Tropical forests are being cleared rapidly, causing between 12 and 20% of all anthropogenic CO2 emissions. This process drives climate change and biodiversity loss. A new mechanism called REDD+ is being developed to pay tropical forest countries to reduce deforestation, and thereby to reduce these negative externalities. To be able to do this, maps of forest carbon stocks and change are fundamental. Policy impact analysis is essential too since REDD+ payments are performance-based. Quantify- ing biodiversity benefits of REDD+ is important too for carbon credit buyers. This thesis addresses these needs on Sumatra. As of 2007, a 7.2Mha study area holds 503 x 10 6 Mg of forest biomass, with the largest stocks in protected and production forests. Other land classes have much lower biomass, suggesting legally exploitable forests are already depleted. What forest remains is being cleared rapidly. Between 2007 and 2009, 229 x 103 ha of forest were cleared, a rate of 1.6% yr−1, and loss of >6% of the 2007 forest biomass, creating emissions of 58 ±12.1 x 10 110 6 Mg CO2e. Yet the deforestation is not uniform. On average protected forests reduce deforestation. However at the extreme, one protected forest area had virtually no forest remaining at all by 2007. By contrast the Berbak Carbon Initiative REDD+ pilot project has significant stocks (34.7 ± 17.3 ±3.5 x 106 Mg forest carbon; 380 x 106 Mg peat carbon). It also supports a population of critically endangered Sumatran tigers (occupancy Ψ=0.14; 95% CI= 0.05:0.33). The project developers hope to conserve tigers and carbon simultaneously. However, following the first year of project activities, compared against control sites, deforestation appears to have increased.

This thesis investigates the role that institutional conditions have on policy for Reducing Emissions from Deforestation and Forest Degradation (REDD+) by applying a New Institutional Economics perspective and a multimethodological approach. It focuses on three specific institutional conditions: property rights, governance and politics, and applies theoretical and empirical techniques. A single case study of Guyana’s innovative REDD+ programme is used for empirical analysis. The thesis provides contributions to normative and evaluative REDD+ literature, especially with regard to early assessments of the design, impacts and effectiveness of national-level REDD+. It makes subsidiary contributions in the areas of small-scale mining, policy design under political influence and environmental governance. Through analytical modelling the thesis finds that design of REDD+ is significantly altered when placed in a general equilibrium setting, along with when political influence is included. Econometric analysis of a unique data-set from Guyana shows effects from electoral cycles on the holding of property rights relating to the main driver of deforestation, mining, along with the introduction of REDD+. Qualitative analysis of interviews and media sources highlight that governance of REDD+ in Guyana has remained predominantly state-centric, with only some evidence that multi-actor, multilevel governance has emerged. Issues such as capacity, political will, electioneering, the retained control of finance by donors and the introduction of complicated systems of safeguards have all affected the emergence of ‘pure’ REDD+ in Guyana. The thesis provides key conclusions on the importance of a cognisance of the institutional landscape on which REDD+ is to be implemented. Including such an institutional perspective raises questions over the perceived cheapness of REDD+ as a mitigation option. It offers guidance for the design and implementation of national level REDD+ policy and highlights the need for a differentiated approach to REDD+, factoring in the relevant institutional conditions prevalent in each jurisdiction.

This thesis conducts an empirical analysis of how ‘reducing emissions from deforestation and forest degradation’ (REDD+) is rendered governable through negotiations under the United Nations Framework Convention on Climate Change. REDD+ is a proposed emissions trading scheme where deforestation in ‘developing’ countries is reduced through monetary incentives, and where this counts and ‘reduced greenhouse gas emissions’ that can be used by ‘developed’ countries to comply with their commitments to reduce emissions. A Foucauldian governmentality perspective is applied to conceptualise the negotiations as a process of contestation where the outcomes validate and target certain governance arrangements, actors and ideas, while subjugating others, with concrete effects for how forest users, forests and the climate will be governed. This process is analysed by drawing on discourse analysis and actor-network theory to consider both social and material contestation throughout the negotiations, which serves to elucidate the contested foundation REDD+ is built on. The process of validation and subjugation analysed throughout the negotiations is argued to manifest a governing strategy that subjugates deviations from how REDD+ was originally conceived, and that polices its borders so as not to jeopardise growth-oriented patterns of production and consumption outside of the scheme.

Mangrove forests are considered one of the most efficient natural carbon sinks and their preservation is thus important in climate change mitigation. However, they are declining at higher rates than terrestrial forests, due to human activities; with Kenyan mangroves being no of exception. One of the main drivers of mangrove decline in Kenya is over-exploitation for wood products. The present study aimed to assess (a) the effects of tree removal on the fluxes of greenhouse gases, surface elevation and other ecosystem functions of mangrove forests and (b) mangrove root production. To explore these objectives two experiments were established in the mangrove forests at Gazi bay, Kenya: (i) tree harvest and (ii) mangrove productivity studies. For the tree harvest experiment, ten 12 m x 12 m plots were established in March 2009 in a Rhizophora mucronata (Lam.) forest. Five plots were randomly selected and all trees within them were girdled in November 2009 and then cut in May 2010. Gas fluxes of CO2 and CH4 were sampled using the chamber technique at monthly intervals from June 2009 to April 2011. Surface elevation dynamics were observed using surface elevation stations (SES). Other variables measured included, macrofaunal abundance and diversity and natural regeneration patterns. For the root productivity experiment, twenty eight 10 m x 10 m plots were established in four mangrove forest types; with each type comprising of Avicennia marina (Forsk) Vierh., Ceriops tagal (Perr) C. B. Robinson, R. mucronata and Sonneratia alba (Sm) forests. Ten of the plots were established in A. marina and R. mucronata forests in Makongeni; while 18 plots comprising all the four species were established at Gazi; six plots each for A. marina and R. mucronata and three plots each for C. tagal and S. alba forests. Root production was estimated using the root in-growth technique (two in-growth trenches per plot), while the aboveground productivity was estimated from measurements of girth increment. Girth increment was measured using dendrometers installed on selected trees, one per plot, in combination with periodic girth measurements of 10 trees per plot. Environmental variables such as height above datum, salinity, grain size and redox potential were measured at the beginning of each experiment and during treatment periods for the tree harvest experiment. Treatment significantly elevated carbon emissions from the mangrove sediments by 14.2 ± 10.3 tCO2 ha-1 (rate of 9.8 ± 7.1 tCO2 ha-1 yr-1) within two years. Similarly, treatment significantly induced subsidence of -51.3 ± 24.3 mm (at a rate of -32.1 ± 8.4 mm yr-1) compared to 11.1±10.1 mm (at a rate of 4.2 ± 1.4 mm yr-1) in control plots in over 2 years after treatment. Decomposition of labile roots in the treated plots was most likely the driver of high emissions of carbon in the treated plots. Soil compaction due to collapse of aeranchyma tissue in roots might have been responsible for subsidence in cut plots. Natural regeneration was drastically affected by cutting, with treated plots having sparse seedlings 450 days after treatment. Gap-preferring ocypodid crabs colonized and became more abundant than sesarmids (usually found in closed canopy forest) in treated plots. There was significant variation in mangrove forest productivity between Makongeni and Gazi sites, with the mangroves in the former having higher production than those of the latter. Rhizophora mucronata forest at Makongeni had a higher aboveground biomass (AGB) than all other forest types. On the other hand A. marina forest at Makongeni had the highest belowground biomass (BGB) production. Differences in microtopographical settings and soil factors might have influence the variation in forest productivity between the two sites and between the forest types. These results underscore the importance of putting in place management options that ensure maintenance of continuous canopy cover and fast regeneration in mangrove forests under wood extraction. In addition, mangrove areas at the seafront should be protected. These results also support other work showing that mangrove forests often allocate a higher proportion of carbon to belowground roots than other forests. A high investment in belowground carbon helps facilitate surface elevation and peat formation, which not only forms important carbon sinks but may also enable mangroves to keep pace with projected sea level rise. Therefore, mangrove management in Kenya and the Western Indian Ocean region should explore options that consider trade-offs between mangrove utilization and minimizing loss of ecosystem functioning such as coastal stabilization and protection. In addition initiatives such as the payment for ecosystem services (PES) schemes e.g. reducing emissions from deforestation and degradation (REDD+) and should be explored as some of the strategies to reverse the declining trend in mangrove forest cover.

Forest conservation based on payments anchored to opportunity costs (OCs) is receiving increasing attention, including for international financial transfers for reduced emissions from deforestation and degradation (REDD+). REDD+ emerged as a payment for environmental service (PES) approach in which conditional payments are made for demonstrable greenhouse gas emission reductions against a business-as-usual baseline. Quantitative assessments of the OCs incurred by forest users of these reductions are lacking. Existing studies are coarse, obscure the heterogeneity of OCs and do not consider how OCs may change over time. An integrated assessment of OCs and carbon benefits under a proposed community forest management (CFM) intervention linked to REDD+ is undertaken in Ethiopia. The OCs of land for the intervention are estimated through household survey and market valuation. Scenarios explore how OCs are likely to change over the intervention given qualitative conservation goals and available land-use change information. The feasibility of OCs payment as a tool for REDD+ is assessed by combining cost with emission reductions estimates generated from direct tree measurements. Households’ environmental attitudes, perceptions and intention to cooperate with the intervention, estimated by a voluntary contribution to improve forest management, are then investigated. Mean OCs of forest conservation are US$334/ha, but highly heterogeneous. Plausible futures of agricultural improvement, forest product commercialisation, and degradation of land uses suggest total OCs could approach US$441 million over a 20-year project. Applying carbon stock estimates of 231tC/ha±52 in moist and 132tC/ha±73 in dry forest, REDD+ revenues may not meet annual cumulative OCs, although more nuanced conservation planning could reduce OCs. Despite OCs all households intend to cooperate in the intervention, with mean contribution of US$11±4/year/household. The expected incomes of households under the Bale REDD+ Project intervention however, were high and expectation management is necessary. Recommendations are made for REDD+ intervention design in Ethiopia.

- / -

Between 1990 and 2010 sub-Saharan Africa experienced some of the highest levels of deforestation anywhere in the world. The problem has been particularly acute in what are known as the miombo woodlands of southern Africa. These occupy a unique ecological niche and are crucial to the livelihoods of millions of people in the region, yet are disappearing rapidly. The aim of this thesis is to identify the structural causes of this phenomenon in two of the miombo countries, Zambia and Mozambique. Standard ‘resource based’ explanations for deforestation in both countries tend to focus primarily on demographic and economic factors, emphasising the impact of economic reforms, population growth, rural migration, poverty, minimal access to electricty and a lack of institutional resources. However I argue that these explanations do not account for Mozambique’s relatively better record on deforestation during the period in question, and that a more convincing account is offered by a ‘governance based’ explanation, which emphasises different forms of forest governance and institutional arrangements affecting the forest sector in each country. Specifically, Mozambique has fared better than Zambia thanks to its more secure system of traditional land tenure, the implementation of more progressive legislation and a sustained commitment to community based natural resource management. The implication is that future initiatives to curb deforestation in these countries should concentrate on addressing institutional and policy based shortcomings before implementing market based mechanisms designed to encourage conservation.

Reducing deforestation and degradation (REDD+) under the rubric of payment for ecosystem services (PES) is being promoted as the most cost-effective mechanism for reducing global greenhouse gas emissions. This process of commodifying forest services will redefine the rhetoric of decentralised forest governance that has evolved in developing countries over the last two decades. This thesis uses ethnographic case studies in two forest villages in West Bengal state in India, along with 294 household surveys and 76 interviews, to examine the impending changes in socio-economic and political arenas with the adoption of a market mechanism like REDD+. I undertook a pilot study for one month in October 2009, which was followed by my main field work in two phases: February to July 2010 and November 2010 to February 2011. First, the analysis of livelihood dependence of forests dwellers shows unequal extraction of forest products by various wealth classes under the current socio-economic and political structures of village societies. Rich and medium class families with their higher assets were higher net users of forests, while poor households had a critical dependence on forests for their daily survival. Second, I examine the impact of livelihood dependence on forests. A majority of key informants did not see the current extraction of forest products, for example, for meeting local subsistence and commercial needs as major detriments to forest and carbon conservation. However, I argue that a number of legal provisions and official guidelines could potentially impose restrictions on the ongoing forest use pattern as a result of REDD+. Third, by comparing the functioning of the village council (with a special focus on the implementation of India’s National Employment Guarantee Scheme) with forest protection committees, I reflect upon the limitations in the decentralised forest management that emerge from the institutional design of the programme. I show that the decentralised forest governance suffers from lack of accountability and transparency over the control of forests by the Forest Department. Finally, this thesis suggests that the institutional design for REDD+ at the national level needs to be based on the democratic partnership of local institutions and the state.

This Thesis is concerned with multilevel and multi-sector forest governance in the Malaysian State of Sabah. It begins by amending the theory of vertical institutional interplay by applying constructivist and historical new institutional theoretical approaches, which contrasts with the more static theoretical foundations that have characterised study of this area to date. It then develops an analytical framework that uses policy frame analysis. This is designed to apply this new theoretical approach to the empirical context of Sabah. This framework analyses empirical subject through three stages. The first stage investigates the development of forest governance institutions at the global level and the state level within Sabah. The second stage then considers how the intersection of these developments, specifically focusing on role of ideas, discourse and agency, created the impetus for new policy initiatives in two local-level empirical examples. The third stage then considers the extent to which these initiatives were successful in institutionalising new forest conservation practices, or conversely how they were impeded by state level historical institutional continuities. The findings of this Thesis differentiate two forms of vertical institutional interplay. The first is the way global institutions affect state level ones where key actors mobilise ideas and discourses to in order to shift the direction of policy and initiate institutional change. The second is where the influence of global institutions is blocked by barriers created by long term historic institutional legacies that have shaped state level institutions. These findings show that vertical institutional interplay has initiated a partial shift in forest institutions and policy in Sabah. This shift varies between different locations according to the relative influence of these two forms of institutional interplay, and has created more dynamism and uncertainty in Sabah’s forest governance institutions. This Thesis contributes to existing literature through its ability to better conceptualise the role of vertical institutional interplay in a way that can account for the tension between the fixed and dynamic aspects of institutions. This contrasts to older approaches that have focused largely on the fixed aspects of institutions. The contribution is also demonstrated in the way this theoretical approach is able to better conceptualise fine grain variations in these dynamics at a local level of scale.

Literature on the integration of cross-cutting issues, or policy integration, has given little attention to how policy-making processes allow for policy integration as well as present barriers. There is also little evidence of how sectoral ministries respond to crosscutting issues and in what way they are affected by pressure to address such issues, including those promoted by ‘competing’ agencies. Climate change presents a significant and important issue for integration into many areas of public policy. Many government ministries and departments are tasked with responding to climate change mitigation and adaptation objectives. Forestry is a key sector in building a response to climate change and so an investigation into how policy-makers and policy-making processes have responded to climate change can shed light on the integration of cross-cutting issues. The thesis reports on investigation into how the Ministry of Forestry, Indonesia, has responded to climate change and into the organisational arrangements developed in the response to climate change. The research developed an innovative framework for the analysis of policy integration, generating conclusions in relation to the policy process, organisational arrangements and the influence of key actors, including policy champions and boundary spanners.