Efficacy of forest resources governance on REDD+ performance in Uvinza district, western Tanzania

This paper presents findings of a study on the effectiveness of forest resource governance on the performance of the REDD+ project in the Masito-Ugalla ecosystem in Uvinza district, western Tanzania through transparency, accountability, participation and integrity. The study was conducted in Ilagala, Karago and Kirando villages with the objective of understanding the effects of forest resources governance on REDD+ project performance as well as examining the best ways of integrating forest resource governance with REDD+ projects. The methods used in the study included household interviews that used a structured questionnaire, semi-structured interviews and focus group discussions. Semi-structured interviews were used to gather qualitative data from key informants. The findings revealed that despite the efforts undertaken by REDD+ initiative in Masito-Ugalla ecosystem, its performance was generally average. Such performance was due to the support rendered by JUWAMMA, the inter-village community based organization, to the REDD+ project activities in the area. Poor forest resources governance had, however, continued to hamper its performance. Minimal duration of operation of the REDD+ Project and dominance of politics in conservation matters were the major reasons for its poor performance. Embezzlement of funds by politicians and senior government officials also contributed to the bad performance of the project. Villagers were also seldom engaged in decision-making. Implementing accountability by setting standards would have improved forest resources governance in the study area. The study concludes that more transparent budgeting and public procurement is inevitable for achieving good REDD+ performance. REDD+ activities also need to be addressed openly, with legal measures being taken by responsible authorities against corruption'.


INTRODUCTION
Understanding the relationship between forest resources management governance and the needs of local communities is an essential step for successful REDD+ performance.Local communities are perceived as direct *Corresponding author: E-mail: jmakunga1@gmail.com.Tel: Mobile: +255 755 676826.
Authors agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License users of forest resources, therefore, improving governance in such resources in inevitable.Forest resources governance aims at addressing governance issues regarding forest resources and the environment with the overall objective of ensuring sustainable economic growth, adaptation to climate change and variability, poverty alleviation, and improving environmental protection through avoided deforestation.The World Bank (2015) revealed that governance can be good or bad.
According to IFAD (1999), good governance is "the manner in which power is exercised in the management of a country"s economic and social resources for development".In contrast, bad governance is the one associated with corruption, distortion of government budgets, inequitable growth, social exclusion, and lack of trust in authorities.Lipari (2011) for example, pointed out that poor governance was among the most important causes of state failure and underdevelopment.
Springate-Baginski and Wollenberg (2010) pointed out that achieving "good forest governance" for REDD+ is an agenda that covers more than the institutional construction for administering funds and monitoring results.To them, good governance involved a range of issues, including recognition and enforcement of forest, land or carbon rights (including those of indigenous groups), participation of forest communities and civil society in policy processes, institutional arrangements and setting management priorities, transparency and accountability.
In Tanzania, URT (2013a) revealed that the national REDD+ strategy of 2013 and action plan have been developed to direct implementation of REDD+ activities in the country.Key issues addressed included: 1. Establishment of baseline, monitoring, reporting and verification system 2. Financial mechanism and incentive schemes 3. Stakeholders engagement 4. Coordination of REDD+ schemes 5. Financing options 6.Governance 7. Training and infrastructure for REDD+ 8. Researches 9. Information and communication system and 10.Strategies to address drivers of deforestation and forest degradation.
The URT (2013a) further reported that the National Framework for REDD+ was the foundation for developing the REDD+ strategy.Burgess et al. (2010) reported that extensive forest cover and alarming deforestation rates in Tanzania were established as the key motivators for the establishment of REDD+ initiatives.
The Masito-Ugalla ecosystem (MUE) in Uvinza and Kigoma Districts has potential for carbon storage and sequestration, and thus significant for climate change and variability mitigation and poverty alleviation if forests are used in a sustainable manner (Indufor, 2014).Because of this potential, one of the pilot REDD+ projects in Tanzania namely, Building REDD+ Readiness in the Masito-Ugalla Ecosystem (MUE) Area in Support of Tanzania"s National REDD+ Strategy (the JGI pilot project), was implemented there.
A report by IUCN-Tanzania (2013), revealed that in most REDD+ Project sites, implementing nongovernmental (NGOs) had successfully facilitated the development of village by-laws to guide the implementation of REDD+ pilot activities that were approved by District Councils.The development of bylaws was followed by the provision of training on good governance to members of the Village Natural Resource Council and Village Governments.In all forest communities, fines acquired from illegal forest activities as stipulated in by-laws were identified as a source of funds for conservation purposes.On the other hand, communities had been facilitated to form and register community-based forest organizations (CBOs) to carry out the management of forest resources now and after the end of pilot REDD+ activities.A good example of these community organizations was the "JUWAMMA" under JGI REDD+ in Kigoma Region.CIFOR (2014) reported that the objective of JUWAMMA in the study area was to reduce deforestation and forest degradation driven by a demand for agricultural land and fuel-wood.Another activity was a replicable methodology for remote sensing/GIS based forest and carbon accounting.Indufor (2014) mentioned such other activities as empowerment of cadre of local trainers comprised of stakeholders that facilitate broad stakeholder participation in REDD+ project design and management; provision of communities and CBOs with the tools and skills to monitor forest biomass and carbon stocks as well as development and practice of community based equitable benefit sharing mechanism.
The goal of this study was to improve understanding of the contribution of forest resources governance in reducing deforestation, achieving biodiversity conservation, reducing climate change and variability and alleviating poverty.Specifically, the study sought to investigate the effects of forest resources governance on the REDD+ project performance and examine the best ways of integrating forest resource governance with the REDD+ project.

Study area
This study was conducted in Ilagala, Karago and Kirando villages along the shoreline of Lake Tanganyika, in Uvinza District, Western Tanzania (Figure 1).This district is confined within the Masito- Ugalla ecosystem (MUE), which is an expansive forested landscape of approximately 10,827 km 2 (Deloitte, 2012).It is within this district, where the pilot REDD+ project had operated for three years (January, 2010to December, 2012), in western Tanzania.
According to Svoboda and McNamara (2009), the study area is characterized by sedimentary rocks and well-cemented sandstones of the uppermost geologic formation, which are resistant to weathering.Other types of rocks include limestone, shale, siltstone, quartzite and volcanic rocks.
The topography of Masito-Ugalla Ecosystem is characterized by canyons, cliffs and flat-topped hills (Moyer et al., 2006).Much of the valley floor landscape contains actively flowing small streams, minerotrophic wetlands, and relatively forested areas.Accordingly, a large part of this ecosystem is drained by the Malagarasi and Lugufu river systems, which flow into Lake Tanganyika.Ugalla River is the major tributary to Malagarasi River in the east (Svoboda and McNamara, 2009) (Figure 1).
The area is characterized by seasonal tropical climate with a distinct long wet season beginning from late October to May and a short dry spell of 2 to 3 weeks in January or February followed by a prolonged dry season (URT, 1998;Svoboda and McNamara, 2009).According to URT (1998), annual rainfall varies from 600 to 1500 mm, being intense in the highlands, intermediate in lower slopes and low in valley bottom and lake-offshore areas.The mean maximum temperatures of between 21°C and 30°C and mean minimum temperatures of between 15°C and 21°C have been described within the larger eco-region (Svoboda and McNamara, 2009).
According to URT (1998) the soils along the lake shore are deep and well drained and comprise dark reddish brown fine sandy loam.The heavy black clay soils are found in permanently water logged areas.In the low relief areas the soils are dark reddish clay loams with fairly good internal drainage while the black and brown alluvial soils are mostly found in areas of high relief (URT, 1998).
The vegetation in Uvinza district comprise closed and open woodlands, which cover about 70% of the land area, as well as bushy grassland and swamps of various coverage (Svoboda and McNamara, 2009).The study area is a dry area dominated by miombo woodlands with Isoberlinia, Julbernadia and Brachystegia trees, with Brachystegia bussei often found on the steep slopes of the eroded canyons (Moyer et al., 2006).In valleys, which are generally steep walled along streams, there are gallery forests, which remain green throughout the year and it is where chimpanzees obtain much of their food.Grasslands or sclerophyllous vegetation are adapted for low rainfall (Moyer et al., 2006;Svoboda and McNamara, 2009).
Agriculture is the predominant economic activity in the study area.The major food crops grown in the district are maize, paddy, cassava, bananas, beans and sorghum while cash crops include cotton, coffee, tobacco, oil palm and groundnuts (URT, 1998).Livestock keeping is another major economic activity in the study area, the most important types of livestock being cattle, sheep and goats.
Forestry products include timber, building materials, charcoal, fuel wood, honey and beeswax.According to the URT (1998), the study area has plenty of fish in the Lake Tanganyika and Malagarasi River.Beekeeping is also another socio-economic activity conducted in the study area.Deloitte (2012) reported that promoting beekeeping as an income generating activity had achieved significant success providing training to villages on improved techniques reaching over 425 beekeepers and financing over 210 bee hives.URT (1998) revealed that people in the study area were also involved in industrial and trade activities including the Uvinza salt mine, soap making, palm oil and palm kernel oil extraction and printing to mention but a few.
Sigunga, Sunuka and Ilagala wards, in which the study villages were sampled have a population of 20,455, 36,023 and 47,026 people, respectively (URT, 2013b).The main indigenous ethnic group for Uvinza District (including all other districts in the region) is the Waha.Other tribes are Bembe, Tongwe and a mixture of ethnic groups such as Nyamwezi, Sukuma, and Fipa from Tanzania mainland and some people with origins from the neighbouring countries such as Congo, Rwanda and Burundi (URT, 1998).

Data collection methods
Different data collection methods and techniques were employed.The methods included household interviews, semi-structured interview and focus group discussion.The use of different methods (that is, triangulation) was meant to allow verification of the results and Improve confidence in the research findings.According to Patton (1999), triangulation also has been viewed as a qualitative research strategy to test validity (that is, checking out the consistency of findings) through the convergence of information from different sources.

Household Interviews
Household interviews were conducted using structured questionnaires in order to capture quantitative information about socio-economic data and other relevant information.The questionnaire consisted of both open and closed-ended questions.Different types of data were collected using this method, including elements of forest resources governance such as participation, transparency, integrity, accountability and institution structures.Other types of data were concerned with effects of natural resources governance on REDD+ performance.
The sample size was drawn at 90% confidence interval corresponding to a level of significance α.With such level of significance, the margin of error (E) or the probability of committing an error was therefore ± 10% or ± 0.1 (Smith, 2013).Households were used for drawing the sample size (n).The total number of households (N) in the three villages was 11,632.This study was conducted without knowing the population"s behaviour, (that is, it was not possible to quantify the population standard deviation).
Therefore, Slovin"s formula (Word Press, 2014) was adopted for obtaining the sample size, such that: n = N/ (1+Ne 2 ), where "n" is the sample size, "N" is the total population (for this case, the households) and "e" is the error tolerance or margin of error.
The sample size was thus determined as follows: (1) Therefore, 99 respondents were considered as the sample size for this study.Two more respondents were included above the calculated sample size, giving a total of 101 respondents due to easy availability of extra respondents before accomplishment of the questionnaire survey exercise.The sampling procedure was accomplished using non-probability random sampling.The sample population was drawn from a list of households obtained from Kigoma Regional Commissioner's Office, each household representing one respondent.In every household only one adult respondent was picked.The actual population was not preferred due to the fact that it included people of different ages, even those aged below 18 years, which was the lower limit of the respondent"s.Thus every household which was accessed was picked and included in the sample population.

Semi-structured interviews
Semi-structured interviews were used to gather qualitative data from key informants.In this study, the key informants comprised of one village chairperson, two village executive officers, one ward executive officer; one "JUWAMMA" leader, one forest monitor responsible for conducting field patrols, one fire breaker responsible for controlling wildfires and a district natural resources officer (DNRO).There were eight ( 8) key informants from whom information was obtained.Different types of data were collected using this method; these included; benefits obtained from the project and effects of natural resources governance on REDD+ performance.

Focus group discussions (FGDs)
Focus group discussions were used to collect qualitative data.The FGDs involved two processes, which were brainstorming and discussions.There were four focus groups comprising 10 individuals each.The first group included farmers; the second one represented women as the major forest resource users; the third group consisted of businessmen; and the last focus group represented honey gatherers.The units of data, which included statements from the focus groups, were recorded using a smart phone; digital camera; notebook and pen.The types of data, which were collected using the FGD method, were the effects of natural resources governance on REDD+ performance.FGDs method (Patton, 1999), was used for comparing viewpoints of different people.

Data analysis
Both quantitative and qualitative data analysis approaches were used in the analysis of the data collected to allow easier interpretation of numerical and respondents" perceptions, respectively.Quantitative data analysis was accomplished using IBM Statistical Package for Social Sciences (SPSS), version 20.0.Qualitative data, which involved those obtained using FGDs and semi-structured interviews, were analyzed using content analysis (GAO, 1989;Stewart and Shamdasani, 2017).
As reported by Hsieh and Shannon (2005) content analysis is a widely used qualitative research technique, and one of the approaches used under this kind of analysis is conventional content analysis in which coding categories were derived directly from the text data.

Performance of REDD+ project in the study area
There were mixed feelings about the performance of Percentage (%) REDD+ in the Masito-ugalla ecosystem.For example when asked whether REDD+ project attained a good performance or not, 24.4% of the respondents revealed that it resulted into a good performance, 22.9% reported that it performed poorly, while only 0.5% were not sure.However, 52.2% of the respondents did not respond on whether or not REDD+ project had attained good performance.Based on the responses on REDD+ performance, respondents were asked to state the reasons for good performance of REDD+ as well as its poor performance.For the good performance, majority (67.3%) said that it had provided them with conservation education, 28.8% of the respondents said it had empowered people to conduct income generating activities.Very few reported that the project had improved social services and also maintained soil fertility.
On the contrary, 53.2% of the respondents revealed that the major reasons for REDD+ project"s poor performance was its minimal operation duration.Another dominant reason reported by 23.4% of the respondents was the dominance of politics in conservation matters.Some 14.9% reported about leaders" selfishness.Other reasons such as inadequate facilities, insufficient education about REDD+ and lack of benefits to people were reported by very few respondents.
Some key informants explained that during REDD+ activities, there was a conflict of interest between Kigoma District Council and JUWAMMA leaders, especially on the setting of allowances during joint meetings, which were held in Kigoma District involving JUWAMMA and government leaders.The District Council allegedly wanted to control the funds for the REDD+ project, based on the fact that Masito forest, which was the focus of REDD+ pilot project was not a village forest; rather it was a general land, meaning that village authorities had less control.Such conflict affected villagers" sense of ownership of the forest, such that they continued engaging in illegal utilization of the forest resource, thus affecting REDD+ performance.In terms of the success attained by the REDD+ project, 42.6% of the respondents said that it had helped villagers in construction of classrooms and village offices (Figure 2).The other success which was reported by 33.7% of the respondents was improvements in conservation awareness.Improving people"s income through establishment of entrepreneurship groups was mentioned by 23.8% people (Figure 2).

Effects of natural resource governance on REDD+ Performance
Natural resource governance parameters that were analyzed were integrity (corruption), transparency, accountability and participation.According to Gberevbie et al. (2014) and O"Donnell (1999), good governance is supposed to be characterized by the respect for the rule of law, transparency, very limited corruption and accountability of all state officials.The findings from the REDD+ project in the Masito-Ugalla ecosystem, however, revealed that implementation of the project was marred by corruption, lack of accountability and limited participation of the local communities as revealed in the sections that follow.

Existence of corruption during REDD+ project implementation
Corruption was one of the impediments to the performance of REDD+ project next to selfishness among government leaders.Majority of the respondents (81.2%) were aware about the existence of corruption during REDD+ project implementation.Only 14.9% of the respondents reported that corruption was not experienced and 4% were not sure whether corruption was experienced or not.When asked about the types of corruption which were experienced during REDD+ implementation, a number of the respondents (30.7%) reported that embezzlement of funds by politicians and senior government officials dominated during REDD+ project implementation (Table 1).This had the detrimental consequences of accelerating the violation of forest resource conservation by-laws among others.One key informant, for example, supported the idea that embezzlement existed during the REDD+ project implementation.He pointed out that such a situation was evident during the construction of Kabuyange Primary School, in Ilagala Village, where village leaders were allegedly involved in embezzlement of the project funds.Out of Tsh 20 million which were allocated for construction of two classrooms and one staff office, Tsh 10 million were allegedly embezzled by a number of local leaders in comparison to other similar activities in other villages within the project area.The above results clearly indicate that some social development projects in the study area, which were supported by REDD+, were mainly constrained by embezzlement.
Fraud was mentioned by a few respondents.Other types of corruption such as land grabbing, nepotism and favouritism were also mentioned by some respondents (Table 1).An example of nepotism was Kabuyange Primary School, whose construction was assigned to a son of one of the government leaders at Ilagala Village, contrary to local contract procedures.
Although Tanzania continues to perform extremely poorly in Transparency International's Corruption Perceptions Index (CPI), ranked 102 nd out of 180 with a score of 3.0, the 2006 afrobarometer survey indicated that the public perceived the level of corruption to have declined between 2003 and 2005.The World Bank"s 2007 Worldwide Governance Indicators as reported by Anti-corruption Resource Centre (2009) also showed positive trends in terms of control of corruption, with a score of 43 compared to 19 in 2003.Progress has also been recorded in terms of voice and accountability while rule of law indicators remained stable (Kaufmann et al., 2008).
The main negative consequences of corruption, which were experienced during implementation of REDD+ project in the study area included an increase in violation of forest resources conservation by-laws, presence of income poverty and lack of social services.Other consequences mentioned were destruction of water catchment areas and deprivation of land tenure rights (Table 2).
The aforementioned findings are supported by Marmon (2009) who reported that in forest management corruption contributed to the uncontrolled depletion of forests and undermined on large scale political efforts to sustainably manage forests.Accordingly, it deformed public policies, leading to the misallocation of public funds.Illegal logging due to corruption caused social disruption due to the destruction of the living space of people and the ensuing deterioration of rural living conditions.It also affected the functioning of environmental services and the people who directly depended on them for their survival (Marmon, 2009)".
Bivariate correlation analysis between reasons for the poor performance of REDD+ and existence of corruption (Table 3) indicated that there was a strong, positive correlation (r = 0.135, N = 46 to 97, p<.371).The strong positive correlation is revealed by the value of "r", which is above 0.05, implying corruption was among the principal factors which exacerbated the poor performance of the REDD+ project (Table 3).

Transparency in sharing REDD+ Information
When asked about transparency criteria during sharing REDD+ information with local communities, 24.9% of the respondents showed that REDD+ information was accessed at minimal or sometimes without any cost.Others reported that transparency was relevant in terms of language and formats (11.2%).Very few individuals reported that information was timely to permit analysis and evaluation (2%) and accurate (1.5%).Also, very few respondents (2.9%) revealed that all four transparency criteria were met during the REDD+ pilot project implementation.However, more than a half of the respondents (57.6%) did not know whether or not there was transparency.
When asked about the reasons as to why transparency was not fully implemented during the REDD+ project in the study area, 93.7% of the respondents had no idea.However, very few of them (4.4%) felt that some beneficiaries restricted some information to be accessed by local communities for reasons known to them.Other reasons were lack of effective communication channels and strategies (1.5%) and inadequacy of experts to meet information demands of local communities (0.5%).
The bivariate correlation analysis of transparency and success of REDD+ in addressing deforestation, climate variability and people's livelihoods (Table 4) revealed that there was a weak positive correlation (r = 0.025, N = 101, p<.803).The weak positive correlation was manifested by the value of "r", which was below 0.05, implying that transparency was not the major factor for the success of REDD+, thus other factors with negative consequences might have impinged the project implementation.Transparency international (TI) (2013) research into the governance of REDD+ in recipient countries has revealed challenges in terms of transparency and accountability of developing REDD+ policy and project implementation and highlighted concerns regarding risks of corruption.For example TI"s governance assessments of the major multi-lateral REDD+ funding mechanisms, the UN-REDD programme, the forest carbon partnership facility and the climate investment fund"s forest investment programme showed that gaps existed in terms of transparency at the policy level and in particular sanctions for corruption (Transparency International, 2013).
An annual report by Byrn et al. (2013) indicated a general lack of disclosure in areas such as land concession contracts, forest management plans and the relationship between provided goods and services and generated revenue.Information on REDD+ financing and benefits was often fragmented and was dependent on often inadequate reporting by financing agencies.Accordingly, transparency in the forest sector was relatively poor.

Accountability
When respondents were asked about ways in which people were accounting for their actions, they mentioned setting standards, answerability and sanction (that is, enforcement) (Table 5) as the major ways.Majority of the respondents (53.5%) said that accountability was implemented only by setting standards.Very few respondents (7.9%) reported that it was very rare for people to be answerable and to be sanctioned (2.0%) due to their actions.However, 34.7% said that accountability was not implemented.
Partial correlation analysis of corruption risks and REDD+ performance while controlling the ways in which accountability was implemented during REDD+ operation, revealed that there was a weak positive partial correlation between the two variables [r=.231,N=101, p<.038], with REDD+ performance being slightly affected by risk of corruption.The zero order correlation (r=.231) suggests that controlling for accountability had no effect on the strength of the relationship between these two variables.
A report by Williams and De Koning (2016) pointed out that in the context of REDD+ programs, broad approaches to accountability incorporate in particular, the importance of institutional frameworks, answerability, and oversight.The authors used the Peru example to show that it would be useful for REDD+ countries to clarify how they would take advantage of existing resources and capacity to resolve conflicts while also designating an entity responsible for overall management and oversight.Such institutions could play a role in providing guidance to existing mechanisms on addressing REDD+related grievances, standardizing procedures for processing and reporting complaints, and analyzing the types of conflicts that are occurring in order to better understand the impacts of REDD+ decisions in practice and develop solutions.

Local community participation
For good governance to be sustainable, it requires partnerships between government and civil society.Thus, people in the study area were asked about the kind of activities in which they participated.The activities mentioned included planning for management of forest resources, revenue collection, benefit sharing planning, election of leaders, decision making and participation in other activities.Majority of the respondents (60.4%) reported that villagers were mainly involved in planning for forest resources management (Figure 3).Very few reported that people were involved in decision-making.Other activities in which the involvement of local people was limited were election of leaders, revenue collection and benefit sharing planning.The focus group discussions revealed that villagers were mainly involved in protection of forest resources against illegal activities, but majority of them were not involved in decision making and managing the project funds, activities which were controlled by project officials and top leaders from the village, district and regional levels.The aforementioned findings are corroborated by the findings from Ratsimbazafy et al. (2012) in the Makira conservation project in Madagascar, Isagar et al. (2013) in India, Griffiths (2007) and Freudenthal et al. (2011).In Madagascar, for example, majority of the respondents claimed that decision making was centered on the management committee members, leaving behind majority of the local people.Isagar et al. (2013) observed that a key lesson of the Joint Forest Management (JFM) experience in India was that involving local communities in forest management could lead to more effective forest protection.Also successful conservation depended on cooperation from the local people and forestry officials, and on legal and institutional backing from the State.Many development projects in the past failed to meet their objectives due to lack of popular participation (Isagar et al., 2013).Griffiths (2007), noted that communities need to be part of the solution if REDD+ project implementation is to be successful.This is due to the risk that REDD+ schemes may result in government, companies, conservation NGOs or speculators carving up forest lands and pursuing forest protection approaches that marginalize forest people.UNREDD programme stresses the importance of Free, Prior and Informed Consent (FPIC) as an operational guideline for the engagement of indigenous people and other dependent communities in the implementation of the UN Declaration on the Rights of Indigenous Peoples (UN-REDD Programme, 2013).
It was revealed by Freudenthal et al. (2011) in Cameroon that there was lack of information sharing even at the national policy-making level, both within the government and between external agencies.Within the government, the Ministry of Environment and Nature Protection (MINEP) was responsible for negotiations at the UNFCCC, while the Ministry of Forests and Wildlife (MINFOF) was involved in the implementation of subnational projects.The dialogue between the two ministries was found to be weak so that experiences from so-called pilot projects would not necessarily feedback the policy-making process at the national and international level.Such situation of lacking public information, meaningful community participation and mechanisms to seek FPIC was also contrary to the institutional obligations and operational policies of many conservation and finance agencies involved in REDD+ project development in Cameroon.

Land tenure conditions in Uvinza district during REDD+ Project
When asked whether they owned land or not, 98.0% of the respondents in the study area affirmed that they owned some land.Only 2% of the respondents reported that they did not own any land at all.However, the largest majority (98%) of the respondents revealed that they did not own title deeds to their land.
Respondents mentioned various reasons for them not to own land.The most dominant narrative was lack of knowledge in interpreting the Land Act and Village Land Act, which was supported by 28.7% of the respondents.Other reasons included weak legal frameworks, weak institutions in solving land disputes, high cost in accessing land tenure instruments, land grabbing and more favourable consideration for investors than local communities (Figure 4).certificate for the land.Despite the law, government officials do not always recognize village land, especially land that has not been demarcated or for which there is no land use plan.

Institutional contribution for JGI REDD+ good performance
A large majority (92.7%) of the respondents did not know any central government institution that contributed or could contribute to REDD+ good performance.Only 7.3% reported that central government institutions had contributed for the REDD+ performance.Few respondents (2%) reported that Uvinza District Council supported the project to attain good performance and only 7.3% reported that central government institutions had contributed to the REDD+ performance.However, 81.9% of the respondents said they knew the Community-Based Organization that played a big role in supporting REDD+ project to acquire a good performance.These were probably referring to the role played by the local JUWAMMA.Strangely enough, some respondents (3.2%) mentioned United Nations Agencies to have helped the REDD+ project attain a good performance (Figure 5).
Different focus group discussions and key informants specifically mentioned JUWAMMA, the inter-village CBO to have highly supported REDD+ project in its activities in the study area.They reported that JUWAMMA, made up of seven village representatives, worked in close association with village leaders and the Uvinza District Council to manage forests and organize field patrols in the Masito-Ugalla Ecosystem.Furthermore, one of the key informants emphasized that the survival of the natural forest resources in the ecosystem would be assured if JUWAMMA persisted.
Such views concur with those of Indufor ( 2014).Accordingly, JUWAMMA used existing village bodies such as the environment committee and village government to undertake community based forest management activities, leading to increased effectiveness.The CBO and village governments were assisted to secure certified credits for carbon emission reductions from REDD+ projects in the voluntary carbon markets.Indufor (2014) further explained that the implementation of the project was clearly based on a collaborative and consultative approach.This was evident from the involvement of the village government, CBO, District Council staff and officials from central government and Sokoine University of Agriculture (SUA).
On the other hand, the UNDP (2012) reported that in Indonesia, the success of the REDD+ agenda was highly dependent on the preparedness of Indonesian institutions and citizens to meet the challenges of achieving the goals set out in the Draft National REDD+ Strategy.The Participatory Governance Assessment (PGA) aimed at assessing factors related to the successful implementation of the REDD+ agenda, which included legal and regulatory frameworks, capacities of key stakeholders (including the government, judicial and legislative bodies, civil society organizations, "adat" (that is, for indigenous communities), the private sector as well as forest management performance.

Best options for using forest resource governance in attaining good REDD+ performance
The study also examined how best forest resource governance could be used to enhance good performance of the REDD+ project.Specific focus was on transparency and accountability.Majority of the respondents (64.4%) revealed that more transparent budgeting and public procurement would help achieve the objectives of the REDD+ project.Only 22.8% preferred independent audit functions as an appropriate means of transparency for achieving REDD+ project, while 12.9% of the respondents reported that formulation of external committee could work better in revealing information necessary for a successful REDD+ project.
Upon being interviewed, key informants and focus groups discussants had different views about the best options in which forest resource governance could improve REDD+ performance.The majority reported that all the principles of forest resources governance, such as transparency, participation, accountability and integrity needed to be integrated.They reported, however, that more transparent budgeting and public procurement would work much better to make REDD+ project successful.
Transparency is essential if government policy processes are to be made more accountable to stakeholders.Transparency provides information that supports public participation and improves planning (Tan et al., 2008).When there is no information about laws and institutions governing forest management, predatory agents or dishonest officials can easily manipulate the law to their advantage (Tan et al., 2008).It was for these reasons that the need for transparency as the element of forest resources governance was crucial for REDD+ performance.
In addition to transparency, majority of the respondents (65.3%) revealed that horizontal accountability was the most reliable type of accountability that would help improve REDD+ performance in the study area.Other types of accountability as reported by respondents were vertical and both horizontal and vertical (Figure 6).All the focus groups discussants and majority of the key informants reported that horizontal accountability (for example, institutions such as judiciary) was the most reliable tool compared to vertical accountability (for example, media and civil society).A member of the focus group had this to say: "…there is more enforcement in the governmental bodies like courts than using the media."(Personal communication, 13th September, 2014).
According to UNDP (2010), vertical accountability is imposed externally on governments, formally through electoral processes or indirectly through citizens and civil

Figure 1 .
Figure 1.Location of the study villages in Uvinza District (Source: IRA Lab, University of Dar es Salaam, 2014).

Figure 5 .
Figure 5. Categories of Institutions that had contributed to REDD+ good performance (Source: Field Data, 2014).

Table 1 .
Types of corruption experienced during implementation of REDD+ activities (N = 101).

Table 3 .
Bivariate correlation analysis between reasons for REDD+ poor performance and existence of corruption (N = 101).

Table 4 .
Bivariate correlations analysis between transparency and success of REDD+ in addressing deforestation, climate variability and people"s livelihoods (N = 101).