Socioeconomic Analysis of Agroforestry Systems and Livelihoods in Small Island Developing States

Material Information

Socioeconomic Analysis of Agroforestry Systems and Livelihoods in Small Island Developing States A Case Study from Pohnpei, Federated States of Micronesia
Drew, Walter
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (128 p.)

Thesis/Dissertation Information

Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Forest Resources and Conservation
Committee Chair:
Alavalapati, Janaki R.
Committee Co-Chair:
Nair, Ramachandr P.
Committee Members:
Popenoe, Hugh L.
Hildebrand, Peter E.
Ewel, Katherine C.
Graduation Date:


Subjects / Keywords:
Agricultural land ( jstor )
Agriculture ( jstor )
Agroforestry ( jstor )
Cash income ( jstor )
Crop value ( jstor )
Crops ( jstor )
Employment ( jstor )
Farmers ( jstor )
Highlands ( jstor )
Sudden infant death syndrome ( jstor )
Forest Resources and Conservation -- Dissertations, Academic -- UF
agroecosystems, agroforestry, islands, livelihoods, micronesia, socioeconomic
Electronic Thesis or Dissertation
born-digital ( sobekcm )
Forest Resources and Conservation thesis, Ph.D.


The Small Island Developing States (SIDS), comprising more than 50 nations and territories worldwide, are challenged to establish sustainable development strategies consistent with their small size, remoteness, and fragile and limited natural resources. Agriculture is considered instrumental in furthering sustainable development of SIDS. Agroforestry, an age-old agricultural practice dominates many landscapes among SIDS, proving its ecological sustainability; yet little is known (quantitatively) about productivity, value, and use of agroforests or the role agroforestry plays in the livelihoods of SIDS. Such information is important to ensure that decision makers are able to establish sustainable programs for economic development. Research conducted on Pohnpei, Federated States of Micronesia, documented livelihood systems, quantified agricultural activities, and simulated the impact of potential policy scenarios on both agroforestry and livelihood systems. Sixty-four percent of households relied on three or more sources for cash income; the main income sources included formal employment, agriculture, remittances/pensions, and fishing. One-hundred percent of households were engaged in subsistence agriculture, and 88 and 90% of households used crops for cash income and ceremonial purposes, respectively. Homegarden agroforestry practices provided between 24 and 98% of the total value of agricultural products. Agroforests contributed, per household, an average of $4,242 yr-1 and island-wide contributed 18% to Pohnpei's GDP (total annual value of $15.7 million). Limited market access was found to increase dependence on crops for cash income. In contrast formal employment was found to reduce dependence on crops for cash income. Using ethnographic linear programming, the results suggested that establishing a price policy and/or expanding market opportunities would contribute directly to households most reliant on agriculture for cash income but would have only marginal impacts on households reliant on fishing, remittances, and formal employment for cash income. Use of an expanded version of the traditional Policy Analysis Matrix methodology allowed for the internalization of environmental externalities contributing to a more accurate analysis of farming systems. Results showed that even with such internalization of economic values, it would require additional policy action to provide sufficient incentives to persuade farmers to halt expansion into Pohnpei's uplands where ecological damage had been occurring. Desires for greater material wealth are threatening traditional cultures and limited natural resources on Pohnpei and SIDS world-wide. For those SIDS where agroforestry remains an important component of livelihood systems, development strategies should emphasize using agroforestry as a foundation for agricultural development as opposed to introducing new agricultural systems that have yet to prove their ecological or economic sustainability and/or superiority. To achieve development goals for all members of an island's society, it is critical to tailor development strategies to specific livelihood systems. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (Ph.D.)--University of Florida, 2008.
Adviser: Alavalapati, Janaki R.
Co-adviser: Nair, Ramachandr P.
Statement of Responsibility:
by Walter Drew.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Drew, Walter. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
LD1780 2008 ( lcc )


This item has the following downloads:

Full Text




2 2008 Walter Mark Drew


3 To Tracy, Hannah Lee, and John Henry Drew, along with the people of Pohnpei who were gracious enough to particip ate in this research.


4 ACKNOWLEDGMENTS First and forem ost, I would like to acknowledge and thank my wife and dearest friend, Tracy Drew. Simply put, this work would neve r have been completed without her sincere support and motivation. I also want to acknowledge my daughter Hannah Lee; and my son, John Henry, for their patience and for putting up with me while away working on this dissertation. I thank all my parents for the encouragement to keep moving forward and focus on completing this work, one step at a time. I would also like to sincerely thank Dr. Janaki Alavalapti, who continually motivated me to complete this dissertation, served as a wonderful Committee Chair, and who I believe has truly served to support students and their professional development. Without his continued support and patience you would not be read ing this today. I owe great deal to Dr. Katherine Ewel, for it was because of Dr. Ewel that I first began working in Micronesia. Dr. Ewel has been there as a supporting Committee member has al so truly been a mentor from the onset of our working relationship dating back to 1996. Dr s. P.K R. Nair, Hugh Popenoe and Pete Hildebrand, I cannot express enough how grateful I am that they did not give up on me. Their patience and willingness to stick by me over the years is sincer ely appreciated and I re alize not without some struggles on your part too. Acknowledgments would not be complete with out addressing the support of friends and working partners in Pohnpei. Mr. Adelino Lo rens and Mr. Jackson Phillips were both instrumental in establishing the research de sign and providing the nece ssary contacts that ultimately enabled me to work with so many wonderful people on Pohnpei. Likewise, Mr. Bill Raynor was critical to the success of this projec t, providing invaluable insight into Pohnpeaen culture as well as providing extremely us eful input all the way to the end.


5 Institutionally, this project would not have been possible without financial support from the University of Floridas Alumni Fellowship Program and the Institute for Pacific Islands Forestry. To both I am very grateful. I also than k the USDA Forest Services Institute of Pacific Islands Forestry that provided financial support. I would also lik e to acknowledge Father Greg Muckenhaupt Director of the Ponape Agricult ure and Trade School (PATS) who was generous enough to provide us with room and board while living on Pohnpei. We were very fortunate to have been able to live and work at PATs during our time on Pohnpei and as a result our experience was enriched significantly. Finally, I would like to acknowle dge all of the kind people of Pohnpei who truly provided a unique and incredible learning opportunity for me and my family. Above all, it is the experience I value most in life and an experience it has been. It is my si ncere hope the results of this work will some how make a positive contri bution to you all and possibly others. For those that I have forgotten, of which I am sure there are plenty please forgive me.


6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.......................................................................................................................10 ABSTRACT...................................................................................................................................11 CHAP TER 1 INTRODUCTION..................................................................................................................13 Small Island Developing States.............................................................................................. 13 Agroforestry as a Component of Agricu lture in Sm all Island Developing States.................. 15 Study Objectives.....................................................................................................................18 Presentation of Dissertation....................................................................................................20 2 RESEARCH SETTING..........................................................................................................22 Geographic and Biophysical Setting...................................................................................... 22 Political Setting.......................................................................................................................23 Socioeconomic and Demographic Setting..............................................................................24 Agricultural Setting................................................................................................................26 Agroforestry on Pohnpei.................................................................................................26 Crops and Pohnpeis Agroforests.................................................................................... 28 3 RESEARCH DESIGN AND DATA COLLECTION............................................................ 33 Research Design.....................................................................................................................33 Methods of Data Collection....................................................................................................36 4 PRODUCTION, USE AND VALUE OF POHNPEIS AGROFORESTS............................39 Introduction................................................................................................................... ..........39 Methods of Analysis............................................................................................................ ...42 Regression Model............................................................................................................... ....43 Results.....................................................................................................................................44 Household Characteristics...............................................................................................44 Household Crop Production............................................................................................ 45 Household Crop Use........................................................................................................46 Household Value of Crops..............................................................................................48 Island-wide Value of Crops Produced Within Pohnpeis Agroforests............................ 48 Island Wide Value by Use............................................................................................... 49 Regression Results........................................................................................................... 49


7 5 MODELING LIVELIHOODS ON POHNPE I USING ET HNOGRAPHIC LINEAR PROGRAMMING.................................................................................................................. 65 Introduction................................................................................................................... ..........65 Methods..................................................................................................................................66 Model Specifications.......................................................................................................68 Production Activities....................................................................................................... 69 Model Constraints...........................................................................................................70 Objective Function.......................................................................................................... 71 Livelihood Strategies....................................................................................................... 71 Scenarios Simulated........................................................................................................ 72 Results.....................................................................................................................................73 Model Validation.............................................................................................................73 Scenario 1: Base Case..................................................................................................... 73 Scenario 2: 25% Crop Price Subsidy............................................................................... 74 Scenario 3: Relaxing Cere m onial Use Constraint........................................................... 75 Scenario 4: Expansion of Fish Markets........................................................................... 75 Discussion...............................................................................................................................76 6 DETERMINING AGROFORESTRY PROF ITABILITY USING THE POLICY ANAL YSIS MATRIX............................................................................................................ 85 Introduction................................................................................................................... ..........85 Background and Methodology...............................................................................................87 Description of Three Farming System s Being Analyzed................................................ 87 Policy Analysis Matrix Methodology............................................................................. 88 Data Specifications and Policy Scenarios....................................................................... 90 Results.....................................................................................................................................94 Discussion...............................................................................................................................99 Concluding Comments on the Policy Analysis Matrix Methodology .................................. 101 7 DISSERTATION SUMMARY............................................................................................ 108 APPENDIX A HOUSEHOLD SURVEY.....................................................................................................113 B RECORD FORM.................................................................................................................. 116 C ETHNOGRAPHIC LINEAR PROGRAM MODEL............................................................ 119 D MATHEMATICAL EXPLANATION OF LINEAR PROGRAMMING........................... 121 LIST OF REFERENCES.............................................................................................................122 BIOGRAPHICAL SKETCH.......................................................................................................128


8 LIST OF TABLES Table page 3-1 Employment status of households participati ng in the survey in the two municipalities of FSM ........................................................................................................38 4-1 Role of homegarden agroforests in h ousehold production of crops in Pohnpei, FSM ...... 61 4-2 Average allocation of househol d production of crops for various uses (cerem onial, cash income, friends and household consumption) in Pohnpei, FSM............................... 61 4-3 Use of crops for ceremonial, cash in com e, friends, and household consumption on Pohnpei, FSM....................................................................................................................61 4-4 Average household value of crops produced in P ohnpeis agroforests, 2002................... 62 4-5 Value of crop used per household from Pohnpeis agroforests......................................... 62 4-6 Island-wide values for crops produced in Pohnpeis agroforests...................................... 62 4-7 Island wide-value of agri cultural crops produced and used for different purposes on Pohnpei, FSM ....................................................................................................................63 4-8 Correlation matrix of explanatory vari ab les used in the model for determining percentage of total income derived from agroforests on Pohnpei, FSM...........................63 4-9 Results of the complete regression model pred icting agriculture as a percentage of total income................................................................................................................... .....63 4-10 Results of the final regression model pred ic ting agriculture as a percentage of total income................................................................................................................................64 4-11 2 X 2 contingency table showing per centage of total household incom e from agriculture for significant explanatory variables............................................................... 64 5-1 Percentage of households per livelihood system involved in various activities on Pohnpei, FSM ....................................................................................................................81 5-2 Impacts of a 25% increase in crop pr ices per livelihood system on Pohnpei, FSM .......... 81 5-3 Impacts of reducing ceremonial cr op constraints by 50% and assum ing an international market exists for surplus crops per livelihood system on Pohnpei, FSM..... 83 5-4 Impacts of reducing fish consumption constraint by 50% and es tab lishing domestic markets to sell surplus fish per livelihood system on Pohnpei, FSM................................ 84 6-1 Policy Analysis Matrix (PAM) framework...................................................................... 104


9 6-2 Policy Analysis Matrix expande d to incorporate externalities ........................................ 104 6-3 Coefficients derived from the Policy Analysis Matrix used to measure the relative im pact(s) of policies or market failures........................................................................... 104 6-4 Estimates of private and social profitab ility ca lculated using the Policy Analysis Matrix (PAM) methodology, for a one-y ear period over the three dominant agroforestry land-use syst ems in Pohnpei, Federated States of Micronesia.................... 105 6-5 Estimates of five-year net present values based on a 10% discount rate for private and social profitability calculated usin g the Policy Analysis Ma trix (PAM) methodology, for the three dominant agroforestry land-use systems in Pohnpei, FSM.. 106 6.6 Estimates of ten-year net present values based on a 10% discount rate for private and social profitability calculated using the Policy A nalysis Matrix methodology, for the three dominant agroforestry land-use systems in Pohnpei, FSM.................................... 107 6-7 Ratio indicators derived from Polic y Analyses M atrices for three dominant agroforestry land-use systems for a ten-ye ar period discounted at 10%, in Pohnpei, FSM..................................................................................................................................107


10 LIST OF FIGURES Figure page 2-1 Location of Pohnpei, Federa ted States of Micronesia ....................................................... 31 2-2. Municipalities of Pohnpei, Fe derated States of Micronesia ............................................. 32 4-1 Comparison of the percent of GDP from agriculture in 2002........................................... 64 5-1 Percentage of cash income from diffe rent sources per recomm endation domain on Pohnpei, FSM....................................................................................................................82 5-2 Scenario 1-Total annual cash income per livelihood system on Pohnpei, FSM................ 83


11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SOCIOECONOMIC ANALYSIS OF AGROFOR ESTRY AND LIVELI HOODS ON A SMALL ISLAND DEVELOPING STATE: A CASE ST UDY OF POHNPEI, FEDERATED STATES OF MICRONESIA By Walter Mark Drew August 2008 Chair: Janaki R.R. Alavalapati Cochair: P.K. Ramachandran Nair Major: Forest Resources and Conservation The Small Island Developing States (SIDS) comprising more than 50 nations and territories worldwide, are challenged to establish sustainable development strategies consistent with their small size, remoteness, and fragile and limited natural resources. Agriculture is considered instrumental in furt hering sustainable development of SIDS. Agroforestry, an age-old agricultural practice dominates many landscapes among SIDS, proving its ecological sustainability; yet little is known (quantitatively) about pr oductivity, value, and use of agroforests or the role agroforestry plays in the livelihoods of SIDS Such information is important to ensure that decision makers are ab le to establish sustaina ble programs for economic development. Research conducted on Pohnpei, Federated States of Micronesia, documented livelihood systems, quantified agri cultural activities, and simulate d the impact of potential policy scenarios on both agroforestry and livelihood system s. Sixty-four percent of households relied on three or more sources for cash income; the main income sources included formal employment, agriculture, remittances/pensions, and fishing. On e-hundred percent of households were engaged in subsistence agriculture, and 88% and 90% of households used crops for cash income and


12 ceremonial purposes, respectively. Homegarden ag roforestry practices provided between 24% and 98% of the total value of agricultural pr oducts. Agroforests contributed, per household, an average of $4,242 yr-1 and island-wide contributed 18% to Pohnpeis GDP (total annual value of $15.7 million). Limited market access was found to increase dependence on crops for cash income. In contrast formal employment wa s found to reduce dependence on crops for cash income. Using ethnographic linear programming, th e results suggested that establishing a price policy and/or expanding market opportunities woul d contribute directly to households most reliant on agriculture for cash income but w ould have only marginal impacts on households reliant on fishing, remittances, and formal employment for cash income. Use of an expanded version of the traditi onal Policy Analysis Matrix methodology allowed for the internalization of environmental externalities contributing to a more accurate analysis of farming systems. The results showed that even with such internalization of economic values, it would require additional policy action to provide sufficient incentives to persuade farmers to halt expansion into Pohnpeis uplands where ecologi cal damage had been occurring. Desires for greater material wealth are threatening trad itional cultures and limited natural resources on Pohnpei and SIDS world-wide. For those SIDS where agroforestry remains an important component of livelihood systems, development st rategies should emphasize using agroforestry as a foundation for agricultural development as oppos ed to introducing new agricultural systems that have yet to prove their eco logical or economic sustainability and/or superiority. To achieve development goals for all members of an island s society, it is critical to tailor development strategies to specific livelihood systems.


13 CHAPTER 1 INTRODUCTION Small Island Developing States More than 50 countries and territories of the world com e under the category of Small Island Developing States (SIDS), the majority being in Latin Am erica/Caribbean (23) and AsiaPacific (22). Although world-wide SIDS encompasses an area of roughly one-sixth of the planet (Food and Agriculture Organization [FAO], 1999) as a group, they have received relatively little attention within the development community un til very recently. At the Earth Summit convened in 1992, the FAO first sponsored the Inter-Regional Conference of Sm all Island Countries on Sustainable Development and Environment in Agriculture, Forestry and Fisheries held in Barbados. The main outcome of this conference was the Barbados Declaration which marked the beginning of formal recognition for the need to formulate an integrated approach to the development of SIDS world-wide (FAO 2004). With subsequent global meetings held in Ba rbados in 1994 (resulting in the Barbados Plan of Action for Small Island Developing States) an d more recently regiona l meetings convened in the Pacific and Caribbean (1996), and finally in Mauritius in 2005 (re sulting in Mauritius Declaration for Small Island Developing States) gl obal awareness of the special needs associated with SIDS has arrived (United Nations 2005). Typically defined as having a population of less than 1.5 million (United Nations, 2003) populations of SIDS range from as low as 10,000 (T uvalu) to upwards of more than four million if including larger island nati ons such as Papua New Guinea (United Nations, 2003). In area, SIDS varies tremendously from Nauru consisting of only 13 km2 to Papua New Guinea covering an area of almost 500,000 km2. Demographically, SIDS are also diverse with many having


14 annual population growth rates above 2.5%, not ably, the Solomon Islands, Vanuatu, Palau, Papua New Guinea in the Pacific region (United Nations, 2003). Recent indicators of development of SIDS illustra te their diversity and that the majority of SIDS are within the extreme groups-either the po orer groups having a large agricultural sector, or the wealthier group having limited agricultu re: between 7%-58% of SIDS populations are involved in commercial agriculture with 67% of the to tal population of SIDS having incomes of less than $4,000 of GDP/cap calculated based on the purchase power parity (FAO, 2004). While diverse in population and geographic scal e, collectively, SIDS share characteristics that differentiate them from land-locked nations and that require special consideration for their development. Economically, SIDS have narrow and limited resource base that when combined with small domestic markets often prohibits re alization of economies of scale. Due to the remoteness of many SIDS, they are frequently is olated from major markets leading to higher transaction costs associated with the importation of productio n inputs and/or export of goods produced locally, both of which impact potential profits. During the 1990s, these characteristics, combined with societal changes have contribut ed to SIDS, as a group, transitioning from being net exporters of agricultural products to net importers (FAO, 2005). Additionally, given their smallness and increasing dependence on imported goods, SIDS are increasingly subject to world market fluctuations increasing their vulnerabi lity to outside forces. Size and remoteness constraining access to markets also limits demand for certain goods and services which in turn limits local employment opportunities. Ecologically, many SIDS are lo cated within the tropics a nd are vulnerable to natural disasters such as typhoons, and sea level ri se resulting from global climate change (Intergovernmental Panel on Climate Change [IPCC], 2001). Given the drive for economic


15 development, natural resources and the ecosyst em goods and services they provide are under constant threat of over-exploitation or out right destruction. In the face of these challenges, SIDS must deal with increasing populations leading to further demands on already limited and fragile resources. The challenge for many SIDS, therefore, is to ensure that these resources ar e developed in a manner that meets the needs of current and future generations while minimizing de trimental effects to their natural resources and the health of an islands population. Agroforestry as a Component of Agriculture in Small Island Developing States Agriculture rem ains the single largest sect or of many island states, often contributing substantially to local employment, income, food security, an d a significant portion of foreign exchange earnings. As a result, agriculture is an essential component of overall economic development strategies for many of these island states (FAO 1994; United Nations Development Programme [UNDP], 1996). For SIDS, developing their agricult ural sectors will entail the introduction of entirely new technologies, im proving upon existing agricultural practices, or some combination of integrating new technol ogies into current practices. Additionally, government interventions thr ough the establishment of public policies, infrastructure improvement or technical assistance are necessa ry to promote change. In many areas limited human and financial resources ha ve hindered the ability of polic y makers from having at their disposal adequate knowledge to guide sound policy development. Nonetheless, before programs to develop agricultural sectors ar e established, decision makers need to be informed of the value and contribution of existing ag ricultural systems and what role agriculture plays in the livelihoods of island populations. In the absence of information to develop acceptable alternatives, and without assura nces that adopting incremental or new technological change will benefit farmers, farming households are likely to resist change. Moreover, in many cases,


16 technologies that are proven successful resulting from farm-trials may not actually be desirable by the farming household due to the lack of farm er participation and involvement in developing such technologies (Adesina and Coulibaly, 1998; Versteeg et al ., 1998). In such circumstances, not only are scarce funds allocated for agricult ural development squandered, often what trust exists between farmers and those prom oting change is eroded further. Agroforestry is an integrated land-use practice incorporating agricultural and tree components. Agroforestry is also an age-old practice in many countries world wide. During the last three decades, increasing research and development in terest has been focused on agroforestry as a land-use option particularly in the tropics. The result of these efforts has drawn attention to the benefits of agroforestry includ ing the production of multiple agricultural and tree crops; enhanced organic matter producti on and nutrient cycling rates (Rao et al ., 1998); more efficient utilization of water and light (Black and Ong, 2000); enhanced crop productivity (Kid and Pimentel, 1992; Nair, 1993); better soil conservation (Pimentel and Wightman 1999; Khamzina et al ., 2006); greater biodiversity conserva tion (Nair, 1993; Kumar and Nair, 2004); accelerated rates of carbon sequestration (Wang and Feng, 1995; Oelbermann et al ., 2004) and greater sustainability (Sanchez, 1995). Increasin gly, there is also a growing consensus that agricultural systems that more closely mimic na tural ecosystems are more ecologically viable, providing for sustainable output of goods and services (Ewel, 1999; Lefroy et al ., 1999). While there have been significant strides in understanding the biophysical benefits of agroforestry, efforts towards understanding the soci oeconomic benefits and role of public policy have been relatively small (Mercer and Mill er, 1997; Nair, 1998; Sanchez, 1999; Franzel et al ., 2001; Montambault and Alavalapati, 2005). More recently, there has been increasing awareness of the need to address socioeconomic aspects of agroforestry and the role of policy in order to


17 identify and promote agricultural technologies that are both econom ically viable and ecologically sound (Sullivan et al ., 1991; Current et al ., 1995; Sanchez, 1995; Nair, 1998; Alavalapti et al ., 2004). Agroforestry is now recognized as being integral to meeting Millennium Development Goals (Garrity, 2004). Agroforestry has been practiced in the Pacifi c for millennia, contributing, at one time to Pacific Islanders being among the most self-su fficient and well-nourished people in the world (Clarke and Thamen, 1993). Indi genous Pacific island agrofore stry includes a tremendous diversity of food production systems, most of which have evolved in response to ecological, cultural, and socioeconomic changes through time (Thaman, 1975; Manner, 1981; Raynor, 1989; Raynor and Fownes, 1991a,b; Clarke and Tham an, 1993; Nair, 1993). Many of these systems have proven their stability, sustainability and s ubstantial contribution to local food security (Thaman, 1975; Manner, 1981; Raynor, 1989; Raynor and Fownes, 1991a,b; Clarke and Thaman, 1993; Drew et al ., 2005). Not fully understanding the ecol ogical and economic value of agroforestry in Pacific island settings has resulted in significant environmen tal damage and concomitant economic costs. In countries such as Tonga and the Cook Islands the promotion and adoption of intensive cash cropping systems comprising of melon ( Cucumis spp.) and pineapple ( Ananas comosus ) production, have resulted in severe environmen tal degradation and lo sses of once productive natural resources (Storey and Murray, 2001). In Fiji, the drive for increased agricultural production through the establishment of monoculture plantations has led to extensive conversion of mangrove forests. Many of the systems that have replaced traditional agroforests are dependent on high energy inputs and mechaniza tion that have proven to be uneconomical and biologically unsustainable (Asian Development Bank [ADB], 1992).


18 The lack of appreciation for existing agroforest ry practices in the Paci fic can be attributed to a lack of understanding about the important role that agroforestry practices currently play in food production, local economies, and ecological bala nce, or the potential of these systems for future agricultural development. Due to increas ing population pressures, the drive for economic development and the detrimental impacts associated with climate change and sea level rise, there is an increasingly urgent need to determine the economic sign ificance of current traditional agricultural technologies that ha ve proven their ecological sust ainability to prevent such occurrences from repeating themselves in othe r Pacific island settings. Moreover, understanding how agriculture fits within livelihood systems can increase the probability that agricultural development programs will achieve their desired objectives. Study Objectives The over all goal of this study is to obtain an understanding of the role and contribution of trad itional agriculture to livelihoods residing on a Small Island Developing State located in the Western Pacific. A secondary goa l of this study is to examine how potential policy scenarios influence cash income, use, and value of traditional crops for different segments of an islands population. This study was guided by several research questions: (1) what is the economic value and contribution to household income derived from tr aditional agriculture? (2) How are traditional agricultural products used and what is the valu e per use? (3) How does market access and formal employment affect crop producti on, use and contribution to h ousehold income? (4) How does total household income based on an array of production activities differ for different segments (livelihood systems) of an island s population? (5) What are the potential impacts of policies on farmer profits and how are these profits influenced by incorporating values attributable to ecosystem services (soil stabi lization and carbon sequestration)?


19 To achieve the above goals and answer gui ding research questions, several specific objectives were established for this study, namely: 1. To quantify the productivity, use and valu e of traditional agricultural systems. 2. To determine significant predictors of the percentage of total household income derived from agriculture. 3. To quantify total household income for differe nt segments of an islands population and model how those incomes would vary de pending on pricing policies and market expansion scenarios. 4. To quantify private and social profits for ag roforests allowing for the impacts of potential policies and market failures to be measured. To address Objective 2 specifically, two hypotheses were te sted: (1) that households having greater access to urban mark ets do not derive a greater per centage of total income than households located in more remote locales a nd (2) Households having at least one member formally employed outside the household do not derive a greater percentage of total income from agriculture than households not having any members formally employed. The significance of this study is that it offers evidence of how important traditional agriculture/agroforestr y is to a Small Island Developing State using Pohnpei as a case study. Many Small Island Developing States have chosen to replace traditional agricultural practices with intensive monoculture agricultural systems w ith the intent of filling niche export markets. Yet in many of these same States, the benefits of such changes have been short lived due to unstable market conditions, rapid deterioration of natu ral resources and/or natural disasters. Indeed, many of these same States (Tonga, for ex ample) are attempting to reintroduce traditional agroforestry systems to address food security and a deterioration of limited natural resources (Wolff, 2001). Understanding the economic importanc e of traditional agriculture is necessary to justify policy decisions that may affect inco me opportunities and live lihoods more broadly. Understanding how traditional crops are used is im portant to more fully appreciate the multiple


20 ways in which agriculture is intertwined in island life and the potential tradeoffs between maintaining, modifying or replacing existing pro duction systems. This research answers these questions. Additionally this research expands on th e traditional Policy Analysis Matrix to include values associated with ecosystem services and thereby provides the means to more fully measure benefits and costs of agrofore stry systems. Apart from cont ributing to the body of knowledge specific to Small Islands Developing States, this study contributes to the broader body of knowledge associated with the economic value a nd biological productivity of agroforests and specifically homegarden agroforestry systems of which is currently limited (Kumar and Nair 2004). It is hoped that the results of this study will provide relevant information to help guide future development and management of Pohnpeis limited natural resources. Using Pohnpei as a case study, the results of this study may have useful application for other SIDs in other regions of the world seeking alternative livelihood strategies in support of economic development. Presentation of Dissertation This dissertation analyzes livelihoods and the role of agriculture on Pohnpei. W hat follows next is a chapter describing Pohnpei, the se tting, followed by Chapter 3 which describes the research design and data collection methods employed for the research conducted. In Chapter 4, the analysis centers on the productivity, value and use of agriculture. The analysis first provides a descriptive statistics of hous eholds and agricultural product ion activities. Chapter 4 also examines the variation among households and agri cultural activities wi th a specific focus on whether market access and wealth are influencin g which crops and how much is being produced. Chapter 5 broadens the focus from agriculture to livelihood systems in order to better understand the role of agriculture within the broader context of househol ds on Pohnpei using ethnographic linear programming. Chapter 6 examines the role of policy and market failures on agricultural


21 profitability using the Policy An alysis Matrix (PAM) framework. A second objective of Chapter 6 is to illustrate the potential of the PA M methodology to analyze multi-cropping systems and expand upon the tradition PAM model to incorporate non-market values to more fully account for the true costs and benefits of agroforestry systems. This dissertation closes with a brief concluding chapter reflecting on th e findings, approaches used and their relevance to other Small Island Developing States. Additi onally, the final chapter includes some comments regarding future research needs.


22 CHAPTER 2 RESEARCH SETTING Geographic and Biophysical Setting Located abo ut three-quarters of the way from Hawaii to Indonesi a, the Federated States of Micronesia is made up of more than 600 islands scattered across over 2.5 million square kilometers of water in the Pacific (Figure 2-1). Though the FSM claims an area of ocean twofifths the size of the U.S. mainland, the is lands themselves combined make up only 450 km2 of land-about a quarter of the size of Rhode Island. Pohnpei is one of four states comprising of the Federated States of Micronesia. At 60 54N latitude and 1580 14E longitude, Pohnpei is the second largest island in th e Caroline Islands group w ith a land-base of 216 km2. It is a high volcanic island having a steep m ountainous interior region, the hi ghest point being 772 m above sea level. Average annual rainfall is approxi mately 4,800 mm evenly distributed over 300 days of the year (National Oceanic Atmos pheric Administration [NOAA], 1987). In 1995, 15% of the island was covered by intact upland forests, with an additional 15% consisting of disturbed upland forests (ADB 199 6). Traditional agricultural systems known on Pohnpei as agroforests covered 37% of the isla nd, with coastal forest s (16%) and secondary forests (12%) comprising much of the remainde r of the island. Extensive mangrove forests and a fringing reef surround much of th e island. In the lowlands, highly weathered Oxisols are the most common soils whereas Inceptisols are mo re commonly found in the interior and upland regions of the island (Laird, 1987). Due to the relative age and isolation of Pohnpei, the upla nd forests are very rich in biodiversity with high levels of flora and fauna only found on Pohnpe i. Thirty-four percent of the 767 plant species that have been recorded on Pohnpei are native to the islands upland forests (Merlin et al ., 1992). One hundred and eleven plant specie s are endemic to the island and 90% of


23 these are found in upland forests. In addition, 16% of the islands 50 bird species are endemic to the island (Merlin et al ., 1992). Pohnpei remains one of the most biologically diverse of any island in Micronesia (Merlin and Raynor, 2005). Political Setting Pohnpei is divided into five m unicipalities, with the islands main urban area of Kolonia located within the municipality of Nett (Figure 2-2). Pohnpei is also home to the capital of the Federated States of Micronesia, Palikir. The formal government structure resembles that of the United States at the national le vel consisting of judicial, legislative and executive branches. Like the U.S. system, most major govern mental functions other than the conduct of foreign affairs and defense are carried out by the State government s. The State Government under their Constitution is structurally similar to that of the U.S. util izing three branches as well. For each of the five municipalities, an elected Mayor governs. Parallel to the formal government of Pohnpei is a traditional political entity ruled by district chiefdoms. Five traditional chiefdoms exist, one within each of the municipalities of Madolenhimw, U, Kitti, Nett, and Sokehs. Within the scheme of the district chiefdoms, the kousapw or village is the primary socio-political unit comprising of a colle ction of farmsteads and ruled by a local chief or soumas (Dahl and Raynor, 1996). Each of the five kingdoms is governed by a Nanmwarki and Nahnken, the former serving as the paramount chief and the latter serving as chief advisor and liaison between the paramount chief and the community. Raynor (1991) describes the traditional chiefdoms in the following manner: Both cheifships are hereditary, with succession based on matrilinity. These leaders exert administrative, judicial, and legisla tive powers by the authority of tradition, and through the control of a stoc k of titles. The lower peopl e in each municipality are hierarchically organized by thes e ranked statuses or titles into a stratified society, although there are no distinct classes. The chiefs also function as the center of the prestige economy, based on complex rituals of redistribution of agricultural products


24 Raynor (1991) goes on to note: The chiefs, through redistribution of agricu ltural tributes, conti nuously encourage the cycle of reciprocity. Subjects who continuously accumulate and increase their tributes are recognized by the chiefs through the bestowi ng of a higher title, thus encouraging continued participation in the traditional system among all subj ects. Thus, the agroforestry system is intimately linked with the local cultural system Thus Pohnpei is politically dualistic, comprising of trad itional chiefdoms having a significant influence over a prestige economy and a modern political system structured around a capitalistic economy. Over the last 15 years Pohnpei has rapidly developed as the most westernized state within the FSM. This is pr imarily a function of the national governments being located on Pohnpei, the urban town of Ko lonia and the increasing travel taking place by Pohnpeaens between Pohnpei, the U.S. mainland Hawaii and Guam. As the trend towards westernization continues the influence of traditio nal chiefs is thought to be eroding, particularly related to the production and use of crops for both ceremonial and commercial purposes. Socioeconomic and Demographic Setting From the early days of U.S. control just af ter World War II, rather than investing in Micronesia to develop the local eco nomies of the island states, the U.S. had emphasized political development ensuring close political relationshi p between Micronesia and the U.S. (Peoples, 1985). While the FSM has achieved a form of so vereignty through Free A ssociation with the U.S, its economy has remained almost entirely dependent on U.S. aid. In 1986, the Federated States of Micronesia entered into a 15-year Compact of Free Association with the U.S., which guaranteed paym ents of approximately $1.5 billion in exchange for exclusive military access to the areas wa terways (Osman, 1995). Just under 60% of these funds were spent in support of the day-to da y operations of the natio nal and state governments primarily through the development of a large civil servant workforce (Osman, 1995; United States Department of Interior 2004). The Compact also provided the means for residents of the


25 FSM to travel and work freely within the U. S and its territories cr eating opportunities for employment and educational outside of the FSM to be realized more easily. The first Compact also provided for the renegotiation of certain te rms of the agreement after a fixed period be conducted in support of the establishment of a second Compact agreement (Compact II). The renegotiation for Compact II formally began in 2003. In 2004 Compact II wa s ratified by all four states and approved by the by the United States While much of the components of Compact I were carried over into the Compact II agreement, one of the major changes of the renegotiations was an overall reduction of overseas funds to the FSM. Compact II wi ll provide decreasing payments to the National and State govern ments through 2023. Beginning in 2023, direct payments from the United States will cease and th e FSM will become reliant on the interest from a Development Fund created as another componen t of the Compact II re negotiations. However, for both the National and State governments, thes e payments will decline during the 20 year Compact period with the objective of slowly en couraging the FSM to become self-reliant by developing its economy and securing other sources of funds as the Compact comes to a close. Osman (1989) presents the results of Compact I succinctly: Since the U.S. has delivered a steady flow of economic assistan ce to the FSM over the last four decades, the distribution of the aid a nd its utilization have become government's predominant business. As a result, there is no primary production economy to speak of in the FSM, at least for which reliable data ex ist. A subsistence production economy that was fairly self-sufficient decades ago has been al most completely replaced by one that is now almost entirely dependent on imports. In other words, U.S. funds are the only real source of income. Of all pre-Compact economic activities, agriculture has probably been affected the most. With an already stagnated private sector, ne w employment not only provided opportunities to earn relatively high wages1 it also reduced the reliance on s ubsistence agriculture. Continued 1 Compared with the private sector public sector salaries are between 1.5 and 2 times that of comparable private sector positions (Government of the Federated States of Micronesia, 2002).


26 decline of Compact funds will result in a furt her reduction of public sector employment. One impact of such a reduction will likely be a more concerted effort to develop a more commercially oriented agricultural sector. The population of Pohnpei in 2000 was 34,400 people and is estimated to increase to over 48,600 by 2014 (Office of the Governor, 1996; Gove rnment of the Federated States of Micronesia, 2002). Pohnpei has one of the highest population densitie s in the Pacific, estimated to be just over 160 persons km2 (Government of the Federated States of Micronesia, 2002). The age distribution is heavily skewed to younger ag es with over 50% of the total population under the age of 25 (Government of the Fe derated States of Micronesia, 2002). In 1994, 72% of the working labor force was em ployed in the formal sector (51% in the public sector), and the re maining 28% worked in the informal sector that include agriculture, fishing, and subsistence activities (Office of th e Governor, 1996). During the period between 1994 and 2000, the percentage involved in the formal work force had declined to only 52% with 48% engaged in informal activities coinciding with the final step down in Compact I funding (Government of the Federated States of Micronesi a, 2002). Indicative of the effects of declining Compact funding and a stagnating private sector, in 2000, the median household income was $6,609, down from $8,039 in 1994. Agricultural Setting Agroforestry on Pohnpei Agriculture has been an im portant part of life on Pohnpei for centuries and remains so today. Agroforestry has been practiced for centuries on Pohnpei (Huan, 1984) and was the predominant land use activity at the time of this research (ADB, 1996). Tr aditionally, almost all production, be it for subsistence, local market, or export is by family units using localized agroforestry systems, providing employment, food security, and income while maintaining the


27 cultural and ecological integrity of the island and its population (Raynor, 1991). Over time, numerous crops have been introduced through co ntinued waves of migration, and more recently, through the efforts of colonial administrations (Barrau, 1961; Ragone et al ., 2001). Homegardens are the most common form of agroforestry on Pohnpei. Homegardens incorporate multipurpose trees and shrubs in intimate association with annual and perennial crops, and often domestic animals, adjacent to homes (Fernandes and Nair, 1985; Raynor, 1989; Falanruw, 1993; Kumar and Nair, 2004). Pohnpei, like many other Pacific islands, is at a crossroads whereby traditional norms are changing to more western oriented desires, particular ly desires for increasing material wealth. Whereas only two decades ago, prestige was la rgely defined by village members who provided the largest and rarest variety of traditional crops such as yams to the local chiefdom, today where possible, these crops are being so ld to purchase cars, television s, and CD players. Moreover, with close to 60% of the popul ation under the age of 25 a nd very limited employment opportunities in the formal economy (Government of the Federated States of Micronesia, 2002), Pohnpeaens are searching for altern ative ways to accommodate thei r changing lifestyles. It has yet to be seen how the dynamic situation will ultimately affect Pohnpei. For the present time these changes are leading to increasing pressure to intensify (or in some cases replace) current agricultural practices in ways that may not be suitable to Pohnpei s ecological, social, cultural, or economic setting (Kostka and Raynor, 2000). Given th e long history of agriculture on Pohnpei and the uncertainties about its economic prospects, future developments of Pohnpeis agricultural sector should be based on a more informed unde rstanding of agricultures current role and how future development strategies may influence fa rmer decisions (Pohnpei Office of Agriculture and Forestry, 1996).


28 A stated objective of the na tional and state governments is to promote import substitution policies to reduce dependence on imported food a nd increase household incomes (Office of the Governor, 1996; Government of the Federated States of Mi cronesia, 1999). Yet specific strategies have not been identified, largely due to the lack of baseline data about the production and economic value and use of existing agricu ltural systems for which comparisons can be measured. Crops and Pohnpeis Agroforests Locally produced indigenous food crops cultivated within agroforests dom inate Pohnpeis agricultural sector. Trad itionally, almost all production, be it for subsistence, or commercial purposes, is by family units using localized agro forestry practices, providing employment, food security, and income while maintaining the cult ural and ecological inte grity of the islands resources (Raynor, 1991). Species common to these systems include in the upper main canopy coconuts (Cocus nucifera) and breadfruit (Artocarpus altilis. Ylang-ylang (Cananga odorata) and yams (Dioscorea spp.) are both common in between the main canopy and sub-canopy layers. Within the sub-canopy bananas and plantains ( Musa spp.), hibiscus ( Hibiscus tiliaceus ), Indian Mulberry ( Morinda citrifolia ), yam vines ( Dioscorea spp.), and soursop ( Annona muricata ) dominate planting activities. Below 2.5 meters, sakau (Piper methysticum) is very common in addition to root crops such as wild taro ( Alocasia macrorrhiza ), sweet taro, ( Colocasia esculenta ) and swamp taro ( Cyrtosperma chamissonis ) (Raynor and Fownes, 1991a,b). Although homegarden practices are the most prominent land use on Pohnpei, they are not the only type of agroforests. Other lands located in the lowlands, that may or may not be in close proximity to homegardens, are al so cultivated. Many of these lands are comprised of some or all of the crops produced in homegardens although they are often not as intensively farmed.


29 Additionally, households may cultivate crops in the highlands, or on lands that are owned, provided access to or that are found within th e States jurisdiction a nd are being cultivated illegally. In those lands being cu ltivated in the highlands or under the States jurisdiction, the primary crop being produced is sakau. Howeve r, in some cases households are producing some of the same crops being produced in homegardens and other lowla nd agroforests in highlands as well. Pohnpeaens recognize two major seasons base d primarily on breadfruit and yams, traditionally the two most important staple foods. The breadfruit season most often falls between April and ends in the fall. The breadfruit seas on is recognized as being the lazy time because many of the tree crops used by humans for one use or another receive lit tle care and attention other than just harvesting thei r fruits. In contrast, the yam s eason, coinciding with the ending of the breadfruit season is generally when bananas, taros and yams are harvested requiring more labor inputs. It is the traditional chiefs on Pohnpe i that decide when a crop season begins and ends, signified by large feasts and driven in large part by the natural production of breadfruit (Raynor, 1991). Sakau is grown and harvested year round. Sakau combined with breadfruit and yams are considered the three most im portant crops for ceremonial purposes. There is very limited production of crops more recently introduced for commercial purposes on Pohnpei. Of the more recently introduced cash crops, the most common are cucumbers ( Cucumis sativus ), cabbage ( Brassica spp.), bell peppers ( Capsicum spp.) and to a lesser extent black pepper ( Pepper nigrum ). Most commonly these crops are produced in small quantities in and around homegardens.


30 Within the traditional or prestige economy agricultural crops, yams ( Dioscorea spp.), and sakau2 ( Piper methysticum) in particular, have played an important role in food security, maintenance of the availability of diverse cu ltivars and social status (Raynor, 1991; Barrau, 1961). However since the beginning of Compact funding, there has been increasing levels of imported food goods as Pohnpeaens have become incr easingly accustomed to the convenience of cooking processed foods (dominated by imported rice), and more western appetites. Pohnpei remains a net importer of goods, dominated by importation of foods and beverages. Reliable estimates of the annual economic valu e of all agricultural products produced on Pohnpei are few. Available estimates state that in 1996, the agricultural sector was responsible for approximately 9% of Pohnpeis gross dom estic product while other sectors including wholesale and retail trad e (30%), government services ( 30%) and transportation, storage and communications (9%) made up the majority of the remaining GDP during the same year (Pohnpei Statistical Office, 1998). One of the prim ary objectives of this study was to provide updated and reliable data pertai ning to the production of crops being used for both commercial and subsistence activities within Pohnpeis agricultural sector. 2 Sakau, is also known throughout the Pacific as kava, kava kava, and yogona. Sakau is in the Piperaceae family. Sakau is a mind altering bevera ge and its preparations has been and remains today, imbibed in a wide range of tr opical Pacific societies (Lebot, 1992)


31 Figure 2-1. Location of Pohnpei, Federate d States of Micronesia (Source: s_oceans_poles/micronesia_pol99.jpg)


32 Figure 2-2. Municipalities of Pohnpei, Federated States of Micronesia (Source: pedia/commons/e/eb/Pohnpei_map.gif)


33 CHAPTER 3 RESEARCH DESIGN AND DATA COLLECTION Research Design The author had spent close to six m onths conducting research cente red on resource use on Kosrae, another of the four main island states of the Federated Stat es of Micronesia. As a result, prior to arriving on Pohnpei, the author had established working contacts with personnel from within the National and State governments, Pohnpeis Forestry and Agricultural Extension offices, The Nature Conservancy (TNC) and fr om the Ponape Agricultural & Trade School. Upon arrival on Pohnpei, the author met on severa l occasion with these established contacts to discuss the objectives of the proposed research. Consultations with re sidents on Pohnpei also provided relevant guidance on appropriate strategi es to consider during the administration of a questionnaire and further data collection propos ed over a 12 month period. Additional meetings were held with village members to communicate the purpose of the research and to informally measure the willingness of participants to participate in the research project. Following multiple meetings and discussions with officials, TNC staff, and several community members, as well as informally spendi ng significant time with members of different villages, a draft household questionnaire form was developed and tested (administered) among those who had participated in prior meetings In response to feedback resulting from the administering of the first iteration of the questionn aire, two more versions were tested before a final household survey was considered accepta ble by the author and those who might be participating in the research project. A parallel process was follo wed in order to establish forms needed for the collection of data relative to agricultural and fishing activities. The household questionnaire was created in Eng lish, however the agricultural and fishing activities forms, while


34 initially created in English, was converted to the Pohnpeaen language by TNC and Agricultural Extension staff. Based on many discussions with established contacts, it was determined that having someone who lived and was respected in the municipality of Nett would be necessary to (1) have a local point of contact that se rved to appease any suspicions about the research being conducted, (2) bridge any cultural gap there might have been and (3) provide continued feedback into how the research was progressing and how best to work with and among household members in Nett. As a result, additional funding was pursued and granted by The Nature Co nservancy to hire a part-time assistant who in turn was intimately invol ved with all aspects of data collection in Nett. This was less of a concern for Madolenhimw be cause this was where the author lived and had already established relationships with community members prior to or very early on during the initial field research activities. The household questionnaire was structured in to three main sections (Appendix A). The first section centered on demographics of each household. Demographic information was needed in order to determine household composition, age of participant etc. in order to more accurately characterize participating households. The s econd portion of the quest ionnaire centered on availability and use of resources other than labor, in order to determine the number of lands and area available for household use in addition to in formation relative to farming practices, e.g., use of fertilizers and/or other production inputs. The last section of the questionnaire centered on economic aspects of each household to document em ployment activities, amounts and sources of cash income and household expenses. Data from a ll three sections were used to quantitatively determine dependent and independent variable enabling regression analysis to be conducted. Additionally, data from the quest ionnaire were critical to de termine unique livelihood systems


35 that in turn provided the basis for analysis of such systems using ethnographic linear programming. Approval of the final survey forms by the Internal Review Board at the University of Florida was sought and obtained. To answer whether or not formal employme nt and access to markets were significant predictors of agricultu ral production, value, and income, a stratified random sample design was established with four strata: (1 ) households located close to mark et center (2) households located far from market center (3) house holds having at least one member formally employed and (4) households having no members formally employed. These strata along with other independent variables were used to conduct re gression analysis to test the two hypotheses. To determine if market access was influencing agricultural activities, two municipalities, Nett and Madolenhimw, were chosen for this research1. Nett is located adjacent and very accessible to Kolonia, Pohnpeis (and that of the FSM) onl y urban center and where most of the islands commerce takes place (Figure 2-2) In contrast, Madolenhimw, separated by the municipality of U, is located to the SE of Nett at abou t an hours driving distance from Nett. To determine which households were selected to participate in this study, first of all, within each of the two municipa lities, 19 villages were randomly selected. The next step was to meet with the chiefs from the selected villages to discuss the purpose of the research and to obtain permission to engage household members from their village in the research study. Once permission for a meeting was granted, the village chief provided a list of all the households, delineated by those having at least one member w ithin each of the household earning a full-time 1 Referring to Figure 2-2, it can be seen that th ere are only six municipa lities comprising of Pohnpei with Nett containing Kolonia, the centr al urban area of the island. Both Kitti and Madolenhimw are furthest from Nett. In order to develop as much ra pport with each of the participating households, the fact that the author was living in Madolenhimw and due to limited resources, Nett only one additional municipa lity (Madolenhimw) were chosen as the municipalities to include in this research.


36 wage (hereafter referred to as formally empl oyed) or not. These lists provided the basis to randomly select households, stratified by proxim ity to Pohnpeis market center of Kolonia and employment status (households within both Nett and Madolenhi mw stratified by employment status). Households were then randomly selected and approached to ask if the families were interested and willing to particip ate in the research. Initially there were over 50 participating households; 33 were completed for the study (Table 3-1). Methods of Data Collection Each of the participating households was adm i nistered the questionnaire. After completing the questionnaire, participants were asked again if they would still be in terested and willing to record information relative to their agricultural and fishery activities. Fo r households interested in participating, a weight scale, clip board, pens and record forms, were provided to heads of households. Members from each household recorded the hours spent working on agriculture, what was harvested and how much, and how harv ested crops were used (Appendix B contains the record form). The use categories included cash income, household consumption, ceremonial activities, and for extended family and friends. Cash income was determined based on how much of any given crop was sold either to markets, sold locally at family owned and run road side markets and/or directly to other village member s. Consumption use was defined as the amount of a given crop consumed within th e household and/or used to feed animals. Ceremonial use was considered to be the amount of crops used in bot h traditional celebrations such as seasonal feasts signifying the beginning or end of breadfruit and yam seasons, tribut e to traditional leaders, the granting of traditional titles of prestige, reli gious ceremonies, funerals, weddings and childbirths as well as more contemporary celebrations su ch as school graduations and athletic events. Finally, use for extended family a nd friends was considered to be crops given to family members and friends residing in households other than the participating households themselves.


37 Agriculture production was record ed either by weight or by number depending on how it was transacted in the local markets and those received by farmers themselves. The size of lands used for agriculture was determined by responses of heads of households. In those instances when heads of households were not exactly sure of th e size of lands, estimates were derived based on using known landmarks as proxies e.g., local soccer fields, meeting loca tions, churchs, etc. Fishing activity was also recorded but limited to the number of hours fished, total catch and total sold. To ensure record forms were being filled out correctly weekly visits to each of the households were carried out. Prices used to determine values of agricultural crops and fish harvested were based on the average prices observ ed in local markets. Data were collected over a period of 12 months during 2002-2003. Data collec ted during the field research were then transferred into a database to allow for descriptiv e statistics, regression anal ysis (Chapter 4) to be conducted and linear programming model si mulations (Chapter 5) to be completed. Additionally, data in corporated into the database were us ed to conduct an analysis comparing private and social profitabilitys using the Po licy Analysis Matrix methodology (Chapter 6). Throughout the duration of the fiel d research, there was a continued and conscious effort to spend as much time as possible visiting friends a nd families participating in the research itself with the aim of learning as much as possibl e about the various liv elihoods of Pohnpeaens. Whether spending time helping to prepare for ce remonial activities such as seasonal feasts, attending funerals, church, or other activities, doing so provided a much more robust opportunity to observe first-hand life on Pohnpei. As importan tly, it provided an oppor tunity to build rapport with household members that were involved with this research.


38 Table 3-1. Employment status of households participating in the su rvey in the two municipalities of FSM Municipality Employed No Employment Total Nett 11 4 15 Madolenhimw 7 11 18 Total 18 15 33


39 CHAPTER 4 PRODUCTION, USE AND VALUE OF POHNPEIS AGROFORESTS Introduction As Pohnpei moves forward and attempts to deve lop its agricultural sector, the efficient use of scarce resources will require that proscr ibed programs have the highest probability for success. Understanding current agricultural activities provides a baseline in which to measure against potential change. Moreover, determining th e relative dependence on agriculture can help determine whether there are vari ations in current practices between different segments of a population and in doing so can help sharpen the focus on which types of programs might most effectively achieve success. For over a century, researchers have been interested in determining the potential variation in agricultural systems and the driving forces behind any such variation. For example, Von Thunen (1826) was interested in determining if spatial distances from agricultural markets had any significant influence on the intensity of agri cultural systems. His theory was that intensity was inversely related to distance to market because the greater the distance, the lower the profits resulting from higher transportation costs for farmers living further from markets (Symons, 1978; Visser, 1980). Others ha ve also drawn attention to the poten tial role that in frastructure and market access may play in the adoption of agricultural systems and the use of crops because they broaden the opportunities for the ad option of new technologies and crops leading to changes in agricultural systems themselv es (Allan, 1986; Turkelboom et al ., 1996; Reardon et al ., 2001). In the case of Pohnpei and prior to this study, it had yet to be determined whether the relative proximity to markets has any influence on the ty pes of crops being produ ced. Moreover it had yet to be determined whether market access in fluences the use of crops produced within agroforests are used for different purpos es, and specifically for cash income.


40 Population density has also been identified as a major influencing factor on agricultural systems. Boserup (1965) theorized that as popula tions increased, land scarcity increased leading to gradual intensification and changes in the use of agricultural systems. An example of this trend is the shift from swidden agriculture ha ving long fallow periods to annual cropping systems as population densities and land sc arcity increase. A counter argum ent to Boserup might be that population density-as measured by the number of members per household-may actually result in increased amounts of available labor to fa rm and produce additional crops. Moreover, households having more land-less scar city-provide greater opportunities to utiliz e larger areas for agricultural purposes, on ce again, providing the means to produc e and use a greater quantity of crops. At what point does population density and land scarcity tri gger transformation of agricultural practices is a question that is very li kely site specific given available resources, local cultures and the relative dependence on agricu lture to meet liveli hood objectives. With increasing populations and with more and more families on Pohnpei needing land to farm, it has yet to be determined whether population densitie s and the scarcity of land are influencing agroforestry practices and th e relative dependence on agroforests for their livelihoods. Of particular interest to this research is examin ing whether population dens ity and land scarcity as measured by the number of members per household (density) and the amount of land available to a given household (scarcity) has any statistically significant influence on th e relative dependence on agriculture for household income. Pohnpeis dualistic economy provides a unique opportunity to examine how the process of monetization may be influencing agricultural activities. Thro ughout the pacific (Brookfield, 1972: Clarke and Thamen, 1993; Powell, 1997), and in Micronesia in particular (Barrau, 1961; Bascom, 1963), production of crops for ritual and ceremonial purposes has been noted as an


41 important component of livelihood activities. As a result of pa rticipating in these types of activities, islanders not only are able to rece ive tangible goods through reci procity of agricultural crops, greater social status among communities ma y also be achieved. Yet there are opportunity costs associated with using pr oducts of economic value for one purpose versus another e.g., for prestige versus cash. Indeed, one can speculate that as cash income and material wealth become more highly valued relative to traditional gains, a greater em phasis on pursuing whatever means possible to achieve such wealth will come at the expense of trad itional activities. In addition to the ideas discussed above, anot her topic having relevance to this study deals with economic development itself and the theory of structural transformation. Structural transformation is a process by which the contribu tion of nonagricultural se ctors to an overall economy rises as the agriculture sectors role declines in relative terms (Johnston, 1970). The inference at the household level is that as househ olds become more activel y engaged in activities outside of the agricultural sect or for economic purposes, the rela tive dependence on agriculture declines. Whether at the household or larger economy-wide scale, in those areas that have not become entirely industrialized, agriculture remain s an important contributor to economic activity but on a relative scale. As Pohnpei continues its pursuit of economic de velopment, in what capacity agriculture might contribute to such development is uncerta in. Is Pohnpei in the midst of a structural transformation and if so, where along the proc ess does Pohnpei stand vis a vis the relationship between agriculture and other economic activities and their relati ve contribution to household incomes? Based on the discussed above, the objectives of this chapter is to (1) quantify the production, use and value of agroforests on Pohnpe i at the household leve l and then scaling up


42 the analysis to determine these same values at the island-wide level and (2) to determine whether market access, number of household members, availa ble land area, ceremonial use of crops and formal employment are statistically significant ex planatory variables of the percentage of total household income derived from agriculture. What follows is first the methods used for this chapters analyses followed by the results that include descriptive statistics of households involved in this study, the producti vity and use of agricultural crops, and economic contribution of agriculture at the household and island-wide sc ales. The Results section also presents findings based on regression analysis of explanatory va riables explaining the percentage of total household income from the sale of crops. The Ch apter concludes with a section comprising of discussion of the results and their possible implications fo r the development of Pohnpeis agricultural sector. Methods of Analysis The data pro vided for the descriptive statisti cal analysis were deri ved from those data collected as described in Chapter 3. Descriptive st atistics presented here include the mean values for household demographics, land ownership, cr op production, crop value and crop use. These data provide the basis to determine the value of agroforestry products and their use, based on production levels and market prices observed during the dur ation of the field re search activities. As described in Chapter 3, the research design was a stratified random sample to assess the potential influence of market access and in come on the productivity, use and value of agroforests, using the variations in agri cultural activities between households in the municipalities of Nett and Madolenhimw, and betw een households having at least one member formally employed as the indicators of these parameters to assess. Correlation analysis was conducted to test fo r multicolinearity and to determine whether positive or negative relationships existed between explanatory variables. Additionally,


43 independent Sample T-tests were conducted to assess if significant differences existed between the four strata. Finally a linear regression model was developed in order to determine significant predictors of the contribution of agriculture to household incomes. Backward elimination regression method was used to determine the most significant model based on five explanatory variables. This method begins with fitting the regression model with a ll of the explanatory variables and eliminates variables one at a time until the best-fit is found. Regression Model To assess the influence of the various theori es discussed above and the influence of the prestige econom y on total cash income, the regression model comprised of the following variables. The dependent variable used was the pe rcentage of total income derived from the sale of agricultural crops (PINCAG) This variable helps to expl ain the relative dependence on agriculture for cash inco me. Presumably, if other opportunitie s for income generation such as formal employment change so too would th e relative dependence on other income earning activities e.g., commercial agriculture. The explanatory variables used in this model include: Relative access to market (MKACES). A dumm y variable was assigne d to households in Nett (1) and for households situated in Madolen himw (0). Nett is situated adjacent to Kolonia, Pohnpeis only urban center and where most commerci al activity on the island is located. Therefore, it was expected that b ecause of the advantage of having market(s) nearby, households in Nett would have a greate r opportunity to sell crops and therefore, ceteris paribus those households would derive a greater percentage of income from agriculture. Moreover, in addi tion to having to drive up to an hour to reach Kolonia, households in Madolenhimw have more lim ited information about market demands for crops and as a result, may be more risk aver se, choosing not to go into Kolonia to sell crops. Total number of persons residing in the household (NHHMEM). Given already limited opportunities for formal employment, the in creasing number of people to feed and financially support as well as th e available agricultural labor, it was expected that there would be a positive relationship between NHHMEM and the dependence on agriculture for cash income.


44 Total available land in hectares (TOTLND). Wi th increasing land area available to farm, it would be expected that more crops could be produced and used for cash income purposes. Therefore, a positive relationship between TOTLND and PINCAG would be expected. Percent of crop value for cerem onial activities (PERCER). It was expected there would be an inverse relationship between the percentage of total crop value resulting from selling and ceremonial use of crops. With a greater percentage of crop value being used for ceremonial activities, there would be a declin ing percentage of crop value for cash earning purposes. Households having at least one member formally employed full-time (EMPLOY). A dummy variable was assigned to households having at least one member employed fulltime (1) and for those that do not (0). T hose households having at least one member formally employed will have a more steady a nd reliable source of income and therefore will not necessarily be as dependent on agricu lture to meet cash income needs. It is expected that there would be a negative re lationship between EMPLOY and the percentage of total income from agriculture. The full equation to estimate the percentage of total income derived from agriculture is: PINCAG= + B1 (MKACES) + B 2 (EMPLOY) + B 3 (NHHMEM) + B 4 (TOTLND) + B 5 (PERCER) Results Household Characteristics The average age of head of household was 47 years and the average age of households where hom egardens were located was 29 years. The average household had 10 members, five adults and five children under th e age of 18 years. Forty-six pe rcent of the households had at least one member having graduated from high sc hool. On Pohnpei, there is a community college providing opportunities for higher education. Nonetheless, onl y 21% of households had any member that had at least some university level education. With the signing of Compact I, Micronesians were provided the freedom to travel and work in the U.S. and its territories. As a resu lt, many of the households (64%) had at least one immediate family member residing off-island (an average of 1.6 persons per household), the majority of which were residing in Hawaii a nd the U.S. mainland. Forty-five percent of


45 households were receiving an average of $106 per month in remittances from family members residing outside of Pohnpei. Households on Pohnpei employ diverse strate gies for generating income. Sixty-four percent of all households generated at least some income from three or more different sources, 24% from only two sources and only 12% depend ed on only once income source. In addition to formal employment, some household members are involved in commercial agriculture (88%), fishing (42%) receive remittances (46%) or collect social security (20%) as a means to acquire cash income. Seventy-nine percent of households having at least one member formally employed considered employment as the most important source of income, primarily due to the insecurity of their employment status. Comparatively, for those households not having any formal employment, 59% noted agriculture 12% noted fishing and 18% noted Social Security or other retirement pensions as their most important source of income. Household Crop Production All the households participating in this res earch had m embers that owned at least one parcel of land, with 40% owning only one, 37% owning two and 23% owning three or more. On Pohnpei, land may be granted for use by extende d family members providing additional areas for crop cultivation. Thus, although only 23% of households owned three or more parcels, 48% had access to at least three land parcel s. However, access doesnt necessa rily reflect actual use. Only 21% of households were cultivating three land parcels compared to 64% cultivating two; 100% of the participating households were cultivati ng at least one land parcel, their homegarden. One hundred percent of households were cultivating multiple crops within their homegardens including sakau (94%), yams (97%), breadfruit (94%), swamp taro (85%), coconuts (100%), betel nut (64%) and bananas ( 100%). There were no households that produced crops exclusively on lands other than their homeg ardens. Moreover, with the exception of sakau,


46 the majority of all crops are produced within homeagardens (Table 41). All participating households stated that their agroforests were currently producing enough food crops for household consumption needs. Within the homegardens, bananas are produced in the greatest qua ntity by weight (689 kg yr-1) followed by breadfruit (387 kg yr-1) and swamp taro (196 kg yr-1) being produced in the smallest amounts. However, when crops produced on all lands are combined, sakau is harvested in the greatest quantity (1106 kg yr-1). In terms of absolute number s, betel nut is produced in far greater quantity (but by fewer households), relativ e to coconuts, the only other crop sold by piece as opposed to weight. Households in Nett pr oduced significantly more swamp taro (P<0.1) relative to households in Madolenh imw, most likely due to the prev alence of freshwater wetlands in Nett. There were no significant differences in production of any crops between households having and not having at least one member employed. Household Crop Use Each of the participating households was asked to record how each of the crops were being used, either by weight or number. The use categories were broken down by whether crops were used for cash income, traditional ceremonial pur poses, provided to extended family and friends or used within the household (consumption by either household members or animals). Although 100% of participating households producing some or all crops used them for more than one activity, the percentage of househol ds per activity varied based on th e particular crop in question. For those households that produced a given crop, th e majority of them utilized at least some the total harvest for ceremonial activities ranging fr om 90% of households producing sakau down to only 12% of households producing betel nut (Table 4-2). Eighty five percent of all households sold at least some of the crops harvested.


47 The majority of these households sold at l east some bananas (64%) and sakau (58%). More than 90% of all households producing a ny crop consumed at leas t a portion of the total harvested. As a whole, the number of households that provided crops to friends and extended family members was low. While the percentage of households using crops for a particular use is informative, the percentage of each crop produced by households used for each activity sheds light on relative importance of each activity to households on Pohnpei. The three main crops used for ceremonial purposes by quantity are yam (42%), sakau ( 24%) and breadfruit (18%) respectively. Although 90% of households use sakau for ceremonial purposes only 24% of the amount of sakau produced is used for ceremonial activities (Tab le 4-3). On the other hand, of the 81% of households using yams for ceremonial activities cl ose to half of all yam production by the same households is used exclusively for this purpose. Relatively little of each crop produced by a household is used for cash income, the exception being betel nut (61%) and sakau (39%). In contrast, the majority of crops are consumed within the household ranging from a high of 91% of coconuts, down to only 25% of betel nut. Although a relatively large percentage of households provides each of the crops produced to friends and extended family, the actual quantity is low relative to other uses (Table 4-3). Households in Nett were consuming signi ficantly more swamp taro (P<.05) compared to those in Madolenhimw. Additionally, house holds having one member employed sold significantly more swamp taro (P<. 05) than those that do not. But for these two exceptions, there were no significant differences in crop use between households in Ne tt and Madolenhimw or depending on employment status of the households.


48 Household Value of Crops For hom egarden-produced crops, the economic value varied from a high of $575 yr -1 for sakau and a low of $85 yr -1 for breadfruit per household in 2002 (Table 4-4). Although the economic value is greatest fo r sakau, both yams and betel nuts have household values over $500 yr -1 indicating that there was no dramatic difference between the values of these three crops produced in homegard ens. Considering the economic value of crops produced on all lands, the order of crop values remained the same as crops produced solely in homegardens, yet the value of sakau ($2,434 yr -1) was four-fold more than that of any other crop Homegardens alone provided goods valued at just over $2000 yr -1. When the values for all crops being produced on all la nds used were combined, the av erage value of agriculture per household was just over $4,000 yr -1. In 1999 at the time of the last census, annual median household income on Pohnpei was $6,354. Thus, the es timated total value of agriculture based on the seven crops considered in this research was equivalent to 67% of median household income. Households engaged in selling agricultural crops generated almost $1,600 in cash annually representing 24% of median household in come (Table 4-5). Sakau and betel nut sales alone explain more than 90% of the source of cash income from commercial agriculture. The total average household c onsumption value was very cl ose to average cash income primarily due to the large contri bution of income from selling saka u. With the exception of sakau and betel nut, all crops had a gr eater value associated with ho usehold consumption relative to other uses. Island-wide Value of Crops Produced Within Pohnpeis Agroforests Using household values determ ined in this research and multiplying them by the total number of households on Pohnpe i it becomes possible to calculat e the total economic importance


49 of Pohnpeis agricultural sector. Homegardens alon e are valued at more than $8 million per year with sakau having the greatest total value of all cr ops considered (Table 4-6). When all lands are combined, the total value of agroforests almost doubled to nearly $16 million. Island Wide Value by Use In term s of the total value of crops and their use, sakau hol ds the highest value regardless of how it is used (Table 4-7). On the other ha nd, coconuts used for ceremonial purposes have the lowest value ($614). In terms of the largest valu e per use, household cons umption of agricultural crops on Pohnpei was valued at just over $7 million per year. Based on a population of 34,000, the seven agricultural crops included in this st udy have a consumption value worth $208 per year per person on the island. Second behind consumption value is cash income valued at just under $5 million yr -1; ceremonial use was valued at close to $3 million yr -1 and friends and extended family use at more than $1 million yr -1. Combining all use values, production of the seven crops included in this study had a total annual approximating 8% of the GDP for all of th e Federated States of Micronesia. This value equates to just be low 18% of Pohnpeis GDP during 2001 when this research was conducted. Based on the census con ducted in 2000, the total value of these seven crops on Pohnpei equates to approximately 50% of total annual household cash income. Regression Results Based on the correlation results (Table 4-8) there is not a con cern for multicolin earity given that none of the explanatory variables were strongly correlated with each other (Table 4-8). Results of the initial and final regression models are provided in Tables 4-9 and 4-10 respectively. The results confirm expected pos itive or negative rela tionships between the explanatory variables and the per centage of total income from agriculture. Households having greater market access, more members and more total land area available to them were positively


50 correlated to the percentage of income from agriculture. On the other hand, employment and the use of crops for ceremonial pur pose both negatively affect th e percentage of income from agriculture. The final model consisted of only mark et access and employment as significant explanatory variables (F = 8.285; df =31; P<.001). The number of household members and thus presumably available labor does not significantly influence the tota l proportion of income due to crop sales. Nor does having more or less land availa ble for agriculture or th e percentage of crops for ceremonial purposes significan tly influence relative income levels. Thus of all of the explanatory variables only households having gr eater market access (P<0.05) and those having at least one member employed (P<0.001) are signifi cant predictors of the total percentage of income derived from agriculture on Pohnpei (N= 33). The final model explains 36% of the percentage of total household income that is derived from agriculture. Both of the significant explanatory variables (defined as P<0.05 or P<0.001) in the model are discrete. Because of this it is difficult extr apolating the model to predict percentage of total income from agriculture. Thus the model prove s very useful in determining which of the explanatory variables are most significant, th eir positive or negative relationship to the dependent variable and the rela tive contribution of the explanatory vari ables in explaining the percentage of total income coming from agricultur e. However, the model is not as useful as a means to predict the percentage of total income from agriculture. Given the shortfall of using the model coe fficients, a two by two contingency table was constructed to show the mean values of each of the four scenarios based on the two significant explanatory variables (Table 4-11). Regardless of relative proximity to markets, households having formal employment ar e deriving only between 7% an d 12% of their income from


51 commercial agriculture. Households in Nett and particularly thos e not having formal means of employment (Table 4-11-Quad 2) are most heav ily dependent on agricultu re as their primary means for cash income. Although not significantly different (at P<.05), households with more limited access to markets and without any formal means of employment (Table 4.11-Quad 1) depend on commercial agriculture for cash income more than households with formal employment regardless of proximity to markets (Table 4-11 -Quads 3 & 4). Those households with formal employment and in more remote locales depend on commercial agriculture for cash income the least of all four groups (Table 4-11-Quad 4). With the exception of households without formal employment and that are in close proximity to urban market, a relatively small percentage of total income is derived from commercial agriculture. For more remote households this may be primarily a function of high transaction costs and the mark et uncertainty. For those househ olds in Nett having formal employment, the relatively small percentage of to tal income from agriculture might be explained simply by limited demand and/or household income levels meeting existing needs. Discussion Like other Sm all Island Devel oping States in remote locatio ns and having limited domestic markets, few opportunities exist for economic development on Pohnpei. Pohnpeaens have developed diverse strategies to meet the n eeds for cash income, food security and cultural obligations. To meet many of these needs, one of the primary activities household members are engaged in is agriculture, employing traditional practices that have been in existence for centuries. The results of this study, po ssibly the first of its kind on Pohnpei, indicate that all households involved with this research were producing multiple crops within Pohnpeis


52 agroforests. With the exception of sakau, the vast majority of the crops (and values) included in this study are derived from homegardens, an indi cation of the importance of these agroforestry systems to Pohnpeaens. With the exception of swamp taro which is produced in significantly greater quantities in Nett then Madolenhimw, the types and quan tities of crops being produced were not significantly different among households having more or less market access or having members formally employed or not (Table 4-1). Given the similarities in the types and quantity of crops produced, this suggests there is a relativel y high level of homogeneity of Pohnpeis agroforests. The level of homogeneity, however, is not the same when it co mes to the particular use of crops. Whereas swamp taro, bananas a nd coconuts are used primarily for household consumption and for feeding animals, yams and sa kau are particularly important for ceremonial activities. Sakau and betel nuts are the primary sources of income from the sale of crops (Table 4-7). Agriculture is contributing substantial valu e in terms of household consumption needs, ceremonial goods and monetary worth to both Po hnpeis formal and informal economy. The sale of seven of the primary crops produced in P ohnpeis agroforests alone represents a large contribution to Pohnpeis local economy estimated at just under $5 million annually (Table 4-5). Moreover, agriculture is providing necessary sources of income in light of very limited formal employment opportunities. In addition to the form al economy, ceremonial use of the same crops is valued at almost $3 million per year, indicative of the relati onship that Pohnpeaens have to their traditional culture. Beyond formal and cultural values, agri culture is providing just over $7 million annually in consumptive values and in doi ng so providing for nutritional needs in light of ever-increasing consumption of imported food goods. Combined with over a $1 million annually being used to provide extended family and friend s with food sources, the results illustrate how


53 agriculture is truly embedded in most aspect s of daily life on Pohnpei. Should the role of agriculture diminish whether in terms of its commercial value, cultural contribution or food security, it would be expected th ere would also be substantial soci etal changes, th e implications of which should be evaluated when consid ering possible development strategies. Comparing agricultures contri bution to Pohnpeis GDP with estimates of agricultures contribution to GDP in other SIDS in the Paci fic regions provides further indication of the relative importance of Pohnpeis ag roforests to the local economy. It can be seen that almost 20% of Pohnpeis total GDP is de rived from crops produced within agroforests (Figure 4-1). This value is relatively high compared to many ot her SIDS in the Pacific region, particularly if one considers the percentages for SIDS in Figur e 4-1 other than Pohnpei include production and value added of agricultural, forestry and fish ery products as opposed to only the seven crops comprising of the nearly 20% of Pohnpeis GDP1. While the commercial forestry sector on Pohnpei, and for all of the Federated States of Micronesia is limited, the fisheries sector in th e FSM remains an important contributor to GDP having an estimated value of $36,900,000 in 1999 (G illet and Lightfoot, 2001). Assuming that only a quarter of this value was attributed to P ohnpei (Pohnpei being one of four states within the FSM), the percentage of Pohnpeis GDP as define d for the other countries in Figure 4-1 would increase to close to 28% indi cating the importance of natural resources to Pohnpeis (and the FSM) economy. 1 Definition: Agriculture for all countries except Pohnpei includes production and value added figures for forestry, hunting, and fi shing, as well as cultivation of crops and livestock production. Value added is the net output of a sector after adding up all outputs and su btracting intermediate inputs. Source: World Bank national accounts da ta, and OECD National Accounts data files.


54 Taking into consideration the total value of a ll agroforests it becomes very clear that sakau is driving both commercial agricultural activitie s, as well as the total value derived from these agroforests (Tables 4-5 & 4-6). However, the importance of sakau is proving to be a mixed blessing. On the one hand, sakau is clearly contributing to household incomes and the formal and prestige economies. On the other hand, because of the substantial local demand for sakau, farmers have cleared extensive areas of upland native forests to plant sakau where it grows more quickly to a harvestable size (Kostka and Ra ynor, 2000). The results of these actions are contributing to the loss of unique and very biologically diverse native forests. Moreover, clearing is causing negative impacts downstream due to er osion and according to local farmers sakau production in the uplands is not sustainable (Sakau in relation to the environment and economy of Pohnpei is explored further in Chapter 6). Pohnpei has established a grow-low campaign to encourage the planting of sakau in the lowlands of Pohnpei which is proving to be somewhat successful (Kostka and Raynor, 2000, Raynor personal communication 2007). Unless an oppor tunity to market a greater assortment of crops becomes possible, the heavy dependence on sakau grown in the uplands will continue to cause detrimental impacts to Pohnpeis upland forests. The heavy dependence on only two crops (Sakau and betel nut) for cash income presents another challenging situation. While these cr ops are being produced within agroforests comprising of many more crops, the reliance on onl y two crops for substantial income increases the economic vulnerability of farmers to signific ant events, be it a result of anthropogenic or natural causes. To reduce the potential vulnerabil ity strategies to establish markets should focus on diversifying as well as increasing potential marketing opportunities.


55 The percentage of total income from agriculture is assumed to be a reasonable indicator of the dependence on agriculture for cash income. Cont rary to what was exp ected, total number of people per household did not appear to influence the relative dependence on agriculture for cash income. Nor did household population have a ny statistically significant bearing on the contribution of crops to total income. One explanation for this could be that enough, if not surplus, labor exists to produce what is needed for household consumption, ceremonial obligations, and what the current market demand fo r crops calls for. It also would suggest that additional labor on Pohnpei is available for income earning activ ities if such opportunities existed. Moreover, available land per household has not proven to be a stat istically significant predictor of the dependence on ag riculture for household income. At this juncture the results suggests that population density (measured by number of people per household) as well as lands available to each household (land scarcity) have not reached a point in which they are influencing agricultura l activities as proposed by Boserup (1965) and specifically the dependence on agriculture for cash income. One may assume that in other SIDS having similar household population densities and av ailable land areas agricu ltural activities are not determined by these metrics. Although all households use a portion of their crops for ceremonial purposes it does not appear that doing so has any si gnificant bearing on the percenta ge of total household income generated from crop sales; Pohnpeaens still believe in the importance of ma intaining traditional cultural ties such that the desire to increase cas h income has not yet taken full precedent over the desire to remain integrated into Pohnpeis prestige economy. The regression model also suggests that those households without employment and having greater market access derive a larger percenta ge of total income, and are therefore more


56 dependent on agriculture for cash income relative to other households. It remains to be seen if farmers will modify their farming activities in response to increased market access (viz. Von Thunen 1826). However based on the fact that households having greater market access and no formal employment generate almost 75% of thei r total income from agricultural sales strongly suggests that having greater market access will broaden the range of opportunities for Pohnpeaens as Allen (1986) and Reardon et al. (2001) suggest. Households having members formally employed generate a significantly less amount of their total income from agriculture relative to households that do not ha ve formal employment (Table 4-11). Assuming these same households re present a segment of Po hnpeis population that is reliant on nonagricultura l activities for the majority of household income, the results point to Pohnpei being in the midst of a structural tran sformation. It would follow that assuming formal employment opportunities were to increase in the future fe wer households would rely on agriculture for cash income. Howe ver based on the same results, it would be expected that if formal employment levels were to level off or even decline in light of an increasing population, households on Pohnpei would look to other opportuni ties for income generation. In doing so, the continued trend towards moving away from th e agricultural sector to support economic development would pause, if not reverse in direction. At the time this research was conducted, limited local demand was acknowledged by many farmers as being a major reason for not increas ing the production of crops with the exception being for sakau. Given Pohnpeis relatively small si ze, the fact that all of the households that participated in this resear ch were producing enough food for their own consumption and ceremonial needs, further research needs to be conducted to look at ma rketing opportunities offisland.


57 After spending more than 3 years living and wo rking on Small-Island States in the Pacific and in particular, through di rect observation of Pohnpeaens over the course of a 12 monthperiod, the author has determined that the modus operandi for many Pacific island farmers, is that they market their crops based on short-term immediate needs. This is in part likely a function of limited markets alt hough it is also likely that cult urally, Pohnpeaens (and many other Pacific island cultures) are not accustomed to producing a consistent flow of high quality products to meet market expectations. There may be the means to produce sufficient quantities of agricultural products to supply off-island ma rkets; however the current manner in which many farmers produce and sell their crops is problematic and risky to those willing to invest the time and finances necessary to establish viable produc tion and marketing chains from the farmgate to markets themselves. There is a need to develop a multi-pronge d strategy aimed at developing both local marketing opportunities as well as devoting sufficien t financial and human resources to train and enable potential farmers and buyers to devel op and implement more agricultural commerce. Conducting marketing research to determine the pot ential for niche markets outside of the FSM combined with research on possible value-added activities should be implemented. Public sector efforts should also center on developing infrastr ucture including impr oving roads and public transportation opportunities, linking rural to urban areas reducing tr ansaction costs that in turn improve potential profitabilitys for commercial ag riculture. Providing financial incentives such as tax relief for capital investment s in processing and market faciliti es, export subsidies to attract private investment, as well as establishing low-in terest credit programs fo r farmers in support of interstate and international agricultural commerce should also be given consideration. Exploring all possible options that center on production and sales of as ma ny traditional crops as possible


58 enabling Pohnpeis agroforests to remain a primary component of agricultura l development should be pursued. Throughout much of the world (and certainly among SIDS too) agricultural development policy has focused on promoting co mmercial agriculture via the intensification of land use and often through the introduc tion of monoculture practices. In su ch situations, intensification and changes in farming practices have most often taken the form of increased use of synthetic inputs and mechanization of labor. The process fundament ally altering traditiona l agricultural practices has in some instances been qu ite successful in achieving bo th economic and food security objectives. However, it has also led to increas ed dependence on producti on inputs and exposure to market fluctuations. For SIDS where such small economies have no influence on market supply and demand, the impacts of implementing such policies substantially increases the vulnerability of an islands ec onomic well being as well as poten tial for significan t deterioration of natural resources. Indeed, for those States and territories where agriculture remains an integral part of every day life, implementing su ch changes leads to fundamental changes not only to the agricultural sector, but potentially to all asp ect of a given population. As Figure 4-1 illustrates, of the countries incl uded, agricultures contribution to Pohnpeis GDP is second highest (17.8%), surpassed onl y by Tonga (28.6%) which derives the vast majority of its agricultural value from the production and export of squash using intensive monoculture practices. The promotion of Tongas squash industry has been at the center of developing its economy and has been the backbone of the transformation of what, only 50 years ago, was a non-monetary economy based on subsistence agriculture to a fully-monetary economy today. In the process of developing T ongas squash industry, the practices employed


59 have resulted in extensive deforestation and the loss of traditional agroforestry systems throughout Tonga during the last three decades (Wolff, 2001). More recently, Tongas squash industry has suffered due to increasing production costs and vulnerability to world demand fluctuations that remain out of the control of Tongan farmers. There are now ongoing efforts to reintroduce and promote traditional agroforestry practices on Tonga as a result of the loss of forests, unsusta inable monoculture practices and the loss of ecosystem services once provided by such traditi onal practices. Tonga provides a useful example of how a Small Island Developing State has transitioned from trad itional agroforestry practices that supported a countrys popula tion for centuries to more intensive monoculture producing a very narrow range of crops for the primary pur pose of developing its eco nomy. Other countries such as the Cook Islands have also experienced a boom a nd bust period of agricultural development but have been more able to respond to the collapse of its pi neapple industry through effective development of a tourism sector. Pohnp ei has yet to make a similar transition as experienced in Tonga, the Cook Islands as well as other developing states but it is well on its way. As Pohnpei continues towards a more fully monetized economy wh ereby greater cash income is sought, the results of this research sugg est that it could be expe cted, particularly given the limited employment opportunities, there would be an increasing trend to ward selling as much agricultural products as possible. If this were to happen it is al so very likely there would be a shift from the use of crops for ceremonial a nd extended family purposes to purely commercial purposes. Doing so would likely strain the exis ting cultural integrity of Pohnpeaens and would certainly test the social resiliency of Pohnpeaen society. It remains to be seen if Pohnpei will pursue similar development strategies as those pu rsued in other pacific islands such as Tonga.


60 Those in policy positions must address the potential tradeoffs between promoting more modern agricultural practices relative to the important ro le and value of existing practices. Based on the results of this study, economic comparisons of th e traditional ag ricultural system s practiced on Pohnpei and more modern technologies used elsewh ere needs to be conducted. Currently, it is unclear whether more modern practices, particul arly given Pohnpeis remoteness and associated transaction high costs, would actually provide greater and lasting benefits. Having a long tradition of being a subsiste nce economy, Pohnpei is now moving towards a more cash based society where tradition and pres tige, based largely on pr oduction of traditional crops, is slowly giving way to we stern ideals and the desires of material wealth. The trajectory and pace of where agriculture on Pohnpei goes from here will largely depend on the availability of markets for those crops currently produced wi thin Pohnpeis agroforests and possibly others that may be introduced. Critical to any analysis that may help shape Pohnpeis future course is the full appreciation and recognition of the total value of current ag ricultural practices as well as the contribution of agriculture to Pohnpeis culture and consum ption patterns. The research reported here is one step in this direction.


61 Table 4-1. Role of homegarden agroforests in household production of crops in Pohnpei, FSM Crop Unit of Measure Homegarden Total % from Homegarden Mean S.E. Mean S.E. Yams kg yr -1 25843.731440.8 76.5 Breadfruit kg yr -1 38760.748094.7 80.7 Swamp taro kg yr -1 19645.2255*55.3 85.9 Coconuts Number yr -1 1041155.91335180.4 81.2 Betel nut Number yr -1 10,1304,114.410,8504,181.2 92.4 Sakau kg yr -1 26151.51106271.6 53.0 Bananas kg yr -1 68966.186782.8 82.3 Nett vs. Madolenhimw: *= P<.10 Table 4-2. Average allocation of household production of crops for various uses (ceremonial, cash income, friends and household consumption) in Pohnpei, FSM % of households using the crop for Crop % of households producing crop Ceremonial Cash income Friends and extended family Household consumption Yam 97 8130 42 97 Breadfruit 94 8824 66 100 Swamp taro 85 4818 48 93 Coconuts 100 2736 48 100 Betel nuts 64 1248 30 64 Sakau 94 9058 60 94 Bananas 100 4864 78 100 Table 4-3. Use of crops for ceremonial, cash income, friends, and household consumption on Pohnpei, FSM Percentage of total crop production per use Crop Ceremonial Cash income Friends and extended family Household consumption Total Yam 42 5 4 49 100 Breadfruit 18 3 9 70 100 Swamp taro 3 19+ 14 64* 100 Coconuts 2 3 4 91 100 Betel nuts 1 61 13 25 100 Sakau 24 39 6 31 100 Bananas 7 3 9 81 100 Nett vs. Madolenhimw: *= P< .05 Employment vs. no-employment: += P< .05


62 Table 4-4. Average household value of crops produced in Pohnpei s agroforests, 2002 Crop Homegarden Total Mean ($ yr -1 ) S.E. ($ yr -1 ) Mean ($ yr -1 ) S.E. ($ yr -1 ) % of value from Homegarden Yams 551 94.069089.8 80 Breadfruit 85 13.610921.2 78 Swamp taro 151 34.825455.4 59 Coconuts 104 15.513318.0 78 Betel nut 506 205.5542209.0 98 Sakau 575 113.32,434597.5 24 Bananas 379 36.447745.5 79 Total* 2,089 234.74,242612.9 *Note that the total value per crop per use is not equal to the sum of individual crop values because not all households produce every crop. The total value reflects the sum of values for households producing a given crop. Table 4.5. Value of crop used per hou sehold from Pohnpeis agroforests Value per crop activity ($ yr-1) Crop Ceremonial Cash income Friends and extended family Household consumption Yam 412 95 56 296 Breadfruit 11 26 18 69 Swamp taro 20 57 46 142 Coconut 10 8 13 121 Betel nut 21 503 161 101 Sakau 606 1705 308 633 Banana 429 63 64 375 Total household value per use* 966 1,538 399 1,556 Note that the total value per use does not equal to the sum of individual crop values because not all households produce every crop. Instead, the total va lue reflects the sum of the average to tal value for each household producing and using a crop for a specified use. Table 4-6. Island-wide values for cr ops produced in Pohnpeis agroforests Crop Homegarden Total Mean ($ yr -1 ) Mean ($ yr -1 ) Yam 2,083,188 2,524,760 Breadfruit 310,597 373,255 Swamp taro 482,793 625,660 Coconuts 396,516 461,650 Beetle nut 1,226,743 1,254,504 Sakau 2,126,520 8,711,975 Banana 1,443,420 1,816,321 TOTAL 8,069,777 15,768,125


63 Table 4-7. Island wide-value of agricultural cr ops produced and used for different purposes on Pohnpei, FSM Crop Ceremonial ($) Cash income ($) Friends and extended family ($) Household consumption ($) Yam 1,269,288 109,356 91,009 1,202,756 Breadfruit 85,297 9,995 46,368 310,597 Swamp taro 105,045 14,227 93,661 455,937 Coconuts 614 13,849 24,192 461,650 Betel nuts 24,094 928,875 184,608 942,046 Sakau 1,337,416 3,701,776 711,013 2,260,412 Bananas 49,876 154,188 182,514 1,428,755 TOTAL 2,871,630 4,932,266 1,333,365 7,062,153 Table 4-8. Correlation matrix of explanatory va riables used in the model for determining percentage of total income derived from agroforests on Pohnpei, FSM Market access Households having formal employment # of household members Total lands available % of crop value used for ceremonial activities Market access 1.000 Households having formal employment 0.277 1.000 # of household members 0.149 0.3771.000 Total lands available -0.064 0.0620.1861.000 % of crop value used for ceremonial activities 0.021 -0.0150.1040.198 1.000 Table 4-9. Results of the complete regression m odel predicting agriculture as a percentage of total income Descriptor CoefficientStandard Error T-stat Intercept 0.252 0.120 2.102 MKACES Market access 0.278 0.095 2.940** EMPLOY Household with employed member -0.347 0.100 -3.453*** NHHMEM Number of household members 0.000 0.006 0.041 TOTLND Total land area available (ha) 0.025 0.020 1.258 PERCER Percent of total value for ceremonial activities -.287 0.254 -1.130 P< 0.10 ** P< 0.05 *** P< .001 R2 = 0.40 Adjusted R2 =.30


64 Table 4-10. Results of the final regression model predicting agriculture as a percentage of total income Descriptor Coefficients Standard Error T-Stat Intercept 0.277 0.073 3.934 MKACES Market access 0.265 0.093 2.847** EMPLOY Household with employed member -0.332 0.093 -3.583*** P< 0.10 ** P< 0.05 *** P< 0.001 R2 = 0.36 Adjusted R2 =.31 Table 4-11. 2 X 2 contingency table showing percentage of total household income from agriculture for significant explanatory variables No employment 72.0% (Quad 2 N=4) 19.8% (Quad 1 N=11) Employment 12.3% (Quad 3 N=11) 6.8% (Quad 4 N=7) Nett Madolenhimw 3.3 17.8 28.6 15.4 15.4 6.4 15.00.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0FijiKiribatiPalauSamoaTongaVanuatuPohnpei Countr y or Territor y Percent of GD P Figure 4-1. Comparison of the percent of GDP from agriculture in 2002 (World Bank, 2008)


65 CHAPTER 5 MODELING LIVELIHOODS ON POHNPE I USING ET HNOGRAPHIC LINEAR PROGRAMMING Introduction In contrast to trad itional economic theory that associates the alloca tion of resources to maximizing profits, small scale, limited resource families often pursue diverse strategies to meet livelihood objectives (Hildebrand et al ., 2008). Pacific islanders often manage their limited resources to meet multiple objectives both moneta ry and non-monetary in nature (Powell, 1998). For example where tradition and cultural norms remain strong, certain crops are often produced for reasons of prestige and acqui sition of social status in addition to income generation. Additionally, households may em phasize production of goods and maintenance of ecosystem services to meet food security objectives. Over centuries, Pacific cultur es have pursued several strategies to meet these multiple objectives, principal among these are the development of traditional agricultural systems (food security, so cial status gains, ceremonial and income earning objectives) fishing (food security and in come earning objectives) and more recently the pursuit of formal employment for cash income objectives (Hazel, 1993; Peoples, 1995). As the process of globalization continue s and subsequent increased de sires for greater cash income occur, the order of importance of these multiple objectives may shift. A livelihood system is defined broadly as the composite of a ll activities ava ilable to all households in the system from which to c hoose to secure their livelihoods (Hildebrand et al., 2008). Within a livelihood system there are often subsets of households selecting diverse strategies based on composition, preferences, skills, location, and household objectives. The activities chosen by a given household from t hose comprising the livelihood system make up what is known as livelih ood strategies (Hildebrand et al., 2008). Modeling households within livelihood systems enables an ex ante assessmen t of possible development programs and is one


66 tool to help better understand the implications of such development scenarios. Having such information can not only help to ensure that appropriate development programs are established, it can also help to ensure the substantial and ofte n very limited resources needed for development programs are used more efficiently. In Chapter 4, descriptive statistics were presented and regression analysis was conducted providing values associated w ith agricultural productivity to determine whether there are significant predictors of agriculture s contribution to total household income. In this chapter, the scope of interest is broadened to examine livelihood strategies being employed to meet household objectives, including pr oduction activities ot her than agriculture. Understanding Pohnpeis agricultural sector is important and can provide useful insight into potential agricultural development programs. However, th e efficacy of such programs may be less than desired without having a comprehensive understa nding of the how resource availability and constraints facing various household groups influence the allocation of limited resources and for what purposes. In the next section, a brief description of the method used to model households on Pohnpei is presented followed by a descri ption of the model specificati ons, the livelihood systems being modeled and three scenarios that the model simulate s. The remainder of the chapter consists of results of the simulations followed by a discus sion based on the results of the simulations conducted. Methods Mathem atical programming has been used in th e agricultural sectors for decades. One such type of mathematical programming is linea r programming. A linear program (LP) is a mathematical optimizing procedure that ma ximizes or minimizes an objective subject to constraints (See Appendix D for a mathemati cal explanation of linear programming).


67 Traditionally, LPs were developed and used in agriculture as a normative tool as opposed simulating what a farmer was ac tually doing and based on current actions, to estimate how might a desired objective (increasing cash income, mini mize land or labor etc.) be achieved. In the early years of linear programming, modelers limited their analysis to those activities specific to agricultural productivity without incorporating other activ ities pursued by households or commercial firms More recently ethnographic linear programming models (ELPs) have been developed as a means to more fully comprehend livelihood sy stems and strategies being employed by households and to assess potential impact resulting from a change in technologies, environmental conditions, policy or livelihood objectives themse lves. Ethnographic linear programs are simply a decision support tool that enables decision make rs to be more informed. The ELP models are developed (1) to help research ers or other decision makers understand the complexities and diversity of livelihood systems and the households that depend on them and ultimately (2) to simulate the livelihood systems and livelihood strate gies of the diverse ho useholds within these systems. At the household level, the process of devel oping ELPs consists of determining resources such as land, labor, and capital (usually cash but can include tr ading) used for a particular production activity, production ou tputs and the total amount of resources available to a household. This information is then used to create a matrix comp rised of resources used (inputs) to produce a given amount of production outputs along with actu al or potential uses of the outputs being produced. Once the ELP s are developed the models themselves can be validated. Validation is most readily done by direct interviews with participating house holds to see if the models structure inputs and output s closely reflect reality. When and where available, validation


68 may also be supported comparing model inputs a nd outputs with secondary literature and data collected by local agricultural extension offices Once validated, the models can be used to simulate and assess potential impacts on diffe rent kinds of households of programs being considered prior to implementation in a given locale or within a pa rticular recommendation domain. To conduct the analysis of livelihoods on P ohnpei, ELP was used. Both descriptive and analytic in scope, ELP allows for relevant ac tivities carried out by individual households along with production, market or other co nstraints to be incorporated in to analyses to determine, ex ante, the potential changes result ing from a proposed action. While it is recognized that every household is engaged in both pr oduction and reproduction activities the analysis conducted for this chapter centers only on producti on activities included in the ELP models as described below. However, in doing so, available adult labor was limited to eight hours per day, five days a week and childrens available labor is limited based on whether or not they were enrolled in school. Additionally, while absolute valu es for time required to prepare food were not included in the model, time and resources required to gr ow food for household consumption was. Model Specifications Data used for the ELPs developed for this ch apter were collected, during a one-year period, based on the m ethods described in Chapter 3. For al l four scenarios discussed below, simulations were conducted for each of the participating hous eholds (n=33). Based on the primary source(s) of income, four livelihood syst ems were identified. Household m odels were then grouped per livelihood system and averages for households w ithin each of the respective systems were calculated.


69 Production Activities Several activities were incorpora ted into the ELPs developed for this analysis. Agricultural activities included produc tion of the seven main crops found w ithin Pohnpeis agroforests (yams, breadfruit, giant taro, coconuts, beetle nut, sakau and bananas) Agricultural activities were broken down based on whether they were conducte d within homegardens or other lands (Land2, Land3) either owned by extended family or are c onsidered public lands. Agricultural crops are used for multiple purposes on Pohnpei including fo r each of the use options of agricultural crops as identified in Chapter 4 (household consumption, given to friends and extended family, commercial sale, use in traditional activities) wa s incorporated into the ELPs to reflect actual uses by individual households. Other activities in the models comprised of production of handicrafts, mangrove crab harvesting, fishing, full and part-time employme nt, receiving retirement/pension income and receiving remittances, either from local or from sources from outside of Pohnpei. Coefficients for agricultural and fishing act ivities in the ELP m odels included production inputs, namely land and labor, broken down by adu lt male, adult female, male child, and female child1. For households with employed members, the standard number of hours worked in the public sector was used unless otherwise indicate d by participants. Income generated from sales of handicrafts, crops, and fish were derive d based on production levels multiplied by local market prices. Income from employment, retirement and remittances, was based on responses provided from household surveys (Appendix C). 1 For this research a household member was considered an adult if s/he was 18 years of age or older.


70 Model Constraints Constraints can be considered re strictions on activities, uses, and/or quantities of an input or output coefficient that often are used to capture realities of households in a given livelihood system For example a constraint can be incorpor ated into a model to reflect minimum needs to produce and use a given crop for household food consumption or a maximum value for the amount of crops that can be sold based on market conditions. Relaxing or reducing constraints is one way to simulate potential effects on a given livelihoods system resulting from the introduction of policies, technolog ical changes or otherwise. For the base case scenario (S cenario 1 below), constraints imposed in the ELPs are based on available resources and how each of the househol ds allocated harvested goods to various uses identified above. Thus, during the course of da ta collection, if a household had recorded 100 kg of yams being used for ceremonial purposes, the model had a constraint requiring at least 100 kg of yams be produced and alloca ted to ceremonial activities. Additionally, constraints limiti ng the amount of crops or fish that can be sold commercially, demand constraints, were imposed based on what each household actually sold. The underlying assumption behind imposing a market constraint is based on the understanding that if local demand were greater, households would sell more of their crops and/or fish. Since not all households participated in every activity, if a particular household did not engage in a particular activity, a constraint was imposed restricting house holds from such activities. Employment constraints were imposed to limit the amount of time, and thus income that can be earned based on the actual number of household pa rticipants that held formal employment positions and the income generated from such activities. Finally, a constraint imposed for cash needs was based on what was identified as cash needs per household per m onth converted to an annual amount. Modification of these constraints, an assumed crop pricing policy and expanded


71 market opportunities provide the basi s for three additional scenarios to be simulated as described below. Objective Function It is recognized that in settings such as Pohnpei, households m ana ge their resources to meet multiple objectives. Among these are those a ssociated with meeting cultural and extended family obligations and food security. However, mo re and more, acquiring cash income to support material wealth has become a primary goal of many households and maximizing cash income a primary objective to achieve th is goal. Therefore the objective function for this model is maximizing end of year cash income. Livelihood Strategies Based on the results of the field research conducted, four household groups were identified (hereafter referred to as Em ployment househol ds, Agriculture households, Fishing households and Remittance households). While many of the participating households were engaged in similar activities (Table 5-1) given the objective function to maximize cash income, households were grouped based on relative dependence on empl oyment, agriculture, fishing or remittances as primary sources of income. Description of Livelihood systems Employment : Households that generate the majo rity of their income from formal employment (N=18). Households in this gr oup include those located in Nett (56%) and Madolenhimw (44%). Eighty nine percent of these households had at least one member employed full time while 11% had at l east one member employed part time. Commercial agriculture : Households that generate the majority of their income from commercial agricultural activities (N=6). There were no households in this domain having members formally employed full time. Sixty se ven percent and 33% of households in this domain are located in Nett and Madolenhimw respectively. Commercial fishing : Households that generate the majority of their income from commercial fishing activities (N=3 ). All households in this doma in are located in one village


72 in the district of Madolenhimw. Two of the households had no members formally employed full time. One household had at least one member employed part-time. Remittance : This recommendation domain comprise d of households th at generated the majority of their income from sources other than formal employment, commercial agriculture or commercial fishing (N=6). Sixty seven percent and 33% of households derived the majority of their income from remittances and retirement sources respectively. All but one household was located in the more remote district of Madolenhimw and none of the households had any members formally employe d full time. One household within this recommendation domain had at least one member employed part-time. Scenarios Simulated Scenario 1: Sim ulates the base case and reflects what households were actually doing at the time and under market conditions that existed when the field research was conducted. Scenario 2: Simulates the effect of a crop pricing policy of 25% in crease in crop prices while holding constant the amount of crops sold in dome stic markets (demand constraint) to that of Scenario 1. Scenario 3: Simulates the effects of an increase of intern ational market oppor tunities in light of declining uses of crops with in the prestige economy. To do so, Scenario 3 simulates the assumption that Pohnpeis local markets for the crops and fish meet local demand and introduces an opportunity to sell crops internat ionally at 50% of actual local market prices2. This scenario also assumes productivity remains constant while ceremonial use of crops is constrained at only 50% of original values, providing surplus for sale to international markets. Scenario 4: Simulates an expansion of fish ma rketing opportunities by relaxing fish consumption constraints by 50% and assuming Pohnpeis domestic market could absorb additional fish sales at current prices. S cenario 4 only includes changes in marketing conditions specific to fishing activities as oppos ed to mangrove crab harvesting and fishing. 2 A 50% decrease in crop price re lative to actual local crop prices used in Scenario 1 simulations.


73 Results Model Validation Given this is the first known work on Pohnpe i that docum ents productivity and use of crops over the course of a years period there ar e very few secondary data available to validate the model. However, the model has been validate d to the extent that results were shown and discussed with household members participating in this research. Based on their feedback the base model (Scenario 1) is an accurate representation of livelihoods on Pohnpei. Scenario 1: Base Case Scenario 1 represen ts the results of the si mulations based on the ac tual data collected. Therefore the results of Scenario 1 most closel y represent current strategies and associated incomes per livelihood system. Households in all four livelihood systems rely on between three and seven different sources of income (Figure 5-1). There is a posit ive correlation between the number of income sources and actual income leve ls among the four livelihood sy stems. However although such a correlation exists, three out of the four system s derive between 65% and 77% of their total household income from only one source. Agricultur e and Fishing households generate more than 70% of their total incomes from agriculture (7 7%) and fishing (72%) re spectively. Households within both systems also depend on part-time em ployment and remittances as their only other sources of income. Households reliant primarily on employment ge nerate income from more sources (total seven) than households within the three other livelihood systems. Even so, these same households are reliant on employment for almost 70% of total annual income. Second in importance to employment is agriculture income providing 12% with the remaining sources of income ranging from between 1%-6% of total income (Figure 5-1).


74 In contrast to the other livelihood systems, households within the Remittance group have more balanced strategies to generate cash income The two most important sources of income are remittances (39%) and retirement (36%) followed by agriculture (17%). The balance of income is generated from part-time employment (5%), handicrafts (2%) and fishing (1%). Fishing households is the only livelihood system that doe s not rely on commercia l agriculture for cash income. Household incomes for the four livelihood systems range between just over $11,500 yr-1 and just under $2,000 yr-1 (Figure 5-2). Employment house holds have significantly greater income (P<.001) relative to all other domains. Households most reliant on fishing ha ve the lowest cash incomes ($1,966 yr-1) followed by commercial agriculture ($3,303 yr-1) and remittance ($4,536 yr-1) livelihood systems. For the remaining three scenarios, production inputs and quantity of crops harvested did not change relative to the base case. Instead, the use of crops a nd/or fish harvested and the value associated with their use changed based on the particular scenario simulated. Scenario 2: 25% Crop Price Subsidy In Scenario 2, prices for crops that were actu ally sold in Scenario 1 were increased by 25% to sim ulate how household incomes would change assuming such a pricing policy was established. Households most de pendent on agriculture would real ize the largest increase in incomes relative to households w ithin the other livelihood system s. The benefit of such an increase in crop prices would contribute an additional $643 yr-1 to Agriculture households, a value equivalent to almost 20% of their total existing incomes (Table 5-2). Income levels for Employment and Remittance households would increase by $343 yr-1 and $178 yr-1 respectively representing an increase of between 3-4% of total income. Fishing households are not currently


75 engaged in any commercial activities. As a result, their income would not change as a result in crop prices being increased by 25%. Scenario 3: Relaxing Ceremonial Use Constraint Table 5-3 summarizes the results of reducing the constraint imposed on cerem onial use of crops by 50% while simultaneously assuming inte rnational markets (at 50% of domestic crop prices) could absorb the surplus crops not used for ceremonial purposes. In this scenario, commercial agricultural households would realize the largest increase in income ($335 yr-1), roughly half of the gain derived based on Scenario 2. An increase in household income of $249 yr-1 would be realized by employment households representing a total increase of 2.0% relative to Scenar io 1 results. Although Remittance households would real ize a greater percentage of increased income relative to Employment households, in absolute terms the increase would only be $153 yr-1. Fishing households would realize very little benefit from relaxing ceremonial constraints because of the small amount of agricultural products produced that are actually used for ceremonial purposes. Scenario 4: Expansion of Fish Markets In Scenario 4 sim ulations were conducted to determine the relative e ffects of increasing the sale of fish in local markets. Those households most reliant on fish for income are also the largest consumers of fish on a per househol d basis. As a result, and working under the assumption that maximizing household cash income was the objective, these same households would benefit the most from having an opportunity to sell as much as 50% of the fish currently being consumed (Table 5-4). The actual increase in income for fishing households is more than double ($477 yr-1) of any other household groups representing an increase of 24% over the Base Case.


76 For both Employment and Remittance households they would realize an increase of household income of approximately $200 yr-1. Agricultural households consuming the least amount of fish of the four groups would only realize a very small increase representing a difference of less than 1% of annual household income or $24 yr-1. Discussion The results of Scenario 1 (exi sting situation) illu strate how the m ajor share of household income is derived from employment, particularly relative to the three other livelihood systems. Second to households within the Employment group, households most reliant on remittances and retirement have the second highest incomes albe it still less than 50% of formal employment households. Those households most reliant prim arily on natural resources (Agriculture and Fishing households) generate the lowest cash income among the four groups. Moreover, in the case of households reliant on fishing their income is less than 20% of the income generated by Employment households illustrating the large in come disparity between these two livelihood systems. Two main conclusions can be drawn from these findings. First, while the vast majority of all households on Pohnpei derive income fr om multiple sources, employment outside the home is by far the largest singl e source of income. With approximately 50% of households on Pohnpei having at least one member employed, formal employment is the driving stimulus of the economic activity on Pohnpei. Secondly, although inco me levels for households most reliant on Pohnpeis natural resources are far less than those having form al employment, commercial agriculture and fishing are providing indispensable incomes gi ven the very limited opportunities for other income ear ning activities. The differences in income earning strategi es and amounts between the four livelihood systems also illustrate the potential vulnerabilities that households within each system must cope with. For Employment, Agricultur e and Fishing households there is a significant reliance (at


77 least 65% of total income) on only one source of income. For households within the Employment system, and with the continued decline in Compact funding, public sector employment opportunities will continue to diminish. Given that close to 50% of formal employment is within the public sector, should a reduc tion in employment opportunities occur, these same households would loose the majority of thei r household income. In the case of Agricultural and/or Fishing households where more than 70% of incomes ar e generated from natural resources, they are particularly vulnerable to natural disasters, disease occurr ence and/or introduced pest infestations3. Such vulnerabilities, both in terms of secure employment and exposure to natural disasters support much of what has been described as unique challenges facing Small Island Developing States. Oddly enough, it may be households that are less reliant on ex isting domestic employment or markets that are positioned best to at least ma intain their current levels of income. For this group of households their income is primarily gene rated from either retirement or remittances. It is highly unlikely that sources of retirement funds (Social Securit y, pensions) will be curtailed in the foreseeable future. Similarly, with the ex ception of a substantia l downturn in the US mainland, Guam, or Hawaiian economies there is litt le reason to expect that remittance levels will decline. The results also illustrate the value of ethnographic linear programming (ELP) as a means to examine, ex ante, the potenti al impacts of different policie s or development strategies on different types of households. Moreover, this study demonstrates the importance of tailoring development strategies if the goal is to pr omote development and increase cash incomes for 3 Such was the case with taro producers in Wester n Samoa that experienced a series of typhoons and then disease that repeated ly destroyed entire taro produc tion during the 1990s (Paulson and Rogers, 1997)


78 different household types in a given locale, and part icularly on Pohnpei. In the case of Scenario 2 where a cropping price subsidy was simulated, house holds most reliant on agriculture for cash income would realize almost a 20% in crease in household income ($643 yr-1) while the other three household groups would realize a much smaller proportional benefit if any at all. Likewise, as simulated in Scenario 4 where marketing opportunities of fish were expanded, with the exception of households most reliant on fish for income the other household groupings would realize only a marginal increase in household incomes. Given the large discrepancy in income among the four livelihood systems, it seems reasonable that policies aimed at improvi ng household income should focus on improving economic opportunities for those households that currently do not have access to formal employment. More specifically, policies aime d at improving cash income through enhanced market opportunities for both fish and agricultural produc ts would benefit those households that do not have formal employment or that are curr ently receiving remittances and/or retirement income. Doing so may also help to offset th e loss of income from the potential reduction in public sector employment that may result from declining Compact funds. Pohnpeis economy mirrors the MIRAB (Migra tion, Remittances/Aid and Bureaucracy) model first discussed by Bertram and Watters (1985). The underlying premise of the MIRAB model is that Pacific islanders have been i nvolved in a global economy for decades, and have responded to the unique characterization of SIDS via emigration of island ers to metropolitan markets as (factors of production), the transfer of remittances back to home islands as (financial transfers) and procurement of s ubstantial aid to fund large bureaucracies resulting in inflated public employment (producti on of non-tradables).


79 With almost two thirds of households on Pohnpei having immediate family residing offisland (emigration), all four livelihood systems are deriving at least some income from remittances (financial transfers). In particular Remittance households are receiving almost 40% of total income is from remittances. This suggests not only the importance of remittance income to support livelihoods on Pohnpei, but also the impor tance of the ability of Micronesians to study and work freely in the United States and its territories as a condition of the Compact negotiations. Compact funds are also pr oviding essential aid responsible for supporting formal employment on Pohnpei (production of non-tradable s). With substantial overseas aid, Compact funds are the primary source s upporting the public employment sector estimated to comprise close to 50% of all formal employment positions. Funds supporting the public employment sector are also providing indi rect funding to Pohnpeis lo cal economy through purchases of goods and services derived from salaries as well as other government expenditures necessary for everyday operations. Outside the bounds of the MIRAB model, Compac t funds are also indirectly linked to the health and viability of Pohnpeis natural resources. The results of this study show an inverse relationship between formal em ployment and dependence on natural resources for total household income. As noted above, the conditions of the Compact are providing both direct and indirect funding supporting livelihoods on Pohnpe i via the support of public sector employment and through the provision of remittance providing opportunities. Without such opportunities, and unless private sector opportunities are it is very likely there w ould be a trend towards the overexploitation of Pohnpei s natural resources.


80 Given the narrow band of opportunities for the development of Small Island Developing States, a strategy for expanding marketing opportun ities for agricultural and fishery products on Pohnpei will have limitations. To complement su ch a strategy, developing a portfolio of additional possible opportunities needs to be pursu ed. One such opportunity may be to research the potential for eco-cultural tourism highlightin g the unique culture, resources and traditional agricultural practices on Pohnpei. Doing so has the potential to support economic development through establishing private sector opportunities for such activ ities as home stays where visitors are exposed to Pohneaen life while al so potentially increasing the demand for locally produced goods. Another strategy is to ensure if not promote through policy incentives the freedom to travel and work abroad enabling employment and in turn remittance income on Pohnpei. Such strategies would serve as an exam ple for other Small Island Developing States pursuing development that complements their ex isting livelihoods as opposed to fundamentally changing their way of life.


81 Table 5-1. Percentage of hous eholds per livelihood system i nvolved in various activities on Pohnpei, FSM Activity Livelihood System Employment Agriculture Fishing Remittance Employmentfull time 89 0 33 0 Employmentpart time 11 173317 Commercial agriculture 94 100086 Subsistence agriculture 100 100 100 100 Commercial fishing 44 17 100 50 Subsistence fishing 44 17 100 100 Remittances 44 33 33 100 Retirement 17 0 0 50 Handicrafts 6 0 0 17 Table 5-2. Impacts of a 25% increase in crop prices per livelihood system on Pohnpei, FSM Livelihood system Base case 25% crop price subsidy Income difference % change Employment $11,537$11,880$343 +2.9 Commercial agriculture $3,303 $3,946$643 +19.4 Commercial fishing $1,966 $1,966$0 0.0 Remittance $4,536 $4,714$178 +3.9


82 Figure 5-1.Percentage of cash income from di fferent sources per recommendation domain on Pohnpei, FSM Commercial Fishing: Percentage of household income earned from different activities on Pohnpei, FSM 7% 21% 72% Employment Remittances Fishing Commercial Agriculture: Percentage of household income earned from different activities on Pohnpei, FSM 77% 15% 8% Agriculture Employment Remittances Employment: Percentage of household income earned from different activities on Pohnpei, FSM 12% 1% 68% 6% 5% 4% 4% Agriculture Handicrafts Employment Retirement Remittances Mangrove fishing Fishing Other: Percentage of household income earned from different activities on Pohnpei, FSM 17% 2% 5% 36% 39% 1% Agriculture Handicrafts Employment Retirement Remittances Fishing


83 $11,541 $3,303 $1,966 $4,536 $0 $2,000 $4,000 $6,000 $8,000 $10,000 $12,000Total average income Formal employment Commercial agriculture Commercial fishing Other Recommendation domain Figure 5-2. Scenario 1-Total annual cash inco me per livelihood system on Pohnpei, FSM Table 5-3. Impacts of reduci ng ceremonial crop constraint s by 50% and assuming an international market exists for surplus crops per livelihood system on Pohnpei, FSM Livelihood system Base case International market/50% reduction in ceremonial use Income difference Percent change Employment $11,537$11,786$249 +2.1 Commercial agriculture $3,303 $3,638$335 +10.1 Commercial fishing $1,966 $1,993$27 +1.3 Remittance $4,536 $4,714$153 +3.3


84 Table 5-4. Impacts of reducing fish consumption constraint by 50% and establishing domestic markets to sell surplus fish per livelihood system on Pohnpei, FSM Livelihood system Base case 50% of fish consumption sold in domestic markets Income difference % change Employment $11,537$11,733$196+1.7 Commercial agriculture $3,303$3,327$ 24<+1.0 Commercial fishing $1,966$2,443$477+24.2 Remittance $4,536$4,746$210+4.6


85 CHAPTER 6 DETERMINING AGROFORESTRY PROFITABILITY USING THE POLICY ANALYSIS MATRIX Introduction In recent times, there has been a push to e xpand sakau production into the upland areas that have cooler tem peratures and gr eater rainfall than in the lowlands; in the uplands, cultivation time before harvests is generally longer than in the uplands. However, the drive to maximize sakau production in the uplands has led to clearing large areas that in turn are exposed to heavy rains and significant soil erosion. Indeed, ba sed on recent evidence (Kostka and Raynor, 2000), commercial production of sakau in Pohnpeis upland fo rests is not sustainable, with an estimated 5% to 10% decline in production annually. In contrast, sakau pr oduction has been practiced in Pohnpeis homegardens for decades suggesting ecol ogical sustainability of the practice (Raynor and Fownes, 1991a,b). Non-governmental organizations on Pohnpei ar e attempting to promote lowland sakau production within homegarden agroforests but co ntinued upland clearing remains a problem. The government of Pohnpei now recognizes the negative impacts associated with upland sakau farming while also recognizing the need on the farmers part to gene rate income. Comparing private and social profits for these three sy stems will provide a building block to develop appropriate strategies for the fu ture use of Pohnpeis resources. The primary objective of this chapter is to examine the profitability of three different agroforestry systems on Pohnpei, FSM using th e Policy Analysis Matrix methodology and to assess how policy may influence their profitab ility. The three systems are: two homegarden systems differentiated by the presence or not of freshwater swampland and a third upland system cultivated primarily for a single cash cropsakau. The Policy Analysis Matrix (PAM) methodology is used to assess two different measur es of profits: private profitability (PP), or


86 those profits that are actually realized by practic ing farmers; and social profitability (SP), or those profits that in theory woul d be realized in the absence of policy distortion and/or market failure. This will provide a means to quantify the impacts of current public sector policies and market failures (or lack thereof) for dominant agroforestry systems on Pohnpei. In addition, essential information will also be provided to better understand the current motives of farmers and how potential government strategies may impact the agricultural sector. Moreover, the comparison of profits reveal s the underlying comparativ e advantage of different commodities/systems in question. A secondary objective of this chapter is to provide a framework to expand the traditional PAM allowing for the internalization of non-market values. During the past few decades, the PAM has been used to quantify profitability and the impacts of policy interventions on dairy system s in Kenya (Staal, 1996), rice in Thailand (Yao, 1997), and rice and sugar in Indonesia (Nelson and Panggabean, 1991; Pearson, et al ., 1991). Whereas the PAM methodology has been used to measure a single crop produced within an agroforestry system (Adesina and Coulibaly, 19 98), to the best of the authors knowledge, the PAM approach has not been used to quantify prof itability of multiple cr ops cultivated within complex agroforestry systems anywhere in the world, let alone within the Pacific region. What follows is first, a brief description of the farm ing systems being analyzed, as a means to further distinguish them from the homegarden and o ther lands discussed in Chapter 3. This is followed by a discussion on the PAM methodology its elf. Next, sources of data and the policy scenarios used within our analysis are presen ted followed by our results and a discussion on the respective findings. The chapter concludes with a few final thoughts regarding the Policy Analysis Matrix as a tool for profitability analyses.


87 Background and Methodology Description of Three Farming Systems Being Analyzed Chapter 2 described general characterist ics of the hom egardens found on Pohnpei. However, in Pohnpei, like in many other island lo cales, people have established homes in areas where wetlands exist, combining wetland food pr oduction practices with those resembling other homegarden systems. Data collected for this research indicated that homegardens on Pohnpei could be further differentiated by the presence of wetland areas that have been converted into taro patches. Together, these two homegarden systems, homegardens (HG) and homegardens with wetlands (HGW), dominate the agricultural la ndscape on Pohnpei and therefore both are examined in this study. A third system of interest is centered on the cultivation of sakau, primarily in the uplands of Pohnpei. The significance of sakau to Pohnpeis econom y is growing in response to increasing demand, primarily from the domestic market. At one time used almost exclusively for traditional ceremonial activities, sakau is now more widely used for both social and ceremonial purposes and therefore its economic importance has incr eased, which has become a mixed blessing. On the one hand, increasing sakau production is c ontributing both directly and indirectly to economic activity on Pohnpei. On the other hand, recent research has shown that production of sakau has led to the conversion of a significan t amount of Pohnpeis re maining intact-upland forests to the upland sakau system (UPS) ( ADB, 1996; Kostka and Raynor, 2000). The UPS is best characterized as being less biologically diverse relative to the other two homegardens systems and having much more emphasis on cu ltivating sakau for ceremonial and commercial needs. While all of the participating househol ds had homegardens 15 of the households had HGW and 18 households had HG systems. Eleven of the participating hous eholds were actively cultivating sakau in the uplands using UPS.


88 Policy Analysis Matrix Methodology Monke and Pearson (1989) developed the Poli cy Analysis Matrix m ethodology in the late 1970s primarily for use in the profitability anal ysis of single crops w ithin specific cropping systems. The PAM has proven to be an effectiv e tool for analyzing the influence of policy distortions on farmer profits and the willingness of farmers to modify existing practices. For example, trade, domestic factor market, as well as exchange rate policies can be analyzed in relation to their impacts on farm profits (Pearson et al ., 1991). Additionally, the PAM allows for in-depth analysis of the domestic or international production ch ain, i.e., activities involved in the movement of products between the farmgate and markets. As a result, the macro-micro linkages can be clearly il lustrated, helping to inform policy ma kers about the potential tradeoffs and impacts of alternative t echnologies, as well as revea ling the underlying comparative advantage of different commodity systems. According to Monke and Pearson (1989), the mechanics of the PAM consists of two accounting identities (Table 6-1). The first measur es profits based on tota l revenues less the costs of tradable inputs and domestic factors. The top row of the PA M represents private costs and returns, where profitability (D) is defined as reve nue (A) less total costs (B+C). The prices used for calculating private profitability are based on observations, inclusive of any current policy interventions and/or market failu res. The second row of the PAM measures social profitability based on the difference between social revenue (E) less social costs (F+G). In contrast to private prices, social prices are efficiency prices or those that would be observed without any policy distortions or market failures. It follows that social revenues and costs reflect the underlying scarcity values and if used in actual decision-making would reflect the optimal allocation of resources from an ec onomic efficiency standpoint. While relevant


89 values for the calculation of private profitability are derived dir ectly from actual observed prices at the farm and markets, determini ng social values is more complex. Social prices for those commodities traded in the international market, both inputs and outputs, are derived using world market prices; CIF (cost, insurance, freight) price for imports and/or the FOB (freight on board) for exports. Bo th CIF and FOB prices provide the basis for calculating the import and export parity prices ne eded for determining farm level revenues and costs. For example, the social farmgate cost of imported fertilizer used within a farming system is its CIF price plus the marketing and distribution costs of moving the goods from the dock to the farmgate. In contrast, farmers exporti ng their goods would expect to realize the FOB price less the costs of processi ng, marketing, and transporting their goods to the dock in a distortion free setting. Calculating pa rity prices allows a policy anal yst to identify the impacts of policy intervention(s) at all stages of production. Social prices for domestic factors are based on social opportunity costs reflecting underlying domes tic factor market conditions. For example, in the case of labor, assuming a f unctioning labor market freedom of movement among laborers, the social price would be equivale nt to the going market wage rate. The second accounting identity (third row of the PAM) allows for the calculation of divergences between private and social revenues, costs, and profits resulting from policy impact(s) and/or market failures. The third row of the PAM also allows for determination of the direction of transfers between producers, government budgets, and consumers. The signs of each of the variables in the third row facilitate determining whethe r producers are being taxed or subsidized. To quantify costs and benefits resu lting specifically from externalities, these costs and benefits can be disaggregated from polic y distortions and labor, land, and credit market failures by incorporating an addition al private and social prices ro w (Table 6-2). For example, in


90 the case of an environmental tax on outputs bein g assessed to mitigate the costs of soil erosion, the farmer would realize these costs, thereby in ternalizing them and, as such, the outcome would be reflected in the private prices2 row within A2. On the other hand, where externalities are generated from, for example, carbon sequestration, yet the farmer is not internalizing benefits derived from this additional output of the farming system, the values of these, externalities would be included in the social prices2 row within E2. The net effects of both policy distortions and market failures are still found with in the second accounting row, but in this case, I would be equal to the (A1+A2) (E1+E2). Likewise, the net transfers for each of the remaining costs and in turn profits would be calculated accordingly. The PAM also allows for the determining polic y parameters for measuring the impact of policies on single-commodities, whole farm production systems, and comparing multiple systems. Most notably among these parameters are the nominal protection coefficient (NPC), effective protection coefficient (EPC), the priv ate cost ratio (PCR), the domestic resource cost (DRC) and the profitability coefficient (PC). Each of these parameters is briefly described in Table 6-3. Data Specifications and Policy Scenarios Data used for this analysis were collected from randomly selected households as described in Chapter 3. In each of the homegardens, all of these crops are being cultivated albeit in different quantities as a function of the presence/absence of the wetlands. In contrast, in most areas where the UPS are located, sakau is the prim ary, if not the only crop being cultivated. Land area per household was based on either owner docume nts or using local landmarks as proxies to estimate land area used/owned by each household. Da ta for all three systems were derived from mature agroforests in which the perennial co mponents had attained fruit-bearing age.


91 Averages for labor inputs and crop outputs pe r hectare were determined for each of the three respective systems providing the basis for cost and revenue estimates. Prices used to determine values for the crops harvested were based on those observed by farmers throughout the same one-year time period. The price of la bor was based on Pohnpeis minimum wage rate1. All farm operations were manual; no form of mechanized farming was used. Only 5% of households indicated that they used any producti on inputs other than labor (the purchase of planting materials is almost nonexistent on Pohnpei where farmers obtain needed planting materials from their own farm pl ots). Farm profits were derived using the total value of outputs less costs of labor. There is no current trade or pricing policies contributing to diverg ences between private and social prices involving th e HG and the HGW systems. Yet, these systems are providing a positive externality in the form of carbon sequestra tion. Farmers practicing either the HG or the HGW systems are not currently receiving payments for carbon being sequ estered within these systems, and therefore, payments for carbon are no t considered within the private profitability calculation. Instead, assuming a f unctioning carbon market in P ohnpei (correcting for a market failure), farmers would receive payment and, thus, values for sequestration are to be included in the social profit calculation. More specifically, given carbon sequest ration is produced within a specified system, in the PAM it is accounted for in E2 or in the social output calculation. For complex tropical agroforests, average values fo r carbon sequestration rates have recently been estimated to range between 1.5 to 4 Mg ha-1 yr-1 (Palm et al ., 2000; Montagnini and Nair, 2004). For carbon sequestration value determination in th e HG system, an average of these values (2.75 1 Pohnpeis minimum wage in 2001 was US$1.35 hr -1. Given the significant unemployment rate, the actual opportunity cost of labor is likely to be clos er to US$0.00. Regardless, there is essentially no hiring of labor for farming pur poses on Pohnpei; instead, the immediate household members, friends, or extende d family provide the labor.


92 Mg ha-1 yr-1) was used. Research conducted on another island in Federated States of Micronesia (Kosrae) has shown that wetland-based agrofore sts similar to those in Pohnpei have carbon sequestration rates approximating 2.3 Mg ha -1 yr -1 (Chimmer and Ewel, 2002), which is used for the HGW system in this analysis. Prices for carbon sequestration are generally between US$10 and US$30 Mg -1 (Niles et al., 2001). In contrast to the two homegarden systems, carbon values were not included for the PAMs derived from the UPS system. The process of cl earing much of the existing vegetation from the land, removing sakau roots and the process of harv esting sakau resulting in substantial soil loss and disturbance, may actually lead to a net carb on loss in these systems. Moreover, cultivating sakau in the uplands has increase d downstream erosion, negatively affecting the coral and fishery resources in local lagoons. Thus, the amount of potential carbon sequestered is compromised resulting from extensive soil disturbance and negative extern alities stemming from erosion generate costs not curren tly internalized by those farming in the uplands. The external cost(s) of erosion resulting from the UPS are incurred by Pohnpeaens yet are not being paid for by the farmers responsible for them. As such, profits realized by the farmers actually reflect greater values than if they were paying the full cost of their cultivation activities (i.e., private profits are greater than soci al profits by an amount equivalent to the costs of the erosion impacts). While exact costs of erosion stemming from upland sakau production has not been determined, for our analysis, we have assumed a difference of 10% between the private and social output price of sakau for the UPS to reflect these costs. This assumption is based on the premise that if farmers in the uplan ds were to be assessed a 10% environmental tax (reflected in the A2 calculation) on the production of sa kau the receipts from such a tax would be sufficient to pay for the downstream impacts associated with upland farming. Profits


93 for these three systems are therefore not only affected by the compositi on and quantity of the different products being produced, but also the effects of the pres ence of two externalities, one positive (carbon sequestration) and one negative (erosion). Based on numerous discussions with local farmers and extension agents on Pohnpei, evidence to date suggests that each of th e two homegarden systems (HG and HMW) are producing a consistent quantity of outputs from year to year using the same level of labor inputs. As such, we assume that the annual data collected for this study represen t inputs and outputs that remain constant from one year to another fo r both of the homegarden systems. On the other hand, Pohnpei farmers reported an unsustainable 10% annual decline in saka u cultivated in the UPS. It is valuable to examine how changes in cr op productivity as well as any proposed policies may influence farmer profits, and in turn, how thes e influences are likely to factor into farming decisions over time. Given the consequences of upland sakau production, policy makers on Pohnpei are interested in determining what strategies may keep farmers from further expanding into the uplands. From an economic policy stand point, pricing policies in the form of taxes and/or subsidies are often used as incentives (or disincentives) to achieve a desired outcome. Here, in addition to examining how a 10% per annum decline in sakau yields in the UPS influences farmer profits, we also explore how two such policies, a subsidy on outputs from the two homegarden systems and an environmental tax on sakau from the UPS are likely to influence future farming activities on Pohnpei. Within the context of the PAM methodology, it is necessary to calculate net present values (whereby future values are discounted into toda ys currency terms) for revenues, tradable and non-tradable inputs, and domestic factors to measure divergences that may exist over time and


94 how they influence profitability outcomes. In th eory, if elasticities of supply and demand were known for Pohnpeis economy, one could examine th e dynamic effects associated with price and production changes over time. However, such data are nonexistent in Pohnpei. As a result, we assume constant output prices fo r all three systems, constant input and output coefficients for the HG and HGW systems, and annual sa kau yields declining by 10% fo r the UPS. In the Federated States of Micronesia, the currency used is the US dollar; therefore a ll financial and economic values are in US dollar denominations. Results Table 6-4 provides the profitability results fo r a single years outco m e under the current and potential policy scenarios of Pohnpei. For a ll three systems, under ex isting policies (first column of results), both private and social prof its are positive, indicating that farmers on Pohnpei have financial incentives to cont inue farming all three systems. Profits are greatest for the UPS ($1,412 ha yr) followed by the HG ($954 ha yr) and HGW ($906 ha yr), explaining why farmers continue to move into the upland areas to farm. Comparing the second column of results, even if a carbon market was established and farmers in the uplands were taxed 10% for th eir erosion impacts, the social profit values generated from the UPS are still greater than the HG and HGW indica ting that even with a 10% erosion tax, further expansion into the uplands is likely to continue. Since policy makers are interest ed in minimizing impacts of sakau planting in the uplands and profits are currently far greater for the UPS than for either the HG or HGW, policy intervention in the form of economic incentive/d isincentives may be necessary. The last three columns in Table 6-4 present how a range of subsidies on outputs from the two homegarden systems and erosion taxes on sakau in the upland system, are likely to influence both potential profits and farmer motivation for land use decisions.


95 First, consider a subsidy of HG and HGW in the absence of sakau erosion taxes. The subsidy rate necessary to persua de farmers to abandon upland activit ies would need to be at least 31%. At a rate of 31%, profits from all three systems would be essentially the same ($1,407 ha yr, $1,417 ha yr, and $1,412 ha yr for the HG, HGW, and UPS respectively). Under this policy, the government would also incu r a budgetary cost of between $398 ha yr for the HG farmers to $464 ha yr for the HGW farmers. Given the ex isting limitations on the government budget on Pohnpei, combined with increasing de mands to minimize government spending as a result of the new Compact agreemen t, this scenario is highly unlik ely to be feasible without the provision of overseas grants. A more practical alternative from a government budgetary perspective would be to assess a tax on sakau produced in the uplands. Under this scenario, the tax rate required to persuade farmers to abandon their upland activities is 27.5%, a value equivalent to a profit reduction of $507 ha yr for upland farmers (Table 6-4). Any ta x rates less than 27.5 % would result in greater profits for the UPS rela tive to the HG and HGW while a ra te higher than 27.5% would, in theory, provide enough of an incentive to pers uade farmers to adopt either of the two homegarden systems. If farmers still wanted to ma intain activities at or near the margin in the uplands, the government could generate revenue s from this policy that are above and beyond what is likely to be required to mitigate erosion impacts from these activities. The results presented above are based on prof its for a 1-year time period. We now turn to how these profits are influenced for two longer time periods, (1 ) a 5year time horizon and (2) a 10-year time horizon, testing the sustainability of these thr ee production systems. We noted earlier that anecdotal evidence suggests that while both of the homegarden systems produce a sustainable flow of outputs, sakau yields in th e uplands decline by about 10 % annually.


96 For the HG and HGW, a sustainable stream of private profits of $954 and $906 for 5 years respectively would produce a ne t present value of $3,616 ha and $3,436 ha if discounted at 10% (Table 6-5). The net present valu e of profits for the UPS is $4,213 ha using the same discount rate and time period but with production declining 10% annually2. Even taking into account the declining productivity of upland sakau over a five-year period, under the existing setting on Pohnpei, farmers in the uplands would still generate gr eater profits relative to the mother systems and therefore will continue to farm in these areas. Although during the one-year pe riod, establishing a carbon market and accounting for erosion costs does not change the order of prof its for each of the three systems, these policy interventions combined with the declining sakau production in the UPS results in the HG generating the greatest profits with the HGW and UPS profits differing by only $20 ha (Table 6-5). Thus, in trying to achieve the governme nts objective of persuading upland farmers to move down to the lowlands, corre cting for the existing market failures considered here would result in profits for all three systems being almo st equal if farmers operate under a 5 year time horizon and a 10% discount rate. If, alternatively, the government preferred a pr icing policy in the fo rm of a subsidy on homegarden outputs, the sustaine d output of crops in the two homegarden systems combined with a decline of sakau production would result in a subsidy rate far below that of the one-year period. Table 6-4 shows that a price subsidy on outputs from the HG and HGW systems of 10% would not provide a sufficient financial incentiv e for upland farmers to halt their activities. 2 For a comparison, the private real present value of $954 ha yr $906 ha yr and $1,412 ha yr incomes for the HG, HGW and UPS systems respectively discounted at 15% would be $3,197 ha $3,038 ha and $3,771 ha : di scounted at 5.0% private real present values would be $4,130 ha $3,924 ha and $4,748 ha for the HG, HGW, and UPS systems respectively.


97 However a subsidy rate of 12.5% or more would re sult in greater profits for the two homegarden systems relative to the upland system and in doing so, provide a financial justification for upland farmers to adopt either of the two homegarden systems. Under the subsidy policy discussed in Table 65, and in order to attract upland farmers to adopt either of the two homegarden systems, th e government would incur costs in the amount of between $485 ha and $602 ha over five years. While for a one-year period, the subsidy approach appears unlikely, payments based on a 12.5% subsidy over a five-year period may be more feasible. Indeed, on average, annual paym ents per hectare would range between $97 for the HG and $120 for the HGW. Considering a tax on sakau produced in the UPS, a rate of 10% would equate to private and social profits being equal at $3,634 ha (Table 6.5). Moreover, with a tax of only 10%, the UPS still generates greater profits than either the HG or HGW (in the absence of lowland subsidies). The impact of a 12.5% tax on sakau would result in the HG being more profitable than the UPS by a margin of $126 ha At the same time, the UPS would still generate greater profits relative to the HGW by a margin of $54 ha Thus in order to generate profits for both the HG and the HGW greater than the UPS, and th ereby inducing upland farmers to adopt either the HG or HGW a tax of 15% would be necessary equating to a reduction in UPS profits by $289 ha As more time is taken into c onsideration, the ecological sust ainability of these systems becomes much more pronounced. Table 6-6 shows that as a result of dec lining sakau yields in the UPS under the current policy setting, the ne t present values for both the HG and the HGW


98 systems becomes greater than for the UPS ($5,861 ha, $5,569 ha, and $5,358 ha respectively)3. For those farmers and policy makers interested in generating greater profits over the longterm, the fact that sakau yields decline at su ch a rate in the UPS would, in theory, provide enough of an impetus to convince more farmers to forgo planting in the upland areas. However, given the limited opportunities to generate inco me, both in the short and long-term time period, farmers may still be inclined to continue farm ing the uplands as a means to generate greater income in the immediate time frame. In order to further persuade farmers to leave the upland areas, it may be necessary for the government to provide additional incentives. Thus far we have examined profitability of three agroforestry systems based on actual dollar values. When comparing systems having identical commodities being produced, both the private and social profits are suitable indicator s of relative competitiven ess. When considering systems producing different outputs, or simila r commodities using differi ng technologies or in different ecological settings, it is us eful to examine the ratios (defin ed in Table 6-3) to ascertain the relative competitiveness of systems and the extent of policy divergences. Three such coefficients, the profitability coefficient (PC), the private cost ratio (PCR), and the domestic resource cost ratio (DECRY), derived from th e PAMs under the current policy setting (ten-year time period) are shown in Table 6-7. For the profitability coefficient, an indicator of the net transfer effects, it is evident that the UPS system, having a value > 1 indicates that farmers are receiving an implicit subsidy whereas farmers engaged in the two homegarden systems are 3 Once again for comparison, the private real pr esent value of $954 ha yr $906 ha yr 1, and $1,412 ha yr incomes for the HG, HGW and UPS systems respectively discounted at 15% would be $4,787 ha $4,549 ha and $4,596 ha : discounted at 5.0% private real present values would be $7,365 ha $6,998 ha and $6,370 ha for the HG, HGW, and UPS systems respectively.


99 implicitly being taxed due to the nonexistent carbon market. Moreover, the relative value of the impacts is greatest for the upland system, inferring that correcting for the lack of a carbon market would have less of an impact on the two homega rden systems than correcting for the external costs associated with erosion. There is less of a relative difference when considering the private cost ratio and domestic resource co st ratio for these three systems. Indeed, even though the UPS syst em has the lowest PCR and DRC ratios implying that for each unit of domestic resources used to produce sa kau in the uplands greater profits are realized, both of these ratios infer that domestic resources are being used to genera te positive profits both in private and social terms4. Discussion The Policy Analysis Matrix m ethodology has been applied here to determine the relative profits and use of domestic resources associated with three agroforestry systems on Pohnpei. In Pohnpei, agroforestry systems have traditionally played an importa nt role, primarily as the main source of food security, and to a lesser degr ee for commercial purposes. As the local economy continues to become more integr ated into the global economy, so have preferences for western goods and subsequently the desi re to increase household incomes for their purchase. To develop Pohnpeis economy, one such strategy that historically has been promoted by government officials involves taking advantage of Compact funds from the United States to inflate the public sector work force, there by providing employment to local citizens. The renegotiated Compact (Compact II) remains the primary source of funding supporting formal employment on Pohnpei, however, the amount of funding for this pur pose will decline as 4 A PCR or DRC value of 1 indicates that a marg inal increase in domestic resource use generate the exact same amount in value-added. Similarl y, a PCR/DRC < 1 reflects a relative efficiency value for the use of domestic resources. A smaller value indicates greater value-added per unit of domestic resources.

PAGE 100

100 Compact II comes to a close in 2023. Nonethele ss, the opportunities that Compact funds have provided, combined with relatively high wages paid for public sector jobs has led to a declining interest in agriculture on the part of the privat e sector, particularly fo r cash income needs. The United States mandated a smaller public se ctor as a condition of the newly negotiated Compact, which directly impacts the governments role as the largest employer on the island. With already very limited private sector opportu nities, local families are turning to the upland areas of Pohnpei establishing unsusta inable agricultural systems to increase income levels and/or replace prior income sources due to the loss of public sector jobs. To date these actions are generating income for local families. Yet, th ey are also negatively impacting the local environment, and in turn other natural res ources heavily depended upon by much of Pohnpeis population. In such situations where limited economic opportunities exist, private individuals often place a greater value on achieving short-term needs, even if it entails forgoing personal potential future benefits, or t hose that may be accrued by other me mbers of society. In contrast, policy makers are often more concerned with poten tial benefits that may be accrued to all of society. In this sense, a policy maker is often mo re inclined to promote agricultural systems that use available resources in economically efficient manner. Results presented here illustrate a dilemma often faced by policy makers, and now facing those on Pohnpei. On the one hand, a stated goal of the government is to promote increasing income by developing the agricultu ral sector. Based on our results, the upland system is most apt to do so in the short-term. On the other hand, the government is interested in economically efficient and ecologically sustainable use of limited lo cal resources. If this last objective is to be realized, it is critical to consider the profitabi lity over a longer time, particularly if a known unsustainable system is currently expanding.

PAGE 101

101 When considering the results of a one-ye ar and a five-year period under the current situation on Pohnpei, profits remain greater for the UPS than either of the two homegarden systems. Assuming the government intends to promote agricultural development farmers practicing the UPS interested in maximizing short-term profits will likely require policy intervention to adopt eith er of the two homegarden systems. In this chapter we have presented how different policy strategies may contribute to achieving this objective, namely (1) correcting for the current market failures, (2) estab lishing a subsidy for outputs derived from both homegarden systems, and (3) levying a tax on sa kau produced in the upland system. On the other hand, for those farmers interest ed in maintaining a steady in come over a longer time period, maintaining, or adopting either of the homegarden systems is likely to ensure such an outcome. In either case, policy makers can now be more informed about why farmers are engaged in their respective systems and what might be done to influence the future of Pohnpeis agricultural sector. Concluding Comments on the Polic y Analysis Matrix Methodology The Policy Analysis Matrix provides a m ean s by which the impacts of policy and market failures can be quantified. In doi ng so, it allows not only for th e quantification of individual policies, but also the quantifica tion of the aggregate impacts of all policies affecting farmers currently engaged in different ag ricultural practices. Having the ab ility to analyze individual as well as the total impacts of policies can provide an important function in settings where different public institutions have established policies in an uncoordinated fashion, all of which have a common objective, but that ultimately work against one another. In this regard, the PAM can aid in untangling policies unsu ccessful in their symbiotic goal. Anot her useful aspect of the PAM is its utility in analyzi ng the potential of agricultural t echnologies under consideration. Once developed, sensitivity analysis can easily be carried out allowing for the impacts of new or

PAGE 102

102 modified policies to be explore d. In addition, the PAM, while so mewhat complicated behind the scenes provides a fairly simple means to exam ine dynamics taking place at all levels of the production chain. This can be particularly impor tant to policy planners when attempting to identify where and how policy inte rvention is most likely to result in the desired outcome. The case study presented here provided a unique opportunity to a pply the PAM methodology within the context of agroforests over a multi-year time frame. This chapter also illustrated how the traditional PAM model could be expanded upon to in clude benefits derived from an ecological service even when a market for them is abse nt (i.e., correcting for the absence of a carbon market). In our case, carbon sequestration represents a benefit realized by society at large but results from an individuals actions Other externalities, be they positive or negative, could also be incorporated into the PAM framework. The di fficulty does not necessarily lie in determining how or where to include the cost or benefit of an externality within the PAM framework, but instead ascertaining relevant and accurate values to be used. One limitation of the PAM methodology stems from the frequent use of fixed input-output coefficients making it more difficult to determ ine the dynamic effects attributable to policy shifts. As noted above, one means to address this dilemma is to determin e price elasticities of supply and demand, as well as cross price elastic ities of demand allowing for estimates of how farmers may respond to various poli cy interventions. Although this may be the case, in settings such as Pohnpei where the cash-economy is still developing, commodity choices are limited and motives of farmers are influenced by traditional norms, reliable estimates of elasticities are often difficult to obtain and thereby may limit the util ity of PAMs under dynamic s cenarios. This is not to say that certain assumptions pertaining to various elasticities coul d not be made and the robustness of the results tested through sensitivity analysis Yet the PAM methodology was not

PAGE 103

103 developed as an ultimate means to derive perf ect results under all circum stances. Instead, it is a tool to empirically determine how public interv ention(s), or lack thereof, may influence the decisions of those participating in the agricultu ral sector and in turn, how those decisions may affect government resources.

PAGE 104

104 Table 6-1.Policy Analysis Matrix (PAM) framework Revenues Costs Output Tradable inputs Domestic factors Profits Private prices A B C D Social prices E F G H Effects of policy distortions and divergences I J K L Private profits D = (A-B-C) Social profits H = (E-F-G) Output transfers I = (A-E) Tr adable input transfers J = (B-F) Factor transfers K = ( C-G) Net transfers L = (I-J-K) or (D-H) Table 6-2.Policy Analysis Matrix ex panded to incorporate externalities Revenues Costs Output Tradable inputs Domestic factors Profits Private prices1 A1 B1 C1 D1 Private prices2 A2 B2 C2 D2 Social prices1 E1 F1 G1 H1 Social prices E2 F2 G2 H2 Effects of policy distortions and divergences I J K L Table 6-3. Coefficients derived from the Policy Analysis Matrix used to measure the relative impact(s) of policies or market failures NPC = A/E: a ratio that contrasts private and so cial output values. A NP C > 1 is indicative of private prices of output being greater than social prices reflecting that producers are positively protected. EPC = (A-B)/(E-F): the ratio of value added measured at privat e vs. social prices. Unlike the NPC and the NPI that measure the effect of di vergences for output and tradable input prices respectively, the EPC measures the total effects of intervention in both markets. The implication of EPC > 1 is that there is an overall artificial incentive to produce a commodity due to the presence or absence of policies. PCR = C/(A-B): the ratio of domes tic factor prices to value added at private prices. It identifies the cost of domestic resources in private prices necessary to produce a unit of value added. A PCR between 0 and 1 indicates that in private terms, domestic reso urces generate more than their value in value added. DRC = G/(E-F): the ratio of domes tic factor prices to value added at social prices. It shows the cost of domestic resources in so cial prices needed to produce a unit of value added. If the DRC > 1 the commodity system is not desirable from an economic efficiency standpoint. PC = D/H: measures the incentive effect of al l policies and provides a ratio to determine the relative net policy transfers. A PC = 1 indicates no net transfers Source: Monke and Pearson (1989)

PAGE 105

105 Table 6-4. Estimates of private and social prof itability calculated using the Policy Analysis Matrix (PAM) methodology, for a one-y ear period over the three dominant agroforestry land-use syst ems in Pohnpei, Federated States of Micronesia Policy scenarios Farming system Existing: no carbon market (US $ ha yr) 10.0% output price subsidy (US $ ha yr) 15.0% output price subsidy (US $ ha yr) 31.0% output price subsidy (US $ ha yr) Home Garden Private Social Divergence $954 $1,009 -$55 $1,100 $1,009 $91 $1,173 $1,009 $164 $1,407 $1,009 $398 Home Garden Wetland Private Social Divergence $906 $953 -$47 $1,071 $953 $118 $1,153 $953 $200 $1,417 $953 $464 Upland System Private Social Divergence $1,412 $1,229 $183 $1,229 $1,229 $0.0 $1,138 $1,229 -$91 $906 $1,229 -$324

PAGE 106

106 Table 6-5. Estimates of five-year net present valu es based on a 10% discount rate for private and social profitability calculated using the Policy Analysis Matrix (PAM) methodology, for the three dominant agroforestry land-use systems in Pohnpei, FSM Policy scenarios Farming system Existing: no carbon market (US $ ha yr) 10.0% output price subsidy (US $ ha yr) 12.5 % output price subsidy (US $ ha yr) 15.0% output price subsidy (US $ ha yr) Home Garden Private Social Divergence $3,616 $3,824 -$208 $4,170 $3,824 $346 $4,309 $3,824 $485 $4,448 $3,824 $623 Home Garden Wetland Private Social Divergence $3,436 $3,614 -$178 $4,060 $3,614 $446 $4,216 $3,614 $602 $4,372 $3,614 $759 Upland System Private Social Divergence $4,213 $3,634 $578 $3,634 $3,634 $0.0 $3,490 $3,634 -$145 $3,345 $3,634 -$289

PAGE 107

107 Table 6.6 Estimates of ten-year net present valu es based on a 10% discount rate for private and social profitability calculated using the Policy Analysis Matrix methodology, for the three dominant agroforestry land-use systems in Pohnpei, FSM Policy scenarios Farming system Existing: no carbon market (US $ ha yr) Existing: with erosion costs not internalized (US $ ha) Home Garden Private Social Divergence $5,861 $6,199 -$338 Home Garden Wetland Private Social Divergence $5,569 $5,852 -$283 Upland System Private Social Divergence $5,358 $4,567 $791 Table 6-7. Ratio indicators derived from Policy Analyses Matrices for three dominant agroforestry land-use systems for a ten-ye ar period discounted at 10%, in Pohnpei, FSM Ratio Indicators HomegardenHomegarden Wetland Upland Sakau Profitability coefficient (PC) 0.960.971.11 Private cost ratio (PCR) 0.350.450.35 Domestic resource cost ratio (DRC) 0.340.440.38

PAGE 108

108 CHAPTER 7 DISSERTATION SUMMARY The populations of m any Small Island Developi ng States (SIDS) ha ve a long history of reliance on natural resources particularly agriculture and fish eries, for their survival. An indepth understanding of SIDS conti nued reliance on agriculture a nd other natural resources is important for formulating development strategies The research conducted for this dissertation illustrated how traditional agriculture is fundament ally intertwined in the lives of people living on Pohnpei, an island state within the Fe derated States of Micronesia (FSM). This research centered on quantifying the productivity, value, and use of Pohnpeis agroforests at the household level. Through the ex trapolation of household-level data, the total productivity and value of Pohnpeis agroforests to the islands economy were determined. In turn, these same data were used to assess th e influence of market access and employment on household income from agriculture. Household-level data also provided the basi s for understanding the role of agriculture within Pohnpeaen livelihoods. Utilizing ethnogra phic linear programming, livelihood systems were analyzed to establish how different livelih ood strategies employing an array of activities can be employed to achieve cash incomes for each of the livelihood systems. The analysis was extended to simulate the effects of different crop-pricing polic ies and of market expansion on incomes for a broader group of households. Household data were also used to quantif y the productivity per unit area of different agroforestry systems on Pohnpei. Using the Policy Analysis Matrix methodology, production values were then used to calculate profits per unit area for the different systems. More specifically, profits for farmers in the presence of market failures (p rivate profitability) and in the absence of market failures (social profits) were de termined. This analysis not only allows one to

PAGE 109

109 measure the tradeoffs of using different agrofore stry systems, it also provides insight into how farmer behavior might change in response to an introduction of a new technology and/or a changing policy environment. Such an analysis can be extended to assess private and social profits at a state or national scale and to a ssess the tradeoffs associated with policies. Pohnpeaen households derived benefits from th eir agroforests equal to nearly half the islands median household income. The majority of these values were derived from homegardens and were based on consumptive values signi fying both the importance of homegarden agroforests and their contribu tion to food security. Moreover, in 2002 agroforests provided between 6%and 72% of total household cash income, depending on whether households have access to off-farm employment, and the extent to which they are involved with fishing and/or receiving remittance and retirement payments. During the same period, the value of Pohnpeis agroforests at the island-wide scale was more than $16 million or nearly 20% of Pohnpeis annual GDP. Greater market access could increas e cash income for some segments of the population but would have only limited impacts on those livelihood systems reliant primarily on fish and/or remittances for cash income. While increased market access could expand opportunities for commercial agriculture, off-farm employment significantly reduces reliance on agriculture for cash income. Given the long history of agroforestry on Pohnpei and the fact that agroforestry is still a dominant component of the island s livelihoods, it is l ogical that agroforest ry should remain at the forefront of future development strategi es. Although there are geog raphic, cultural, and economic differences among the other three stat es constituting the Federated States of Micronesia (Yap, Chuuk and Kosrae), livelihood strategies and the contribution that agroforests make to the populations of these islands ar e similar to those found on Pohnpei. As such,

PAGE 110

110 development strategies, whether infrastruct ural improvement, expansion of off-farm employment, market expansion or changes in a crop-pricing policy, would will have similar ramifications on the livelihoods throughout the FSM. Livelihoods in Pohnpei and that of the FSM which, in te rms of per capita GDP, lies midway along the continuum of SIDS worldwide center around agriculture and the public sector. However, this is not necessarily the case for all other SIDS. Indeed, in many other regions of the world, SIDS that are economically advanced may pursue different development strategies. In general, agriculture plays only a minor part in the livelihoods in SIDS that are at the upper end of the development continuum (e.g., the British Virgin Islands). On the other end of the development continuum, other SIDS remain highl y dependent on other sources of income (e.g., Kiribati). Although all SIDS must pursue an in tegrated approach fo r achieving sustainable development, each one has a unique pattern of traditional resource use and contemporary livelihood systems. Once such an understandi ng of distinct livelihood systems are realized, future development will be more informed and ap t to focus on programs more appropriate for the local setting. The approach employed for this research was based on the need to meet specific research goals as they related to SIDS in the western Pacific. The approach builds on the foundation of household data to analyze livelihood system and strategies and to se rve as a basis for profitability analysis, and it has applicability to SIDS worldwide. Critical to the success of implementing such an approach elsewhere is th e recognition that in al l settings there will be unique and, often challenging circumstances that require flexibilit y and an adaptive approach in order to achieve research objectives.

PAGE 111

111 Future research needs : This study quantified the value a nd use of agroforestry products on Pohnpei based on seven of the most commonly cultivated crops. However, past research conducted throughout the Pacific has documente d the cultivation and use of many more agroforestry products throughout the region. The estimation of the total value of agroforests on Pohnpei can therefore be consider ed conservative. Building on this study and relevant past research could be a first step in fully accounting for the total va lue of Pohnpeis agroforests. The simulations conducted using ethnographic linear programming provided insight into the impacts of household income relative to changing policy and mark et conditions. However additional insight could be gain ed by simulating technological ch anges such as the introduction of new crops and/or increasing production of existing crops through the use of either improved crop varieties or the use of synt hetic inputs such as fertilizers. Before doing so, research to determine potential marketing opportunities would need to gauge additional demand for existing and/or new crops. Furthermore, research conducte d in consultation with farmers and extension agents to determine which crops might be appropriate, given the biophysical and cultural setting, would be necessary. Once poten tial technological changes have been identified, using the ethnographic linear programming m odels developed for this rese arch along with the Policy Analysis Matrix methodology, further analyses coul d be conducted to determine the potential for profitability and adoptio n of those technologies. Further research is needed on identifying to what extent, if any overseas aid aimed primarily at economic development is contri buting to the conserva tion of Small Island Developing States natural resources. Within th e FSM, overseas aid in th e form of Compact II funding is largely responsible for maintaining formal employment. The present study has shown that households having members formally employed derive the vast majority of their income

PAGE 112

112 from employment earnings. Additionally, thes e same households have shown less of a dependence on agricultural and fish ery resources relative to other livelihood systems. Therefore it could be inferred that Compact II funding is helping to reduce pre ssure on Pohnpeis limited and fragile resource base for economic gain. As Compact II funding continue s to decline in the coming years, it will be important to determine if similar trends associ ated with the relative dependence on employment and natural resources change, and, if so, in what direction. Understanding how overseas aid contributes to local economies and the relative dependence on natural resources is necessary to evaluate the potential tradeoffs of reducing aid packages and the need to ensure the health and viabil ity of an islands natural resources.

PAGE 113

113 APPENDIX A HOUSEHOLD SURVEY Household name: Kousapw: Interviewee age: Age of house: Number of adults (18 yrs. or older): Male: 1 2 3 4 5 6 7 8 9 10 Female: 1 2 3 4 5 6 7 8 9 10 Education grade completed for each adult: Grade for each male and female Males Females Number of children between ages 8-18 years: Male: 0 1 2 3 4 5 6 7 8 9 10 Female: 0 1 2 3 4 5 6 7 8 9 10 Number of children younger than 8 yrs: Male: 0 1 2 3 4 5 6 7 8 9 10 Female: 0 1 2 3 4 5 6 7 8 9 10 Number of children attending school Monday Friday: 0 1 2 3 4 5 6 7 8 9 10 Number of land parcels owned by members of households: 0 1 2 3 4 5 Name of land parcel where house is located: ______________________________ Name of each parcel owned by a member of the household not including where house is located: Parcel name Kousapw Municipality Agroforests? Y or N 1. 2. 3. Number of agroforests used by family members but not owned? 0 1 2 3 4 5 Parcel name Kousapw Municipality 1. 2. Are there any crops that you plant only in the home garden plot and not in other agroforests? Y N Plant name Reason/Why 1. 2. Number of Livestock: Livestock type Number per household Sell ? Y or N Chickens Pigs Dogs Other How much time do you spend feeding pigs each day?

PAGE 114

114 Who in the family feeds the pigs most often? Adult male Adult female Male child Female child Do you purchase any items that you use for your farm? Input name Type purchased Estimated weekly use (quantity) Fertilizer Seedlings Pig feed Other What is the main method of cooking? Fire Kerosene Electric stove Um If fire, where do you obtain most of your fuel wood from? Purchase Home garden Agroforests Mangroves Other Do your agroforests lands produce enough food for your household needs? Y N If no, what more would you like to be able to produce in your agroforest? Plant name/type Purpose Reason why not currently producing crop/plant 1. 2. How important is your home garden to your household? 1 2 3 4 NA For what reason(s)? Food Prestige Income Other How important are your other agroforests to your household? 1 2 3 4 NA For what reason(s)? Food Prestige Income Other What is your future plan for your land (s)? What are the biggest constraints to farming for your family? Labor Money Land Soil fertility Markets Transportation Does anyone in this household own a business? Y N Do you hire employees outside of family? Y N Food shop Gas station Sakau bar Other Number employed Hours per week Hourly wage Number of adults employed or retired: 0 1 2 3 4 5 6 Number employed part time: 0 1 2 3 4 5 Govt Private Family business Retired Govt Private Family business Male Female Approximate total monthly income from employment or retirement: Employment: $ Retirement:$ Do any family members fish? Y N Sell? Y N Estimated number of pounds sold per month: Sell agriculture or livestock products? Y N Estimated monthly income from selling ag. products? $

PAGE 115

115 Sell handicrafts? Y N Type: Estimated monthly income from selling handicrafts? $ Number of immediate family members living off island: 0 1 2 3 4 5 6 7 8 9 10 FSM Guam Hawaii Mainland Other Estimated amount of remittances received per month: $ What are the five most important sources of cash income? 1. 2. 3. 4. 5. Rank your familys top five monthly expenses? 1. 2. 3. 4. 5. What is the approximate monthly cost for each of the top five expenses? 1. 2. 3. 4. 5. What, if any, are your familys major cr edit sources (shops, banks, private etc.)?

PAGE 116


PAGE 117

117 Household name Kousapw Mark Drew's home phone: 320-7539 Land parcel name Date: Number of adults working on farm(18 yrs+) Number of children working on farm Planting activities Total hours spent for: Hours spent on farm Male FemaleMale Female Plant name Variety Number planted Sakau Bananas Yams

PAGE 118

118 Date: Products harvested Amount harvested Amount of pounds or number of each crop used for each purpose If sold Name Variety Pounds Number Household consumption Give to other family/friends Funeral Sell to market Traditional ceremony Animal feed Trade Market name Selling price per product or pound

PAGE 119


PAGE 120

120Activities>AFHomegardenLAND2Sell YamsYam CerimonialYam fam/friends Yam trn.Sell brdfruitbrdfruit Cerimonialbrdfruit fam/friends brdfruit trn.Sell GTaroGTaro CerimonialGTaro fam/friendsGTaro trnSell CoconutsCoconuts fam/frien d Coconuts trnSell Bnut Bnut CerimonialBnut fam/friendsBnut trn Sell SakauSakau CerimonialSakau fam/frinedsAF Sakau trnSell BananaB anana CerimonialBanana fam/frinedsBanana trnM-FT employmentM-PT employmentF-FT employmentF-PT employmentremittancesFish Sell Fis h Fish trn Cash trn. Variables> 3.00002.9894 8.755206.00010.000115.0000 0.2766 50.0000 0.0000 384.0000 0.0000 0.0000 0.0000 6.0000 0.0000 36.00002044.0000 0.0000 20.0000 0.0000 245.0000 0.0000 81.0000 0.0000 416.0000 0.0000 10.0000 18.0000 735.0000 0.0000 0.0000 0.0000 7.5356 0.0000 0.0000 0.0000 0.00Resou rce Use Constraint Unit hectarehectarekg kg kg kg kg kg kg kg kg kg kg kg # # # # # # # kg kg kg kg kg kg kg kg month month month month $/month day kg kg $ RHSConstraint HG land Hectares 1 3.00<= 3HG land Land2 Hectares 1 2.99<= 3Land2 Adult male laborm-hours 69 56 0 0 134 80 7 0 374.40<= 6240 Adult male labor Adult female laborF-hours 4 1 0 0 134 60 04 6 7 1 2 < = 4160 Adult female labor Child male laborCM-hours 82 119 601.73<= 1664 Child male labor Child female laborCF-hours 2 1 0 0 08 9 9 < = 1664 Child female labor Yam acct. kg -87 -23 1 1 1 0 0.00<= 0Yam acct. Yam consumptionkg 0 0 -1 0 -115.00<=-115Yam consumption Yam Cerimonialkg -1 -206.00<=-206Yam Cerimonial Yam fam/friendskg -1 -10.00<=-10Yam fam/friends brdfruit acct kg -77 -68 11 1 0.00<= brdfruit acct brdfruit consumptionlkg -1 -384.00<=-384brdfruit consumption brdfruit Cerimonialkg -1 -50.00<=-50brdfruit Cerimonial brdfruit fam/friendskg -1 0.00<=0brdfruit fam/friends GTaro acct kg -2 0 11 1 0.00<= GTaro acct GTaro consumptionkg -1 -6.00<=-6GTaro consumption GTaro Cerminialkg -1 0.00<= GTaro Cerminial GTaro fam/friendskg -1 0.00<= GTaro fam/friends Coconuts acct# -477 -278 11 -218.04<= Coconuts acct Coconuts consumption# -1 -2044.00<=-2044Coconuts consumption Coconuts fam/friends# -1 -36.00<=-36Coconuts fam/friends Bnut acct # -1499 -1406 11 1 -8435.04<= Bnut acct Bnut consumption# -1 -245.00<=-245Bnut consumption Bnut Cerimonial# -1 -20.00<=-20Bnut Cerimonial Bnut fam/friends # -1 0.00<= Bnut fam/friends Sakau acct kg -72 -94 11 1 0.00<= Sakau acct Sakau consumptionkg -1 -416.00<=-416Sakau consumption Sakau Cerimonialkg -1 -81.00<=-81Sakau Cerimonial Sakau fam/frinedskg -1 0.00<= Sakau fam/frineds Banana acct kg -118 -138 11 1 -21.53<= Banana acct Banana consumptionkg -1 735.00<=-735Banana consumption Banana Cerimonialkg -1 -10.00<=-10Banana Cerimonial Banana fam/frinedkg -1 -18.00<=-18Banana fam/frined Fish acct kg -13 1 1 0.00<= Fish acct Fish consumptionkg -1 0.00<= Fish consumption Fishing constraintkg -1 0.00<= Fishing constraint cash beg. Season$US 0 0 0 10 0 0 < = 100 cash beg. Season Total cash end of seaso n $US 0 2 0 0.22 0 0.77 0 0.1 0 0.05 2.2 0 0.55 250 240 360 300 85 -5 1.98 1 2280.00>= 2280 Total cash end of season

PAGE 121

121 APPENDIX D MATHEMATICAL EXPLANATION OF LINEAR PROGRAMMING Linear programm ing can be stated mathem atically as (taken from Hildebrand et al., 2008): Max (or Min): = jCjXj (j = 1 n) Subject to: iAijXj <= Ri (i = 1 m) and Xi >= 0 is the variable objective to be minimized or maximized, Cj is the cost (debit) or returns (credit) of each of the n activities Xj, Aij is the set of input or output coefficients for each activity j and resource or constraint i, and Ri is the set of m minimum or maximum constraints or restrictions. The program is set up in a matrix (such as Excel and solved with the add-in solver). The resources in this simple matrix are land, labor and beginning cash. The activities are the production of maize, hay and goats. Inside the matrix are the input or output coefficient, and the objective to be maximized, is cash income.

PAGE 122

122 LIST OF REFERENCES ADB. (1992) Environm ent and Development: Paci fic Island Perspective. Asian Development Bank, Manila. ADB. (1996) Pohnpeis Watershed Spatial Plan a nd Management Guidelines. Prepared for the Asian Development Bank T. A. No. FSM-1925: Watershed Management and Environment. Adesina, A.A. and Coulibaly, O.N. (1998) Polic y and competitiveness of agroforestry-based technologies for maize production in Cameroon: An application of policy analysis matrix. Agricultural Economics 19: 1-13. Alavalapti, J. R. R., Shrestha, R. K., Stainb ack, G.A., and Matta, J.R. (2004) Agroforestry development: An environmen tal economic perspective. Agroforestry Systems 61: 299-310 Barrau, J. (1961) Subsistence Agriculture in Polynesia and Micronesia. Bishop Museum Bulletin 223. Honolulu, HI. Bascom, W. (1963) Ponape: A Pacific Economy in Transition. Anthropological Records V 22. University of California Press, Berkeley and Los Angeles, CA Bertram, I.G. and Watters, R.F. (1985) The MIRAB economy in South Pacific microstates. Pacific Viewpoint 26(3): 497-519 Black, C. and Ong, C. (2000) Utilisation of light and water in tropical agriculture. Agriculture and Forest Meteorology 104: 25-47. Boserup, E. (1965) The Conditions of Agricultural Gr owth: Economics of Agrarian Change under Population Pressure Earthscan Publications, London. Brookfield, H.C. (1972) Intensification and disintensification in Pacific agriculture. Pacific Viewpoint 13:30-41 Clarke, W.C. and Thaman, R.R. (1993) Agroforestry in the Paci fic Islands; systems for sustainability. New York: United Nations University Press,. Current, D., Lutz, E. and Scherr, S. (eds). ( 1995) Costs, Benefits, a nd Farmer Adoption of Agroforestry: Project experience in Centra l America and the Caribbean. A CATIE-IFPRIWorld Bank Project. World Bank Environmental Paper No. 14. Drew, W.M., Ewel, K.C., Naylor, R.L. and Sigrah, A. (2005) A tropical freshwater wetland: III. Direct use values and other goods and services. Wetlands Ecology and Management 13: 685-693 Ewel, J.J. (1999) Natural systems as models for the design of sustainable systems of land use. Agroforestry Systems 45: 1-21.

PAGE 123

123 FAO. (1994) The Role of Agriculture, Forestry a nd Fisheries in the Sustainable Development of Small Island Developing States. Food and Agricultural Organization, Rome. FAO. (1999) Status of environment and natural resources in Small Isla nd Developing States. Sustainable Development Department, F ood and Agricultural Organization, Rome. FAO. (2004) FAO and SIDS: Challenges and emerging issues in agriculture, forestry and fisheries. Prepared by FAO on the Occasion of the Inter-Regional Conference of Small Island Developing States, Bahamas Jan. 26-30th, 2004. FAO (2005) Small Island Developing States, ag ricultural production and trade, policy and preferences. Commodities a nd Trade Technical Paper No. 7. Food and Agricultural Organization, Rome. Falanruw, M.V.C. (1993) Micronesian Agroforestry : Evidence from the past, implications for the future. Proceedings of the workshop on rese arch methodologies and applications for Pacific Islands agroforestry; July 16-20, 1990: Kolonia, Pohnpei, Federated States of Micronesia. Gen. Tech. Rep. PSA-GTR-140. Paci fic southwest Resear ch Station, Forest Service, U.S. Department of Agriculture. Fernandes, E. C. M. and Nair, P. K.R. (1986) An evaluation of the structure and function of tropical homegardens. Agroforestry Systems 21: 279-301. Franzel, S., Coe, R., Cooper, P., Place, F., Sche rr, S.J. (2001) Assessing th e adoption potential of agroforestry practices in sub-Sharan. Agricultural Systems. 69: 37-62. Garrity, D.P. (2004) Agroforestry and the achieve ment of the Millennium Development Goals. Agroforestry Systems 61: 1-17 Gillet, R. and Lightfoot, C. (2001) The contribu tion of fisheries to the economies of Pacific Island countries: A Report Prepared for the Asian Development Bank, the Forum Fishery Agency and the World Bank. Government of the Federated States of Mi cronesia. (1999) Proceedings of the 2nd FSM Economic Summit, Federated States of Micronesia. September 13-17, 1999. Government of the Federated States of Microne sia. (2002) Federated States of Micronesia 2000 Population and Housing Census Report. Department of Economic Affairs, FSM National Government, Palikir, Pohnpei, Fede rated States of Micronesia. Haun, A. (1984) Prehistoric subsistence, population, and socio-political evolution on Ponape, Micronesia PhD dissertation. University of Oregon. Hazel, F. X. (1993) Culture in crisis: trends in the Pacific today. Micronesia Counselor #10 Hildebrand, P.E., Sullivan, A. J., Breuer, N.E., and Cabrera, V.E. (2008) Modeling diverse livelihood strategies in small farm liv elihood systems using ethnographic linear programming. Human Values and Agriculture. Manuscript submitted

PAGE 124

124 IPCC. 2001. Small Island States. IPCC Third Assessment Report-Climate Change (2001) Impacts, Adaptation and vulnerability Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press Johnston, B.F. (1970) Agriculture and structural transformation in developing countries: a survey of research. Journal of Economic Literature 8 (2):369-404 Khamzina, A., Lamers, J.P.A., Worbes, M., Botm an, E., and Vleck, P.L.G. (2006) Assessing the potential of trees for affore station of degraded landscapes in the Aral Sea Basin of Uzbekistan. Agroforestry Systems 66: 121-149 Kidd, C.V. and Pimentel D. (eds). (1992) Inte grated resource management: Agroforestry for development. London: Academic Press Inc. Kostka, M. and Raynor, W.C. (2000) Pohnpei Forest Monitoring Report. The Nature Conservancy, Pohnpei. 10pp. Kumar, B.M. and Nair, P.K.R (2004) Th e enigma of tropical homegardens. Agroforestry Systems 61: 135-152 Lafroy, E.C., Hobbs, R.J., O'Connor, M.H. and Pate, J.S. (eds). (1999) What can agriculture learn from natural ecosystems? Agroforestry Systems 45: 426-436. Laird, W.E. (1987) Soil survey of Island of Pona pe, Federated States of Micronesia. USDA Soil Conservation Service. Leboy, V., Merlin, M., and Lindstrom, L. (1992). Kava: The Pacific drug New Haven, CT: Yale University Press Manner, H.I. (199) A review of traditional ag roforestry in Micronesia. Proceedings of the workshop on research methodologies and applicati ons for Pacific Islands agroforestry; July 16-20, 1900: Kolonia, Pohnpei, Federated Stat es of Micronesia. Gen. Tech. Rep. PSAGTR-140. Pacific southwest Research Station, Forest Service, U.S. Department of Agriculture. Mercer, D.E. and Miller, R.P. (1997) Socioeconomic research in agroforestry: progress, prospects, and priorities. Agroforestry Systems 38: 177-193. Merlin, M., Jano, D., Raynor, W.C, Keene, T., Juvik, J., and Sebastian, B. (1992) Tuhke en Pohnpei: Plants of Pohnpei. Environment and Policy Institute, East-West Center, Honolulu, Hawaii. Merlin, M and Raynor, W.C. (2005) Kava cu ltivation, native species conservation, and integrated watershed resource management on Pohnpei island. Pacific Science 59 (2): 241260 Monke, E.A. and Pearson, S.R. (1989) The Policy Analysis Matrix for Agricultural Development Ithaca, NY; Cornell University Press.

PAGE 125

125 Montagnini, F. and Nair P.K.R. (2004) Carbon sequestration: an underexploited environmental benefit of agroforestry systems. Agroforestry Systems 61: 281-295 Montambault, J.R and Alavalapati, J.R.R. (2005) Socioeconomic research in agroforestry: A decade in review. Agroforestry Systems 65: 151-161 National Oceanic and Atmospheric Administrati on. (1987) Local climatologically data: annual summary with comparative data: Pohnpei, Eastern Caroline Islands, Pacific. NOAA, National Climatic Data Center, Asheville, North Carolina. Nair, P.K.R. (1993) An Introduction to Agroforestry Kluwer Academic Press, in cooperation with ICRAF. Nair, P.K.R. (1998) Directions in tropical agroforestry resear ch: past, present, and future. Agroforestry Systems 38: 223-245. Nair, P.K.R. (2000) Do tropical homegardens elude science, or is it the other way round? Retrieved on July 15th, 2007, from Website: Nelson, G.C. and Panggabean M. (1991) The costs of Indonesia sugar policy: A policy analysis matrix approach. American Journal of Agricultural Economics Aug : 703-712. Oelbermann, M., Voroney, R.P., and Gordon, A.M. (2004) Carbon sequestra tion in tropical and temperate agroforestry systems: A review w ith examples from Costa Rica and Southern Canada. Agriculture, Ecosystems and Environment. 104(3): 359-377 Office of the Governor. (1996) National Census Report, 1994 FSM Census of Population and Housing. National Government, Federated States of Micronesia. Palikir, Pohnpei, Federated States of Micronesia. Osman, W. M. (1995) Federated States of Micronesia Economic Report. Bank of Hawaii, Honolulu, Hawaii, USA. Pohnpei Statistical Office. (1998) Gross Domestic Product, Pohnpei State 1996: Conceptual framework, methodologies and results. Federated States of Micronesia. Paulson, D. and Rogers, S. (1997) Maintaining subsistence security in Western Samoa. Geoforum 28 (2): 173-187. Pearson, S. R., Falcon, W. P., Heytens, P., Monke, E. and Naylor, R. (1991) Rice Policy in Indonesia Ithaca, NY: Cornell University Press, Peoples, James G. (1985) Islands in Trust, Culture Change and Dependence in a Micronesia Economy Boulder, CO: Westview Press.

PAGE 126

126 Pimentel, D. and Wightman, A. (1999) Economic and environmental benefits of agroforestry in food and fuelwood production. In L. E. Buck, J. P. Lassoie and E. C. M. Fernandes (eds.). Agroforestry in sustainable agriculture systems, pp 295-317. New York: Lewis Publishers. Pohnpei Office of Agriculture and Forestry. ( 1996) Five-Year Strate gic Action Plan (19972001). University of Oregon Micronesia and South Pacific Program and the Pohnpei Office of Agriculture and Forestry, Federated States of Micronesia. Powell, P.T. (1998) Traditional production, communal land tenure, and policies for environmental preservation in the South Pacific. Ecological Economics 24: 89-101 Ragone, D., Lorence, D.H., and Flynn, T. (2001) History of plant intr oductions to Pohnpei, Micronesia and the role of th e Pohnpei Agricultural Station. Economic Botany 55 (2): 290324. Rao M. R., Nair, P.K.R., and Ong, C.K. (1998) Bi ophysical interactions in tropical agroforestry systems. Agroforestry Systems 38: 3-50. Raynor, W. C. (1989) Structure, Production, and Seasonalit y in an Indigenous Pacific Island Agroforestry System: A Case Example on Pohnpei Island, FSM Thesis, Department of Agronomy and Soil Science, Univ ersity of Hawaii, Honolulu, HI. Raynor, W.C. (1991) Agroforestry Systems in Pohnpei Practices and Strategies for Development. RAS/86/036, Field Document 4. South Pacific Forestry Development Program. United Nations Development Program. Raynor, W.C. and Fownes, J.H. (1991a) Indigenous agroforestry of Pohnpei: 1. Plant species and cultivars. Agroforestry Systems 16: 139-157. Raynor, W. C. and Fownes, J. H. (1991b) Indi genous agroforestry of Pohnpei: 2. Spatial and successional vegetation patterns. Agroforestry Systems 16: 159-165. Reardon, T., Barrett, C.B., Kelly, V., and Savadogo, K. (2001) Sustainable versus unsustainable agricultural intensif ication in Africa: focus on policy reforms and market conditions In: D. R. Lee and C.B. Barrett (eds), Tradeoffs or Synergies? Agricultural Intensification, Economic Development and the Environment Wallingford, UK: CABI Publishing. Sanchez, P. A. (1995) Science in Agroforestry. Agroforestry Systems 30: 5-55. Sanchez, P.A. (1999) Delivering on the Promise of Agroforestry. Environment, Development and Sustainability 1: 275-284. Staal, S. J. (1996) Periurban dairying and pub lic policy in Ethiopia and Kenya: A comparative economic and institutional analysis. Dissertati on, University of Florida, Gainesville, FL. Storey, D and Murray, W. E. (2001) Dilemmas of development in Oceania: The political economy of the Tongan agro-export sector. The Geographical Journal 167 (4): 291-304

PAGE 127

127 Sullivan, G. M., Huke, S. M. and Fox, J. M. (eds). (1992) Financial and Economic Analysis of Agroforestry Systems: Proceedings of a workshop held in Honolulu, Hawaii USA, July 1991. Nitrogen Fixing Tree Assoc., Paia, HI. Symons, L. (1978) Agricultural Geography third ed. London: Bell and Hyman Limited Thaman, R. (1975) The Tongan Agricultural System: With special emphasis on plant assemblages. Ph.D Thesis, Universi ty of California, Los Angeles. Thaman, R, Elevitch, C. and Wilkinson, K. (2000) Protecting and Expanding Traditional Agroforests in the Pacific. The Overstory # 51 Turkelboom, F., Van, K.K., Ongpraseri, S., Sutigoolabud, P. and Pelletier, J. (1996) The changing landscape of the Northe rn Thai hills: adaptive st rategies to increasing land pressure. Montane Mainland Sout heast Asia in Transition Chang Mai University, Thailand. United Nations. (1994) Report of the Global Co nference on the Sustainable Development of Small Island Developing States. Brid egetown, Barbados, April 25-May 6, 1994. United Nations. (2003) World Stat istic Pocketbook: Small Island Developing States. Department of Economic Affairs Statistics Division. Series V No. 24 SIDS United Nations. (2005) Report of th e International Meeting to Revi ew the Implementation of the Programme of Action for the Sustainable Deve lopment of Small Island Developing States. Port Louis, Mauritius, January 10-14, 2005. U.S. Department of the Interi or. (2004) Comments presented at FSM Economic Summit, March. Washington D.C. Versteeg, M.N., Amadju, F., Eteka, A., Goga n, A., Koudokpon, V. (1998) Farmers adoptability of Mucuna fallowing and agroforestry techno logies in the Coasta l Savanna of Benin. Agricultural Systems 56:269-287. Visser, S. (1980) Technologica l change and the spatial stru cture of agriculture. Economic Geography 56:311-319. von Thunen, J.H. (1826) Der Isolierte St aat in Beziehung auf Landwirtschaft und Nationalekonomie. Stuttgart Wang, X. and Feng, Z. (1995) Atmo spheric carbon sequest ration through agroforestry in China. Energy 20(2): 117-121. Wolff, D. (2001) Underlying causes of deforestati on and forest degradati on-Oceania and Pacific: Deforestation and forest degrada tion in the Kingdom of Tonga. World Rainforest Movements Bulletin No. 45 World Bank (2008) as visited in April 14, 2008.

PAGE 128

128 BIOGRAPHICAL SKETCH Mark Drew curren tly serves as California Tr outs Eastern Sierra program manager. Before joining California Trout, Mark worked as Th e Nature Conservancys (TNCs) Southeastern Caribbean Programs protected areas specialist. As the protected areas specialist, Mark managed TNCs conservation programs in the US and British Virgin Islands. Before working for The Nature Conservancy, Mark spent almost 15 years involved, both academically and professionally, in resource conservation in the US mainland, Oceania, and the Asia region. Mark holds a B.S. degree in forestry and natural res ources management from California Polytechnic State University and a M.A. degree in inte rnational development policy from Stanford University. Mark is now in the fi nal stages of obtaining his Ph.D. at the University of Florida (in forest resources and conservation) with an em phasis on developing conser vation strategies in Micronesia. Mark spent 3 years as a Peace Corps volunteer in th e Cook Islands and Nepal where he implemented extensive conservation program s, aided in the development of resource management plans, and conducted numerous co mmunity-based projects. After completing his M.A. degree at Stanford, Mark worked for Stanfords Institute for International Studies for 3 years, designing and conducting research on th e socioeconomic and ecological importance of mangrove and freshwater wetlands in Micronesia. He has conducted additional research centered on the economic and ecological dynamics of coasta l aquaculture systems in Indonesia, Thailand, and Vietnam.