










^1 

o 

o 


On 

4- 

O 






















00 


la 


LA 


on 


-J 




00 


00 




LA 




On 


-J 










NO 


la 


U> 


0C 


to 




o 
o 


to 
-J 




LA 
LA 




nC 


o 
o 










1—1 










^j 




-J 


ON. 






OC 


00 


LA 


LA 






NO 










LO 
LA 




o 
o 


to 
-J 






o 
^1 


to 


00 


to 

O 






U> 








0C 




00 


















LA 


ON. 


NO 








la 




— 

NO 


















4- 


O 

o 


-p^ 


On 


ON 
















LA 


LA 


LA 


-J 


00 






NO 


4^ 

o 


to 














1 LA 1 — 
jOC |ON 


LA 


LA 
-J 


it 






LA 


-J 






OC 










^1 




OC 


00 










NO 


u> 






LA 
OC 










LA 

00 




L^> 
~~1 


^o 










On 








o> a> ui 


LA 




NO 










NO 


- 


LA 


LA 




-J 






LA 


LA 














ON. 


On 


^1 


-J 


NO 






4^ 


LA 














4^ 


LO 


4- 


LA 

00 


00 


00 








On 




-J 


-J 


00 












NO 


4^ 

00 








00 




o 

LA 


to 

LA 


00 












<D 


















laI 


LA 


LA 


O* 


On 


0C 


00 


<D 


| | | | | | | ILaI 




4^ 

O 


LA 
LA 




o 


00 


O 




-J 


-J 


O 


00 


^1 




00 




On 














NO 



Pair 


Distance 


Sorensen Classic 


Chao-Sorensen-Est 
Incidence-based 


PI 





0.788 


0.894 


P2 





0.638 


0.944 


P3 





0.81 


0.926 


P4 





0.615 


0.864 


CI 





(>.7 l )2 


0.929 


C2 





0.638 


1.000 


C3 





0.64 


0.944 


C4 





0.72 


0.856 


PI-P2 


2154 


0.654 


0.859 


P1-P3 


1335 


0.622 


0.778 


P1-P4 


1418 


0.596 


0.748 


P2-P3 


1088 


0.632 


0.834 


P2-P4 


1442 


0.746 


0.918 


P3-P4 


121 


0.615 


0.849 


C1-C2 


573 


0.754 


0.973 


C 1 -C3 


1098 


0.667 


0.837 


C1-C4 


845 


0.787 


0.944 


C2-( ' 3 


566 


0.788 


0.963 


C2-C4 


695 


0.844 


0.973 


C3-C4 


442 


0.788 


0.928 


Pl-Cl 


2395 


0.49 


0.522 


P1-C2 


1718 


0.577 


0.722 


P1-C3 


1184 


0.593 


0.724 


P1-C4 


1730 


0.577 


0.577 


P2-C1 


1120 


0.557 


0.677 


P2-C2 


730 


0.656 


0.888 


P2-C3 


1361 


0.636 


0.806 


P2-C4 


1589 


0.656 


0.74 


P3-C1 


891 


0.519 


0.547 


P3-C2 


322 


0.561 


0.754 


P3-C3 


18 


0.61 


0.733 


P3-C4 


388 


0.596 


0.68 


P4-C1 


896 


0.536 


0.729 


P4-C2 


527 


0.678 


0.892 


P4-C3 


162 


0.656 


0.826 


P4-C4 


193 


0.61 


0.715 



Table B.2. Similarity indexes for all site pairs, and for plot pairs within sites. 



183 



O H 

o ta 

3 2 

<L » 

S 53 

g S" 

o 13 

S <T 

ft m 

H? " 

lli w 

o o 

5 § 

< £L 

r^-" CO 

^ &. 

co O 

t=t o 

co O 

w O 

g a, 

*" s- 

* fl> 

3 

R <w 

« O 

K> O 

<§' 3 

a >t3 

» r 1 

CD 3 

5 era 



11 ? 

o S- 

ft 00 

!— O 

i-l 

r. 

3 



n 



3 



3 

n 

3" 

O 
i 

o 

-5 

ft 

3 



c 

3 



> 


> 


00 


O 


o 






p 




o 
3 


Cfi 


Cfl 


3 

CD 




r*- 


r*- 




>-i 


T3 


CC 


CD 


tr 


3' 


S5 


►T3 


►T3 


a; 

P 


00 

tr 

p 

a; 


3 

CB 

o 



►13 

3" 




►13 
CD 


P 




ga 




r^- 


re 


►13 




"cT 


►T3 




o" 




era 


0> 




r^- 






,. — v 




T3 




j 


*a 




ta 




T3 


p 








03 


C/3 




>-t' 




C/3 


rj 




m 




r*- 


c 




3 




1-1 


CD 




r*- 




re 


< 




tr 




< 


C/3 








ga 






C3 






O 




en 




■a 


o 








p 


CD 




IZ3 




CD 


CD 


O 




























Is 


^1 


vo 


VO 


VO 


a 2 


'^ 


ON 


OO 


(-/> 


« s 


o 


-£> 


H "' 


to 


n S3 


OJ 


IO 


H 


H 


Z 


ON 


00 


to 


ON 










C/3 










O 










5 ^ 










2 re 










S3 S3 










SE 1 










2 CB 










SS o 


* 

1 


1 


1 


| 


«s 


o 


O 


o 


o 


as o 


Lo 


to 


Lo 


'-t^ 


S ^ 


ON 


-J 


OO 


o 




LO 


to 


VO 


^1 


ic (2-t 


o 


o 


o 


o 


S. 


o 


J— ■ 


to 


I—. 


» ^ 


to 


ON 


1^ 


K^ 




VO 


00 


H 


OO 










9^ 

E CD 










as aj 










^ i 


* 






* 


© 53 

l-J 


o 


o 


o 


o 


CD rt 



References 

Aguilar, B. J. 1992. Consequences of the Suspension of Economic Clauses in the 

International Coffee Agreement: A Free Coffee Market? University of Georgia, 
Athens, GA. 

Albertin, A., and P. K. R. Nair. 2004. Farmers' perspectives on the role of shade trees in 
coffee production systems: An assessment from the Nicoya Peninsula, Costa Rica. 
Human Ecology 32:443-463. 

Alfaro, B. S. 1992. Ante la crisis cafetalera. La Nacion, San Jose, Costa Rica, August 9. 

American Ornithologists Union. 2005. List of the 2,037 bird species (with scientific and 
English names) known from the A.O.U. check-list area. American Ornithologists 
Union, McLean, VA. http://www.aou.org/checklist/birdlist46.pdf. Accessed 
November 30, 2005. 

Anonymous. 1992a. Crisis social y caficultura. La Republica, San Jose, Costa Rica, July 

25. 

Anonymous. 1992b. La Crisis: de mercado y estructura. La Prensa Libre, San Jose, Costa 
Rica, February 1 1 . 

Anonymous. 1995. La crisis del cafe. La Prensa Libre, San Jose, Costa Rica, July 8. 

Anonymous. 2002. Crisis del cafe con efectos explosivos. La Republica, San Jose, Costa 
Rica, May 22. 

Anonymous. 2005. Starbucks premia a productores costarricenses de cafe organico. La 
Nacion, San Jose, Costa Rica, August 11. 

Antongiovanni, M., and J. P. Metzger. 2005. Influence of matrix habitats on the 
occurrence of insectivorous bird species in Amazonian forest fragments. 
Biological Conservation 122:441-451. 

Arellano, L., M. E. Favila, and C. Huerta. 2005. Diversity of dung and carrion beetles in 
a disturbed Mexican tropical montane cloud forest and on shade coffee 
plantations. Biodiversity and Conservation 14:601-615. 

Arguedas M., J. M. 1988. Coto Bras, canton olvidado. La Nacion, San Jose, Costa Rica, 
December 1. 

Armbrecht, I., and I. Perfecto. 2003. Litter-twig dwelling ant species richness and 

predation potential within a forest fragment and neighboring coffee plantations of 

185 



contrasting habitat quality in Mexico. Agriculture, Ecosystems and Environment 
97:107-115. 

Associated Press Costa Rica. 2005. Starbucks ofrece en el mercado cafe de Tarrazu. La 
Nacion, San Jose, Costa Rica, July 9. 

Associated Press Guatemala. 1994. Productores de banano no lograron consenso. La 
Nacion, San Jose, Costa Rica, January 30. 

Bandeira, F. P., C. Martorell, J. A. Meave, and J. Cabellero. 2005. The role of rustic 
coffee plantations in the conservation of wild tree diversity in the Chinantec 
region of Mexico. Biodiversity and Conservation 14:1225-1240. 

Barquero S., M. 1998. Caficultores temen crisis. La Nacion, San Jose, Costa Rica, 
December 1 1 . 

Barquero S., M. 1999. Caficultores temen crisis. La Nacion, San Jose, Costa Rica, 
August 24. 

Barquero S., M. 2002. Beneficios en calle sin salida. La Nacion, San Jose, Costa Rica, 
December 16. 

Barquero S., M. 2004a. Crisis del cafe cobra otra victima. La Nacion, San Jose, Costa 
Rica, August 6. 

Barquero S., M. 2004b. Denuncian "robo" de marca cafetera. La Nacion, San Jose, Costa 
Rica, December 8. 

Barquero S., M. 2004c. Divisiones diferentes. La Nacion, San Jose, Costa Rica, 
December 8. 

Barquero S., M. 2004d. Firmas pagan 50% mas por cafe nacional de alta calidad. La 
Nacion, San Jose, Costa Rica, November 11. 

Barquero S., M. 2004e. Temen cierre de una cooperativa cafetalera. La Nacion, San Jose, 
Costa Rica, July 3. 

Barquero S., M. 2005. Caficultores afrontan aun graves secuelas de crisis. La Nacion, 
San Jose, Costa Rica, May 2. 

Barrantes, M. G., and J. M. Pana. 2005. Migration interna en Costa Rica en el perfodo 
1927-2000. Instituto Nacional de Estadistica y Censos, San Jose, Costa Rica. 
http://www.inec.go.cr/06Publicaciones/07SerieCensal/21%20Migraci6n%20inter 
na%20en%20Costa%20Rica%20en%20el%20periodo%201927- 
2000/pl7%20gomez.pdf. Accessed March 22, 2005. 

186 



Beer, J. 1987. Advantages, disadvantages and desirable characteristics of shade trees for 
coffee, cacao and tea. Agroforestry Systems 5:3-13. 

Beer, J., R. Muschler, D. Kass, and E. Somarriba. 1998. Shade management in coffee and 
cacao plantations. Agroforestry Systems 38:139-164. 

Bevan, D. L., P. Collier, and J. Gunning. 1990. Fiscal response to a temporary trade 
shock; the aftermath of the Kenyan coffee boom. The World Bank Economic 
Review 3. 

Borrero, H., and J. Ignacio. 1986. La substitucion de cafetales de sombrio por caturalles y 
su efect negativo sobre la fauna de vertebrados. Caldasia 15:71-75. 

Boucher, D. H. 1983. Coffee (Cafe). Pages 86-88 in D. H. Janzen, editor. Costa Rican 
Natural History. University of Chicago Press, Chicago, IL. 

Bray, D. B., J. L. P. Sanchez, and E. C. Murphy. 2002. Social dimensions of organic 
cofee production in Mexico: Lessons for eco-labeling initiatives. Society and 
Natural Resources 15:429-446. 

Brenes, L. 2002. Nadie escapa de la crisis del cafe. La Nation, San Jose, Costa Rica, 
December 16. 

Briceno, J. A. 1992. Crisis del cafe y medidas compensatorias. La Republica, San Jose, 
Costa Rica, March 5. 

Brignoli, H. P. 1994. Naturaleza de las economias de exportacion, con especial referenda 
a las economia cafetaleras. Pages 27-35 in M. S. K., editor. Crisis y perspectivas 
del cafe latinoamericano. Instituto del Cafe/Universidad National, San Jose, 
Costa Rica. 

Burke, D. M., and E. Nol. 1998. Influence of food abundance, nest-site habitat, and forest 
fragmentation on breeding overnbirds. Auk 115:96-104. 

Burnham, K. P., and W. S. Overton. 1978. Estimation of the size of a closed population 
when capture probabilities vary among animals. Biometrika 65:623-633. 

Burnham, K. P., and W. S. Overton. 1979. Robust estimation of the size of a closed 

population when capture probabilities vary among animals. Ecology 60:927-936. 

Caramori, P. H., A. Androcioli-Filho, and A. C. Leal. 1996. Coffee shade with Mimosa 
scabrella Benth. for frost protection in southern Brazil. Agroforestry Systems 
33:205-214. 

Cardoso, C. 1973. La formation de la hacienda cafetalera en Costa Rica. Revista de 
Estudios Sociales Centroamericanos 6. 

187 



Chao, A. 1987. Estimating the population size for capture-recapture data with unequal 
catchability. Biometrics 43:783-791. 

Chao, A. 2004. Species Richness Estimation in S. Kotz, B. Vidakovic, and N. 

Balakrishnan, editors. Encyclopedia of Statistical Sciences. John Wiley & Sons, 
New York, NY. 

Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R. Guariguata. 1998. Statistical 
methods for estimating species richness of woody regeneration in primary and 
secondary rain forests of NE Costa Rica. Pages 285-309 in F. Dallmeier, and J. A. 
Komiskey, editors. Forest biodiversity research, monitoring and modeling: 
conceptual background and Old World case studies. Parthenon Publishing, Paris. 

The coffee crisis in the western hemisphere. 2002. Page 156. Subcommittee on the 

Western Hemisphere, Committee on International Relations, Washington, D.C. 

Colwell, R. K. 2001. User's Guide to Estimates : Statistical estimation of species richness 
and shared species from samples. Department of Ecology and Evolutionary 
Biology, University of Connecticut, Storrs, CT. 

http://viceroy.eeb.uconn.edu/Estimates6/EstimateS%20Pages/UsersGuideHTML/ 
EstiMateS6Guide.html. Accessed June 30, 2003. 

Colwell, R. K. 2005a. Estimates 7.5 User's Guide. University of Connecticut, Storrs, CT. 
http://viceroy.eeb.uconn.edu/EstimateS7Pages/EstS7UsersGuide/EstimateS7User 
sGuide.htm. Accessed June 6, 2005. 

Colwell, R. K. 2005b. Estimates : Statistical estimation of species richness and shared 
species from samples. University of Conneticut, Storrs, CT. 

Cordero, F., R. Garcia, and A. Izquierdo 1987. Economia y geographia del desarollo en 
America Latina. Fondo de Cultura Economica, Mexico. 

Corrales B., J. M. 1992. Crisis del cafe y riesgo democratico. La Nacion, San Jose, Costa 
Rica, May 5. 

Coste, R. 2001. Coffee Production in I. Encyclopedia Britannica, editor. Britannica 2001 
Standard Edition CD-ROM. Encyclopedia Britannica, Inc. Accessed October 15, 
2004. 

Daily, G. C, P. R. Ehrlich, and G. A. Sanchez-Azofeifa. 2001. Countryside 

biogeography: use of human-dominated habitats by the avifauna of southern 
Costa Rica. Ecological Applications 11:1-13. 



188 



Daviron, B. 1994. La crisis del mercado cafetalero internacional an una perspectiva de 
largo plazo. Pages 11-26 in M. S. K., editor. Crisis y perspectivas del cafe 
latinoamericano. Instituto del Cafe/Universidad Nacional, San Jose, Costa Rica. 

Daviron, B., and F. Lerin. 1986. Le marche mondial du cafe. Le Monde Diplomatique. 

Defries, R., A. Hansen, A. C. Newton, and M. C. Hansen. 2005. Increasing isolation of 
protected areas in tropical forests over the past twenty years. Ecological 
Applications 15:19-26. 

DGEC 1974. Costa Rica: Censo Nacional de Poblacion y Vivienda 1973. Direccion 
General de Estadistica y Censos, San Jose, Costa Rica. 

DGEC 1987. Censo Agropecuario 1984. Ministerio de Economia, Industria y Comercio, 
San Jose, Costa Rica. 

Dicum, G., and N. Luttinger 1999. The Coffee Book: anatomy of an industry from crop 
to the last drop. New Press, New York, NY. 

Dieni, J. S., and S. L. Jones. 2002. A field test of the area search method for measuring 
breeding bird populations. Journal of Field Ornithology 73:253-257. 

DiFulvio, A. 1947. Le Cafe dans le Monde. UN Food and Agriculture Oranization, 
Rome, Italy. 

Donald, P. F. 2004. Biodiversity impacts of some agricultural commodity production 
systems. Conservation Biology 18:17-37. 

Dumas, M. 1992. Coffee into the next century: a trader's perspective. Page 17. 8th 
International Coffee Congress, Jerez de la Frontera. 

Dunlap, W. P. 1981. An interactive FORTRAN IV program for calculating power, 

sample size, or detectable differences in means. Behavior Research Methods & 
Instrumentation 13:757-759. 

Edwards, S. 1984. Coffee, money and inflation in Colombia. World Development 11/12. 

Eisemann, P. 1982. L'organization Internationale du Commerce des Produits de Base. 
Publication de la Faculte de Droit de Paris VI, Paris, France. 

EIU. 1991. Coffee to 1995: recovery without crutches. Special Report no. 21161, 
London. 

Engstrom, R. T., and F. C. James. 1981. Plot size as a factor in winter bird population 
studies. Condor 83:34-41. 

189 



Estrada, A., R. Coates-Estrada, and D. A. Meritt Jr. 1997. Anthropogenic landscape 
changes and avian diversity at Los Tuxtlas, Mexico. Biodiversity and 
Conservation 6:19-43. 

Federation de Cafeteros de Colombia. 1978. 50 Anos, 1927-1977. Boletin de 
information de estadistica sobre cafe. 

Fernandez, J. B. M., and G. P. Alvarez. 2004. Precio de Liqudacion Final Cosecha 2003- 
2004. Insituto del Cafe de Costa Rica, San Jose, Costa Rica. 

Fernandez, J. B. M., E. R. Rojas, and M. A. A. Molina 2000. Informe Sobre la Actividad 
Cafetalera de Costa Rica. Instituto del Cafe de Costa Rica, San Jose, Costa Rica. 

Fernandez, J. B. M., E. R. Rojas, and M. A. A. Molina 2002. Informe Sobre la Actividad 
Cafetalera de Costa Rica. Instituto del Cafe de Costa Rica, San Jose, Costa Rica. 

FLO. 2005. Fair Trade Labeling Organizations International web site. Fair Trade 

Labeling Organizations International, Bonn, Germany, http://www.fairtrade.net. 
Accessed December 4, 2005. 

Geupel, G. R., C. J. Ralph, and S. L. Jones. 2003. The area search bird census: a habitat 
based method. Unpublished work. 

Gillison, A. N., N. Liswanti, S. Budidarsono, M. v. Noordwijk, and T. P. Tomich. 2004. 
Impact of cropping methods on biodiversity in coffee agroecosystems in Sumatra, 
Indonesia. Ecology and Society 9:7-38. 

Gobbi, J. A. 2000. Is biodiversity-friendly coffee financially viable? An analysis of five 
different coffee production systems in western El Salvador. Ecological Economics 
33:267-281. 

Gotelli, N. J., and R. K. Colwell. 2001. Quantifying biodiversity: procedures and pitfalls 
in the measurement and comparison of species richness. Ecology Letters 4:379- 
391. 

Greenberg, R., P. Bichier, A. C. Angon, and R. Reitsma. 1996. Bird populations in shade 
and sun coffee plantations in Central Guatemala. Conservation Biology 11:448- 
459. 

Greenberg, R., P. Bichier, and J. Sterling. 1997. Bird populations in rustic and planted 
shade coffee plantations of Eastern Chiapas, Mexico. Biotropica 29:501-514. 

Gresser, C, and S. Tickell. 2002. Mugged: Poverty in your cup. 60 pp. Oxfam 
International, Oxford, UK. 



190 



Griffith, D. M. 2000. Agroforestry: a refuge for tropical biodiversity after fire. 
Conservation Biology 14:325-326. 

Grosser, K., J. Hudakand, and T. Ralston. 2005. The Community Agroecology Network 
(CAN) and The Evergreen State College: A presentation to the Farm to Table 
academic program. The Evergreen State College, Olympia, WA. 

Guevara M., J. D. 1993. Expertos analizan crisis cafetalera. La Nation, San Jose, Costa 
Rica, July 14. 

Gutierrez, J. C. 1992. La crisis de la economia cafetera internacional y su impacto en los 
paises productores. VIII Congreso Internacional del Cafe, Jerez de la Frontera. 

Harper, L. 1989. The persistence of ant- following birds in small Amazonian forest 
fragments. Acta Amazonica 19:249-263. 

Hartshorn, G., L. Hartshorn, A. Atmella, L. D. Gomez, A. Mata, R. Morales, R. Ocampo, 
D. Pool, C. Quesada, C. Solera, R. Solorzano, G. Stiles, J. Joseph Tosi, A. 
Umana, C. Villalobos, and R. Wells 1982. Costa Rica: Country Environmental 
Profile. Tropical Science Center, San Jose, Costa Rica. 

Hayek, L. C, and M. A. Buzas 1996. Surveying natural populations. Columbia 
University Press, New York. 

Heltshe, J., and N. E. Forrester. 1983. Estimating species richness using the jackknife 
procedure. Biometrics 39:1-11. 

Hernandez Navarro, L. 1995. Cafe con aroma de burocracia. La Jornada, Mexico City, 
March 14. 

Hidalgo, G. 1983. Productividad de dos niveles technologicos usados en la production de 
cafe en Costa Rica. VI Simposio de caficultura de Centroamerica, Panama. 

Hietz, P. 2005. Conservation of vascular epiphyte diversity in Mexican coffee 
plantations. Conservation Biology 19:391-399. 

Hilje, B., C. Naranjo, and M. S. Kutschbach. 1994. "No se puede dejar perder y no paga 
para los gastos." Testimonios de caficultores costarricenses ante la crisis. Pages 
163-219 in M. S. K., editor. Crisis y perspectivas del cafe latinoamericano. 
Instituto del Cafe/Universidad National, San Jose, Costa Rica. 

Holdridge, L. R. 1967. Life Zone Ecology. Tropical Science Center, San Jose, Costa 
Rica. 

Holdridge, L. R. 1971. Forest environments in tropical life zones: a pilot study. Pergamon 
Press, Oxford, New York. 

191 



Holmes, B. 1996. Life on the edge. New Scientist 151:38-39. 

Hughes, J. B., G. C. Daily, and P. R. Ehrlich. 2002. Conservation of tropical forest birds 
in countryside habitats. Ecology Letters 5:121-129. 

ICAFE. 1993. Informe Sobre la Actividad Cafetalera de Costa Rica. XXII Congreso 
Nacional Cafetalero. Instituto del Cafe de Costa Rica, San Jose, Costa Rica. 

ICAFE. 1997. Informe Sobre la Actividad Cafetalera de Costa Rica. XXVI Congreso 
Nacional Cafetalero. Instituto del Cafe de Costa Rica, San Jose, Costa Rica. 

ICAFE. 2003. Boletin de Precios y Mercado Internacional. Instituto del Cafe de Costa 
Rica, San Jose, Costa Rica. 

ICAFE. 2005a. Cafe recibido acumulado: cosecha 0405. Instituto del Cafe de Costa Rica, 
San Jose, Costa Rica. 

ICAFE. 2005b. Comportamiento Produccion de Cafe. Instituto del Cafe de Costa Rica, 
San Jose, Costa Rica. http://www.icafe.go.cr/newsite.nsf/repint0257OpenPage. 
Accessed January 4, 2005. 

ICAFE. 2005c. Sistema de Information. Instituto del Cafe de Costa Rica, San Jose, Costa 
Rica, http://www.icafe.go.cr/newsite.nsf. Accessed January 8, 2005. 

ICO. 2005a. Breakdown of exports of arabica and robusta for countries exporting both 
types of coffee. International Coffee Organization, London, England. 
http://www.ico.org/prices/ml-a.htm. Accessed May 24, 2005. 

ICO. 2005b. Coffee Trade Statistics. International Coffee Organization, London, 
England, http://www.ico.org. Accessed October 31, 2005. 

ICO. 2005c. Exports by exporting countries to all destinations. International Coffee 
Organization, London, England, http://www.ico.org/prices/ml.htm. Accessed 
November 21, 2005. 

ICO. 2005d. ICO Indicator prices: monthly and annual averages. Coffee Prices. 
International Coffee Organization, London, England. 
http://www.ico.org/prices/p2.htm. Accessed November 6, 2005. 

ICO. 2005e. Total production of exporting countries: Crop years 1999/00 to 2004/05. 
International Coffee Organization, http://www.ico.org/prices/po.htm. Accessed 
May 24, 2005. 

IGNCR. 1975. Canas Gordas; Hoja 3642 III. Instituto Geografico Nacional de Costa 
Rica, San Jose, Costa Rica. 

192 



IGNCR. 1998. Costa Rica L13 1:40,000 98-12-26 Terra 14-15. Institute Geografico 
Nacional de Costa Rica, San Jose, Costa Rica. 

INEC 2001. IX Censo Nacional de Poblacion y V de Vivienda del 2000: Resultados 
Generales. Institute Nacional de Estadistica y Censos, San Jose, Costa Rica. 

INEC. 2003a. Encuesta de hogares de propositos multiples: Cifras basicas sobre pobreza 
e ingresos. Page 6. Institute Nacional de Estadistica y Censos, San Jose, Costa 
Rica. 

INEC. 2003b. Encuesta de hogares de propositos multiples: Cifras basicas sobre 
pobreza e ingresos. Institute Nacional de Estadistica y Censos, San Jose, Costa Rica. 

INEC 2004. Censo Cafetalero: Aspectos metodologicos y principales resultados. Institute 
Nacional de Estadistica y Censos, San Jose, Costa Rica. 

Inter-American Development Bank, World Bank, and U. S. A. f. I. Development. 2002. 
Managing the competitive transition of the coffee sector in Central America. 38 
pp. The coffee crisis and its impact in Central America: situation and lines of 
action, Antigua, Guatemala. 

IUCN. 2004. IUCN Red List of Threatened Species. International Union for the 

Conservation of Nature and Resources, Cambridge, UK. www.iucnredlist.org. 
Accessed December 1, 2005. 

Jimenez, M. F. 1994. From plantation to cup: the beverage of North American capitalism 
1830-1930 in W. Roseberry, L. Gudmundson, and M. Samper, editors. Coffee, 
Society and Power in Latin America. Johns Hopkins, Baltimore, MD. 

Johnson, M. D. 2000. Effects of shade-tree species and crop structure on the winter 

arthropod and bird communities in a Jamaican shade coffee plantation. Biotropica 
32:133-145. 

Jules, E. S., and P. Shahani. 2003. A broader ecological context to habitat fragmentation: 
Why matrix habitat is more important than we thought. Journal of Vegetation 
Science 14:459-464. 

Karnes, T. L. 2001. Costa Rica in I. Encyclopedia Britannica, editor. Encyclopedia 
Britannica. Encyclopedia Britannica, Inc. Accessed October 15, 2004. 

Klein, A.-M., I. Steffan-Dewenter, D. Buchori, and T. Tscharntke. 2004. Effects of land- 
use intensity in tropical agroforestry systems on coffee flower-visiting and trap- 
nesting bees and wasps. Conservation Biology 16:1003-1014. 



193 



Komar, O. 2000. Effects of coffee shade trees on migratory bird abundance. IV Congress 
of the Mesoamerican Society for Biology and Conservation, Panama City, 
Panama. 

Komar, O. 2001. Effects on resident birds of shade trees in coffee plantations, and 

applications to coffee certification. Proceedings of the American Ornithologist's 
Union Conference, Seattle, WA, August 18. 

Kutschbach, M. S. 1994a. Crisis e Historia. Pages 1 1-26 in M. S. K., editor. Crisis y 

perspectivas del cafe latinoamericano. Instituto del Cafe/Universidad Nacional, 
San Jose, Costa Rica. 

Kutschbach, M. S. 1994b. Las caficulturas latinoamericanas en coyunturas criticas. Pages 
79-131 in M. S. K., editor. Crisis y perspectivas del cafe latinoamericano. 
Instituto del Cafe/Universidad Nacional, San Jose, Costa Rica. 

Kutschbach, M. S. 1994c. Las fronteras del cafe: respuestas a la crisis u nuevas 

exploraciones. Pages 223-259 in M. S. K., editor. Crisis y perspectivas del cafe 
latinoamericano. Instituto del Cafe/Universidad Nacional, San Jose, Costa Rica. 

Lee, S. M., and A. Chao. 1994. Estimating population size via sample coverage for 
closed capture-recapture models. Biometrics 50:1-1 1. 

Lindell, C. A., W. H. Chomentowski, and J. R. Zook. 2004. Characteristics of bird 
species using forest and agricultural land covers in southern Costa Rica. 
Biodiversity and Conservation 13:2419-2441. 

Lindell, C. A., and M. Smith. 2003. Nesting bird species in sun coffee, pasture, and 

understory forest in southern Costa Rica. Biodiversity and Conservation 12:423- 
440. 

Lindsey, B. 2003. Grounds for Complaint? Understanding the "coffee crisis." Trade 
briefing paper No. 16. The Cato Institute, Washington, D.C. 

Lovejoy, T. E., R. O. Bierregaard, A. B. Rylands, J. R. Malcolm, C. E. Quintela, L. H. 
Harper, K. S. Brown, A. H. Powell, G. V. N. Powell, H. O. R. Schuburt, and M. 
B. Hays. 1986. Edge and other effects of isolation on Amazonian forest fragments 
In M. E. Soule, editor. Conservation Biology: the science of scarcity and 
diversity. Sinauer Associates, Sunderland, Massachusetts. 

Luck, G. W., and G. C. Daily. 2003. Tropical countryside bird assemblages: Richness, 
composition, and foraging differ by landscape context. Ecological Applications 

13:235-247. 



194 



Lugo, A. E. 1988. Estimating reductions in the diversity of tropical forest species. Pages 
58-76 in E. O. Wilson, editor. Biodiversity. National Academy Press, 
Washington, D.C. 

Lyngbaek, A. E., R. G. Muschler, and F. L. Sinclair. 1999. Productivity, labour and 

variable costs of organic versus conventional coffee smallholdings in Costa Rica. 
Pages 216-219. Actas IV Semana Cientifica del CATIE. CATIE, Turrialba, Costa 
Rica, Turrialba, Costa Rica. 

Magurran, A. E. 1988. Ecological diversity and its measurement. Princeton University 
Press, Princeton. 

Magurran, A. E. 2004. Measuring Biological Diversity. Blackwell Publishing, Maiden, 
MA. 

Manger, W. F. 1992. Colonization on the southern frontier of Costa Rica: a historical- 
cultural landscape. 251 pp. Department of Geography. Memphis State University, 
Memphis, TN. 

Marcano-Vega, H., T. M. Aide, and D. Baez. 2002. Forest regeneration in abandoned 
coffee plantations and pastures in the Cordillera Central of Puerto Rico. Plant 
Ecology 161:75-87. 

Mason, D. 1996. Responses of Venezuelan understory birds to selective logging, 
enrichment strips, and vine cutting. Biotropica 28:296-309. 

Messer, K. D., M. J. Kotchen, and M. R. Moore. 2000. Can shade-grown coffee help 

conserve tropical biodiversity? A market perspective. Endagered Species Update 
17:125-131. 

MGP 1986. Censo de poblacion 1984. Ministerio de Gobernacion y Policia, Imprenta 
Nacional, San Jose, Costa Rica. 

Moguel, P., and V. M. Toledo. 1999. Biodiversity conservation in traditional coffee 
systems of Mexico. Conservation Biology 13:11-21. 

Montoya, Y. 2003. Cafe con aroma de mujer. La Nation, San Jose, Costa Rica, May 15. 

Mora, A. C. 2001a. Crisis sentencia a pequeno caficultores. La Republica, San Jose, 
Costa Rica, July 15. 

Mora, C. 2001b. Disminuye uso de agroqufmicos. La Republica, San Jose, Costa Rica, 
January 27. 

Morales Ch., S. 1999. Preven crisis cafetalera. La Prensa Libre, San Jose, Costa Rica, 
August 14. 

195 



Morales, S. 2000. Cafe y banano debiles ante crisis. La Republica, San Jose, Costa Rica, 
December 19. 

Munoz N., M. 1997. FONECAFE debe ser para crisis. La Prensa Libre, San Jose, Costa 
Rica, December 4. 

Murillo, N. 1994. Buscan salida a crisis de caficultura. La Nacion, San Jose, Costa Rica, 
January 30. 

Murillo, N. 1996. Vivir del cafe. Pages 21-22. Rumbo. 

Murillo, N. 1999. Crisis ahoga productos tradicionales. La Republica, San Jose, Costa 
Rica, November 14. 

Murillo, W. 1992. Crisis cafetalera y bananera golpea a la America Latina. La Republica, 
San Jose, Costa Rica, June 10. 

Muschler, R. G. 1998. Tree-crop compatibility in agroforestry: Production and quality of 
coffee grown under managed tree shade in Costa Rica. 221 pp. Ph.D. dissertation, 
University of Florida. 

Muschler, R. G. 2001. Shade improves coffee quality in a sub-optimal zone of Costa 
Rica. Agroforestry Systems 51:131-139. 

Nestel, D., F. Dickschen, and M. A. Altieri. 1993. Diversity of soil macro-Coleoptera in 
agroecosys terns: an indication of habitat perturbation. Biodiversity and 
Conservation 2:70-78. 

Numa, C, J. R. Verdu, and P. Sanchez-Palomino. 2005. Phyllostomid bat diversity in a 
variegated coffee landscape. Biological Conservation 122:151-158. 

Nyhus, P., and R. Tilson. 2004. Agroforestry, elephants, and tigers: balancing 

conservation theory and practice in human-dominated landscapes of Southeast 
Asia. Agriculture, Ecosystems and Environment 104:87-97. 

O'Brien, T. G., and M. F. Kinnaird. 2003. Caffeine and Conservation. Science 300:587. 

O'Brien, T. G., and M. F. Kinnaird. 2004. Response to Dietsch, et. al. Science 303:625- 
626. 

Osorio, N. 2005a. Coffee Market Report: May 2005. International Coffee Organization, 
London, England. 

Osorio, N. 2005b. Coffee Market Report: September 2005. International Coffee 
Organization, London, England. 

196 



Osorio, N. 2005c. ICO submission to the G-8 Summit, Gleneagles, Scotland, 6-8 July 
2005. International Coffee Organization, London, England. 

Owens, I. P., and P. M. Bennett. 2000. Ecological basis of extinction risk in birds: habitat 
loss versus human persecution and introduced predators. Proceedings of the 
National Academy of Sciences 97:12144-12148. 

Palmer, M. W. 1991. Estimating species richness: the second-order jackknife 
reconsidered. Ecology 72:1512-1513. 

Pelupessy, W. 1991. Problemas relacionados con el nuevo Acuerdo Internacional del 
Cafe. Aportes al Debate 1:18-31. 

Pelupessy, W. 1993. El Mercado Mundial del Cafe: El Caso de El Salvador. Editorial 
Departamento de Investigaciones (DEI), San Jose, Costa Rica. 

Pelupessy, W. 1997. The sustainability of international coffee chains in Cote d'lvoire and 
Costa Rica. 38 pp. Symposium on Agricultural Marketing in Tropical Africa, 
Afrika Studicentrum & Vakgroep 'Marktkunde LUW. 

Pelupessy, W., and E. v. Tilburg. 1994. El mercado solidario de cafe y el pequeno 
productor en Centroamerica. Pages 239-258 in M. S. K., editor. Crisis y 
perspectivas del cafe latinoamericano. Instituto del Cafe/Universidad Nacional, 
San Jose, Costa Rica. 

Pendergrast, M. 1999. Uncommon grounds: The history of coffee and how it transformed 
our world. Basic Books, New York, NY. 

Perfecto, I., A. Mas, T. Dietsch, and J. Vandermeer. 2003. Conservation of biodiversity 
in coffee agroecosystems: a tri-taxa comparison in southern Mexico. Biodiversity 
and Conservation 12:1239-1252. 

Perfecto, I., R. A. Rice, R. Greenberg, and M. E. V. d. Voort. 1996. Shade coffee: a 
disappearing refuge for biodiversity. Bioscience 46:598-606. 

Perfecto, I., and J. Vandermeer. 1996. Microclimatic changes and the indirect loss of ant 
diversity in a tropical agroecosystem. Oecologia 108:577-582. 

Perfecto, I., and J. Vandermeer. 2002. Quality of agroecological matrix in a tropical 

montane landscape: ants in coffee plantations in Southern Mexico. Conservation 
Biology 16:174-182. 

Perfecto, I., J. Vandermeer, P. Hanson, and V. Cartin. 1997. Arthropod biodiversity loss 
and the transformation of a tropical agro-ecosystem. Biodiversity and 
Conservation 6:935-945. 

197 



Petit, L. J., and D. R. Petit. 2003. Evaluating the importance of human-modified lands for 
neotropical bird conservation. Conservation Biology 17:687-694. 

Philpott, S. M., and T. Dietsch. 2003. Coffee and conservation: a global context and the 
value of farmer involvement. Conservation Biology 17:1844-1846. 

Pielou, E. C. 1975. Ecological Diversity. John Wiley & Sons, Inc., New York, NY. 

Pimentel, D., and U. Stachow. 1992. Conserving biological diversity in 
agricultural/forestry systems. Bioscience 42:354-362. 

Pineda, E., and G. Halffter. 2004. Species diversity and habitat fragmentation: frogs in a 
tropical montane landscape in Mexico. Biological Conservation 117:499-508. 

Pineda, E., C. Moreno, F. Escobar, and G. Halffter. 2005. Frog, bat, and dung beetle 
diversity in the cloud forest and coffee agroecosystems of Veracruz, Mexico. 
Conservation Biology 19:400-410. 

Portilla R., W., and M. A. Araya M. 2002. Modelo de costos de produccion agricola de 
cafe. Page 20. Instituto del Cafe de COsta Rica, San Jose, Costa Rica. 

Portillo, L. 1993. El Convenio Internacional del Cafe y la crisis del mercado. Comercio 
Exterior 43:378-391. 

Prendas, G. 2001a. Crisis cafetalera extiende sus tentaculos. La Republica, San Jose, 
Costa Rica, November 23. 

Prendas, G. 2001b. Pocas esperanzas para cafetaleros. La Republica, San Jose, Costa 
Rica, August 28. 

Prendas, G. 2002. Cafetaleros preven recuperation de precios. La Republica, San Jose, 
Costa Rica, February 6. 

Puelschen, L., and H. J. Lutzyer. 1994. Ecological and economic conditions of organic 

coffee production in Latin America and Papua New Guinea. Angewandte Botanik 
67:204-208. 

Quesada, and Marin. 2001. Cafetaleros preocupados. Al Dia, San Jose, Costa Rica, 
December 10. 

Raaijmakers, J. G. W. 1987. Statistical analysis of the Michaelis-Menten equation. 
Biometrics 43:793-803. 

Ramirez, O. A., and R. Sosa. 2000. Assessing the financial risks of diversified coffee 

production systems: an alternative nonnormal CDF estimation approach. Journal 
of Agriculture and Resource Economics 25:267-285. 

198 



Rappole, J. H., D. I. King, and J. H. V. Rivera. 2003a. Coffee and conservation. 
Conservation Biology 17:334-336. 

Rappole, J. H., D. I. King, and J. H. V. Rivera. 2003b. Coffee and conservation III: Reply 
to Philpott and Dietsch. Conservation Biology 17:1847-1849. 

Reeves, L. H., and R. J. Lilieholm. 1993. Reducing financial risk in agroforestry 
planning: a case study in Costa Rica. Agroforestry Systems 21:169-175. 

Renjifo, L. M. 1999. Composition changes in a Subandean avifauna after long-term forest 
fragmentation. Conservation Biology 13:1 124-1 139. 

Rice, R. A. 1999. A place unbecoming: the coffee farm of northern Latin America. 
Geographical Review 89:554-579. 

Rice, R. A. 2001. Noble goals and challenging terrain: organic and fair trade coffee 
movements. Journal of Agricultural and Environmental Ethics 14:39-66. 

Rice, R. A., and J. R. Ward. 1997. Coffee, conservation, and commerce in the Western 
Hemisphere. Smithsonian Migratory Bird Center and Natural Resources Defense 
Council, Washington, D.C. 

Roberts, D. L., R. J. Cooper, and L. J. Petit. 2000a. Flock characteristics of ant-following 
birds in premontane moist forest and coffee agroecosystems. Ecological 
applications 10:1414-1425. 

Roberts, D. L., R. J. Cooper, and L. J. Petit. 2000b. Use of premontane moist forest and 
shade coffee agroecosystems by army ants in Western Panama. Conservation 
Biology 14:192-199. ' 

Rogers, T. 2003. Starbucks to help Costa Rica weather world coffee crisis. The Tico 
Times, San Jose, Costa Rica, September 19. 

Rojas, J. R. 1988. Algunas observaciones sobre la caficultura de Coto Bras. Page 3. 
Noticiero del Cafe. 

Rojas, L., C. Godoy, P. Hanson, C. Kleinn, and L. Hilje. 2001. Hopper (Homoptera: 

Auchenorrhyncha) diversity in shaded coffee systems of Turrialba, Costa Rica. 
Agroforestry Systems 53:171-177. 

Rojas, M. A. 2000. Pais ante dura crisis. Al Dia, San Jose, Costa Rica, October 2. 

Rosemeyer, M., N. Viane, H. Swartz, and J. Kettler. 2000. The effect of slash/mulch and 
alleycropping bean production systems on soil microbiota in the tropics. Applied 
Soil Ecology 15:49-59. 

199 



Rowe, J. W. F. 1963. The World's Coffee. Her Majesty's Stationery Office, London, UK. 

Sanchez-Azofeifa, G. A. 2001. Deforestation in Costa Rica: A quantitative analysis using 
remote sensing imagery. Biotropica 33:378-384. 

Sandoval, J. A. S. 2002. Coto Bras y el cafe: una breve resena. Boletin Informativo de La 
Regional Coto Brus, ICAFE 2:3-4. 

Sekercioglu, C. H. 2002. Forest fragmentation hits insectivorous birds hard. Directions in 
Science 1:62-64. 

Sekercioglu, C. H., P. R. Ehrlich, G. C. Daily, D. Aygen, D. Goehring, and R. F. Sandi. 
2002. Disappearance of insectivorous birds from tropical forest fragments. 
Proceedings of the National Academy of Sciences 99:263-267. 

Sendieu, D. O. 1996. Impacts de la production du cafe sur lenvironnement en Cote 

d'lvoire in I. C. Organization, editor. Seminar on Coffee and the Environment. 

SEPSA 2002. Boletin Estadistico Numero 13. Secretaria Ejecutiva de Planification 
Sectorial Agropecuaria, San Jose, Costa Rica. 

Sherry, T. W. 2000. Shade coffee: a good brew even in small doses. The Auk 117:563- 
568. 

Siebert, S. F. 2002. From shade- to sun-grown perennial crops in Sulawesi, Indonesia: 
implications for biodiversity conservation and soil fertility. Biodiversity and 
Conservation 11:1889-1902. 

Smith, E. P., and G. v. Belle. 1984. Nonparametric estimation of species richness. 
Biometrics 40:119-129. 

Sokal, R. R., and F. J. Rohlf 1995. Biometry. Freeman, New York, NY. 

Solorzano, J. D. T., and F. M. Fonseca. 2005. Costa Rica: una mapa de carencias criticas 
para el ano 2000. Instituto Nacional de Estadistica y Censos, San Jose, Costa 
Rica. 

http://www.inec.go.cr/06Publicaciones/07SerieCensal/ll%20Costa%20Rica%20 
Un%20mapa%20de%20carencias%20criticas%20para%20el%20ano%202000/p7 
%20mendez.pdf. Accessed March 22, 2005. 

Soto-Pinto, L., I. Perfecto, J. Castillo-Hernandez, and J. Caballero-Nieto. 2000. Shade 
effect on coffee production at the northern Tzeltal zone of the state of Chiapas, 
Mexico. Agriculture, Ecosystems and Environment 80:61-69. 

Stiles, F. G., G. P. Hewson, and L. D. Gomez 1996. Checklist of the birds of Las Cruces 

Biological Station. Organization for Tropical Studies, San Jose, Costa Rica. 
200 



Stiles, G., and A. F. Skutch 1989. A Guide to the Birds of Costa Rica. Cornell University 
Press, Ithaca, New York. 

Stone, S. 1976. La dinastia de los conquistadores. Editorial Universitaria 
Centroamericana, San Jose, Costa Rica. 

Stouffer, P. C. 1998. Survival of the ant followers. Natural History 107:40. 

Stouffer, P. C, and R. O. J. Bierregaard. 1995. Use of Amazonian forest fragments by 
understory insectivorous birds. Ecology 76:2429-2445. 

Stratford, J. A., and P. C. Stouffer. 1999. Local extinctions of terrestrial insectivorous 
birds in a fragmented landscape near Manaus, Brazil. Conservation Biology 
13:1416-1423. 

Thiollay, J. M. 1995. The role of traditional agrofo rests in the conservation of rain and 
forest bird diversity in Sumatra. Conservation Biology 9:335-353. 

Thompson, J., N. Brokaw, J. K. Zimmerman, R. B. Waide, E. M. E. Ill, D. J. Lodge, C. 
M. Taylor, D. Garcia-Montiel, and M. Fluet. 2002. Land use history, 
environment, and tree composition in a tropical forest. Ecological Applications 
12:1344-1363. 

TransFairUSA. 2004a. Backgrounder: Fair Trade Certified™ coffee. TransFairUSA, 
Oakland, CA. http://www.transfairusa.org/pdfs/backgrounder_coffee.pdf. 
Accessed December 4, 2005. 

TransFairUSA. 2004b. Fast facts: Fair Trade Certified™ coffee. TransFairUSA, Oakland, 
CA. http://www.transfairusa.org/pdfs/fastfacts_coffee.pdf. Accessed December 4, 
2005. 

TransFairUSA. 2005. Fair trade certification. TransFairUSA, Oakland, CA. 

http://www.transfairusa.org/pdfs/certification_overview(4.05).pdf. Accessed 
December 4, 2005. 

U.S. Agency for International Development. 2003. USAID's response to the global coffee 
crisis (fact sheet). U.S. Agency for International Development, Washington, D.C. 

U.S. Department of Labor. 2004. Consumer Price Index 1913-2004. U.S. Department of 
Labor Bureau of Labor Statistics, ftp://ftp.bls.gov. Accessed September 17, 2004. 

U.S. Naval Observatory. 2005. Complete sun and moon data for one day. U.S. Naval 
Observatory, Astronomical Applications Department, Washington, D.C. 
http://aa.usno.navy.mil/data/docs/RS_OneDay.html. Accessed October 30, 2005. 

Ulate, A. A. 1997. Fe en el cafe. Pages 16-17. Rumbo. 

201 



Ulett, L. 2001. Cafes finos solucionarfan crisis. La Republica, San Jose, Costa Rica, July 
4. 

UNFAO 1995. Production Yearbook. UN Food and Agriculture Organization, Rome, 
Italy. 

UNFAO 1999. State of the world's forests 1999. United Nations Food and Agriculture 
Organization, Rome, Italy. 

UNFAO 2001. State of the World's Forests 2001. United Nations Food and Agriculture 
Organization, Forestry Department, Rome, Italy. 

UNFAO. 2005. FAOSTAT Agriculture Data. United Nations Food and Agriculture 
Organization, http://apps.fao.org/default.jsp. Accessed October 31, 2005. 

USAID. 2003. USAID fact sheet: USAID's response to the global coffee crisis. U.S. 
Agency for International Development, Washington, D.C. 

USAID. 2004. USAID fact sheet: USAID supports coffee growers around the globe. U.S. 
Agency for International Development, Washington, D.C. 

Valverde, L. 2002. Crisis atasco desarrollo en zonas cafetaleras. La Republica, San Jose, 
Costa Rica, August 5. 

Vandermeer, J., and I. Perfecto. 1997. The agroecosystem: a need for the conservation 
biologist's lens. Conservation Biology 11:1-3. 

Varangis, P., P. Siegel, D. Giovannucci, and B. Lewin. 2003. Dealing with the coffee 

crisis in Central America: impacts and strategies. Policy research working paper 
2993. World Bank, Washington, D.C. 

Vargas, R. R. 1993. Experimentos en cafe para enfrentar la crisis. La Republica, San 
Jose, Costa Rica, June 19. 

Vega M., L. 1992. Medidas para conjurar crisis cafetalera no deben demorarse. La 
Republica, San Jose, Costa Rica, February 11. 

Viquez, A. S. 1992a. Crisis amenaza a los pequenos cafetaleros. La Republica, San Jose, 
Costa Rica, July 2. 

Viquez, A. S. 1992b. Se agrava crisis en Coto Brus. La Republica, San Jose, Costa Rica, 
July 24. 

VNKT. 1987. Jaarverslag. 



202 



Vorst, C. R. 1986. Desarrollo economico, estructura y contradicciones sociales an la 
production de cafe. Revista de Historia 14:179-195. 

Wa-chong, T. G. 2001. Cafetaleros sin dinero para pagarle a recolectores. La Prensa 
Libre, San Jose, Costa Rica, October 30. 

Wasserman, M. 2002. Trouble in coffee lands. Federal Reserve Bank of Boston Regional 
Review 12:4-13. 

Wattel, C. J., M. v. d. Boogaard, R. Ruben, and J. P. d. Groot 1991. Preferentiele 
Handelsvoorwaarden voor Kleine Koffieboeren, San Jose, Costa 
Rica/Amsterdam, Netherlands. 

Wild, A. 2004. Coffee: a dark history. W.W. Norton & Company, New York, NY. 

Williams, R. G. 1994. States and social evolution: coffee and the rise of national 

governments in Central America. University of North Carolina Press, Chapel Hill, 

NC. 

Wilson, S. 2001. Coca invades Colombia's coffee fields: falling prices push farmers to 
plant illegal crops, threatening U.S. drug war. The Washington Post, October 30. 

Winson, A. 1989. Coffee and Democracy in Modern Costa Rica. Between the Lines, 
Toronto, Ontario. 

Woodroffe, R., and J. R. Ginsberg. 1998. Edge effects and the extinction of populations 
inside protected areas. Science 280:2126-2128. 

World Bank Group. 2003. Costa Rica at a Glance. The World Bank Group. 

World Bank Group. 2005a. Costa Rica at a glance. World Bank Group. 

http://devdata.worldbank.org/AAG/cri_aag.pdf. Accessed November 18, 2005. 

World Bank Group. 2005b. World Development Indicators Database. The World Bank 
Group, http://www.worldbank.org. Accessed November 18, 2005. 

Young, C. M. 2003. Coffee agroforestry systems for conservation and economic 
development: A case study of the AMISCONDE initiative in a buffer zone 
community of Costa Rica. Journal of Sustainable Forestry 6:39-63. 

Zanette, L., P. Doyle, and S. M. Tremont. 2000. Food shortage in small fragments: 
evidence from an area-sensitive passerine. Ecology 81:1654-1666. 

Zeledon, V. F. 1987. Campesinos luchan contra selva y lodo en Coto Bras. La Republica, 
San Jose, Costa Rica, August 17. 

203