Introduction

Tropical deforestation and rural poverty are among the major apprehension of developing nations in the 21(st) century. The protectionist paradigm that had dominated nature conservation since the 19th century is replaced by a strong notion that poverty reduction and environmental protection should go hand in hand. Decentralization is among the key polices devised to achieve such dual objectives of forest governance. It has also been a pivotal forest policy in Ethiopia since 1991 when the current regime took power. The policy aimed at addressing the two key problems of the country - mitigating the ongoing alarming deforestation and forest degradation as well as contributes to poverty alleviation. Despite being crucial policy for the country, there was scarcity of studies conducted to analytically characterize the decentralization policy, to compare an extent of democratization achieved in different forms of decentralization, and to clarify the social and environmental outcomes from them.

Objectives

The study aimed to fill the aforementioned insufficiency empirically as well as theoretically in the study on decentralized forest governance with the following four objectives: to construct a modified actor-power-accountability framework to atone for drawbacks of the original framework developed by Agrawal and Ribot in 1999; to analytically characterize and assess an extent of democratization in different forms of decentralized forest governance in Ethiopia; to investigate local level social and environmental outcomes from different forms of decentralization; and to draw policy implication to improve forest policy of Ethiopia.

Research Approach

Multiple cases with embedded units design is used for the study. Case study is preferred for the study because of its numerous advantages including ability to utilize qualitative and quantitative data; applicability when the research relies on multiple sources of evidence, when the context cannot be separated from the subject, and when the study benefit from prior development of theoretical preposition to guide data collection and analysis etc. Multiple cases over single case is preferred as the key topic of the study is analysis and comparison of outputs and outcome from three forms of decentralization.

Analytical framework: Modified actor-power-accountability framework

The actor- power- accountability framework developed by Agrawal and Ribot (1999) has been one of the most influential tools in analyzing decentralized forest governance. Despite being one of the most extensively used analytical tool, the framework failed to give sufficient attention to shift in property right due to decentralization and largely neglected to link an extent of democratization with outcomes from decentralization. This study constructed a modified framework to atone for the aforementioned shortcomings. The modified framework assesses the extent of democratization, which is defined as transfer of decision making power to local people or lower level leaders that are accountable to the grassroots, by investigating (1) the actor involved, (2) decision making power on rule making, implementation and adjudication of three major property rights on the forest i.e., management, exclusion and withdraw right, and (3) accountability mechanism and direction among these actors. It also links the extent of decentralization with local level social and environmental outcomes from the decentralization reform. The framework is employed to investigate the extent of democratization as well as local level outcome in Oromia region of Ethiopia, a rarely researched country.

Research sites and data collection

The research is conducted in Oromia regional state where one third of the population as well as about 70% of the forest cover of the country are located. From Oromia region, West Shoa zone is purposively selected for the case study as three major forms of decentralization in Ethiopia, or 'deconcentration', 'devolution' and 'delegation' forms of decentralization, can be seen. Chilimo forest and nearest peasant association Gare Arera containing of Chilimo and Mesalemiya forest cooperative is selected for the study of 'devolution'. For the study of 'delegation', Jibat forest and nearest peasant association Tutu is selected. Both of these forests were previously managed under 'deconcentration' type of decentralization before the latest policy reforms. Each of the household in Chilimo, Mesalemiya and Tutu are categorized into four wealth category by using the method of simple wealth ranking.

Data is collected by using document such as policies, proclamations, regulations, agreements etc; archive records such as forest enterprise and forest cooperatives archive on seedling planting, harvesting, job creations etc; semi structured questionnaire; structured questionnaire; and satellite images. A total of fifteen documents as well as thirteen archival records are reviewed to collect qualitative and quantitative data. Open ended, semi structured and structured interview is conducted with a total of 210 actors/ households. The data on environmental outcome is triangulated with data from satellite imagery analysis. Direct observation is also major source of information throughout the data collection period. Discourse analysis is used to analyze the drivers of decentralization reform. The extent of decentralization is analyzed using pattern matching while cover change analysis of satellite imagery by ArcGIS 10, qualitative analysis and descriptive statistics are used to analyze social and environmental outcomes.

Deconcentration: Management by the lower governments

In deconcentration form of decentralization, decision-making power was transferred to regional and woreda (equivalent to district level) government and their administrative branch; woreda judiciary; and forest guards. The forest guards who were upwardly accountable to woreda administrative office was the only actor who were closer to the local people among the decision makers . Election, or the only mechanism for the local people to make the woreda and regional government be accountable, was found to be weak for three major reasons. Firstly, the election happened once in five years and sometimes it may be skipped at woreda government levels. Secondly, the local people elected representatives in the council among whom the government is elected. Thirdly, the numbers of the people significantly affected by decision on the forest, i.e. people living in and near the forest, are miniature of the total population of woreda. This makes enforcing their voice through election next to impossible.

The earlier stage of deconcentration resulted in indiscriminate destruction of the forest due to insufficient law enforcement during regime transitions. In the latter stage, natural forest recovery continued, particularly in Chilimo, albeit with gradual encroachment into the forest near the boundary to settlements. Forest guarding was poor due to lack of guards' number, insufficient financial incentive, and social opportunity cost of suing community members. Local people had stronger use right on the forest than ever.

Delegation: Managed by parastatal company

In delegation form of decentralization, Oromia Forest and Wildlife Enterprise (OFWE) made all key decisions through its offices at different levels. There was no pragmatic involvement of lower level actors like peasant associations or local people in the forest governance decision making. Vertical accountability mechanism in delegation is also found to be weak. Horizontal accountability, which is the only possible mechanism to keep decision makers accountable to the grassroots in current condition, is not functioning. Within the enterprise, the lower level offices, i.e. district level bureau, are upwardly accountable to central office.

Delegation did not affect physical capital, the most vital capital for income generation and coping mechanism in Tutu peasant association. Nonetheless, it significant altered natural capital. Except for use right on none timber product, the delegation either reduced or completely excluded local people from their key withdrawal rights on the forest. Job creation as well as investment on local development was also found to be minimal. The environmental outcome was positive because planting exceeded harvesting. Lack of incentive to manage the natural forest which is major home for biodiversity as well as the recent decline in proportion of indigenous species panted, however, were major concern.

Devolution: Management by the local people

The local people and leaders of forest cooperatives (forest user groups) have gained significant discretionary decision making power on management and exclusion rights on the forest as well as withdrawal right on logs for subsistence use and fuelwood for income generation in devolution. For income generation from selling standing volume of plantation and/ or logs, however, they need the blessing of stuffs' of woreda natural resource work process or district level staffs of Oromia Forest and Wildlife Enterprise (OFWE). In addition to the power structure, the accountability mechanism was also found to be partially in line with the requirements of democratization. The local people can make their decisions, i.e. leaders of cooperative (FUG) downwardly accountable through periodic evaluation and voting.

Social outcome analysis showed that livelihood strategies based on human capital, physical capital and natural capital are vital as income source as well as coping mechanisms from different vulnerabilities of which draught and associated shortage of food are major one. Of these, natural capital (forest) is salient for the poor community, Chilimo. Within the communities, forest is salient for the poor and very poor wealth categories. The two important physical capitals, i.e. agriculture land and livestock as well as housing, were strengthened in both communities after devolution. Devolution boosted use rights on natural capital, i.e. forest. Benefit sharing from it, nonetheless, privileged political and economic elites at the expense of the poor and very poor who were significantly dependent on the forest. Human capital is also enhanced after devolution through training, albeit the participants in these trainings were also dominated by the political and economic elites. Although the trend in environmental outcome has been good, the recent significant increase in harvesting of the forest for income generation without subsequent increase in planting after OFWE took power is a precaution for the future environmental outcome.

Conclusion and policy implication

Although transfer of meaningful discretionary power to local people or lower-tier governments that are accountable to the populace is a prerequisite to achieve positive outcomes from decentralization, this precondition is realized in none of the three forms of decentralization. However, the devolution form of decentralization is found to be in proximity to the process of democratization.. The delegation and deconcentration form of decentralization reforms shifted the power to middle and higher level actors that are upwardly accountable. The local level social and environmental outcome have shown that achieving positive social and environmental outcome is very complicated and needs more than just achieving democratization. Such complexity that requires thinking beyond democratization is better captured by the modified actor-power-accountability framework createded in this study. Restructuring of the existing power and accountability relations from local household to the federal level that also involves bringing new actors into the play ground; creating different incentives to alleviate/ minimize patronage networking, rent seeking etc; setting agreed minimum environmental standards; and selective empowering of the poor and very poor can play significant role in improving outputs and outcomes from each of three forms of decentralization.

本研究は、エチオピアのオロミア州を事例とし、3類型の分権型ガバナンスの民主化の度合と社会面・環境面での成果について検討した理論的・実証的な研究である。

1章では背景と課題を提示した。自然環境の保全においては19世紀から人為を排する保護パラダイムが優勢であったが、近年では自然保護と貧困削減の両者ともに重要であるという認識が取って代わった。分権型森林ガバナンスは、まさに貧困削減と自然保護の両立という目標達成のために編み出された仕組みの一つである。エチオピアでは1991年に分権型森林政策が導入されたが、まだほとんど研究されていない。本研究の課題は、(1)改良型「アクター・権力・アカウンタビリティ」フレームワークを構築し、(2)3類型(deconcentration, devolution, delegation)の分権型ガバナンスの民主化の度合を評価し、(3) 地元での社会面・環境面での成果について検討し、(4)エチオピアの森林政策への含意を得ることである。

2−5章で分析枠組み及びデータ収集の方法を解説している。分権型森林ガバナンスの分析ツールとしては「アクター・権力・アカウンタビリティ」フレームワークがよく知られている。しかし、この分析枠組みでは分権化に伴う所有権の変化を把握することが困難であり、かつ民主化の進展度合と分権化の成果を関連づけることができない。そこで、本研究ではこの分析枠組みを改良したうえで、事例分析のツールとして活用した。

調査対象地としてオロミア州を選んだのは、国家の森林面積の70%が分布しているからである。オロミア州の中では、分権の3類型すべてが見られる西ショア地区を調査地とした。チリモの森とその周辺の農民集団(チリモ、メサレミヤ)をdevolutionの事例として、ジバットの森とその周辺の農民集団(トゥトゥ)をdelegationの事例として、近年の政策改革以前の両者をdeconcentrationの事例として位置づけた。そして、それぞれの農民集団に属する世帯を、simple wealth ranking法により4つの豊かさ類型に区分した。

フィールド調査では、210名に対して自由回答型・半構造型・構造型インタビューを実施した。また、15の文書(政策文書、声明文、規則、合意文書など)、および13の記録(種子の播種、収穫、雇用創出など森林公社や協同組合の書類)から質的・量的なデータを入手した。環境への影響については衛星画像分析結果を活用して住民から得たデータを検証した。分権型改革の要因については言説分析の手法を、地方分権の程度についてはパターンマッチングの手法を、森林被覆の変化についてはArcGIS 10の衛星画像解析を、社会的環境的な結果については質的分析法を活用した。

6章でDeconcentration(下位の政府機関による管理)の分析結果、7章でDelegation(半官半民の企業による管理)の分析結果、8章でDevolution(地元住民による管理)の分析結果を示したうえで、9−10章で比較検討および全体の結論と政策的含意が検討されている。まずは、意味のある意志決定権を地元住民あるいは下位の政府機関へと委譲することが分権の前提であるにもかかわらず、現実には3類型の分権のどれもこの前提を満たしていないことがわかった。しかし、devolution型の分権は民主化の過程に類似していることが示された。Delegation型の分権では地元レベルの社会的環境的な成果は複雑であり、民主化の側面だけで評価することはできない。本研究で提示した改良型「アクター・権力・アカウンタビリティ」フレームワークを用いれば、民主主義の側面以外を含む複雑な実態を把握することが可能となる。世帯レベルから連邦レベルまで現存する権力と説明責任の関係とを再構築すること、レントシーキングや互恵的なネットワークを縮小するインセンティブを創出すること、合意可能な最小限の環境基準を設定すること、そして貧者たちを選択的にエンパワーすることが、3類型の分権の結果と成果を改善するためには重要である。

以上のように、本研究は複雑な側面を抱える3類型の分権のあり方について理論と現実の往復から評価・検討し、将来の政策に示唆的な結論を得ることに成功しており、学術上および政策上の貢献が大きい。よって審査委員一同は、本論文が博士(農学)の学位論文として価値あるものと認めた。