Land and Natural Resources Tenure Security Learning Initiative for East and Southern Africa Country Report Rwanda

Land and Natural Resources Tenure Security Learning Initiative for East and Southern Africa Country Report Rwanda SECURING LAND AND PROPERTY RIGHTS FOR ALL

ACRONYMS CLGS DLB GEF GIS GLTN GoR IFAD KM KWAMP LTR MINAGRI MINALOC MINIRENA MINITERE NGO NLP NR OLL PCU RCMRD RNRA SRM TSLI-ESA WMP Local Committees for Management and Supervision District Land Bureaux Global Environmental Facility Geographic Information System Global Land Tool Network Government of Rwanda International Fund for Agricultural Development Knowledge Management Kirehe Community-Based Watershed Management Project Land Tenure Regularisation Ministry of Agriculture Ministry of Local Government, Good Governance, Community Development and Social Affairs Ministry of Natural Resources Ministry of Lands, Environment Forestry, Water and Mines Non-Governmental Organisation National Land Policy Natural Resources Organic Land Law Project Coordination Unit Regional Centre for Mapping of Resources for Development Rwanda Natural Resources Authority Strategic Road Map Tenure Security Learning Initiative Eastern & Southern Africa Watershed Management Plans i

CONTENTS ACRONYMS... I INTRODUCTION... 3 DESCRIPTION & AIMS OF THE PROJECT... 4 TENURE SECURITY CONTEXT IN RWANDA... 4 LAND POLICY AND REFORM PROCESSES... 6 Principles for the Land Tenure Regularisation Process... 7 Progress, Achievements and Challenges in National Land Tenure Regularisation... 8 IFAD Involvement and Support for Land Tenure Regularisation... 9 SOME CURRENT ISSUES... 10 LAND AND WATER RIGHTS: Issues, Challenges, Tools and Strategies... 12 GROUP RIGHTS: Issues, Challenges, Tools and Strategies... 15 MAPPING: Issues, Challenges, Tools and Strategies... 16 WOMEN S ACCESS: Issues, Challenges, Tools and Strategies... 17 KNOWLEDGE MANAGEMENT... 17 ADDITIONAL READING... 18 Tenure Security Learning Initiative East and Southern Africa... 18 Rwanda Land Policy... 18 ii

INTRODUCTION The aim of this report is to capture some lessons from IFAD projects being implemented in Rwanda. It is produced as part of the Tenure Security Learning Initiative - Eastern & Southern Africa (TSLI-ESA). It is one of a number of reports on different country visits or contexts, the objective of which is to develop a better understanding of the tenure issues faced by IFAD-supported projects in these countries and to examine some of the tools and strategies adopted to deal with these. These visits and country reports form part of the overall strategy of the TSLI-ESA, which is to identify common issues and to enhance lesson sharing and knowledge management on land-related tools amongst the various projects and country partners. The process of lesson-sharing, and of developing a strategy for disseminating information on tools and strategies amongst the IFAD-supported projects, was initiated in May 2012 through a regional workshop held in Nairobi. 1 These country reports are designed to further this process, by providing the participants with further insights into potentially useful tools and strategies for dealing with land-related issues and challenges. They also provide contact points within the country, from which further details and assistance can be sought. In addition, the reports identify some further potential activities or initiatives that could be undertaken in order to assist the IFAD projects, either in the specific country, or through the further development of knowledge management processes within the broader framework of the TSLI-ESA. As an introduction, the report provides some description and context to the relevant IFAD projects in Rwanda and describes, in general terms, some of the key national policy and legal issues that form the background to their implementation. The specific land and natural resource issues that are identified here are examined through the lens of the five thematic areas that form the framework for the approach of the TSLI-ESA, each of which forms a heading for the report. These thematic areas are: MAPPING: Using technically advanced geographic information technologies, such as aerial photography, remote sensing technology and Geographic Information Systems (GIS) for mapping land and natural resource rights, use and management. LAND & WATER RIGHTS: Recognizing and documenting small-scale farmers land and water rights in irrigation schemes. GROUP RIGHTS: Recognizing and documenting group rights, focusing on range/grazing lands, forests and artisanal fishing communities. WOMEN S ACCESS: Strengthening women s access to land. INCLUSIVE BUSINESS: Documenting best practices in securing land and natural resource rights through business partnerships between small-scale farmers and outside investors. The separate headings above capture both a description of the issues being faced within each thematic area, and identify the tools and strategies that the projects incorporate, or that have been adopted during implementation. Each section also contains some short 1 See Final Proceedings: Regional Learning Workshop on Land and Natural Resources Tenure Security, 29-31 May 2012, Nairobi, Kenya. 3

comments in respect to knowledge management within the projects. The final part of the report provides some conclusions and proposed follow up actions in the context of ongoing work in this and potential subsequent phases of the TSLI. The central focus of this report is the Kirehe Community-Based Watershed Management Project (KWAMP), implemented by the Rwandan Ministry of Agriculture (MINAGRI). Its overall objective is the development of sustainable and profitable small-scale commercial agriculture in Kirehe District, in Eastern Rwanda. 2 The key objective of this report is to draw lessons on how KWAMP and partners are addressing land and natural resource tenure and management issues in the context of the development of a number of irrigation schemes. DESCRIPTION & AIMS OF THE PROJECT The project supported in Rwanda by IFAD and examined in this report is: KWAMP (Kirehe Community-Based Watershed Management Project) aims to empower farming communities to become local planning and implementation partners, and to create a strong district administration providing key public services required for local economic development processes. The project aims at reaching 22,500 direct households. It includes a Water and Land Use Management component that assists with watershed management and planning process in 18 watersheds, as well as the regularisation of land tenure in the district generally, as well as specifically within both marshland and hillside irrigation schemes, prioritising the allocation of land for landless in the productive wetlands. TENURE SECURITY CONTEXT IN RWANDA Rwanda s land resources, patterns of land occupation and existing land tenure systems are characterised by high population growth, severe land pressure and an increasing number of small fragmented land plots. 3 Until recently these plots were held under customary arrangements by the majority of rural Rwandans, but a massive regularisation process since 2010 has seen the systematic identification, adjudication and demarcation of all land in Rwanda, with a view to the issuance of long-term, tradable, state leaseholds to the legitimate occupants. It is this nationwide process that sets Rwanda apart from the rest of the region; no other country in Sub-Saharan Africa has attempted a process of formalisation and documentation of customary land rights on such a comprehensive scale. The process was conducted under the framework of the Organic Land Law (OLL), introduced in 2005 (see below). The OLL was informed by a number of concerns: Rwanda is a small country and in 2005 had an estimated growth rate of 3.1 per cent. The overall area is 2.6 million hectares, of which 52 per cent is arable land. The population density on arable land was particularly high. Quoted figures give estimates of over 300 persons per square kilometre in 2001, with some districts 2 The KWAMP is co-financed by IFAD, WFP and the Government of Rwanda. It became effective on 30 April 2009 and is due for completion in June 2016. 3 These characteristics have been well-documented and have underpinned the nature and direction of new policies and legislation introduced in the last few years; the policy, for example, emphasises decentralisation, villagisation and urbanisation, with improved land-use planning and management practises. 4

containing up to 820 personss per square kilometre. Approximately 91 per cent of the population was involved in agriculture. Population pressures and inheritance practices in Rwanda had resulted in increasing fragmentation of land parcels, which the Government of Rwandaa (GoR) was concerned to discourage. Land fragmentation was considered to be inefficient and to leave households with holdings too small to support them. In the 1950s, 50 per cent of households had more than 2 hectares of land, whilst by 2001 almost 60 per cent of households had less than 0.5 hectares of land, with only 6 per cent having more than 2 hectares of land. Parcel sizes nationwide in 2008 varied from 0.06 ha (Northern Province) to 0.025 (Western Province), to 1.0 ha (Eastern Province). Households routinely held more than one parcel, with an average total land holding of 0.75 ha scattered over five or six land parcels ( Strategic Road Map for Land Tenure Reform in Rwanda, 2008). Table 1 below shows the average area of the individual parcels that have since been approved through the regularisation process (partial data from 2011). Table 1: Average area of approved parcels, Aug 2011 Source: Norfolk & De Wit, 2011 Province West 0.14 East 0.28 North 0.14 South 0.22 Kigali 0.23 Average area (ha) The majority of Rwanda s land was still held under customary or local tenure; successive waves of violence between 1959 and 1994 and the progressive commodification of land had weakened these informal structures. Population pressure, land scarcity and economic development were increasing the demand for, and hence the value of land, and an active informal market in land had arisen to meet that demand, with localised pressures resulting in increased prices. Even in poor rural areas a field sample of 3,000 land holdings noted annual inflation in land prices of 15 per cent; in urban areas the price inflation was up to 25 per cent (MINITERE Field Trials, 2007). Resettlement practices after the extensive violence in 1994 involved a process of land sharing between land occupants and returnees, conducted in the interests of promoting national unity, equity and reconciliation. This however contributed to land fragmentation and, althoughh generally accepted by original land claimants, left some feelings of grievance and alienation. The GoR wanted to discontinue this practice of land sharing, while accepting and formalising the patterns of land rights which resulted from the process, and avoiding the problems that would be generated by seeking legal resolution of restitution claims and counter claims through the courts. The solutions for remaining returnees and landless consequently lay in re-settlement in different areas where sufficient land could be made available to establish new agricultural villages, and in the provision of house plots and alternative off-farm livelihoods. Past subdivisionss for the purpose of land sharing were allowed to stand, 5

despite the general proscription of the sub-division of holdings less than 1 hectare. The State is, however, responsible for providing land for the landless and specifies the categories of land which may be allocated to them ( Strategic Road Map for Land Tenure Reform in Rwanda, 2008). There were in existence a number of different forms of land rights, including old titles to freehold issued in the colonial period; the country had complex systems of customary and written tenure exacerbated by the effects of war and genocide and the GoR wished to introduce a unitary system of forms of land holding. LAND POLICY AND REFORM PROCESSES Rwanda has been engaged in a comprehensive land reform process since the introduction of a new Constitution in 2003. The National Land Policy (NLP) of 2004 was followed by the passing of the OLL in 2005 and, starting in 2006, the GoR began introducing a number of orders and decrees that set out the new framework for planning and land administration. Implementation of this new framework has been on-going to date. The 2004 NLP put forward a number of general principles, strongly emphasising the optimum use of available land resources. The OLL aimed to set out the legal and institutional framework for land administration and management, and encapsulated a number of strategic approaches to achieve the stated policy objectives. One of the key approaches was the establishment of a single statutory system of land tenure and the replacement of the customary system by formal and registered land rights across the whole of the national territory. The OLL thus recognises land rights acquired through customary law, but obliges the conversion of these into legally registered rights. It also mandated the establishment of a national cadastral system, linked to a registry which is designed to record and guarantee the integrity of subsequent transactions. This need to convert informal customary rights into formally registered rights is the key driver behind the Rwandan Land Tenure Regularisation (LTR) programme. The initial imperative set by the NLP and OLL was to establish a suitable administrative procedure to adequately and efficiently recognise existing rights over land, and to convert these into legally registered rights. The Rwandan government and its partners were also mindful of the need to satisfy a number of further imperatives: Developing a methodology that would permit this process to go ahead at more or less the same time across the entire country; Ensuring that the process was transparent, legitimate and did not lead to people (especially the more vulnerable) being dispossessed of their land, and; Ensuring that the process could adequately deal with any resulting disputes. Consequently, in 2005, the Ministry of Lands, Environment Forestry, Water and Mines (MINITERE), which at that time was charged with implementing the OLL, embarked upon a process of developing and fine tuning a methodology for LTR through a pilot exercise in 4 cells (the lowest administrative units), chosen to reflect typical situations encountered throughout the country, before embarking on a national roll-out. This process began in 2006 and concluded in April 2009. The key objective of the pilot exercises was to develop a Strategic Road Map (SRM) for land reform in the country. The SRM would set out a framework for a land reform process that secures the rights of all 6

citizens, including the poor and vulnerable, whilst also supporting national economic development and promoting environmental sustainability (DFID, 2005). A team of consultants were recruited to assist the MINITERE, which embarked on a series of field consultations to gain an understanding of what was required. The consultations provided direction for a series of subsequent field trials conducted in 2007/8. On completion of these trials, the project conducted in-depth analyses of both the process and the results of the trials, which were designed to test acceptability, levels of buy-in and the appropriateness of the systems. 4 The resulting SRM was approved by the Rwandan Cabinet in March 2008. The road map laid out an ambitious programme and the Ministry of Natural Resources (MINIRENA), which had since become responsible for the overall programme, adopted a 5 year plan for full registration, within which there would be a two year field programme of LTR. 5 Principles for the Land Tenure Regularisation Process Systematic LTR provides one of the principal means of bringing land to first registration. The objective of the LTR process was to record all existing rights in land and clarify their status under the OLL. The rights that were validated under this process could then be converted into a form that was recognised under the OLL and registered. As required by the law, this applied to all land in Rwanda: including private land, State land in the Private Domain, and State land in the Public Domain. There were two key aspects to the task: The conversion of titles held under written law (Civil Code) to the new forms of the OLL, and their transfer from the existing national (or city) registers to the new Rwanda National Land Registry; The recording all customary or informal rights and issuing new titles under the OLL, followed by registration on the national land register. The second task was much greater, involving nearly 100 times more land parcels. The only way to effectively record the land rights was by making a field visit to each and every parcel. The SRM set out the following guiding principles for the LTR process: The protection of existing rights: the fundamental objective of LTR is to establish what rights already exist over a given piece of land. The rights are protected under the OLL and were created prior to any decisions or plans for land use or land development. The adoption of appropriate standards of evidence: many Rwandans had no formal evidence of their land rights. Simple LTR procedures for proving legitimate occupation, such as attestation by a quorum of community members, including a local land committee, allowed all valid rights to be verified and recorded. 4 They also calculated work rates, costs and other resource requirements for a rollout of the process across the country. Throughout this period, a considerable amount of work was also directed towards the legal and policy studies necessary for completing the secondary legislation for LTR, as well as developing strategies for communications, monitoring and evaluation, and capacity building. 5 Initial drafts of the SRM proposed a phased introduction of the LTR process at national level, concentrating initially in registration hotspots and extending over a period of 15 to 20 years. 7

The hearing all claims: as part of the process, a period for publication, review, objection and correction to claims would be needed. The resolving of disputed claims: disputes in relation to claims needed to be resolved quickly but effectively. If many cases were to end up in the courts, regularisation would be severely delayed. The LTR process therefore needed to include a fair and transparent mechanism for settling competing claims in the community through mediation and agreement. A public and transparent process: the LTR process depends on informed choices by landholders, by the community and by the officials doing the work. The public had to be informed, through the media, open public meetings and individual advice in respect to the procedures being followed and how to participate in the process. A just and accountable process: the LTR should not usurp any person s valid right over land and everyone must receive a fair hearing. Officials involved in LTR would be held legally accountable for their decisions and obliged to make them in a manner that was reasonable and well justified. Vulnerable groups, in particular, were to be supported to protect their rights through the LTR process. A key element of the process was the adoption of the General Boundaries system, allowing the boundaries of plots to be identified and demarcated utilising high resolution imagery, rather than through the surveying of the exact boundaries. 6 The LTR process designed by the MINITERE was therefore a recognised, well tried process, adapted and applied to local conditions, aimed at maximising local level participation, which gave practical effect to the government s policy and strategic initiatives. Progress, Achievements and Challenges in National Land Tenure Regularisation The process of demarcating the customary land parcels was completed towards the end of 2012. More than 10 million parcels were demarcated and adjudicated in less than three years. Although the LTR Programme has been considered a success in respect to demarcating and adjudicating the land parcels, there are, however, indications that the all-important final stage of issuing the titles into the hands of the rightful owners is facing some significant challenges (Baldwin, 2012; Norfolk & De Wit, 2011). In particular, it seems that the payment of fees by landowners may have created barriers to full engagement of beneficiaries in the programme, particularly the fees for 1 st registration, which may have dissuaded people from coming to collect the certificates. As Baldwin notes, Elsewhere, in systematic registration programmes, there are no fees precisely because of the problems that have arisen in Rwanda. This is not so much a lesson learnt, as a lesson reinforced (Baldwin, 2012). The requirements for maintaining the register are also important variables for future sustainability. It is very important that there are affordable and suitable services that keep landowners engaged with the formal system that has now been established; if there are barriers to access for use of the system (cumbersome procedures and requirements, inaccessible offices, high fees), there is a significant danger that transactions on land will lapse back into the informal sphere (Norfolk & De Wit, 2011). See below for further details on these issues. 6 A title plan with 'general boundaries' shows the boundary of a property in relation to a given physical feature on the ground such as a wall or hedge. In this concept, the precise position of the legal boundary between adjoining parcels is effectively left undetermined and unrecorded on the title plan, until such time as it may become necessary to do so. 8

The governance of land administration is shared by a set of institutions from both central and local government. These responsibilities can be broadly categorised as policy (Ministry of Natural Resources - formerly MINITERE, now MINIRENA), implementation (Land Commissions), technical (the Rwanda Natural Resources Authority Land & Mapping Department) and administrative (Ministry of Local Government, Good Governance, Community Development and Social Affairs - MINALOC and the Districts). The responsibility for maintaining the land register is a joint one technically the process is managed from central government by the Rwanda Natural Resources Authority (RNRA), but in terms of the delivery of services to customers, the responsibility lies with the District Land Bureaux, which are administratively the responsibility of the District local government. In the short term, the LTR programme has hired hundreds of temporary staff to complete the work, but as the focus switches to maintenance of the land register it will become crucial to consider organisational roles, their specific responsibilities and how different organisations are to work together. Linked to this is the need to consider the products and services being provided and to examine whether the administrative structures and the capacity within the organisations involved are adequate and if not, what needs to be done? These are the current challenges of the RNRA. IFAD Involvement and Support for Land Tenure Regularisation In 2005, the Ministry of Lands, Environment Forestry, Water and Mines (MINITERE) embarked upon a process of developing and fine tuning a methodology for land tenure reform through a pilot exercise in four cells, chosen to reflect typical situations encountered throughout the country, before embarking on a national roll-out. The IFAD supported Kirehe Community-based Watershed Management Project (KWAMP) has built on this successful land regularisation trial. The pilot was carried out by the Kirehe District Administration and National Land Centre (NLC) from October 2007 to March 2008 in Mwoga Cell in Mahama Sector as one of four trials done in the country. About 2,850 parcels were demarcated, adjudicated and issued certificates in six weeks, with an additional three months for the lodging and resolving objections. The Department for Lands and Mapping (DLM), formerly National Land Centre (NLC), through a MoU with Ministry of Agriculture and Animal resources (MINAGRI), has support the Kirehe District Land Bureau (DLB) to upscale land regularisation in Sectors covering the 15 watersheds within the Project area. Demarcation has been completed in the whole district, with almost 240,000 plots which have been demarcated (initially around 150,000 land parcels were estimated to be registered in Kirehe). Of these plots about 202,000 have been adjudicated (84%) and 176,000 lease titles have been printed (73%) and 137,000 issued (57%) as at the beginning of October 2012. The land regularisation process has been contributing positively to development in Kirehe, in particular: to the reduction of disputes over land, to improving people s ability to access credit, to recognising women s land rights and to providing a good foundation for private investments in the land, including reforestation and soil and water conservation measures. However, the Project has yet to assess the exact impact. It is expected that Kirehe will still be the first rural district in Rwanda to issue all land titles, which has contributed positively to the district s performance against its Performance Contract with the President. Lessons learned from KWAMP have contributed significantly to the national implementation roll-out in other parts of the country. At national level, the process of demarcating the customary land parcels was completed 9

towards the end of 2012. More than 10 million parcels were demarcated and adjudicated in less than three years. Despite some challenges, the rate of land regularisation has been impressive and is increasingly recognised internationally as being exemplary in respect to demarcating and adjudicating the land parcels. SOME CURRENT ISSUES Although the major work of demarcating all the parcels in the field has been completed, the GoR is still facing a number of significant challenges in ensuring that the work completed so far is sustainable and that the new, comprehensive register of leaseholds can be maintained. These include the following (see (Norfolk & De Wit, 2011): There has been a lower than expected rate of collection of the new lease documents. In October last year, the RNRA admitted that three million land title deeds remain unclaimed at the various district offices (Karake, 2012). The same report also indicated that although the RNRA had demarcated 10.4 million plots, it was holding full information for only 8.7 million of those. The collection rate of the title deeds in the Eastern Province has been reasonable [but] in other provinces the collection is very low (ibid). Part of the problem arises from the lease fees. For the majority of Rwandans, this is not a concern, because for land under two hectares in rural areas, the fees are not applicable. In addition, the government runs an Ubudehe scheme, which waives fees among citizens regarded as the most vulnerable (Karake, 2012); they represent 22.28 per cent of the total population. However, in any future transactions relating to land, the new leaseholders are required to travel to the District Land Office, and present a number of documents, including proof of payment of other taxes. Rural Rwandans transact regularly in land; swapping, and purchasing, dividing and consolidating parcels on an annual basis, as part of a dynamic livelihood management process. 7 Travelling to a district centre each time may be acting as a powerful disincentive to the new title holders. Further reasons that the public may not use the registration system, and may instead favour informal recognition and transactions, are: because the transaction is unlawful (such as selling subdivisions of land below the 1 hectare limit); to avoid the imposition of formal inheritance rules; to avoid payment of transfer or inheritance taxes; or because the registry is expensive, time-consuming or not user-friendly. Some of the additional costs that landowners will have to face in engaging with the formal land registration system include: transport and opportunity costs involved in travelling to the district in order to complete transactions; 7 The DFID 2011 MidTerm Review notes that in Ruhengeri cell in excess of 200 transfers had accumulated over a period of one year. In Nyarutarama cell (Gicumbi district), over a single morning, the land committee overseeing local LTR processes received 20 requests for the transfer of whole parcels and 15 requests for subdivisions (10 sales and 5 inheritances) (Norfolk & De Wit, 2011). 10

official and opportunity costs for the seller in obtaining a Tax Identification Number (TIN) and a Tax Clearance Certificate from the Rwanda Revenue Authority (it is unclear if this available in every district); a fixed transaction fee of FRw 20,000 or up to 30% of the land value. Compare this to the relative ease and low cost of the informal system currently in operation, which is even considered by many sophisticated urban dwellers in Kigali as being perfectly adequate in terms of tenure security; this involves handwritten or typed sales and purchase agreements, known as Icyemezo cy Ubuguzi, and the witnessing of such agreements by the local institutions. This situation is exacerbated by the fact that the receipts issued to land holders in the field (once the demarcation has been completed, but before all the information has been registered) have legal force. This means that transactions based on these receipts are also legally valid. Beneficiaries of the programme do not always perceive the status and importance of a certificate. Receiving a paper that proves land ownership is extremely important, but land owners can already obtain a paper in the form of the claim receipt, or even the receipt from the para-surveyor. There is evidence that, for several land owners, including some urban sophisticates that are well versed in the OLL, the claim receipt provides the necessary security; it is even used as a widely accepted document to transfer land parcels. The legal status of village (umudugudu) land has prevented the issue of leases to residents of these areas. Although some of these residential areas are being established through land swaps and compensatory deals between private persons, the leases are being issued for the whole village, with those allocated a parcel merely being registered as persons of interest, on the grounds that the land where umudugudu are being built becomes public (district) land. This approach has the potential to create a level of tenure insecurity; most people are aware that what is given by the state can be taken away by the state. Residents in the umudugudu, particularly those that have provided alternative land to those displaced through the process of establishment, expect the issuance of leases for their individual parcels. The law is currently under revision in this respect. There has been some cost-cutting in the LTR programme which may have had a negative impact on the sustainability of the resulting cadastre. There is a statutory period for people to lodge objections and corrections to the demarcation process (Ministerial Order N 002/2008), which is sixty working days. However, a review in 2011 found that the RNRA teams were only allowing for a period of two calendar weeks. As the review document states: The O&C phase of the LTR process is critical to ensure the transparency of the title demarcation and award. O&C needs to be considered as the point at which Rwandese society has the opportunity to express grievances on real or perceived land allocation, fraud or wrongful or inequitable registration. Currently, it is predominantly treated as a technical exercise, designed merely to correct names and ID card numbers (Norfolk & De Wit, 2011). This is related to the uncertainty about the exact nature and designation of the first field phase of the LTR process. Sometimes it is referred to as information dissemination and public awareness, whilst at other times it is called mobilisation. The 2011 DFID review found that the LTR activities had more of the flavour of mobilisation, than of creating public awareness. There is much emphasis on, and use of, the tools of mobilisation (megaphones, radio messages, songs, announcements 11

and public notices) rather than on the creation of the necessary conditions for engagement and awareness creation (workshops, meetings, events at which people ask questions and receive advice) (Norfolk & De Wit, 2011). There are on-going concerns about the demarcation and status of wetlands. A recent national demarcation exercise, designed to identify wetland areas from which people are supposed to be evacuated, has had to be re-evaluated after a number of mistakes appear to have been made. A press report quotes an RNRA spokesman as saying that some places that had been highlighted as wetlands in the GIS report, were actually found to be hills and valleys" (Asiimwe, 2012). This is an emotive issue, since for many people the most productive areas of land that they own fall within areas classified as wetlands. LAND AND WATER RIGHTS: Issues, Challenges, Tools and Strategies The Water and Land Use Management component of KWAMP has supported the regularisation of land tenure in the district generally, as well as specifically within both marshland and hillside irrigation schemes, prioritising the allocation of land for landless in the productive wetlands. Overall, KWAMP aims to support the irrigation of 1,500 hectares of government-owned productive marshlands 8 and 1,500 hectares of mainly customary-owned hillside 9 land, whilst ensuring that access rights within these areas are distributed as equitably as possible. These processes have been conducted within the ambit of the national LTR programme. It is this area of KWAMP s work in which there is a number of land and water rights issues. Some of the key issues and challenges that have been faced by the project in these tasks include the following: The starting point for the LTR process has been to recognise existing, legitimately owned and occupied parcels of land. However, the KWAMP is implementing improvements to certain areas of land, including the construction of new irrigation infrastructure. Inevitably, in this process, there are those who will be negatively affected and those who may benefit disproportionately, given the existing patterns of land holdings. A key challenge for KWAMP is therefore: to understand the dimensions of this challenge (that is, to identify those who are negatively affected, those who will benefit to a greater extent than others, to assess the capacity for redistribution, etc.); and, to create the conditions for a process that results, as far as possible, in an equitable sharing of the benefits and the burdens of the schemes implementation. The first part of the challenge relates to understanding the current situation. Although the land demarcation and adjudication process has been completed (for 208,000 parcels), as with other areas of the country, the level of lease issuance has been much lower. About 16 per cent of the records were incomplete, and a significant number of owners were not identified at the time of demarcation (see below). Although surveying and demarcation has now been fully completed for the entire Kirehe district; according to the KWAMP staff, only 60 per cent of leases at district level have been issued (2012 data). 8 Marshlands, refers to private state land that is eligible for productive use. This land is typically under the management of MINAGRI. 9 Hillside refers mainly to customary, privately-owned land and is a bit of a misnomer as the land is typically located in valley bottoms and will most likely be irrigated through gravity-fed schemes. 12

The main reasons for the incomplete records are that either the owners are absent or they do not have identity documents. Absentee owners are: those that may have been present during the demarcation and adjudication process, but have since not been located, possibly because they are working elsewhere; those that may be out of the country, possibly a significant percentage being people who have not returned since the genocide; those that may be in jail this may include former perpetrators of the genocide. In cases where owners are not identified, the Government will hold the land for up to 30 years, after which it will take over ownership and make it available for redistribution. The non-identification of owners within these schemes has the potential to significantly impact on the utilisation of land, which has implications for KWAMP s investment, especially in the irrigation schemes on private hillside land. Initial data from the schemes suggest that there is: a significant difference in the amount of land owned by different owners; and, there could be a significant amount of land either which is owned by the government, or which could be without identified owners for some time. The first point implies that a few larger landowners could benefit significantly more from the irrigation scheme. It also raises questions as to whether these owners will have the sufficient resources, in particular labour, to utilise the irrigated land to its fullest potential. The second point implies that there could be some government land and possibly even some privately-owned land available for allocation to KWAMP s target groups. Some of the land under government ownership could be immediately available for reallocation to KWAMP s target groups. In the case of privately-owned land where the identification of owners may take some time, the civil law mandates that such land be held by the State for a 30 year period, after which it could take over ownership and redistribute to a different owner(s). This raises concerns as to whether the land will be utilised in the interim and who would use it. It is possible that in an interim period the land could be used by KWAMP s target groups; one option could be for land to be leased annually by the WUA. In cases where there are identified owners with larger amounts of land that they are not able to fully exploit, one option for them could to employ additional people to work the land or, perhaps preferably from KWAMP s perspective, to lease out parcels or portions of parcels to KWAMP s target groups on a long-term lease basis (for example 3 to 5 years). The equitable distribution of parcels within the schemes has also been a prime concern. KWAMP has compiled registers of relevant parcel holders from data obtained during the LTR process, which are then reviewed to ascertain whether there is an equitable allocation of parcels; a re-allocation of parcels from people with more than two parcels to upstream users affected by the dams and to the poorest households, in particular landless households, forms part of the options available to the project. 13

Some initial analysis of land holdings within the planned hillside schemes, done with the aid of GIS (see the Mapping section below for more), shows that some schemes have greater proportions of parcels above 1 ha than others (Table 2). This gives an indication of the potential for redistribution within the new schemes. Table 2 - Number and area of parcels > 1 ha Scheme # Parcels Total Area (ha) % of Scheme > 1ha Area Kinoni 1 19 34.42 30.82% Kinoni 2 26 33.30 18.23% Mahama 90 116.50 24.20% Nayamugali 53 78.33 35.33% However, KWAMP has identified that a more useful way of approaching the potential for redistribution is to look at the number of unique owners whose total holdings amount to areas of over 1 ha (Table 3), rather than looking at individual parcel sizes. This has required analysis on the basis of the ownership records. Table 3 - Number of owners with total land holdings > 1 ha Scheme # owners # parcels Total Area (ha) > 1ha Kinoni 1 23 50 36.01 Kinoni 2 38 100 61.27 Mahama 98 149 128.77 Nayamugali 60 96 119.99 KWAMP has also been able to analyse the relative concentrations of land ownership within the scheme areas; this is shown by a comparison of the Gini Coefficients applied to the land holdings of unique owners (where 0 represents an equal share of land and 1 represents all the land being owned by one person). This shows (Table 4) that land concentration is highest in Nayamugali and most evenly distributed in Mahama. Parcels with no registered owners have been removed from the equations. 10 Table 4 - Gini Coefficient applied to land holdings of individuals (total area of all parcels) within the schemes Scheme # unique # parcels Gini Coefficient owners (area in ha) Kinoni 1 216 305 0.562 Kinoni 2 341 528 0.542 Mahama 527 641 0.408 Nayamugali 308 384 0.580 The second part of the challenge facing KWAMP is to address some of the inequities in how the benefits and burdens are accruing to the future scheme users. Compensation for people losing land as a result of the implementation of 10 Note that this calculation is applied to the registered owners of the parcels rather than being based upon household or other shared ownership; thus some unique owners may in fact be married to each other or otherwise sharing the land areas. If this is the case, it would serve to increase the Gini Coefficient further towards 1, revealing a greater level of inequality. 14

the irrigation schemes has also been an inevitable part of the project design. In one scheme, affected parties have all been allocated parcels in the downstream scheme, but at others, although some affected parties already had access to parcels in the downstream scheme, it has not been possible to allocate parcels to others. One means of reducing the negative impact on affected parties has been to ensure that land owners affected by the reservoir buffer zones are supported in the cultivation of fodder grass and agro-forestry and benefit from their use. Financial compensation for crops grown on government-owned land that will be inundated by water due to the dam construction has been paid out in the irrigation schemes on government-owned land at Sagatare and Cynuzi. In the case of Sagatare, which is a new scheme, the affected upstream users have also been allocated land parcels in the downstream scheme, but at Cynuzi, which was already being cultivated, while certain affected upstream users seem already to have plots in the downstream scheme, others apparently do not. GROUP RIGHTS: Issues, Challenges, Tools and Strategies KWAMP is assisting with the formulation of comprehensive Watershed Management Plans (WMP) and the establishment of permanent public/private institutions, known as Local Committees for Management and Supervision (CLGS) to manage the development of each watershed, including the implementation of soil and water conservation activities. Members of the CLGS include representative from each WUA in the watershed, and representatives from other water user groups (e.g. livestock producers, borehole committee); village representatives concerned with water for domestic use and with the management of other natural resources; representatives of any other groups dealing with natural resources, for example groups managing tree nurseries; representatives of cooperative societies that operate in the watershed; and representatives of special interest groups, including women and youth. As a community institution the CLGS would be independent from government, although government staff, such as the In-charge of Agriculture and the Executive Secretary at Sector level could join as co-opted members, to offer technical advice or assist in problem resolution. The CLGS sends selected members for an inspection tour of the watershed on a regular basis, and meets afterwards to hear their report. These tours are used to assess the water availability in the (sub)catchment, the condition of roads especially those rehabilitated under the project, whether different water users adhere to the rules and regulations in the watershed management plan, and whether management systems (both physical infrastructure and management tools such as rotation schedules) are functioning properly. The condition of the watershed in terms of erosion and the implementation of soil and water conservation measures would also be assessed. In addition, the CLGS have ad-hoc meetings as necessary to address any conflict over water or the use of other natural resources, and in response to calamities such as flooding or landslides in the area. Finally, annual workshops are foreseen to review and amend the watershed management plan. These CLGS represent a form of comanagement institution. Community-led mapping exercises are being conducted to identify the extent and present land use in the watersheds, including an inventory of physical, economic and social attributes. 15

Some of the key issues and challenges that have been faced by the project in working with these groups include the following: The presidents of the CLGS also occupy an official function in the decentralized state structure (Executive Secretary of Sectors or Cells) which is helpful in ensuring the appropriation of KWAMP activities in the district performance plan and for their future sustainability. The experience of the CLGS to date is that they are functional and their members fully committed to serving their communities and sharing information with the more decentralized level (i.e. the villages). Membership and roles within these groups has been an issue in the past. The CLGS have had to be revised to include representatives of the main interest groups involved in the watersheds. In addition, service providers, which were previously included in CLGS as voting-members, and sometimes even as the secretary of the CLGS, have been moved to advisory roles and became non-voting members. A new group of CLGSs comprise from 9 to 13 members, all elected except the youth and women representatives who are part of an existing decentralized structure (IFAD, Sep 2011). MAPPING: Issues, Challenges, Tools and Strategies The KWAMP have already made considerable progress in respect to obtaining spatial data for use in both watershed planning and in assessing the implications of the hillside schemes. The use of GIS has assisted the CLGS and the project staff. In the watershed planning processes, printed maps (at scales of 1:10,000 to 1:20,000 depending on the watershed area) have been used in the participatory planning process, during which a range of important features were identified and spatially located by the members of the CLGS. These include existing erosion control measures, erosion risk areas, livestock grazing routes & access and social infrastructure. KWAMP consultants then undertake the process of digitising these features in a GIS programme, after which the ortho photo-based maps, with the additional imagery, are re-printed and provided to the relevant CLGS, with copies for the district authorities. So far this process has seen the updating of 7 initial watershed management plans (WMPs), and in the participatory preparation of 11 new WMPs; geo-referenced ortho photos in digital format have been obtained for all the watershed areas and have been used to produce A0 maps printed on paper. For the hillside schemes, GIS and spatial data were used to assess some of the implications of the schemes implementation. Data regarding demarcated parcel boundaries and ownership details was sourced from the RNRA. Additional spatial data, showing infrastructure plans for the schemes, was available from the contractors. The shape files for the irrigation command areas can therefore be overlaid on top of the shape files for all land parcels in the relevant areas. The overlay permits an identification of those parcels that fall either completely or partially within the boundaries of the command areas. The shape files are then analysed to calculate the areas of any new parcels (created where only a portion of a registered parcel falls within the command area boundary). The data was then exported to Excel files, providing a complete list of affected parcels for each scheme, with the data in respect to ownership, status, land shares, etc. Combined with high resolution imagery for the areas (also sourced from the RNRA), this kind of visualisation permits an easy way of identifying where there may 16

be issues that need to be addressed. It allows for assessments of the parcel size distribution within each scheme, the production of lists of land owners with total land holdings > 1 ha, and other data relevant to addressing some of the challenges noted above. A challenge when dealing with this data is the large amount of hard disk space required; the images also require a reasonable level of processing power in order to visualise and manipulate. The District s GIS and mapping capacity has been strengthened with the training of a GIS technical officer and basic maps of KWAMP s activities are being produced. The GIS technician responsible will receive additional training in GIS and map production using the ortho-photo shape files with support from the PCU and the DLM. (Sep 2011) WOMEN S ACCESS: Issues, Challenges, Tools and Strategies KWAMP is being implemented within the context of a number of laws that are designed to ensure women s participation, particularly in land-related matters. These include: The Organic Land Law; Law N 22/99 O to Supplement Book I of the Civil Code and to Institute Part Five Regarding Matrimonial Regimes, Liberalities and Successions. O.G. Nº 22 of 15/11/1999; Law N 59/2008 Of 10/09/2008 On Prevention And Punishment Of Gender- Based Violence and the Civil Code. The rights of married women and their children are well protected in law. The OLL prohibits any discrimination based on sex or origin in matters relating to ownership of rights over land; wife and husband have equal rights over land (Art. 4). The OLL further seems to protect the rights of other family members in the case of a land transfer, such as sale, donation or exchange, by a representative of the family. Article 34 conditions a land transfer with the written consent (Art. 37) of all other members of the family who are joint owners of these rights. These family members are defined as being legally married spouses, children who have attained majority age, minors represented by their guardians and the incompetent children represented by their tutors. Rights of family members of a legal marriage in case of inheritance (transfers upon death of one of the joint owners) are regulated by the 1999 Succession Law. This law provides reserved portions for children or surviving spouse that are high, at 80% of the patrimony. Inheritance division as stipulated in the law can be conditioned by a will, written or oral. The will constitutes an important tool for including the surviving partner of an unrecognised marriage union in the inheritance arrangements. In the absence of a will, the surviving partner of an unrecognised marriage does not get anything. If there are children from the union, the surviving partner has the duty to administer the property of any minor children and will maintain a right of usufruct. KNOWLEDGE MANAGEMENT The PCU of the KWAMP has a Knowledge Management action plan under implementation. The PCU has started documenting success stories on some of its interventions, which is being done with support from PROCASUR; this includes documenting aspects such as functionality, the potential for replication, lessons learned, challenges and recommendations. To reach a broad number of potential users, the PCU is translating 17